{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9999765939518772,
  "global_step": 42723,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1.6355140186915887e-06,
      "loss": 10.7373,
      "theoretical_loss": 20.81281780154715,
      "tokens_seen": 65536
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.2710280373831774e-06,
      "loss": 10.7003,
      "theoretical_loss": 17.566201104328645,
      "tokens_seen": 131072
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.906542056074766e-06,
      "loss": 10.5217,
      "theoretical_loss": 15.939477092836569,
      "tokens_seen": 196608
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.542056074766355e-06,
      "loss": 10.1915,
      "theoretical_loss": 14.89231675598857,
      "tokens_seen": 262144
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.177570093457943e-06,
      "loss": 10.0415,
      "theoretical_loss": 14.136216937762974,
      "tokens_seen": 327680
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.813084112149533e-06,
      "loss": 9.8739,
      "theoretical_loss": 13.552561472550224,
      "tokens_seen": 393216
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.144859813084112e-05,
      "loss": 9.5509,
      "theoretical_loss": 13.08180900140119,
      "tokens_seen": 458752
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.308411214953271e-05,
      "loss": 9.7099,
      "theoretical_loss": 12.690129625483323,
      "tokens_seen": 524288
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.4719626168224297e-05,
      "loss": 9.3894,
      "theoretical_loss": 12.356592463873625,
      "tokens_seen": 589824
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.6355140186915887e-05,
      "loss": 8.9112,
      "theoretical_loss": 12.067412607035077,
      "tokens_seen": 655360
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.7990654205607474e-05,
      "loss": 9.1517,
      "theoretical_loss": 11.813066231101676,
      "tokens_seen": 720896
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.9626168224299065e-05,
      "loss": 8.9402,
      "theoretical_loss": 11.586719208706729,
      "tokens_seen": 786432
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.1261682242990653e-05,
      "loss": 9.2393,
      "theoretical_loss": 11.383314140186787,
      "tokens_seen": 851968
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.289719626168224e-05,
      "loss": 9.1026,
      "theoretical_loss": 11.199011702111871,
      "tokens_seen": 917504
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.453271028037383e-05,
      "loss": 9.0532,
      "theoretical_loss": 11.030833917977912,
      "tokens_seen": 983040
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.616822429906542e-05,
      "loss": 8.8591,
      "theoretical_loss": 10.87642808645695,
      "tokens_seen": 1048576
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.7803738317757007e-05,
      "loss": 9.0862,
      "theoretical_loss": 10.733905740062724,
      "tokens_seen": 1114112
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.9439252336448595e-05,
      "loss": 8.7767,
      "theoretical_loss": 10.60172987623028,
      "tokens_seen": 1179648
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.107476635514018e-05,
      "loss": 8.9771,
      "theoretical_loss": 10.478634172356642,
      "tokens_seen": 1245184
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.2710280373831774e-05,
      "loss": 8.8099,
      "theoretical_loss": 10.36356394376333,
      "tokens_seen": 1310720
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.4345794392523365e-05,
      "loss": 8.9064,
      "theoretical_loss": 10.255632220896747,
      "tokens_seen": 1376256
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.598130841121495e-05,
      "loss": 8.4657,
      "theoretical_loss": 10.15408655327002,
      "tokens_seen": 1441792
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.761682242990654e-05,
      "loss": 8.5643,
      "theoretical_loss": 10.058283561732598,
      "tokens_seen": 1507328
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.925233644859813e-05,
      "loss": 8.9142,
      "theoretical_loss": 9.967669178840278,
      "tokens_seen": 1572864
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.088785046728972e-05,
      "loss": 8.2228,
      "theoretical_loss": 9.881763126393109,
      "tokens_seen": 1638400
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.2523364485981306e-05,
      "loss": 8.6888,
      "theoretical_loss": 9.80014659154056,
      "tokens_seen": 1703936
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.415887850467289e-05,
      "loss": 8.7567,
      "theoretical_loss": 9.722452346907446,
      "tokens_seen": 1769472
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.579439252336448e-05,
      "loss": 8.1503,
      "theoretical_loss": 9.648356759081546,
      "tokens_seen": 1835008
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.742990654205607e-05,
      "loss": 8.5098,
      "theoretical_loss": 9.577573271145639,
      "tokens_seen": 1900544
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.906542056074766e-05,
      "loss": 8.3572,
      "theoretical_loss": 9.509847046764852,
      "tokens_seen": 1966080
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.070093457943925e-05,
      "loss": 8.2511,
      "theoretical_loss": 9.444950537631936,
      "tokens_seen": 2031616
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.233644859813084e-05,
      "loss": 8.2922,
      "theoretical_loss": 9.382679790910457,
      "tokens_seen": 2097152
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.397196261682242e-05,
      "loss": 8.5234,
      "theoretical_loss": 9.32285135423398,
      "tokens_seen": 2162688
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.5607476635514014e-05,
      "loss": 8.053,
      "theoretical_loss": 9.265299666660276,
      "tokens_seen": 2228224
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.7242990654205605e-05,
      "loss": 8.3058,
      "theoretical_loss": 9.209874847444755,
      "tokens_seen": 2293760
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.887850467289719e-05,
      "loss": 8.0853,
      "theoretical_loss": 9.156440812508292,
      "tokens_seen": 2359296
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.051401869158878e-05,
      "loss": 7.9382,
      "theoretical_loss": 9.10487366241335,
      "tokens_seen": 2424832
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.214953271028036e-05,
      "loss": 8.3957,
      "theoretical_loss": 9.055060296533734,
      "tokens_seen": 2490368
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.378504672897196e-05,
      "loss": 7.8667,
      "theoretical_loss": 9.006897216643829,
      "tokens_seen": 2555904
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.542056074766355e-05,
      "loss": 8.0708,
      "theoretical_loss": 8.960289489909357,
      "tokens_seen": 2621440
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.705607476635514e-05,
      "loss": 7.9402,
      "theoretical_loss": 8.915149846640611,
      "tokens_seen": 2686976
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.869158878504673e-05,
      "loss": 7.9024,
      "theoretical_loss": 8.871397892478225,
      "tokens_seen": 2752512
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.032710280373832e-05,
      "loss": 7.788,
      "theoretical_loss": 8.828959418153499,
      "tokens_seen": 2818048
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.19626168224299e-05,
      "loss": 7.8857,
      "theoretical_loss": 8.787765792778412,
      "tokens_seen": 2883584
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.359813084112149e-05,
      "loss": 7.7723,
      "theoretical_loss": 8.747753428911455,
      "tokens_seen": 2949120
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.523364485981308e-05,
      "loss": 7.6216,
      "theoretical_loss": 8.708863309520833,
      "tokens_seen": 3014656
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.686915887850467e-05,
      "loss": 7.6382,
      "theoretical_loss": 8.671040568508847,
      "tokens_seen": 3080192
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.850467289719626e-05,
      "loss": 7.652,
      "theoretical_loss": 8.634234117735474,
      "tokens_seen": 3145728
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.014018691588785e-05,
      "loss": 7.8272,
      "theoretical_loss": 8.598396314536323,
      "tokens_seen": 3211264
    },
    {
      "epoch": 0.0,
      "objective/train/avg_token_score": 0.01812920905649662,
      "objective/train/avg_weight": 0.9787099957466125,
      "objective/train/docs_used": 13098,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 7.596808910369873,
      "objective/train/std_weight": 0.1134423315525055,
      "objective/train/theoretical_loss": 8.563482664611069,
      "objective/train/tokens_used": 23736800,
      "theoretical_loss": 8.563482664611069,
      "tokens_seen": 3276800
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.177570093457944e-05,
      "loss": 7.5096,
      "theoretical_loss": 8.563482664611069,
      "tokens_seen": 3276800
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.341121495327102e-05,
      "loss": 7.4326,
      "theoretical_loss": 8.529451555895115,
      "tokens_seen": 3342336
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.504672897196261e-05,
      "loss": 7.4954,
      "theoretical_loss": 8.496264019646002,
      "tokens_seen": 3407872
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.66822429906542e-05,
      "loss": 7.2682,
      "theoretical_loss": 8.463883515497187,
      "tokens_seen": 3473408
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.831775700934578e-05,
      "loss": 7.0752,
      "theoretical_loss": 8.432275737672779,
      "tokens_seen": 3538944
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.995327102803737e-05,
      "loss": 7.0729,
      "theoretical_loss": 8.401408439930716,
      "tokens_seen": 3604480
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.158878504672896e-05,
      "loss": 7.291,
      "theoretical_loss": 8.371251277120209,
      "tokens_seen": 3670016
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.322429906542055e-05,
      "loss": 7.3838,
      "theoretical_loss": 8.341775661511075,
      "tokens_seen": 3735552
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.485981308411215e-05,
      "loss": 7.356,
      "theoretical_loss": 8.31295463228533,
      "tokens_seen": 3801088
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.649532710280372e-05,
      "loss": 6.8428,
      "theoretical_loss": 8.284762736781182,
      "tokens_seen": 3866624
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.813084112149531e-05,
      "loss": 7.067,
      "theoretical_loss": 8.257175922251864,
      "tokens_seen": 3932160
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.97663551401869e-05,
      "loss": 6.8971,
      "theoretical_loss": 8.230171437050114,
      "tokens_seen": 3997696
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001014018691588785,
      "loss": 7.1514,
      "theoretical_loss": 8.20372774027797,
      "tokens_seen": 4063232
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010303738317757009,
      "loss": 6.7335,
      "theoretical_loss": 8.177824419053046,
      "tokens_seen": 4128768
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010467289719626168,
      "loss": 7.0069,
      "theoretical_loss": 8.152442112639616,
      "tokens_seen": 4194304
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010630841121495327,
      "loss": 6.9053,
      "theoretical_loss": 8.1275624427775,
      "tokens_seen": 4259840
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010794392523364485,
      "loss": 6.7783,
      "theoretical_loss": 8.10316794961571,
      "tokens_seen": 4325376
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010957943925233644,
      "loss": 6.9707,
      "theoretical_loss": 8.07924203272264,
      "tokens_seen": 4390912
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011121495327102803,
      "loss": 6.8147,
      "theoretical_loss": 8.055768896701416,
      "tokens_seen": 4456448
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011285046728971962,
      "loss": 6.8489,
      "theoretical_loss": 8.032733500989007,
      "tokens_seen": 4521984
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011448598130841121,
      "loss": 6.7427,
      "theoretical_loss": 8.010121513461836,
      "tokens_seen": 4587520
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001161214953271028,
      "loss": 6.6035,
      "theoretical_loss": 7.987919267509379,
      "tokens_seen": 4653056
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011775700934579438,
      "loss": 6.5195,
      "theoretical_loss": 7.966113722271801,
      "tokens_seen": 4718592
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011939252336448597,
      "loss": 6.4394,
      "theoretical_loss": 7.944692425767988,
      "tokens_seen": 4784128
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012102803738317756,
      "loss": 6.7287,
      "theoretical_loss": 7.9236434806675184,
      "tokens_seen": 4849664
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012266355140186914,
      "loss": 6.6607,
      "theoretical_loss": 7.902955512484067,
      "tokens_seen": 4915200
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012429906542056073,
      "loss": 6.6581,
      "theoretical_loss": 7.882617639989203,
      "tokens_seen": 4980736
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012593457943925232,
      "loss": 6.4161,
      "theoretical_loss": 7.862619447664628,
      "tokens_seen": 5046272
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001275700934579439,
      "loss": 6.3955,
      "theoretical_loss": 7.842950960027937,
      "tokens_seen": 5111808
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001292056074766355,
      "loss": 6.546,
      "theoretical_loss": 7.823602617682313,
      "tokens_seen": 5177344
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001308411214953271,
      "loss": 6.2445,
      "theoretical_loss": 7.804565254954165,
      "tokens_seen": 5242880
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013247663551401869,
      "loss": 6.6228,
      "theoretical_loss": 7.7858300789950725,
      "tokens_seen": 5308416
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013411214953271028,
      "loss": 6.3481,
      "theoretical_loss": 7.767388650235364,
      "tokens_seen": 5373952
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013574766355140187,
      "loss": 6.7136,
      "theoretical_loss": 7.749232864086619,
      "tokens_seen": 5439488
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013738317757009346,
      "loss": 6.0513,
      "theoretical_loss": 7.731354933799318,
      "tokens_seen": 5505024
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013901869158878505,
      "loss": 6.4743,
      "theoretical_loss": 7.71374737438992,
      "tokens_seen": 5570560
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014065420560747664,
      "loss": 6.3359,
      "theoretical_loss": 7.696402987558934,
      "tokens_seen": 5636096
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014228971962616823,
      "loss": 6.4461,
      "theoretical_loss": 7.679314847528181,
      "tokens_seen": 5701632
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001439252336448598,
      "loss": 6.0117,
      "theoretical_loss": 7.662476287731328,
      "tokens_seen": 5767168
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001455607476635514,
      "loss": 6.0534,
      "theoretical_loss": 7.645880888297279,
      "tokens_seen": 5832704
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014719626168224298,
      "loss": 6.3175,
      "theoretical_loss": 7.629522464270861,
      "tokens_seen": 5898240
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014883177570093457,
      "loss": 5.9341,
      "theoretical_loss": 7.613395054519696,
      "tokens_seen": 5963776
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015046728971962616,
      "loss": 6.5093,
      "theoretical_loss": 7.59749291128028,
      "tokens_seen": 6029312
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015210280373831775,
      "loss": 6.3375,
      "theoretical_loss": 7.581810490299888,
      "tokens_seen": 6094848
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015373831775700934,
      "loss": 5.9975,
      "theoretical_loss": 7.5663424415343705,
      "tokens_seen": 6160384
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015537383177570093,
      "loss": 6.448,
      "theoretical_loss": 7.551083600364949,
      "tokens_seen": 6225920
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015700934579439252,
      "loss": 6.1765,
      "theoretical_loss": 7.536028979299919,
      "tokens_seen": 6291456
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015864485981308411,
      "loss": 6.1477,
      "theoretical_loss": 7.521173760129762,
      "tokens_seen": 6356992
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001602803738317757,
      "loss": 6.011,
      "theoretical_loss": 7.506513286506497,
      "tokens_seen": 6422528
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001619158878504673,
      "loss": 6.399,
      "theoretical_loss": 7.492043056920249,
      "tokens_seen": 6488064
    },
    {
      "epoch": 0.0,
      "objective/train/avg_token_score": 0.04697641730308533,
      "objective/train/avg_weight": 0.9447228908538818,
      "objective/train/docs_used": 15224,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 6.395876884460449,
      "objective/train/std_weight": 0.2162683606147766,
      "objective/train/theoretical_loss": 7.4777587180480305,
      "objective/train/tokens_used": 27013600,
      "theoretical_loss": 7.4777587180480305,
      "tokens_seen": 6553600
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001635514018691589,
      "loss": 6.2428,
      "theoretical_loss": 7.4777587180480305,
      "tokens_seen": 6553600
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016518691588785045,
      "loss": 6.0588,
      "theoretical_loss": 7.463656058451462,
      "tokens_seen": 6619136
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016682242990654204,
      "loss": 5.9645,
      "theoretical_loss": 7.449731002601916,
      "tokens_seen": 6684672
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016845794392523363,
      "loss": 5.9992,
      "theoretical_loss": 7.435979605213019,
      "tokens_seen": 6750208
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017009345794392523,
      "loss": 6.1663,
      "theoretical_loss": 7.422398045861905,
      "tokens_seen": 6815744
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017172897196261682,
      "loss": 5.9906,
      "theoretical_loss": 7.408982623881875,
      "tokens_seen": 6881280
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001733644859813084,
      "loss": 5.8915,
      "theoretical_loss": 7.395729753510345,
      "tokens_seen": 6946816
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000175,
      "loss": 6.0481,
      "theoretical_loss": 7.3826359592770325,
      "tokens_seen": 7012352
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017663551401869156,
      "loss": 6.0684,
      "theoretical_loss": 7.369697871618373,
      "tokens_seen": 7077888
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017827102803738318,
      "loss": 6.0159,
      "theoretical_loss": 7.3569122227050885,
      "tokens_seen": 7143424
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017990654205607474,
      "loss": 6.0677,
      "theoretical_loss": 7.3442758424706875,
      "tokens_seen": 7208960
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018154205607476636,
      "loss": 6.0673,
      "theoretical_loss": 7.331785654829519,
      "tokens_seen": 7274496
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018317757009345793,
      "loss": 6.1199,
      "theoretical_loss": 7.319438674073677,
      "tokens_seen": 7340032
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018481308411214954,
      "loss": 5.8234,
      "theoretical_loss": 7.307232001438824,
      "tokens_seen": 7405568
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001864485981308411,
      "loss": 5.869,
      "theoretical_loss": 7.295162821829564,
      "tokens_seen": 7471104
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001880841121495327,
      "loss": 6.2231,
      "theoretical_loss": 7.283228400695652,
      "tokens_seen": 7536640
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001897196261682243,
      "loss": 6.0009,
      "theoretical_loss": 7.271426081050832,
      "tokens_seen": 7602176
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019135514018691588,
      "loss": 6.0854,
      "theoretical_loss": 7.259753280626623,
      "tokens_seen": 7667712
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019299065420560745,
      "loss": 5.9703,
      "theoretical_loss": 7.24820748915387,
      "tokens_seen": 7733248
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019462616822429906,
      "loss": 5.8284,
      "theoretical_loss": 7.236786265765262,
      "tokens_seen": 7798784
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019626168224299063,
      "loss": 5.8319,
      "theoretical_loss": 7.225487236512497,
      "tokens_seen": 7864320
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019789719626168225,
      "loss": 5.9159,
      "theoretical_loss": 7.21430809199212,
      "tokens_seen": 7929856
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001995327102803738,
      "loss": 6.3316,
      "theoretical_loss": 7.2032465850744005,
      "tokens_seen": 7995392
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020116822429906543,
      "loss": 5.9887,
      "theoretical_loss": 7.192300528730015,
      "tokens_seen": 8060928
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000202803738317757,
      "loss": 5.8817,
      "theoretical_loss": 7.1814677939495155,
      "tokens_seen": 8126464
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002044392523364486,
      "loss": 6.1222,
      "theoretical_loss": 7.1707463077509646,
      "tokens_seen": 8192000
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020607476635514017,
      "loss": 6.0845,
      "theoretical_loss": 7.160134051271272,
      "tokens_seen": 8257536
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020771028037383176,
      "loss": 5.9122,
      "theoretical_loss": 7.149629057937138,
      "tokens_seen": 8323072
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020934579439252336,
      "loss": 6.0076,
      "theoretical_loss": 7.139229411711638,
      "tokens_seen": 8388608
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021098130841121495,
      "loss": 5.6421,
      "theoretical_loss": 7.128933245412794,
      "tokens_seen": 8454144
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021261682242990654,
      "loss": 6.0391,
      "theoretical_loss": 7.118738739100616,
      "tokens_seen": 8519680
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021425233644859813,
      "loss": 5.8951,
      "theoretical_loss": 7.1086441185293445,
      "tokens_seen": 8585216
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002158878504672897,
      "loss": 5.94,
      "theoretical_loss": 7.09864765366177,
      "tokens_seen": 8650752
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002175233644859813,
      "loss": 5.8941,
      "theoretical_loss": 7.088747657242693,
      "tokens_seen": 8716288
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021915887850467288,
      "loss": 5.9848,
      "theoretical_loss": 7.078942483428749,
      "tokens_seen": 8781824
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002207943925233645,
      "loss": 6.0512,
      "theoretical_loss": 7.069230526471966,
      "tokens_seen": 8847360
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022242990654205606,
      "loss": 6.0352,
      "theoretical_loss": 7.059610219454568,
      "tokens_seen": 8912896
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022406542056074768,
      "loss": 5.8118,
      "theoretical_loss": 7.0500800330726685,
      "tokens_seen": 8978432
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022570093457943924,
      "loss": 6.0794,
      "theoretical_loss": 7.040638474466625,
      "tokens_seen": 9043968
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022733644859813086,
      "loss": 5.3759,
      "theoretical_loss": 7.031284086095933,
      "tokens_seen": 9109504
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022897196261682242,
      "loss": 6.0015,
      "theoretical_loss": 7.022015444656678,
      "tokens_seen": 9175040
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000230607476635514,
      "loss": 5.9016,
      "theoretical_loss": 7.012831160039609,
      "tokens_seen": 9240576
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002322429906542056,
      "loss": 5.8157,
      "theoretical_loss": 7.003729874327071,
      "tokens_seen": 9306112
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002338785046728972,
      "loss": 5.9676,
      "theoretical_loss": 6.994710260827057,
      "tokens_seen": 9371648
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023551401869158876,
      "loss": 5.7144,
      "theoretical_loss": 6.98577102314278,
      "tokens_seen": 9437184
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023714953271028038,
      "loss": 5.9299,
      "theoretical_loss": 6.976910894276189,
      "tokens_seen": 9502720
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023878504672897194,
      "loss": 5.7533,
      "theoretical_loss": 6.968128635764015,
      "tokens_seen": 9568256
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024042056074766356,
      "loss": 5.8261,
      "theoretical_loss": 6.959423036844894,
      "tokens_seen": 9633792
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024205607476635512,
      "loss": 5.7374,
      "theoretical_loss": 6.950792913656309,
      "tokens_seen": 9699328
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024369158878504674,
      "loss": 5.6275,
      "theoretical_loss": 6.942237108460029,
      "tokens_seen": 9764864
    },
    {
      "epoch": 0.0,
      "objective/train/avg_token_score": 0.021589580923318863,
      "objective/train/avg_weight": 0.9745916128158569,
      "objective/train/docs_used": 16449,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.943325996398926,
      "objective/train/std_weight": 0.09623614698648453,
      "objective/train/theoretical_loss": 6.9337544888949,
      "objective/train/tokens_used": 30290400,
      "theoretical_loss": 6.9337544888949,
      "tokens_seen": 9830400
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002453271028037383,
      "loss": 5.9083,
      "theoretical_loss": 6.9337544888949,
      "tokens_seen": 9830400
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002469626168224299,
      "loss": 5.6951,
      "theoretical_loss": 6.925343947255817,
      "tokens_seen": 9895936
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024859813084112146,
      "loss": 5.7893,
      "theoretical_loss": 6.917004399797798,
      "tokens_seen": 9961472
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002502336448598131,
      "loss": 5.4462,
      "theoretical_loss": 6.908734786064147,
      "tokens_seen": 10027008
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00025186915887850464,
      "loss": 5.7981,
      "theoretical_loss": 6.900534068237688,
      "tokens_seen": 10092544
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002535046728971963,
      "loss": 5.919,
      "theoretical_loss": 6.89240123051416,
      "tokens_seen": 10158080
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002551401869158878,
      "loss": 5.5707,
      "theoretical_loss": 6.884335278496871,
      "tokens_seen": 10223616
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00025677570093457947,
      "loss": 5.3321,
      "theoretical_loss": 6.87633523861175,
      "tokens_seen": 10289152
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000258411214953271,
      "loss": 5.6088,
      "theoretical_loss": 6.868400157541997,
      "tokens_seen": 10354688
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002600467289719626,
      "loss": 5.7315,
      "theoretical_loss": 6.860529101681551,
      "tokens_seen": 10420224
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002616822429906542,
      "loss": 5.4568,
      "theoretical_loss": 6.85272115660663,
      "tokens_seen": 10485760
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002633177570093458,
      "loss": 5.3406,
      "theoretical_loss": 6.844975426564642,
      "tokens_seen": 10551296
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00026495327102803737,
      "loss": 5.5645,
      "theoretical_loss": 6.8372910339797945,
      "tokens_seen": 10616832
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00026658878504672896,
      "loss": 5.3677,
      "theoretical_loss": 6.829667118974749,
      "tokens_seen": 10682368
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00026822429906542055,
      "loss": 5.4701,
      "theoretical_loss": 6.8221028389077185,
      "tokens_seen": 10747904
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00026985981308411214,
      "loss": 5.8299,
      "theoretical_loss": 6.814597367924395,
      "tokens_seen": 10813440
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00027149532710280373,
      "loss": 5.7607,
      "theoretical_loss": 6.807149896524181,
      "tokens_seen": 10878976
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002731308411214953,
      "loss": 5.9118,
      "theoretical_loss": 6.799759631140145,
      "tokens_seen": 10944512
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002747663551401869,
      "loss": 5.6398,
      "theoretical_loss": 6.7924257937322245,
      "tokens_seen": 11010048
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002764018691588785,
      "loss": 5.7278,
      "theoretical_loss": 6.785147621393148,
      "tokens_seen": 11075584
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002780373831775701,
      "loss": 5.731,
      "theoretical_loss": 6.777924365966638,
      "tokens_seen": 11141120
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002796728971962617,
      "loss": 5.2525,
      "theoretical_loss": 6.770755293677423,
      "tokens_seen": 11206656
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002813084112149533,
      "loss": 5.8404,
      "theoretical_loss": 6.763639684772625,
      "tokens_seen": 11272192
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00028294392523364487,
      "loss": 5.5347,
      "theoretical_loss": 6.756576833174123,
      "tokens_seen": 11337728
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00028457943925233646,
      "loss": 5.455,
      "theoretical_loss": 6.749566046141486,
      "tokens_seen": 11403264
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00028621495327102805,
      "loss": 5.7178,
      "theoretical_loss": 6.7426066439450905,
      "tokens_seen": 11468800
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002878504672897196,
      "loss": 5.5751,
      "theoretical_loss": 6.735697959549075,
      "tokens_seen": 11534336
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00028948598130841124,
      "loss": 5.6354,
      "theoretical_loss": 6.728839338303761,
      "tokens_seen": 11599872
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002911214953271028,
      "loss": 5.4645,
      "theoretical_loss": 6.722030137647226,
      "tokens_seen": 11665408
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002927570093457944,
      "loss": 5.5628,
      "theoretical_loss": 6.715269726815689,
      "tokens_seen": 11730944
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00029439252336448596,
      "loss": 5.6322,
      "theoretical_loss": 6.7085574865624125,
      "tokens_seen": 11796480
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002960280373831776,
      "loss": 5.7197,
      "theoretical_loss": 6.701892808884824,
      "tokens_seen": 11862016
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00029766355140186914,
      "loss": 5.4419,
      "theoretical_loss": 6.695275096759559,
      "tokens_seen": 11927552
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002992990654205608,
      "loss": 5.4932,
      "theoretical_loss": 6.68870376388518,
      "tokens_seen": 11993088
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003009345794392523,
      "loss": 5.4887,
      "theoretical_loss": 6.682178234432274,
      "tokens_seen": 12058624
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003025700934579439,
      "loss": 5.3856,
      "theoretical_loss": 6.675697942800715,
      "tokens_seen": 12124160
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003042056074766355,
      "loss": 5.7199,
      "theoretical_loss": 6.669262333383815,
      "tokens_seen": 12189696
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003058411214953271,
      "loss": 5.522,
      "theoretical_loss": 6.662870860339158,
      "tokens_seen": 12255232
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003074766355140187,
      "loss": 5.4893,
      "theoretical_loss": 6.656522987365879,
      "tokens_seen": 12320768
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003091121495327103,
      "loss": 5.0631,
      "theoretical_loss": 6.6502181874881705,
      "tokens_seen": 12386304
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00031074766355140187,
      "loss": 5.4963,
      "theoretical_loss": 6.643955942844831,
      "tokens_seen": 12451840
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00031238317757009346,
      "loss": 5.2365,
      "theoretical_loss": 6.637735744484626,
      "tokens_seen": 12517376
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00031401869158878505,
      "loss": 5.4391,
      "theoretical_loss": 6.631557092167304,
      "tokens_seen": 12582912
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00031565420560747664,
      "loss": 5.2507,
      "theoretical_loss": 6.625419494170049,
      "tokens_seen": 12648448
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00031728971962616823,
      "loss": 5.5124,
      "theoretical_loss": 6.619322467099223,
      "tokens_seen": 12713984
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003189252336448598,
      "loss": 5.8388,
      "theoretical_loss": 6.613265535707211,
      "tokens_seen": 12779520
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003205607476635514,
      "loss": 5.1539,
      "theoretical_loss": 6.607248232714213,
      "tokens_seen": 12845056
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000322196261682243,
      "loss": 5.4408,
      "theoretical_loss": 6.60127009863481,
      "tokens_seen": 12910592
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003238317757009346,
      "loss": 5.2484,
      "theoretical_loss": 6.59533068160918,
      "tokens_seen": 12976128
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003254672897196262,
      "loss": 5.3267,
      "theoretical_loss": 6.589429537238785,
      "tokens_seen": 13041664
    },
    {
      "epoch": 0.0,
      "objective/train/avg_token_score": 0.011510718613862991,
      "objective/train/avg_weight": 0.9864480495452881,
      "objective/train/docs_used": 18197,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.6189093589782715,
      "objective/train/std_weight": 0.06611253321170807,
      "objective/train/theoretical_loss": 6.583566228426414,
      "objective/train/tokens_used": 33567200,
      "theoretical_loss": 6.583566228426414,
      "tokens_seen": 13107200
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003271028037383178,
      "loss": 5.4239,
      "theoretical_loss": 6.583566228426414,
      "tokens_seen": 13107200
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00032873831775700937,
      "loss": 5.4964,
      "theoretical_loss": 6.5777403252204305,
      "tokens_seen": 13172736
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003303738317757009,
      "loss": 5.5019,
      "theoretical_loss": 6.571951404663098,
      "tokens_seen": 13238272
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00033200934579439255,
      "loss": 5.3374,
      "theoretical_loss": 6.566199050642863,
      "tokens_seen": 13303808
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003336448598130841,
      "loss": 5.2617,
      "theoretical_loss": 6.560482853750463,
      "tokens_seen": 13369344
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00033528037383177573,
      "loss": 5.3161,
      "theoretical_loss": 6.554802411138745,
      "tokens_seen": 13434880
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00033691588785046727,
      "loss": 5.3924,
      "theoretical_loss": 6.549157326386091,
      "tokens_seen": 13500416
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003385514018691589,
      "loss": 5.4244,
      "theoretical_loss": 6.54354720936333,
      "tokens_seen": 13565952
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00034018691588785045,
      "loss": 5.1808,
      "theoretical_loss": 6.537971676104026,
      "tokens_seen": 13631488
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003418224299065421,
      "loss": 5.2809,
      "theoretical_loss": 6.532430348678068,
      "tokens_seen": 13697024
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00034345794392523363,
      "loss": 5.5477,
      "theoretical_loss": 6.5269228550684195,
      "tokens_seen": 13762560
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003450934579439252,
      "loss": 5.3971,
      "theoretical_loss": 6.521448829050978,
      "tokens_seen": 13828096
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003467289719626168,
      "loss": 5.4314,
      "theoretical_loss": 6.516007910077416,
      "tokens_seen": 13893632
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003483644859813084,
      "loss": 5.406,
      "theoretical_loss": 6.51059974316095,
      "tokens_seen": 13959168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035,
      "loss": 5.3586,
      "theoretical_loss": 6.50522397876491,
      "tokens_seen": 14024704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003516355140186916,
      "loss": 5.6176,
      "theoretical_loss": 6.499880272694068,
      "tokens_seen": 14090240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003532710280373831,
      "loss": 5.3208,
      "theoretical_loss": 6.494568285988618,
      "tokens_seen": 14155776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035490654205607477,
      "loss": 5.4386,
      "theoretical_loss": 6.489287684820745,
      "tokens_seen": 14221312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035654205607476636,
      "loss": 5.3735,
      "theoretical_loss": 6.484038140393699,
      "tokens_seen": 14286848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035817757009345795,
      "loss": 5.5309,
      "theoretical_loss": 6.4788193288433105,
      "tokens_seen": 14352384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003598130841121495,
      "loss": 5.2901,
      "theoretical_loss": 6.473630931141869,
      "tokens_seen": 14417920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003614485981308411,
      "loss": 5.1968,
      "theoretical_loss": 6.468472633004308,
      "tokens_seen": 14483456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003630841121495327,
      "loss": 5.2695,
      "theoretical_loss": 6.463344124796616,
      "tokens_seen": 14548992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003647196261682243,
      "loss": 5.6746,
      "theoretical_loss": 6.45824510144643,
      "tokens_seen": 14614528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036635514018691585,
      "loss": 5.5118,
      "theoretical_loss": 6.45317526235573,
      "tokens_seen": 14680064
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036799065420560744,
      "loss": 5.365,
      "theoretical_loss": 6.448134311315593,
      "tokens_seen": 14745600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003696261682242991,
      "loss": 5.2057,
      "theoretical_loss": 6.443121956422939,
      "tokens_seen": 14811136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003712616822429907,
      "loss": 5.2961,
      "theoretical_loss": 6.438137909999214,
      "tokens_seen": 14876672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003728971962616822,
      "loss": 5.5135,
      "theoretical_loss": 6.433181888510964,
      "tokens_seen": 14942208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003745327102803738,
      "loss": 5.4459,
      "theoretical_loss": 6.428253612492239,
      "tokens_seen": 15007744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003761682242990654,
      "loss": 5.3314,
      "theoretical_loss": 6.4233528064687855,
      "tokens_seen": 15073280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037780373831775704,
      "loss": 5.4039,
      "theoretical_loss": 6.418479198883969,
      "tokens_seen": 15138816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003794392523364486,
      "loss": 5.4166,
      "theoretical_loss": 6.413632522026391,
      "tokens_seen": 15204352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038107476635514017,
      "loss": 5.1591,
      "theoretical_loss": 6.40881251195914,
      "tokens_seen": 15269888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038271028037383176,
      "loss": 5.1202,
      "theoretical_loss": 6.404018908450656,
      "tokens_seen": 15335424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003843457943925234,
      "loss": 5.527,
      "theoretical_loss": 6.399251454907132,
      "tokens_seen": 15400960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003859813084112149,
      "loss": 5.4969,
      "theoretical_loss": 6.394509898306452,
      "tokens_seen": 15466496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038761682242990654,
      "loss": 5.2881,
      "theoretical_loss": 6.389793989133574,
      "tokens_seen": 15532032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038925233644859813,
      "loss": 5.3295,
      "theoretical_loss": 6.385103481317387,
      "tokens_seen": 15597568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003908878504672897,
      "loss": 5.5285,
      "theoretical_loss": 6.380438132168923,
      "tokens_seen": 15663104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039252336448598126,
      "loss": 5.331,
      "theoretical_loss": 6.375797702320966,
      "tokens_seen": 15728640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003941588785046729,
      "loss": 5.0246,
      "theoretical_loss": 6.371181955668966,
      "tokens_seen": 15794176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003957943925233645,
      "loss": 5.1401,
      "theoretical_loss": 6.366590659313248,
      "tokens_seen": 15859712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003974299065420561,
      "loss": 5.2412,
      "theoretical_loss": 6.36202358350248,
      "tokens_seen": 15925248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003990654205607476,
      "loss": 5.414,
      "theoretical_loss": 6.357480501578371,
      "tokens_seen": 15990784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004007009345794392,
      "loss": 5.3819,
      "theoretical_loss": 6.352961189921553,
      "tokens_seen": 16056320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040233644859813086,
      "loss": 5.1708,
      "theoretical_loss": 6.348465427898629,
      "tokens_seen": 16121856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040397196261682245,
      "loss": 5.2611,
      "theoretical_loss": 6.343992997810366,
      "tokens_seen": 16187392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000405607476635514,
      "loss": 5.3726,
      "theoretical_loss": 6.33954368484097,
      "tokens_seen": 16252928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004072429906542056,
      "loss": 5.0317,
      "theoretical_loss": 6.33511727700846,
      "tokens_seen": 16318464
    },
    {
      "epoch": 0.01,
      "objective/train/avg_token_score": 0.019595211371779442,
      "objective/train/avg_weight": 0.9769594669342041,
      "objective/train/docs_used": 19928,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.429431438446045,
      "objective/train/std_weight": 0.09315425157546997,
      "objective/train/theoretical_loss": 6.330713565116083,
      "objective/train/tokens_used": 36844000,
      "theoretical_loss": 6.330713565116083,
      "tokens_seen": 16384000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004088785046728972,
      "loss": 5.4247,
      "theoretical_loss": 6.330713565116083,
      "tokens_seen": 16384000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004105140186915888,
      "loss": 5.4742,
      "theoretical_loss": 6.326332342704751,
      "tokens_seen": 16449536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00041214953271028035,
      "loss": 5.1381,
      "theoretical_loss": 6.32197340600647,
      "tokens_seen": 16515072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00041378504672897194,
      "loss": 5.0167,
      "theoretical_loss": 6.3176365538987636,
      "tokens_seen": 16580608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00041542056074766353,
      "loss": 5.3232,
      "theoretical_loss": 6.313321587860021,
      "tokens_seen": 16646144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004170560747663552,
      "loss": 5.1792,
      "theoretical_loss": 6.309028311925785,
      "tokens_seen": 16711680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004186915887850467,
      "loss": 5.4146,
      "theoretical_loss": 6.304756532645939,
      "tokens_seen": 16777216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004203271028037383,
      "loss": 5.2124,
      "theoretical_loss": 6.300506059042775,
      "tokens_seen": 16842752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004219626168224299,
      "loss": 5.3321,
      "theoretical_loss": 6.296276702569918,
      "tokens_seen": 16908288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042359813084112154,
      "loss": 5.26,
      "theoretical_loss": 6.292068277072099,
      "tokens_seen": 16973824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004252336448598131,
      "loss": 5.2664,
      "theoretical_loss": 6.28788059874573,
      "tokens_seen": 17039360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042686915887850467,
      "loss": 5.2215,
      "theoretical_loss": 6.283713486100297,
      "tokens_seen": 17104896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042850467289719626,
      "loss": 5.2296,
      "theoretical_loss": 6.279566759920507,
      "tokens_seen": 17170432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043014018691588785,
      "loss": 5.1681,
      "theoretical_loss": 6.275440243229228,
      "tokens_seen": 17235968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004317757009345794,
      "loss": 5.2843,
      "theoretical_loss": 6.271333761251142,
      "tokens_seen": 17301504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043341121495327103,
      "loss": 5.1078,
      "theoretical_loss": 6.267247141377137,
      "tokens_seen": 17367040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004350467289719626,
      "loss": 5.3769,
      "theoretical_loss": 6.2631802131294085,
      "tokens_seen": 17432576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004366822429906542,
      "loss": 5.2802,
      "theoretical_loss": 6.259132808127246,
      "tokens_seen": 17498112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043831775700934575,
      "loss": 5.1869,
      "theoretical_loss": 6.255104760053497,
      "tokens_seen": 17563648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004399532710280374,
      "loss": 5.1366,
      "theoretical_loss": 6.251095904621689,
      "tokens_seen": 17629184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000441588785046729,
      "loss": 5.1189,
      "theoretical_loss": 6.247106079543801,
      "tokens_seen": 17694720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004432242990654206,
      "loss": 5.2507,
      "theoretical_loss": 6.243135124498652,
      "tokens_seen": 17760256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004448598130841121,
      "loss": 5.2812,
      "theoretical_loss": 6.239182881100916,
      "tokens_seen": 17825792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004464953271028037,
      "loss": 5.5138,
      "theoretical_loss": 6.235249192870732,
      "tokens_seen": 17891328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044813084112149535,
      "loss": 4.9504,
      "theoretical_loss": 6.231333905203899,
      "tokens_seen": 17956864
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044976635514018694,
      "loss": 5.0152,
      "theoretical_loss": 6.227436865342643,
      "tokens_seen": 18022400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004514018691588785,
      "loss": 5.2435,
      "theoretical_loss": 6.223557922346955,
      "tokens_seen": 18087936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045303738317757007,
      "loss": 5.069,
      "theoretical_loss": 6.219696927066456,
      "tokens_seen": 18153472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004546728971962617,
      "loss": 5.2707,
      "theoretical_loss": 6.215853732112821,
      "tokens_seen": 18219008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004563084112149533,
      "loss": 5.011,
      "theoretical_loss": 6.212028191832702,
      "tokens_seen": 18284544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045794392523364484,
      "loss": 5.0889,
      "theoretical_loss": 6.208220162281178,
      "tokens_seen": 18350080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045957943925233643,
      "loss": 5.0665,
      "theoretical_loss": 6.204429501195701,
      "tokens_seen": 18415616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000461214953271028,
      "loss": 5.0969,
      "theoretical_loss": 6.20065606797053,
      "tokens_seen": 18481152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046285046728971967,
      "loss": 5.0774,
      "theoretical_loss": 6.19689972363164,
      "tokens_seen": 18546688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004644859813084112,
      "loss": 5.0749,
      "theoretical_loss": 6.1931603308120975,
      "tokens_seen": 18612224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004661214953271028,
      "loss": 5.231,
      "theoretical_loss": 6.189437753727901,
      "tokens_seen": 18677760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004677570093457944,
      "loss": 5.2637,
      "theoretical_loss": 6.185731858154261,
      "tokens_seen": 18743296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046939252336448603,
      "loss": 5.2598,
      "theoretical_loss": 6.182042511402313,
      "tokens_seen": 18808832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004710280373831775,
      "loss": 5.0632,
      "theoretical_loss": 6.17836958229627,
      "tokens_seen": 18874368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047266355140186916,
      "loss": 4.9935,
      "theoretical_loss": 6.1747129411509825,
      "tokens_seen": 18939904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047429906542056075,
      "loss": 5.0129,
      "theoretical_loss": 6.171072459749913,
      "tokens_seen": 19005440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047593457943925234,
      "loss": 5.2207,
      "theoretical_loss": 6.1674480113235095,
      "tokens_seen": 19070976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004775700934579439,
      "loss": 5.1481,
      "theoretical_loss": 6.163839470527964,
      "tokens_seen": 19136512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004792056074766355,
      "loss": 5.2321,
      "theoretical_loss": 6.160246713424372,
      "tokens_seen": 19202048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004808411214953271,
      "loss": 4.9259,
      "theoretical_loss": 6.156669617458243,
      "tokens_seen": 19267584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004824766355140187,
      "loss": 5.3024,
      "theoretical_loss": 6.153108061439397,
      "tokens_seen": 19333120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048411214953271025,
      "loss": 5.1281,
      "theoretical_loss": 6.149561925522211,
      "tokens_seen": 19398656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048574766355140184,
      "loss": 5.2467,
      "theoretical_loss": 6.146031091186222,
      "tokens_seen": 19464192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004873831775700935,
      "loss": 4.9875,
      "theoretical_loss": 6.142515441217064,
      "tokens_seen": 19529728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004890186915887851,
      "loss": 5.1153,
      "theoretical_loss": 6.1390148596877605,
      "tokens_seen": 19595264
    },
    {
      "epoch": 0.01,
      "objective/train/avg_token_score": 0.024301357567310333,
      "objective/train/avg_weight": 0.9714186191558838,
      "objective/train/docs_used": 21444,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.313518524169922,
      "objective/train/std_weight": 0.11250856518745422,
      "objective/train/theoretical_loss": 6.135529231940326,
      "objective/train/tokens_used": 40120800,
      "theoretical_loss": 6.135529231940326,
      "tokens_seen": 19660800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004906542056074766,
      "loss": 5.178,
      "theoretical_loss": 6.135529231940326,
      "tokens_seen": 19660800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004922897196261683,
      "loss": 5.258,
      "theoretical_loss": 6.132058444567705,
      "tokens_seen": 19726336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004939252336448598,
      "loss": 5.1252,
      "theoretical_loss": 6.128602385396022,
      "tokens_seen": 19791872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004955607476635514,
      "loss": 5.2444,
      "theoretical_loss": 6.125160943467138,
      "tokens_seen": 19857408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004971962616822429,
      "loss": 5.2126,
      "theoretical_loss": 6.121734009021521,
      "tokens_seen": 19922944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988317757009346,
      "loss": 5.2346,
      "theoretical_loss": 6.118321473481398,
      "tokens_seen": 19988480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005004672897196262,
      "loss": 4.9806,
      "theoretical_loss": 6.114923229434213,
      "tokens_seen": 20054016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005021028037383178,
      "loss": 5.148,
      "theoretical_loss": 6.111539170616359,
      "tokens_seen": 20119552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005037383177570093,
      "loss": 5.0447,
      "theoretical_loss": 6.108169191897195,
      "tokens_seen": 20185088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005053738317757009,
      "loss": 4.9784,
      "theoretical_loss": 6.104813189263336,
      "tokens_seen": 20250624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005070093457943926,
      "loss": 5.162,
      "theoretical_loss": 6.101471059803204,
      "tokens_seen": 20316160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005086448598130842,
      "loss": 4.9211,
      "theoretical_loss": 6.098142701691856,
      "tokens_seen": 20381696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005102803738317756,
      "loss": 5.1214,
      "theoretical_loss": 6.094828014176053,
      "tokens_seen": 20447232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005119158878504672,
      "loss": 5.108,
      "theoretical_loss": 6.091526897559593,
      "tokens_seen": 20512768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005135514018691589,
      "loss": 5.0679,
      "theoretical_loss": 6.088239253188885,
      "tokens_seen": 20578304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005151869158878505,
      "loss": 5.1838,
      "theoretical_loss": 6.084964983438763,
      "tokens_seen": 20643840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000516822429906542,
      "loss": 4.8267,
      "theoretical_loss": 6.0817039916985465,
      "tokens_seen": 20709376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005184579439252336,
      "loss": 5.1442,
      "theoretical_loss": 6.078456182358325,
      "tokens_seen": 20774912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005200934579439252,
      "loss": 5.108,
      "theoretical_loss": 6.075221460795472,
      "tokens_seen": 20840448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005217289719626169,
      "loss": 4.9474,
      "theoretical_loss": 6.071999733361386,
      "tokens_seen": 20905984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005233644859813084,
      "loss": 4.9546,
      "theoretical_loss": 6.068790907368448,
      "tokens_seen": 20971520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000525,
      "loss": 5.3031,
      "theoretical_loss": 6.0655948910771915,
      "tokens_seen": 21037056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005266355140186916,
      "loss": 5.1576,
      "theoretical_loss": 6.062411593683687,
      "tokens_seen": 21102592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005282710280373831,
      "loss": 5.0048,
      "theoretical_loss": 6.059240925307134,
      "tokens_seen": 21168128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005299065420560747,
      "loss": 5.0791,
      "theoretical_loss": 6.056082796977648,
      "tokens_seen": 21233664
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005315420560747663,
      "loss": 4.87,
      "theoretical_loss": 6.052937120624258,
      "tokens_seen": 21299200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005331775700934579,
      "loss": 4.9193,
      "theoretical_loss": 6.049803809063083,
      "tokens_seen": 21364736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005348130841121495,
      "loss": 4.8748,
      "theoretical_loss": 6.0466827759857145,
      "tokens_seen": 21430272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005364485981308411,
      "loss": 4.8228,
      "theoretical_loss": 6.04357393594778,
      "tokens_seen": 21495808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005380841121495327,
      "loss": 4.9232,
      "theoretical_loss": 6.040477204357686,
      "tokens_seen": 21561344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005397196261682243,
      "loss": 5.0226,
      "theoretical_loss": 6.037392497465552,
      "tokens_seen": 21626880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005413551401869159,
      "loss": 5.0393,
      "theoretical_loss": 6.034319732352309,
      "tokens_seen": 21692416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005429906542056075,
      "loss": 4.9405,
      "theoretical_loss": 6.031258826918979,
      "tokens_seen": 21757952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005446261682242991,
      "loss": 5.1435,
      "theoretical_loss": 6.0282096998761245,
      "tokens_seen": 21823488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005462616822429907,
      "loss": 4.8956,
      "theoretical_loss": 6.025172270733464,
      "tokens_seen": 21889024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005478971962616822,
      "loss": 5.1063,
      "theoretical_loss": 6.0221464597896475,
      "tokens_seen": 21954560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005495327102803738,
      "loss": 4.8378,
      "theoretical_loss": 6.0191321881221995,
      "tokens_seen": 22020096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005511682242990654,
      "loss": 5.0983,
      "theoretical_loss": 6.016129377577614,
      "tokens_seen": 22085632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000552803738317757,
      "loss": 4.6722,
      "theoretical_loss": 6.01313795076161,
      "tokens_seen": 22151168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005544392523364486,
      "loss": 5.0122,
      "theoretical_loss": 6.010157831029533,
      "tokens_seen": 22216704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005560747663551402,
      "loss": 4.6042,
      "theoretical_loss": 6.007188942476907,
      "tokens_seen": 22282240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005577102803738318,
      "loss": 5.085,
      "theoretical_loss": 6.0042312099301425,
      "tokens_seen": 22347776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005593457943925234,
      "loss": 4.7879,
      "theoretical_loss": 6.001284558937368,
      "tokens_seen": 22413312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005609813084112149,
      "loss": 5.1437,
      "theoretical_loss": 5.998348915759426,
      "tokens_seen": 22478848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005626168224299066,
      "loss": 4.8091,
      "theoretical_loss": 5.995424207360987,
      "tokens_seen": 22544384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005642523364485982,
      "loss": 5.0561,
      "theoretical_loss": 5.992510361401818,
      "tokens_seen": 22609920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005658878504672897,
      "loss": 5.1344,
      "theoretical_loss": 5.989607306228168,
      "tokens_seen": 22675456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005675233644859812,
      "loss": 5.282,
      "theoretical_loss": 5.986714970864292,
      "tokens_seen": 22740992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005691588785046729,
      "loss": 4.8961,
      "theoretical_loss": 5.983833285004112,
      "tokens_seen": 22806528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005707943925233645,
      "loss": 4.7289,
      "theoretical_loss": 5.980962179002983,
      "tokens_seen": 22872064
    },
    {
      "epoch": 0.01,
      "objective/train/avg_token_score": 0.03267163410782814,
      "objective/train/avg_weight": 0.961552619934082,
      "objective/train/docs_used": 23342,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.749721050262451,
      "objective/train/std_weight": 0.12489069998264313,
      "objective/train/theoretical_loss": 5.978101583869607,
      "objective/train/tokens_used": 43397600,
      "theoretical_loss": 5.978101583869607,
      "tokens_seen": 22937600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005724299065420561,
      "loss": 4.8266,
      "theoretical_loss": 5.978101583869607,
      "tokens_seen": 22937600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005740654205607476,
      "loss": 5.073,
      "theoretical_loss": 5.975251431258057,
      "tokens_seen": 23003136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005757009345794392,
      "loss": 4.9607,
      "theoretical_loss": 5.972411653459913,
      "tokens_seen": 23068672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005773364485981309,
      "loss": 4.954,
      "theoretical_loss": 5.9695821833965335,
      "tokens_seen": 23134208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005789719626168225,
      "loss": 4.9903,
      "theoretical_loss": 5.966762954611432,
      "tokens_seen": 23199744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000580607476635514,
      "loss": 4.918,
      "theoretical_loss": 5.963953901262764,
      "tokens_seen": 23265280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005822429906542055,
      "loss": 4.8353,
      "theoretical_loss": 5.961154958115937,
      "tokens_seen": 23330816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005838785046728972,
      "loss": 4.6741,
      "theoretical_loss": 5.958366060536315,
      "tokens_seen": 23396352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005855140186915888,
      "loss": 4.8514,
      "theoretical_loss": 5.955587144482044,
      "tokens_seen": 23461888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005871495327102803,
      "loss": 4.9463,
      "theoretical_loss": 5.952818146496978,
      "tokens_seen": 23527424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005887850467289719,
      "loss": 4.7774,
      "theoretical_loss": 5.950059003703704,
      "tokens_seen": 23592960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005904205607476635,
      "loss": 5.0155,
      "theoretical_loss": 5.94730965379668,
      "tokens_seen": 23658496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005920560747663552,
      "loss": 4.8992,
      "theoretical_loss": 5.944570035035458,
      "tokens_seen": 23724032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005936915887850467,
      "loss": 4.8901,
      "theoretical_loss": 5.941840086238027,
      "tokens_seen": 23789568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005953271028037383,
      "loss": 4.7879,
      "theoretical_loss": 5.939119746774228,
      "tokens_seen": 23855104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005969626168224299,
      "loss": 4.9298,
      "theoretical_loss": 5.936408956559284,
      "tokens_seen": 23920640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005985981308411216,
      "loss": 5.1531,
      "theoretical_loss": 5.933707656047414,
      "tokens_seen": 23986176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000600233644859813,
      "loss": 4.8991,
      "theoretical_loss": 5.93101578622554,
      "tokens_seen": 24051712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006018691588785046,
      "loss": 5.0655,
      "theoretical_loss": 5.928333288607086,
      "tokens_seen": 24117248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006035046728971962,
      "loss": 4.9448,
      "theoretical_loss": 5.925660105225867,
      "tokens_seen": 24182784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006051401869158878,
      "loss": 5.1153,
      "theoretical_loss": 5.92299617863006,
      "tokens_seen": 24248320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006067757009345794,
      "loss": 5.1179,
      "theoretical_loss": 5.920341451876267,
      "tokens_seen": 24313856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000608411214953271,
      "loss": 4.9149,
      "theoretical_loss": 5.9176958685236585,
      "tokens_seen": 24379392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006100467289719626,
      "loss": 5.0208,
      "theoretical_loss": 5.9150593726282015,
      "tokens_seen": 24444928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006116822429906542,
      "loss": 4.9617,
      "theoretical_loss": 5.912431908736972,
      "tokens_seen": 24510464
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006133177570093458,
      "loss": 5.06,
      "theoretical_loss": 5.909813421882534,
      "tokens_seen": 24576000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006149532710280374,
      "loss": 5.0619,
      "theoretical_loss": 5.907203857577422,
      "tokens_seen": 24641536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000616588785046729,
      "loss": 4.878,
      "theoretical_loss": 5.9046031618086765,
      "tokens_seen": 24707072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006182242990654205,
      "loss": 5.0219,
      "theoretical_loss": 5.902011281032472,
      "tokens_seen": 24772608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006198598130841121,
      "loss": 4.9154,
      "theoretical_loss": 5.899428162168808,
      "tokens_seen": 24838144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006214953271028037,
      "loss": 5.0189,
      "theoretical_loss": 5.896853752596286,
      "tokens_seen": 24903680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006231308411214953,
      "loss": 4.9928,
      "theoretical_loss": 5.894288000146949,
      "tokens_seen": 24969216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006247663551401869,
      "loss": 4.9262,
      "theoretical_loss": 5.891730853101199,
      "tokens_seen": 25034752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006264018691588785,
      "loss": 4.8464,
      "theoretical_loss": 5.88918226018278,
      "tokens_seen": 25100288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006280373831775701,
      "loss": 4.9285,
      "theoretical_loss": 5.8866421705538325,
      "tokens_seen": 25165824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006296728971962617,
      "loss": 4.9531,
      "theoretical_loss": 5.8841105338100155,
      "tokens_seen": 25231360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006313084112149533,
      "loss": 5.0287,
      "theoretical_loss": 5.881587299975694,
      "tokens_seen": 25296896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006329439252336449,
      "loss": 5.0161,
      "theoretical_loss": 5.8790724194991935,
      "tokens_seen": 25362432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006345794392523365,
      "loss": 4.9106,
      "theoretical_loss": 5.876565843248124,
      "tokens_seen": 25427968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000636214953271028,
      "loss": 4.9445,
      "theoretical_loss": 5.8740675225047525,
      "tokens_seen": 25493504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006378504672897196,
      "loss": 4.8723,
      "theoretical_loss": 5.871577408961457,
      "tokens_seen": 25559040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006394859813084112,
      "loss": 4.9034,
      "theoretical_loss": 5.869095454716231,
      "tokens_seen": 25624576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006411214953271028,
      "loss": 4.8989,
      "theoretical_loss": 5.866621612268246,
      "tokens_seen": 25690112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006427570093457944,
      "loss": 4.5686,
      "theoretical_loss": 5.864155834513486,
      "tokens_seen": 25755648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000644392523364486,
      "loss": 5.0447,
      "theoretical_loss": 5.8616980747404295,
      "tokens_seen": 25821184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006460280373831775,
      "loss": 4.5405,
      "theoretical_loss": 5.859248286625787,
      "tokens_seen": 25886720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006476635514018692,
      "loss": 4.7458,
      "theoretical_loss": 5.856806424230314,
      "tokens_seen": 25952256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006492990654205608,
      "loss": 4.7213,
      "theoretical_loss": 5.854372441994654,
      "tokens_seen": 26017792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006509345794392524,
      "loss": 5.0534,
      "theoretical_loss": 5.851946294735258,
      "tokens_seen": 26083328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006525700934579439,
      "loss": 5.0044,
      "theoretical_loss": 5.849527937640345,
      "tokens_seen": 26148864
    },
    {
      "epoch": 0.01,
      "objective/train/avg_token_score": 0.021748121827840805,
      "objective/train/avg_weight": 0.9743937253952026,
      "objective/train/docs_used": 25475,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.305008888244629,
      "objective/train/std_weight": 0.10435178875923157,
      "objective/train/theoretical_loss": 5.8471173262659235,
      "objective/train/tokens_used": 46674400,
      "theoretical_loss": 5.8471173262659235,
      "tokens_seen": 26214400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006542056074766356,
      "loss": 4.8294,
      "theoretical_loss": 5.8471173262659235,
      "tokens_seen": 26214400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006558411214953271,
      "loss": 4.7635,
      "theoretical_loss": 5.84471441653186,
      "tokens_seen": 26279936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006574766355140187,
      "loss": 4.9873,
      "theoretical_loss": 5.842319164718004,
      "tokens_seen": 26345472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006591121495327102,
      "loss": 4.9752,
      "theoretical_loss": 5.83993152746036,
      "tokens_seen": 26411008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006607476635514018,
      "loss": 4.7146,
      "theoretical_loss": 5.83755146174731,
      "tokens_seen": 26476544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006623831775700935,
      "loss": 4.5248,
      "theoretical_loss": 5.835178924915889,
      "tokens_seen": 26542080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006640186915887851,
      "loss": 4.7758,
      "theoretical_loss": 5.832813874648102,
      "tokens_seen": 26607616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006656542056074766,
      "loss": 4.64,
      "theoretical_loss": 5.8304562689673,
      "tokens_seen": 26673152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006672897196261682,
      "loss": 4.7524,
      "theoretical_loss": 5.828106066234588,
      "tokens_seen": 26738688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006689252336448599,
      "loss": 4.8509,
      "theoretical_loss": 5.825763225145295,
      "tokens_seen": 26804224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006705607476635515,
      "loss": 4.6327,
      "theoretical_loss": 5.823427704725473,
      "tokens_seen": 26869760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006721962616822429,
      "loss": 4.7313,
      "theoretical_loss": 5.82109946432846,
      "tokens_seen": 26935296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006738317757009345,
      "loss": 4.712,
      "theoretical_loss": 5.818778463631473,
      "tokens_seen": 27000832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006754672897196261,
      "loss": 4.9721,
      "theoretical_loss": 5.816464662632243,
      "tokens_seen": 27066368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006771028037383178,
      "loss": 4.8358,
      "theoretical_loss": 5.8141580216457065,
      "tokens_seen": 27131904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006787383177570093,
      "loss": 4.9087,
      "theoretical_loss": 5.811858501300729,
      "tokens_seen": 27197440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006803738317757009,
      "loss": 4.8568,
      "theoretical_loss": 5.809566062536868,
      "tokens_seen": 27262976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006820093457943925,
      "loss": 4.6859,
      "theoretical_loss": 5.807280666601191,
      "tokens_seen": 27328512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006836448598130842,
      "loss": 4.9171,
      "theoretical_loss": 5.805002275045111,
      "tokens_seen": 27394048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006852803738317757,
      "loss": 4.8244,
      "theoretical_loss": 5.8027308497212875,
      "tokens_seen": 27459584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006869158878504673,
      "loss": 4.9344,
      "theoretical_loss": 5.800466352780546,
      "tokens_seen": 27525120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006885514018691589,
      "loss": 4.7146,
      "theoretical_loss": 5.798208746668847,
      "tokens_seen": 27590656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006901869158878504,
      "loss": 4.7534,
      "theoretical_loss": 5.795957994124291,
      "tokens_seen": 27656192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000691822429906542,
      "loss": 4.6933,
      "theoretical_loss": 5.7937140581741575,
      "tokens_seen": 27721728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006934579439252336,
      "loss": 4.6152,
      "theoretical_loss": 5.791476902131985,
      "tokens_seen": 27787264
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006950934579439252,
      "loss": 4.7267,
      "theoretical_loss": 5.789246489594688,
      "tokens_seen": 27852800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006967289719626168,
      "loss": 5.0969,
      "theoretical_loss": 5.787022784439701,
      "tokens_seen": 27918336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006983644859813084,
      "loss": 4.6755,
      "theoretical_loss": 5.784805750822171,
      "tokens_seen": 27983872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0007,
      "loss": 4.7763,
      "theoretical_loss": 5.782595353172176,
      "tokens_seen": 28049408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006999834499716285,
      "loss": 4.77,
      "theoretical_loss": 5.780391556191977,
      "tokens_seen": 28114944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006999668999432571,
      "loss": 4.5739,
      "theoretical_loss": 5.778194324853311,
      "tokens_seen": 28180480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006999503499148856,
      "loss": 4.6671,
      "theoretical_loss": 5.776003624394711,
      "tokens_seen": 28246016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006999337998865141,
      "loss": 4.881,
      "theoretical_loss": 5.773819420318858,
      "tokens_seen": 28311552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006999172498581426,
      "loss": 4.8747,
      "theoretical_loss": 5.771641678389971,
      "tokens_seen": 28377088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006999006998297711,
      "loss": 5.0562,
      "theoretical_loss": 5.769470364631225,
      "tokens_seen": 28442624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006998841498013997,
      "loss": 4.7224,
      "theoretical_loss": 5.767305445322201,
      "tokens_seen": 28508160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006998675997730282,
      "loss": 4.7727,
      "theoretical_loss": 5.765146886996363,
      "tokens_seen": 28573696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006998510497446567,
      "loss": 4.7819,
      "theoretical_loss": 5.762994656438579,
      "tokens_seen": 28639232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006998344997162853,
      "loss": 4.5865,
      "theoretical_loss": 5.760848720682651,
      "tokens_seen": 28704768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006998179496879137,
      "loss": 4.8867,
      "theoretical_loss": 5.758709047008894,
      "tokens_seen": 28770304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006998013996595423,
      "loss": 4.7616,
      "theoretical_loss": 5.756575602941732,
      "tokens_seen": 28835840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006997848496311708,
      "loss": 4.2786,
      "theoretical_loss": 5.75444835624733,
      "tokens_seen": 28901376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006997682996027993,
      "loss": 4.7136,
      "theoretical_loss": 5.752327274931249,
      "tokens_seen": 28966912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006997517495744279,
      "loss": 4.8781,
      "theoretical_loss": 5.750212327236129,
      "tokens_seen": 29032448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006997351995460564,
      "loss": 4.8317,
      "theoretical_loss": 5.7481034816394105,
      "tokens_seen": 29097984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006997186495176849,
      "loss": 4.809,
      "theoretical_loss": 5.7460007068510635,
      "tokens_seen": 29163520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006997020994893135,
      "loss": 4.4846,
      "theoretical_loss": 5.74390397181136,
      "tokens_seen": 29229056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006996855494609419,
      "loss": 4.6994,
      "theoretical_loss": 5.741813245688668,
      "tokens_seen": 29294592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006996689994325705,
      "loss": 4.7301,
      "theoretical_loss": 5.739728497877267,
      "tokens_seen": 29360128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000699652449404199,
      "loss": 5.0214,
      "theoretical_loss": 5.737649697995197,
      "tokens_seen": 29425664
    },
    {
      "epoch": 0.01,
      "objective/train/avg_token_score": 0.04862821102142334,
      "objective/train/avg_weight": 0.9428002834320068,
      "objective/train/docs_used": 27675,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.549453258514404,
      "objective/train/std_weight": 0.142616868019104,
      "objective/train/theoretical_loss": 5.7355768158821245,
      "objective/train/tokens_used": 49951200,
      "theoretical_loss": 5.7355768158821245,
      "tokens_seen": 29491200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006996358993758275,
      "loss": 4.5547,
      "theoretical_loss": 5.7355768158821245,
      "tokens_seen": 29491200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006996193493474561,
      "loss": 4.8639,
      "theoretical_loss": 5.73350982159724,
      "tokens_seen": 29556736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006996027993190846,
      "loss": 4.6242,
      "theoretical_loss": 5.731448685417178,
      "tokens_seen": 29622272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000699586249290713,
      "loss": 4.8399,
      "theoretical_loss": 5.729393377833956,
      "tokens_seen": 29687808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006995696992623416,
      "loss": 4.7645,
      "theoretical_loss": 5.7273438695529535,
      "tokens_seen": 29753344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006995531492339701,
      "loss": 4.6744,
      "theoretical_loss": 5.725300131490888,
      "tokens_seen": 29818880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006995365992055987,
      "loss": 4.865,
      "theoretical_loss": 5.7232621347738455,
      "tokens_seen": 29884416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006995200491772272,
      "loss": 4.6753,
      "theoretical_loss": 5.721229850735305,
      "tokens_seen": 29949952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006995034991488557,
      "loss": 4.8826,
      "theoretical_loss": 5.719203250914208,
      "tokens_seen": 30015488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006994869491204842,
      "loss": 4.6157,
      "theoretical_loss": 5.717182307053037,
      "tokens_seen": 30081024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006994703990921127,
      "loss": 4.918,
      "theoretical_loss": 5.715166991095922,
      "tokens_seen": 30146560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006994538490637412,
      "loss": 4.4597,
      "theoretical_loss": 5.713157275186761,
      "tokens_seen": 30212096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006994372990353697,
      "loss": 4.6848,
      "theoretical_loss": 5.71115313166738,
      "tokens_seen": 30277632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006994207490069982,
      "loss": 4.9726,
      "theoretical_loss": 5.709154533075688,
      "tokens_seen": 30343168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006994041989786268,
      "loss": 4.9124,
      "theoretical_loss": 5.707161452143879,
      "tokens_seen": 30408704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006993876489502553,
      "loss": 4.6662,
      "theoretical_loss": 5.7051738617966326,
      "tokens_seen": 30474240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006993710989218839,
      "loss": 4.8423,
      "theoretical_loss": 5.7031917351493515,
      "tokens_seen": 30539776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006993545488935124,
      "loss": 4.3859,
      "theoretical_loss": 5.701215045506411,
      "tokens_seen": 30605312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006993379988651408,
      "loss": 5.0413,
      "theoretical_loss": 5.699243766359421,
      "tokens_seen": 30670848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006993214488367694,
      "loss": 4.8113,
      "theoretical_loss": 5.697277871385534,
      "tokens_seen": 30736384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006993048988083979,
      "loss": 4.6223,
      "theoretical_loss": 5.695317334445736,
      "tokens_seen": 30801920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006992883487800264,
      "loss": 4.6236,
      "theoretical_loss": 5.693362129583184,
      "tokens_seen": 30867456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000699271798751655,
      "loss": 4.6444,
      "theoretical_loss": 5.691412231021549,
      "tokens_seen": 30932992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006992552487232835,
      "loss": 4.6815,
      "theoretical_loss": 5.689467613163388,
      "tokens_seen": 30998528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006992386986949121,
      "loss": 4.6858,
      "theoretical_loss": 5.687528250588518,
      "tokens_seen": 31064064
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006992221486665406,
      "loss": 4.6606,
      "theoretical_loss": 5.6855941180524265,
      "tokens_seen": 31129600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000699205598638169,
      "loss": 4.7533,
      "theoretical_loss": 5.683665190484683,
      "tokens_seen": 31195136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006991890486097976,
      "loss": 4.9801,
      "theoretical_loss": 5.681741442987381,
      "tokens_seen": 31260672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006991724985814261,
      "loss": 4.9544,
      "theoretical_loss": 5.679822850833591,
      "tokens_seen": 31326208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006991559485530546,
      "loss": 4.5476,
      "theoretical_loss": 5.677909389465831,
      "tokens_seen": 31391744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006991393985246832,
      "loss": 4.7129,
      "theoretical_loss": 5.676001034494554,
      "tokens_seen": 31457280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006991228484963117,
      "loss": 4.7291,
      "theoretical_loss": 5.674097761696653,
      "tokens_seen": 31522816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006991062984679403,
      "loss": 4.3524,
      "theoretical_loss": 5.672199547013983,
      "tokens_seen": 31588352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006990897484395687,
      "loss": 4.5745,
      "theoretical_loss": 5.670306366551898,
      "tokens_seen": 31653888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006990731984111972,
      "loss": 4.7082,
      "theoretical_loss": 5.6684181965778,
      "tokens_seen": 31719424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006990566483828258,
      "loss": 4.6352,
      "theoretical_loss": 5.666535013519715,
      "tokens_seen": 31784960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006990400983544543,
      "loss": 4.8316,
      "theoretical_loss": 5.6646567939648715,
      "tokens_seen": 31850496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006990235483260828,
      "loss": 4.8965,
      "theoretical_loss": 5.6627835146583045,
      "tokens_seen": 31916032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006990069982977114,
      "loss": 4.5165,
      "theoretical_loss": 5.660915152501465,
      "tokens_seen": 31981568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006989904482693399,
      "loss": 4.5354,
      "theoretical_loss": 5.659051684550857,
      "tokens_seen": 32047104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006989738982409685,
      "loss": 4.4489,
      "theoretical_loss": 5.657193088016677,
      "tokens_seen": 32112640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006989573482125969,
      "loss": 5.0073,
      "theoretical_loss": 5.655339340261474,
      "tokens_seen": 32178176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006989407981842254,
      "loss": 4.7338,
      "theoretical_loss": 5.653490418798825,
      "tokens_seen": 32243712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000698924248155854,
      "loss": 4.766,
      "theoretical_loss": 5.651646301292022,
      "tokens_seen": 32309248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006989076981274825,
      "loss": 4.5283,
      "theoretical_loss": 5.649806965552774,
      "tokens_seen": 32374784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000698891148099111,
      "loss": 4.5273,
      "theoretical_loss": 5.6479723895399205,
      "tokens_seen": 32440320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006988745980707396,
      "loss": 4.6118,
      "theoretical_loss": 5.6461425513581665,
      "tokens_seen": 32505856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000698858048042368,
      "loss": 4.2993,
      "theoretical_loss": 5.6443174292568195,
      "tokens_seen": 32571392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006988414980139966,
      "loss": 4.7995,
      "theoretical_loss": 5.6424970016285485,
      "tokens_seen": 32636928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006988249479856251,
      "loss": 4.8587,
      "theoretical_loss": 5.640681247008156,
      "tokens_seen": 32702464
    },
    {
      "epoch": 0.01,
      "objective/train/avg_token_score": 0.0061433082446455956,
      "objective/train/avg_weight": 0.9927703142166138,
      "objective/train/docs_used": 29692,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.961858749389648,
      "objective/train/std_weight": 0.03961006924510002,
      "objective/train/theoretical_loss": 5.638870144071353,
      "objective/train/tokens_used": 53228000,
      "theoretical_loss": 5.638870144071353,
      "tokens_seen": 32768000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006988083979572536,
      "loss": 4.8442,
      "theoretical_loss": 5.638870144071353,
      "tokens_seen": 32768000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006987918479288822,
      "loss": 4.5169,
      "theoretical_loss": 5.637063671633564,
      "tokens_seen": 32833536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006987752979005107,
      "loss": 4.7815,
      "theoretical_loss": 5.635261808648728,
      "tokens_seen": 32899072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006987587478721392,
      "loss": 4.4051,
      "theoretical_loss": 5.6334645342081195,
      "tokens_seen": 32964608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006987421978437678,
      "loss": 4.6124,
      "theoretical_loss": 5.631671827539186,
      "tokens_seen": 33030144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006987256478153962,
      "loss": 4.6997,
      "theoretical_loss": 5.629883668004389,
      "tokens_seen": 33095680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006987090977870248,
      "loss": 4.6398,
      "theoretical_loss": 5.628100035100061,
      "tokens_seen": 33161216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006986925477586533,
      "loss": 4.5393,
      "theoretical_loss": 5.626320908455279,
      "tokens_seen": 33226752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006986759977302818,
      "loss": 4.4909,
      "theoretical_loss": 5.6245462678307385,
      "tokens_seen": 33292288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006986594477019104,
      "loss": 4.6716,
      "theoretical_loss": 5.622776093117652,
      "tokens_seen": 33357824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006986428976735389,
      "loss": 4.5331,
      "theoretical_loss": 5.621010364336651,
      "tokens_seen": 33423360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006986263476451674,
      "loss": 4.5143,
      "theoretical_loss": 5.619249061636698,
      "tokens_seen": 33488896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000698609797616796,
      "loss": 4.6772,
      "theoretical_loss": 5.61749216529402,
      "tokens_seen": 33554432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006985932475884244,
      "loss": 4.6454,
      "theoretical_loss": 5.615739655711037,
      "tokens_seen": 33619968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000698576697560053,
      "loss": 4.7753,
      "theoretical_loss": 5.61399151341532,
      "tokens_seen": 33685504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006985601475316815,
      "loss": 4.4263,
      "theoretical_loss": 5.6122477190585425,
      "tokens_seen": 33751040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00069854359750331,
      "loss": 4.6613,
      "theoretical_loss": 5.610508253415453,
      "tokens_seen": 33816576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006985270474749386,
      "loss": 4.6859,
      "theoretical_loss": 5.6087730973828585,
      "tokens_seen": 33882112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000698510497446567,
      "loss": 4.4016,
      "theoretical_loss": 5.6070422319786095,
      "tokens_seen": 33947648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006984939474181956,
      "loss": 4.5694,
      "theoretical_loss": 5.605315638340606,
      "tokens_seen": 34013184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000698477397389824,
      "loss": 4.7871,
      "theoretical_loss": 5.603593297725807,
      "tokens_seen": 34078720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006984608473614526,
      "loss": 4.6551,
      "theoretical_loss": 5.601875191509249,
      "tokens_seen": 34144256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006984442973330811,
      "loss": 4.5821,
      "theoretical_loss": 5.600161301183084,
      "tokens_seen": 34209792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006984277473047096,
      "loss": 4.6527,
      "theoretical_loss": 5.598451608355614,
      "tokens_seen": 34275328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006984111972763382,
      "loss": 4.4284,
      "theoretical_loss": 5.596746094750342,
      "tokens_seen": 34340864
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006983946472479667,
      "loss": 4.4011,
      "theoretical_loss": 5.595044742205037,
      "tokens_seen": 34406400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006983780972195952,
      "loss": 4.5266,
      "theoretical_loss": 5.5933475326707995,
      "tokens_seen": 34471936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006983615471912237,
      "loss": 4.3793,
      "theoretical_loss": 5.591654448211143,
      "tokens_seen": 34537472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006983449971628522,
      "loss": 4.6342,
      "theoretical_loss": 5.589965471001077,
      "tokens_seen": 34603008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006983284471344808,
      "loss": 4.6609,
      "theoretical_loss": 5.5882805833262115,
      "tokens_seen": 34668544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006983118971061093,
      "loss": 4.5533,
      "theoretical_loss": 5.586599767581859,
      "tokens_seen": 34734080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006982953470777378,
      "loss": 4.55,
      "theoretical_loss": 5.584923006272151,
      "tokens_seen": 34799616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006982787970493664,
      "loss": 4.7867,
      "theoretical_loss": 5.583250282009159,
      "tokens_seen": 34865152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006982622470209949,
      "loss": 4.6793,
      "theoretical_loss": 5.581581577512031,
      "tokens_seen": 34930688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006982456969926233,
      "loss": 4.3533,
      "theoretical_loss": 5.579916875606134,
      "tokens_seen": 34996224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006982291469642519,
      "loss": 4.4249,
      "theoretical_loss": 5.578256159222196,
      "tokens_seen": 35061760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006982125969358804,
      "loss": 4.9595,
      "theoretical_loss": 5.576599411395472,
      "tokens_seen": 35127296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000698196046907509,
      "loss": 4.4967,
      "theoretical_loss": 5.574946615264906,
      "tokens_seen": 35192832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006981794968791375,
      "loss": 4.7252,
      "theoretical_loss": 5.5732977540723105,
      "tokens_seen": 35258368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000698162946850766,
      "loss": 4.5479,
      "theoretical_loss": 5.571652811161542,
      "tokens_seen": 35323904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006981463968223946,
      "loss": 4.6196,
      "theoretical_loss": 5.570011769977693,
      "tokens_seen": 35389440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006981298467940231,
      "loss": 4.6719,
      "theoretical_loss": 5.568374614066299,
      "tokens_seen": 35454976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006981132967656515,
      "loss": 4.8718,
      "theoretical_loss": 5.566741327072535,
      "tokens_seen": 35520512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006980967467372801,
      "loss": 4.6118,
      "theoretical_loss": 5.565111892740433,
      "tokens_seen": 35586048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006980801967089086,
      "loss": 4.1763,
      "theoretical_loss": 5.563486294912105,
      "tokens_seen": 35651584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006980636466805372,
      "loss": 4.6928,
      "theoretical_loss": 5.56186451752697,
      "tokens_seen": 35717120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006980470966521657,
      "loss": 4.4074,
      "theoretical_loss": 5.560246544620993,
      "tokens_seen": 35782656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006980305466237942,
      "loss": 4.656,
      "theoretical_loss": 5.558632360325929,
      "tokens_seen": 35848192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006980139965954228,
      "loss": 4.6657,
      "theoretical_loss": 5.557021948868571,
      "tokens_seen": 35913728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006979974465670512,
      "loss": 4.3525,
      "theoretical_loss": 5.555415294570011,
      "tokens_seen": 35979264
    },
    {
      "epoch": 0.01,
      "objective/train/avg_token_score": 0.014716659672558308,
      "objective/train/avg_weight": 0.982679545879364,
      "objective/train/docs_used": 31846,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.995543479919434,
      "objective/train/std_weight": 0.09380307048559189,
      "objective/train/theoretical_loss": 5.553812381844907,
      "objective/train/tokens_used": 56504800,
      "theoretical_loss": 5.553812381844907,
      "tokens_seen": 36044800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006979808965386797,
      "loss": 4.5074,
      "theoretical_loss": 5.553812381844907,
      "tokens_seen": 36044800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006979643465103083,
      "loss": 4.4142,
      "theoretical_loss": 5.552213195200755,
      "tokens_seen": 36110336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006979477964819368,
      "loss": 4.5336,
      "theoretical_loss": 5.550617719237167,
      "tokens_seen": 36175872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006979312464535654,
      "loss": 4.4024,
      "theoretical_loss": 5.549025938645155,
      "tokens_seen": 36241408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006979146964251939,
      "loss": 4.4963,
      "theoretical_loss": 5.547437838206435,
      "tokens_seen": 36306944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006978981463968224,
      "loss": 4.5633,
      "theoretical_loss": 5.545853402792717,
      "tokens_seen": 36372480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000697881596368451,
      "loss": 4.7642,
      "theoretical_loss": 5.544272617365014,
      "tokens_seen": 36438016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006978650463400794,
      "loss": 4.4152,
      "theoretical_loss": 5.542695466972956,
      "tokens_seen": 36503552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006978484963117079,
      "loss": 4.497,
      "theoretical_loss": 5.541121936754111,
      "tokens_seen": 36569088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006978319462833365,
      "loss": 4.5549,
      "theoretical_loss": 5.539552011933312,
      "tokens_seen": 36634624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000697815396254965,
      "loss": 4.7109,
      "theoretical_loss": 5.537985677821986,
      "tokens_seen": 36700160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006977988462265936,
      "loss": 4.7176,
      "theoretical_loss": 5.536422919817495,
      "tokens_seen": 36765696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006977822961982221,
      "loss": 4.695,
      "theoretical_loss": 5.5348637234024824,
      "tokens_seen": 36831232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006977657461698506,
      "loss": 4.3335,
      "theoretical_loss": 5.53330807414422,
      "tokens_seen": 36896768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006977491961414791,
      "loss": 4.2846,
      "theoretical_loss": 5.5317559576939725,
      "tokens_seen": 36962304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006977326461131076,
      "loss": 4.3704,
      "theoretical_loss": 5.530207359786353,
      "tokens_seen": 37027840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006977160960847361,
      "loss": 4.7271,
      "theoretical_loss": 5.5286622662386975,
      "tokens_seen": 37093376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006976995460563647,
      "loss": 4.5593,
      "theoretical_loss": 5.52712066295044,
      "tokens_seen": 37158912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006976829960279932,
      "loss": 4.3478,
      "theoretical_loss": 5.525582535902489,
      "tokens_seen": 37224448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006976664459996218,
      "loss": 4.6174,
      "theoretical_loss": 5.524047871156618,
      "tokens_seen": 37289984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006976498959712503,
      "loss": 4.3986,
      "theoretical_loss": 5.52251665485486,
      "tokens_seen": 37355520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006976333459428787,
      "loss": 4.5249,
      "theoretical_loss": 5.520988873218897,
      "tokens_seen": 37421056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006976167959145073,
      "loss": 4.1813,
      "theoretical_loss": 5.519464512549478,
      "tokens_seen": 37486592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006976002458861358,
      "loss": 4.5434,
      "theoretical_loss": 5.5179435592258095,
      "tokens_seen": 37552128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006975836958577643,
      "loss": 4.4571,
      "theoretical_loss": 5.516425999704987,
      "tokens_seen": 37617664
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006975671458293929,
      "loss": 4.6872,
      "theoretical_loss": 5.514911820521407,
      "tokens_seen": 37683200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006975505958010214,
      "loss": 4.4437,
      "theoretical_loss": 5.5134010082861895,
      "tokens_seen": 37748736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006975340457726499,
      "loss": 4.2075,
      "theoretical_loss": 5.511893549686616,
      "tokens_seen": 37814272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006975174957442783,
      "loss": 4.5671,
      "theoretical_loss": 5.51038943148556,
      "tokens_seen": 37879808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006975009457159069,
      "loss": 4.4977,
      "theoretical_loss": 5.508888640520928,
      "tokens_seen": 37945344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006974843956875354,
      "loss": 4.5493,
      "theoretical_loss": 5.50739116370511,
      "tokens_seen": 38010880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006974678456591639,
      "loss": 4.3631,
      "theoretical_loss": 5.505896988024423,
      "tokens_seen": 38076416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006974512956307925,
      "loss": 4.6148,
      "theoretical_loss": 5.5044061005385725,
      "tokens_seen": 38141952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000697434745602421,
      "loss": 4.4654,
      "theoretical_loss": 5.502918488380116,
      "tokens_seen": 38207488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006974181955740496,
      "loss": 4.2282,
      "theoretical_loss": 5.501434138753918,
      "tokens_seen": 38273024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006974016455456781,
      "loss": 4.7714,
      "theoretical_loss": 5.499953038936635,
      "tokens_seen": 38338560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006973850955173065,
      "loss": 4.4798,
      "theoretical_loss": 5.498475176276176,
      "tokens_seen": 38404096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006973685454889351,
      "loss": 4.6407,
      "theoretical_loss": 5.497000538191195,
      "tokens_seen": 38469632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006973519954605636,
      "loss": 4.7017,
      "theoretical_loss": 5.495529112170568,
      "tokens_seen": 38535168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006973354454321921,
      "loss": 4.5794,
      "theoretical_loss": 5.494060885772887,
      "tokens_seen": 38600704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006973188954038207,
      "loss": 4.7515,
      "theoretical_loss": 5.492595846625951,
      "tokens_seen": 38666240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006973023453754492,
      "loss": 4.4655,
      "theoretical_loss": 5.491133982426266,
      "tokens_seen": 38731776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006972857953470778,
      "loss": 4.2478,
      "theoretical_loss": 5.489675280938547,
      "tokens_seen": 38797312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006972692453187062,
      "loss": 4.5672,
      "theoretical_loss": 5.488219729995227,
      "tokens_seen": 38862848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006972526952903347,
      "loss": 4.3841,
      "theoretical_loss": 5.486767317495966,
      "tokens_seen": 38928384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006972361452619633,
      "loss": 4.2779,
      "theoretical_loss": 5.48531803140717,
      "tokens_seen": 38993920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006972195952335918,
      "loss": 4.4864,
      "theoretical_loss": 5.483871859761511,
      "tokens_seen": 39059456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006972030452052203,
      "loss": 4.4961,
      "theoretical_loss": 5.482428790657449,
      "tokens_seen": 39124992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006971864951768489,
      "loss": 4.4443,
      "theoretical_loss": 5.480988812258763,
      "tokens_seen": 39190528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006971699451484774,
      "loss": 4.4742,
      "theoretical_loss": 5.479551912794086,
      "tokens_seen": 39256064
    },
    {
      "epoch": 0.01,
      "objective/train/avg_token_score": 0.014626851305365562,
      "objective/train/avg_weight": 0.9827902317047119,
      "objective/train/docs_used": 33055,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.184968948364258,
      "objective/train/std_weight": 0.08116675913333893,
      "objective/train/theoretical_loss": 5.478118080556438,
      "objective/train/tokens_used": 59781600,
      "theoretical_loss": 5.478118080556438,
      "tokens_seen": 39321600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000697153395120106,
      "loss": 4.3536,
      "theoretical_loss": 5.478118080556438,
      "tokens_seen": 39321600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006971368450917344,
      "loss": 4.4976,
      "theoretical_loss": 5.476687303902768,
      "tokens_seen": 39387136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006971202950633629,
      "loss": 4.3944,
      "theoretical_loss": 5.475259571253502,
      "tokens_seen": 39452672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006971037450349915,
      "loss": 4.7297,
      "theoretical_loss": 5.473834871092089,
      "tokens_seen": 39518208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00069708719500662,
      "loss": 4.2532,
      "theoretical_loss": 5.4724131919645576,
      "tokens_seen": 39583744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006970706449782485,
      "loss": 4.33,
      "theoretical_loss": 5.470994522479069,
      "tokens_seen": 39649280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006970540949498771,
      "loss": 4.1657,
      "theoretical_loss": 5.4695788513054815,
      "tokens_seen": 39714816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006970375449215056,
      "loss": 4.5991,
      "theoretical_loss": 5.468166167174912,
      "tokens_seen": 39780352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006970209948931341,
      "loss": 4.3302,
      "theoretical_loss": 5.466756458879306,
      "tokens_seen": 39845888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006970044448647626,
      "loss": 4.5392,
      "theoretical_loss": 5.465349715271013,
      "tokens_seen": 39911424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006969878948363911,
      "loss": 4.3954,
      "theoretical_loss": 5.463945925262355,
      "tokens_seen": 39976960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006969713448080197,
      "loss": 4.2732,
      "theoretical_loss": 5.462545077825214,
      "tokens_seen": 40042496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006969547947796482,
      "loss": 4.6529,
      "theoretical_loss": 5.461147161990611,
      "tokens_seen": 40108032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006969382447512767,
      "loss": 4.4683,
      "theoretical_loss": 5.459752166848292,
      "tokens_seen": 40173568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006969216947229053,
      "loss": 4.5611,
      "theoretical_loss": 5.458360081546321,
      "tokens_seen": 40239104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006969051446945337,
      "loss": 4.328,
      "theoretical_loss": 5.456970895290674,
      "tokens_seen": 40304640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006968885946661623,
      "loss": 4.2401,
      "theoretical_loss": 5.455584597344835,
      "tokens_seen": 40370176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006968720446377908,
      "loss": 4.1584,
      "theoretical_loss": 5.454201177029395,
      "tokens_seen": 40435712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006968554946094193,
      "loss": 4.0171,
      "theoretical_loss": 5.452820623721662,
      "tokens_seen": 40501248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006968389445810479,
      "loss": 4.5929,
      "theoretical_loss": 5.45144292685526,
      "tokens_seen": 40566784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006968223945526764,
      "loss": 4.4385,
      "theoretical_loss": 5.450068075919752,
      "tokens_seen": 40632320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006968058445243049,
      "loss": 4.4343,
      "theoretical_loss": 5.44869606046024,
      "tokens_seen": 40697856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006967892944959335,
      "loss": 4.2101,
      "theoretical_loss": 5.447326870076996,
      "tokens_seen": 40763392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006967727444675619,
      "loss": 4.4287,
      "theoretical_loss": 5.445960494425072,
      "tokens_seen": 40828928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006967561944391905,
      "loss": 4.4485,
      "theoretical_loss": 5.444596923213931,
      "tokens_seen": 40894464
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000696739644410819,
      "loss": 4.333,
      "theoretical_loss": 5.443236146207074,
      "tokens_seen": 40960000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006967230943824475,
      "loss": 4.5789,
      "theoretical_loss": 5.441878153221662,
      "tokens_seen": 41025536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006967065443540761,
      "loss": 4.2159,
      "theoretical_loss": 5.440522934128164,
      "tokens_seen": 41091072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006966899943257046,
      "loss": 4.6049,
      "theoretical_loss": 5.439170478849976,
      "tokens_seen": 41156608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006966734442973331,
      "loss": 4.5168,
      "theoretical_loss": 5.437820777363078,
      "tokens_seen": 41222144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006966568942689616,
      "loss": 4.3722,
      "theoretical_loss": 5.4364738196956655,
      "tokens_seen": 41287680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006966403442405901,
      "loss": 4.5328,
      "theoretical_loss": 5.435129595927794,
      "tokens_seen": 41353216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006966237942122187,
      "loss": 4.326,
      "theoretical_loss": 5.433788096191039,
      "tokens_seen": 41418752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006966072441838472,
      "loss": 4.2985,
      "theoretical_loss": 5.432449310668134,
      "tokens_seen": 41484288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006965906941554757,
      "loss": 4.6112,
      "theoretical_loss": 5.4311132295926345,
      "tokens_seen": 41549824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006965741441271043,
      "loss": 4.5714,
      "theoretical_loss": 5.42977984324857,
      "tokens_seen": 41615360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006965575940987327,
      "loss": 4.7828,
      "theoretical_loss": 5.428449141970107,
      "tokens_seen": 41680896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006965410440703612,
      "loss": 4.4037,
      "theoretical_loss": 5.427121116141212,
      "tokens_seen": 41746432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006965244940419897,
      "loss": 4.3501,
      "theoretical_loss": 5.42579575619531,
      "tokens_seen": 41811968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006965079440136182,
      "loss": 4.2668,
      "theoretical_loss": 5.424473052614967,
      "tokens_seen": 41877504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0006964913939852468,
      "loss": 4.5753,
      "theoretical_loss": 5.423152995931552,
      "tokens_seen": 41943040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006964748439568753,
      "loss": 4.6173,
      "theoretical_loss": 5.421835576724906,
      "tokens_seen": 42008576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006964582939285039,
      "loss": 4.5196,
      "theoretical_loss": 5.420520785623031,
      "tokens_seen": 42074112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006964417439001324,
      "loss": 4.2612,
      "theoretical_loss": 5.4192086133017625,
      "tokens_seen": 42139648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006964251938717608,
      "loss": 4.4192,
      "theoretical_loss": 5.417899050484451,
      "tokens_seen": 42205184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006964086438433894,
      "loss": 4.2164,
      "theoretical_loss": 5.416592087941646,
      "tokens_seen": 42270720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006963920938150179,
      "loss": 4.6111,
      "theoretical_loss": 5.415287716490787,
      "tokens_seen": 42336256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006963755437866464,
      "loss": 4.2953,
      "theoretical_loss": 5.413985926995892,
      "tokens_seen": 42401792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000696358993758275,
      "loss": 4.4642,
      "theoretical_loss": 5.412686710367245,
      "tokens_seen": 42467328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006963424437299035,
      "loss": 4.4528,
      "theoretical_loss": 5.411390057561097,
      "tokens_seen": 42532864
    },
    {
      "epoch": 0.02,
      "objective/train/avg_token_score": 0.012977970764040947,
      "objective/train/avg_weight": 0.984731137752533,
      "objective/train/docs_used": 34814,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.334832191467285,
      "objective/train/std_weight": 0.06617426127195358,
      "objective/train/theoretical_loss": 5.410095959579362,
      "objective/train/tokens_used": 63058400,
      "theoretical_loss": 5.410095959579362,
      "tokens_seen": 42598400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006963258937015321,
      "loss": 4.3725,
      "theoretical_loss": 5.410095959579362,
      "tokens_seen": 42598400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006963093436731606,
      "loss": 4.5467,
      "theoretical_loss": 5.408804407469308,
      "tokens_seen": 42663936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000696292793644789,
      "loss": 4.5378,
      "theoretical_loss": 5.407515392323276,
      "tokens_seen": 42729472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006962762436164176,
      "loss": 4.3338,
      "theoretical_loss": 5.406228905278368,
      "tokens_seen": 42795008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006962596935880461,
      "loss": 4.299,
      "theoretical_loss": 5.404944937516161,
      "tokens_seen": 42860544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006962431435596746,
      "loss": 4.7761,
      "theoretical_loss": 5.403663480262418,
      "tokens_seen": 42926080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006962265935313032,
      "loss": 4.457,
      "theoretical_loss": 5.402384524786797,
      "tokens_seen": 42991616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006962100435029317,
      "loss": 4.4925,
      "theoretical_loss": 5.401108062402562,
      "tokens_seen": 43057152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006961934934745603,
      "loss": 4.5406,
      "theoretical_loss": 5.399834084466306,
      "tokens_seen": 43122688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006961769434461888,
      "loss": 4.499,
      "theoretical_loss": 5.398562582377666,
      "tokens_seen": 43188224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006961603934178172,
      "loss": 4.3341,
      "theoretical_loss": 5.397293547579041,
      "tokens_seen": 43253760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006961438433894458,
      "loss": 4.6055,
      "theoretical_loss": 5.396026971555319,
      "tokens_seen": 43319296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006961272933610743,
      "loss": 4.6323,
      "theoretical_loss": 5.394762845833601,
      "tokens_seen": 43384832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006961107433327028,
      "loss": 4.2861,
      "theoretical_loss": 5.393501161982926,
      "tokens_seen": 43450368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006960941933043314,
      "loss": 4.3503,
      "theoretical_loss": 5.392241911614005,
      "tokens_seen": 43515904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006960776432759599,
      "loss": 4.2815,
      "theoretical_loss": 5.390985086378949,
      "tokens_seen": 43581440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006960610932475885,
      "loss": 4.3826,
      "theoretical_loss": 5.389730677971002,
      "tokens_seen": 43646976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006960445432192169,
      "loss": 4.3955,
      "theoretical_loss": 5.388478678124285,
      "tokens_seen": 43712512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006960279931908454,
      "loss": 4.3965,
      "theoretical_loss": 5.387229078613521,
      "tokens_seen": 43778048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000696011443162474,
      "loss": 4.4863,
      "theoretical_loss": 5.385981871253785,
      "tokens_seen": 43843584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006959948931341025,
      "loss": 4.3053,
      "theoretical_loss": 5.384737047900243,
      "tokens_seen": 43909120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006959783431057311,
      "loss": 4.1314,
      "theoretical_loss": 5.3834946004478965,
      "tokens_seen": 43974656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006959617930773596,
      "loss": 4.2661,
      "theoretical_loss": 5.382254520831328,
      "tokens_seen": 44040192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006959452430489881,
      "loss": 4.1895,
      "theoretical_loss": 5.381016801024449,
      "tokens_seen": 44105728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006959286930206167,
      "loss": 4.3734,
      "theoretical_loss": 5.379781433040252,
      "tokens_seen": 44171264
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006959121429922451,
      "loss": 4.5317,
      "theoretical_loss": 5.378548408930558,
      "tokens_seen": 44236800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006958955929638736,
      "loss": 4.4048,
      "theoretical_loss": 5.377317720785777,
      "tokens_seen": 44302336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006958790429355022,
      "loss": 4.3595,
      "theoretical_loss": 5.37608936073466,
      "tokens_seen": 44367872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006958624929071307,
      "loss": 4.3133,
      "theoretical_loss": 5.374863320944057,
      "tokens_seen": 44433408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006958459428787593,
      "loss": 4.5435,
      "theoretical_loss": 5.373639593618675,
      "tokens_seen": 44498944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006958293928503878,
      "loss": 4.4868,
      "theoretical_loss": 5.372418171000847,
      "tokens_seen": 44564480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006958128428220163,
      "loss": 4.3651,
      "theoretical_loss": 5.371199045370283,
      "tokens_seen": 44630016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006957962927936448,
      "loss": 4.2728,
      "theoretical_loss": 5.369982209043851,
      "tokens_seen": 44695552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006957797427652733,
      "loss": 4.277,
      "theoretical_loss": 5.368767654375327,
      "tokens_seen": 44761088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006957631927369018,
      "loss": 4.4265,
      "theoretical_loss": 5.367555373755179,
      "tokens_seen": 44826624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006957466427085304,
      "loss": 4.0888,
      "theoretical_loss": 5.366345359610327,
      "tokens_seen": 44892160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006957300926801589,
      "loss": 4.5792,
      "theoretical_loss": 5.365137604403923,
      "tokens_seen": 44957696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006957135426517875,
      "loss": 4.3524,
      "theoretical_loss": 5.363932100635117,
      "tokens_seen": 45023232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000695696992623416,
      "loss": 4.4885,
      "theoretical_loss": 5.362728840838843,
      "tokens_seen": 45088768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006956804425950444,
      "loss": 4.2669,
      "theoretical_loss": 5.361527817585586,
      "tokens_seen": 45154304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000695663892566673,
      "loss": 4.4064,
      "theoretical_loss": 5.360329023481169,
      "tokens_seen": 45219840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006956473425383015,
      "loss": 4.3629,
      "theoretical_loss": 5.359132451166534,
      "tokens_seen": 45285376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00069563079250993,
      "loss": 4.1924,
      "theoretical_loss": 5.357938093317518,
      "tokens_seen": 45350912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006956142424815586,
      "loss": 4.3838,
      "theoretical_loss": 5.356745942644645,
      "tokens_seen": 45416448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006955976924531871,
      "loss": 4.3628,
      "theoretical_loss": 5.355555991892905,
      "tokens_seen": 45481984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006955811424248157,
      "loss": 4.1631,
      "theoretical_loss": 5.35436823384155,
      "tokens_seen": 45547520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000695564592396444,
      "loss": 4.5134,
      "theoretical_loss": 5.353182661303873,
      "tokens_seen": 45613056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006955480423680726,
      "loss": 3.8195,
      "theoretical_loss": 5.35199926712701,
      "tokens_seen": 45678592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006955314923397011,
      "loss": 4.196,
      "theoretical_loss": 5.350818044191721,
      "tokens_seen": 45744128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006955149423113296,
      "loss": 4.5571,
      "theoretical_loss": 5.349638985412193,
      "tokens_seen": 45809664
    },
    {
      "epoch": 0.02,
      "objective/train/avg_token_score": 0.008400598540902138,
      "objective/train/avg_weight": 0.9901174902915955,
      "objective/train/docs_used": 36830,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.275974750518799,
      "objective/train/std_weight": 0.03861531987786293,
      "objective/train/theoretical_loss": 5.348462083735834,
      "objective/train/tokens_used": 66335200,
      "theoretical_loss": 5.348462083735834,
      "tokens_seen": 45875200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006954983922829582,
      "loss": 4.1835,
      "theoretical_loss": 5.348462083735834,
      "tokens_seen": 45875200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006954818422545867,
      "loss": 4.2501,
      "theoretical_loss": 5.347287332143064,
      "tokens_seen": 45940736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006954652922262152,
      "loss": 4.3239,
      "theoretical_loss": 5.346114723647119,
      "tokens_seen": 46006272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006954487421978438,
      "loss": 4.5505,
      "theoretical_loss": 5.344944251293852,
      "tokens_seen": 46071808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006954321921694722,
      "loss": 4.3584,
      "theoretical_loss": 5.343775908161532,
      "tokens_seen": 46137344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006954156421411008,
      "loss": 3.9772,
      "theoretical_loss": 5.342609687360644,
      "tokens_seen": 46202880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006953990921127293,
      "loss": 4.2646,
      "theoretical_loss": 5.341445582033705,
      "tokens_seen": 46268416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006953825420843578,
      "loss": 4.1213,
      "theoretical_loss": 5.3402835853550545,
      "tokens_seen": 46333952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006953659920559864,
      "loss": 4.1654,
      "theoretical_loss": 5.339123690530673,
      "tokens_seen": 46399488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006953494420276149,
      "loss": 4.2532,
      "theoretical_loss": 5.337965890797989,
      "tokens_seen": 46465024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006953328919992434,
      "loss": 4.1621,
      "theoretical_loss": 5.336810179425685,
      "tokens_seen": 46530560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006953163419708719,
      "loss": 4.4152,
      "theoretical_loss": 5.335656549713516,
      "tokens_seen": 46596096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006952997919425004,
      "loss": 4.1885,
      "theoretical_loss": 5.334504994992115,
      "tokens_seen": 46661632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000695283241914129,
      "loss": 4.1638,
      "theoretical_loss": 5.333355508622814,
      "tokens_seen": 46727168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006952666918857575,
      "loss": 4.2939,
      "theoretical_loss": 5.332208083997459,
      "tokens_seen": 46792704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000695250141857386,
      "loss": 4.0653,
      "theoretical_loss": 5.33106271453822,
      "tokens_seen": 46858240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006952335918290146,
      "loss": 4.1552,
      "theoretical_loss": 5.329919393697422,
      "tokens_seen": 46923776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006952170418006431,
      "loss": 4.4098,
      "theoretical_loss": 5.328778114957351,
      "tokens_seen": 46989312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006952004917722715,
      "loss": 4.034,
      "theoretical_loss": 5.327638871830089,
      "tokens_seen": 47054848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006951839417439001,
      "loss": 4.1252,
      "theoretical_loss": 5.326501657857326,
      "tokens_seen": 47120384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006951673917155286,
      "loss": 4.3107,
      "theoretical_loss": 5.32536646661019,
      "tokens_seen": 47185920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006951508416871572,
      "loss": 4.305,
      "theoretical_loss": 5.324233291689069,
      "tokens_seen": 47251456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006951342916587857,
      "loss": 4.2402,
      "theoretical_loss": 5.323102126723439,
      "tokens_seen": 47316992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006951177416304142,
      "loss": 4.2407,
      "theoretical_loss": 5.321972965371691,
      "tokens_seen": 47382528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006951011916020428,
      "loss": 4.3933,
      "theoretical_loss": 5.320845801320959,
      "tokens_seen": 47448064
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006950846415736713,
      "loss": 4.0686,
      "theoretical_loss": 5.319720628286955,
      "tokens_seen": 47513600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006950680915452997,
      "loss": 4.5263,
      "theoretical_loss": 5.318597440013795,
      "tokens_seen": 47579136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006950515415169283,
      "loss": 4.1664,
      "theoretical_loss": 5.317476230273831,
      "tokens_seen": 47644672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006950349914885568,
      "loss": 4.2979,
      "theoretical_loss": 5.316356992867491,
      "tokens_seen": 47710208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006950184414601854,
      "loss": 4.527,
      "theoretical_loss": 5.31523972162311,
      "tokens_seen": 47775744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006950018914318139,
      "loss": 4.3181,
      "theoretical_loss": 5.314124410396767,
      "tokens_seen": 47841280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006949853414034424,
      "loss": 4.5066,
      "theoretical_loss": 5.31301105307212,
      "tokens_seen": 47906816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000694968791375071,
      "loss": 4.1648,
      "theoretical_loss": 5.311899643560251,
      "tokens_seen": 47972352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006949522413466994,
      "loss": 3.9802,
      "theoretical_loss": 5.310790175799497,
      "tokens_seen": 48037888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006949356913183279,
      "loss": 4.1301,
      "theoretical_loss": 5.3096826437553,
      "tokens_seen": 48103424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006949191412899565,
      "loss": 4.1808,
      "theoretical_loss": 5.308577041420046,
      "tokens_seen": 48168960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000694902591261585,
      "loss": 4.0957,
      "theoretical_loss": 5.3074733628129005,
      "tokens_seen": 48234496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006948860412332136,
      "loss": 4.3032,
      "theoretical_loss": 5.3063716019796665,
      "tokens_seen": 48300032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006948694912048421,
      "loss": 4.3625,
      "theoretical_loss": 5.305271752992619,
      "tokens_seen": 48365568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006948529411764706,
      "loss": 4.2262,
      "theoretical_loss": 5.304173809950358,
      "tokens_seen": 48431104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006948363911480992,
      "loss": 4.2869,
      "theoretical_loss": 5.303077766977653,
      "tokens_seen": 48496640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006948198411197276,
      "loss": 4.1831,
      "theoretical_loss": 5.3019836182252895,
      "tokens_seen": 48562176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006948032910913561,
      "loss": 3.9292,
      "theoretical_loss": 5.300891357869929,
      "tokens_seen": 48627712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006947867410629847,
      "loss": 4.1565,
      "theoretical_loss": 5.299800980113945,
      "tokens_seen": 48693248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006947701910346132,
      "loss": 4.2108,
      "theoretical_loss": 5.298712479185288,
      "tokens_seen": 48758784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006947536410062418,
      "loss": 3.9904,
      "theoretical_loss": 5.297625849337331,
      "tokens_seen": 48824320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006947370909778703,
      "loss": 4.2971,
      "theoretical_loss": 5.296541084848727,
      "tokens_seen": 48889856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006947205409494988,
      "loss": 4.1573,
      "theoretical_loss": 5.295458180023262,
      "tokens_seen": 48955392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006947039909211273,
      "loss": 4.0508,
      "theoretical_loss": 5.294377129189715,
      "tokens_seen": 49020928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006946874408927558,
      "loss": 4.1256,
      "theoretical_loss": 5.293297926701706,
      "tokens_seen": 49086464
    },
    {
      "epoch": 0.02,
      "objective/train/avg_token_score": 0.008058790117502213,
      "objective/train/avg_weight": 0.9905455112457275,
      "objective/train/docs_used": 38907,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.9186224937438965,
      "objective/train/std_weight": 0.05724533647298813,
      "objective/train/theoretical_loss": 5.292220566937567,
      "objective/train/tokens_used": 69612000,
      "theoretical_loss": 5.292220566937567,
      "tokens_seen": 49152000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006946708908643843,
      "loss": 3.9716,
      "theoretical_loss": 5.292220566937567,
      "tokens_seen": 49152000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006946543408360129,
      "loss": 4.493,
      "theoretical_loss": 5.29114504430019,
      "tokens_seen": 49217536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006946377908076414,
      "loss": 3.9271,
      "theoretical_loss": 5.290071353216895,
      "tokens_seen": 49283072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00069462124077927,
      "loss": 4.1743,
      "theoretical_loss": 5.288999488139284,
      "tokens_seen": 49348608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006946046907508985,
      "loss": 4.3637,
      "theoretical_loss": 5.28792944354311,
      "tokens_seen": 49414144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006945881407225269,
      "loss": 4.1597,
      "theoretical_loss": 5.286861213928137,
      "tokens_seen": 49479680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006945715906941554,
      "loss": 4.0044,
      "theoretical_loss": 5.285794793817999,
      "tokens_seen": 49545216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006945550406657839,
      "loss": 4.1636,
      "theoretical_loss": 5.284730177760077,
      "tokens_seen": 49610752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006945384906374125,
      "loss": 4.4841,
      "theoretical_loss": 5.283667360325351,
      "tokens_seen": 49676288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000694521940609041,
      "loss": 4.2939,
      "theoretical_loss": 5.2826063361082785,
      "tokens_seen": 49741824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006945053905806696,
      "loss": 4.3237,
      "theoretical_loss": 5.281547099726654,
      "tokens_seen": 49807360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006944888405522981,
      "loss": 4.2169,
      "theoretical_loss": 5.280489645821483,
      "tokens_seen": 49872896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006944722905239265,
      "loss": 4.2478,
      "theoretical_loss": 5.279433969056848,
      "tokens_seen": 49938432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006944557404955551,
      "loss": 4.407,
      "theoretical_loss": 5.278380064119782,
      "tokens_seen": 50003968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006944391904671836,
      "loss": 4.1532,
      "theoretical_loss": 5.277327925720137,
      "tokens_seen": 50069504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006944226404388121,
      "loss": 4.0955,
      "theoretical_loss": 5.276277548590457,
      "tokens_seen": 50135040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006944060904104407,
      "loss": 4.2986,
      "theoretical_loss": 5.275228927485855,
      "tokens_seen": 50200576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006943895403820692,
      "loss": 4.2384,
      "theoretical_loss": 5.2741820571838804,
      "tokens_seen": 50266112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006943729903536978,
      "loss": 4.0566,
      "theoretical_loss": 5.273136932484399,
      "tokens_seen": 50331648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006943564403253263,
      "loss": 4.0992,
      "theoretical_loss": 5.272093548209467,
      "tokens_seen": 50397184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006943398902969547,
      "loss": 4.1551,
      "theoretical_loss": 5.271051899203207,
      "tokens_seen": 50462720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006943233402685833,
      "loss": 3.9363,
      "theoretical_loss": 5.270011980331685,
      "tokens_seen": 50528256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006943067902402118,
      "loss": 4.443,
      "theoretical_loss": 5.268973786482794,
      "tokens_seen": 50593792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006942902402118403,
      "loss": 3.9722,
      "theoretical_loss": 5.267937312566123,
      "tokens_seen": 50659328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006942736901834689,
      "loss": 4.0481,
      "theoretical_loss": 5.266902553512847,
      "tokens_seen": 50724864
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006942571401550974,
      "loss": 4.4073,
      "theoretical_loss": 5.265869504275602,
      "tokens_seen": 50790400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000694240590126726,
      "loss": 4.3928,
      "theoretical_loss": 5.264838159828369,
      "tokens_seen": 50855936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006942240400983544,
      "loss": 4.15,
      "theoretical_loss": 5.263808515166355,
      "tokens_seen": 50921472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006942074900699829,
      "loss": 4.1729,
      "theoretical_loss": 5.262780565305875,
      "tokens_seen": 50987008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006941909400416115,
      "loss": 4.1796,
      "theoretical_loss": 5.261754305284241,
      "tokens_seen": 51052544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00069417439001324,
      "loss": 4.2629,
      "theoretical_loss": 5.260729730159641,
      "tokens_seen": 51118080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006941578399848685,
      "loss": 4.264,
      "theoretical_loss": 5.259706835011027,
      "tokens_seen": 51183616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006941412899564971,
      "loss": 4.0966,
      "theoretical_loss": 5.2586856149380035,
      "tokens_seen": 51249152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006941247399281256,
      "loss": 4.2137,
      "theoretical_loss": 5.257666065060709,
      "tokens_seen": 51314688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006941081898997542,
      "loss": 3.9817,
      "theoretical_loss": 5.256648180519708,
      "tokens_seen": 51380224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006940916398713826,
      "loss": 4.3113,
      "theoretical_loss": 5.255631956475881,
      "tokens_seen": 51445760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006940750898430111,
      "loss": 4.3104,
      "theoretical_loss": 5.25461738811031,
      "tokens_seen": 51511296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006940585398146397,
      "loss": 4.161,
      "theoretical_loss": 5.25360447062417,
      "tokens_seen": 51576832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006940419897862682,
      "loss": 4.1749,
      "theoretical_loss": 5.252593199238619,
      "tokens_seen": 51642368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006940254397578967,
      "loss": 3.9366,
      "theoretical_loss": 5.2515835691946915,
      "tokens_seen": 51707904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006940088897295253,
      "loss": 4.3712,
      "theoretical_loss": 5.2505755757531904,
      "tokens_seen": 51773440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006939923397011538,
      "loss": 3.9246,
      "theoretical_loss": 5.24956921419458,
      "tokens_seen": 51838976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006939757896727823,
      "loss": 4.0444,
      "theoretical_loss": 5.248564479818876,
      "tokens_seen": 51904512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006939592396444108,
      "loss": 4.1468,
      "theoretical_loss": 5.247561367945544,
      "tokens_seen": 51970048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006939426896160393,
      "loss": 4.3775,
      "theoretical_loss": 5.246559873913396,
      "tokens_seen": 52035584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006939261395876679,
      "loss": 4.0843,
      "theoretical_loss": 5.245559993080484,
      "tokens_seen": 52101120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006939095895592964,
      "loss": 4.102,
      "theoretical_loss": 5.24456172082399,
      "tokens_seen": 52166656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006938930395309249,
      "loss": 4.1781,
      "theoretical_loss": 5.243565052540136,
      "tokens_seen": 52232192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006938764895025535,
      "loss": 4.2092,
      "theoretical_loss": 5.242569983644074,
      "tokens_seen": 52297728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000693859939474182,
      "loss": 4.2355,
      "theoretical_loss": 5.241576509569784,
      "tokens_seen": 52363264
    },
    {
      "epoch": 0.02,
      "objective/train/avg_token_score": 0.03005906008183956,
      "objective/train/avg_weight": 0.9646203517913818,
      "objective/train/docs_used": 40119,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.443134307861328,
      "objective/train/std_weight": 0.12583357095718384,
      "objective/train/theoretical_loss": 5.240584625769978,
      "objective/train/tokens_used": 72888800,
      "theoretical_loss": 5.240584625769978,
      "tokens_seen": 52428800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006938433894458105,
      "loss": 4.2754,
      "theoretical_loss": 5.240584625769978,
      "tokens_seen": 52428800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000693826839417439,
      "loss": 4.0679,
      "theoretical_loss": 5.239594327715992,
      "tokens_seen": 52494336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006938102893890675,
      "loss": 4.2109,
      "theoretical_loss": 5.238605610897698,
      "tokens_seen": 52559872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006937937393606961,
      "loss": 4.3482,
      "theoretical_loss": 5.237618470823394,
      "tokens_seen": 52625408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006937771893323246,
      "loss": 4.1702,
      "theoretical_loss": 5.2366329030197125,
      "tokens_seen": 52690944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006937606393039531,
      "loss": 4.1957,
      "theoretical_loss": 5.235648903031521,
      "tokens_seen": 52756480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006937440892755817,
      "loss": 4.2322,
      "theoretical_loss": 5.2346664664218245,
      "tokens_seen": 52822016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006937275392472101,
      "loss": 4.097,
      "theoretical_loss": 5.233685588771669,
      "tokens_seen": 52887552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006937109892188387,
      "loss": 4.3008,
      "theoretical_loss": 5.232706265680049,
      "tokens_seen": 52953088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006936944391904672,
      "loss": 4.323,
      "theoretical_loss": 5.231728492763811,
      "tokens_seen": 53018624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006936778891620957,
      "loss": 4.3315,
      "theoretical_loss": 5.230752265657554,
      "tokens_seen": 53084160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006936613391337243,
      "loss": 3.9943,
      "theoretical_loss": 5.229777580013545,
      "tokens_seen": 53149696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006936447891053528,
      "loss": 4.1854,
      "theoretical_loss": 5.228804431501619,
      "tokens_seen": 53215232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006936282390769813,
      "loss": 4.1715,
      "theoretical_loss": 5.227832815809087,
      "tokens_seen": 53280768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006936116890486098,
      "loss": 3.7258,
      "theoretical_loss": 5.226862728640651,
      "tokens_seen": 53346304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006935951390202382,
      "loss": 4.0137,
      "theoretical_loss": 5.2258941657183,
      "tokens_seen": 53411840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006935785889918668,
      "loss": 4.3328,
      "theoretical_loss": 5.2249271227812315,
      "tokens_seen": 53477376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006935620389634953,
      "loss": 3.9421,
      "theoretical_loss": 5.223961595585755,
      "tokens_seen": 53542912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006935454889351239,
      "loss": 4.0251,
      "theoretical_loss": 5.222997579905204,
      "tokens_seen": 53608448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006935289389067524,
      "loss": 4.0949,
      "theoretical_loss": 5.222035071529845,
      "tokens_seen": 53673984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006935123888783809,
      "loss": 4.2455,
      "theoretical_loss": 5.2210740662667945,
      "tokens_seen": 53739520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006934958388500094,
      "loss": 4.2954,
      "theoretical_loss": 5.220114559939923,
      "tokens_seen": 53805056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006934792888216379,
      "loss": 4.055,
      "theoretical_loss": 5.219156548389775,
      "tokens_seen": 53870592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006934627387932665,
      "loss": 3.7133,
      "theoretical_loss": 5.218200027473481,
      "tokens_seen": 53936128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000693446188764895,
      "loss": 4.1167,
      "theoretical_loss": 5.217244993064664,
      "tokens_seen": 54001664
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006934296387365235,
      "loss": 4.1388,
      "theoretical_loss": 5.216291441053366,
      "tokens_seen": 54067200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006934130887081521,
      "loss": 4.3164,
      "theoretical_loss": 5.215339367345955,
      "tokens_seen": 54132736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006933965386797806,
      "loss": 4.1705,
      "theoretical_loss": 5.214388767865036,
      "tokens_seen": 54198272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000693379988651409,
      "loss": 4.4317,
      "theoretical_loss": 5.2134396385493815,
      "tokens_seen": 54263808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006933634386230376,
      "loss": 3.9078,
      "theoretical_loss": 5.212491975353835,
      "tokens_seen": 54329344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006933468885946661,
      "loss": 4.1712,
      "theoretical_loss": 5.211545774249233,
      "tokens_seen": 54394880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006933303385662947,
      "loss": 4.3638,
      "theoretical_loss": 5.210601031222324,
      "tokens_seen": 54460416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006933137885379232,
      "loss": 4.0157,
      "theoretical_loss": 5.209657742275683,
      "tokens_seen": 54525952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006932972385095517,
      "loss": 4.1828,
      "theoretical_loss": 5.208715903427631,
      "tokens_seen": 54591488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006932806884811803,
      "loss": 4.0533,
      "theoretical_loss": 5.207775510712159,
      "tokens_seen": 54657024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006932641384528088,
      "loss": 4.3225,
      "theoretical_loss": 5.2068365601788384,
      "tokens_seen": 54722560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006932475884244372,
      "loss": 4.1862,
      "theoretical_loss": 5.205899047892753,
      "tokens_seen": 54788096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006932310383960658,
      "loss": 3.9013,
      "theoretical_loss": 5.2049629699344075,
      "tokens_seen": 54853632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006932144883676943,
      "loss": 4.033,
      "theoretical_loss": 5.204028322399658,
      "tokens_seen": 54919168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006931979383393229,
      "loss": 4.1877,
      "theoretical_loss": 5.203095101399628,
      "tokens_seen": 54984704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006931813883109514,
      "loss": 4.2303,
      "theoretical_loss": 5.202163303060633,
      "tokens_seen": 55050240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006931648382825799,
      "loss": 4.183,
      "theoretical_loss": 5.201232923524104,
      "tokens_seen": 55115776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006931482882542085,
      "loss": 4.1661,
      "theoretical_loss": 5.20030395894651,
      "tokens_seen": 55181312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000693131738225837,
      "loss": 3.9613,
      "theoretical_loss": 5.199376405499277,
      "tokens_seen": 55246848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006931151881974654,
      "loss": 3.9902,
      "theoretical_loss": 5.198450259368721,
      "tokens_seen": 55312384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000693098638169094,
      "loss": 4.1145,
      "theoretical_loss": 5.197525516755965,
      "tokens_seen": 55377920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006930820881407225,
      "loss": 4.2884,
      "theoretical_loss": 5.196602173876867,
      "tokens_seen": 55443456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006930655381123511,
      "loss": 4.132,
      "theoretical_loss": 5.195680226961947,
      "tokens_seen": 55508992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006930489880839796,
      "loss": 4.0873,
      "theoretical_loss": 5.194759672256309,
      "tokens_seen": 55574528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006930324380556081,
      "loss": 3.6269,
      "theoretical_loss": 5.19384050601957,
      "tokens_seen": 55640064
    },
    {
      "epoch": 0.02,
      "objective/train/avg_token_score": 0.014826079830527306,
      "objective/train/avg_weight": 0.9825499057769775,
      "objective/train/docs_used": 42175,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.764860153198242,
      "objective/train/std_weight": 0.0845070630311966,
      "objective/train/theoretical_loss": 5.192922724525789,
      "objective/train/tokens_used": 76165600,
      "theoretical_loss": 5.192922724525789,
      "tokens_seen": 55705600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006930158880272367,
      "loss": 3.9019,
      "theoretical_loss": 5.192922724525789,
      "tokens_seen": 55705600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006929993379988651,
      "loss": 4.2445,
      "theoretical_loss": 5.19200632406339,
      "tokens_seen": 55771136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006929827879704936,
      "loss": 4.1435,
      "theoretical_loss": 5.19109130093509,
      "tokens_seen": 55836672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006929662379421222,
      "loss": 4.3442,
      "theoretical_loss": 5.190177651457833,
      "tokens_seen": 55902208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006929496879137507,
      "loss": 4.1968,
      "theoretical_loss": 5.189265371962712,
      "tokens_seen": 55967744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006929331378853793,
      "loss": 4.0468,
      "theoretical_loss": 5.188354458794902,
      "tokens_seen": 56033280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006929165878570078,
      "loss": 4.0182,
      "theoretical_loss": 5.187444908313586,
      "tokens_seen": 56098816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006929000378286363,
      "loss": 4.1566,
      "theoretical_loss": 5.186536716891892,
      "tokens_seen": 56164352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006928834878002648,
      "loss": 3.9491,
      "theoretical_loss": 5.185629880916814,
      "tokens_seen": 56229888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006928669377718933,
      "loss": 3.934,
      "theoretical_loss": 5.18472439678915,
      "tokens_seen": 56295424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006928503877435218,
      "loss": 4.3137,
      "theoretical_loss": 5.18382026092343,
      "tokens_seen": 56360960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006928338377151504,
      "loss": 3.8034,
      "theoretical_loss": 5.182917469747851,
      "tokens_seen": 56426496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006928172876867789,
      "loss": 3.9922,
      "theoretical_loss": 5.182016019704204,
      "tokens_seen": 56492032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006928007376584075,
      "loss": 4.0951,
      "theoretical_loss": 5.1811159072478095,
      "tokens_seen": 56557568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000692784187630036,
      "loss": 3.8332,
      "theoretical_loss": 5.180217128847451,
      "tokens_seen": 56623104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006927676376016644,
      "loss": 4.165,
      "theoretical_loss": 5.17931968098531,
      "tokens_seen": 56688640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000692751087573293,
      "loss": 4.0083,
      "theoretical_loss": 5.178423560156894,
      "tokens_seen": 56754176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006927345375449215,
      "loss": 4.1422,
      "theoretical_loss": 5.177528762870973,
      "tokens_seen": 56819712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00069271798751655,
      "loss": 4.0735,
      "theoretical_loss": 5.176635285649521,
      "tokens_seen": 56885248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006927014374881786,
      "loss": 4.0276,
      "theoretical_loss": 5.175743125027638,
      "tokens_seen": 56950784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006926848874598071,
      "loss": 4.0588,
      "theoretical_loss": 5.174852277553498,
      "tokens_seen": 57016320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006926683374314357,
      "loss": 3.8868,
      "theoretical_loss": 5.173962739788276,
      "tokens_seen": 57081856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006926517874030642,
      "loss": 3.9795,
      "theoretical_loss": 5.17307450830609,
      "tokens_seen": 57147392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006926352373746926,
      "loss": 4.2626,
      "theoretical_loss": 5.172187579693933,
      "tokens_seen": 57212928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006926186873463211,
      "loss": 4.1594,
      "theoretical_loss": 5.1713019505516105,
      "tokens_seen": 57278464
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006926021373179496,
      "loss": 4.1611,
      "theoretical_loss": 5.170417617491682,
      "tokens_seen": 57344000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006925855872895782,
      "loss": 4.2428,
      "theoretical_loss": 5.169534577139395,
      "tokens_seen": 57409536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006925690372612067,
      "loss": 3.9263,
      "theoretical_loss": 5.168652826132623,
      "tokens_seen": 57475072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006925524872328352,
      "loss": 4.1791,
      "theoretical_loss": 5.167772361121805,
      "tokens_seen": 57540608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006925359372044638,
      "loss": 4.1246,
      "theoretical_loss": 5.166893178769884,
      "tokens_seen": 57606144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006925193871760922,
      "loss": 4.1984,
      "theoretical_loss": 5.1660152757522475,
      "tokens_seen": 57671680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006925028371477208,
      "loss": 3.6761,
      "theoretical_loss": 5.165138648756665,
      "tokens_seen": 57737216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006924862871193493,
      "loss": 4.1385,
      "theoretical_loss": 5.164263294483226,
      "tokens_seen": 57802752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006924697370909778,
      "loss": 4.124,
      "theoretical_loss": 5.163389209644287,
      "tokens_seen": 57868288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006924531870626064,
      "loss": 4.1967,
      "theoretical_loss": 5.162516390964408,
      "tokens_seen": 57933824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006924366370342349,
      "loss": 3.9058,
      "theoretical_loss": 5.1616448351802875,
      "tokens_seen": 57999360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006924200870058634,
      "loss": 3.9278,
      "theoretical_loss": 5.160774539040716,
      "tokens_seen": 58064896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000692403536977492,
      "loss": 4.1848,
      "theoretical_loss": 5.159905499306511,
      "tokens_seen": 58130432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006923869869491204,
      "loss": 3.9331,
      "theoretical_loss": 5.159037712750455,
      "tokens_seen": 58195968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000692370436920749,
      "loss": 3.9552,
      "theoretical_loss": 5.158171176157245,
      "tokens_seen": 58261504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006923538868923775,
      "loss": 4.1885,
      "theoretical_loss": 5.157305886323435,
      "tokens_seen": 58327040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000692337336864006,
      "loss": 3.7721,
      "theoretical_loss": 5.156441840057371,
      "tokens_seen": 58392576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006923207868356346,
      "loss": 4.0939,
      "theoretical_loss": 5.155579034179144,
      "tokens_seen": 58458112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006923042368072631,
      "loss": 4.0097,
      "theoretical_loss": 5.15471746552053,
      "tokens_seen": 58523648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006922876867788915,
      "loss": 3.9944,
      "theoretical_loss": 5.153857130924929,
      "tokens_seen": 58589184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006922711367505201,
      "loss": 4.3663,
      "theoretical_loss": 5.1529980272473175,
      "tokens_seen": 58654720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006922545867221486,
      "loss": 3.8272,
      "theoretical_loss": 5.152140151354191,
      "tokens_seen": 58720256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006922380366937772,
      "loss": 4.0076,
      "theoretical_loss": 5.151283500123505,
      "tokens_seen": 58785792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006922214866654057,
      "loss": 3.9584,
      "theoretical_loss": 5.150428070444621,
      "tokens_seen": 58851328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006922049366370342,
      "loss": 4.0342,
      "theoretical_loss": 5.149573859218261,
      "tokens_seen": 58916864
    },
    {
      "epoch": 0.02,
      "objective/train/avg_token_score": 0.04295344278216362,
      "objective/train/avg_weight": 0.9494563937187195,
      "objective/train/docs_used": 44036,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.746087074279785,
      "objective/train/std_weight": 0.20767179131507874,
      "objective/train/theoretical_loss": 5.1487208633564405,
      "objective/train/tokens_used": 79442400,
      "theoretical_loss": 5.1487208633564405,
      "tokens_seen": 58982400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006921883866086628,
      "loss": 3.7038,
      "theoretical_loss": 5.1487208633564405,
      "tokens_seen": 58982400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006921718365802913,
      "loss": 3.7517,
      "theoretical_loss": 5.147869079782423,
      "tokens_seen": 59047936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006921552865519197,
      "loss": 4.0537,
      "theoretical_loss": 5.147018505430666,
      "tokens_seen": 59113472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006921387365235483,
      "loss": 3.9347,
      "theoretical_loss": 5.146169137246765,
      "tokens_seen": 59179008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006921221864951768,
      "loss": 3.8096,
      "theoretical_loss": 5.145320972187402,
      "tokens_seen": 59244544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006921056364668054,
      "loss": 4.2075,
      "theoretical_loss": 5.144474007220293,
      "tokens_seen": 59310080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006920890864384339,
      "loss": 3.9704,
      "theoretical_loss": 5.143628239324139,
      "tokens_seen": 59375616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006920725364100624,
      "loss": 4.0612,
      "theoretical_loss": 5.142783665488567,
      "tokens_seen": 59441152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000692055986381691,
      "loss": 4.2251,
      "theoretical_loss": 5.1419402827140885,
      "tokens_seen": 59506688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006920394363533195,
      "loss": 4.1385,
      "theoretical_loss": 5.141098088012036,
      "tokens_seen": 59572224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006920228863249479,
      "loss": 3.9277,
      "theoretical_loss": 5.140257078404524,
      "tokens_seen": 59637760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006920063362965765,
      "loss": 4.1021,
      "theoretical_loss": 5.13941725092439,
      "tokens_seen": 59703296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000691989786268205,
      "loss": 4.0529,
      "theoretical_loss": 5.138578602615146,
      "tokens_seen": 59768832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006919732362398336,
      "loss": 4.0784,
      "theoretical_loss": 5.137741130530934,
      "tokens_seen": 59834368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006919566862114621,
      "loss": 3.8994,
      "theoretical_loss": 5.1369048317364685,
      "tokens_seen": 59899904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006919401361830906,
      "loss": 4.1785,
      "theoretical_loss": 5.13606970330699,
      "tokens_seen": 59965440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006919235861547192,
      "loss": 4.0059,
      "theoretical_loss": 5.135235742328217,
      "tokens_seen": 60030976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006919070361263476,
      "loss": 3.9191,
      "theoretical_loss": 5.134402945896297,
      "tokens_seen": 60096512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006918904860979762,
      "loss": 4.1226,
      "theoretical_loss": 5.133571311117755,
      "tokens_seen": 60162048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006918739360696047,
      "loss": 3.9939,
      "theoretical_loss": 5.132740835109448,
      "tokens_seen": 60227584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006918573860412332,
      "loss": 3.9732,
      "theoretical_loss": 5.131911514998518,
      "tokens_seen": 60293120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006918408360128618,
      "loss": 3.7944,
      "theoretical_loss": 5.131083347922338,
      "tokens_seen": 60358656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006918242859844903,
      "loss": 3.7417,
      "theoretical_loss": 5.130256331028474,
      "tokens_seen": 60424192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006918077359561188,
      "loss": 3.8553,
      "theoretical_loss": 5.129430461474628,
      "tokens_seen": 60489728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006917911859277474,
      "loss": 3.7923,
      "theoretical_loss": 5.128605736428597,
      "tokens_seen": 60555264
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006917746358993758,
      "loss": 3.7437,
      "theoretical_loss": 5.127782153068225,
      "tokens_seen": 60620800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006917580858710044,
      "loss": 3.8784,
      "theoretical_loss": 5.126959708581356,
      "tokens_seen": 60686336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006917415358426329,
      "loss": 3.8498,
      "theoretical_loss": 5.1261384001657895,
      "tokens_seen": 60751872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006917249858142614,
      "loss": 3.9966,
      "theoretical_loss": 5.125318225029231,
      "tokens_seen": 60817408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00069170843578589,
      "loss": 3.8502,
      "theoretical_loss": 5.124499180389249,
      "tokens_seen": 60882944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006916918857575185,
      "loss": 3.7597,
      "theoretical_loss": 5.12368126347323,
      "tokens_seen": 60948480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000691675335729147,
      "loss": 4.1789,
      "theoretical_loss": 5.122864471518334,
      "tokens_seen": 61014016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006916587857007755,
      "loss": 3.6936,
      "theoretical_loss": 5.122048801771443,
      "tokens_seen": 61079552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006916422356724039,
      "loss": 3.8891,
      "theoretical_loss": 5.121234251489128,
      "tokens_seen": 61145088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006916256856440325,
      "loss": 3.9953,
      "theoretical_loss": 5.120420817937591,
      "tokens_seen": 61210624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000691609135615661,
      "loss": 4.0396,
      "theoretical_loss": 5.119608498392633,
      "tokens_seen": 61276160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006915925855872896,
      "loss": 3.8553,
      "theoretical_loss": 5.118797290139605,
      "tokens_seen": 61341696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006915760355589181,
      "loss": 3.9816,
      "theoretical_loss": 5.117987190473361,
      "tokens_seen": 61407232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006915594855305466,
      "loss": 3.6918,
      "theoretical_loss": 5.1171781966982195,
      "tokens_seen": 61472768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006915429355021751,
      "loss": 4.082,
      "theoretical_loss": 5.116370306127921,
      "tokens_seen": 61538304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006915263854738036,
      "loss": 3.919,
      "theoretical_loss": 5.11556351608558,
      "tokens_seen": 61603840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006915098354454321,
      "loss": 4.06,
      "theoretical_loss": 5.114757823903647,
      "tokens_seen": 61669376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006914932854170607,
      "loss": 3.7171,
      "theoretical_loss": 5.113953226923864,
      "tokens_seen": 61734912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006914767353886892,
      "loss": 3.9781,
      "theoretical_loss": 5.113149722497221,
      "tokens_seen": 61800448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006914601853603178,
      "loss": 3.9143,
      "theoretical_loss": 5.112347307983919,
      "tokens_seen": 61865984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006914436353319463,
      "loss": 3.8755,
      "theoretical_loss": 5.111545980753322,
      "tokens_seen": 61931520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006914270853035747,
      "loss": 3.9072,
      "theoretical_loss": 5.110745738183919,
      "tokens_seen": 61997056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006914105352752033,
      "loss": 3.6445,
      "theoretical_loss": 5.109946577663284,
      "tokens_seen": 62062592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006913939852468318,
      "loss": 3.6633,
      "theoretical_loss": 5.109148496588032,
      "tokens_seen": 62128128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006913774352184603,
      "loss": 3.6048,
      "theoretical_loss": 5.108351492363779,
      "tokens_seen": 62193664
    },
    {
      "epoch": 0.02,
      "objective/train/avg_token_score": 0.008287075906991959,
      "objective/train/avg_weight": 0.990248441696167,
      "objective/train/docs_used": 45869,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.876835346221924,
      "objective/train/std_weight": 0.06015734747052193,
      "objective/train/theoretical_loss": 5.107555562405102,
      "objective/train/tokens_used": 82719200,
      "theoretical_loss": 5.107555562405102,
      "tokens_seen": 62259200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006913608851900889,
      "loss": 3.7264,
      "theoretical_loss": 5.107555562405102,
      "tokens_seen": 62259200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006913443351617174,
      "loss": 3.5827,
      "theoretical_loss": 5.106760704135499,
      "tokens_seen": 62324736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000691327785133346,
      "loss": 3.6777,
      "theoretical_loss": 5.105966914987349,
      "tokens_seen": 62390272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006913112351049745,
      "loss": 3.9192,
      "theoretical_loss": 5.1051741924018685,
      "tokens_seen": 62455808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006912946850766029,
      "loss": 3.5905,
      "theoretical_loss": 5.10438253382908,
      "tokens_seen": 62521344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006912781350482315,
      "loss": 3.7266,
      "theoretical_loss": 5.103591936727762,
      "tokens_seen": 62586880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00069126158501986,
      "loss": 3.9492,
      "theoretical_loss": 5.102802398565418,
      "tokens_seen": 62652416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006912450349914885,
      "loss": 3.919,
      "theoretical_loss": 5.102013916818235,
      "tokens_seen": 62717952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006912284849631171,
      "loss": 4.0317,
      "theoretical_loss": 5.101226488971042,
      "tokens_seen": 62783488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006912119349347456,
      "loss": 4.0277,
      "theoretical_loss": 5.100440112517276,
      "tokens_seen": 62849024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006911953849063742,
      "loss": 3.8511,
      "theoretical_loss": 5.09965478495894,
      "tokens_seen": 62914560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006911788348780026,
      "loss": 3.8332,
      "theoretical_loss": 5.098870503806567,
      "tokens_seen": 62980096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006911622848496311,
      "loss": 3.9677,
      "theoretical_loss": 5.09808726657918,
      "tokens_seen": 63045632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006911457348212597,
      "loss": 3.932,
      "theoretical_loss": 5.097305070804255,
      "tokens_seen": 63111168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006911291847928882,
      "loss": 3.7431,
      "theoretical_loss": 5.096523914017688,
      "tokens_seen": 63176704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006911126347645167,
      "loss": 3.7376,
      "theoretical_loss": 5.095743793763747,
      "tokens_seen": 63242240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006910960847361453,
      "loss": 3.8638,
      "theoretical_loss": 5.094964707595047,
      "tokens_seen": 63307776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006910795347077738,
      "loss": 3.8856,
      "theoretical_loss": 5.094186653072505,
      "tokens_seen": 63373312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006910629846794024,
      "loss": 3.9528,
      "theoretical_loss": 5.093409627765306,
      "tokens_seen": 63438848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006910464346510308,
      "loss": 3.8506,
      "theoretical_loss": 5.092633629250866,
      "tokens_seen": 63504384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006910298846226593,
      "loss": 3.909,
      "theoretical_loss": 5.091858655114796,
      "tokens_seen": 63569920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006910133345942879,
      "loss": 3.9093,
      "theoretical_loss": 5.091084702950868,
      "tokens_seen": 63635456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006909967845659164,
      "loss": 3.7696,
      "theoretical_loss": 5.090311770360971,
      "tokens_seen": 63700992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006909802345375449,
      "loss": 3.789,
      "theoretical_loss": 5.089539854955088,
      "tokens_seen": 63766528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006909636845091735,
      "loss": 3.8335,
      "theoretical_loss": 5.088768954351249,
      "tokens_seen": 63832064
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000690947134480802,
      "loss": 3.6775,
      "theoretical_loss": 5.087999066175502,
      "tokens_seen": 63897600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006909305844524305,
      "loss": 3.7804,
      "theoretical_loss": 5.0872301880618735,
      "tokens_seen": 63963136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000690914034424059,
      "loss": 3.8243,
      "theoretical_loss": 5.086462317652341,
      "tokens_seen": 64028672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006908974843956875,
      "loss": 3.7675,
      "theoretical_loss": 5.085695452596788,
      "tokens_seen": 64094208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006908809343673161,
      "loss": 3.842,
      "theoretical_loss": 5.084929590552976,
      "tokens_seen": 64159744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006908643843389446,
      "loss": 3.8265,
      "theoretical_loss": 5.0841647291865115,
      "tokens_seen": 64225280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006908478343105731,
      "loss": 3.8074,
      "theoretical_loss": 5.083400866170806,
      "tokens_seen": 64290816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006908312842822017,
      "loss": 3.9583,
      "theoretical_loss": 5.082637999187046,
      "tokens_seen": 64356352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006908147342538301,
      "loss": 3.8096,
      "theoretical_loss": 5.081876125924159,
      "tokens_seen": 64421888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006907981842254587,
      "loss": 3.5749,
      "theoretical_loss": 5.0811152440787755,
      "tokens_seen": 64487424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006907816341970872,
      "loss": 3.843,
      "theoretical_loss": 5.0803553513552036,
      "tokens_seen": 64552960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006907650841687157,
      "loss": 3.6934,
      "theoretical_loss": 5.079596445465386,
      "tokens_seen": 64618496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006907485341403443,
      "loss": 3.8961,
      "theoretical_loss": 5.078838524128878,
      "tokens_seen": 64684032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006907319841119728,
      "loss": 4.1123,
      "theoretical_loss": 5.078081585072802,
      "tokens_seen": 64749568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006907154340836013,
      "loss": 3.7531,
      "theoretical_loss": 5.077325626031826,
      "tokens_seen": 64815104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006906988840552299,
      "loss": 3.6968,
      "theoretical_loss": 5.076570644748123,
      "tokens_seen": 64880640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006906823340268583,
      "loss": 3.9381,
      "theoretical_loss": 5.075816638971341,
      "tokens_seen": 64946176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006906657839984869,
      "loss": 3.6544,
      "theoretical_loss": 5.075063606458576,
      "tokens_seen": 65011712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006906492339701153,
      "loss": 3.4993,
      "theoretical_loss": 5.074311544974331,
      "tokens_seen": 65077248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006906326839417439,
      "loss": 3.6908,
      "theoretical_loss": 5.07356045229049,
      "tokens_seen": 65142784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006906161339133724,
      "loss": 3.7965,
      "theoretical_loss": 5.072810326186285,
      "tokens_seen": 65208320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006905995838850009,
      "loss": 3.8404,
      "theoretical_loss": 5.072061164448261,
      "tokens_seen": 65273856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006905830338566295,
      "loss": 3.8435,
      "theoretical_loss": 5.071312964870252,
      "tokens_seen": 65339392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006905664838282579,
      "loss": 3.8376,
      "theoretical_loss": 5.070565725253344,
      "tokens_seen": 65404928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006905499337998865,
      "loss": 3.8723,
      "theoretical_loss": 5.069819443405842,
      "tokens_seen": 65470464
    },
    {
      "epoch": 0.02,
      "objective/train/avg_token_score": 0.02166317217051983,
      "objective/train/avg_weight": 0.974501371383667,
      "objective/train/docs_used": 47778,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.8713741302490234,
      "objective/train/std_weight": 0.09755385667085648,
      "objective/train/theoretical_loss": 5.069074117143246,
      "objective/train/tokens_used": 85996000,
      "theoretical_loss": 5.069074117143246,
      "tokens_seen": 65536000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000690533383771515,
      "loss": 3.8,
      "theoretical_loss": 5.069074117143246,
      "tokens_seen": 65536000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006905168337431435,
      "loss": 3.6949,
      "theoretical_loss": 5.068329744288216,
      "tokens_seen": 65601536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006905002837147721,
      "loss": 3.7228,
      "theoretical_loss": 5.067586322670541,
      "tokens_seen": 65667072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006904837336864006,
      "loss": 3.7891,
      "theoretical_loss": 5.0668438501271105,
      "tokens_seen": 65732608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000690467183658029,
      "loss": 3.941,
      "theoretical_loss": 5.066102324501883,
      "tokens_seen": 65798144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006904506336296576,
      "loss": 3.8617,
      "theoretical_loss": 5.065361743645855,
      "tokens_seen": 65863680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006904340836012861,
      "loss": 3.72,
      "theoretical_loss": 5.064622105417033,
      "tokens_seen": 65929216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006904175335729147,
      "loss": 4.0502,
      "theoretical_loss": 5.063883407680405,
      "tokens_seen": 65994752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006904009835445432,
      "loss": 3.6957,
      "theoretical_loss": 5.063145648307904,
      "tokens_seen": 66060288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006903844335161717,
      "loss": 3.7182,
      "theoretical_loss": 5.062408825178388,
      "tokens_seen": 66125824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006903678834878003,
      "loss": 3.7784,
      "theoretical_loss": 5.061672936177604,
      "tokens_seen": 66191360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006903513334594288,
      "loss": 3.6848,
      "theoretical_loss": 5.06093797919816,
      "tokens_seen": 66256896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006903347834310572,
      "loss": 3.9823,
      "theoretical_loss": 5.060203952139497,
      "tokens_seen": 66322432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006903182334026858,
      "loss": 3.8234,
      "theoretical_loss": 5.059470852907861,
      "tokens_seen": 66387968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006903016833743143,
      "loss": 3.7332,
      "theoretical_loss": 5.0587386794162725,
      "tokens_seen": 66453504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006902851333459429,
      "loss": 3.616,
      "theoretical_loss": 5.058007429584498,
      "tokens_seen": 66519040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006902685833175714,
      "loss": 3.6886,
      "theoretical_loss": 5.057277101339023,
      "tokens_seen": 66584576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006902520332891999,
      "loss": 3.758,
      "theoretical_loss": 5.056547692613021,
      "tokens_seen": 66650112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006902354832608285,
      "loss": 3.9053,
      "theoretical_loss": 5.055819201346331,
      "tokens_seen": 66715648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000690218933232457,
      "loss": 3.6385,
      "theoretical_loss": 5.055091625485421,
      "tokens_seen": 66781184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006902023832040854,
      "loss": 4.0574,
      "theoretical_loss": 5.054364962983367,
      "tokens_seen": 66846720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000690185833175714,
      "loss": 4.0204,
      "theoretical_loss": 5.053639211799824,
      "tokens_seen": 66912256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006901692831473425,
      "loss": 3.8911,
      "theoretical_loss": 5.052914369900997,
      "tokens_seen": 66977792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006901527331189711,
      "loss": 3.8315,
      "theoretical_loss": 5.052190435259614,
      "tokens_seen": 67043328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006901361830905996,
      "loss": 3.7451,
      "theoretical_loss": 5.051467405854897,
      "tokens_seen": 67108864
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006901196330622281,
      "loss": 3.7621,
      "theoretical_loss": 5.05074527967254,
      "tokens_seen": 67174400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006901030830338567,
      "loss": 3.8774,
      "theoretical_loss": 5.050024054704677,
      "tokens_seen": 67239936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006900865330054851,
      "loss": 3.7939,
      "theoretical_loss": 5.049303728949859,
      "tokens_seen": 67305472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006900699829771136,
      "loss": 3.8488,
      "theoretical_loss": 5.048584300413019,
      "tokens_seen": 67371008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006900534329487422,
      "loss": 3.8907,
      "theoretical_loss": 5.04786576710546,
      "tokens_seen": 67436544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006900368829203707,
      "loss": 3.6678,
      "theoretical_loss": 5.0471481270448155,
      "tokens_seen": 67502080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006900203328919993,
      "loss": 3.8831,
      "theoretical_loss": 5.046431378255027,
      "tokens_seen": 67567616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006900037828636278,
      "loss": 3.7127,
      "theoretical_loss": 5.045715518766322,
      "tokens_seen": 67633152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006899872328352563,
      "loss": 3.7037,
      "theoretical_loss": 5.0450005466151815,
      "tokens_seen": 67698688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006899706828068849,
      "loss": 3.7676,
      "theoretical_loss": 5.044286459844319,
      "tokens_seen": 67764224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006899541327785133,
      "loss": 3.7438,
      "theoretical_loss": 5.043573256502652,
      "tokens_seen": 67829760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006899375827501418,
      "loss": 3.8359,
      "theoretical_loss": 5.0428609346452795,
      "tokens_seen": 67895296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006899210327217704,
      "loss": 3.8242,
      "theoretical_loss": 5.042149492333452,
      "tokens_seen": 67960832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006899044826933989,
      "loss": 3.8242,
      "theoretical_loss": 5.041438927634549,
      "tokens_seen": 68026368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006898879326650275,
      "loss": 3.8219,
      "theoretical_loss": 5.040729238622053,
      "tokens_seen": 68091904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000689871382636656,
      "loss": 4.0827,
      "theoretical_loss": 5.040020423375525,
      "tokens_seen": 68157440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006898548326082845,
      "loss": 3.7776,
      "theoretical_loss": 5.039312479980579,
      "tokens_seen": 68222976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000689838282579913,
      "loss": 3.632,
      "theoretical_loss": 5.038605406528857,
      "tokens_seen": 68288512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006898217325515415,
      "loss": 3.8848,
      "theoretical_loss": 5.037899201118005,
      "tokens_seen": 68354048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00068980518252317,
      "loss": 3.4744,
      "theoretical_loss": 5.037193861851646,
      "tokens_seen": 68419584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006897886324947986,
      "loss": 3.7576,
      "theoretical_loss": 5.03648938683936,
      "tokens_seen": 68485120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006897720824664271,
      "loss": 3.9633,
      "theoretical_loss": 5.035785774196654,
      "tokens_seen": 68550656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006897555324380557,
      "loss": 3.5832,
      "theoretical_loss": 5.035083022044944,
      "tokens_seen": 68616192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006897389824096842,
      "loss": 3.5699,
      "theoretical_loss": 5.034381128511525,
      "tokens_seen": 68681728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006897224323813126,
      "loss": 3.6594,
      "theoretical_loss": 5.0336800917295506,
      "tokens_seen": 68747264
    },
    {
      "epoch": 0.02,
      "objective/train/avg_token_score": 0.009921571239829063,
      "objective/train/avg_weight": 0.9883232712745667,
      "objective/train/docs_used": 49496,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.83107852935791,
      "objective/train/std_weight": 0.05008544400334358,
      "objective/train/theoretical_loss": 5.032979909838007,
      "objective/train/tokens_used": 89272800,
      "theoretical_loss": 5.032979909838007,
      "tokens_seen": 68812800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006897058823529412,
      "loss": 3.8138,
      "theoretical_loss": 5.032979909838007,
      "tokens_seen": 68812800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006896893323245697,
      "loss": 3.9847,
      "theoretical_loss": 5.032280580981691,
      "tokens_seen": 68878336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006896727822961982,
      "loss": 3.6918,
      "theoretical_loss": 5.031582103311187,
      "tokens_seen": 68943872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006896562322678267,
      "loss": 3.8223,
      "theoretical_loss": 5.030884474982842,
      "tokens_seen": 69009408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006896396822394552,
      "loss": 3.624,
      "theoretical_loss": 5.030187694158739,
      "tokens_seen": 69074944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006896231322110838,
      "loss": 3.7397,
      "theoretical_loss": 5.02949175900668,
      "tokens_seen": 69140480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006896065821827122,
      "loss": 3.9202,
      "theoretical_loss": 5.028796667700159,
      "tokens_seen": 69206016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006895900321543408,
      "loss": 3.9324,
      "theoretical_loss": 5.0281024184183405,
      "tokens_seen": 69271552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006895734821259693,
      "loss": 3.9112,
      "theoretical_loss": 5.0274090093460355,
      "tokens_seen": 69337088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006895569320975978,
      "loss": 3.8159,
      "theoretical_loss": 5.026716438673677,
      "tokens_seen": 69402624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006895403820692264,
      "loss": 3.5544,
      "theoretical_loss": 5.0260247045973045,
      "tokens_seen": 69468160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006895238320408549,
      "loss": 3.9137,
      "theoretical_loss": 5.02533380531853,
      "tokens_seen": 69533696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006895072820124835,
      "loss": 3.5748,
      "theoretical_loss": 5.024643739044526,
      "tokens_seen": 69599232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000689490731984112,
      "loss": 3.5986,
      "theoretical_loss": 5.023954503987998,
      "tokens_seen": 69664768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006894741819557404,
      "loss": 3.8723,
      "theoretical_loss": 5.023266098367161,
      "tokens_seen": 69730304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000689457631927369,
      "loss": 3.669,
      "theoretical_loss": 5.022578520405721,
      "tokens_seen": 69795840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006894410818989975,
      "loss": 3.5709,
      "theoretical_loss": 5.0218917683328534,
      "tokens_seen": 69861376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000689424531870626,
      "loss": 3.8255,
      "theoretical_loss": 5.021205840383175,
      "tokens_seen": 69926912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0006894079818422546,
      "loss": 3.6235,
      "theoretical_loss": 5.020520734796728,
      "tokens_seen": 69992448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006893914318138831,
      "loss": 3.4081,
      "theoretical_loss": 5.019836449818957,
      "tokens_seen": 70057984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006893748817855117,
      "loss": 3.6605,
      "theoretical_loss": 5.019152983700687,
      "tokens_seen": 70123520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006893583317571401,
      "loss": 3.706,
      "theoretical_loss": 5.018470334698101,
      "tokens_seen": 70189056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006893417817287686,
      "loss": 3.7298,
      "theoretical_loss": 5.01778850107272,
      "tokens_seen": 70254592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006893252317003972,
      "loss": 3.7154,
      "theoretical_loss": 5.017107481091379,
      "tokens_seen": 70320128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006893086816720257,
      "loss": 3.7976,
      "theoretical_loss": 5.016427273026212,
      "tokens_seen": 70385664
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006892921316436542,
      "loss": 3.8711,
      "theoretical_loss": 5.015747875154622,
      "tokens_seen": 70451200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006892755816152828,
      "loss": 3.5807,
      "theoretical_loss": 5.015069285759269,
      "tokens_seen": 70516736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006892590315869113,
      "loss": 3.8519,
      "theoretical_loss": 5.01439150312804,
      "tokens_seen": 70582272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006892424815585399,
      "loss": 3.8338,
      "theoretical_loss": 5.0137145255540405,
      "tokens_seen": 70647808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006892259315301683,
      "loss": 3.8396,
      "theoretical_loss": 5.013038351335559,
      "tokens_seen": 70713344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006892093815017968,
      "loss": 3.8973,
      "theoretical_loss": 5.012362978776057,
      "tokens_seen": 70778880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006891928314734254,
      "loss": 3.8575,
      "theoretical_loss": 5.011688406184147,
      "tokens_seen": 70844416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006891762814450539,
      "loss": 3.5637,
      "theoretical_loss": 5.011014631873566,
      "tokens_seen": 70909952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006891597314166824,
      "loss": 3.6933,
      "theoretical_loss": 5.010341654163167,
      "tokens_seen": 70975488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000689143181388311,
      "loss": 3.5602,
      "theoretical_loss": 5.009669471376882,
      "tokens_seen": 71041024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006891266313599395,
      "loss": 3.6911,
      "theoretical_loss": 5.008998081843721,
      "tokens_seen": 71106560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000689110081331568,
      "loss": 3.7986,
      "theoretical_loss": 5.008327483897736,
      "tokens_seen": 71172096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006890935313031965,
      "loss": 3.7517,
      "theoretical_loss": 5.00765767587801,
      "tokens_seen": 71237632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000689076981274825,
      "loss": 3.9759,
      "theoretical_loss": 5.006988656128635,
      "tokens_seen": 71303168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006890604312464536,
      "loss": 3.7954,
      "theoretical_loss": 5.006320422998691,
      "tokens_seen": 71368704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006890438812180821,
      "loss": 3.6831,
      "theoretical_loss": 5.00565297484223,
      "tokens_seen": 71434240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006890273311897106,
      "loss": 3.6897,
      "theoretical_loss": 5.004986310018252,
      "tokens_seen": 71499776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006890107811613392,
      "loss": 3.6683,
      "theoretical_loss": 5.004320426890686,
      "tokens_seen": 71565312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006889942311329676,
      "loss": 3.4581,
      "theoretical_loss": 5.003655323828376,
      "tokens_seen": 71630848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006889776811045962,
      "loss": 3.6333,
      "theoretical_loss": 5.002990999205057,
      "tokens_seen": 71696384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006889611310762247,
      "loss": 3.7539,
      "theoretical_loss": 5.002327451399335,
      "tokens_seen": 71761920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006889445810478532,
      "loss": 4.0036,
      "theoretical_loss": 5.001664678794671,
      "tokens_seen": 71827456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006889280310194818,
      "loss": 3.6134,
      "theoretical_loss": 5.001002679779363,
      "tokens_seen": 71892992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006889114809911103,
      "loss": 3.5746,
      "theoretical_loss": 5.0003414527465235,
      "tokens_seen": 71958528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006888949309627388,
      "loss": 3.5362,
      "theoretical_loss": 4.99968099609406,
      "tokens_seen": 72024064
    },
    {
      "epoch": 0.03,
      "objective/train/avg_token_score": 0.013204310089349747,
      "objective/train/avg_weight": 0.9844692349433899,
      "objective/train/docs_used": 51476,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.624199390411377,
      "objective/train/std_weight": 0.05366871878504753,
      "objective/train/theoretical_loss": 4.999021308224664,
      "objective/train/tokens_used": 92549600,
      "theoretical_loss": 4.999021308224664,
      "tokens_seen": 72089600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006888783809343674,
      "loss": 3.5738,
      "theoretical_loss": 4.999021308224664,
      "tokens_seen": 72089600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006888618309059958,
      "loss": 3.6446,
      "theoretical_loss": 4.998362387545782,
      "tokens_seen": 72155136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006888452808776244,
      "loss": 3.7694,
      "theoretical_loss": 4.997704232469606,
      "tokens_seen": 72220672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006888287308492529,
      "loss": 3.5828,
      "theoretical_loss": 4.997046841413049,
      "tokens_seen": 72286208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006888121808208814,
      "loss": 3.7638,
      "theoretical_loss": 4.996390212797728,
      "tokens_seen": 72351744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00068879563079251,
      "loss": 3.4573,
      "theoretical_loss": 4.995734345049949,
      "tokens_seen": 72417280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006887790807641385,
      "loss": 3.7883,
      "theoretical_loss": 4.995079236600686,
      "tokens_seen": 72482816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000688762530735767,
      "loss": 3.8957,
      "theoretical_loss": 4.994424885885564,
      "tokens_seen": 72548352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006887459807073955,
      "loss": 3.8065,
      "theoretical_loss": 4.993771291344839,
      "tokens_seen": 72613888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000688729430679024,
      "loss": 3.6693,
      "theoretical_loss": 4.993118451423381,
      "tokens_seen": 72679424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006887128806506526,
      "loss": 3.4174,
      "theoretical_loss": 4.992466364570659,
      "tokens_seen": 72744960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006886963306222811,
      "loss": 3.738,
      "theoretical_loss": 4.991815029240721,
      "tokens_seen": 72810496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006886797805939096,
      "loss": 3.5045,
      "theoretical_loss": 4.991164443892175,
      "tokens_seen": 72876032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006886632305655381,
      "loss": 3.6336,
      "theoretical_loss": 4.990514606988173,
      "tokens_seen": 72941568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006886466805371666,
      "loss": 3.8963,
      "theoretical_loss": 4.989865516996396,
      "tokens_seen": 73007104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006886301305087951,
      "loss": 3.8094,
      "theoretical_loss": 4.98921717238903,
      "tokens_seen": 73072640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006886135804804236,
      "loss": 3.7429,
      "theoretical_loss": 4.988569571642756,
      "tokens_seen": 73138176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006885970304520521,
      "loss": 3.5191,
      "theoretical_loss": 4.98792271323873,
      "tokens_seen": 73203712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006885804804236807,
      "loss": 3.649,
      "theoretical_loss": 4.9872765956625615,
      "tokens_seen": 73269248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006885639303953092,
      "loss": 3.5774,
      "theoretical_loss": 4.9866312174043035,
      "tokens_seen": 73334784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006885473803669378,
      "loss": 3.6099,
      "theoretical_loss": 4.9859865769584335,
      "tokens_seen": 73400320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006885308303385663,
      "loss": 3.5048,
      "theoretical_loss": 4.9853426728238315,
      "tokens_seen": 73465856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006885142803101947,
      "loss": 3.7514,
      "theoretical_loss": 4.984699503503771,
      "tokens_seen": 73531392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006884977302818233,
      "loss": 3.5395,
      "theoretical_loss": 4.984057067505898,
      "tokens_seen": 73596928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006884811802534518,
      "loss": 3.9503,
      "theoretical_loss": 4.9834153633422105,
      "tokens_seen": 73662464
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006884646302250803,
      "loss": 3.6216,
      "theoretical_loss": 4.982774389529053,
      "tokens_seen": 73728000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006884480801967089,
      "loss": 3.4013,
      "theoretical_loss": 4.9821341445870875,
      "tokens_seen": 73793536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006884315301683374,
      "loss": 3.6569,
      "theoretical_loss": 4.981494627041286,
      "tokens_seen": 73859072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000688414980139966,
      "loss": 3.8729,
      "theoretical_loss": 4.98085583542091,
      "tokens_seen": 73924608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006883984301115945,
      "loss": 3.6342,
      "theoretical_loss": 4.980217768259496,
      "tokens_seen": 73990144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006883818800832229,
      "loss": 3.4577,
      "theoretical_loss": 4.979580424094836,
      "tokens_seen": 74055680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006883653300548515,
      "loss": 3.5283,
      "theoretical_loss": 4.978943801468967,
      "tokens_seen": 74121216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00068834878002648,
      "loss": 3.7837,
      "theoretical_loss": 4.978307898928149,
      "tokens_seen": 74186752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006883322299981085,
      "loss": 3.6566,
      "theoretical_loss": 4.977672715022855,
      "tokens_seen": 74252288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006883156799697371,
      "loss": 3.5953,
      "theoretical_loss": 4.97703824830775,
      "tokens_seen": 74317824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006882991299413656,
      "loss": 3.7783,
      "theoretical_loss": 4.976404497341676,
      "tokens_seen": 74383360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006882825799129942,
      "loss": 3.5949,
      "theoretical_loss": 4.975771460687641,
      "tokens_seen": 74448896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006882660298846226,
      "loss": 3.8052,
      "theoretical_loss": 4.975139136912794,
      "tokens_seen": 74514432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006882494798562511,
      "loss": 3.5775,
      "theoretical_loss": 4.974507524588424,
      "tokens_seen": 74579968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006882329298278797,
      "loss": 3.8187,
      "theoretical_loss": 4.973876622289927,
      "tokens_seen": 74645504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006882163797995082,
      "loss": 3.6292,
      "theoretical_loss": 4.973246428596802,
      "tokens_seen": 74711040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006881998297711367,
      "loss": 3.665,
      "theoretical_loss": 4.972616942092634,
      "tokens_seen": 74776576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006881832797427653,
      "loss": 3.8113,
      "theoretical_loss": 4.971988161365077,
      "tokens_seen": 74842112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006881667297143938,
      "loss": 3.7201,
      "theoretical_loss": 4.9713600850058395,
      "tokens_seen": 74907648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006881501796860224,
      "loss": 3.7812,
      "theoretical_loss": 4.970732711610667,
      "tokens_seen": 74973184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006881336296576508,
      "loss": 3.7295,
      "theoretical_loss": 4.97010603977933,
      "tokens_seen": 75038720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006881170796292793,
      "loss": 3.5628,
      "theoretical_loss": 4.96948006811561,
      "tokens_seen": 75104256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006881005296009079,
      "loss": 3.6559,
      "theoretical_loss": 4.968854795227281,
      "tokens_seen": 75169792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006880839795725364,
      "loss": 3.4903,
      "theoretical_loss": 4.968230219726093,
      "tokens_seen": 75235328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006880674295441649,
      "loss": 3.485,
      "theoretical_loss": 4.967606340227765,
      "tokens_seen": 75300864
    },
    {
      "epoch": 0.03,
      "objective/train/avg_token_score": 0.013501526787877083,
      "objective/train/avg_weight": 0.9841073751449585,
      "objective/train/docs_used": 53412,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.83372163772583,
      "objective/train/std_weight": 0.0673174113035202,
      "objective/train/theoretical_loss": 4.966983155351962,
      "objective/train/tokens_used": 95826400,
      "theoretical_loss": 4.966983155351962,
      "tokens_seen": 75366400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006880508795157935,
      "loss": 3.7213,
      "theoretical_loss": 4.966983155351962,
      "tokens_seen": 75366400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000688034329487422,
      "loss": 3.6288,
      "theoretical_loss": 4.966360663722287,
      "tokens_seen": 75431936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006880177794590505,
      "loss": 3.6686,
      "theoretical_loss": 4.96573886396626,
      "tokens_seen": 75497472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000688001229430679,
      "loss": 3.6954,
      "theoretical_loss": 4.965117754715307,
      "tokens_seen": 75563008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006879846794023075,
      "loss": 3.7692,
      "theoretical_loss": 4.964497334604748,
      "tokens_seen": 75628544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006879681293739361,
      "loss": 3.7547,
      "theoretical_loss": 4.963877602273776,
      "tokens_seen": 75694080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006879515793455646,
      "loss": 3.7887,
      "theoretical_loss": 4.963258556365449,
      "tokens_seen": 75759616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006879350293171932,
      "loss": 3.5267,
      "theoretical_loss": 4.962640195526673,
      "tokens_seen": 75825152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006879184792888217,
      "loss": 3.6024,
      "theoretical_loss": 4.962022518408183,
      "tokens_seen": 75890688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006879019292604501,
      "loss": 3.5887,
      "theoretical_loss": 4.96140552366454,
      "tokens_seen": 75956224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006878853792320787,
      "loss": 3.6631,
      "theoretical_loss": 4.9607892099541075,
      "tokens_seen": 76021760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006878688292037072,
      "loss": 3.4075,
      "theoretical_loss": 4.9601735759390415,
      "tokens_seen": 76087296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006878522791753357,
      "loss": 3.8834,
      "theoretical_loss": 4.959558620285274,
      "tokens_seen": 76152832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006878357291469643,
      "loss": 3.3976,
      "theoretical_loss": 4.958944341662502,
      "tokens_seen": 76218368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006878191791185928,
      "loss": 3.6609,
      "theoretical_loss": 4.958330738744172,
      "tokens_seen": 76283904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006878026290902214,
      "loss": 3.6951,
      "theoretical_loss": 4.957717810207466,
      "tokens_seen": 76349440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006877860790618499,
      "loss": 3.6287,
      "theoretical_loss": 4.957105554733289,
      "tokens_seen": 76414976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006877695290334783,
      "loss": 3.6605,
      "theoretical_loss": 4.956493971006253,
      "tokens_seen": 76480512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006877529790051069,
      "loss": 3.6084,
      "theoretical_loss": 4.955883057714669,
      "tokens_seen": 76546048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006877364289767354,
      "loss": 3.4722,
      "theoretical_loss": 4.955272813550524,
      "tokens_seen": 76611584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006877198789483639,
      "loss": 3.5048,
      "theoretical_loss": 4.954663237209477,
      "tokens_seen": 76677120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006877033289199924,
      "loss": 3.4239,
      "theoretical_loss": 4.954054327390841,
      "tokens_seen": 76742656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006876867788916209,
      "loss": 3.4794,
      "theoretical_loss": 4.9534460827975675,
      "tokens_seen": 76808192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006876702288632495,
      "loss": 3.5659,
      "theoretical_loss": 4.952838502136241,
      "tokens_seen": 76873728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006876536788348779,
      "loss": 3.5284,
      "theoretical_loss": 4.952231584117056,
      "tokens_seen": 76939264
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006876371288065065,
      "loss": 3.6231,
      "theoretical_loss": 4.951625327453812,
      "tokens_seen": 77004800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000687620578778135,
      "loss": 3.5373,
      "theoretical_loss": 4.951019730863894,
      "tokens_seen": 77070336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006876040287497635,
      "loss": 3.7111,
      "theoretical_loss": 4.950414793068266,
      "tokens_seen": 77135872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006875874787213921,
      "loss": 3.6972,
      "theoretical_loss": 4.94981051279145,
      "tokens_seen": 77201408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006875709286930206,
      "loss": 3.5202,
      "theoretical_loss": 4.94920688876152,
      "tokens_seen": 77266944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006875543786646491,
      "loss": 3.64,
      "theoretical_loss": 4.948603919710088,
      "tokens_seen": 77332480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006875378286362777,
      "loss": 3.583,
      "theoretical_loss": 4.948001604372287,
      "tokens_seen": 77398016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006875212786079061,
      "loss": 3.4153,
      "theoretical_loss": 4.947399941486762,
      "tokens_seen": 77463552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006875047285795347,
      "loss": 3.778,
      "theoretical_loss": 4.946798929795658,
      "tokens_seen": 77529088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006874881785511632,
      "loss": 3.6162,
      "theoretical_loss": 4.946198568044602,
      "tokens_seen": 77594624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006874716285227917,
      "loss": 3.9088,
      "theoretical_loss": 4.945598854982698,
      "tokens_seen": 77660160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006874550784944203,
      "loss": 3.3088,
      "theoretical_loss": 4.944999789362508,
      "tokens_seen": 77725696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006874385284660488,
      "loss": 3.5686,
      "theoretical_loss": 4.944401369940043,
      "tokens_seen": 77791232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006874219784376773,
      "loss": 3.5744,
      "theoretical_loss": 4.9438035954747495,
      "tokens_seen": 77856768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006874054284093058,
      "loss": 3.509,
      "theoretical_loss": 4.9432064647294975,
      "tokens_seen": 77922304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006873888783809343,
      "loss": 3.4697,
      "theoretical_loss": 4.942609976470566,
      "tokens_seen": 77987840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006873723283525629,
      "loss": 3.5172,
      "theoretical_loss": 4.942014129467637,
      "tokens_seen": 78053376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006873557783241914,
      "loss": 3.4169,
      "theoretical_loss": 4.941418922493774,
      "tokens_seen": 78118912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006873392282958199,
      "loss": 3.4419,
      "theoretical_loss": 4.940824354325419,
      "tokens_seen": 78184448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006873226782674485,
      "loss": 3.5538,
      "theoretical_loss": 4.940230423742372,
      "tokens_seen": 78249984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000687306128239077,
      "loss": 3.334,
      "theoretical_loss": 4.939637129527789,
      "tokens_seen": 78315520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006872895782107054,
      "loss": 3.6022,
      "theoretical_loss": 4.939044470468156,
      "tokens_seen": 78381056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000687273028182334,
      "loss": 3.8116,
      "theoretical_loss": 4.938452445353294,
      "tokens_seen": 78446592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006872564781539625,
      "loss": 3.7126,
      "theoretical_loss": 4.937861052976332,
      "tokens_seen": 78512128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006872399281255911,
      "loss": 3.2962,
      "theoretical_loss": 4.937270292133704,
      "tokens_seen": 78577664
    },
    {
      "epoch": 0.03,
      "objective/train/avg_token_score": 0.033612485975027084,
      "objective/train/avg_weight": 0.9604661464691162,
      "objective/train/docs_used": 55300,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.508220672607422,
      "objective/train/std_weight": 0.174960657954216,
      "objective/train/theoretical_loss": 4.9366801616251355,
      "objective/train/tokens_used": 99103200,
      "theoretical_loss": 4.9366801616251355,
      "tokens_seen": 78643200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006872233780972196,
      "loss": 3.5492,
      "theoretical_loss": 4.9366801616251355,
      "tokens_seen": 78643200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006872068280688481,
      "loss": 3.4809,
      "theoretical_loss": 4.93609066025363,
      "tokens_seen": 78708736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006871902780404767,
      "loss": 3.5281,
      "theoretical_loss": 4.935501786825457,
      "tokens_seen": 78774272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006871737280121052,
      "loss": 3.4527,
      "theoretical_loss": 4.934913540150143,
      "tokens_seen": 78839808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006871571779837336,
      "loss": 3.5737,
      "theoretical_loss": 4.934325919040461,
      "tokens_seen": 78905344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006871406279553622,
      "loss": 3.4639,
      "theoretical_loss": 4.933738922312413,
      "tokens_seen": 78970880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006871240779269907,
      "loss": 3.597,
      "theoretical_loss": 4.933152548785222,
      "tokens_seen": 79036416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006871075278986193,
      "loss": 3.7061,
      "theoretical_loss": 4.932566797281324,
      "tokens_seen": 79101952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006870909778702478,
      "loss": 3.6442,
      "theoretical_loss": 4.931981666626351,
      "tokens_seen": 79167488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006870744278418763,
      "loss": 3.6729,
      "theoretical_loss": 4.931397155649121,
      "tokens_seen": 79233024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006870578778135049,
      "loss": 3.6287,
      "theoretical_loss": 4.930813263181631,
      "tokens_seen": 79298560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006870413277851333,
      "loss": 3.8565,
      "theoretical_loss": 4.93022998805904,
      "tokens_seen": 79364096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006870247777567618,
      "loss": 3.7185,
      "theoretical_loss": 4.929647329119659,
      "tokens_seen": 79429632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006870082277283904,
      "loss": 3.3918,
      "theoretical_loss": 4.9290652852049455,
      "tokens_seen": 79495168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006869916777000189,
      "loss": 3.9657,
      "theoretical_loss": 4.928483855159485,
      "tokens_seen": 79560704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006869751276716475,
      "loss": 3.3307,
      "theoretical_loss": 4.927903037830983,
      "tokens_seen": 79626240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000686958577643276,
      "loss": 3.6048,
      "theoretical_loss": 4.9273228320702565,
      "tokens_seen": 79691776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006869420276149045,
      "loss": 3.51,
      "theoretical_loss": 4.926743236731218,
      "tokens_seen": 79757312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000686925477586533,
      "loss": 3.484,
      "theoretical_loss": 4.926164250670868,
      "tokens_seen": 79822848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006869089275581615,
      "loss": 3.5629,
      "theoretical_loss": 4.925585872749284,
      "tokens_seen": 79888384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00068689237752979,
      "loss": 3.4576,
      "theoretical_loss": 4.925008101829608,
      "tokens_seen": 79953920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006868758275014186,
      "loss": 3.5592,
      "theoretical_loss": 4.9244309367780374,
      "tokens_seen": 80019456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006868592774730471,
      "loss": 3.3405,
      "theoretical_loss": 4.923854376463816,
      "tokens_seen": 80084992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006868427274446757,
      "loss": 3.3673,
      "theoretical_loss": 4.923278419759217,
      "tokens_seen": 80150528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006868261774163042,
      "loss": 3.5515,
      "theoretical_loss": 4.92270306553954,
      "tokens_seen": 80216064
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006868096273879327,
      "loss": 3.6659,
      "theoretical_loss": 4.922128312683096,
      "tokens_seen": 80281600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006867930773595612,
      "loss": 3.7412,
      "theoretical_loss": 4.921554160071194,
      "tokens_seen": 80347136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006867765273311897,
      "loss": 3.502,
      "theoretical_loss": 4.920980606588142,
      "tokens_seen": 80412672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006867599773028182,
      "loss": 3.3174,
      "theoretical_loss": 4.920407651121222,
      "tokens_seen": 80478208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006867434272744468,
      "loss": 3.623,
      "theoretical_loss": 4.919835292560689,
      "tokens_seen": 80543744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006867268772460753,
      "loss": 3.5926,
      "theoretical_loss": 4.919263529799759,
      "tokens_seen": 80609280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006867103272177038,
      "loss": 3.5255,
      "theoretical_loss": 4.918692361734598,
      "tokens_seen": 80674816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006866937771893323,
      "loss": 3.7263,
      "theoretical_loss": 4.91812178726431,
      "tokens_seen": 80740352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006866772271609608,
      "loss": 3.5766,
      "theoretical_loss": 4.917551805290929,
      "tokens_seen": 80805888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006866606771325893,
      "loss": 3.5895,
      "theoretical_loss": 4.916982414719408,
      "tokens_seen": 80871424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006866441271042178,
      "loss": 3.6744,
      "theoretical_loss": 4.9164136144576105,
      "tokens_seen": 80936960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006866275770758464,
      "loss": 3.6662,
      "theoretical_loss": 4.915845403416299,
      "tokens_seen": 81002496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006866110270474749,
      "loss": 3.7778,
      "theoretical_loss": 4.915277780509124,
      "tokens_seen": 81068032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006865944770191035,
      "loss": 3.5948,
      "theoretical_loss": 4.914710744652614,
      "tokens_seen": 81133568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000686577926990732,
      "loss": 3.3938,
      "theoretical_loss": 4.914144294766169,
      "tokens_seen": 81199104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006865613769623604,
      "loss": 3.7815,
      "theoretical_loss": 4.913578429772047,
      "tokens_seen": 81264640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000686544826933989,
      "loss": 3.7276,
      "theoretical_loss": 4.913013148595355,
      "tokens_seen": 81330176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006865282769056175,
      "loss": 3.533,
      "theoretical_loss": 4.912448450164041,
      "tokens_seen": 81395712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000686511726877246,
      "loss": 3.5464,
      "theoretical_loss": 4.91188433340888,
      "tokens_seen": 81461248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006864951768488746,
      "loss": 3.5504,
      "theoretical_loss": 4.911320797263471,
      "tokens_seen": 81526784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006864786268205031,
      "loss": 3.6409,
      "theoretical_loss": 4.910757840664219,
      "tokens_seen": 81592320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006864620767921317,
      "loss": 3.3887,
      "theoretical_loss": 4.910195462550334,
      "tokens_seen": 81657856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006864455267637602,
      "loss": 3.48,
      "theoretical_loss": 4.909633661863811,
      "tokens_seen": 81723392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006864289767353886,
      "loss": 3.6,
      "theoretical_loss": 4.909072437549434,
      "tokens_seen": 81788928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006864124267070172,
      "loss": 3.3673,
      "theoretical_loss": 4.908511788554753,
      "tokens_seen": 81854464
    },
    {
      "epoch": 0.03,
      "objective/train/avg_token_score": 0.03725932911038399,
      "objective/train/avg_weight": 0.9561554789543152,
      "objective/train/docs_used": 57273,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.0814690589904785,
      "objective/train/std_weight": 0.17060506343841553,
      "objective/train/theoretical_loss": 4.907951713830082,
      "objective/train/tokens_used": 102380000,
      "theoretical_loss": 4.907951713830082,
      "tokens_seen": 81920000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006863958766786457,
      "loss": 3.7398,
      "theoretical_loss": 4.907951713830082,
      "tokens_seen": 81920000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006863793266502742,
      "loss": 3.5007,
      "theoretical_loss": 4.907392212328489,
      "tokens_seen": 81985536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006863627766219028,
      "loss": 3.7045,
      "theoretical_loss": 4.906833283005785,
      "tokens_seen": 82051072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006863462265935313,
      "loss": 3.5731,
      "theoretical_loss": 4.906274924820515,
      "tokens_seen": 82116608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006863296765651599,
      "loss": 3.5126,
      "theoretical_loss": 4.90571713673395,
      "tokens_seen": 82182144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006863131265367883,
      "loss": 3.5197,
      "theoretical_loss": 4.905159917710073,
      "tokens_seen": 82247680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006862965765084168,
      "loss": 3.2709,
      "theoretical_loss": 4.904603266715578,
      "tokens_seen": 82313216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006862800264800454,
      "loss": 3.533,
      "theoretical_loss": 4.904047182719854,
      "tokens_seen": 82378752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006862634764516739,
      "loss": 3.6268,
      "theoretical_loss": 4.903491664694977,
      "tokens_seen": 82444288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006862469264233024,
      "loss": 3.5237,
      "theoretical_loss": 4.902936711615702,
      "tokens_seen": 82509824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000686230376394931,
      "loss": 3.3451,
      "theoretical_loss": 4.902382322459456,
      "tokens_seen": 82575360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006862138263665595,
      "loss": 3.5681,
      "theoretical_loss": 4.901828496206322,
      "tokens_seen": 82640896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000686197276338188,
      "loss": 3.5942,
      "theoretical_loss": 4.90127523183904,
      "tokens_seen": 82706432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006861807263098165,
      "loss": 3.609,
      "theoretical_loss": 4.900722528342988,
      "tokens_seen": 82771968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000686164176281445,
      "loss": 3.1466,
      "theoretical_loss": 4.900170384706181,
      "tokens_seen": 82837504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006861476262530736,
      "loss": 3.3287,
      "theoretical_loss": 4.899618799919256,
      "tokens_seen": 82903040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006861310762247021,
      "loss": 3.6306,
      "theoretical_loss": 4.899067772975469,
      "tokens_seen": 82968576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006861145261963306,
      "loss": 3.4636,
      "theoretical_loss": 4.898517302870679,
      "tokens_seen": 83034112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006860979761679592,
      "loss": 3.6707,
      "theoretical_loss": 4.897967388603346,
      "tokens_seen": 83099648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006860814261395877,
      "loss": 3.4774,
      "theoretical_loss": 4.897418029174519,
      "tokens_seen": 83165184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006860648761112162,
      "loss": 3.4464,
      "theoretical_loss": 4.896869223587828,
      "tokens_seen": 83230720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006860483260828447,
      "loss": 3.749,
      "theoretical_loss": 4.896320970849472,
      "tokens_seen": 83296256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006860317760544732,
      "loss": 3.6061,
      "theoretical_loss": 4.895773269968219,
      "tokens_seen": 83361792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006860152260261018,
      "loss": 3.6093,
      "theoretical_loss": 4.895226119955386,
      "tokens_seen": 83427328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006859986759977303,
      "loss": 3.6034,
      "theoretical_loss": 4.894679519824841,
      "tokens_seen": 83492864
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006859821259693588,
      "loss": 3.4276,
      "theoretical_loss": 4.894133468592984,
      "tokens_seen": 83558400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006859655759409874,
      "loss": 3.7008,
      "theoretical_loss": 4.8935879652787495,
      "tokens_seen": 83623936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006859490259126158,
      "loss": 3.7126,
      "theoretical_loss": 4.893043008903591,
      "tokens_seen": 83689472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006859324758842444,
      "loss": 3.6456,
      "theoretical_loss": 4.892498598491473,
      "tokens_seen": 83755008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006859159258558729,
      "loss": 3.5704,
      "theoretical_loss": 4.891954733068863,
      "tokens_seen": 83820544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006858993758275014,
      "loss": 3.6004,
      "theoretical_loss": 4.891411411664727,
      "tokens_seen": 83886080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00068588282579913,
      "loss": 3.2014,
      "theoretical_loss": 4.890868633310515,
      "tokens_seen": 83951616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006858662757707585,
      "loss": 3.3478,
      "theoretical_loss": 4.890326397040158,
      "tokens_seen": 84017152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000685849725742387,
      "loss": 3.6477,
      "theoretical_loss": 4.889784701890056,
      "tokens_seen": 84082688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006858331757140156,
      "loss": 3.5796,
      "theoretical_loss": 4.8892435468990705,
      "tokens_seen": 84148224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000685816625685644,
      "loss": 3.4562,
      "theoretical_loss": 4.88870293110852,
      "tokens_seen": 84213760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006858000756572726,
      "loss": 3.4888,
      "theoretical_loss": 4.888162853562166,
      "tokens_seen": 84279296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006857835256289011,
      "loss": 3.907,
      "theoretical_loss": 4.88762331330621,
      "tokens_seen": 84344832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006857669756005296,
      "loss": 3.3962,
      "theoretical_loss": 4.88708430938928,
      "tokens_seen": 84410368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006857504255721582,
      "loss": 3.6206,
      "theoretical_loss": 4.8865458408624285,
      "tokens_seen": 84475904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006857338755437866,
      "loss": 3.4882,
      "theoretical_loss": 4.8860079067791204,
      "tokens_seen": 84541440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006857173255154152,
      "loss": 3.3364,
      "theoretical_loss": 4.885470506195227,
      "tokens_seen": 84606976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006857007754870436,
      "loss": 3.6288,
      "theoretical_loss": 4.884933638169014,
      "tokens_seen": 84672512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006856842254586721,
      "loss": 3.523,
      "theoretical_loss": 4.88439730176114,
      "tokens_seen": 84738048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006856676754303007,
      "loss": 3.4849,
      "theoretical_loss": 4.883861496034644,
      "tokens_seen": 84803584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006856511254019292,
      "loss": 3.5019,
      "theoretical_loss": 4.88332622005494,
      "tokens_seen": 84869120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006856345753735578,
      "loss": 3.7374,
      "theoretical_loss": 4.8827914728898065,
      "tokens_seen": 84934656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006856180253451863,
      "loss": 3.4323,
      "theoretical_loss": 4.88225725360938,
      "tokens_seen": 85000192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006856014753168148,
      "loss": 3.4762,
      "theoretical_loss": 4.881723561286149,
      "tokens_seen": 85065728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006855849252884433,
      "loss": 3.4858,
      "theoretical_loss": 4.881190394994943,
      "tokens_seen": 85131264
    },
    {
      "epoch": 0.03,
      "objective/train/avg_token_score": 0.02042163908481598,
      "objective/train/avg_weight": 0.9759610891342163,
      "objective/train/docs_used": 58950,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3498282432556152,
      "objective/train/std_weight": 0.0983477458357811,
      "objective/train/theoretical_loss": 4.880657753812926,
      "objective/train/tokens_used": 105656800,
      "theoretical_loss": 4.880657753812926,
      "tokens_seen": 85196800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006855683752600718,
      "loss": 3.3724,
      "theoretical_loss": 4.880657753812926,
      "tokens_seen": 85196800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006855518252317003,
      "loss": 3.6657,
      "theoretical_loss": 4.880125636819594,
      "tokens_seen": 85262336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006855352752033289,
      "loss": 3.6941,
      "theoretical_loss": 4.879594043096755,
      "tokens_seen": 85327872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006855187251749574,
      "loss": 3.7739,
      "theoretical_loss": 4.879062971728534,
      "tokens_seen": 85393408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000685502175146586,
      "loss": 3.6545,
      "theoretical_loss": 4.87853242180136,
      "tokens_seen": 85458944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006854856251182145,
      "loss": 3.4227,
      "theoretical_loss": 4.878002392403959,
      "tokens_seen": 85524480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000685469075089843,
      "loss": 3.5688,
      "theoretical_loss": 4.877472882627343,
      "tokens_seen": 85590016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006854525250614715,
      "loss": 3.6964,
      "theoretical_loss": 4.8769438915648085,
      "tokens_seen": 85655552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006854359750331,
      "loss": 3.3415,
      "theoretical_loss": 4.876415418311928,
      "tokens_seen": 85721088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006854194250047286,
      "loss": 3.6659,
      "theoretical_loss": 4.875887461966537,
      "tokens_seen": 85786624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006854028749763571,
      "loss": 3.8205,
      "theoretical_loss": 4.875360021628733,
      "tokens_seen": 85852160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006853863249479856,
      "loss": 3.6497,
      "theoretical_loss": 4.874833096400865,
      "tokens_seen": 85917696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006853697749196142,
      "loss": 3.5826,
      "theoretical_loss": 4.874306685387525,
      "tokens_seen": 85983232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006853532248912427,
      "loss": 3.4348,
      "theoretical_loss": 4.873780787695547,
      "tokens_seen": 86048768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006853366748628711,
      "loss": 3.5359,
      "theoretical_loss": 4.87325540243399,
      "tokens_seen": 86114304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006853201248344997,
      "loss": 3.6071,
      "theoretical_loss": 4.872730528714139,
      "tokens_seen": 86179840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006853035748061282,
      "loss": 3.3871,
      "theoretical_loss": 4.872206165649493,
      "tokens_seen": 86245376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006852870247777568,
      "loss": 3.5959,
      "theoretical_loss": 4.871682312355761,
      "tokens_seen": 86310912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006852704747493853,
      "loss": 3.5889,
      "theoretical_loss": 4.871158967950852,
      "tokens_seen": 86376448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006852539247210138,
      "loss": 3.1002,
      "theoretical_loss": 4.870636131554869,
      "tokens_seen": 86441984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006852373746926424,
      "loss": 3.3298,
      "theoretical_loss": 4.8701138022901045,
      "tokens_seen": 86507520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006852208246642708,
      "loss": 3.6435,
      "theoretical_loss": 4.869591979281028,
      "tokens_seen": 86573056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006852042746358993,
      "loss": 3.5984,
      "theoretical_loss": 4.8690706616542805,
      "tokens_seen": 86638592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006851877246075279,
      "loss": 3.5643,
      "theoretical_loss": 4.868549848538675,
      "tokens_seen": 86704128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006851711745791564,
      "loss": 3.665,
      "theoretical_loss": 4.868029539065176,
      "tokens_seen": 86769664
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000685154624550785,
      "loss": 3.4772,
      "theoretical_loss": 4.867509732366907,
      "tokens_seen": 86835200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006851380745224135,
      "loss": 3.7947,
      "theoretical_loss": 4.866990427579129,
      "tokens_seen": 86900736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000685121524494042,
      "loss": 3.2814,
      "theoretical_loss": 4.866471623839248,
      "tokens_seen": 86966272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006851049744656706,
      "loss": 3.5332,
      "theoretical_loss": 4.8659533202867955,
      "tokens_seen": 87031808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000685088424437299,
      "loss": 3.5144,
      "theoretical_loss": 4.86543551606343,
      "tokens_seen": 87097344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006850718744089275,
      "loss": 3.7937,
      "theoretical_loss": 4.864918210312927,
      "tokens_seen": 87162880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006850553243805561,
      "loss": 3.5001,
      "theoretical_loss": 4.864401402181173,
      "tokens_seen": 87228416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006850387743521846,
      "loss": 3.3351,
      "theoretical_loss": 4.863885090816158,
      "tokens_seen": 87293952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006850222243238132,
      "loss": 3.2574,
      "theoretical_loss": 4.863369275367968,
      "tokens_seen": 87359488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006850056742954417,
      "loss": 3.7066,
      "theoretical_loss": 4.862853954988781,
      "tokens_seen": 87425024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006849891242670702,
      "loss": 3.5393,
      "theoretical_loss": 4.862339128832857,
      "tokens_seen": 87490560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006849725742386987,
      "loss": 3.5543,
      "theoretical_loss": 4.861824796056533,
      "tokens_seen": 87556096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006849560242103272,
      "loss": 3.5436,
      "theoretical_loss": 4.861310955818219,
      "tokens_seen": 87621632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006849394741819557,
      "loss": 3.7373,
      "theoretical_loss": 4.860797607278385,
      "tokens_seen": 87687168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006849229241535843,
      "loss": 3.258,
      "theoretical_loss": 4.86028474959956,
      "tokens_seen": 87752704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006849063741252128,
      "loss": 3.3747,
      "theoretical_loss": 4.859772381946323,
      "tokens_seen": 87818240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006848898240968414,
      "loss": 3.3223,
      "theoretical_loss": 4.859260503485298,
      "tokens_seen": 87883776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006848732740684699,
      "loss": 3.4048,
      "theoretical_loss": 4.858749113385144,
      "tokens_seen": 87949312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006848567240400983,
      "loss": 3.4787,
      "theoretical_loss": 4.858238210816554,
      "tokens_seen": 88014848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006848401740117269,
      "loss": 3.5143,
      "theoretical_loss": 4.8577277949522415,
      "tokens_seen": 88080384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006848236239833554,
      "loss": 3.5868,
      "theoretical_loss": 4.857217864966943,
      "tokens_seen": 88145920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006848070739549839,
      "loss": 3.6964,
      "theoretical_loss": 4.856708420037402,
      "tokens_seen": 88211456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006847905239266125,
      "loss": 3.4618,
      "theoretical_loss": 4.8561994593423705,
      "tokens_seen": 88276992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000684773973898241,
      "loss": 3.2704,
      "theoretical_loss": 4.8556909820625975,
      "tokens_seen": 88342528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006847574238698695,
      "loss": 3.4807,
      "theoretical_loss": 4.855182987380823,
      "tokens_seen": 88408064
    },
    {
      "epoch": 0.03,
      "objective/train/avg_token_score": 0.017664186656475067,
      "objective/train/avg_weight": 0.9792220592498779,
      "objective/train/docs_used": 60915,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.430917263031006,
      "objective/train/std_weight": 0.0741313099861145,
      "objective/train/theoretical_loss": 4.854675474481779,
      "objective/train/tokens_used": 108933600,
      "theoretical_loss": 4.854675474481779,
      "tokens_seen": 88473600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000684740873841498,
      "loss": 3.3831,
      "theoretical_loss": 4.854675474481779,
      "tokens_seen": 88473600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006847243238131265,
      "loss": 3.5037,
      "theoretical_loss": 4.8541684425521705,
      "tokens_seen": 88539136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000684707773784755,
      "loss": 3.5658,
      "theoretical_loss": 4.85366189078068,
      "tokens_seen": 88604672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006846912237563835,
      "loss": 3.4308,
      "theoretical_loss": 4.853155818357957,
      "tokens_seen": 88670208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006846746737280121,
      "loss": 3.4933,
      "theoretical_loss": 4.852650224476609,
      "tokens_seen": 88735744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006846581236996406,
      "loss": 3.5777,
      "theoretical_loss": 4.852145108331205,
      "tokens_seen": 88801280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006846415736712691,
      "loss": 3.3987,
      "theoretical_loss": 4.851640469118255,
      "tokens_seen": 88866816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006846250236428977,
      "loss": 3.4603,
      "theoretical_loss": 4.851136306036219,
      "tokens_seen": 88932352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006846084736145261,
      "loss": 3.5885,
      "theoretical_loss": 4.850632618285486,
      "tokens_seen": 88997888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006845919235861547,
      "loss": 3.1274,
      "theoretical_loss": 4.850129405068383,
      "tokens_seen": 89063424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006845753735577832,
      "loss": 3.443,
      "theoretical_loss": 4.849626665589156,
      "tokens_seen": 89128960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006845588235294117,
      "loss": 3.181,
      "theoretical_loss": 4.849124399053969,
      "tokens_seen": 89194496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006845422735010403,
      "loss": 3.387,
      "theoretical_loss": 4.8486226046709024,
      "tokens_seen": 89260032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006845257234726688,
      "loss": 3.396,
      "theoretical_loss": 4.8481212816499415,
      "tokens_seen": 89325568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006845091734442973,
      "loss": 3.7163,
      "theoretical_loss": 4.847620429202967,
      "tokens_seen": 89391104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006844926234159258,
      "loss": 3.2938,
      "theoretical_loss": 4.847120046543763,
      "tokens_seen": 89456640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006844760733875543,
      "loss": 3.521,
      "theoretical_loss": 4.846620132887992,
      "tokens_seen": 89522176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006844595233591829,
      "loss": 3.5584,
      "theoretical_loss": 4.8461206874532055,
      "tokens_seen": 89587712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006844429733308114,
      "loss": 3.3438,
      "theoretical_loss": 4.845621709458831,
      "tokens_seen": 89653248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006844264233024399,
      "loss": 3.2269,
      "theoretical_loss": 4.845123198126162,
      "tokens_seen": 89718784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006844098732740685,
      "loss": 3.4366,
      "theoretical_loss": 4.844625152678364,
      "tokens_seen": 89784320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000684393323245697,
      "loss": 3.277,
      "theoretical_loss": 4.844127572340455,
      "tokens_seen": 89849856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006843767732173254,
      "loss": 3.753,
      "theoretical_loss": 4.84363045633931,
      "tokens_seen": 89915392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000684360223188954,
      "loss": 3.4894,
      "theoretical_loss": 4.843133803903651,
      "tokens_seen": 89980928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006843436731605825,
      "loss": 3.1454,
      "theoretical_loss": 4.84263761426404,
      "tokens_seen": 90046464
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006843271231322111,
      "loss": 3.0453,
      "theoretical_loss": 4.842141886652876,
      "tokens_seen": 90112000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006843105731038396,
      "loss": 3.4578,
      "theoretical_loss": 4.841646620304388,
      "tokens_seen": 90177536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006842940230754681,
      "loss": 3.4765,
      "theoretical_loss": 4.841151814454632,
      "tokens_seen": 90243072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006842774730470967,
      "loss": 3.3346,
      "theoretical_loss": 4.840657468341476,
      "tokens_seen": 90308608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006842609230187252,
      "loss": 3.5955,
      "theoretical_loss": 4.84016358120461,
      "tokens_seen": 90374144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006842443729903536,
      "loss": 3.6155,
      "theoretical_loss": 4.839670152285526,
      "tokens_seen": 90439680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006842278229619822,
      "loss": 3.3196,
      "theoretical_loss": 4.8391771808275195,
      "tokens_seen": 90505216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006842112729336107,
      "loss": 3.2122,
      "theoretical_loss": 4.838684666075682,
      "tokens_seen": 90570752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006841947229052393,
      "loss": 3.4673,
      "theoretical_loss": 4.838192607276896,
      "tokens_seen": 90636288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006841781728768678,
      "loss": 3.3446,
      "theoretical_loss": 4.837701003679829,
      "tokens_seen": 90701824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006841616228484963,
      "loss": 3.5423,
      "theoretical_loss": 4.8372098545349305,
      "tokens_seen": 90767360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006841450728201249,
      "loss": 3.493,
      "theoretical_loss": 4.836719159094422,
      "tokens_seen": 90832896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006841285227917533,
      "loss": 3.3207,
      "theoretical_loss": 4.836228916612292,
      "tokens_seen": 90898432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006841119727633818,
      "loss": 3.6016,
      "theoretical_loss": 4.835739126344298,
      "tokens_seen": 90963968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006840954227350104,
      "loss": 3.502,
      "theoretical_loss": 4.8352497875479505,
      "tokens_seen": 91029504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006840788727066389,
      "loss": 3.2982,
      "theoretical_loss": 4.834760899482514,
      "tokens_seen": 91095040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006840623226782675,
      "loss": 3.4451,
      "theoretical_loss": 4.834272461409001,
      "tokens_seen": 91160576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000684045772649896,
      "loss": 3.3806,
      "theoretical_loss": 4.833784472590165,
      "tokens_seen": 91226112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006840292226215245,
      "loss": 3.2449,
      "theoretical_loss": 4.833296932290495,
      "tokens_seen": 91291648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006840126725931531,
      "loss": 3.3548,
      "theoretical_loss": 4.832809839776213,
      "tokens_seen": 91357184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006839961225647815,
      "loss": 3.3444,
      "theoretical_loss": 4.832323194315265,
      "tokens_seen": 91422720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006839795725364101,
      "loss": 3.4999,
      "theoretical_loss": 4.831836995177319,
      "tokens_seen": 91488256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006839630225080386,
      "loss": 3.5225,
      "theoretical_loss": 4.831351241633756,
      "tokens_seen": 91553792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006839464724796671,
      "loss": 3.5833,
      "theoretical_loss": 4.8308659329576695,
      "tokens_seen": 91619328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006839299224512957,
      "loss": 3.5492,
      "theoretical_loss": 4.830381068423856,
      "tokens_seen": 91684864
    },
    {
      "epoch": 0.03,
      "objective/train/avg_token_score": 0.013729576021432877,
      "objective/train/avg_weight": 0.9838443994522095,
      "objective/train/docs_used": 62413,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.567444324493408,
      "objective/train/std_weight": 0.10627894848585129,
      "objective/train/theoretical_loss": 4.8298966473088125,
      "objective/train/tokens_used": 112210400,
      "theoretical_loss": 4.8298966473088125,
      "tokens_seen": 91750400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006839133724229242,
      "loss": 3.5146,
      "theoretical_loss": 4.8298966473088125,
      "tokens_seen": 91750400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006838968223945527,
      "loss": 3.6577,
      "theoretical_loss": 4.829412668890729,
      "tokens_seen": 91815936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006838802723661812,
      "loss": 3.2537,
      "theoretical_loss": 4.8289291324494865,
      "tokens_seen": 91881472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006838637223378097,
      "loss": 3.4543,
      "theoretical_loss": 4.828446037266647,
      "tokens_seen": 91947008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006838471723094383,
      "loss": 3.4777,
      "theoretical_loss": 4.827963382625454,
      "tokens_seen": 92012544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006838306222810668,
      "loss": 3.5426,
      "theoretical_loss": 4.827481167810825,
      "tokens_seen": 92078080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006838140722526953,
      "loss": 3.3838,
      "theoretical_loss": 4.826999392109344,
      "tokens_seen": 92143616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006837975222243239,
      "loss": 3.6945,
      "theoretical_loss": 4.826518054809259,
      "tokens_seen": 92209152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006837809721959524,
      "loss": 3.3187,
      "theoretical_loss": 4.826037155200478,
      "tokens_seen": 92274688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006837644221675808,
      "loss": 3.5344,
      "theoretical_loss": 4.825556692574562,
      "tokens_seen": 92340224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006837478721392093,
      "loss": 3.7428,
      "theoretical_loss": 4.825076666224717,
      "tokens_seen": 92405760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006837313221108378,
      "loss": 3.5813,
      "theoretical_loss": 4.824597075445799,
      "tokens_seen": 92471296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006837147720824664,
      "loss": 3.493,
      "theoretical_loss": 4.824117919534297,
      "tokens_seen": 92536832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006836982220540949,
      "loss": 3.4876,
      "theoretical_loss": 4.823639197788334,
      "tokens_seen": 92602368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006836816720257235,
      "loss": 3.4847,
      "theoretical_loss": 4.823160909507665,
      "tokens_seen": 92667904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000683665121997352,
      "loss": 3.4209,
      "theoretical_loss": 4.822683053993664,
      "tokens_seen": 92733440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006836485719689804,
      "loss": 3.4711,
      "theoretical_loss": 4.822205630549329,
      "tokens_seen": 92798976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000683632021940609,
      "loss": 3.203,
      "theoretical_loss": 4.821728638479267,
      "tokens_seen": 92864512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006836154719122375,
      "loss": 3.4941,
      "theoretical_loss": 4.821252077089696,
      "tokens_seen": 92930048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000683598921883866,
      "loss": 3.4233,
      "theoretical_loss": 4.820775945688437,
      "tokens_seen": 92995584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006835823718554946,
      "loss": 3.2547,
      "theoretical_loss": 4.820300243584913,
      "tokens_seen": 93061120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006835658218271231,
      "loss": 3.4876,
      "theoretical_loss": 4.819824970090138,
      "tokens_seen": 93126656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006835492717987517,
      "loss": 3.4344,
      "theoretical_loss": 4.819350124516717,
      "tokens_seen": 93192192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006835327217703802,
      "loss": 3.49,
      "theoretical_loss": 4.818875706178841,
      "tokens_seen": 93257728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006835161717420086,
      "loss": 3.307,
      "theoretical_loss": 4.818401714392279,
      "tokens_seen": 93323264
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006834996217136372,
      "loss": 3.3212,
      "theoretical_loss": 4.817928148474378,
      "tokens_seen": 93388800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006834830716852657,
      "loss": 3.2645,
      "theoretical_loss": 4.817455007744052,
      "tokens_seen": 93454336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006834665216568942,
      "loss": 3.4429,
      "theoretical_loss": 4.816982291521785,
      "tokens_seen": 93519872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006834499716285228,
      "loss": 3.5009,
      "theoretical_loss": 4.816509999129618,
      "tokens_seen": 93585408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006834334216001513,
      "loss": 3.5296,
      "theoretical_loss": 4.816038129891151,
      "tokens_seen": 93650944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006834168715717799,
      "loss": 3.4918,
      "theoretical_loss": 4.815566683131536,
      "tokens_seen": 93716480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006834003215434084,
      "loss": 3.6185,
      "theoretical_loss": 4.815095658177472,
      "tokens_seen": 93782016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006833837715150368,
      "loss": 3.2304,
      "theoretical_loss": 4.814625054357199,
      "tokens_seen": 93847552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006833672214866654,
      "loss": 3.1996,
      "theoretical_loss": 4.814154871000497,
      "tokens_seen": 93913088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006833506714582939,
      "loss": 3.6303,
      "theoretical_loss": 4.813685107438679,
      "tokens_seen": 93978624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006833341214299224,
      "loss": 3.175,
      "theoretical_loss": 4.813215763004585,
      "tokens_seen": 94044160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000683317571401551,
      "loss": 3.572,
      "theoretical_loss": 4.812746837032582,
      "tokens_seen": 94109696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006833010213731795,
      "loss": 3.1892,
      "theoretical_loss": 4.812278328858554,
      "tokens_seen": 94175232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006832844713448081,
      "loss": 3.3737,
      "theoretical_loss": 4.811810237819904,
      "tokens_seen": 94240768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006832679213164365,
      "loss": 3.6076,
      "theoretical_loss": 4.81134256325554,
      "tokens_seen": 94306304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000683251371288065,
      "loss": 3.5255,
      "theoretical_loss": 4.810875304505881,
      "tokens_seen": 94371840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006832348212596936,
      "loss": 3.6813,
      "theoretical_loss": 4.810408460912846,
      "tokens_seen": 94437376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006832182712313221,
      "loss": 3.2712,
      "theoretical_loss": 4.809942031819853,
      "tokens_seen": 94502912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006832017212029506,
      "loss": 3.8311,
      "theoretical_loss": 4.809476016571809,
      "tokens_seen": 94568448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006831851711745792,
      "loss": 3.84,
      "theoretical_loss": 4.809010414515113,
      "tokens_seen": 94633984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006831686211462077,
      "loss": 3.3595,
      "theoretical_loss": 4.808545224997644,
      "tokens_seen": 94699520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006831520711178363,
      "loss": 3.3876,
      "theoretical_loss": 4.808080447368766,
      "tokens_seen": 94765056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006831355210894647,
      "loss": 3.2379,
      "theoretical_loss": 4.807616080979315,
      "tokens_seen": 94830592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006831189710610932,
      "loss": 3.5315,
      "theoretical_loss": 4.807152125181597,
      "tokens_seen": 94896128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006831024210327218,
      "loss": 3.2348,
      "theoretical_loss": 4.806688579329387,
      "tokens_seen": 94961664
    },
    {
      "epoch": 0.03,
      "objective/train/avg_token_score": 0.013349258340895176,
      "objective/train/avg_weight": 0.984282374382019,
      "objective/train/docs_used": 64216,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.048253059387207,
      "objective/train/std_weight": 0.08914466947317123,
      "objective/train/theoretical_loss": 4.8062254427779205,
      "objective/train/tokens_used": 115487200,
      "theoretical_loss": 4.8062254427779205,
      "tokens_seen": 95027200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006830858710043503,
      "loss": 3.6868,
      "theoretical_loss": 4.8062254427779205,
      "tokens_seen": 95027200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006830693209759788,
      "loss": 3.3632,
      "theoretical_loss": 4.80576271488389,
      "tokens_seen": 95092736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006830527709476074,
      "loss": 3.306,
      "theoretical_loss": 4.805300395005444,
      "tokens_seen": 95158272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006830362209192359,
      "loss": 3.4065,
      "theoretical_loss": 4.804838482502181,
      "tokens_seen": 95223808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006830196708908644,
      "loss": 3.4275,
      "theoretical_loss": 4.8043769767351385,
      "tokens_seen": 95289344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006830031208624929,
      "loss": 3.5507,
      "theoretical_loss": 4.8039158770668005,
      "tokens_seen": 95354880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006829865708341214,
      "loss": 3.3524,
      "theoretical_loss": 4.803455182861087,
      "tokens_seen": 95420416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00068297002080575,
      "loss": 3.4962,
      "theoretical_loss": 4.802994893483348,
      "tokens_seen": 95485952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006829534707773785,
      "loss": 3.5186,
      "theoretical_loss": 4.802535008300364,
      "tokens_seen": 95551488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000682936920749007,
      "loss": 3.3367,
      "theoretical_loss": 4.802075526680335,
      "tokens_seen": 95617024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006829203707206356,
      "loss": 3.1991,
      "theoretical_loss": 4.801616447992888,
      "tokens_seen": 95682560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000682903820692264,
      "loss": 3.3588,
      "theoretical_loss": 4.801157771609061,
      "tokens_seen": 95748096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006828872706638926,
      "loss": 3.4808,
      "theoretical_loss": 4.8006994969013,
      "tokens_seen": 95813632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006828707206355211,
      "loss": 3.2459,
      "theoretical_loss": 4.800241623243467,
      "tokens_seen": 95879168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006828541706071496,
      "loss": 3.5291,
      "theoretical_loss": 4.799784150010819,
      "tokens_seen": 95944704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006828376205787782,
      "loss": 3.4193,
      "theoretical_loss": 4.799327076580017,
      "tokens_seen": 96010240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006828210705504067,
      "loss": 3.5068,
      "theoretical_loss": 4.798870402329115,
      "tokens_seen": 96075776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006828045205220352,
      "loss": 3.3694,
      "theoretical_loss": 4.798414126637558,
      "tokens_seen": 96141312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006827879704936636,
      "loss": 3.4897,
      "theoretical_loss": 4.797958248886179,
      "tokens_seen": 96206848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006827714204652922,
      "loss": 3.4107,
      "theoretical_loss": 4.797502768457193,
      "tokens_seen": 96272384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006827548704369207,
      "loss": 3.3764,
      "theoretical_loss": 4.797047684734192,
      "tokens_seen": 96337920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006827383204085492,
      "loss": 3.2787,
      "theoretical_loss": 4.796592997102147,
      "tokens_seen": 96403456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006827217703801778,
      "loss": 3.5012,
      "theoretical_loss": 4.796138704947397,
      "tokens_seen": 96468992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006827052203518063,
      "loss": 3.4531,
      "theoretical_loss": 4.795684807657649,
      "tokens_seen": 96534528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006826886703234348,
      "loss": 3.1963,
      "theoretical_loss": 4.795231304621968,
      "tokens_seen": 96600064
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006826721202950634,
      "loss": 3.2685,
      "theoretical_loss": 4.794778195230787,
      "tokens_seen": 96665600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006826555702666918,
      "loss": 3.4672,
      "theoretical_loss": 4.794325478875885,
      "tokens_seen": 96731136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006826390202383204,
      "loss": 3.6089,
      "theoretical_loss": 4.793873154950399,
      "tokens_seen": 96796672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006826224702099489,
      "loss": 3.5225,
      "theoretical_loss": 4.793421222848808,
      "tokens_seen": 96862208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006826059201815774,
      "loss": 3.412,
      "theoretical_loss": 4.7929696819669365,
      "tokens_seen": 96927744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000682589370153206,
      "loss": 3.3439,
      "theoretical_loss": 4.792518531701948,
      "tokens_seen": 96993280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006825728201248345,
      "loss": 3.3691,
      "theoretical_loss": 4.792067771452341,
      "tokens_seen": 97058816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000682556270096463,
      "loss": 3.3195,
      "theoretical_loss": 4.791617400617948,
      "tokens_seen": 97124352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006825397200680915,
      "loss": 3.4026,
      "theoretical_loss": 4.791167418599925,
      "tokens_seen": 97189888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00068252317003972,
      "loss": 3.2916,
      "theoretical_loss": 4.790717824800755,
      "tokens_seen": 97255424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006825066200113486,
      "loss": 3.4642,
      "theoretical_loss": 4.790268618624239,
      "tokens_seen": 97320960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006824900699829771,
      "loss": 3.3448,
      "theoretical_loss": 4.789819799475499,
      "tokens_seen": 97386496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006824735199546056,
      "loss": 3.478,
      "theoretical_loss": 4.789371366760961,
      "tokens_seen": 97452032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006824569699262342,
      "loss": 3.6651,
      "theoretical_loss": 4.788923319888369,
      "tokens_seen": 97517568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006824404198978627,
      "loss": 3.4305,
      "theoretical_loss": 4.788475658266766,
      "tokens_seen": 97583104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006824238698694911,
      "loss": 3.0006,
      "theoretical_loss": 4.788028381306497,
      "tokens_seen": 97648640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006824073198411197,
      "loss": 3.5459,
      "theoretical_loss": 4.787581488419207,
      "tokens_seen": 97714176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006823907698127482,
      "loss": 3.5026,
      "theoretical_loss": 4.787134979017832,
      "tokens_seen": 97779712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006823742197843768,
      "loss": 3.3352,
      "theoretical_loss": 4.786688852516599,
      "tokens_seen": 97845248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006823576697560053,
      "loss": 3.2877,
      "theoretical_loss": 4.786243108331024,
      "tokens_seen": 97910784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0006823411197276338,
      "loss": 3.4104,
      "theoretical_loss": 4.7857977458779,
      "tokens_seen": 97976320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006823245696992624,
      "loss": 3.4694,
      "theoretical_loss": 4.785352764575304,
      "tokens_seen": 98041856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006823080196708909,
      "loss": 3.2699,
      "theoretical_loss": 4.784908163842585,
      "tokens_seen": 98107392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006822914696425193,
      "loss": 3.5738,
      "theoretical_loss": 4.784463943100367,
      "tokens_seen": 98172928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006822749196141479,
      "loss": 3.5712,
      "theoretical_loss": 4.7840201017705395,
      "tokens_seen": 98238464
    },
    {
      "epoch": 0.04,
      "objective/train/avg_token_score": 0.016760189086198807,
      "objective/train/avg_weight": 0.9802879691123962,
      "objective/train/docs_used": 66562,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2249441146850586,
      "objective/train/std_weight": 0.08385981619358063,
      "objective/train/theoretical_loss": 4.783576639276257,
      "objective/train/tokens_used": 118764000,
      "theoretical_loss": 4.783576639276257,
      "tokens_seen": 98304000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006822583695857764,
      "loss": 3.2653,
      "theoretical_loss": 4.783576639276257,
      "tokens_seen": 98304000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000682241819557405,
      "loss": 3.4688,
      "theoretical_loss": 4.783133555041934,
      "tokens_seen": 98369536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006822252695290335,
      "loss": 3.5089,
      "theoretical_loss": 4.782690848493245,
      "tokens_seen": 98435072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000682208719500662,
      "loss": 3.8299,
      "theoretical_loss": 4.7822485190571165,
      "tokens_seen": 98500608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006821921694722906,
      "loss": 3.4318,
      "theoretical_loss": 4.781806566161723,
      "tokens_seen": 98566144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000682175619443919,
      "loss": 3.2447,
      "theoretical_loss": 4.781364989236488,
      "tokens_seen": 98631680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006821590694155475,
      "loss": 3.5454,
      "theoretical_loss": 4.78092378771208,
      "tokens_seen": 98697216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006821425193871761,
      "loss": 3.4734,
      "theoretical_loss": 4.780482961020402,
      "tokens_seen": 98762752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006821259693588046,
      "loss": 3.3503,
      "theoretical_loss": 4.780042508594596,
      "tokens_seen": 98828288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006821094193304332,
      "loss": 3.7548,
      "theoretical_loss": 4.779602429869035,
      "tokens_seen": 98893824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006820928693020617,
      "loss": 3.5865,
      "theoretical_loss": 4.779162724279324,
      "tokens_seen": 98959360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006820763192736902,
      "loss": 3.4755,
      "theoretical_loss": 4.7787233912622895,
      "tokens_seen": 99024896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006820597692453188,
      "loss": 3.3081,
      "theoretical_loss": 4.778284430255981,
      "tokens_seen": 99090432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006820432192169472,
      "loss": 3.2137,
      "theoretical_loss": 4.77784584069967,
      "tokens_seen": 99155968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006820266691885757,
      "loss": 3.4775,
      "theoretical_loss": 4.777407622033838,
      "tokens_seen": 99221504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006820101191602043,
      "loss": 3.2992,
      "theoretical_loss": 4.776969773700181,
      "tokens_seen": 99287040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006819935691318328,
      "loss": 3.4413,
      "theoretical_loss": 4.776532295141601,
      "tokens_seen": 99352576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006819770191034614,
      "loss": 3.2457,
      "theoretical_loss": 4.776095185802211,
      "tokens_seen": 99418112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006819604690750899,
      "loss": 3.4575,
      "theoretical_loss": 4.775658445127318,
      "tokens_seen": 99483648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006819439190467184,
      "loss": 3.287,
      "theoretical_loss": 4.775222072563429,
      "tokens_seen": 99549184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006819273690183469,
      "loss": 3.3679,
      "theoretical_loss": 4.7747860675582485,
      "tokens_seen": 99614720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006819108189899754,
      "loss": 3.5858,
      "theoretical_loss": 4.77435042956067,
      "tokens_seen": 99680256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006818942689616039,
      "loss": 3.582,
      "theoretical_loss": 4.773915158020776,
      "tokens_seen": 99745792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006818777189332325,
      "loss": 3.4494,
      "theoretical_loss": 4.773480252389831,
      "tokens_seen": 99811328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000681861168904861,
      "loss": 3.4353,
      "theoretical_loss": 4.773045712120284,
      "tokens_seen": 99876864
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006818446188764896,
      "loss": 3.5594,
      "theoretical_loss": 4.77261153666576,
      "tokens_seen": 99942400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006818280688481181,
      "loss": 3.5072,
      "theoretical_loss": 4.772177725481062,
      "tokens_seen": 100007936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006818115188197465,
      "loss": 3.492,
      "theoretical_loss": 4.77174427802216,
      "tokens_seen": 100073472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000681794968791375,
      "loss": 3.4573,
      "theoretical_loss": 4.771311193746191,
      "tokens_seen": 100139008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006817784187630035,
      "loss": 3.395,
      "theoretical_loss": 4.770878472111465,
      "tokens_seen": 100204544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006817618687346321,
      "loss": 3.2948,
      "theoretical_loss": 4.770446112577445,
      "tokens_seen": 100270080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006817453187062606,
      "loss": 3.3892,
      "theoretical_loss": 4.770014114604756,
      "tokens_seen": 100335616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006817287686778891,
      "loss": 3.79,
      "theoretical_loss": 4.769582477655177,
      "tokens_seen": 100401152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006817122186495177,
      "loss": 3.5211,
      "theoretical_loss": 4.769151201191641,
      "tokens_seen": 100466688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006816956686211461,
      "loss": 3.3944,
      "theoretical_loss": 4.768720284678228,
      "tokens_seen": 100532224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006816791185927747,
      "loss": 3.3523,
      "theoretical_loss": 4.768289727580161,
      "tokens_seen": 100597760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006816625685644032,
      "loss": 3.2406,
      "theoretical_loss": 4.767859529363809,
      "tokens_seen": 100663296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006816460185360317,
      "loss": 3.7703,
      "theoretical_loss": 4.767429689496682,
      "tokens_seen": 100728832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006816294685076603,
      "loss": 3.4608,
      "theoretical_loss": 4.767000207447417,
      "tokens_seen": 100794368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006816129184792888,
      "loss": 3.7484,
      "theoretical_loss": 4.766571082685794,
      "tokens_seen": 100859904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006815963684509173,
      "loss": 3.1587,
      "theoretical_loss": 4.766142314682716,
      "tokens_seen": 100925440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006815798184225459,
      "loss": 3.5163,
      "theoretical_loss": 4.765713902910214,
      "tokens_seen": 100990976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006815632683941743,
      "loss": 3.0615,
      "theoretical_loss": 4.765285846841444,
      "tokens_seen": 101056512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006815467183658029,
      "loss": 3.3799,
      "theoretical_loss": 4.76485814595068,
      "tokens_seen": 101122048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006815301683374314,
      "loss": 3.351,
      "theoretical_loss": 4.764430799713314,
      "tokens_seen": 101187584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006815136183090599,
      "loss": 3.3401,
      "theoretical_loss": 4.764003807605853,
      "tokens_seen": 101253120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006814970682806885,
      "loss": 3.5429,
      "theoretical_loss": 4.763577169105912,
      "tokens_seen": 101318656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000681480518252317,
      "loss": 3.1936,
      "theoretical_loss": 4.763150883692218,
      "tokens_seen": 101384192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006814639682239456,
      "loss": 3.5129,
      "theoretical_loss": 4.762724950844598,
      "tokens_seen": 101449728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000681447418195574,
      "loss": 3.3457,
      "theoretical_loss": 4.762299370043984,
      "tokens_seen": 101515264
    },
    {
      "epoch": 0.04,
      "objective/train/avg_token_score": 0.033091478049755096,
      "objective/train/avg_weight": 0.9610720872879028,
      "objective/train/docs_used": 67858,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.037550449371338,
      "objective/train/std_weight": 0.1469828337430954,
      "objective/train/theoretical_loss": 4.761874140772408,
      "objective/train/tokens_used": 122040800,
      "theoretical_loss": 4.761874140772408,
      "tokens_seen": 101580800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006814308681672025,
      "loss": 3.2426,
      "theoretical_loss": 4.761874140772408,
      "tokens_seen": 101580800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006814143181388311,
      "loss": 3.4807,
      "theoretical_loss": 4.761449262512993,
      "tokens_seen": 101646336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006813977681104596,
      "loss": 3.463,
      "theoretical_loss": 4.761024734749958,
      "tokens_seen": 101711872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006813812180820881,
      "loss": 3.6557,
      "theoretical_loss": 4.76060055696861,
      "tokens_seen": 101777408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006813646680537167,
      "loss": 3.5392,
      "theoretical_loss": 4.760176728655345,
      "tokens_seen": 101842944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006813481180253452,
      "loss": 3.3912,
      "theoretical_loss": 4.75975324929764,
      "tokens_seen": 101908480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006813315679969738,
      "loss": 3.2869,
      "theoretical_loss": 4.759330118384053,
      "tokens_seen": 101974016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006813150179686022,
      "loss": 3.3123,
      "theoretical_loss": 4.758907335404221,
      "tokens_seen": 102039552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006812984679402307,
      "loss": 3.593,
      "theoretical_loss": 4.758484899848854,
      "tokens_seen": 102105088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006812819179118593,
      "loss": 3.6043,
      "theoretical_loss": 4.7580628112097365,
      "tokens_seen": 102170624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006812653678834878,
      "loss": 3.4618,
      "theoretical_loss": 4.7576410689797175,
      "tokens_seen": 102236160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006812488178551163,
      "loss": 3.3889,
      "theoretical_loss": 4.757219672652717,
      "tokens_seen": 102301696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006812322678267449,
      "loss": 3.4028,
      "theoretical_loss": 4.756798621723712,
      "tokens_seen": 102367232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006812157177983734,
      "loss": 3.2917,
      "theoretical_loss": 4.756377915688748,
      "tokens_seen": 102432768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000681199167770002,
      "loss": 3.4656,
      "theoretical_loss": 4.755957554044917,
      "tokens_seen": 102498304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006811826177416304,
      "loss": 3.4789,
      "theoretical_loss": 4.755537536290373,
      "tokens_seen": 102563840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006811660677132589,
      "loss": 3.4085,
      "theoretical_loss": 4.755117861924321,
      "tokens_seen": 102629376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006811495176848875,
      "loss": 3.3001,
      "theoretical_loss": 4.754698530447009,
      "tokens_seen": 102694912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000681132967656516,
      "loss": 3.5567,
      "theoretical_loss": 4.754279541359738,
      "tokens_seen": 102760448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006811164176281445,
      "loss": 3.4275,
      "theoretical_loss": 4.753860894164845,
      "tokens_seen": 102825984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006810998675997731,
      "loss": 3.2401,
      "theoretical_loss": 4.75344258836571,
      "tokens_seen": 102891520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006810833175714015,
      "loss": 3.5437,
      "theoretical_loss": 4.753024623466752,
      "tokens_seen": 102957056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006810667675430301,
      "loss": 3.3504,
      "theoretical_loss": 4.752606998973421,
      "tokens_seen": 103022592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006810502175146586,
      "loss": 3.2456,
      "theoretical_loss": 4.752189714392202,
      "tokens_seen": 103088128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006810336674862871,
      "loss": 3.5934,
      "theoretical_loss": 4.7517727692306035,
      "tokens_seen": 103153664
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006810171174579157,
      "loss": 3.5144,
      "theoretical_loss": 4.751356162997164,
      "tokens_seen": 103219200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006810005674295442,
      "loss": 3.414,
      "theoretical_loss": 4.750939895201443,
      "tokens_seen": 103284736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006809840174011727,
      "loss": 3.5823,
      "theoretical_loss": 4.750523965354024,
      "tokens_seen": 103350272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006809674673728013,
      "loss": 3.4069,
      "theoretical_loss": 4.750108372966501,
      "tokens_seen": 103415808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006809509173444297,
      "loss": 3.5531,
      "theoretical_loss": 4.749693117551491,
      "tokens_seen": 103481344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006809343673160583,
      "loss": 3.3942,
      "theoretical_loss": 4.749278198622617,
      "tokens_seen": 103546880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006809178172876868,
      "loss": 3.5112,
      "theoretical_loss": 4.748863615694514,
      "tokens_seen": 103612416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006809012672593153,
      "loss": 3.5657,
      "theoretical_loss": 4.748449368282822,
      "tokens_seen": 103677952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006808847172309439,
      "loss": 3.4296,
      "theoretical_loss": 4.748035455904185,
      "tokens_seen": 103743488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006808681672025724,
      "loss": 3.4556,
      "theoretical_loss": 4.747621878076252,
      "tokens_seen": 103809024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006808516171742009,
      "loss": 3.3401,
      "theoretical_loss": 4.747208634317664,
      "tokens_seen": 103874560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006808350671458294,
      "loss": 3.5101,
      "theoretical_loss": 4.746795724148061,
      "tokens_seen": 103940096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006808185171174578,
      "loss": 3.4798,
      "theoretical_loss": 4.746383147088078,
      "tokens_seen": 104005632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006808019670890864,
      "loss": 3.4844,
      "theoretical_loss": 4.745970902659338,
      "tokens_seen": 104071168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006807854170607149,
      "loss": 3.5339,
      "theoretical_loss": 4.745558990384451,
      "tokens_seen": 104136704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006807688670323435,
      "loss": 3.1863,
      "theoretical_loss": 4.7451474097870125,
      "tokens_seen": 104202240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000680752317003972,
      "loss": 3.1652,
      "theoretical_loss": 4.744736160391602,
      "tokens_seen": 104267776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006807357669756005,
      "loss": 3.1986,
      "theoretical_loss": 4.744325241723777,
      "tokens_seen": 104333312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000680719216947229,
      "loss": 3.4505,
      "theoretical_loss": 4.743914653310073,
      "tokens_seen": 104398848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006807026669188575,
      "loss": 3.5288,
      "theoretical_loss": 4.743504394678,
      "tokens_seen": 104464384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000680686116890486,
      "loss": 3.4208,
      "theoretical_loss": 4.743094465356039,
      "tokens_seen": 104529920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006806695668621146,
      "loss": 3.6038,
      "theoretical_loss": 4.742684864873641,
      "tokens_seen": 104595456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006806530168337431,
      "loss": 3.3705,
      "theoretical_loss": 4.742275592761223,
      "tokens_seen": 104660992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006806364668053717,
      "loss": 3.3817,
      "theoretical_loss": 4.741866648550168,
      "tokens_seen": 104726528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006806199167770002,
      "loss": 3.5512,
      "theoretical_loss": 4.741458031772817,
      "tokens_seen": 104792064
    },
    {
      "epoch": 0.04,
      "objective/train/avg_token_score": 0.012615805491805077,
      "objective/train/avg_weight": 0.9851570725440979,
      "objective/train/docs_used": 69339,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.517286777496338,
      "objective/train/std_weight": 0.05006233975291252,
      "objective/train/theoretical_loss": 4.741049741962473,
      "objective/train/tokens_used": 125317600,
      "theoretical_loss": 4.741049741962473,
      "tokens_seen": 104857600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006806033667486286,
      "loss": 3.3572,
      "theoretical_loss": 4.741049741962473,
      "tokens_seen": 104857600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006805868167202572,
      "loss": 3.4188,
      "theoretical_loss": 4.740641778653395,
      "tokens_seen": 104923136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006805702666918857,
      "loss": 3.3917,
      "theoretical_loss": 4.740234141380794,
      "tokens_seen": 104988672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006805537166635142,
      "loss": 3.3981,
      "theoretical_loss": 4.739826829680833,
      "tokens_seen": 105054208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006805371666351428,
      "loss": 3.3335,
      "theoretical_loss": 4.739419843090626,
      "tokens_seen": 105119744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006805206166067713,
      "loss": 3.3934,
      "theoretical_loss": 4.739013181148229,
      "tokens_seen": 105185280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006805040665783999,
      "loss": 3.3408,
      "theoretical_loss": 4.738606843392644,
      "tokens_seen": 105250816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006804875165500284,
      "loss": 3.2739,
      "theoretical_loss": 4.738200829363815,
      "tokens_seen": 105316352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006804709665216568,
      "loss": 3.217,
      "theoretical_loss": 4.737795138602624,
      "tokens_seen": 105381888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006804544164932854,
      "loss": 3.6013,
      "theoretical_loss": 4.737389770650887,
      "tokens_seen": 105447424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006804378664649139,
      "loss": 3.2252,
      "theoretical_loss": 4.736984725051357,
      "tokens_seen": 105512960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006804213164365424,
      "loss": 3.2865,
      "theoretical_loss": 4.736580001347717,
      "tokens_seen": 105578496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000680404766408171,
      "loss": 3.2738,
      "theoretical_loss": 4.736175599084576,
      "tokens_seen": 105644032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006803882163797995,
      "loss": 3.5137,
      "theoretical_loss": 4.735771517807473,
      "tokens_seen": 105709568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006803716663514281,
      "loss": 3.5158,
      "theoretical_loss": 4.735367757062869,
      "tokens_seen": 105775104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006803551163230565,
      "loss": 3.4878,
      "theoretical_loss": 4.734964316398148,
      "tokens_seen": 105840640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000680338566294685,
      "loss": 3.2551,
      "theoretical_loss": 4.734561195361609,
      "tokens_seen": 105906176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006803220162663136,
      "loss": 3.4591,
      "theoretical_loss": 4.734158393502471,
      "tokens_seen": 105971712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006803054662379421,
      "loss": 3.5389,
      "theoretical_loss": 4.733755910370867,
      "tokens_seen": 106037248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006802889162095706,
      "loss": 3.4312,
      "theoretical_loss": 4.73335374551784,
      "tokens_seen": 106102784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006802723661811992,
      "loss": 3.3739,
      "theoretical_loss": 4.732951898495341,
      "tokens_seen": 106168320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006802558161528277,
      "loss": 3.4301,
      "theoretical_loss": 4.7325503688562325,
      "tokens_seen": 106233856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006802392661244563,
      "loss": 3.5548,
      "theoretical_loss": 4.732149156154276,
      "tokens_seen": 106299392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006802227160960847,
      "loss": 3.3241,
      "theoretical_loss": 4.731748259944139,
      "tokens_seen": 106364928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006802061660677132,
      "loss": 3.6091,
      "theoretical_loss": 4.731347679781386,
      "tokens_seen": 106430464
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006801896160393418,
      "loss": 3.4927,
      "theoretical_loss": 4.730947415222481,
      "tokens_seen": 106496000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006801730660109703,
      "loss": 3.3634,
      "theoretical_loss": 4.730547465824781,
      "tokens_seen": 106561536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006801565159825988,
      "loss": 3.3017,
      "theoretical_loss": 4.730147831146537,
      "tokens_seen": 106627072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006801399659542274,
      "loss": 3.7818,
      "theoretical_loss": 4.72974851074689,
      "tokens_seen": 106692608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006801234159258559,
      "loss": 3.6431,
      "theoretical_loss": 4.729349504185867,
      "tokens_seen": 106758144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006801068658974844,
      "loss": 3.4547,
      "theoretical_loss": 4.728950811024383,
      "tokens_seen": 106823680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006800903158691129,
      "loss": 3.302,
      "theoretical_loss": 4.7285524308242355,
      "tokens_seen": 106889216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006800737658407414,
      "loss": 3.2739,
      "theoretical_loss": 4.728154363148102,
      "tokens_seen": 106954752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00068005721581237,
      "loss": 3.4643,
      "theoretical_loss": 4.72775660755954,
      "tokens_seen": 107020288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006800406657839985,
      "loss": 3.6336,
      "theoretical_loss": 4.72735916362298,
      "tokens_seen": 107085824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000680024115755627,
      "loss": 3.3129,
      "theoretical_loss": 4.7269620309037315,
      "tokens_seen": 107151360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006800075657272556,
      "loss": 3.4835,
      "theoretical_loss": 4.726565208967973,
      "tokens_seen": 107216896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000679991015698884,
      "loss": 3.1504,
      "theoretical_loss": 4.726168697382751,
      "tokens_seen": 107282432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006799744656705126,
      "loss": 3.2854,
      "theoretical_loss": 4.725772495715983,
      "tokens_seen": 107347968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006799579156421411,
      "loss": 3.5653,
      "theoretical_loss": 4.725376603536446,
      "tokens_seen": 107413504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006799413656137696,
      "loss": 3.4233,
      "theoretical_loss": 4.724981020413787,
      "tokens_seen": 107479040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006799248155853982,
      "loss": 3.2364,
      "theoretical_loss": 4.724585745918505,
      "tokens_seen": 107544576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006799082655570267,
      "loss": 3.2818,
      "theoretical_loss": 4.7241907796219635,
      "tokens_seen": 107610112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006798917155286553,
      "loss": 3.5029,
      "theoretical_loss": 4.723796121096381,
      "tokens_seen": 107675648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006798751655002838,
      "loss": 3.3081,
      "theoretical_loss": 4.723401769914824,
      "tokens_seen": 107741184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006798586154719122,
      "loss": 3.6838,
      "theoretical_loss": 4.723007725651219,
      "tokens_seen": 107806720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006798420654435407,
      "loss": 3.4143,
      "theoretical_loss": 4.722613987880335,
      "tokens_seen": 107872256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006798255154151692,
      "loss": 3.2929,
      "theoretical_loss": 4.722220556177792,
      "tokens_seen": 107937792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006798089653867978,
      "loss": 3.5624,
      "theoretical_loss": 4.721827430120053,
      "tokens_seen": 108003328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006797924153584263,
      "loss": 3.3907,
      "theoretical_loss": 4.721434609284424,
      "tokens_seen": 108068864
    },
    {
      "epoch": 0.04,
      "objective/train/avg_token_score": 0.014366760849952698,
      "objective/train/avg_weight": 0.983095109462738,
      "objective/train/docs_used": 71265,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.71353816986084,
      "objective/train/std_weight": 0.07395368069410324,
      "objective/train/theoretical_loss": 4.721042093249051,
      "objective/train/tokens_used": 128594400,
      "theoretical_loss": 4.721042093249051,
      "tokens_seen": 108134400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006797758653300548,
      "loss": 3.5775,
      "theoretical_loss": 4.721042093249051,
      "tokens_seen": 108134400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006797593153016834,
      "loss": 3.2632,
      "theoretical_loss": 4.720649881592919,
      "tokens_seen": 108199936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006797427652733118,
      "loss": 3.4468,
      "theoretical_loss": 4.7202579738958494,
      "tokens_seen": 108265472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006797262152449404,
      "loss": 3.4108,
      "theoretical_loss": 4.7198663697384955,
      "tokens_seen": 108331008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006797096652165689,
      "loss": 3.3991,
      "theoretical_loss": 4.719475068702346,
      "tokens_seen": 108396544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006796931151881974,
      "loss": 3.5498,
      "theoretical_loss": 4.719084070369714,
      "tokens_seen": 108462080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000679676565159826,
      "loss": 3.4994,
      "theoretical_loss": 4.718693374323747,
      "tokens_seen": 108527616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006796600151314545,
      "loss": 3.3266,
      "theoretical_loss": 4.718302980148412,
      "tokens_seen": 108593152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000679643465103083,
      "loss": 3.1781,
      "theoretical_loss": 4.717912887428501,
      "tokens_seen": 108658688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006796269150747115,
      "loss": 3.3395,
      "theoretical_loss": 4.717523095749626,
      "tokens_seen": 108724224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00067961036504634,
      "loss": 3.4558,
      "theoretical_loss": 4.717133604698222,
      "tokens_seen": 108789760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006795938150179686,
      "loss": 3.2568,
      "theoretical_loss": 4.7167444138615355,
      "tokens_seen": 108855296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006795772649895971,
      "loss": 3.5368,
      "theoretical_loss": 4.716355522827633,
      "tokens_seen": 108920832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006795607149612256,
      "loss": 3.3238,
      "theoretical_loss": 4.715966931185388,
      "tokens_seen": 108986368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006795441649328542,
      "loss": 3.2825,
      "theoretical_loss": 4.715578638524491,
      "tokens_seen": 109051904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006795276149044827,
      "loss": 3.4265,
      "theoretical_loss": 4.715190644435435,
      "tokens_seen": 109117440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006795110648761111,
      "loss": 3.6616,
      "theoretical_loss": 4.714802948509522,
      "tokens_seen": 109182976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006794945148477397,
      "loss": 3.3835,
      "theoretical_loss": 4.71441555033886,
      "tokens_seen": 109248512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006794779648193682,
      "loss": 3.4042,
      "theoretical_loss": 4.714028449516356,
      "tokens_seen": 109314048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006794614147909968,
      "loss": 3.1198,
      "theoretical_loss": 4.713641645635718,
      "tokens_seen": 109379584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006794448647626253,
      "loss": 3.36,
      "theoretical_loss": 4.713255138291454,
      "tokens_seen": 109445120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006794283147342538,
      "loss": 3.4081,
      "theoretical_loss": 4.712868927078868,
      "tokens_seen": 109510656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006794117647058824,
      "loss": 3.5164,
      "theoretical_loss": 4.712483011594056,
      "tokens_seen": 109576192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006793952146775109,
      "loss": 3.6761,
      "theoretical_loss": 4.7120973914339075,
      "tokens_seen": 109641728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006793786646491393,
      "loss": 3.3632,
      "theoretical_loss": 4.7117120661961005,
      "tokens_seen": 109707264
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006793621146207679,
      "loss": 3.3443,
      "theoretical_loss": 4.711327035479103,
      "tokens_seen": 109772800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006793455645923964,
      "loss": 3.5539,
      "theoretical_loss": 4.710942298882169,
      "tokens_seen": 109838336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000679329014564025,
      "loss": 3.2816,
      "theoretical_loss": 4.710557856005335,
      "tokens_seen": 109903872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006793124645356535,
      "loss": 3.3935,
      "theoretical_loss": 4.710173706449419,
      "tokens_seen": 109969408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000679295914507282,
      "loss": 3.6501,
      "theoretical_loss": 4.709789849816021,
      "tokens_seen": 110034944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006792793644789106,
      "loss": 3.476,
      "theoretical_loss": 4.7094062857075185,
      "tokens_seen": 110100480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000679262814450539,
      "loss": 3.3124,
      "theoretical_loss": 4.709023013727063,
      "tokens_seen": 110166016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006792462644221675,
      "loss": 3.204,
      "theoretical_loss": 4.708640033478584,
      "tokens_seen": 110231552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006792297143937961,
      "loss": 3.3594,
      "theoretical_loss": 4.708257344566778,
      "tokens_seen": 110297088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006792131643654246,
      "loss": 3.4931,
      "theoretical_loss": 4.7078749465971175,
      "tokens_seen": 110362624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006791966143370532,
      "loss": 3.4515,
      "theoretical_loss": 4.707492839175837,
      "tokens_seen": 110428160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006791800643086817,
      "loss": 3.5227,
      "theoretical_loss": 4.707111021909941,
      "tokens_seen": 110493696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006791635142803102,
      "loss": 3.2203,
      "theoretical_loss": 4.706729494407197,
      "tokens_seen": 110559232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006791469642519388,
      "loss": 3.3841,
      "theoretical_loss": 4.706348256276138,
      "tokens_seen": 110624768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006791304142235672,
      "loss": 3.5182,
      "theoretical_loss": 4.705967307126051,
      "tokens_seen": 110690304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006791138641951957,
      "loss": 3.2962,
      "theoretical_loss": 4.705586646566987,
      "tokens_seen": 110755840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006790973141668243,
      "loss": 3.4855,
      "theoretical_loss": 4.705206274209751,
      "tokens_seen": 110821376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006790807641384528,
      "loss": 3.4976,
      "theoretical_loss": 4.704826189665905,
      "tokens_seen": 110886912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006790642141100814,
      "loss": 3.3693,
      "theoretical_loss": 4.704446392547759,
      "tokens_seen": 110952448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006790476640817099,
      "loss": 3.3425,
      "theoretical_loss": 4.7040668824683785,
      "tokens_seen": 111017984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006790311140533384,
      "loss": 3.306,
      "theoretical_loss": 4.7036876590415755,
      "tokens_seen": 111083520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000679014564024967,
      "loss": 3.2351,
      "theoretical_loss": 4.7033087218819105,
      "tokens_seen": 111149056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006789980139965954,
      "loss": 3.2696,
      "theoretical_loss": 4.7029300706046895,
      "tokens_seen": 111214592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006789814639682239,
      "loss": 3.528,
      "theoretical_loss": 4.702551704825957,
      "tokens_seen": 111280128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006789649139398525,
      "loss": 3.4421,
      "theoretical_loss": 4.702173624162507,
      "tokens_seen": 111345664
    },
    {
      "epoch": 0.04,
      "objective/train/avg_token_score": 0.010272081010043621,
      "objective/train/avg_weight": 0.9879071116447449,
      "objective/train/docs_used": 72747,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.5214505195617676,
      "objective/train/std_weight": 0.060382235795259476,
      "objective/train/theoretical_loss": 4.701795828231866,
      "objective/train/tokens_used": 131871200,
      "theoretical_loss": 4.701795828231866,
      "tokens_seen": 111411200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000678948363911481,
      "loss": 3.3459,
      "theoretical_loss": 4.701795828231866,
      "tokens_seen": 111411200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006789318138831096,
      "loss": 3.3855,
      "theoretical_loss": 4.701418316652299,
      "tokens_seen": 111476736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006789152638547381,
      "loss": 3.7471,
      "theoretical_loss": 4.701041089042813,
      "tokens_seen": 111542272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006788987138263666,
      "loss": 3.4769,
      "theoretical_loss": 4.700664145023142,
      "tokens_seen": 111607808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006788821637979951,
      "loss": 3.487,
      "theoretical_loss": 4.700287484213753,
      "tokens_seen": 111673344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006788656137696236,
      "loss": 3.2516,
      "theoretical_loss": 4.699911106235849,
      "tokens_seen": 111738880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006788490637412521,
      "loss": 3.5063,
      "theoretical_loss": 4.6995350107113545,
      "tokens_seen": 111804416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006788325137128806,
      "loss": 3.598,
      "theoretical_loss": 4.699159197262922,
      "tokens_seen": 111869952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006788159636845092,
      "loss": 3.381,
      "theoretical_loss": 4.698783665513934,
      "tokens_seen": 111935488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006787994136561377,
      "loss": 3.4983,
      "theoretical_loss": 4.698408415088491,
      "tokens_seen": 112001024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006787828636277662,
      "loss": 3.3171,
      "theoretical_loss": 4.698033445611415,
      "tokens_seen": 112066560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006787663135993947,
      "loss": 3.3635,
      "theoretical_loss": 4.6976587567082495,
      "tokens_seen": 112132096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006787497635710232,
      "loss": 3.4945,
      "theoretical_loss": 4.697284348005253,
      "tokens_seen": 112197632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006787332135426517,
      "loss": 3.3634,
      "theoretical_loss": 4.696910219129402,
      "tokens_seen": 112263168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006787166635142803,
      "loss": 3.1058,
      "theoretical_loss": 4.696536369708386,
      "tokens_seen": 112328704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006787001134859088,
      "loss": 3.3002,
      "theoretical_loss": 4.696162799370606,
      "tokens_seen": 112394240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006786835634575374,
      "loss": 3.4106,
      "theoretical_loss": 4.695789507745176,
      "tokens_seen": 112459776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006786670134291659,
      "loss": 3.4174,
      "theoretical_loss": 4.695416494461917,
      "tokens_seen": 112525312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006786504634007943,
      "loss": 3.2949,
      "theoretical_loss": 4.695043759151353,
      "tokens_seen": 112590848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006786339133724229,
      "loss": 3.6219,
      "theoretical_loss": 4.694671301444722,
      "tokens_seen": 112656384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006786173633440514,
      "loss": 3.4265,
      "theoretical_loss": 4.694299120973957,
      "tokens_seen": 112721920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006786008133156799,
      "loss": 3.5429,
      "theoretical_loss": 4.693927217371698,
      "tokens_seen": 112787456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006785842632873085,
      "loss": 3.28,
      "theoretical_loss": 4.693555590271282,
      "tokens_seen": 112852992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000678567713258937,
      "loss": 3.2972,
      "theoretical_loss": 4.693184239306744,
      "tokens_seen": 112918528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006785511632305656,
      "loss": 3.4319,
      "theoretical_loss": 4.692813164112819,
      "tokens_seen": 112984064
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000678534613202194,
      "loss": 3.4595,
      "theoretical_loss": 4.692442364324931,
      "tokens_seen": 113049600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006785180631738225,
      "loss": 3.5689,
      "theoretical_loss": 4.692071839579201,
      "tokens_seen": 113115136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006785015131454511,
      "loss": 3.5903,
      "theoretical_loss": 4.6917015895124425,
      "tokens_seen": 113180672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006784849631170796,
      "loss": 3.2777,
      "theoretical_loss": 4.691331613762153,
      "tokens_seen": 113246208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006784684130887081,
      "loss": 3.1893,
      "theoretical_loss": 4.690961911966523,
      "tokens_seen": 113311744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006784518630603367,
      "loss": 3.5089,
      "theoretical_loss": 4.690592483764427,
      "tokens_seen": 113377280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006784353130319652,
      "loss": 3.2201,
      "theoretical_loss": 4.690223328795424,
      "tokens_seen": 113442816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006784187630035938,
      "loss": 3.3114,
      "theoretical_loss": 4.689854446699757,
      "tokens_seen": 113508352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006784022129752222,
      "loss": 3.4761,
      "theoretical_loss": 4.689485837118347,
      "tokens_seen": 113573888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006783856629468507,
      "loss": 3.4177,
      "theoretical_loss": 4.689117499692798,
      "tokens_seen": 113639424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006783691129184793,
      "loss": 3.3841,
      "theoretical_loss": 4.688749434065389,
      "tokens_seen": 113704960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006783525628901078,
      "loss": 3.5932,
      "theoretical_loss": 4.688381639879076,
      "tokens_seen": 113770496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006783360128617363,
      "loss": 2.9831,
      "theoretical_loss": 4.68801411677749,
      "tokens_seen": 113836032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006783194628333649,
      "loss": 3.2593,
      "theoretical_loss": 4.687646864404934,
      "tokens_seen": 113901568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006783029128049934,
      "loss": 3.349,
      "theoretical_loss": 4.687279882406381,
      "tokens_seen": 113967104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000678286362776622,
      "loss": 3.5735,
      "theoretical_loss": 4.686913170427477,
      "tokens_seen": 114032640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006782698127482504,
      "loss": 3.3558,
      "theoretical_loss": 4.68654672811453,
      "tokens_seen": 114098176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006782532627198789,
      "loss": 3.6124,
      "theoretical_loss": 4.68618055511452,
      "tokens_seen": 114163712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006782367126915075,
      "loss": 3.3006,
      "theoretical_loss": 4.685814651075088,
      "tokens_seen": 114229248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000678220162663136,
      "loss": 3.2073,
      "theoretical_loss": 4.685449015644537,
      "tokens_seen": 114294784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006782036126347645,
      "loss": 3.3643,
      "theoretical_loss": 4.685083648471835,
      "tokens_seen": 114360320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006781870626063931,
      "loss": 3.2823,
      "theoretical_loss": 4.684718549206607,
      "tokens_seen": 114425856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006781705125780216,
      "loss": 3.216,
      "theoretical_loss": 4.6843537174991345,
      "tokens_seen": 114491392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006781539625496501,
      "loss": 3.4131,
      "theoretical_loss": 4.6839891530003595,
      "tokens_seen": 114556928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006781374125212786,
      "loss": 3.2002,
      "theoretical_loss": 4.683624855361876,
      "tokens_seen": 114622464
    },
    {
      "epoch": 0.04,
      "objective/train/avg_token_score": 0.010599654167890549,
      "objective/train/avg_weight": 0.9875220060348511,
      "objective/train/docs_used": 74963,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.458974838256836,
      "objective/train/std_weight": 0.06394177675247192,
      "objective/train/theoretical_loss": 4.68326082423593,
      "objective/train/tokens_used": 135148000,
      "theoretical_loss": 4.68326082423593,
      "tokens_seen": 114688000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006781208624929071,
      "loss": 3.4232,
      "theoretical_loss": 4.68326082423593,
      "tokens_seen": 114688000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006781043124645357,
      "loss": 3.4455,
      "theoretical_loss": 4.682897059275422,
      "tokens_seen": 114753536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006780877624361642,
      "loss": 3.2785,
      "theoretical_loss": 4.682533560133901,
      "tokens_seen": 114819072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006780712124077927,
      "loss": 3.554,
      "theoretical_loss": 4.682170326465565,
      "tokens_seen": 114884608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006780546623794213,
      "loss": 3.2857,
      "theoretical_loss": 4.681807357925257,
      "tokens_seen": 114950144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006780381123510497,
      "loss": 3.457,
      "theoretical_loss": 4.681444654168468,
      "tokens_seen": 115015680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006780215623226783,
      "loss": 3.265,
      "theoretical_loss": 4.68108221485133,
      "tokens_seen": 115081216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006780050122943068,
      "loss": 3.2632,
      "theoretical_loss": 4.680720039630617,
      "tokens_seen": 115146752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006779884622659353,
      "loss": 3.3197,
      "theoretical_loss": 4.680358128163747,
      "tokens_seen": 115212288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006779719122375639,
      "loss": 3.4245,
      "theoretical_loss": 4.679996480108773,
      "tokens_seen": 115277824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006779553622091924,
      "loss": 3.3016,
      "theoretical_loss": 4.6796350951243895,
      "tokens_seen": 115343360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006779388121808209,
      "loss": 3.1627,
      "theoretical_loss": 4.679273972869922,
      "tokens_seen": 115408896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006779222621524495,
      "loss": 3.4023,
      "theoretical_loss": 4.678913113005333,
      "tokens_seen": 115474432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006779057121240779,
      "loss": 3.32,
      "theoretical_loss": 4.6785525151912175,
      "tokens_seen": 115539968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006778891620957065,
      "loss": 3.4606,
      "theoretical_loss": 4.678192179088802,
      "tokens_seen": 115605504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006778726120673349,
      "loss": 3.0879,
      "theoretical_loss": 4.6778321043599425,
      "tokens_seen": 115671040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006778560620389635,
      "loss": 3.291,
      "theoretical_loss": 4.677472290667122,
      "tokens_seen": 115736576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000677839512010592,
      "loss": 3.199,
      "theoretical_loss": 4.677112737673453,
      "tokens_seen": 115802112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006778229619822205,
      "loss": 3.1871,
      "theoretical_loss": 4.676753445042669,
      "tokens_seen": 115867648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000677806411953849,
      "loss": 3.4287,
      "theoretical_loss": 4.676394412439132,
      "tokens_seen": 115933184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006777898619254775,
      "loss": 3.1629,
      "theoretical_loss": 4.6760356395278215,
      "tokens_seen": 115998720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000677773311897106,
      "loss": 3.4632,
      "theoretical_loss": 4.675677125974339,
      "tokens_seen": 116064256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006777567618687346,
      "loss": 3.4824,
      "theoretical_loss": 4.675318871444908,
      "tokens_seen": 116129792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006777402118403631,
      "loss": 3.6116,
      "theoretical_loss": 4.674960875606366,
      "tokens_seen": 116195328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006777236618119917,
      "loss": 3.2575,
      "theoretical_loss": 4.674603138126168,
      "tokens_seen": 116260864
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006777071117836202,
      "loss": 3.1646,
      "theoretical_loss": 4.674245658672382,
      "tokens_seen": 116326400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006776905617552487,
      "loss": 3.5151,
      "theoretical_loss": 4.673888436913694,
      "tokens_seen": 116391936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006776740117268772,
      "loss": 3.5362,
      "theoretical_loss": 4.673531472519397,
      "tokens_seen": 116457472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006776574616985057,
      "loss": 3.5749,
      "theoretical_loss": 4.673174765159393,
      "tokens_seen": 116523008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006776409116701342,
      "loss": 3.1624,
      "theoretical_loss": 4.672818314504198,
      "tokens_seen": 116588544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006776243616417628,
      "loss": 3.5592,
      "theoretical_loss": 4.6724621202249335,
      "tokens_seen": 116654080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006776078116133913,
      "loss": 3.2612,
      "theoretical_loss": 4.672106181993324,
      "tokens_seen": 116719616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006775912615850199,
      "loss": 3.472,
      "theoretical_loss": 4.6717504994817,
      "tokens_seen": 116785152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006775747115566484,
      "loss": 3.445,
      "theoretical_loss": 4.671395072362996,
      "tokens_seen": 116850688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006775581615282768,
      "loss": 3.261,
      "theoretical_loss": 4.671039900310747,
      "tokens_seen": 116916224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006775416114999054,
      "loss": 3.2307,
      "theoretical_loss": 4.670684982999088,
      "tokens_seen": 116981760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006775250614715339,
      "loss": 3.4988,
      "theoretical_loss": 4.670330320102753,
      "tokens_seen": 117047296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006775085114431624,
      "loss": 3.2906,
      "theoretical_loss": 4.669975911297072,
      "tokens_seen": 117112832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000677491961414791,
      "loss": 3.379,
      "theoretical_loss": 4.669621756257971,
      "tokens_seen": 117178368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006774754113864195,
      "loss": 3.4057,
      "theoretical_loss": 4.669267854661973,
      "tokens_seen": 117243904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006774588613580481,
      "loss": 3.3071,
      "theoretical_loss": 4.668914206186189,
      "tokens_seen": 117309440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006774423113296766,
      "loss": 3.4041,
      "theoretical_loss": 4.6685608105083265,
      "tokens_seen": 117374976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000677425761301305,
      "loss": 3.3082,
      "theoretical_loss": 4.66820766730668,
      "tokens_seen": 117440512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006774092112729336,
      "loss": 3.4138,
      "theoretical_loss": 4.667854776260132,
      "tokens_seen": 117506048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006773926612445621,
      "loss": 3.2075,
      "theoretical_loss": 4.667502137048155,
      "tokens_seen": 117571584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006773761112161907,
      "loss": 3.3093,
      "theoretical_loss": 4.667149749350805,
      "tokens_seen": 117637120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006773595611878192,
      "loss": 3.6402,
      "theoretical_loss": 4.666797612848723,
      "tokens_seen": 117702656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006773430111594477,
      "loss": 3.5363,
      "theoretical_loss": 4.666445727223134,
      "tokens_seen": 117768192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006773264611310763,
      "loss": 3.4251,
      "theoretical_loss": 4.666094092155843,
      "tokens_seen": 117833728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006773099111027047,
      "loss": 3.2238,
      "theoretical_loss": 4.665742707329238,
      "tokens_seen": 117899264
    },
    {
      "epoch": 0.04,
      "objective/train/avg_token_score": 0.0178232304751873,
      "objective/train/avg_weight": 0.9790148735046387,
      "objective/train/docs_used": 77191,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.600983142852783,
      "objective/train/std_weight": 0.10139288753271103,
      "objective/train/theoretical_loss": 4.665391572426282,
      "objective/train/tokens_used": 138424800,
      "theoretical_loss": 4.665391572426282,
      "tokens_seen": 117964800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006772933610743332,
      "loss": 3.3414,
      "theoretical_loss": 4.665391572426282,
      "tokens_seen": 117964800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006772768110459618,
      "loss": 3.4759,
      "theoretical_loss": 4.665040687130518,
      "tokens_seen": 118030336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006772602610175903,
      "loss": 3.1881,
      "theoretical_loss": 4.664690051126065,
      "tokens_seen": 118095872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006772437109892189,
      "loss": 3.2646,
      "theoretical_loss": 4.664339664097617,
      "tokens_seen": 118161408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006772271609608474,
      "loss": 3.3701,
      "theoretical_loss": 4.66398952573044,
      "tokens_seen": 118226944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006772106109324759,
      "loss": 3.3418,
      "theoretical_loss": 4.663639635710373,
      "tokens_seen": 118292480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006771940609041045,
      "loss": 3.4871,
      "theoretical_loss": 4.663289993723826,
      "tokens_seen": 118358016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006771775108757329,
      "loss": 3.4466,
      "theoretical_loss": 4.662940599457777,
      "tokens_seen": 118423552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006771609608473614,
      "loss": 3.2738,
      "theoretical_loss": 4.662591452599774,
      "tokens_seen": 118489088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00067714441081899,
      "loss": 3.294,
      "theoretical_loss": 4.662242552837929,
      "tokens_seen": 118554624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006771278607906185,
      "loss": 3.4097,
      "theoretical_loss": 4.661893899860923,
      "tokens_seen": 118620160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006771113107622471,
      "loss": 3.4192,
      "theoretical_loss": 4.6615454933579965,
      "tokens_seen": 118685696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006770947607338756,
      "loss": 3.3621,
      "theoretical_loss": 4.661197333018957,
      "tokens_seen": 118751232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006770782107055041,
      "loss": 3.5622,
      "theoretical_loss": 4.66084941853417,
      "tokens_seen": 118816768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006770616606771326,
      "loss": 3.2512,
      "theoretical_loss": 4.6605017495945615,
      "tokens_seen": 118882304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006770451106487611,
      "loss": 3.4807,
      "theoretical_loss": 4.660154325891618,
      "tokens_seen": 118947840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006770285606203896,
      "loss": 3.2467,
      "theoretical_loss": 4.659807147117382,
      "tokens_seen": 119013376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006770120105920182,
      "loss": 3.2403,
      "theoretical_loss": 4.6594602129644525,
      "tokens_seen": 119078912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006769954605636467,
      "loss": 3.3513,
      "theoretical_loss": 4.659113523125981,
      "tokens_seen": 119144448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006769789105352753,
      "loss": 3.442,
      "theoretical_loss": 4.6587670772956775,
      "tokens_seen": 119209984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006769623605069038,
      "loss": 3.5298,
      "theoretical_loss": 4.658420875167799,
      "tokens_seen": 119275520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006769458104785322,
      "loss": 3.4019,
      "theoretical_loss": 4.658074916437155,
      "tokens_seen": 119341056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006769292604501608,
      "loss": 3.3011,
      "theoretical_loss": 4.657729200799105,
      "tokens_seen": 119406592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006769127104217893,
      "loss": 3.2877,
      "theoretical_loss": 4.657383727949558,
      "tokens_seen": 119472128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006768961603934178,
      "loss": 3.3069,
      "theoretical_loss": 4.657038497584967,
      "tokens_seen": 119537664
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006768796103650463,
      "loss": 3.0861,
      "theoretical_loss": 4.656693509402331,
      "tokens_seen": 119603200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006768630603366748,
      "loss": 3.4621,
      "theoretical_loss": 4.6563487630991975,
      "tokens_seen": 119668736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006768465103083034,
      "loss": 3.5319,
      "theoretical_loss": 4.656004258373651,
      "tokens_seen": 119734272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006768299602799318,
      "loss": 3.53,
      "theoretical_loss": 4.655659994924323,
      "tokens_seen": 119799808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006768134102515604,
      "loss": 3.5892,
      "theoretical_loss": 4.655315972450383,
      "tokens_seen": 119865344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006767968602231889,
      "loss": 3.4837,
      "theoretical_loss": 4.65497219065154,
      "tokens_seen": 119930880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006767803101948174,
      "loss": 3.5044,
      "theoretical_loss": 4.654628649228041,
      "tokens_seen": 119996416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000676763760166446,
      "loss": 3.1617,
      "theoretical_loss": 4.654285347880672,
      "tokens_seen": 120061952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006767472101380745,
      "loss": 3.2647,
      "theoretical_loss": 4.653942286310749,
      "tokens_seen": 120127488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000676730660109703,
      "loss": 3.1961,
      "theoretical_loss": 4.653599464220129,
      "tokens_seen": 120193024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006767141100813316,
      "loss": 3.3048,
      "theoretical_loss": 4.653256881311198,
      "tokens_seen": 120258560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00067669756005296,
      "loss": 3.091,
      "theoretical_loss": 4.6529145372868745,
      "tokens_seen": 120324096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006766810100245886,
      "loss": 3.288,
      "theoretical_loss": 4.652572431850608,
      "tokens_seen": 120389632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006766644599962171,
      "loss": 3.168,
      "theoretical_loss": 4.652230564706377,
      "tokens_seen": 120455168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006766479099678456,
      "loss": 3.2049,
      "theoretical_loss": 4.651888935558688,
      "tokens_seen": 120520704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006766313599394742,
      "loss": 3.1102,
      "theoretical_loss": 4.651547544112575,
      "tokens_seen": 120586240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006766148099111027,
      "loss": 3.0657,
      "theoretical_loss": 4.651206390073597,
      "tokens_seen": 120651776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006765982598827312,
      "loss": 3.3227,
      "theoretical_loss": 4.650865473147837,
      "tokens_seen": 120717312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006765817098543597,
      "loss": 3.197,
      "theoretical_loss": 4.650524793041903,
      "tokens_seen": 120782848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006765651598259882,
      "loss": 3.264,
      "theoretical_loss": 4.650184349462922,
      "tokens_seen": 120848384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006765486097976168,
      "loss": 3.1299,
      "theoretical_loss": 4.649844142118544,
      "tokens_seen": 120913920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006765320597692453,
      "loss": 3.0494,
      "theoretical_loss": 4.6495041707169396,
      "tokens_seen": 120979456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006765155097408738,
      "loss": 3.4312,
      "theoretical_loss": 4.649164434966794,
      "tokens_seen": 121044992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006764989597125024,
      "loss": 3.3713,
      "theoretical_loss": 4.648824934577313,
      "tokens_seen": 121110528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006764824096841309,
      "loss": 3.3474,
      "theoretical_loss": 4.648485669258216,
      "tokens_seen": 121176064
    },
    {
      "epoch": 0.04,
      "objective/train/avg_token_score": 0.031265705823898315,
      "objective/train/avg_weight": 0.9632101655006409,
      "objective/train/docs_used": 78964,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.45198130607605,
      "objective/train/std_weight": 0.1388011872768402,
      "objective/train/theoretical_loss": 4.648146638719739,
      "objective/train/tokens_used": 141701600,
      "theoretical_loss": 4.648146638719739,
      "tokens_seen": 121241600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006764658596557593,
      "loss": 3.3364,
      "theoretical_loss": 4.648146638719739,
      "tokens_seen": 121241600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006764493096273879,
      "loss": 3.4242,
      "theoretical_loss": 4.647807842672631,
      "tokens_seen": 121307136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006764327595990164,
      "loss": 3.3819,
      "theoretical_loss": 4.647469280828153,
      "tokens_seen": 121372672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000676416209570645,
      "loss": 3.361,
      "theoretical_loss": 4.647130952898077,
      "tokens_seen": 121438208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006763996595422735,
      "loss": 3.355,
      "theoretical_loss": 4.646792858594686,
      "tokens_seen": 121503744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000676383109513902,
      "loss": 3.1114,
      "theoretical_loss": 4.64645499763077,
      "tokens_seen": 121569280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006763665594855306,
      "loss": 3.4007,
      "theoretical_loss": 4.646117369719629,
      "tokens_seen": 121634816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006763500094571591,
      "loss": 3.296,
      "theoretical_loss": 4.645779974575069,
      "tokens_seen": 121700352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006763334594287875,
      "loss": 3.3096,
      "theoretical_loss": 4.6454428119113995,
      "tokens_seen": 121765888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006763169094004161,
      "loss": 3.348,
      "theoretical_loss": 4.6451058814434365,
      "tokens_seen": 121831424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006763003593720446,
      "loss": 3.3218,
      "theoretical_loss": 4.644769182886495,
      "tokens_seen": 121896960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006762838093436732,
      "loss": 3.2782,
      "theoretical_loss": 4.644432715956399,
      "tokens_seen": 121962496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006762672593153017,
      "loss": 3.5958,
      "theoretical_loss": 4.644096480369466,
      "tokens_seen": 122028032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006762507092869302,
      "loss": 3.2856,
      "theoretical_loss": 4.643760475842518,
      "tokens_seen": 122093568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006762341592585588,
      "loss": 3.4076,
      "theoretical_loss": 4.6434247020928705,
      "tokens_seen": 122159104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006762176092301872,
      "loss": 3.2544,
      "theoretical_loss": 4.643089158838341,
      "tokens_seen": 122224640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006762010592018157,
      "loss": 3.4624,
      "theoretical_loss": 4.642753845797243,
      "tokens_seen": 122290176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006761845091734443,
      "loss": 3.3304,
      "theoretical_loss": 4.642418762688379,
      "tokens_seen": 122355712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006761679591450728,
      "loss": 3.3515,
      "theoretical_loss": 4.642083909231053,
      "tokens_seen": 122421248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006761514091167014,
      "loss": 3.0565,
      "theoretical_loss": 4.641749285145057,
      "tokens_seen": 122486784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006761348590883299,
      "loss": 3.3485,
      "theoretical_loss": 4.641414890150675,
      "tokens_seen": 122552320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006761183090599584,
      "loss": 3.2329,
      "theoretical_loss": 4.641080723968684,
      "tokens_seen": 122617856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000676101759031587,
      "loss": 3.3034,
      "theoretical_loss": 4.6407467863203475,
      "tokens_seen": 122683392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006760852090032154,
      "loss": 3.3707,
      "theoretical_loss": 4.640413076927418,
      "tokens_seen": 122748928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006760686589748439,
      "loss": 3.5231,
      "theoretical_loss": 4.6400795955121374,
      "tokens_seen": 122814464
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006760521089464725,
      "loss": 3.294,
      "theoretical_loss": 4.639746341797229,
      "tokens_seen": 122880000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000676035558918101,
      "loss": 3.4259,
      "theoretical_loss": 4.639413315505905,
      "tokens_seen": 122945536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006760190088897296,
      "loss": 3.2659,
      "theoretical_loss": 4.639080516361861,
      "tokens_seen": 123011072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006760024588613581,
      "loss": 3.4748,
      "theoretical_loss": 4.638747944089273,
      "tokens_seen": 123076608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006759859088329866,
      "loss": 3.5453,
      "theoretical_loss": 4.638415598412799,
      "tokens_seen": 123142144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006759693588046151,
      "loss": 3.4086,
      "theoretical_loss": 4.638083479057579,
      "tokens_seen": 123207680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006759528087762436,
      "loss": 3.238,
      "theoretical_loss": 4.637751585749234,
      "tokens_seen": 123273216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006759362587478722,
      "loss": 3.2474,
      "theoretical_loss": 4.6374199182138565,
      "tokens_seen": 123338752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006759197087195007,
      "loss": 3.1992,
      "theoretical_loss": 4.637088476178025,
      "tokens_seen": 123404288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006759031586911292,
      "loss": 3.3656,
      "theoretical_loss": 4.636757259368787,
      "tokens_seen": 123469824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006758866086627577,
      "loss": 3.2696,
      "theoretical_loss": 4.636426267513668,
      "tokens_seen": 123535360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006758700586343862,
      "loss": 3.1573,
      "theoretical_loss": 4.636095500340669,
      "tokens_seen": 123600896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006758535086060147,
      "loss": 3.3956,
      "theoretical_loss": 4.635764957578261,
      "tokens_seen": 123666432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006758369585776432,
      "loss": 3.3668,
      "theoretical_loss": 4.635434638955388,
      "tokens_seen": 123731968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006758204085492717,
      "loss": 3.4725,
      "theoretical_loss": 4.635104544201465,
      "tokens_seen": 123797504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006758038585209003,
      "loss": 3.4506,
      "theoretical_loss": 4.634774673046376,
      "tokens_seen": 123863040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006757873084925288,
      "loss": 3.2885,
      "theoretical_loss": 4.634445025220475,
      "tokens_seen": 123928576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006757707584641574,
      "loss": 3.4437,
      "theoretical_loss": 4.634115600454582,
      "tokens_seen": 123994112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006757542084357859,
      "loss": 3.4673,
      "theoretical_loss": 4.633786398479983,
      "tokens_seen": 124059648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006757376584074143,
      "loss": 2.9906,
      "theoretical_loss": 4.6334574190284314,
      "tokens_seen": 124125184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006757211083790429,
      "loss": 3.0863,
      "theoretical_loss": 4.633128661832145,
      "tokens_seen": 124190720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006757045583506714,
      "loss": 3.3945,
      "theoretical_loss": 4.632800126623803,
      "tokens_seen": 124256256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006756880083222999,
      "loss": 3.387,
      "theoretical_loss": 4.632471813136547,
      "tokens_seen": 124321792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006756714582939285,
      "loss": 3.3522,
      "theoretical_loss": 4.632143721103983,
      "tokens_seen": 124387328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000675654908265557,
      "loss": 3.2613,
      "theoretical_loss": 4.631815850260173,
      "tokens_seen": 124452864
    },
    {
      "epoch": 0.04,
      "objective/train/avg_token_score": 0.022483162581920624,
      "objective/train/avg_weight": 0.9735393524169922,
      "objective/train/docs_used": 81017,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.546703815460205,
      "objective/train/std_weight": 0.07191955298185349,
      "objective/train/theoretical_loss": 4.631488200339643,
      "objective/train/tokens_used": 144978400,
      "theoretical_loss": 4.631488200339643,
      "tokens_seen": 124518400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006756383582371856,
      "loss": 3.4109,
      "theoretical_loss": 4.631488200339643,
      "tokens_seen": 124518400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006756218082088141,
      "loss": 3.4406,
      "theoretical_loss": 4.63116077107737,
      "tokens_seen": 124583936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006756052581804425,
      "loss": 3.2448,
      "theoretical_loss": 4.630833562208797,
      "tokens_seen": 124649472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006755887081520711,
      "loss": 3.3641,
      "theoretical_loss": 4.630506573469815,
      "tokens_seen": 124715008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006755721581236996,
      "loss": 3.4978,
      "theoretical_loss": 4.630179804596775,
      "tokens_seen": 124780544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006755556080953281,
      "loss": 3.3144,
      "theoretical_loss": 4.629853255326481,
      "tokens_seen": 124846080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006755390580669567,
      "loss": 3.4717,
      "theoretical_loss": 4.629526925396189,
      "tokens_seen": 124911616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006755225080385852,
      "loss": 3.4841,
      "theoretical_loss": 4.6292008145436085,
      "tokens_seen": 124977152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006755059580102138,
      "loss": 3.2636,
      "theoretical_loss": 4.628874922506897,
      "tokens_seen": 125042688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006754894079818422,
      "loss": 3.4188,
      "theoretical_loss": 4.628549249024666,
      "tokens_seen": 125108224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006754728579534707,
      "loss": 3.4419,
      "theoretical_loss": 4.628223793835975,
      "tokens_seen": 125173760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006754563079250993,
      "loss": 3.1614,
      "theoretical_loss": 4.627898556680327,
      "tokens_seen": 125239296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006754397578967278,
      "loss": 3.3443,
      "theoretical_loss": 4.627573537297678,
      "tokens_seen": 125304832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006754232078683563,
      "loss": 2.9854,
      "theoretical_loss": 4.627248735428427,
      "tokens_seen": 125370368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006754066578399849,
      "loss": 3.2646,
      "theoretical_loss": 4.6269241508134185,
      "tokens_seen": 125435904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006753901078116134,
      "loss": 3.5994,
      "theoretical_loss": 4.6265997831939405,
      "tokens_seen": 125501440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000675373557783242,
      "loss": 3.56,
      "theoretical_loss": 4.6262756323117245,
      "tokens_seen": 125566976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006753570077548704,
      "loss": 3.4692,
      "theoretical_loss": 4.625951697908944,
      "tokens_seen": 125632512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006753404577264989,
      "loss": 3.3719,
      "theoretical_loss": 4.625627979728212,
      "tokens_seen": 125698048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006753239076981275,
      "loss": 3.3993,
      "theoretical_loss": 4.625304477512584,
      "tokens_seen": 125763584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000675307357669756,
      "loss": 3.5418,
      "theoretical_loss": 4.624981191005554,
      "tokens_seen": 125829120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006752908076413845,
      "loss": 3.3205,
      "theoretical_loss": 4.624658119951052,
      "tokens_seen": 125894656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0006752742576130131,
      "loss": 3.2628,
      "theoretical_loss": 4.624335264093447,
      "tokens_seen": 125960192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006752577075846416,
      "loss": 3.399,
      "theoretical_loss": 4.624012623177544,
      "tokens_seen": 126025728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006752411575562701,
      "loss": 3.494,
      "theoretical_loss": 4.623690196948582,
      "tokens_seen": 126091264
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006752246075278986,
      "loss": 3.4548,
      "theoretical_loss": 4.623367985152234,
      "tokens_seen": 126156800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006752080574995271,
      "loss": 2.9678,
      "theoretical_loss": 4.623045987534609,
      "tokens_seen": 126222336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006751915074711557,
      "loss": 3.2589,
      "theoretical_loss": 4.622724203842246,
      "tokens_seen": 126287872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006751749574427842,
      "loss": 3.1326,
      "theoretical_loss": 4.622402633822114,
      "tokens_seen": 126353408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006751584074144127,
      "loss": 3.133,
      "theoretical_loss": 4.622081277221616,
      "tokens_seen": 126418944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006751418573860413,
      "loss": 3.2025,
      "theoretical_loss": 4.62176013378858,
      "tokens_seen": 126484480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006751253073576697,
      "loss": 3.6294,
      "theoretical_loss": 4.621439203271267,
      "tokens_seen": 126550016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006751087573292983,
      "loss": 3.3886,
      "theoretical_loss": 4.621118485418362,
      "tokens_seen": 126615552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006750922073009268,
      "loss": 3.2715,
      "theoretical_loss": 4.620797979978978,
      "tokens_seen": 126681088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006750756572725553,
      "loss": 3.262,
      "theoretical_loss": 4.620477686702651,
      "tokens_seen": 126746624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006750591072441839,
      "loss": 3.3274,
      "theoretical_loss": 4.620157605339347,
      "tokens_seen": 126812160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006750425572158124,
      "loss": 3.054,
      "theoretical_loss": 4.619837735639452,
      "tokens_seen": 126877696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006750260071874409,
      "loss": 3.531,
      "theoretical_loss": 4.619518077353776,
      "tokens_seen": 126943232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006750094571590695,
      "loss": 3.3474,
      "theoretical_loss": 4.619198630233547,
      "tokens_seen": 127008768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006749929071306979,
      "loss": 3.3633,
      "theoretical_loss": 4.6188793940304205,
      "tokens_seen": 127074304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006749763571023265,
      "loss": 3.2994,
      "theoretical_loss": 4.618560368496466,
      "tokens_seen": 127139840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000674959807073955,
      "loss": 3.404,
      "theoretical_loss": 4.618241553384175,
      "tokens_seen": 127205376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006749432570455835,
      "loss": 3.445,
      "theoretical_loss": 4.617922948446459,
      "tokens_seen": 127270912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000674926707017212,
      "loss": 3.423,
      "theoretical_loss": 4.617604553436642,
      "tokens_seen": 127336448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006749101569888405,
      "loss": 3.5437,
      "theoretical_loss": 4.617286368108466,
      "tokens_seen": 127401984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006748936069604691,
      "loss": 3.5567,
      "theoretical_loss": 4.6169683922160925,
      "tokens_seen": 127467520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006748770569320975,
      "loss": 3.5488,
      "theoretical_loss": 4.616650625514091,
      "tokens_seen": 127533056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006748605069037261,
      "loss": 3.2665,
      "theoretical_loss": 4.616333067757449,
      "tokens_seen": 127598592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006748439568753546,
      "loss": 3.3029,
      "theoretical_loss": 4.616015718701563,
      "tokens_seen": 127664128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006748274068469831,
      "loss": 3.4949,
      "theoretical_loss": 4.615698578102245,
      "tokens_seen": 127729664
    },
    {
      "epoch": 0.05,
      "objective/train/avg_token_score": 0.013609668239951134,
      "objective/train/avg_weight": 0.983987033367157,
      "objective/train/docs_used": 82924,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.299387216567993,
      "objective/train/std_weight": 0.04530864208936691,
      "objective/train/theoretical_loss": 4.615381645715717,
      "objective/train/tokens_used": 148255200,
      "theoretical_loss": 4.615381645715717,
      "tokens_seen": 127795200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006748108568186117,
      "loss": 3.1499,
      "theoretical_loss": 4.615381645715717,
      "tokens_seen": 127795200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006747943067902402,
      "loss": 3.2652,
      "theoretical_loss": 4.615064921298608,
      "tokens_seen": 127860736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006747777567618687,
      "loss": 3.3192,
      "theoretical_loss": 4.61474840460796,
      "tokens_seen": 127926272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006747612067334973,
      "loss": 3.2324,
      "theoretical_loss": 4.614432095401219,
      "tokens_seen": 127991808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006747446567051257,
      "loss": 3.2781,
      "theoretical_loss": 4.614115993436242,
      "tokens_seen": 128057344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006747281066767543,
      "loss": 3.4055,
      "theoretical_loss": 4.613800098471291,
      "tokens_seen": 128122880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006747115566483828,
      "loss": 3.4645,
      "theoretical_loss": 4.613484410265032,
      "tokens_seen": 128188416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006746950066200113,
      "loss": 3.5301,
      "theoretical_loss": 4.613168928576538,
      "tokens_seen": 128253952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006746784565916399,
      "loss": 3.3399,
      "theoretical_loss": 4.612853653165283,
      "tokens_seen": 128319488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006746619065632684,
      "loss": 3.5154,
      "theoretical_loss": 4.612538583791146,
      "tokens_seen": 128385024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006746453565348969,
      "loss": 3.1842,
      "theoretical_loss": 4.612223720214407,
      "tokens_seen": 128450560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006746288065065254,
      "loss": 3.3704,
      "theoretical_loss": 4.611909062195749,
      "tokens_seen": 128516096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006746122564781539,
      "loss": 3.3855,
      "theoretical_loss": 4.61159460949625,
      "tokens_seen": 128581632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006745957064497825,
      "loss": 3.3535,
      "theoretical_loss": 4.611280361877393,
      "tokens_seen": 128647168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000674579156421411,
      "loss": 3.3647,
      "theoretical_loss": 4.610966319101056,
      "tokens_seen": 128712704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006745626063930395,
      "loss": 3.287,
      "theoretical_loss": 4.610652480929515,
      "tokens_seen": 128778240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006745460563646681,
      "loss": 3.322,
      "theoretical_loss": 4.610338847125445,
      "tokens_seen": 128843776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006745295063362966,
      "loss": 2.8977,
      "theoretical_loss": 4.610025417451913,
      "tokens_seen": 128909312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000674512956307925,
      "loss": 3.3799,
      "theoretical_loss": 4.6097121916723856,
      "tokens_seen": 128974848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006744964062795536,
      "loss": 3.3479,
      "theoretical_loss": 4.609399169550718,
      "tokens_seen": 129040384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006744798562511821,
      "loss": 3.4278,
      "theoretical_loss": 4.609086350851165,
      "tokens_seen": 129105920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006744633062228107,
      "loss": 3.3086,
      "theoretical_loss": 4.6087737353383655,
      "tokens_seen": 129171456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006744467561944392,
      "loss": 3.3838,
      "theoretical_loss": 4.6084613227773605,
      "tokens_seen": 129236992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006744302061660677,
      "loss": 3.0932,
      "theoretical_loss": 4.608149112933571,
      "tokens_seen": 129302528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006744136561376963,
      "loss": 3.3349,
      "theoretical_loss": 4.607837105572816,
      "tokens_seen": 129368064
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006743971061093248,
      "loss": 3.2678,
      "theoretical_loss": 4.607525300461299,
      "tokens_seen": 129433600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006743805560809532,
      "loss": 3.5554,
      "theoretical_loss": 4.607213697365613,
      "tokens_seen": 129499136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006743640060525818,
      "loss": 3.2279,
      "theoretical_loss": 4.606902296052739,
      "tokens_seen": 129564672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006743474560242103,
      "loss": 3.2397,
      "theoretical_loss": 4.6065910962900425,
      "tokens_seen": 129630208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006743309059958389,
      "loss": 3.422,
      "theoretical_loss": 4.606280097845277,
      "tokens_seen": 129695744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006743143559674674,
      "loss": 3.352,
      "theoretical_loss": 4.60596930048658,
      "tokens_seen": 129761280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006742978059390959,
      "loss": 3.2104,
      "theoretical_loss": 4.605658703982471,
      "tokens_seen": 129826816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006742812559107245,
      "loss": 3.4532,
      "theoretical_loss": 4.6053483081018545,
      "tokens_seen": 129892352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006742647058823529,
      "loss": 3.0227,
      "theoretical_loss": 4.605038112614018,
      "tokens_seen": 129957888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006742481558539814,
      "loss": 3.4461,
      "theoretical_loss": 4.604728117288631,
      "tokens_seen": 130023424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00067423160582561,
      "loss": 2.7931,
      "theoretical_loss": 4.604418321895739,
      "tokens_seen": 130088960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006742150557972385,
      "loss": 3.0268,
      "theoretical_loss": 4.604108726205774,
      "tokens_seen": 130154496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006741985057688671,
      "loss": 3.3372,
      "theoretical_loss": 4.60379932998954,
      "tokens_seen": 130220032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006741819557404956,
      "loss": 3.3281,
      "theoretical_loss": 4.6034901330182265,
      "tokens_seen": 130285568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006741654057121241,
      "loss": 3.2123,
      "theoretical_loss": 4.603181135063394,
      "tokens_seen": 130351104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006741488556837527,
      "loss": 3.4802,
      "theoretical_loss": 4.6028723358969845,
      "tokens_seen": 130416640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006741323056553811,
      "loss": 3.0854,
      "theoretical_loss": 4.602563735291312,
      "tokens_seen": 130482176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006741157556270096,
      "loss": 3.3232,
      "theoretical_loss": 4.602255333019068,
      "tokens_seen": 130547712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006740992055986382,
      "loss": 3.1927,
      "theoretical_loss": 4.6019471288533165,
      "tokens_seen": 130613248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006740826555702667,
      "loss": 2.9752,
      "theoretical_loss": 4.601639122567497,
      "tokens_seen": 130678784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006740661055418953,
      "loss": 3.1402,
      "theoretical_loss": 4.601331313935418,
      "tokens_seen": 130744320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006740495555135238,
      "loss": 3.323,
      "theoretical_loss": 4.601023702731264,
      "tokens_seen": 130809856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006740330054851523,
      "loss": 3.2923,
      "theoretical_loss": 4.600716288729587,
      "tokens_seen": 130875392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006740164554567808,
      "loss": 3.4908,
      "theoretical_loss": 4.600409071705312,
      "tokens_seen": 130940928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006739999054284093,
      "loss": 3.0826,
      "theoretical_loss": 4.60010205143373,
      "tokens_seen": 131006464
    },
    {
      "epoch": 0.05,
      "objective/train/avg_token_score": 0.01248980313539505,
      "objective/train/avg_weight": 0.9853021502494812,
      "objective/train/docs_used": 84906,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1505138874053955,
      "objective/train/std_weight": 0.07924885302782059,
      "objective/train/theoretical_loss": 4.599795227690505,
      "objective/train/tokens_used": 151532000,
      "theoretical_loss": 4.599795227690505,
      "tokens_seen": 131072000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006739833554000378,
      "loss": 3.1072,
      "theoretical_loss": 4.599795227690505,
      "tokens_seen": 131072000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006739668053716664,
      "loss": 3.0892,
      "theoretical_loss": 4.5994886002516635,
      "tokens_seen": 131137536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006739502553432949,
      "loss": 3.4964,
      "theoretical_loss": 4.599182168893604,
      "tokens_seen": 131203072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006739337053149234,
      "loss": 3.4707,
      "theoretical_loss": 4.598875933393089,
      "tokens_seen": 131268608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006739171552865519,
      "loss": 3.3255,
      "theoretical_loss": 4.5985698935272445,
      "tokens_seen": 131334144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006739006052581804,
      "loss": 3.3944,
      "theoretical_loss": 4.598264049073565,
      "tokens_seen": 131399680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006738840552298089,
      "loss": 3.3378,
      "theoretical_loss": 4.597958399809908,
      "tokens_seen": 131465216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006738675052014374,
      "loss": 3.3531,
      "theoretical_loss": 4.59765294551449,
      "tokens_seen": 131530752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000673850955173066,
      "loss": 3.4117,
      "theoretical_loss": 4.597347685965897,
      "tokens_seen": 131596288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006738344051446945,
      "loss": 3.4614,
      "theoretical_loss": 4.597042620943069,
      "tokens_seen": 131661824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000673817855116323,
      "loss": 3.2928,
      "theoretical_loss": 4.596737750225311,
      "tokens_seen": 131727360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006738013050879516,
      "loss": 3.3988,
      "theoretical_loss": 4.596433073592289,
      "tokens_seen": 131792896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00067378475505958,
      "loss": 3.2741,
      "theoretical_loss": 4.596128590824026,
      "tokens_seen": 131858432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006737682050312086,
      "loss": 3.1204,
      "theoretical_loss": 4.595824301700904,
      "tokens_seen": 131923968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006737516550028371,
      "loss": 3.411,
      "theoretical_loss": 4.595520206003663,
      "tokens_seen": 131989504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006737351049744656,
      "loss": 3.0964,
      "theoretical_loss": 4.595216303513399,
      "tokens_seen": 132055040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006737185549460942,
      "loss": 3.4585,
      "theoretical_loss": 4.594912594011566,
      "tokens_seen": 132120576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006737020049177227,
      "loss": 3.32,
      "theoretical_loss": 4.594609077279973,
      "tokens_seen": 132186112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006736854548893512,
      "loss": 3.2168,
      "theoretical_loss": 4.594305753100782,
      "tokens_seen": 132251648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006736689048609798,
      "loss": 3.3101,
      "theoretical_loss": 4.594002621256511,
      "tokens_seen": 132317184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006736523548326082,
      "loss": 3.2215,
      "theoretical_loss": 4.59369968153003,
      "tokens_seen": 132382720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006736358048042368,
      "loss": 3.288,
      "theoretical_loss": 4.593396933704562,
      "tokens_seen": 132448256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006736192547758653,
      "loss": 3.3153,
      "theoretical_loss": 4.593094377563681,
      "tokens_seen": 132513792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006736027047474938,
      "loss": 3.4509,
      "theoretical_loss": 4.592792012891314,
      "tokens_seen": 132579328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006735861547191224,
      "loss": 3.7005,
      "theoretical_loss": 4.592489839471735,
      "tokens_seen": 132644864
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006735696046907509,
      "loss": 3.3348,
      "theoretical_loss": 4.592187857089571,
      "tokens_seen": 132710400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006735530546623794,
      "loss": 3.2826,
      "theoretical_loss": 4.591886065529795,
      "tokens_seen": 132775936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006735365046340079,
      "loss": 3.22,
      "theoretical_loss": 4.591584464577728,
      "tokens_seen": 132841472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006735199546056364,
      "loss": 3.2512,
      "theoretical_loss": 4.591283054019041,
      "tokens_seen": 132907008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000673503404577265,
      "loss": 3.3664,
      "theoretical_loss": 4.5909818336397485,
      "tokens_seen": 132972544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006734868545488935,
      "loss": 3.3486,
      "theoretical_loss": 4.590680803226213,
      "tokens_seen": 133038080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000673470304520522,
      "loss": 3.3694,
      "theoretical_loss": 4.590379962565141,
      "tokens_seen": 133103616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006734537544921506,
      "loss": 3.2306,
      "theoretical_loss": 4.590079311443583,
      "tokens_seen": 133169152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006734372044637791,
      "loss": 3.1966,
      "theoretical_loss": 4.589778849648934,
      "tokens_seen": 133234688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006734206544354077,
      "loss": 3.4411,
      "theoretical_loss": 4.589478576968932,
      "tokens_seen": 133300224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006734041044070361,
      "loss": 3.2543,
      "theoretical_loss": 4.589178493191655,
      "tokens_seen": 133365760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006733875543786646,
      "loss": 3.2894,
      "theoretical_loss": 4.588878598105527,
      "tokens_seen": 133431296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006733710043502932,
      "loss": 3.3215,
      "theoretical_loss": 4.588578891499308,
      "tokens_seen": 133496832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006733544543219217,
      "loss": 3.1946,
      "theoretical_loss": 4.588279373162101,
      "tokens_seen": 133562368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006733379042935502,
      "loss": 3.1186,
      "theoretical_loss": 4.587980042883347,
      "tokens_seen": 133627904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006733213542651788,
      "loss": 3.2237,
      "theoretical_loss": 4.587680900452824,
      "tokens_seen": 133693440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006733048042368073,
      "loss": 3.3602,
      "theoretical_loss": 4.587381945660653,
      "tokens_seen": 133758976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006732882542084358,
      "loss": 3.1881,
      "theoretical_loss": 4.587083178297288,
      "tokens_seen": 133824512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006732717041800643,
      "loss": 3.4178,
      "theoretical_loss": 4.5867845981535185,
      "tokens_seen": 133890048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006732551541516928,
      "loss": 3.1325,
      "theoretical_loss": 4.586486205020474,
      "tokens_seen": 133955584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006732386041233214,
      "loss": 3.2069,
      "theoretical_loss": 4.586187998689616,
      "tokens_seen": 134021120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006732220540949499,
      "loss": 3.2783,
      "theoretical_loss": 4.585889978952741,
      "tokens_seen": 134086656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006732055040665784,
      "loss": 3.1266,
      "theoretical_loss": 4.58559214560198,
      "tokens_seen": 134152192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000673188954038207,
      "loss": 3.5458,
      "theoretical_loss": 4.585294498429796,
      "tokens_seen": 134217728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006731724040098354,
      "loss": 3.3352,
      "theoretical_loss": 4.584997037228986,
      "tokens_seen": 134283264
    },
    {
      "epoch": 0.05,
      "objective/train/avg_token_score": 0.012648314237594604,
      "objective/train/avg_weight": 0.98512202501297,
      "objective/train/docs_used": 86934,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3703534603118896,
      "objective/train/std_weight": 0.06850993633270264,
      "objective/train/theoretical_loss": 4.584699761792674,
      "objective/train/tokens_used": 154808800,
      "theoretical_loss": 4.584699761792674,
      "tokens_seen": 134348800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000673155853981464,
      "loss": 3.27,
      "theoretical_loss": 4.584699761792674,
      "tokens_seen": 134348800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006731393039530925,
      "loss": 3.4574,
      "theoretical_loss": 4.5844026719143205,
      "tokens_seen": 134414336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000673122753924721,
      "loss": 3.5042,
      "theoretical_loss": 4.5841057673877135,
      "tokens_seen": 134479872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006731062038963496,
      "loss": 3.33,
      "theoretical_loss": 4.5838090480069695,
      "tokens_seen": 134545408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006730896538679781,
      "loss": 3.4397,
      "theoretical_loss": 4.5835125135665375,
      "tokens_seen": 134610944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006730731038396066,
      "loss": 3.2247,
      "theoretical_loss": 4.583216163861191,
      "tokens_seen": 134676480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006730565538112352,
      "loss": 3.0804,
      "theoretical_loss": 4.58291999868603,
      "tokens_seen": 134742016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006730400037828636,
      "loss": 3.32,
      "theoretical_loss": 4.582624017836489,
      "tokens_seen": 134807552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006730234537544922,
      "loss": 3.3253,
      "theoretical_loss": 4.582328221108318,
      "tokens_seen": 134873088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006730069037261207,
      "loss": 3.2295,
      "theoretical_loss": 4.5820326082976,
      "tokens_seen": 134938624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006729903536977492,
      "loss": 3.1209,
      "theoretical_loss": 4.581737179200739,
      "tokens_seen": 135004160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006729738036693778,
      "loss": 3.2666,
      "theoretical_loss": 4.581441933614466,
      "tokens_seen": 135069696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006729572536410062,
      "loss": 3.274,
      "theoretical_loss": 4.581146871335832,
      "tokens_seen": 135135232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006729407036126348,
      "loss": 3.3469,
      "theoretical_loss": 4.580851992162214,
      "tokens_seen": 135200768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006729241535842632,
      "loss": 3.2187,
      "theoretical_loss": 4.5805572958913086,
      "tokens_seen": 135266304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006729076035558917,
      "loss": 3.3451,
      "theoretical_loss": 4.580262782321135,
      "tokens_seen": 135331840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006728910535275203,
      "loss": 3.3705,
      "theoretical_loss": 4.579968451250032,
      "tokens_seen": 135397376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006728745034991488,
      "loss": 3.233,
      "theoretical_loss": 4.579674302476661,
      "tokens_seen": 135462912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006728579534707774,
      "loss": 3.1377,
      "theoretical_loss": 4.579380335800001,
      "tokens_seen": 135528448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006728414034424059,
      "loss": 3.2981,
      "theoretical_loss": 4.579086551019348,
      "tokens_seen": 135593984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006728248534140344,
      "loss": 3.1261,
      "theoretical_loss": 4.5787929479343195,
      "tokens_seen": 135659520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000672808303385663,
      "loss": 3.4668,
      "theoretical_loss": 4.578499526344848,
      "tokens_seen": 135725056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006727917533572914,
      "loss": 3.4522,
      "theoretical_loss": 4.578206286051184,
      "tokens_seen": 135790592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006727752033289199,
      "loss": 3.2669,
      "theoretical_loss": 4.5779132268538945,
      "tokens_seen": 135856128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006727586533005485,
      "loss": 3.073,
      "theoretical_loss": 4.577620348553859,
      "tokens_seen": 135921664
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000672742103272177,
      "loss": 3.4576,
      "theoretical_loss": 4.577327650952276,
      "tokens_seen": 135987200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006727255532438056,
      "loss": 2.8302,
      "theoretical_loss": 4.5770351338506545,
      "tokens_seen": 136052736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006727090032154341,
      "loss": 3.0392,
      "theoretical_loss": 4.57674279705082,
      "tokens_seen": 136118272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006726924531870625,
      "loss": 3.2055,
      "theoretical_loss": 4.57645064035491,
      "tokens_seen": 136183808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006726759031586911,
      "loss": 3.5626,
      "theoretical_loss": 4.576158663565371,
      "tokens_seen": 136249344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006726593531303196,
      "loss": 3.1795,
      "theoretical_loss": 4.575866866484967,
      "tokens_seen": 136314880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006726428031019481,
      "loss": 3.1808,
      "theoretical_loss": 4.575575248916767,
      "tokens_seen": 136380416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006726262530735767,
      "loss": 3.3395,
      "theoretical_loss": 4.575283810664155,
      "tokens_seen": 136445952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006726097030452052,
      "loss": 3.2797,
      "theoretical_loss": 4.574992551530822,
      "tokens_seen": 136511488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006725931530168338,
      "loss": 3.3452,
      "theoretical_loss": 4.574701471320768,
      "tokens_seen": 136577024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006725766029884623,
      "loss": 3.1374,
      "theoretical_loss": 4.574410569838304,
      "tokens_seen": 136642560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006725600529600907,
      "loss": 3.3728,
      "theoretical_loss": 4.574119846888045,
      "tokens_seen": 136708096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006725435029317193,
      "loss": 3.4457,
      "theoretical_loss": 4.573829302274915,
      "tokens_seen": 136773632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006725269529033478,
      "loss": 3.3588,
      "theoretical_loss": 4.573538935804146,
      "tokens_seen": 136839168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006725104028749763,
      "loss": 2.9501,
      "theoretical_loss": 4.573248747281273,
      "tokens_seen": 136904704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006724938528466049,
      "loss": 3.0974,
      "theoretical_loss": 4.5729587365121365,
      "tokens_seen": 136970240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006724773028182334,
      "loss": 3.3543,
      "theoretical_loss": 4.572668903302886,
      "tokens_seen": 137035776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000672460752789862,
      "loss": 3.2222,
      "theoretical_loss": 4.572379247459969,
      "tokens_seen": 137101312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006724442027614904,
      "loss": 3.1941,
      "theoretical_loss": 4.57208976879014,
      "tokens_seen": 137166848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006724276527331189,
      "loss": 3.5945,
      "theoretical_loss": 4.571800467100456,
      "tokens_seen": 137232384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006724111027047475,
      "loss": 3.2209,
      "theoretical_loss": 4.5715113421982725,
      "tokens_seen": 137297920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000672394552676376,
      "loss": 3.1278,
      "theoretical_loss": 4.571222393891253,
      "tokens_seen": 137363456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006723780026480045,
      "loss": 3.2807,
      "theoretical_loss": 4.570933621987356,
      "tokens_seen": 137428992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006723614526196331,
      "loss": 3.2206,
      "theoretical_loss": 4.570645026294844,
      "tokens_seen": 137494528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006723449025912616,
      "loss": 3.0663,
      "theoretical_loss": 4.570356606622278,
      "tokens_seen": 137560064
    },
    {
      "epoch": 0.05,
      "objective/train/avg_token_score": 0.010820876806974411,
      "objective/train/avg_weight": 0.9872872233390808,
      "objective/train/docs_used": 89023,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.452348470687866,
      "objective/train/std_weight": 0.051426272839307785,
      "objective/train/theoretical_loss": 4.570068362778516,
      "objective/train/tokens_used": 158085600,
      "theoretical_loss": 4.570068362778516,
      "tokens_seen": 137625600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006723283525628902,
      "loss": 3.3175,
      "theoretical_loss": 4.570068362778516,
      "tokens_seen": 137625600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006723118025345186,
      "loss": 3.1992,
      "theoretical_loss": 4.569780294572718,
      "tokens_seen": 137691136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006722952525061471,
      "loss": 3.2359,
      "theoretical_loss": 4.569492401814339,
      "tokens_seen": 137756672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006722787024777757,
      "loss": 3.1201,
      "theoretical_loss": 4.569204684313133,
      "tokens_seen": 137822208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006722621524494042,
      "loss": 3.2457,
      "theoretical_loss": 4.568917141879149,
      "tokens_seen": 137887744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006722456024210327,
      "loss": 3.3986,
      "theoretical_loss": 4.568629774322736,
      "tokens_seen": 137953280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006722290523926613,
      "loss": 3.3989,
      "theoretical_loss": 4.568342581454532,
      "tokens_seen": 138018816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006722125023642898,
      "loss": 3.2283,
      "theoretical_loss": 4.568055563085476,
      "tokens_seen": 138084352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006721959523359183,
      "loss": 3.3473,
      "theoretical_loss": 4.567768719026797,
      "tokens_seen": 138149888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006721794023075468,
      "loss": 3.169,
      "theoretical_loss": 4.567482049090019,
      "tokens_seen": 138215424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006721628522791753,
      "loss": 3.3483,
      "theoretical_loss": 4.567195553086961,
      "tokens_seen": 138280960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006721463022508039,
      "loss": 3.4332,
      "theoretical_loss": 4.566909230829729,
      "tokens_seen": 138346496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006721297522224324,
      "loss": 3.2389,
      "theoretical_loss": 4.566623082130729,
      "tokens_seen": 138412032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006721132021940609,
      "loss": 3.1587,
      "theoretical_loss": 4.566337106802651,
      "tokens_seen": 138477568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006720966521656895,
      "loss": 3.0632,
      "theoretical_loss": 4.56605130465848,
      "tokens_seen": 138543104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000672080102137318,
      "loss": 3.3083,
      "theoretical_loss": 4.565765675511487,
      "tokens_seen": 138608640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006720635521089465,
      "loss": 3.0383,
      "theoretical_loss": 4.565480219175237,
      "tokens_seen": 138674176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000672047002080575,
      "loss": 3.3131,
      "theoretical_loss": 4.56519493546358,
      "tokens_seen": 138739712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006720304520522035,
      "loss": 3.4541,
      "theoretical_loss": 4.56490982419066,
      "tokens_seen": 138805248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006720139020238321,
      "loss": 3.3006,
      "theoretical_loss": 4.564624885170902,
      "tokens_seen": 138870784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006719973519954606,
      "loss": 3.4936,
      "theoretical_loss": 4.564340118219022,
      "tokens_seen": 138936320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006719808019670892,
      "loss": 3.2611,
      "theoretical_loss": 4.56405552315002,
      "tokens_seen": 139001856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006719642519387175,
      "loss": 3.2125,
      "theoretical_loss": 4.563771099779187,
      "tokens_seen": 139067392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006719477019103461,
      "loss": 3.4071,
      "theoretical_loss": 4.563486847922093,
      "tokens_seen": 139132928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006719311518819746,
      "loss": 3.3687,
      "theoretical_loss": 4.563202767394597,
      "tokens_seen": 139198464
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006719146018536031,
      "loss": 3.1747,
      "theoretical_loss": 4.562918858012843,
      "tokens_seen": 139264000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006718980518252317,
      "loss": 3.281,
      "theoretical_loss": 4.562635119593255,
      "tokens_seen": 139329536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006718815017968602,
      "loss": 3.2813,
      "theoretical_loss": 4.562351551952542,
      "tokens_seen": 139395072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006718649517684887,
      "loss": 3.4842,
      "theoretical_loss": 4.5620681549076965,
      "tokens_seen": 139460608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006718484017401173,
      "loss": 3.4465,
      "theoretical_loss": 4.561784928275992,
      "tokens_seen": 139526144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006718318517117457,
      "loss": 3.3164,
      "theoretical_loss": 4.561501871874984,
      "tokens_seen": 139591680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006718153016833743,
      "loss": 3.147,
      "theoretical_loss": 4.561218985522507,
      "tokens_seen": 139657216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006717987516550028,
      "loss": 3.2495,
      "theoretical_loss": 4.560936269036679,
      "tokens_seen": 139722752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006717822016266313,
      "loss": 3.3709,
      "theoretical_loss": 4.560653722235895,
      "tokens_seen": 139788288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006717656515982599,
      "loss": 2.9967,
      "theoretical_loss": 4.560371344938831,
      "tokens_seen": 139853824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006717491015698884,
      "loss": 3.3588,
      "theoretical_loss": 4.560089136964439,
      "tokens_seen": 139919360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006717325515415169,
      "loss": 3.1494,
      "theoretical_loss": 4.559807098131953,
      "tokens_seen": 139984896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006717160015131454,
      "loss": 3.1717,
      "theoretical_loss": 4.559525228260882,
      "tokens_seen": 140050432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006716994514847739,
      "loss": 3.3764,
      "theoretical_loss": 4.559243527171011,
      "tokens_seen": 140115968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006716829014564025,
      "loss": 3.4461,
      "theoretical_loss": 4.558961994682403,
      "tokens_seen": 140181504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000671666351428031,
      "loss": 3.4839,
      "theoretical_loss": 4.558680630615397,
      "tokens_seen": 140247040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006716498013996595,
      "loss": 3.3972,
      "theoretical_loss": 4.558399434790607,
      "tokens_seen": 140312576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006716332513712881,
      "loss": 3.174,
      "theoretical_loss": 4.558118407028921,
      "tokens_seen": 140378112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006716167013429166,
      "loss": 3.3071,
      "theoretical_loss": 4.557837547151502,
      "tokens_seen": 140443648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000671600151314545,
      "loss": 3.4222,
      "theoretical_loss": 4.557556854979786,
      "tokens_seen": 140509184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006715836012861736,
      "loss": 3.3908,
      "theoretical_loss": 4.5572763303354815,
      "tokens_seen": 140574720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006715670512578021,
      "loss": 3.2932,
      "theoretical_loss": 4.556995973040574,
      "tokens_seen": 140640256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006715505012294307,
      "loss": 3.457,
      "theoretical_loss": 4.556715782917314,
      "tokens_seen": 140705792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006715339512010592,
      "loss": 3.1398,
      "theoretical_loss": 4.556435759788229,
      "tokens_seen": 140771328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006715174011726877,
      "loss": 3.1001,
      "theoretical_loss": 4.556155903476114,
      "tokens_seen": 140836864
    },
    {
      "epoch": 0.05,
      "objective/train/avg_token_score": 0.013479014858603477,
      "objective/train/avg_weight": 0.9841359853744507,
      "objective/train/docs_used": 91007,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9468722343444824,
      "objective/train/std_weight": 0.0776786282658577,
      "objective/train/theoretical_loss": 4.555876213804037,
      "objective/train/tokens_used": 161362400,
      "theoretical_loss": 4.555876213804037,
      "tokens_seen": 140902400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006715008511443163,
      "loss": 3.009,
      "theoretical_loss": 4.555876213804037,
      "tokens_seen": 140902400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006714843011159448,
      "loss": 3.1555,
      "theoretical_loss": 4.555596690595333,
      "tokens_seen": 140967936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006714677510875732,
      "loss": 3.0145,
      "theoretical_loss": 4.555317333673611,
      "tokens_seen": 141033472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006714512010592018,
      "loss": 3.3164,
      "theoretical_loss": 4.555038142862742,
      "tokens_seen": 141099008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006714346510308303,
      "loss": 3.2595,
      "theoretical_loss": 4.5547591179868725,
      "tokens_seen": 141164544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006714181010024589,
      "loss": 3.2374,
      "theoretical_loss": 4.554480258870409,
      "tokens_seen": 141230080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006714015509740874,
      "loss": 3.4366,
      "theoretical_loss": 4.554201565338033,
      "tokens_seen": 141295616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006713850009457159,
      "loss": 3.0689,
      "theoretical_loss": 4.5539230372146875,
      "tokens_seen": 141361152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006713684509173445,
      "loss": 3.5831,
      "theoretical_loss": 4.553644674325584,
      "tokens_seen": 141426688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000671351900888973,
      "loss": 3.3294,
      "theoretical_loss": 4.553366476496198,
      "tokens_seen": 141492224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006713353508606014,
      "loss": 3.3677,
      "theoretical_loss": 4.553088443552269,
      "tokens_seen": 141557760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00067131880083223,
      "loss": 3.2724,
      "theoretical_loss": 4.552810575319806,
      "tokens_seen": 141623296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006713022508038585,
      "loss": 3.5094,
      "theoretical_loss": 4.552532871625077,
      "tokens_seen": 141688832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006712857007754871,
      "loss": 3.4322,
      "theoretical_loss": 4.5522553322946155,
      "tokens_seen": 141754368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006712691507471156,
      "loss": 3.145,
      "theoretical_loss": 4.551977957155217,
      "tokens_seen": 141819904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006712526007187441,
      "loss": 3.1994,
      "theoretical_loss": 4.5517007460339425,
      "tokens_seen": 141885440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006712360506903727,
      "loss": 3.3525,
      "theoretical_loss": 4.551423698758111,
      "tokens_seen": 141950976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006712195006620011,
      "loss": 3.2194,
      "theoretical_loss": 4.551146815155304,
      "tokens_seen": 142016512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006712029506336296,
      "loss": 3.3822,
      "theoretical_loss": 4.550870095053366,
      "tokens_seen": 142082048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006711864006052582,
      "loss": 3.2819,
      "theoretical_loss": 4.550593538280398,
      "tokens_seen": 142147584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006711698505768867,
      "loss": 3.3036,
      "theoretical_loss": 4.550317144664766,
      "tokens_seen": 142213120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006711533005485153,
      "loss": 3.366,
      "theoretical_loss": 4.55004091403509,
      "tokens_seen": 142278656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006711367505201438,
      "loss": 3.2946,
      "theoretical_loss": 4.5497648462202545,
      "tokens_seen": 142344192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006711202004917723,
      "loss": 3.0611,
      "theoretical_loss": 4.549488941049397,
      "tokens_seen": 142409728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006711036504634008,
      "loss": 3.2784,
      "theoretical_loss": 4.549213198351914,
      "tokens_seen": 142475264
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006710871004350293,
      "loss": 3.4699,
      "theoretical_loss": 4.548937617957463,
      "tokens_seen": 142540800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006710705504066578,
      "loss": 3.4366,
      "theoretical_loss": 4.548662199695954,
      "tokens_seen": 142606336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006710540003782864,
      "loss": 3.2834,
      "theoretical_loss": 4.548386943397556,
      "tokens_seen": 142671872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006710374503499149,
      "loss": 3.4094,
      "theoretical_loss": 4.548111848892693,
      "tokens_seen": 142737408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006710209003215435,
      "loss": 3.4395,
      "theoretical_loss": 4.547836916012042,
      "tokens_seen": 142802944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000671004350293172,
      "loss": 3.3795,
      "theoretical_loss": 4.547562144586539,
      "tokens_seen": 142868480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006709878002648004,
      "loss": 3.0551,
      "theoretical_loss": 4.547287534447372,
      "tokens_seen": 142934016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006709712502364289,
      "loss": 3.122,
      "theoretical_loss": 4.5470130854259825,
      "tokens_seen": 142999552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006709547002080574,
      "loss": 3.2366,
      "theoretical_loss": 4.546738797354065,
      "tokens_seen": 143065088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000670938150179686,
      "loss": 3.0637,
      "theoretical_loss": 4.546464670063569,
      "tokens_seen": 143130624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006709216001513145,
      "loss": 3.3183,
      "theoretical_loss": 4.546190703386695,
      "tokens_seen": 143196160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006709050501229431,
      "loss": 3.131,
      "theoretical_loss": 4.545916897155894,
      "tokens_seen": 143261696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006708885000945716,
      "loss": 3.2143,
      "theoretical_loss": 4.54564325120387,
      "tokens_seen": 143327232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006708719500662,
      "loss": 3.3951,
      "theoretical_loss": 4.545369765363578,
      "tokens_seen": 143392768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006708554000378286,
      "loss": 3.1986,
      "theoretical_loss": 4.545096439468223,
      "tokens_seen": 143458304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006708388500094571,
      "loss": 3.1921,
      "theoretical_loss": 4.544823273351257,
      "tokens_seen": 143523840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006708222999810856,
      "loss": 2.8971,
      "theoretical_loss": 4.544550266846388,
      "tokens_seen": 143589376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006708057499527142,
      "loss": 3.2634,
      "theoretical_loss": 4.544277419787566,
      "tokens_seen": 143654912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006707891999243427,
      "loss": 3.1916,
      "theoretical_loss": 4.544004732008993,
      "tokens_seen": 143720448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006707726498959713,
      "loss": 3.5607,
      "theoretical_loss": 4.543732203345119,
      "tokens_seen": 143785984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006707560998675998,
      "loss": 3.2747,
      "theoretical_loss": 4.543459833630639,
      "tokens_seen": 143851520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006707395498392282,
      "loss": 3.3526,
      "theoretical_loss": 4.543187622700497,
      "tokens_seen": 143917056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006707229998108568,
      "loss": 3.1754,
      "theoretical_loss": 4.542915570389884,
      "tokens_seen": 143982592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006707064497824853,
      "loss": 3.2145,
      "theoretical_loss": 4.542643676534234,
      "tokens_seen": 144048128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006706898997541138,
      "loss": 3.2383,
      "theoretical_loss": 4.542371940969231,
      "tokens_seen": 144113664
    },
    {
      "epoch": 0.05,
      "objective/train/avg_token_score": 0.019538067281246185,
      "objective/train/avg_weight": 0.9769995212554932,
      "objective/train/docs_used": 93093,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1995291709899902,
      "objective/train/std_weight": 0.10487210005521774,
      "objective/train/theoretical_loss": 4.542100363530799,
      "objective/train/tokens_used": 164639200,
      "theoretical_loss": 4.542100363530799,
      "tokens_seen": 144179200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006706733497257424,
      "loss": 3.1116,
      "theoretical_loss": 4.542100363530799,
      "tokens_seen": 144179200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006706567996973709,
      "loss": 3.067,
      "theoretical_loss": 4.54182894405511,
      "tokens_seen": 144244736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006706402496689995,
      "loss": 3.3223,
      "theoretical_loss": 4.5415576823785795,
      "tokens_seen": 144310272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000670623699640628,
      "loss": 3.1418,
      "theoretical_loss": 4.541286578337866,
      "tokens_seen": 144375808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006706071496122564,
      "loss": 2.9121,
      "theoretical_loss": 4.541015631769872,
      "tokens_seen": 144441344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000670590599583885,
      "loss": 3.232,
      "theoretical_loss": 4.5407448425117405,
      "tokens_seen": 144506880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006705740495555135,
      "loss": 3.1733,
      "theoretical_loss": 4.540474210400859,
      "tokens_seen": 144572416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000670557499527142,
      "loss": 3.3522,
      "theoretical_loss": 4.540203735274855,
      "tokens_seen": 144637952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006705409494987706,
      "loss": 3.2233,
      "theoretical_loss": 4.5399334169716,
      "tokens_seen": 144703488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006705243994703991,
      "loss": 3.2754,
      "theoretical_loss": 4.539663255329202,
      "tokens_seen": 144769024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006705078494420277,
      "loss": 3.1759,
      "theoretical_loss": 4.539393250186015,
      "tokens_seen": 144834560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006704912994136561,
      "loss": 3.1936,
      "theoretical_loss": 4.539123401380625,
      "tokens_seen": 144900096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006704747493852846,
      "loss": 3.5228,
      "theoretical_loss": 4.538853708751866,
      "tokens_seen": 144965632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006704581993569132,
      "loss": 3.2062,
      "theoretical_loss": 4.538584172138804,
      "tokens_seen": 145031168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006704416493285417,
      "loss": 3.3369,
      "theoretical_loss": 4.538314791380748,
      "tokens_seen": 145096704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006704250993001702,
      "loss": 3.2215,
      "theoretical_loss": 4.538045566317242,
      "tokens_seen": 145162240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006704085492717988,
      "loss": 3.1469,
      "theoretical_loss": 4.537776496788071,
      "tokens_seen": 145227776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006703919992434273,
      "loss": 3.2053,
      "theoretical_loss": 4.537507582633253,
      "tokens_seen": 145293312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006703754492150559,
      "loss": 3.0842,
      "theoretical_loss": 4.537238823693045,
      "tokens_seen": 145358848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006703588991866843,
      "loss": 3.2643,
      "theoretical_loss": 4.536970219807939,
      "tokens_seen": 145424384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006703423491583128,
      "loss": 3.3802,
      "theoretical_loss": 4.536701770818665,
      "tokens_seen": 145489920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006703257991299414,
      "loss": 3.2576,
      "theoretical_loss": 4.536433476566185,
      "tokens_seen": 145555456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006703092491015699,
      "loss": 3.1188,
      "theoretical_loss": 4.536165336891699,
      "tokens_seen": 145620992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006702926990731984,
      "loss": 3.3943,
      "theoretical_loss": 4.535897351636638,
      "tokens_seen": 145686528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000670276149044827,
      "loss": 3.2663,
      "theoretical_loss": 4.53562952064267,
      "tokens_seen": 145752064
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006702595990164555,
      "loss": 2.8304,
      "theoretical_loss": 4.535361843751696,
      "tokens_seen": 145817600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000670243048988084,
      "loss": 3.3734,
      "theoretical_loss": 4.535094320805847,
      "tokens_seen": 145883136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006702264989597125,
      "loss": 3.1985,
      "theoretical_loss": 4.534826951647489,
      "tokens_seen": 145948672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000670209948931341,
      "loss": 3.2417,
      "theoretical_loss": 4.5345597361192205,
      "tokens_seen": 146014208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006701933989029696,
      "loss": 3.2552,
      "theoretical_loss": 4.53429267406387,
      "tokens_seen": 146079744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006701768488745981,
      "loss": 3.2999,
      "theoretical_loss": 4.5340257653244995,
      "tokens_seen": 146145280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006701602988462266,
      "loss": 3.2579,
      "theoretical_loss": 4.5337590097444,
      "tokens_seen": 146210816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006701437488178552,
      "loss": 3.0928,
      "theoretical_loss": 4.533492407167093,
      "tokens_seen": 146276352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006701271987894836,
      "loss": 3.3161,
      "theoretical_loss": 4.53322595743633,
      "tokens_seen": 146341888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006701106487611122,
      "loss": 3.4366,
      "theoretical_loss": 4.5329596603960916,
      "tokens_seen": 146407424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006700940987327407,
      "loss": 3.3765,
      "theoretical_loss": 4.53269351589059,
      "tokens_seen": 146472960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006700775487043692,
      "loss": 3.2626,
      "theoretical_loss": 4.532427523764261,
      "tokens_seen": 146538496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006700609986759978,
      "loss": 3.3427,
      "theoretical_loss": 4.532161683861773,
      "tokens_seen": 146604032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006700444486476263,
      "loss": 3.3287,
      "theoretical_loss": 4.5318959960280205,
      "tokens_seen": 146669568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006700278986192548,
      "loss": 3.1837,
      "theoretical_loss": 4.531630460108125,
      "tokens_seen": 146735104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006700113485908832,
      "loss": 3.171,
      "theoretical_loss": 4.531365075947434,
      "tokens_seen": 146800640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006699947985625117,
      "loss": 3.2083,
      "theoretical_loss": 4.531099843391524,
      "tokens_seen": 146866176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006699782485341403,
      "loss": 3.2198,
      "theoretical_loss": 4.5308347622861955,
      "tokens_seen": 146931712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006699616985057688,
      "loss": 3.2982,
      "theoretical_loss": 4.5305698324774735,
      "tokens_seen": 146997248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006699451484773974,
      "loss": 3.1992,
      "theoretical_loss": 4.530305053811611,
      "tokens_seen": 147062784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006699285984490259,
      "loss": 3.359,
      "theoretical_loss": 4.530040426135084,
      "tokens_seen": 147128320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006699120484206544,
      "loss": 3.2868,
      "theoretical_loss": 4.529775949294593,
      "tokens_seen": 147193856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000669895498392283,
      "loss": 3.2725,
      "theoretical_loss": 4.529511623137061,
      "tokens_seen": 147259392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006698789483639114,
      "loss": 3.3968,
      "theoretical_loss": 4.529247447509637,
      "tokens_seen": 147324928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006698623983355399,
      "loss": 3.2024,
      "theoretical_loss": 4.528983422259691,
      "tokens_seen": 147390464
    },
    {
      "epoch": 0.05,
      "objective/train/avg_token_score": 0.012383748777210712,
      "objective/train/avg_weight": 0.9854242205619812,
      "objective/train/docs_used": 95084,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.629666328430176,
      "objective/train/std_weight": 0.09040210396051407,
      "objective/train/theoretical_loss": 4.528719547234816,
      "objective/train/tokens_used": 167916000,
      "theoretical_loss": 4.528719547234816,
      "tokens_seen": 147456000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006698458483071685,
      "loss": 3.5782,
      "theoretical_loss": 4.528719547234816,
      "tokens_seen": 147456000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000669829298278797,
      "loss": 3.1834,
      "theoretical_loss": 4.528455822282828,
      "tokens_seen": 147521536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006698127482504256,
      "loss": 3.2623,
      "theoretical_loss": 4.528192247251763,
      "tokens_seen": 147587072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006697961982220541,
      "loss": 3.2957,
      "theoretical_loss": 4.52792882198988,
      "tokens_seen": 147652608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006697796481936826,
      "loss": 3.1489,
      "theoretical_loss": 4.527665546345656,
      "tokens_seen": 147718144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006697630981653111,
      "loss": 3.395,
      "theoretical_loss": 4.5274024201677925,
      "tokens_seen": 147783680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006697465481369396,
      "loss": 3.2943,
      "theoretical_loss": 4.527139443305209,
      "tokens_seen": 147849216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006697299981085681,
      "loss": 3.3231,
      "theoretical_loss": 4.526876615607042,
      "tokens_seen": 147914752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006697134480801967,
      "loss": 3.302,
      "theoretical_loss": 4.526613936922654,
      "tokens_seen": 147980288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006696968980518252,
      "loss": 3.0376,
      "theoretical_loss": 4.526351407101618,
      "tokens_seen": 148045824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006696803480234538,
      "loss": 3.1229,
      "theoretical_loss": 4.526089025993732,
      "tokens_seen": 148111360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006696637979950823,
      "loss": 3.1173,
      "theoretical_loss": 4.525826793449008,
      "tokens_seen": 148176896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006696472479667107,
      "loss": 3.1225,
      "theoretical_loss": 4.525564709317678,
      "tokens_seen": 148242432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006696306979383393,
      "loss": 3.3502,
      "theoretical_loss": 4.525302773450187,
      "tokens_seen": 148307968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006696141479099678,
      "loss": 3.1978,
      "theoretical_loss": 4.525040985697203,
      "tokens_seen": 148373504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006695975978815963,
      "loss": 3.3738,
      "theoretical_loss": 4.524779345909604,
      "tokens_seen": 148439040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006695810478532249,
      "loss": 3.1129,
      "theoretical_loss": 4.524517853938489,
      "tokens_seen": 148504576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006695644978248534,
      "loss": 3.3215,
      "theoretical_loss": 4.524256509635169,
      "tokens_seen": 148570112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000669547947796482,
      "loss": 3.3302,
      "theoretical_loss": 4.523995312851174,
      "tokens_seen": 148635648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006695313977681105,
      "loss": 3.2271,
      "theoretical_loss": 4.523734263438241,
      "tokens_seen": 148701184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006695148477397389,
      "loss": 3.1881,
      "theoretical_loss": 4.52347336124833,
      "tokens_seen": 148766720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006694982977113675,
      "loss": 3.2077,
      "theoretical_loss": 4.52321260613361,
      "tokens_seen": 148832256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000669481747682996,
      "loss": 2.9924,
      "theoretical_loss": 4.522951997946466,
      "tokens_seen": 148897792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006694651976546246,
      "loss": 3.1829,
      "theoretical_loss": 4.522691536539492,
      "tokens_seen": 148963328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006694486476262531,
      "loss": 3.4841,
      "theoretical_loss": 4.522431221765498,
      "tokens_seen": 149028864
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006694320975978816,
      "loss": 3.3016,
      "theoretical_loss": 4.522171053477507,
      "tokens_seen": 149094400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006694155475695102,
      "loss": 3.3852,
      "theoretical_loss": 4.5219110315287505,
      "tokens_seen": 149159936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006693989975411386,
      "loss": 3.1109,
      "theoretical_loss": 4.521651155772675,
      "tokens_seen": 149225472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006693824475127671,
      "loss": 3.2475,
      "theoretical_loss": 4.521391426062934,
      "tokens_seen": 149291008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006693658974843957,
      "loss": 3.1225,
      "theoretical_loss": 4.521131842253396,
      "tokens_seen": 149356544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006693493474560242,
      "loss": 3.1138,
      "theoretical_loss": 4.520872404198139,
      "tokens_seen": 149422080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006693327974276528,
      "loss": 3.3876,
      "theoretical_loss": 4.520613111751445,
      "tokens_seen": 149487616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006693162473992813,
      "loss": 3.103,
      "theoretical_loss": 4.520353964767814,
      "tokens_seen": 149553152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006692996973709098,
      "loss": 3.235,
      "theoretical_loss": 4.5200949631019505,
      "tokens_seen": 149618688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006692831473425384,
      "loss": 3.0898,
      "theoretical_loss": 4.519836106608768,
      "tokens_seen": 149684224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006692665973141668,
      "loss": 3.2133,
      "theoretical_loss": 4.519577395143388,
      "tokens_seen": 149749760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006692500472857953,
      "loss": 3.063,
      "theoretical_loss": 4.519318828561142,
      "tokens_seen": 149815296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006692334972574239,
      "loss": 3.1379,
      "theoretical_loss": 4.519060406717565,
      "tokens_seen": 149880832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006692169472290524,
      "loss": 2.9823,
      "theoretical_loss": 4.518802129468405,
      "tokens_seen": 149946368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000669200397200681,
      "loss": 2.9909,
      "theoretical_loss": 4.51854399666961,
      "tokens_seen": 150011904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006691838471723095,
      "loss": 3.2517,
      "theoretical_loss": 4.518286008177341,
      "tokens_seen": 150077440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000669167297143938,
      "loss": 3.1792,
      "theoretical_loss": 4.51802816384796,
      "tokens_seen": 150142976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006691507471155665,
      "loss": 3.3512,
      "theoretical_loss": 4.517770463538038,
      "tokens_seen": 150208512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000669134197087195,
      "loss": 3.1702,
      "theoretical_loss": 4.517512907104347,
      "tokens_seen": 150274048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006691176470588235,
      "loss": 3.1356,
      "theoretical_loss": 4.517255494403868,
      "tokens_seen": 150339584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006691010970304521,
      "loss": 3.0778,
      "theoretical_loss": 4.516998225293785,
      "tokens_seen": 150405120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006690845470020806,
      "loss": 3.0681,
      "theoretical_loss": 4.516741099631485,
      "tokens_seen": 150470656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006690679969737092,
      "loss": 3.3494,
      "theoretical_loss": 4.51648411727456,
      "tokens_seen": 150536192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006690514469453377,
      "loss": 3.2784,
      "theoretical_loss": 4.5162272780808035,
      "tokens_seen": 150601728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006690348969169661,
      "loss": 3.308,
      "theoretical_loss": 4.515970581908216,
      "tokens_seen": 150667264
    },
    {
      "epoch": 0.05,
      "objective/train/avg_token_score": 0.008557405322790146,
      "objective/train/avg_weight": 0.9899294972419739,
      "objective/train/docs_used": 96518,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0413498878479004,
      "objective/train/std_weight": 0.05580054968595505,
      "objective/train/theoretical_loss": 4.515714028614996,
      "objective/train/tokens_used": 171192800,
      "theoretical_loss": 4.515714028614996,
      "tokens_seen": 150732800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006690183468885946,
      "loss": 3.0952,
      "theoretical_loss": 4.515714028614996,
      "tokens_seen": 150732800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006690017968602231,
      "loss": 3.1197,
      "theoretical_loss": 4.515457618059546,
      "tokens_seen": 150798336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006689852468318517,
      "loss": 3.2579,
      "theoretical_loss": 4.515201350100471,
      "tokens_seen": 150863872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006689686968034802,
      "loss": 3.4278,
      "theoretical_loss": 4.514945224596577,
      "tokens_seen": 150929408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006689521467751087,
      "loss": 3.3717,
      "theoretical_loss": 4.5146892414068684,
      "tokens_seen": 150994944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006689355967467373,
      "loss": 3.2736,
      "theoretical_loss": 4.514433400390554,
      "tokens_seen": 151060480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006689190467183657,
      "loss": 3.2727,
      "theoretical_loss": 4.514177701407042,
      "tokens_seen": 151126016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006689024966899943,
      "loss": 3.3136,
      "theoretical_loss": 4.51392214431594,
      "tokens_seen": 151191552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006688859466616228,
      "loss": 3.4371,
      "theoretical_loss": 4.513666728977054,
      "tokens_seen": 151257088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006688693966332513,
      "loss": 3.0314,
      "theoretical_loss": 4.51341145525039,
      "tokens_seen": 151322624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006688528466048799,
      "loss": 3.2292,
      "theoretical_loss": 4.513156322996155,
      "tokens_seen": 151388160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006688362965765084,
      "loss": 3.0818,
      "theoretical_loss": 4.512901332074751,
      "tokens_seen": 151453696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006688197465481369,
      "loss": 3.2962,
      "theoretical_loss": 4.5126464823467805,
      "tokens_seen": 151519232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006688031965197655,
      "loss": 3.4124,
      "theoretical_loss": 4.512391773673042,
      "tokens_seen": 151584768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006687866464913939,
      "loss": 3.1299,
      "theoretical_loss": 4.5121372059145335,
      "tokens_seen": 151650304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006687700964630225,
      "loss": 3.0594,
      "theoretical_loss": 4.511882778932447,
      "tokens_seen": 151715840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000668753546434651,
      "loss": 3.2027,
      "theoretical_loss": 4.511628492588174,
      "tokens_seen": 151781376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006687369964062795,
      "loss": 3.1525,
      "theoretical_loss": 4.5113743467433,
      "tokens_seen": 151846912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006687204463779081,
      "loss": 3.1634,
      "theoretical_loss": 4.511120341259608,
      "tokens_seen": 151912448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006687038963495366,
      "loss": 3.3019,
      "theoretical_loss": 4.510866475999077,
      "tokens_seen": 151977984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000668687346321165,
      "loss": 3.0829,
      "theoretical_loss": 4.510612750823878,
      "tokens_seen": 152043520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006686707962927936,
      "loss": 3.2253,
      "theoretical_loss": 4.5103591655963795,
      "tokens_seen": 152109056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006686542462644221,
      "loss": 2.9987,
      "theoretical_loss": 4.510105720179144,
      "tokens_seen": 152174592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006686376962360507,
      "loss": 3.3013,
      "theoretical_loss": 4.5098524144349295,
      "tokens_seen": 152240128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006686211462076792,
      "loss": 2.9738,
      "theoretical_loss": 4.509599248226683,
      "tokens_seen": 152305664
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006686045961793077,
      "loss": 3.2746,
      "theoretical_loss": 4.509346221417552,
      "tokens_seen": 152371200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006685880461509363,
      "loss": 2.9407,
      "theoretical_loss": 4.509093333870869,
      "tokens_seen": 152436736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006685714961225648,
      "loss": 3.1001,
      "theoretical_loss": 4.508840585450166,
      "tokens_seen": 152502272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006685549460941932,
      "loss": 3.3558,
      "theoretical_loss": 4.508587976019164,
      "tokens_seen": 152567808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006685383960658218,
      "loss": 3.2936,
      "theoretical_loss": 4.508335505441774,
      "tokens_seen": 152633344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006685218460374503,
      "loss": 3.1358,
      "theoretical_loss": 4.508083173582105,
      "tokens_seen": 152698880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006685052960090789,
      "loss": 3.4045,
      "theoretical_loss": 4.507830980304451,
      "tokens_seen": 152764416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006684887459807074,
      "loss": 3.3668,
      "theoretical_loss": 4.5075789254733,
      "tokens_seen": 152829952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006684721959523359,
      "loss": 3.2517,
      "theoretical_loss": 4.507327008953329,
      "tokens_seen": 152895488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006684556459239645,
      "loss": 3.4368,
      "theoretical_loss": 4.507075230609407,
      "tokens_seen": 152961024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000668439095895593,
      "loss": 2.9808,
      "theoretical_loss": 4.506823590306591,
      "tokens_seen": 153026560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006684225458672214,
      "loss": 3.3349,
      "theoretical_loss": 4.506572087910127,
      "tokens_seen": 153092096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00066840599583885,
      "loss": 3.0714,
      "theoretical_loss": 4.506320723285455,
      "tokens_seen": 153157632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006683894458104785,
      "loss": 3.282,
      "theoretical_loss": 4.506069496298198,
      "tokens_seen": 153223168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006683728957821071,
      "loss": 3.1557,
      "theoretical_loss": 4.5058184068141705,
      "tokens_seen": 153288704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006683563457537356,
      "loss": 3.2302,
      "theoretical_loss": 4.505567454699373,
      "tokens_seen": 153354240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006683397957253641,
      "loss": 3.1381,
      "theoretical_loss": 4.505316639819997,
      "tokens_seen": 153419776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006683232456969927,
      "loss": 2.9087,
      "theoretical_loss": 4.505065962042418,
      "tokens_seen": 153485312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006683066956686211,
      "loss": 3.2657,
      "theoretical_loss": 4.504815421233202,
      "tokens_seen": 153550848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006682901456402496,
      "loss": 3.3894,
      "theoretical_loss": 4.504565017259097,
      "tokens_seen": 153616384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006682735956118782,
      "loss": 3.2539,
      "theoretical_loss": 4.504314749987044,
      "tokens_seen": 153681920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006682570455835067,
      "loss": 3.2433,
      "theoretical_loss": 4.504064619284163,
      "tokens_seen": 153747456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006682404955551353,
      "loss": 3.1482,
      "theoretical_loss": 4.503814625017766,
      "tokens_seen": 153812992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006682239455267638,
      "loss": 3.2569,
      "theoretical_loss": 4.5035647670553445,
      "tokens_seen": 153878528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0006682073954983923,
      "loss": 3.4582,
      "theoretical_loss": 4.503315045264581,
      "tokens_seen": 153944064
    },
    {
      "epoch": 0.05,
      "objective/train/avg_token_score": 0.009818554855883121,
      "objective/train/avg_weight": 0.988447904586792,
      "objective/train/docs_used": 98118,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3341314792633057,
      "objective/train/std_weight": 0.04770763963460922,
      "objective/train/theoretical_loss": 4.503065459513339,
      "objective/train/tokens_used": 174469600,
      "theoretical_loss": 4.503065459513339,
      "tokens_seen": 154009600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006681908454700209,
      "loss": 3.1674,
      "theoretical_loss": 4.503065459513339,
      "tokens_seen": 154009600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006681742954416493,
      "loss": 3.3122,
      "theoretical_loss": 4.502816009669665,
      "tokens_seen": 154075136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006681577454132778,
      "loss": 3.3258,
      "theoretical_loss": 4.502566695601795,
      "tokens_seen": 154140672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006681411953849064,
      "loss": 3.2261,
      "theoretical_loss": 4.502317517178142,
      "tokens_seen": 154206208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006681246453565349,
      "loss": 3.2944,
      "theoretical_loss": 4.502068474267309,
      "tokens_seen": 154271744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006681080953281635,
      "loss": 3.3748,
      "theoretical_loss": 4.501819566738076,
      "tokens_seen": 154337280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000668091545299792,
      "loss": 3.1597,
      "theoretical_loss": 4.501570794459411,
      "tokens_seen": 154402816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006680749952714205,
      "loss": 3.2503,
      "theoretical_loss": 4.501322157300461,
      "tokens_seen": 154468352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000668058445243049,
      "loss": 3.3669,
      "theoretical_loss": 4.501073655130554,
      "tokens_seen": 154533888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006680418952146774,
      "loss": 3.1101,
      "theoretical_loss": 4.500825287819205,
      "tokens_seen": 154599424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000668025345186306,
      "loss": 3.2273,
      "theoretical_loss": 4.500577055236104,
      "tokens_seen": 154664960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006680087951579345,
      "loss": 3.342,
      "theoretical_loss": 4.500328957251128,
      "tokens_seen": 154730496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006679922451295631,
      "loss": 3.1812,
      "theoretical_loss": 4.500080993734329,
      "tokens_seen": 154796032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006679756951011916,
      "loss": 3.1331,
      "theoretical_loss": 4.499833164555944,
      "tokens_seen": 154861568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006679591450728201,
      "loss": 3.1012,
      "theoretical_loss": 4.499585469586387,
      "tokens_seen": 154927104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006679425950444486,
      "loss": 3.0874,
      "theoretical_loss": 4.499337908696255,
      "tokens_seen": 154992640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006679260450160771,
      "loss": 2.9089,
      "theoretical_loss": 4.499090481756321,
      "tokens_seen": 155058176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006679094949877056,
      "loss": 3.2953,
      "theoretical_loss": 4.498843188637538,
      "tokens_seen": 155123712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006678929449593342,
      "loss": 3.3101,
      "theoretical_loss": 4.498596029211041,
      "tokens_seen": 155189248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006678763949309627,
      "loss": 3.0968,
      "theoretical_loss": 4.498349003348137,
      "tokens_seen": 155254784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006678598449025913,
      "loss": 3.3392,
      "theoretical_loss": 4.4981021109203185,
      "tokens_seen": 155320320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006678432948742198,
      "loss": 3.3681,
      "theoretical_loss": 4.49785535179925,
      "tokens_seen": 155385856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006678267448458482,
      "loss": 3.17,
      "theoretical_loss": 4.497608725856776,
      "tokens_seen": 155451392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006678101948174768,
      "loss": 3.1506,
      "theoretical_loss": 4.497362232964919,
      "tokens_seen": 155516928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006677936447891053,
      "loss": 3.3492,
      "theoretical_loss": 4.497115872995876,
      "tokens_seen": 155582464
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006677770947607338,
      "loss": 3.3569,
      "theoretical_loss": 4.496869645822022,
      "tokens_seen": 155648000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006677605447323624,
      "loss": 3.1448,
      "theoretical_loss": 4.496623551315908,
      "tokens_seen": 155713536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006677439947039909,
      "loss": 3.3314,
      "theoretical_loss": 4.496377589350261,
      "tokens_seen": 155779072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006677274446756195,
      "loss": 3.1825,
      "theoretical_loss": 4.496131759797984,
      "tokens_seen": 155844608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000667710894647248,
      "loss": 2.9718,
      "theoretical_loss": 4.495886062532153,
      "tokens_seen": 155910144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006676943446188764,
      "loss": 3.2791,
      "theoretical_loss": 4.495640497426023,
      "tokens_seen": 155975680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000667677794590505,
      "loss": 3.2665,
      "theoretical_loss": 4.495395064353019,
      "tokens_seen": 156041216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006676612445621335,
      "loss": 3.1148,
      "theoretical_loss": 4.4951497631867445,
      "tokens_seen": 156106752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000667644694533762,
      "loss": 3.3648,
      "theoretical_loss": 4.494904593800973,
      "tokens_seen": 156172288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006676281445053906,
      "loss": 3.1224,
      "theoretical_loss": 4.4946595560696565,
      "tokens_seen": 156237824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006676115944770191,
      "loss": 3.2874,
      "theoretical_loss": 4.494414649866915,
      "tokens_seen": 156303360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006675950444486477,
      "loss": 3.2063,
      "theoretical_loss": 4.494169875067046,
      "tokens_seen": 156368896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006675784944202761,
      "loss": 3.0616,
      "theoretical_loss": 4.493925231544516,
      "tokens_seen": 156434432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006675619443919046,
      "loss": 3.3659,
      "theoretical_loss": 4.493680719173968,
      "tokens_seen": 156499968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006675453943635332,
      "loss": 3.1235,
      "theoretical_loss": 4.4934363378302145,
      "tokens_seen": 156565504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006675288443351617,
      "loss": 3.3044,
      "theoretical_loss": 4.493192087388239,
      "tokens_seen": 156631040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006675122943067902,
      "loss": 3.1246,
      "theoretical_loss": 4.4929479677232,
      "tokens_seen": 156696576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006674957442784188,
      "loss": 3.4389,
      "theoretical_loss": 4.4927039787104235,
      "tokens_seen": 156762112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006674791942500473,
      "loss": 3.3871,
      "theoretical_loss": 4.4924601202254095,
      "tokens_seen": 156827648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006674626442216759,
      "loss": 3.4837,
      "theoretical_loss": 4.492216392143826,
      "tokens_seen": 156893184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006674460941933043,
      "loss": 3.2316,
      "theoretical_loss": 4.491972794341514,
      "tokens_seen": 156958720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006674295441649328,
      "loss": 3.1416,
      "theoretical_loss": 4.49172932669448,
      "tokens_seen": 157024256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006674129941365614,
      "loss": 3.3576,
      "theoretical_loss": 4.491485989078906,
      "tokens_seen": 157089792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006673964441081899,
      "loss": 3.2039,
      "theoretical_loss": 4.491242781371138,
      "tokens_seen": 157155328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006673798940798184,
      "loss": 3.3419,
      "theoretical_loss": 4.490999703447697,
      "tokens_seen": 157220864
    },
    {
      "epoch": 0.06,
      "objective/train/avg_token_score": 0.01150190643966198,
      "objective/train/avg_weight": 0.9864696264266968,
      "objective/train/docs_used": 100657,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9648029804229736,
      "objective/train/std_weight": 0.05343075469136238,
      "objective/train/theoretical_loss": 4.4907567551852665,
      "objective/train/tokens_used": 177746400,
      "theoretical_loss": 4.4907567551852665,
      "tokens_seen": 157286400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000667363344051447,
      "loss": 2.8764,
      "theoretical_loss": 4.4907567551852665,
      "tokens_seen": 157286400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006673467940230755,
      "loss": 3.1593,
      "theoretical_loss": 4.490513936460702,
      "tokens_seen": 157351936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000667330243994704,
      "loss": 3.1925,
      "theoretical_loss": 4.490271247151027,
      "tokens_seen": 157417472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006673136939663325,
      "loss": 3.2353,
      "theoretical_loss": 4.490028687133432,
      "tokens_seen": 157483008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000667297143937961,
      "loss": 3.1541,
      "theoretical_loss": 4.489786256285276,
      "tokens_seen": 157548544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006672805939095896,
      "loss": 3.1465,
      "theoretical_loss": 4.489543954484084,
      "tokens_seen": 157614080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006672640438812181,
      "loss": 3.27,
      "theoretical_loss": 4.489301781607551,
      "tokens_seen": 157679616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006672474938528466,
      "loss": 3.0735,
      "theoretical_loss": 4.489059737533534,
      "tokens_seen": 157745152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006672309438244752,
      "loss": 3.4063,
      "theoretical_loss": 4.48881782214006,
      "tokens_seen": 157810688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006672143937961036,
      "loss": 3.5199,
      "theoretical_loss": 4.48857603530532,
      "tokens_seen": 157876224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006671978437677322,
      "loss": 3.2295,
      "theoretical_loss": 4.488334376907673,
      "tokens_seen": 157941760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006671812937393607,
      "loss": 3.2308,
      "theoretical_loss": 4.4880928468256425,
      "tokens_seen": 158007296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006671647437109892,
      "loss": 3.4689,
      "theoretical_loss": 4.487851444937916,
      "tokens_seen": 158072832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006671481936826178,
      "loss": 2.9126,
      "theoretical_loss": 4.487610171123347,
      "tokens_seen": 158138368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006671316436542463,
      "loss": 3.3481,
      "theoretical_loss": 4.487369025260954,
      "tokens_seen": 158203904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006671150936258748,
      "loss": 3.3663,
      "theoretical_loss": 4.48712800722992,
      "tokens_seen": 158269440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006670985435975034,
      "loss": 3.2244,
      "theoretical_loss": 4.48688711690959,
      "tokens_seen": 158334976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006670819935691318,
      "loss": 2.9837,
      "theoretical_loss": 4.486646354179475,
      "tokens_seen": 158400512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006670654435407604,
      "loss": 3.1302,
      "theoretical_loss": 4.48640571891925,
      "tokens_seen": 158466048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006670488935123888,
      "loss": 3.2701,
      "theoretical_loss": 4.48616521100875,
      "tokens_seen": 158531584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006670323434840174,
      "loss": 3.0534,
      "theoretical_loss": 4.485924830327974,
      "tokens_seen": 158597120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006670157934556459,
      "loss": 3.0284,
      "theoretical_loss": 4.485684576757087,
      "tokens_seen": 158662656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006669992434272744,
      "loss": 3.2884,
      "theoretical_loss": 4.485444450176413,
      "tokens_seen": 158728192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000666982693398903,
      "loss": 3.2871,
      "theoretical_loss": 4.485204450466437,
      "tokens_seen": 158793728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006669661433705314,
      "loss": 3.3242,
      "theoretical_loss": 4.484964577507808,
      "tokens_seen": 158859264
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00066694959334216,
      "loss": 3.0597,
      "theoretical_loss": 4.484724831181337,
      "tokens_seen": 158924800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006669330433137885,
      "loss": 3.2998,
      "theoretical_loss": 4.4844852113679945,
      "tokens_seen": 158990336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000666916493285417,
      "loss": 3.0637,
      "theoretical_loss": 4.484245717948913,
      "tokens_seen": 159055872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006668999432570456,
      "loss": 3.2732,
      "theoretical_loss": 4.484006350805385,
      "tokens_seen": 159121408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006668833932286741,
      "loss": 3.2501,
      "theoretical_loss": 4.483767109818862,
      "tokens_seen": 159186944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006668668432003026,
      "loss": 3.281,
      "theoretical_loss": 4.483527994870958,
      "tokens_seen": 159252480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006668502931719311,
      "loss": 3.2702,
      "theoretical_loss": 4.483289005843445,
      "tokens_seen": 159318016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006668337431435596,
      "loss": 3.2178,
      "theoretical_loss": 4.483050142618255,
      "tokens_seen": 159383552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006668171931151882,
      "loss": 3.3121,
      "theoretical_loss": 4.482811405077482,
      "tokens_seen": 159449088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006668006430868167,
      "loss": 3.2025,
      "theoretical_loss": 4.482572793103373,
      "tokens_seen": 159514624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006667840930584452,
      "loss": 3.2929,
      "theoretical_loss": 4.482334306578339,
      "tokens_seen": 159580160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006667675430300738,
      "loss": 3.2835,
      "theoretical_loss": 4.482095945384946,
      "tokens_seen": 159645696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006667509930017023,
      "loss": 3.1261,
      "theoretical_loss": 4.481857709405919,
      "tokens_seen": 159711232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006667344429733307,
      "loss": 3.2226,
      "theoretical_loss": 4.4816195985241425,
      "tokens_seen": 159776768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006667178929449593,
      "loss": 3.1585,
      "theoretical_loss": 4.481381612622657,
      "tokens_seen": 159842304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006667013429165878,
      "loss": 3.1325,
      "theoretical_loss": 4.481143751584659,
      "tokens_seen": 159907840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006666847928882164,
      "loss": 3.1596,
      "theoretical_loss": 4.480906015293505,
      "tokens_seen": 159973376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006666682428598449,
      "loss": 2.8774,
      "theoretical_loss": 4.480668403632706,
      "tokens_seen": 160038912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006666516928314734,
      "loss": 3.3158,
      "theoretical_loss": 4.480430916485929,
      "tokens_seen": 160104448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000666635142803102,
      "loss": 3.0443,
      "theoretical_loss": 4.480193553736999,
      "tokens_seen": 160169984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006666185927747305,
      "loss": 3.1391,
      "theoretical_loss": 4.479956315269897,
      "tokens_seen": 160235520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006666020427463589,
      "loss": 3.4163,
      "theoretical_loss": 4.479719200968757,
      "tokens_seen": 160301056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006665854927179875,
      "loss": 3.2214,
      "theoretical_loss": 4.479482210717871,
      "tokens_seen": 160366592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000666568942689616,
      "loss": 3.1135,
      "theoretical_loss": 4.479245344401685,
      "tokens_seen": 160432128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006665523926612446,
      "loss": 3.1908,
      "theoretical_loss": 4.479008601904798,
      "tokens_seen": 160497664
    },
    {
      "epoch": 0.06,
      "objective/train/avg_token_score": 0.018276888877153397,
      "objective/train/avg_weight": 0.9784814119338989,
      "objective/train/docs_used": 101862,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.5705292224884033,
      "objective/train/std_weight": 0.09657934308052063,
      "objective/train/theoretical_loss": 4.478771983111967,
      "objective/train/tokens_used": 181023200,
      "theoretical_loss": 4.478771983111967,
      "tokens_seen": 160563200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006665358426328731,
      "loss": 3.3993,
      "theoretical_loss": 4.478771983111967,
      "tokens_seen": 160563200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006665192926045016,
      "loss": 2.9832,
      "theoretical_loss": 4.478535487908101,
      "tokens_seen": 160628736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006665027425761302,
      "loss": 3.3581,
      "theoretical_loss": 4.478299116178265,
      "tokens_seen": 160694272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006664861925477586,
      "loss": 3.2449,
      "theoretical_loss": 4.478062867807674,
      "tokens_seen": 160759808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006664696425193871,
      "loss": 3.1495,
      "theoretical_loss": 4.4778267426817,
      "tokens_seen": 160825344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006664530924910157,
      "loss": 3.1841,
      "theoretical_loss": 4.477590740685867,
      "tokens_seen": 160890880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006664365424626442,
      "loss": 3.2555,
      "theoretical_loss": 4.47735486170585,
      "tokens_seen": 160956416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006664199924342728,
      "loss": 3.1557,
      "theoretical_loss": 4.47711910562748,
      "tokens_seen": 161021952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006664034424059013,
      "loss": 3.2915,
      "theoretical_loss": 4.4768834723367394,
      "tokens_seen": 161087488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006663868923775298,
      "loss": 3.1535,
      "theoretical_loss": 4.4766479617197605,
      "tokens_seen": 161153024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006663703423491584,
      "loss": 2.9657,
      "theoretical_loss": 4.476412573662829,
      "tokens_seen": 161218560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006663537923207868,
      "loss": 3.4011,
      "theoretical_loss": 4.4761773080523835,
      "tokens_seen": 161284096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006663372422924153,
      "loss": 3.1689,
      "theoretical_loss": 4.475942164775013,
      "tokens_seen": 161349632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006663206922640439,
      "loss": 3.3353,
      "theoretical_loss": 4.475707143717455,
      "tokens_seen": 161415168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006663041422356724,
      "loss": 3.1996,
      "theoretical_loss": 4.475472244766601,
      "tokens_seen": 161480704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000666287592207301,
      "loss": 3.0965,
      "theoretical_loss": 4.475237467809492,
      "tokens_seen": 161546240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006662710421789295,
      "loss": 3.2059,
      "theoretical_loss": 4.47500281273332,
      "tokens_seen": 161611776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000666254492150558,
      "loss": 3.0405,
      "theoretical_loss": 4.474768279425424,
      "tokens_seen": 161677312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006662379421221866,
      "loss": 3.084,
      "theoretical_loss": 4.474533867773299,
      "tokens_seen": 161742848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000666221392093815,
      "loss": 3.2622,
      "theoretical_loss": 4.474299577664581,
      "tokens_seen": 161808384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006662048420654435,
      "loss": 3.222,
      "theoretical_loss": 4.474065408987063,
      "tokens_seen": 161873920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006661882920370721,
      "loss": 3.0791,
      "theoretical_loss": 4.473831361628682,
      "tokens_seen": 161939456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006661717420087006,
      "loss": 3.4657,
      "theoretical_loss": 4.473597435477526,
      "tokens_seen": 162004992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006661551919803292,
      "loss": 3.3437,
      "theoretical_loss": 4.473363630421831,
      "tokens_seen": 162070528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006661386419519577,
      "loss": 3.0019,
      "theoretical_loss": 4.473129946349982,
      "tokens_seen": 162136064
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006661220919235862,
      "loss": 3.233,
      "theoretical_loss": 4.472896383150508,
      "tokens_seen": 162201600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006661055418952147,
      "loss": 3.0797,
      "theoretical_loss": 4.472662940712091,
      "tokens_seen": 162267136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006660889918668432,
      "loss": 3.2555,
      "theoretical_loss": 4.472429618923558,
      "tokens_seen": 162332672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006660724418384717,
      "loss": 3.3301,
      "theoretical_loss": 4.472196417673883,
      "tokens_seen": 162398208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006660558918101002,
      "loss": 3.0906,
      "theoretical_loss": 4.471963336852187,
      "tokens_seen": 162463744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006660393417817287,
      "loss": 3.3497,
      "theoretical_loss": 4.471730376347738,
      "tokens_seen": 162529280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006660227917533573,
      "loss": 3.1168,
      "theoretical_loss": 4.4714975360499505,
      "tokens_seen": 162594816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006660062417249858,
      "loss": 3.2121,
      "theoretical_loss": 4.471264815848384,
      "tokens_seen": 162660352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006659896916966143,
      "loss": 3.1819,
      "theoretical_loss": 4.471032215632746,
      "tokens_seen": 162725888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006659731416682428,
      "loss": 3.166,
      "theoretical_loss": 4.470799735292889,
      "tokens_seen": 162791424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006659565916398713,
      "loss": 3.1882,
      "theoretical_loss": 4.470567374718808,
      "tokens_seen": 162856960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006659400416114999,
      "loss": 3.1614,
      "theoretical_loss": 4.470335133800649,
      "tokens_seen": 162922496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006659234915831284,
      "loss": 3.0884,
      "theoretical_loss": 4.470103012428696,
      "tokens_seen": 162988032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006659069415547569,
      "loss": 3.4059,
      "theoretical_loss": 4.469871010493383,
      "tokens_seen": 163053568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006658903915263855,
      "loss": 3.4092,
      "theoretical_loss": 4.469639127885287,
      "tokens_seen": 163119104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006658738414980139,
      "loss": 3.0451,
      "theoretical_loss": 4.4694073644951295,
      "tokens_seen": 163184640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006658572914696425,
      "loss": 3.2997,
      "theoretical_loss": 4.469175720213771,
      "tokens_seen": 163250176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000665840741441271,
      "loss": 3.1725,
      "theoretical_loss": 4.468944194932225,
      "tokens_seen": 163315712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006658241914128995,
      "loss": 3.1188,
      "theoretical_loss": 4.468712788541639,
      "tokens_seen": 163381248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006658076413845281,
      "loss": 3.1186,
      "theoretical_loss": 4.46848150093331,
      "tokens_seen": 163446784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006657910913561566,
      "loss": 3.3982,
      "theoretical_loss": 4.468250331998676,
      "tokens_seen": 163512320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006657745413277851,
      "loss": 3.3819,
      "theoretical_loss": 4.468019281629316,
      "tokens_seen": 163577856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006657579912994137,
      "loss": 3.1185,
      "theoretical_loss": 4.467788349716955,
      "tokens_seen": 163643392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006657414412710421,
      "loss": 2.9936,
      "theoretical_loss": 4.467557536153457,
      "tokens_seen": 163708928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006657248912426707,
      "loss": 2.9604,
      "theoretical_loss": 4.467326840830829,
      "tokens_seen": 163774464
    },
    {
      "epoch": 0.06,
      "objective/train/avg_token_score": 0.027871694415807724,
      "objective/train/avg_weight": 0.9672291278839111,
      "objective/train/docs_used": 104000,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.282362937927246,
      "objective/train/std_weight": 0.09280717372894287,
      "objective/train/theoretical_loss": 4.467096263641219,
      "objective/train/tokens_used": 184300000,
      "theoretical_loss": 4.467096263641219,
      "tokens_seen": 163840000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006657083412142992,
      "loss": 3.1671,
      "theoretical_loss": 4.467096263641219,
      "tokens_seen": 163840000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006656917911859277,
      "loss": 3.3193,
      "theoretical_loss": 4.466865804476919,
      "tokens_seen": 163905536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006656752411575563,
      "loss": 3.1862,
      "theoretical_loss": 4.466635463230359,
      "tokens_seen": 163971072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006656586911291848,
      "loss": 3.3512,
      "theoretical_loss": 4.466405239794113,
      "tokens_seen": 164036608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006656421411008133,
      "loss": 3.0521,
      "theoretical_loss": 4.466175134060894,
      "tokens_seen": 164102144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006656255910724418,
      "loss": 3.348,
      "theoretical_loss": 4.465945145923554,
      "tokens_seen": 164167680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006656090410440703,
      "loss": 2.7042,
      "theoretical_loss": 4.4657152752750875,
      "tokens_seen": 164233216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006655924910156989,
      "loss": 3.2474,
      "theoretical_loss": 4.465485522008629,
      "tokens_seen": 164298752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006655759409873274,
      "loss": 3.3797,
      "theoretical_loss": 4.465255886017452,
      "tokens_seen": 164364288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006655593909589559,
      "loss": 3.5352,
      "theoretical_loss": 4.465026367194971,
      "tokens_seen": 164429824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006655428409305845,
      "loss": 3.3241,
      "theoretical_loss": 4.464796965434738,
      "tokens_seen": 164495360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000665526290902213,
      "loss": 3.3737,
      "theoretical_loss": 4.464567680630443,
      "tokens_seen": 164560896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006655097408738414,
      "loss": 3.4016,
      "theoretical_loss": 4.464338512675919,
      "tokens_seen": 164626432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00066549319084547,
      "loss": 3.1707,
      "theoretical_loss": 4.464109461465133,
      "tokens_seen": 164691968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006654766408170985,
      "loss": 3.3989,
      "theoretical_loss": 4.4638805268921935,
      "tokens_seen": 164757504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006654600907887271,
      "loss": 3.3173,
      "theoretical_loss": 4.463651708851346,
      "tokens_seen": 164823040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006654435407603556,
      "loss": 3.2126,
      "theoretical_loss": 4.463423007236974,
      "tokens_seen": 164888576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006654269907319841,
      "loss": 3.2484,
      "theoretical_loss": 4.4631944219436,
      "tokens_seen": 164954112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006654104407036127,
      "loss": 3.1014,
      "theoretical_loss": 4.462965952865879,
      "tokens_seen": 165019648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006653938906752412,
      "loss": 3.1969,
      "theoretical_loss": 4.46273759989861,
      "tokens_seen": 165085184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006653773406468697,
      "loss": 2.9568,
      "theoretical_loss": 4.462509362936723,
      "tokens_seen": 165150720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006653607906184982,
      "loss": 3.3019,
      "theoretical_loss": 4.46228124187529,
      "tokens_seen": 165216256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006653442405901267,
      "loss": 3.2152,
      "theoretical_loss": 4.462053236609516,
      "tokens_seen": 165281792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006653276905617553,
      "loss": 3.1752,
      "theoretical_loss": 4.461825347034742,
      "tokens_seen": 165347328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006653111405333838,
      "loss": 3.1498,
      "theoretical_loss": 4.461597573046449,
      "tokens_seen": 165412864
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006652945905050123,
      "loss": 3.2566,
      "theoretical_loss": 4.461369914540247,
      "tokens_seen": 165478400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006652780404766409,
      "loss": 3.1834,
      "theoretical_loss": 4.4611423714118885,
      "tokens_seen": 165543936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006652614904482693,
      "loss": 3.2147,
      "theoretical_loss": 4.460914943557256,
      "tokens_seen": 165609472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006652449404198979,
      "loss": 3.2871,
      "theoretical_loss": 4.460687630872371,
      "tokens_seen": 165675008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006652283903915264,
      "loss": 3.3004,
      "theoretical_loss": 4.46046043325339,
      "tokens_seen": 165740544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006652118403631549,
      "loss": 2.9414,
      "theoretical_loss": 4.460233350596599,
      "tokens_seen": 165806080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006651952903347835,
      "loss": 3.441,
      "theoretical_loss": 4.460006382798425,
      "tokens_seen": 165871616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000665178740306412,
      "loss": 3.403,
      "theoretical_loss": 4.459779529755423,
      "tokens_seen": 165937152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006651621902780405,
      "loss": 3.1059,
      "theoretical_loss": 4.459552791364288,
      "tokens_seen": 166002688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000665145640249669,
      "loss": 3.2416,
      "theoretical_loss": 4.459326167521844,
      "tokens_seen": 166068224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006651290902212975,
      "loss": 2.9871,
      "theoretical_loss": 4.4590996581250515,
      "tokens_seen": 166133760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006651125401929261,
      "loss": 3.1613,
      "theoretical_loss": 4.458873263071002,
      "tokens_seen": 166199296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006650959901645545,
      "loss": 3.2153,
      "theoretical_loss": 4.458646982256921,
      "tokens_seen": 166264832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006650794401361831,
      "loss": 3.4034,
      "theoretical_loss": 4.458420815580169,
      "tokens_seen": 166330368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006650628901078116,
      "loss": 3.4192,
      "theoretical_loss": 4.458194762938234,
      "tokens_seen": 166395904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006650463400794401,
      "loss": 3.257,
      "theoretical_loss": 4.457968824228743,
      "tokens_seen": 166461440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006650297900510687,
      "loss": 3.2947,
      "theoretical_loss": 4.457742999349449,
      "tokens_seen": 166526976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006650132400226971,
      "loss": 3.3087,
      "theoretical_loss": 4.4575172881982414,
      "tokens_seen": 166592512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006649966899943256,
      "loss": 3.3482,
      "theoretical_loss": 4.457291690673139,
      "tokens_seen": 166658048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006649801399659542,
      "loss": 3.3064,
      "theoretical_loss": 4.457066206672291,
      "tokens_seen": 166723584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006649635899375827,
      "loss": 3.2407,
      "theoretical_loss": 4.456840836093983,
      "tokens_seen": 166789120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006649470399092113,
      "loss": 3.3958,
      "theoretical_loss": 4.456615578836625,
      "tokens_seen": 166854656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006649304898808398,
      "loss": 3.1842,
      "theoretical_loss": 4.456390434798762,
      "tokens_seen": 166920192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006649139398524683,
      "loss": 3.1279,
      "theoretical_loss": 4.45616540387907,
      "tokens_seen": 166985728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006648973898240968,
      "loss": 3.0779,
      "theoretical_loss": 4.4559404859763525,
      "tokens_seen": 167051264
    },
    {
      "epoch": 0.06,
      "objective/train/avg_token_score": 0.020470954477787018,
      "objective/train/avg_weight": 0.975911557674408,
      "objective/train/docs_used": 106106,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3556699752807617,
      "objective/train/std_weight": 0.10160351544618607,
      "objective/train/theoretical_loss": 4.455715680989545,
      "objective/train/tokens_used": 187576800,
      "theoretical_loss": 4.455715680989545,
      "tokens_seen": 167116800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006648808397957253,
      "loss": 3.08,
      "theoretical_loss": 4.455715680989545,
      "tokens_seen": 167116800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006648642897673538,
      "loss": 3.3836,
      "theoretical_loss": 4.455490988817713,
      "tokens_seen": 167182336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006648477397389824,
      "loss": 3.0199,
      "theoretical_loss": 4.4552664093600525,
      "tokens_seen": 167247872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006648311897106109,
      "loss": 3.0426,
      "theoretical_loss": 4.455041942515887,
      "tokens_seen": 167313408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006648146396822395,
      "loss": 3.1294,
      "theoretical_loss": 4.454817588184669,
      "tokens_seen": 167378944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000664798089653868,
      "loss": 3.0983,
      "theoretical_loss": 4.454593346265984,
      "tokens_seen": 167444480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006647815396254964,
      "loss": 3.2074,
      "theoretical_loss": 4.454369216659542,
      "tokens_seen": 167510016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000664764989597125,
      "loss": 3.3887,
      "theoretical_loss": 4.454145199265183,
      "tokens_seen": 167575552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006647484395687535,
      "loss": 3.293,
      "theoretical_loss": 4.453921293982877,
      "tokens_seen": 167641088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000664731889540382,
      "loss": 3.4525,
      "theoretical_loss": 4.453697500712722,
      "tokens_seen": 167706624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006647153395120106,
      "loss": 3.0582,
      "theoretical_loss": 4.453473819354942,
      "tokens_seen": 167772160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006646987894836391,
      "loss": 3.2045,
      "theoretical_loss": 4.453250249809889,
      "tokens_seen": 167837696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006646822394552677,
      "loss": 3.1561,
      "theoretical_loss": 4.453026791978045,
      "tokens_seen": 167903232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006646656894268962,
      "loss": 3.1603,
      "theoretical_loss": 4.4528034457600185,
      "tokens_seen": 167968768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006646491393985246,
      "loss": 3.2589,
      "theoretical_loss": 4.452580211056542,
      "tokens_seen": 168034304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006646325893701532,
      "loss": 3.2651,
      "theoretical_loss": 4.452357087768481,
      "tokens_seen": 168099840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006646160393417817,
      "loss": 3.0346,
      "theoretical_loss": 4.45213407579682,
      "tokens_seen": 168165376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006645994893134102,
      "loss": 3.1331,
      "theoretical_loss": 4.451911175042679,
      "tokens_seen": 168230912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006645829392850388,
      "loss": 3.0936,
      "theoretical_loss": 4.451688385407296,
      "tokens_seen": 168296448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006645663892566673,
      "loss": 3.1706,
      "theoretical_loss": 4.451465706792041,
      "tokens_seen": 168361984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006645498392282959,
      "loss": 3.2893,
      "theoretical_loss": 4.4512431390984055,
      "tokens_seen": 168427520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006645332891999243,
      "loss": 3.3689,
      "theoretical_loss": 4.451020682228011,
      "tokens_seen": 168493056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006645167391715528,
      "loss": 2.9804,
      "theoretical_loss": 4.450798336082601,
      "tokens_seen": 168558592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006645001891431814,
      "loss": 3.1325,
      "theoretical_loss": 4.450576100564046,
      "tokens_seen": 168624128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006644836391148099,
      "loss": 3.2502,
      "theoretical_loss": 4.450353975574341,
      "tokens_seen": 168689664
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006644670890864384,
      "loss": 3.0385,
      "theoretical_loss": 4.450131961015606,
      "tokens_seen": 168755200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000664450539058067,
      "loss": 3.3478,
      "theoretical_loss": 4.449910056790086,
      "tokens_seen": 168820736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006644339890296955,
      "loss": 3.1543,
      "theoretical_loss": 4.44968826280015,
      "tokens_seen": 168886272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000664417439001324,
      "loss": 3.2174,
      "theoretical_loss": 4.4494665789482895,
      "tokens_seen": 168951808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006644008889729525,
      "loss": 3.0368,
      "theoretical_loss": 4.449245005137125,
      "tokens_seen": 169017344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000664384338944581,
      "loss": 3.1134,
      "theoretical_loss": 4.449023541269395,
      "tokens_seen": 169082880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006643677889162096,
      "loss": 3.1271,
      "theoretical_loss": 4.448802187247966,
      "tokens_seen": 169148416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006643512388878381,
      "loss": 3.2603,
      "theoretical_loss": 4.448580942975825,
      "tokens_seen": 169213952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006643346888594666,
      "loss": 3.1004,
      "theoretical_loss": 4.448359808356084,
      "tokens_seen": 169279488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006643181388310952,
      "loss": 2.9647,
      "theoretical_loss": 4.448138783291979,
      "tokens_seen": 169345024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006643015888027237,
      "loss": 3.2013,
      "theoretical_loss": 4.447917867686863,
      "tokens_seen": 169410560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006642850387743522,
      "loss": 3.4839,
      "theoretical_loss": 4.44769706144422,
      "tokens_seen": 169476096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006642684887459807,
      "loss": 2.9615,
      "theoretical_loss": 4.44747636446765,
      "tokens_seen": 169541632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006642519387176092,
      "loss": 3.3597,
      "theoretical_loss": 4.447255776660878,
      "tokens_seen": 169607168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006642353886892378,
      "loss": 3.1862,
      "theoretical_loss": 4.44703529792775,
      "tokens_seen": 169672704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006642188386608663,
      "loss": 3.1348,
      "theoretical_loss": 4.446814928172234,
      "tokens_seen": 169738240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006642022886324948,
      "loss": 2.9983,
      "theoretical_loss": 4.446594667298421,
      "tokens_seen": 169803776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006641857386041234,
      "loss": 3.3036,
      "theoretical_loss": 4.446374515210521,
      "tokens_seen": 169869312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006641691885757518,
      "loss": 3.1833,
      "theoretical_loss": 4.446154471812866,
      "tokens_seen": 169934848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006641526385473804,
      "loss": 3.1467,
      "theoretical_loss": 4.445934537009911,
      "tokens_seen": 170000384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006641360885190089,
      "loss": 2.9322,
      "theoretical_loss": 4.445714710706228,
      "tokens_seen": 170065920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006641195384906374,
      "loss": 3.0934,
      "theoretical_loss": 4.445494992806513,
      "tokens_seen": 170131456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006641029884622659,
      "loss": 3.1525,
      "theoretical_loss": 4.44527538321558,
      "tokens_seen": 170196992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006640864384338944,
      "loss": 3.0492,
      "theoretical_loss": 4.445055881838365,
      "tokens_seen": 170262528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000664069888405523,
      "loss": 3.1472,
      "theoretical_loss": 4.444836488579924,
      "tokens_seen": 170328064
    },
    {
      "epoch": 0.06,
      "objective/train/avg_token_score": 0.02504768595099449,
      "objective/train/avg_weight": 0.9705386161804199,
      "objective/train/docs_used": 107805,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3389647006988525,
      "objective/train/std_weight": 0.13012516498565674,
      "objective/train/theoretical_loss": 4.44461720334543,
      "objective/train/tokens_used": 190853600,
      "theoretical_loss": 4.44461720334543,
      "tokens_seen": 170393600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006640533383771514,
      "loss": 3.2825,
      "theoretical_loss": 4.44461720334543,
      "tokens_seen": 170393600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00066403678834878,
      "loss": 3.1707,
      "theoretical_loss": 4.444398026040179,
      "tokens_seen": 170459136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006640202383204085,
      "loss": 3.3802,
      "theoretical_loss": 4.444178956569585,
      "tokens_seen": 170524672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000664003688292037,
      "loss": 3.1514,
      "theoretical_loss": 4.443959994839181,
      "tokens_seen": 170590208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006639871382636656,
      "loss": 3.2603,
      "theoretical_loss": 4.44374114075462,
      "tokens_seen": 170655744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006639705882352941,
      "loss": 3.228,
      "theoretical_loss": 4.443522394221671,
      "tokens_seen": 170721280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006639540382069226,
      "loss": 2.9141,
      "theoretical_loss": 4.443303755146225,
      "tokens_seen": 170786816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006639374881785512,
      "loss": 3.0717,
      "theoretical_loss": 4.443085223434291,
      "tokens_seen": 170852352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006639209381501796,
      "loss": 3.0437,
      "theoretical_loss": 4.442866798991993,
      "tokens_seen": 170917888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006639043881218082,
      "loss": 3.0325,
      "theoretical_loss": 4.442648481725577,
      "tokens_seen": 170983424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006638878380934367,
      "loss": 3.0316,
      "theoretical_loss": 4.442430271541404,
      "tokens_seen": 171048960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006638712880650652,
      "loss": 3.3467,
      "theoretical_loss": 4.442212168345956,
      "tokens_seen": 171114496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006638547380366938,
      "loss": 3.2496,
      "theoretical_loss": 4.4419941720458285,
      "tokens_seen": 171180032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006638381880083223,
      "loss": 2.9451,
      "theoretical_loss": 4.441776282547736,
      "tokens_seen": 171245568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006638216379799508,
      "loss": 3.2779,
      "theoretical_loss": 4.441558499758511,
      "tokens_seen": 171311104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006638050879515793,
      "loss": 3.1254,
      "theoretical_loss": 4.441340823585101,
      "tokens_seen": 171376640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006637885379232078,
      "loss": 3.0422,
      "theoretical_loss": 4.441123253934572,
      "tokens_seen": 171442176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006637719878948364,
      "loss": 2.8681,
      "theoretical_loss": 4.440905790714105,
      "tokens_seen": 171507712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006637554378664649,
      "loss": 3.1686,
      "theoretical_loss": 4.440688433830999,
      "tokens_seen": 171573248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006637388878380934,
      "loss": 3.0124,
      "theoretical_loss": 4.440471183192667,
      "tokens_seen": 171638784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000663722337809722,
      "loss": 2.8907,
      "theoretical_loss": 4.440254038706639,
      "tokens_seen": 171704320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006637057877813505,
      "loss": 3.0594,
      "theoretical_loss": 4.440037000280561,
      "tokens_seen": 171769856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000663689237752979,
      "loss": 3.1469,
      "theoretical_loss": 4.439820067822195,
      "tokens_seen": 171835392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006636726877246075,
      "loss": 2.8506,
      "theoretical_loss": 4.439603241239416,
      "tokens_seen": 171900928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000663656137696236,
      "loss": 2.947,
      "theoretical_loss": 4.439386520440218,
      "tokens_seen": 171966464
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006636395876678646,
      "loss": 3.1906,
      "theoretical_loss": 4.439169905332706,
      "tokens_seen": 172032000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006636230376394931,
      "loss": 3.0894,
      "theoretical_loss": 4.438953395825102,
      "tokens_seen": 172097536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006636064876111216,
      "loss": 3.1167,
      "theoretical_loss": 4.438736991825744,
      "tokens_seen": 172163072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006635899375827502,
      "loss": 3.4215,
      "theoretical_loss": 4.438520693243079,
      "tokens_seen": 172228608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006635733875543787,
      "loss": 3.1676,
      "theoretical_loss": 4.4383044999856756,
      "tokens_seen": 172294144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006635568375260071,
      "loss": 3.0309,
      "theoretical_loss": 4.438088411962211,
      "tokens_seen": 172359680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006635402874976357,
      "loss": 3.2625,
      "theoretical_loss": 4.437872429081477,
      "tokens_seen": 172425216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006635237374692642,
      "loss": 3.1831,
      "theoretical_loss": 4.437656551252381,
      "tokens_seen": 172490752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006635071874408928,
      "loss": 3.049,
      "theoretical_loss": 4.4374407783839445,
      "tokens_seen": 172556288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006634906374125213,
      "loss": 2.8811,
      "theoretical_loss": 4.437225110385297,
      "tokens_seen": 172621824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006634740873841498,
      "loss": 3.1852,
      "theoretical_loss": 4.4370095471656885,
      "tokens_seen": 172687360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006634575373557784,
      "loss": 3.081,
      "theoretical_loss": 4.436794088634477,
      "tokens_seen": 172752896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006634409873274068,
      "loss": 3.1503,
      "theoretical_loss": 4.4365787347011345,
      "tokens_seen": 172818432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006634244372990353,
      "loss": 3.3142,
      "theoretical_loss": 4.436363485275246,
      "tokens_seen": 172883968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006634078872706639,
      "loss": 3.3477,
      "theoretical_loss": 4.436148340266508,
      "tokens_seen": 172949504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006633913372422924,
      "loss": 2.9501,
      "theoretical_loss": 4.435933299584729,
      "tokens_seen": 173015040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000663374787213921,
      "loss": 3.2128,
      "theoretical_loss": 4.4357183631398325,
      "tokens_seen": 173080576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006633582371855495,
      "loss": 3.0143,
      "theoretical_loss": 4.435503530841849,
      "tokens_seen": 173146112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000663341687157178,
      "loss": 3.1879,
      "theoretical_loss": 4.435288802600926,
      "tokens_seen": 173211648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006633251371288066,
      "loss": 3.0329,
      "theoretical_loss": 4.4350741783273175,
      "tokens_seen": 173277184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000663308587100435,
      "loss": 3.0737,
      "theoretical_loss": 4.434859657931392,
      "tokens_seen": 173342720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006632920370720635,
      "loss": 2.9814,
      "theoretical_loss": 4.434645241323629,
      "tokens_seen": 173408256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006632754870436921,
      "loss": 3.1096,
      "theoretical_loss": 4.434430928414617,
      "tokens_seen": 173473792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006632589370153206,
      "loss": 3.2479,
      "theoretical_loss": 4.434216719115057,
      "tokens_seen": 173539328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006632423869869492,
      "loss": 3.1014,
      "theoretical_loss": 4.43400261333576,
      "tokens_seen": 173604864
    },
    {
      "epoch": 0.06,
      "objective/train/avg_token_score": 0.013684557750821114,
      "objective/train/avg_weight": 0.9838986396789551,
      "objective/train/docs_used": 109422,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.967532157897949,
      "objective/train/std_weight": 0.07505863904953003,
      "objective/train/theoretical_loss": 4.433788610987646,
      "objective/train/tokens_used": 194130400,
      "theoretical_loss": 4.433788610987646,
      "tokens_seen": 173670400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006632258369585777,
      "loss": 3.133,
      "theoretical_loss": 4.433788610987646,
      "tokens_seen": 173670400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006632092869302062,
      "loss": 3.0836,
      "theoretical_loss": 4.433574711981749,
      "tokens_seen": 173735936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006631927369018347,
      "loss": 3.1011,
      "theoretical_loss": 4.433360916229209,
      "tokens_seen": 173801472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006631761868734632,
      "loss": 2.9726,
      "theoretical_loss": 4.433147223641278,
      "tokens_seen": 173867008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006631596368450917,
      "loss": 3.109,
      "theoretical_loss": 4.432933634129318,
      "tokens_seen": 173932544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006631430868167203,
      "loss": 2.9435,
      "theoretical_loss": 4.4327201476047975,
      "tokens_seen": 173998080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006631265367883487,
      "loss": 3.4222,
      "theoretical_loss": 4.432506763979299,
      "tokens_seen": 174063616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006631099867599773,
      "loss": 3.2546,
      "theoretical_loss": 4.432293483164512,
      "tokens_seen": 174129152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006630934367316058,
      "loss": 3.0363,
      "theoretical_loss": 4.432080305072233,
      "tokens_seen": 174194688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006630768867032343,
      "loss": 3.0761,
      "theoretical_loss": 4.43186722961437,
      "tokens_seen": 174260224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006630603366748628,
      "loss": 3.1559,
      "theoretical_loss": 4.431654256702938,
      "tokens_seen": 174325760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006630437866464913,
      "loss": 3.0947,
      "theoretical_loss": 4.431441386250063,
      "tokens_seen": 174391296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006630272366181199,
      "loss": 2.8949,
      "theoretical_loss": 4.4312286181679745,
      "tokens_seen": 174456832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006630106865897484,
      "loss": 3.1656,
      "theoretical_loss": 4.431015952369016,
      "tokens_seen": 174522368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000662994136561377,
      "loss": 3.1645,
      "theoretical_loss": 4.430803388765636,
      "tokens_seen": 174587904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006629775865330055,
      "loss": 3.1168,
      "theoretical_loss": 4.430590927270388,
      "tokens_seen": 174653440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000662961036504634,
      "loss": 3.1849,
      "theoretical_loss": 4.430378567795938,
      "tokens_seen": 174718976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006629444864762625,
      "loss": 3.1754,
      "theoretical_loss": 4.430166310255057,
      "tokens_seen": 174784512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000662927936447891,
      "loss": 3.1784,
      "theoretical_loss": 4.429954154560624,
      "tokens_seen": 174850048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006629113864195195,
      "loss": 3.1933,
      "theoretical_loss": 4.429742100625624,
      "tokens_seen": 174915584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006628948363911481,
      "loss": 3.3832,
      "theoretical_loss": 4.429530148363151,
      "tokens_seen": 174981120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006628782863627766,
      "loss": 3.3819,
      "theoretical_loss": 4.429318297686402,
      "tokens_seen": 175046656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006628617363344052,
      "loss": 3.2392,
      "theoretical_loss": 4.429106548508685,
      "tokens_seen": 175112192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006628451863060337,
      "loss": 3.2745,
      "theoretical_loss": 4.428894900743411,
      "tokens_seen": 175177728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006628286362776621,
      "loss": 3.1606,
      "theoretical_loss": 4.428683354304098,
      "tokens_seen": 175243264
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006628120862492907,
      "loss": 2.948,
      "theoretical_loss": 4.428471909104372,
      "tokens_seen": 175308800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006627955362209192,
      "loss": 2.9845,
      "theoretical_loss": 4.428260565057964,
      "tokens_seen": 175374336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006627789861925477,
      "loss": 2.9489,
      "theoretical_loss": 4.428049322078708,
      "tokens_seen": 175439872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006627624361641763,
      "loss": 3.2785,
      "theoretical_loss": 4.427838180080547,
      "tokens_seen": 175505408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006627458861358048,
      "loss": 3.159,
      "theoretical_loss": 4.4276271389775275,
      "tokens_seen": 175570944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006627293361074334,
      "loss": 3.1883,
      "theoretical_loss": 4.427416198683803,
      "tokens_seen": 175636480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006627127860790618,
      "loss": 3.3046,
      "theoretical_loss": 4.427205359113629,
      "tokens_seen": 175702016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006626962360506903,
      "loss": 3.141,
      "theoretical_loss": 4.42699462018137,
      "tokens_seen": 175767552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006626796860223189,
      "loss": 2.9397,
      "theoretical_loss": 4.42678398180149,
      "tokens_seen": 175833088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006626631359939474,
      "loss": 3.1104,
      "theoretical_loss": 4.426573443888563,
      "tokens_seen": 175898624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006626465859655759,
      "loss": 3.079,
      "theoretical_loss": 4.426363006357263,
      "tokens_seen": 175964160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006626300359372045,
      "loss": 3.2815,
      "theoretical_loss": 4.426152669122374,
      "tokens_seen": 176029696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000662613485908833,
      "loss": 3.221,
      "theoretical_loss": 4.425942432098774,
      "tokens_seen": 176095232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006625969358804616,
      "loss": 3.0516,
      "theoretical_loss": 4.425732295201455,
      "tokens_seen": 176160768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00066258038585209,
      "loss": 3.1433,
      "theoretical_loss": 4.425522258345508,
      "tokens_seen": 176226304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006625638358237185,
      "loss": 3.2076,
      "theoretical_loss": 4.425312321446127,
      "tokens_seen": 176291840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006625472857953471,
      "loss": 3.0773,
      "theoretical_loss": 4.425102484418613,
      "tokens_seen": 176357376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006625307357669756,
      "loss": 3.2316,
      "theoretical_loss": 4.424892747178365,
      "tokens_seen": 176422912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006625141857386041,
      "loss": 3.333,
      "theoretical_loss": 4.42468310964089,
      "tokens_seen": 176488448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006624976357102327,
      "loss": 3.3154,
      "theoretical_loss": 4.424473571721794,
      "tokens_seen": 176553984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006624810856818612,
      "loss": 3.5263,
      "theoretical_loss": 4.42426413333679,
      "tokens_seen": 176619520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006624645356534897,
      "loss": 3.2573,
      "theoretical_loss": 4.424054794401689,
      "tokens_seen": 176685056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006624479856251182,
      "loss": 3.2284,
      "theoretical_loss": 4.423845554832406,
      "tokens_seen": 176750592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006624314355967467,
      "loss": 3.2937,
      "theoretical_loss": 4.42363641454496,
      "tokens_seen": 176816128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006624148855683753,
      "loss": 3.1035,
      "theoretical_loss": 4.423427373455471,
      "tokens_seen": 176881664
    },
    {
      "epoch": 0.06,
      "objective/train/avg_token_score": 0.025635868310928345,
      "objective/train/avg_weight": 0.9698352217674255,
      "objective/train/docs_used": 111276,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.405101776123047,
      "objective/train/std_weight": 0.11062265187501907,
      "objective/train/theoretical_loss": 4.42321843148016,
      "objective/train/tokens_used": 197407200,
      "theoretical_loss": 4.42321843148016,
      "tokens_seen": 176947200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006623983355400038,
      "loss": 3.1566,
      "theoretical_loss": 4.42321843148016,
      "tokens_seen": 176947200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006623817855116323,
      "loss": 3.0197,
      "theoretical_loss": 4.423009588535351,
      "tokens_seen": 177012736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006623652354832609,
      "loss": 3.3254,
      "theoretical_loss": 4.422800844537466,
      "tokens_seen": 177078272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006623486854548893,
      "loss": 3.0643,
      "theoretical_loss": 4.422592199403036,
      "tokens_seen": 177143808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006623321354265179,
      "loss": 3.042,
      "theoretical_loss": 4.422383653048685,
      "tokens_seen": 177209344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006623155853981464,
      "loss": 3.3402,
      "theoretical_loss": 4.422175205391145,
      "tokens_seen": 177274880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006622990353697749,
      "loss": 3.1458,
      "theoretical_loss": 4.421966856347243,
      "tokens_seen": 177340416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006622824853414035,
      "loss": 3.2145,
      "theoretical_loss": 4.421758605833912,
      "tokens_seen": 177405952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000662265935313032,
      "loss": 3.327,
      "theoretical_loss": 4.421550453768181,
      "tokens_seen": 177471488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006622493852846605,
      "loss": 3.0393,
      "theoretical_loss": 4.421342400067183,
      "tokens_seen": 177537024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006622328352562891,
      "loss": 2.9938,
      "theoretical_loss": 4.42113444464815,
      "tokens_seen": 177602560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006622162852279175,
      "loss": 2.982,
      "theoretical_loss": 4.420926587428411,
      "tokens_seen": 177668096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006621997351995461,
      "loss": 3.2446,
      "theoretical_loss": 4.420718828325403,
      "tokens_seen": 177733632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006621831851711746,
      "loss": 3.1504,
      "theoretical_loss": 4.420511167256656,
      "tokens_seen": 177799168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006621666351428031,
      "loss": 2.9281,
      "theoretical_loss": 4.4203036041398,
      "tokens_seen": 177864704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006621500851144317,
      "loss": 3.1397,
      "theoretical_loss": 4.420096138892568,
      "tokens_seen": 177930240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006621335350860601,
      "loss": 3.0477,
      "theoretical_loss": 4.419888771432789,
      "tokens_seen": 177995776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006621169850576887,
      "loss": 2.8434,
      "theoretical_loss": 4.419681501678395,
      "tokens_seen": 178061312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006621004350293171,
      "loss": 3.0587,
      "theoretical_loss": 4.419474329547413,
      "tokens_seen": 178126848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006620838850009456,
      "loss": 3.157,
      "theoretical_loss": 4.419267254957971,
      "tokens_seen": 178192384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006620673349725742,
      "loss": 3.2264,
      "theoretical_loss": 4.419060277828295,
      "tokens_seen": 178257920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006620507849442027,
      "loss": 3.2272,
      "theoretical_loss": 4.41885339807671,
      "tokens_seen": 178323456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006620342349158313,
      "loss": 3.294,
      "theoretical_loss": 4.4186466156216415,
      "tokens_seen": 178388992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006620176848874598,
      "loss": 3.0892,
      "theoretical_loss": 4.418439930381609,
      "tokens_seen": 178454528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006620011348590883,
      "loss": 2.9434,
      "theoretical_loss": 4.418233342275233,
      "tokens_seen": 178520064
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006619845848307169,
      "loss": 3.5029,
      "theoretical_loss": 4.418026851221231,
      "tokens_seen": 178585600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006619680348023453,
      "loss": 3.0243,
      "theoretical_loss": 4.4178204571384185,
      "tokens_seen": 178651136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006619514847739738,
      "loss": 3.2403,
      "theoretical_loss": 4.41761415994571,
      "tokens_seen": 178716672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006619349347456024,
      "loss": 3.2399,
      "theoretical_loss": 4.417407959562116,
      "tokens_seen": 178782208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006619183847172309,
      "loss": 3.0115,
      "theoretical_loss": 4.417201855906742,
      "tokens_seen": 178847744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006619018346888595,
      "loss": 3.0629,
      "theoretical_loss": 4.416995848898797,
      "tokens_seen": 178913280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000661885284660488,
      "loss": 3.1427,
      "theoretical_loss": 4.4167899384575815,
      "tokens_seen": 178978816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006618687346321165,
      "loss": 2.9744,
      "theoretical_loss": 4.416584124502495,
      "tokens_seen": 179044352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000661852184603745,
      "loss": 2.6337,
      "theoretical_loss": 4.416378406953033,
      "tokens_seen": 179109888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006618356345753735,
      "loss": 3.2106,
      "theoretical_loss": 4.41617278572879,
      "tokens_seen": 179175424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000661819084547002,
      "loss": 3.226,
      "theoretical_loss": 4.4159672607494524,
      "tokens_seen": 179240960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006618025345186306,
      "loss": 3.0218,
      "theoretical_loss": 4.415761831934808,
      "tokens_seen": 179306496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006617859844902591,
      "loss": 3.1698,
      "theoretical_loss": 4.415556499204737,
      "tokens_seen": 179372032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006617694344618877,
      "loss": 2.9655,
      "theoretical_loss": 4.415351262479216,
      "tokens_seen": 179437568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006617528844335162,
      "loss": 2.9228,
      "theoretical_loss": 4.415146121678321,
      "tokens_seen": 179503104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006617363344051446,
      "loss": 3.0621,
      "theoretical_loss": 4.414941076722219,
      "tokens_seen": 179568640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006617197843767732,
      "loss": 3.0557,
      "theoretical_loss": 4.4147361275311745,
      "tokens_seen": 179634176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006617032343484017,
      "loss": 3.0145,
      "theoretical_loss": 4.414531274025548,
      "tokens_seen": 179699712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006616866843200302,
      "loss": 3.1282,
      "theoretical_loss": 4.414326516125795,
      "tokens_seen": 179765248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006616701342916588,
      "loss": 3.2644,
      "theoretical_loss": 4.414121853752466,
      "tokens_seen": 179830784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006616535842632873,
      "loss": 3.3151,
      "theoretical_loss": 4.413917286826205,
      "tokens_seen": 179896320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006616370342349159,
      "loss": 3.0624,
      "theoretical_loss": 4.413712815267752,
      "tokens_seen": 179961856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006616204842065444,
      "loss": 3.0954,
      "theoretical_loss": 4.413508438997944,
      "tokens_seen": 180027392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006616039341781728,
      "loss": 3.0593,
      "theoretical_loss": 4.4133041579377075,
      "tokens_seen": 180092928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006615873841498014,
      "loss": 3.223,
      "theoretical_loss": 4.413099972008068,
      "tokens_seen": 180158464
    },
    {
      "epoch": 0.06,
      "objective/train/avg_token_score": 0.006056290119886398,
      "objective/train/avg_weight": 0.9928745627403259,
      "objective/train/docs_used": 113383,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.104196071624756,
      "objective/train/std_weight": 0.052710287272930145,
      "objective/train/theoretical_loss": 4.412895881130142,
      "objective/train/tokens_used": 200684000,
      "theoretical_loss": 4.412895881130142,
      "tokens_seen": 180224000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006615708341214299,
      "loss": 3.2041,
      "theoretical_loss": 4.412895881130142,
      "tokens_seen": 180224000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006615542840930584,
      "loss": 3.3057,
      "theoretical_loss": 4.412691885225141,
      "tokens_seen": 180289536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000661537734064687,
      "loss": 3.1489,
      "theoretical_loss": 4.412487984214373,
      "tokens_seen": 180355072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006615211840363155,
      "loss": 3.1658,
      "theoretical_loss": 4.412284178019235,
      "tokens_seen": 180420608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006615046340079441,
      "loss": 3.0601,
      "theoretical_loss": 4.412080466561221,
      "tokens_seen": 180486144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006614880839795725,
      "loss": 3.1505,
      "theoretical_loss": 4.411876849761917,
      "tokens_seen": 180551680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000661471533951201,
      "loss": 3.3309,
      "theoretical_loss": 4.411673327543005,
      "tokens_seen": 180617216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006614549839228296,
      "loss": 3.1318,
      "theoretical_loss": 4.4114698998262565,
      "tokens_seen": 180682752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006614384338944581,
      "loss": 3.2469,
      "theoretical_loss": 4.411266566533539,
      "tokens_seen": 180748288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006614218838660867,
      "loss": 3.2488,
      "theoretical_loss": 4.41106332758681,
      "tokens_seen": 180813824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006614053338377152,
      "loss": 2.9121,
      "theoretical_loss": 4.41086018290812,
      "tokens_seen": 180879360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006613887838093437,
      "loss": 3.1954,
      "theoretical_loss": 4.410657132419617,
      "tokens_seen": 180944896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006613722337809723,
      "loss": 3.2068,
      "theoretical_loss": 4.410454176043537,
      "tokens_seen": 181010432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006613556837526007,
      "loss": 3.0495,
      "theoretical_loss": 4.410251313702208,
      "tokens_seen": 181075968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006613391337242292,
      "loss": 3.1232,
      "theoretical_loss": 4.410048545318052,
      "tokens_seen": 181141504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006613225836958578,
      "loss": 3.1708,
      "theoretical_loss": 4.409845870813582,
      "tokens_seen": 181207040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006613060336674863,
      "loss": 3.0634,
      "theoretical_loss": 4.409643290111404,
      "tokens_seen": 181272576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006612894836391149,
      "loss": 3.006,
      "theoretical_loss": 4.409440803134215,
      "tokens_seen": 181338112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006612729336107434,
      "loss": 3.2263,
      "theoretical_loss": 4.409238409804804,
      "tokens_seen": 181403648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006612563835823719,
      "loss": 3.2891,
      "theoretical_loss": 4.409036110046051,
      "tokens_seen": 181469184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006612398335540004,
      "loss": 2.8954,
      "theoretical_loss": 4.408833903780926,
      "tokens_seen": 181534720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006612232835256289,
      "loss": 3.2015,
      "theoretical_loss": 4.408631790932494,
      "tokens_seen": 181600256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006612067334972574,
      "loss": 3.1792,
      "theoretical_loss": 4.408429771423909,
      "tokens_seen": 181665792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000661190183468886,
      "loss": 2.9139,
      "theoretical_loss": 4.408227845178414,
      "tokens_seen": 181731328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006611736334405145,
      "loss": 3.2003,
      "theoretical_loss": 4.408026012119344,
      "tokens_seen": 181796864
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000661157083412143,
      "loss": 2.9455,
      "theoretical_loss": 4.407824272170128,
      "tokens_seen": 181862400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006611405333837715,
      "loss": 2.9932,
      "theoretical_loss": 4.407622625254279,
      "tokens_seen": 181927936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0006611239833554,
      "loss": 2.8737,
      "theoretical_loss": 4.407421071295406,
      "tokens_seen": 181993472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006611074333270285,
      "loss": 3.382,
      "theoretical_loss": 4.407219610217206,
      "tokens_seen": 182059008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000661090883298657,
      "loss": 3.1095,
      "theoretical_loss": 4.407018241943467,
      "tokens_seen": 182124544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006610743332702856,
      "loss": 3.1675,
      "theoretical_loss": 4.406816966398064,
      "tokens_seen": 182190080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006610577832419141,
      "loss": 2.9678,
      "theoretical_loss": 4.406615783504965,
      "tokens_seen": 182255616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006610412332135426,
      "loss": 3.2725,
      "theoretical_loss": 4.4064146931882275,
      "tokens_seen": 182321152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006610246831851712,
      "loss": 3.0948,
      "theoretical_loss": 4.406213695371996,
      "tokens_seen": 182386688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006610081331567996,
      "loss": 3.3041,
      "theoretical_loss": 4.406012789980506,
      "tokens_seen": 182452224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006609915831284282,
      "loss": 3.1464,
      "theoretical_loss": 4.405811976938084,
      "tokens_seen": 182517760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006609750331000567,
      "loss": 3.0872,
      "theoretical_loss": 4.405611256169143,
      "tokens_seen": 182583296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006609584830716852,
      "loss": 3.1469,
      "theoretical_loss": 4.405410627598185,
      "tokens_seen": 182648832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006609419330433138,
      "loss": 3.08,
      "theoretical_loss": 4.405210091149802,
      "tokens_seen": 182714368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006609253830149423,
      "loss": 3.2134,
      "theoretical_loss": 4.405009646748674,
      "tokens_seen": 182779904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006609088329865708,
      "loss": 3.1686,
      "theoretical_loss": 4.404809294319572,
      "tokens_seen": 182845440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006608922829581994,
      "loss": 3.254,
      "theoretical_loss": 4.40460903378735,
      "tokens_seen": 182910976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006608757329298278,
      "loss": 3.2244,
      "theoretical_loss": 4.404408865076955,
      "tokens_seen": 182976512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006608591829014564,
      "loss": 3.0339,
      "theoretical_loss": 4.404208788113422,
      "tokens_seen": 183042048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006608426328730849,
      "loss": 3.07,
      "theoretical_loss": 4.404008802821871,
      "tokens_seen": 183107584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006608260828447134,
      "loss": 3.1312,
      "theoretical_loss": 4.4038089091275125,
      "tokens_seen": 183173120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000660809532816342,
      "loss": 3.1888,
      "theoretical_loss": 4.403609106955645,
      "tokens_seen": 183238656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006607929827879705,
      "loss": 3.2764,
      "theoretical_loss": 4.403409396231651,
      "tokens_seen": 183304192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000660776432759599,
      "loss": 3.3707,
      "theoretical_loss": 4.403209776881004,
      "tokens_seen": 183369728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006607598827312275,
      "loss": 3.3648,
      "theoretical_loss": 4.403010248829265,
      "tokens_seen": 183435264
    },
    {
      "epoch": 0.07,
      "objective/train/avg_token_score": 0.018270255997776985,
      "objective/train/avg_weight": 0.9784905910491943,
      "objective/train/docs_used": 115260,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0628373622894287,
      "objective/train/std_weight": 0.09968066215515137,
      "objective/train/theoretical_loss": 4.4028108120020795,
      "objective/train/tokens_used": 203960800,
      "theoretical_loss": 4.4028108120020795,
      "tokens_seen": 183500800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000660743332702856,
      "loss": 3.1884,
      "theoretical_loss": 4.4028108120020795,
      "tokens_seen": 183500800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006607267826744846,
      "loss": 3.3144,
      "theoretical_loss": 4.402611466325182,
      "tokens_seen": 183566336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006607102326461131,
      "loss": 3.1307,
      "theoretical_loss": 4.4024122117243945,
      "tokens_seen": 183631872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006606936826177416,
      "loss": 3.3724,
      "theoretical_loss": 4.402213048125624,
      "tokens_seen": 183697408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006606771325893702,
      "loss": 3.0158,
      "theoretical_loss": 4.4020139754548655,
      "tokens_seen": 183762944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006606605825609987,
      "loss": 3.2025,
      "theoretical_loss": 4.401814993638199,
      "tokens_seen": 183828480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006606440325326271,
      "loss": 2.7951,
      "theoretical_loss": 4.4016161026017935,
      "tokens_seen": 183894016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006606274825042557,
      "loss": 3.3498,
      "theoretical_loss": 4.401417302271902,
      "tokens_seen": 183959552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006606109324758842,
      "loss": 3.1649,
      "theoretical_loss": 4.401218592574865,
      "tokens_seen": 184025088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006605943824475128,
      "loss": 3.0798,
      "theoretical_loss": 4.401019973437108,
      "tokens_seen": 184090624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006605778324191413,
      "loss": 3.3058,
      "theoretical_loss": 4.400821444785143,
      "tokens_seen": 184156160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006605612823907698,
      "loss": 3.1927,
      "theoretical_loss": 4.400623006545567,
      "tokens_seen": 184221696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006605447323623984,
      "loss": 2.9802,
      "theoretical_loss": 4.400424658645065,
      "tokens_seen": 184287232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006605281823340269,
      "loss": 3.1677,
      "theoretical_loss": 4.400226401010404,
      "tokens_seen": 184352768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006605116323056553,
      "loss": 3.2805,
      "theoretical_loss": 4.40002823356844,
      "tokens_seen": 184418304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006604950822772839,
      "loss": 2.9902,
      "theoretical_loss": 4.39983015624611,
      "tokens_seen": 184483840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006604785322489124,
      "loss": 3.2976,
      "theoretical_loss": 4.39963216897044,
      "tokens_seen": 184549376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000660461982220541,
      "loss": 3.1335,
      "theoretical_loss": 4.3994342716685395,
      "tokens_seen": 184614912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006604454321921695,
      "loss": 3.0963,
      "theoretical_loss": 4.399236464267602,
      "tokens_seen": 184680448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000660428882163798,
      "loss": 3.3216,
      "theoretical_loss": 4.399038746694908,
      "tokens_seen": 184745984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006604123321354266,
      "loss": 3.1979,
      "theoretical_loss": 4.398841118877819,
      "tokens_seen": 184811520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000660395782107055,
      "loss": 3.0926,
      "theoretical_loss": 4.398643580743785,
      "tokens_seen": 184877056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006603792320786835,
      "loss": 3.2267,
      "theoretical_loss": 4.398446132220338,
      "tokens_seen": 184942592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006603626820503121,
      "loss": 3.1268,
      "theoretical_loss": 4.3982487732350934,
      "tokens_seen": 185008128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006603461320219406,
      "loss": 3.1746,
      "theoretical_loss": 4.398051503715753,
      "tokens_seen": 185073664
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006603295819935692,
      "loss": 3.19,
      "theoretical_loss": 4.397854323590102,
      "tokens_seen": 185139200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006603130319651977,
      "loss": 3.2632,
      "theoretical_loss": 4.397657232786008,
      "tokens_seen": 185204736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006602964819368262,
      "loss": 3.2011,
      "theoretical_loss": 4.397460231231424,
      "tokens_seen": 185270272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006602799319084548,
      "loss": 3.0008,
      "theoretical_loss": 4.397263318854384,
      "tokens_seen": 185335808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006602633818800832,
      "loss": 2.9914,
      "theoretical_loss": 4.39706649558301,
      "tokens_seen": 185401344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006602468318517117,
      "loss": 3.252,
      "theoretical_loss": 4.396869761345503,
      "tokens_seen": 185466880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006602302818233403,
      "loss": 3.1581,
      "theoretical_loss": 4.396673116070147,
      "tokens_seen": 185532416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006602137317949688,
      "loss": 3.2836,
      "theoretical_loss": 4.396476559685315,
      "tokens_seen": 185597952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006601971817665974,
      "loss": 3.1798,
      "theoretical_loss": 4.396280092119455,
      "tokens_seen": 185663488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006601806317382258,
      "loss": 3.0478,
      "theoretical_loss": 4.3960837133011035,
      "tokens_seen": 185729024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006601640817098544,
      "loss": 2.9282,
      "theoretical_loss": 4.395887423158877,
      "tokens_seen": 185794560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006601475316814828,
      "loss": 3.1809,
      "theoretical_loss": 4.395691221621476,
      "tokens_seen": 185860096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006601309816531113,
      "loss": 2.9871,
      "theoretical_loss": 4.395495108617682,
      "tokens_seen": 185925632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006601144316247399,
      "loss": 2.944,
      "theoretical_loss": 4.39529908407636,
      "tokens_seen": 185991168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006600978815963684,
      "loss": 3.2264,
      "theoretical_loss": 4.3951031479264575,
      "tokens_seen": 186056704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000660081331567997,
      "loss": 3.0406,
      "theoretical_loss": 4.394907300097002,
      "tokens_seen": 186122240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006600647815396255,
      "loss": 2.9709,
      "theoretical_loss": 4.394711540517106,
      "tokens_seen": 186187776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000660048231511254,
      "loss": 3.1321,
      "theoretical_loss": 4.39451586911596,
      "tokens_seen": 186253312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006600316814828825,
      "loss": 3.2023,
      "theoretical_loss": 4.39432028582284,
      "tokens_seen": 186318848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000660015131454511,
      "loss": 3.27,
      "theoretical_loss": 4.394124790567101,
      "tokens_seen": 186384384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006599985814261395,
      "loss": 3.4215,
      "theoretical_loss": 4.3939293832781825,
      "tokens_seen": 186449920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006599820313977681,
      "loss": 3.2849,
      "theoretical_loss": 4.393734063885599,
      "tokens_seen": 186515456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006599654813693966,
      "loss": 3.2054,
      "theoretical_loss": 4.3935388323189555,
      "tokens_seen": 186580992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006599489313410252,
      "loss": 3.0835,
      "theoretical_loss": 4.39334368850793,
      "tokens_seen": 186646528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006599323813126537,
      "loss": 2.9941,
      "theoretical_loss": 4.3931486323822835,
      "tokens_seen": 186712064
    },
    {
      "epoch": 0.07,
      "objective/train/avg_token_score": 0.024841099977493286,
      "objective/train/avg_weight": 0.9707555174827576,
      "objective/train/docs_used": 117211,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.107152223587036,
      "objective/train/std_weight": 0.1440294235944748,
      "objective/train/theoretical_loss": 4.392953663871862,
      "objective/train/tokens_used": 207237600,
      "theoretical_loss": 4.392953663871862,
      "tokens_seen": 186777600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006599158312842821,
      "loss": 2.9809,
      "theoretical_loss": 4.392953663871862,
      "tokens_seen": 186777600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006598992812559107,
      "loss": 3.1144,
      "theoretical_loss": 4.392758782906586,
      "tokens_seen": 186843136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006598827312275392,
      "loss": 2.9598,
      "theoretical_loss": 4.392563989416462,
      "tokens_seen": 186908672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006598661811991677,
      "loss": 3.4255,
      "theoretical_loss": 4.392369283331574,
      "tokens_seen": 186974208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006598496311707963,
      "loss": 3.1204,
      "theoretical_loss": 4.392174664582085,
      "tokens_seen": 187039744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006598330811424248,
      "loss": 3.0607,
      "theoretical_loss": 4.391980133098244,
      "tokens_seen": 187105280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006598165311140534,
      "loss": 2.9689,
      "theoretical_loss": 4.391785688810373,
      "tokens_seen": 187170816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006597999810856819,
      "loss": 3.0589,
      "theoretical_loss": 4.391591331648879,
      "tokens_seen": 187236352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006597834310573103,
      "loss": 3.2886,
      "theoretical_loss": 4.391397061544247,
      "tokens_seen": 187301888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006597668810289389,
      "loss": 3.251,
      "theoretical_loss": 4.391202878427042,
      "tokens_seen": 187367424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006597503310005674,
      "loss": 2.9487,
      "theoretical_loss": 4.3910087822279085,
      "tokens_seen": 187432960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006597337809721959,
      "loss": 3.1792,
      "theoretical_loss": 4.390814772877571,
      "tokens_seen": 187498496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006597172309438245,
      "loss": 3.067,
      "theoretical_loss": 4.390620850306832,
      "tokens_seen": 187564032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000659700680915453,
      "loss": 3.2124,
      "theoretical_loss": 4.390427014446575,
      "tokens_seen": 187629568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006596841308870816,
      "loss": 3.3088,
      "theoretical_loss": 4.390233265227764,
      "tokens_seen": 187695104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00065966758085871,
      "loss": 2.8368,
      "theoretical_loss": 4.390039602581437,
      "tokens_seen": 187760640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006596510308303385,
      "loss": 3.1249,
      "theoretical_loss": 4.389846026438715,
      "tokens_seen": 187826176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006596344808019671,
      "loss": 3.2422,
      "theoretical_loss": 4.3896525367307975,
      "tokens_seen": 187891712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006596179307735956,
      "loss": 3.0593,
      "theoretical_loss": 4.389459133388962,
      "tokens_seen": 187957248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006596013807452241,
      "loss": 3.2302,
      "theoretical_loss": 4.3892658163445635,
      "tokens_seen": 188022784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006595848307168527,
      "loss": 3.2379,
      "theoretical_loss": 4.389072585529037,
      "tokens_seen": 188088320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006595682806884812,
      "loss": 3.1248,
      "theoretical_loss": 4.388879440873897,
      "tokens_seen": 188153856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006595517306601098,
      "loss": 3.3185,
      "theoretical_loss": 4.388686382310732,
      "tokens_seen": 188219392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006595351806317382,
      "loss": 3.0142,
      "theoretical_loss": 4.388493409771213,
      "tokens_seen": 188284928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006595186306033667,
      "loss": 3.2929,
      "theoretical_loss": 4.388300523187087,
      "tokens_seen": 188350464
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006595020805749953,
      "loss": 3.0178,
      "theoretical_loss": 4.3881077224901786,
      "tokens_seen": 188416000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006594855305466238,
      "loss": 3.0836,
      "theoretical_loss": 4.38791500761239,
      "tokens_seen": 188481536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006594689805182523,
      "loss": 3.2027,
      "theoretical_loss": 4.387722378485703,
      "tokens_seen": 188547072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006594524304898809,
      "loss": 2.8479,
      "theoretical_loss": 4.3875298350421765,
      "tokens_seen": 188612608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006594358804615094,
      "loss": 3.0677,
      "theoretical_loss": 4.387337377213943,
      "tokens_seen": 188678144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000659419330433138,
      "loss": 3.2455,
      "theoretical_loss": 4.387145004933218,
      "tokens_seen": 188743680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006594027804047664,
      "loss": 3.2741,
      "theoretical_loss": 4.38695271813229,
      "tokens_seen": 188809216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006593862303763949,
      "loss": 3.352,
      "theoretical_loss": 4.386760516743526,
      "tokens_seen": 188874752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006593696803480235,
      "loss": 3.2365,
      "theoretical_loss": 4.38656840069937,
      "tokens_seen": 188940288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000659353130319652,
      "loss": 2.7961,
      "theoretical_loss": 4.386376369932344,
      "tokens_seen": 189005824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006593365802912805,
      "loss": 3.1299,
      "theoretical_loss": 4.386184424375044,
      "tokens_seen": 189071360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006593200302629091,
      "loss": 2.94,
      "theoretical_loss": 4.385992563960145,
      "tokens_seen": 189136896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006593034802345375,
      "loss": 3.2045,
      "theoretical_loss": 4.385800788620397,
      "tokens_seen": 189202432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006592869302061661,
      "loss": 3.3168,
      "theoretical_loss": 4.385609098288628,
      "tokens_seen": 189267968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006592703801777946,
      "loss": 2.9308,
      "theoretical_loss": 4.385417492897741,
      "tokens_seen": 189333504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006592538301494231,
      "loss": 3.1603,
      "theoretical_loss": 4.385225972380715,
      "tokens_seen": 189399040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006592372801210517,
      "loss": 3.1525,
      "theoretical_loss": 4.385034536670606,
      "tokens_seen": 189464576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006592207300926802,
      "loss": 3.3267,
      "theoretical_loss": 4.384843185700544,
      "tokens_seen": 189530112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006592041800643087,
      "loss": 3.0528,
      "theoretical_loss": 4.384651919403739,
      "tokens_seen": 189595648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006591876300359371,
      "loss": 3.1569,
      "theoretical_loss": 4.384460737713471,
      "tokens_seen": 189661184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006591710800075656,
      "loss": 2.9377,
      "theoretical_loss": 4.384269640563101,
      "tokens_seen": 189726720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006591545299791942,
      "loss": 3.1563,
      "theoretical_loss": 4.384078627886062,
      "tokens_seen": 189792256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006591379799508227,
      "loss": 3.2032,
      "theoretical_loss": 4.383887699615863,
      "tokens_seen": 189857792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006591214299224513,
      "loss": 3.2693,
      "theoretical_loss": 4.38369685568609,
      "tokens_seen": 189923328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006591048798940798,
      "loss": 3.1688,
      "theoretical_loss": 4.383506096030401,
      "tokens_seen": 189988864
    },
    {
      "epoch": 0.07,
      "objective/train/avg_token_score": 0.03682224825024605,
      "objective/train/avg_weight": 0.9566702246665955,
      "objective/train/docs_used": 119241,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.858774423599243,
      "objective/train/std_weight": 0.15509475767612457,
      "objective/train/theoretical_loss": 4.383315420582533,
      "objective/train/tokens_used": 210514400,
      "theoretical_loss": 4.383315420582533,
      "tokens_seen": 190054400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006590883298657083,
      "loss": 2.9603,
      "theoretical_loss": 4.383315420582533,
      "tokens_seen": 190054400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006590717798373369,
      "loss": 3.0445,
      "theoretical_loss": 4.383124829276294,
      "tokens_seen": 190119936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006590552298089653,
      "loss": 3.1161,
      "theoretical_loss": 4.38293432204557,
      "tokens_seen": 190185472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006590386797805938,
      "loss": 3.2938,
      "theoretical_loss": 4.382743898824321,
      "tokens_seen": 190251008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006590221297522224,
      "loss": 3.3207,
      "theoretical_loss": 4.3825535595465785,
      "tokens_seen": 190316544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006590055797238509,
      "loss": 3.1285,
      "theoretical_loss": 4.382363304146453,
      "tokens_seen": 190382080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006589890296954795,
      "loss": 3.1284,
      "theoretical_loss": 4.382173132558126,
      "tokens_seen": 190447616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000658972479667108,
      "loss": 3.3484,
      "theoretical_loss": 4.381983044715856,
      "tokens_seen": 190513152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006589559296387365,
      "loss": 3.089,
      "theoretical_loss": 4.381793040553973,
      "tokens_seen": 190578688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000658939379610365,
      "loss": 3.3373,
      "theoretical_loss": 4.381603120006883,
      "tokens_seen": 190644224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006589228295819935,
      "loss": 3.26,
      "theoretical_loss": 4.381413283009065,
      "tokens_seen": 190709760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006589062795536221,
      "loss": 2.967,
      "theoretical_loss": 4.381223529495073,
      "tokens_seen": 190775296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006588897295252506,
      "loss": 3.4113,
      "theoretical_loss": 4.381033859399532,
      "tokens_seen": 190840832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006588731794968791,
      "loss": 3.2272,
      "theoretical_loss": 4.380844272657145,
      "tokens_seen": 190906368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006588566294685077,
      "loss": 3.2678,
      "theoretical_loss": 4.380654769202683,
      "tokens_seen": 190971904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006588400794401362,
      "loss": 3.209,
      "theoretical_loss": 4.380465348970995,
      "tokens_seen": 191037440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006588235294117646,
      "loss": 3.2978,
      "theoretical_loss": 4.380276011897003,
      "tokens_seen": 191102976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006588069793833932,
      "loss": 2.9309,
      "theoretical_loss": 4.380086757915698,
      "tokens_seen": 191168512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006587904293550217,
      "loss": 3.0703,
      "theoretical_loss": 4.379897586962148,
      "tokens_seen": 191234048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006587738793266503,
      "loss": 3.1311,
      "theoretical_loss": 4.379708498971494,
      "tokens_seen": 191299584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006587573292982788,
      "loss": 3.1374,
      "theoretical_loss": 4.379519493878948,
      "tokens_seen": 191365120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006587407792699073,
      "loss": 3.1717,
      "theoretical_loss": 4.379330571619795,
      "tokens_seen": 191430656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006587242292415359,
      "loss": 2.9421,
      "theoretical_loss": 4.379141732129394,
      "tokens_seen": 191496192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006587076792131644,
      "loss": 3.0792,
      "theoretical_loss": 4.378952975343175,
      "tokens_seen": 191561728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006586911291847928,
      "loss": 3.0287,
      "theoretical_loss": 4.378764301196642,
      "tokens_seen": 191627264
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006586745791564214,
      "loss": 3.3077,
      "theoretical_loss": 4.37857570962537,
      "tokens_seen": 191692800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006586580291280499,
      "loss": 3.3375,
      "theoretical_loss": 4.378387200565006,
      "tokens_seen": 191758336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006586414790996785,
      "loss": 3.2138,
      "theoretical_loss": 4.378198773951272,
      "tokens_seen": 191823872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000658624929071307,
      "loss": 3.292,
      "theoretical_loss": 4.378010429719957,
      "tokens_seen": 191889408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006586083790429355,
      "loss": 3.0554,
      "theoretical_loss": 4.377822167806928,
      "tokens_seen": 191954944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006585918290145641,
      "loss": 3.0855,
      "theoretical_loss": 4.377633988148117,
      "tokens_seen": 192020480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006585752789861925,
      "loss": 3.2947,
      "theoretical_loss": 4.377445890679534,
      "tokens_seen": 192086016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000658558728957821,
      "loss": 3.1896,
      "theoretical_loss": 4.377257875337257,
      "tokens_seen": 192151552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006585421789294496,
      "loss": 3.3454,
      "theoretical_loss": 4.377069942057436,
      "tokens_seen": 192217088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006585256289010781,
      "loss": 3.2128,
      "theoretical_loss": 4.376882090776293,
      "tokens_seen": 192282624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006585090788727067,
      "loss": 2.9753,
      "theoretical_loss": 4.376694321430121,
      "tokens_seen": 192348160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006584925288443352,
      "loss": 3.155,
      "theoretical_loss": 4.376506633955286,
      "tokens_seen": 192413696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006584759788159637,
      "loss": 2.8734,
      "theoretical_loss": 4.376319028288219,
      "tokens_seen": 192479232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006584594287875923,
      "loss": 3.1692,
      "theoretical_loss": 4.37613150436543,
      "tokens_seen": 192544768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006584428787592207,
      "loss": 3.0772,
      "theoretical_loss": 4.375944062123496,
      "tokens_seen": 192610304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006584263287308492,
      "loss": 3.1536,
      "theoretical_loss": 4.375756701499063,
      "tokens_seen": 192675840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006584097787024778,
      "loss": 3.1167,
      "theoretical_loss": 4.3755694224288515,
      "tokens_seen": 192741376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006583932286741063,
      "loss": 3.2348,
      "theoretical_loss": 4.375382224849648,
      "tokens_seen": 192806912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006583766786457349,
      "loss": 3.2104,
      "theoretical_loss": 4.375195108698316,
      "tokens_seen": 192872448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006583601286173634,
      "loss": 3.1406,
      "theoretical_loss": 4.375008073911781,
      "tokens_seen": 192937984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006583435785889919,
      "loss": 2.9954,
      "theoretical_loss": 4.374821120427047,
      "tokens_seen": 193003520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006583270285606204,
      "loss": 3.2136,
      "theoretical_loss": 4.374634248181182,
      "tokens_seen": 193069056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006583104785322489,
      "loss": 3.0939,
      "theoretical_loss": 4.3744474571113265,
      "tokens_seen": 193134592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006582939285038774,
      "loss": 3.206,
      "theoretical_loss": 4.374260747154692,
      "tokens_seen": 193200128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000658277378475506,
      "loss": 3.3662,
      "theoretical_loss": 4.374074118248559,
      "tokens_seen": 193265664
    },
    {
      "epoch": 0.07,
      "objective/train/avg_token_score": 0.013809370808303356,
      "objective/train/avg_weight": 0.9837430715560913,
      "objective/train/docs_used": 121008,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2585206031799316,
      "objective/train/std_weight": 0.0666441097855568,
      "objective/train/theoretical_loss": 4.373887570330275,
      "objective/train/tokens_used": 213791200,
      "theoretical_loss": 4.373887570330275,
      "tokens_seen": 193331200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006582608284471345,
      "loss": 3.2142,
      "theoretical_loss": 4.373887570330275,
      "tokens_seen": 193331200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006582442784187631,
      "loss": 2.787,
      "theoretical_loss": 4.373701103337263,
      "tokens_seen": 193396736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006582277283903916,
      "loss": 3.1828,
      "theoretical_loss": 4.373514717207009,
      "tokens_seen": 193462272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00065821117836202,
      "loss": 3.2643,
      "theoretical_loss": 4.373328411877073,
      "tokens_seen": 193527808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006581946283336485,
      "loss": 3.0889,
      "theoretical_loss": 4.373142187285083,
      "tokens_seen": 193593344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000658178078305277,
      "loss": 3.4174,
      "theoretical_loss": 4.372956043368736,
      "tokens_seen": 193658880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006581615282769056,
      "loss": 3.0152,
      "theoretical_loss": 4.372769980065797,
      "tokens_seen": 193724416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006581449782485341,
      "loss": 3.198,
      "theoretical_loss": 4.372583997314104,
      "tokens_seen": 193789952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006581284282201626,
      "loss": 3.2252,
      "theoretical_loss": 4.372398095051559,
      "tokens_seen": 193855488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006581118781917912,
      "loss": 3.1903,
      "theoretical_loss": 4.372212273216136,
      "tokens_seen": 193921024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006580953281634196,
      "loss": 3.2049,
      "theoretical_loss": 4.372026531745877,
      "tokens_seen": 193986560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006580787781350482,
      "loss": 3.3349,
      "theoretical_loss": 4.371840870578891,
      "tokens_seen": 194052096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006580622281066767,
      "loss": 3.1832,
      "theoretical_loss": 4.37165528965336,
      "tokens_seen": 194117632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006580456780783052,
      "loss": 3.1234,
      "theoretical_loss": 4.371469788907529,
      "tokens_seen": 194183168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006580291280499338,
      "loss": 3.2177,
      "theoretical_loss": 4.371284368279714,
      "tokens_seen": 194248704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006580125780215623,
      "loss": 3.2629,
      "theoretical_loss": 4.3710990277083,
      "tokens_seen": 194314240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006579960279931908,
      "loss": 2.9788,
      "theoretical_loss": 4.3709137671317375,
      "tokens_seen": 194379776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006579794779648194,
      "loss": 3.0926,
      "theoretical_loss": 4.37072858648855,
      "tokens_seen": 194445312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006579629279364478,
      "loss": 3.2941,
      "theoretical_loss": 4.370543485717322,
      "tokens_seen": 194510848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006579463779080764,
      "loss": 3.4197,
      "theoretical_loss": 4.370358464756713,
      "tokens_seen": 194576384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006579298278797049,
      "loss": 3.1111,
      "theoretical_loss": 4.370173523545443,
      "tokens_seen": 194641920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006579132778513334,
      "loss": 3.1774,
      "theoretical_loss": 4.3699886620223065,
      "tokens_seen": 194707456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000657896727822962,
      "loss": 3.0885,
      "theoretical_loss": 4.369803880126162,
      "tokens_seen": 194772992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006578801777945905,
      "loss": 3.2288,
      "theoretical_loss": 4.3696191777959354,
      "tokens_seen": 194838528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000657863627766219,
      "loss": 3.2316,
      "theoretical_loss": 4.369434554970621,
      "tokens_seen": 194904064
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006578470777378476,
      "loss": 3.1186,
      "theoretical_loss": 4.369250011589279,
      "tokens_seen": 194969600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000657830527709476,
      "loss": 3.1571,
      "theoretical_loss": 4.369065547591038,
      "tokens_seen": 195035136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006578139776811046,
      "loss": 2.8633,
      "theoretical_loss": 4.368881162915095,
      "tokens_seen": 195100672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006577974276527331,
      "loss": 3.1444,
      "theoretical_loss": 4.36869685750071,
      "tokens_seen": 195166208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006577808776243616,
      "loss": 2.8825,
      "theoretical_loss": 4.3685126312872145,
      "tokens_seen": 195231744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006577643275959902,
      "loss": 2.9839,
      "theoretical_loss": 4.368328484214002,
      "tokens_seen": 195297280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006577477775676187,
      "loss": 3.1039,
      "theoretical_loss": 4.368144416220538,
      "tokens_seen": 195362816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006577312275392472,
      "loss": 3.1468,
      "theoretical_loss": 4.3679604272463495,
      "tokens_seen": 195428352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006577146775108757,
      "loss": 3.4244,
      "theoretical_loss": 4.367776517231033,
      "tokens_seen": 195493888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006576981274825042,
      "loss": 2.9499,
      "theoretical_loss": 4.367592686114252,
      "tokens_seen": 195559424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006576815774541328,
      "loss": 2.9999,
      "theoretical_loss": 4.367408933835733,
      "tokens_seen": 195624960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006576650274257613,
      "loss": 3.0627,
      "theoretical_loss": 4.367225260335272,
      "tokens_seen": 195690496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006576484773973898,
      "loss": 3.0668,
      "theoretical_loss": 4.36704166555273,
      "tokens_seen": 195756032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006576319273690184,
      "loss": 3.1117,
      "theoretical_loss": 4.366858149428032,
      "tokens_seen": 195821568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006576153773406469,
      "loss": 3.2581,
      "theoretical_loss": 4.366674711901173,
      "tokens_seen": 195887104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006575988273122753,
      "loss": 3.2142,
      "theoretical_loss": 4.366491352912211,
      "tokens_seen": 195952640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006575822772839039,
      "loss": 2.9181,
      "theoretical_loss": 4.366308072401271,
      "tokens_seen": 196018176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006575657272555324,
      "loss": 2.8985,
      "theoretical_loss": 4.366124870308541,
      "tokens_seen": 196083712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000657549177227161,
      "loss": 3.1106,
      "theoretical_loss": 4.365941746574278,
      "tokens_seen": 196149248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006575326271987895,
      "loss": 3.1396,
      "theoretical_loss": 4.3657587011388035,
      "tokens_seen": 196214784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000657516077170418,
      "loss": 2.9164,
      "theoretical_loss": 4.365575733942503,
      "tokens_seen": 196280320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006574995271420466,
      "loss": 3.1045,
      "theoretical_loss": 4.365392844925829,
      "tokens_seen": 196345856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000657482977113675,
      "loss": 3.09,
      "theoretical_loss": 4.365210034029298,
      "tokens_seen": 196411392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006574664270853036,
      "loss": 3.0893,
      "theoretical_loss": 4.365027301193491,
      "tokens_seen": 196476928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006574498770569321,
      "loss": 3.1255,
      "theoretical_loss": 4.364844646359056,
      "tokens_seen": 196542464
    },
    {
      "epoch": 0.07,
      "objective/train/avg_token_score": 0.008739003911614418,
      "objective/train/avg_weight": 0.9897218346595764,
      "objective/train/docs_used": 122985,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2708659172058105,
      "objective/train/std_weight": 0.06166733801364899,
      "objective/train/theoretical_loss": 4.364662069466704,
      "objective/train/tokens_used": 217068000,
      "theoretical_loss": 4.364662069466704,
      "tokens_seen": 196608000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006574333270285606,
      "loss": 3.2246,
      "theoretical_loss": 4.364662069466704,
      "tokens_seen": 196608000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006574167770001892,
      "loss": 3.0676,
      "theoretical_loss": 4.364479570457213,
      "tokens_seen": 196673536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006574002269718177,
      "loss": 3.0943,
      "theoretical_loss": 4.364297149271423,
      "tokens_seen": 196739072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006573836769434462,
      "loss": 3.2819,
      "theoretical_loss": 4.3641148058502415,
      "tokens_seen": 196804608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006573671269150748,
      "loss": 3.2736,
      "theoretical_loss": 4.363932540134638,
      "tokens_seen": 196870144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006573505768867032,
      "loss": 3.1331,
      "theoretical_loss": 4.363750352065647,
      "tokens_seen": 196935680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006573340268583318,
      "loss": 3.2282,
      "theoretical_loss": 4.363568241584368,
      "tokens_seen": 197001216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006573174768299603,
      "loss": 3.1818,
      "theoretical_loss": 4.363386208631966,
      "tokens_seen": 197066752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006573009268015888,
      "loss": 3.0321,
      "theoretical_loss": 4.363204253149667,
      "tokens_seen": 197132288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006572843767732174,
      "loss": 3.0756,
      "theoretical_loss": 4.3630223750787644,
      "tokens_seen": 197197824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006572678267448459,
      "loss": 3.1079,
      "theoretical_loss": 4.362840574360612,
      "tokens_seen": 197263360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006572512767164744,
      "loss": 3.3191,
      "theoretical_loss": 4.362658850936631,
      "tokens_seen": 197328896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000657234726688103,
      "loss": 3.0622,
      "theoretical_loss": 4.362477204748305,
      "tokens_seen": 197394432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006572181766597313,
      "loss": 2.9964,
      "theoretical_loss": 4.362295635737179,
      "tokens_seen": 197459968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006572016266313599,
      "loss": 3.1809,
      "theoretical_loss": 4.362114143844867,
      "tokens_seen": 197525504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006571850766029884,
      "loss": 3.0136,
      "theoretical_loss": 4.3619327290130405,
      "tokens_seen": 197591040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000657168526574617,
      "loss": 3.2796,
      "theoretical_loss": 4.3617513911834385,
      "tokens_seen": 197656576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006571519765462455,
      "loss": 3.0917,
      "theoretical_loss": 4.361570130297863,
      "tokens_seen": 197722112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000657135426517874,
      "loss": 3.1222,
      "theoretical_loss": 4.3613889462981765,
      "tokens_seen": 197787648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006571188764895026,
      "loss": 3.3036,
      "theoretical_loss": 4.361207839126308,
      "tokens_seen": 197853184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000657102326461131,
      "loss": 2.9454,
      "theoretical_loss": 4.361026808724247,
      "tokens_seen": 197918720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006570857764327595,
      "loss": 2.992,
      "theoretical_loss": 4.360845855034049,
      "tokens_seen": 197984256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006570692264043881,
      "loss": 3.1847,
      "theoretical_loss": 4.360664977997828,
      "tokens_seen": 198049792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006570526763760166,
      "loss": 3.2333,
      "theoretical_loss": 4.360484177557766,
      "tokens_seen": 198115328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006570361263476452,
      "loss": 3.1284,
      "theoretical_loss": 4.360303453656103,
      "tokens_seen": 198180864
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006570195763192737,
      "loss": 2.6822,
      "theoretical_loss": 4.360122806235145,
      "tokens_seen": 198246400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006570030262909022,
      "loss": 3.2369,
      "theoretical_loss": 4.359942235237257,
      "tokens_seen": 198311936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006569864762625307,
      "loss": 3.207,
      "theoretical_loss": 4.359761740604871,
      "tokens_seen": 198377472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006569699262341592,
      "loss": 3.1505,
      "theoretical_loss": 4.359581322280479,
      "tokens_seen": 198443008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006569533762057877,
      "loss": 3.123,
      "theoretical_loss": 4.359400980206634,
      "tokens_seen": 198508544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006569368261774163,
      "loss": 3.0522,
      "theoretical_loss": 4.359220714325954,
      "tokens_seen": 198574080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006569202761490448,
      "loss": 3.2259,
      "theoretical_loss": 4.359040524581116,
      "tokens_seen": 198639616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006569037261206734,
      "loss": 3.0882,
      "theoretical_loss": 4.358860410914861,
      "tokens_seen": 198705152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006568871760923019,
      "loss": 2.9455,
      "theoretical_loss": 4.358680373269993,
      "tokens_seen": 198770688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006568706260639303,
      "loss": 3.0858,
      "theoretical_loss": 4.358500411589375,
      "tokens_seen": 198836224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006568540760355589,
      "loss": 3.1393,
      "theoretical_loss": 4.358320525815934,
      "tokens_seen": 198901760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006568375260071874,
      "loss": 3.0857,
      "theoretical_loss": 4.358140715892658,
      "tokens_seen": 198967296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006568209759788159,
      "loss": 3.2692,
      "theoretical_loss": 4.357960981762595,
      "tokens_seen": 199032832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006568044259504445,
      "loss": 3.0033,
      "theoretical_loss": 4.357781323368857,
      "tokens_seen": 199098368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000656787875922073,
      "loss": 3.3749,
      "theoretical_loss": 4.357601740654617,
      "tokens_seen": 199163904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006567713258937016,
      "loss": 3.0396,
      "theoretical_loss": 4.357422233563106,
      "tokens_seen": 199229440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00065675477586533,
      "loss": 3.1333,
      "theoretical_loss": 4.357242802037623,
      "tokens_seen": 199294976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006567382258369585,
      "loss": 3.2875,
      "theoretical_loss": 4.35706344602152,
      "tokens_seen": 199360512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006567216758085871,
      "loss": 3.1299,
      "theoretical_loss": 4.356884165458217,
      "tokens_seen": 199426048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006567051257802156,
      "loss": 3.2723,
      "theoretical_loss": 4.356704960291191,
      "tokens_seen": 199491584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006566885757518441,
      "loss": 3.2191,
      "theoretical_loss": 4.35652583046398,
      "tokens_seen": 199557120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006566720257234727,
      "loss": 3.0116,
      "theoretical_loss": 4.356346775920185,
      "tokens_seen": 199622656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006566554756951012,
      "loss": 3.0359,
      "theoretical_loss": 4.356167796603467,
      "tokens_seen": 199688192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006566389256667298,
      "loss": 3.2042,
      "theoretical_loss": 4.355988892457546,
      "tokens_seen": 199753728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006566223756383582,
      "loss": 3.0755,
      "theoretical_loss": 4.355810063426204,
      "tokens_seen": 199819264
    },
    {
      "epoch": 0.07,
      "objective/train/avg_token_score": 0.016947783529758453,
      "objective/train/avg_weight": 0.9800531268119812,
      "objective/train/docs_used": 124865,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.051406145095825,
      "objective/train/std_weight": 0.09366383403539658,
      "objective/train/theoretical_loss": 4.355631309453283,
      "objective/train/tokens_used": 220344800,
      "theoretical_loss": 4.355631309453283,
      "tokens_seen": 199884800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006566058256099867,
      "loss": 3.023,
      "theoretical_loss": 4.355631309453283,
      "tokens_seen": 199884800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006565892755816153,
      "loss": 2.9565,
      "theoretical_loss": 4.355452630482685,
      "tokens_seen": 199950336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006565727255532438,
      "loss": 3.1016,
      "theoretical_loss": 4.355274026458375,
      "tokens_seen": 200015872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006565561755248723,
      "loss": 3.1353,
      "theoretical_loss": 4.355095497324373,
      "tokens_seen": 200081408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006565396254965009,
      "loss": 3.1813,
      "theoretical_loss": 4.354917043024765,
      "tokens_seen": 200146944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006565230754681294,
      "loss": 3.1296,
      "theoretical_loss": 4.354738663503692,
      "tokens_seen": 200212480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000656506525439758,
      "loss": 3.0806,
      "theoretical_loss": 4.354560358705358,
      "tokens_seen": 200278016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006564899754113864,
      "loss": 3.1247,
      "theoretical_loss": 4.354382128574027,
      "tokens_seen": 200343552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006564734253830149,
      "loss": 2.9845,
      "theoretical_loss": 4.35420397305402,
      "tokens_seen": 200409088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006564568753546435,
      "loss": 3.1587,
      "theoretical_loss": 4.35402589208972,
      "tokens_seen": 200474624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000656440325326272,
      "loss": 2.9947,
      "theoretical_loss": 4.353847885625571,
      "tokens_seen": 200540160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006564237752979005,
      "loss": 2.9029,
      "theoretical_loss": 4.353669953606072,
      "tokens_seen": 200605696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006564072252695291,
      "loss": 3.37,
      "theoretical_loss": 4.353492095975787,
      "tokens_seen": 200671232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006563906752411576,
      "loss": 3.1724,
      "theoretical_loss": 4.353314312679333,
      "tokens_seen": 200736768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006563741252127861,
      "loss": 3.1489,
      "theoretical_loss": 4.353136603661392,
      "tokens_seen": 200802304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006563575751844146,
      "loss": 2.9855,
      "theoretical_loss": 4.352958968866704,
      "tokens_seen": 200867840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006563410251560431,
      "loss": 3.1687,
      "theoretical_loss": 4.352781408240065,
      "tokens_seen": 200933376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006563244751276717,
      "loss": 3.3378,
      "theoretical_loss": 4.352603921726334,
      "tokens_seen": 200998912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006563079250993002,
      "loss": 2.8935,
      "theoretical_loss": 4.352426509270425,
      "tokens_seen": 201064448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006562913750709287,
      "loss": 2.9528,
      "theoretical_loss": 4.352249170817315,
      "tokens_seen": 201129984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006562748250425573,
      "loss": 3.1029,
      "theoretical_loss": 4.352071906312037,
      "tokens_seen": 201195520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006562582750141857,
      "loss": 3.2121,
      "theoretical_loss": 4.351894715699684,
      "tokens_seen": 201261056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006562417249858142,
      "loss": 3.0713,
      "theoretical_loss": 4.351717598925406,
      "tokens_seen": 201326592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006562251749574427,
      "loss": 3.0281,
      "theoretical_loss": 4.351540555934414,
      "tokens_seen": 201392128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006562086249290713,
      "loss": 3.2604,
      "theoretical_loss": 4.351363586671976,
      "tokens_seen": 201457664
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006561920749006998,
      "loss": 3.0746,
      "theoretical_loss": 4.351186691083417,
      "tokens_seen": 201523200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006561755248723283,
      "loss": 3.0675,
      "theoretical_loss": 4.351009869114124,
      "tokens_seen": 201588736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006561589748439569,
      "loss": 3.1692,
      "theoretical_loss": 4.350833120709539,
      "tokens_seen": 201654272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006561424248155853,
      "loss": 3.0994,
      "theoretical_loss": 4.350656445815164,
      "tokens_seen": 201719808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006561258747872139,
      "loss": 3.1096,
      "theoretical_loss": 4.350479844376557,
      "tokens_seen": 201785344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006561093247588424,
      "loss": 3.0867,
      "theoretical_loss": 4.350303316339337,
      "tokens_seen": 201850880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006560927747304709,
      "loss": 3.0192,
      "theoretical_loss": 4.350126861649178,
      "tokens_seen": 201916416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006560762247020995,
      "loss": 3.112,
      "theoretical_loss": 4.349950480251813,
      "tokens_seen": 201981952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000656059674673728,
      "loss": 3.0123,
      "theoretical_loss": 4.349774172093033,
      "tokens_seen": 202047488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006560431246453565,
      "loss": 3.014,
      "theoretical_loss": 4.349597937118687,
      "tokens_seen": 202113024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000656026574616985,
      "loss": 2.9305,
      "theoretical_loss": 4.3494217752746795,
      "tokens_seen": 202178560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006560100245886135,
      "loss": 3.0342,
      "theoretical_loss": 4.349245686506976,
      "tokens_seen": 202244096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006559934745602421,
      "loss": 2.8373,
      "theoretical_loss": 4.349069670761597,
      "tokens_seen": 202309632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006559769245318706,
      "loss": 3.1335,
      "theoretical_loss": 4.348893727984619,
      "tokens_seen": 202375168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006559603745034991,
      "loss": 2.9881,
      "theoretical_loss": 4.348717858122178,
      "tokens_seen": 202440704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006559438244751277,
      "loss": 3.2555,
      "theoretical_loss": 4.348542061120469,
      "tokens_seen": 202506240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006559272744467562,
      "loss": 3.1164,
      "theoretical_loss": 4.348366336925739,
      "tokens_seen": 202571776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006559107244183847,
      "loss": 2.9599,
      "theoretical_loss": 4.3481906854842975,
      "tokens_seen": 202637312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006558941743900132,
      "loss": 2.9273,
      "theoretical_loss": 4.348015106742507,
      "tokens_seen": 202702848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006558776243616417,
      "loss": 3.0634,
      "theoretical_loss": 4.347839600646786,
      "tokens_seen": 202768384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006558610743332703,
      "loss": 3.0639,
      "theoretical_loss": 4.347664167143615,
      "tokens_seen": 202833920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006558445243048988,
      "loss": 3.156,
      "theoretical_loss": 4.347488806179528,
      "tokens_seen": 202899456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006558279742765273,
      "loss": 3.0367,
      "theoretical_loss": 4.347313517701114,
      "tokens_seen": 202964992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006558114242481559,
      "loss": 3.1842,
      "theoretical_loss": 4.347138301655021,
      "tokens_seen": 203030528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006557948742197844,
      "loss": 3.0741,
      "theoretical_loss": 4.346963157987954,
      "tokens_seen": 203096064
    },
    {
      "epoch": 0.07,
      "objective/train/avg_token_score": 0.024062547832727432,
      "objective/train/avg_weight": 0.9716774225234985,
      "objective/train/docs_used": 126478,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.322425603866577,
      "objective/train/std_weight": 0.1419600397348404,
      "objective/train/theoretical_loss": 4.346788086646671,
      "objective/train/tokens_used": 223621600,
      "theoretical_loss": 4.346788086646671,
      "tokens_seen": 203161600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006557783241914128,
      "loss": 3.0572,
      "theoretical_loss": 4.346788086646671,
      "tokens_seen": 203161600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006557617741630414,
      "loss": 3.0174,
      "theoretical_loss": 4.346613087577991,
      "tokens_seen": 203227136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006557452241346699,
      "loss": 3.0933,
      "theoretical_loss": 4.346438160728785,
      "tokens_seen": 203292672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006557286741062985,
      "loss": 3.0344,
      "theoretical_loss": 4.346263306045983,
      "tokens_seen": 203358208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000655712124077927,
      "loss": 3.027,
      "theoretical_loss": 4.346088523476569,
      "tokens_seen": 203423744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006556955740495555,
      "loss": 3.3949,
      "theoretical_loss": 4.345913812967584,
      "tokens_seen": 203489280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006556790240211841,
      "loss": 3.3346,
      "theoretical_loss": 4.345739174466127,
      "tokens_seen": 203554816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006556624739928126,
      "loss": 3.2075,
      "theoretical_loss": 4.345564607919348,
      "tokens_seen": 203620352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000655645923964441,
      "loss": 3.1623,
      "theoretical_loss": 4.3453901132744575,
      "tokens_seen": 203685888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006556293739360696,
      "loss": 2.9024,
      "theoretical_loss": 4.345215690478719,
      "tokens_seen": 203751424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006556128239076981,
      "loss": 3.1622,
      "theoretical_loss": 4.345041339479453,
      "tokens_seen": 203816960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006555962738793267,
      "loss": 3.0407,
      "theoretical_loss": 4.3448670602240345,
      "tokens_seen": 203882496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006555797238509552,
      "loss": 3.2627,
      "theoretical_loss": 4.344692852659895,
      "tokens_seen": 203948032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006555631738225837,
      "loss": 3.1313,
      "theoretical_loss": 4.34451871673452,
      "tokens_seen": 204013568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006555466237942123,
      "loss": 3.2857,
      "theoretical_loss": 4.344344652395451,
      "tokens_seen": 204079104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006555300737658407,
      "loss": 2.9221,
      "theoretical_loss": 4.3441706595902865,
      "tokens_seen": 204144640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006555135237374692,
      "loss": 3.1121,
      "theoretical_loss": 4.343996738266677,
      "tokens_seen": 204210176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006554969737090978,
      "loss": 3.3923,
      "theoretical_loss": 4.343822888372331,
      "tokens_seen": 204275712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006554804236807263,
      "loss": 3.0616,
      "theoretical_loss": 4.343649109855009,
      "tokens_seen": 204341248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006554638736523549,
      "loss": 3.1175,
      "theoretical_loss": 4.343475402662529,
      "tokens_seen": 204406784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006554473236239834,
      "loss": 3.1258,
      "theoretical_loss": 4.343301766742763,
      "tokens_seen": 204472320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006554307735956119,
      "loss": 3.2925,
      "theoretical_loss": 4.343128202043638,
      "tokens_seen": 204537856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006554142235672405,
      "loss": 2.9715,
      "theoretical_loss": 4.342954708513136,
      "tokens_seen": 204603392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006553976735388689,
      "loss": 3.1518,
      "theoretical_loss": 4.342781286099291,
      "tokens_seen": 204668928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006553811235104974,
      "loss": 2.9769,
      "theoretical_loss": 4.3426079347501965,
      "tokens_seen": 204734464
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000655364573482126,
      "loss": 3.0575,
      "theoretical_loss": 4.342434654413995,
      "tokens_seen": 204800000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006553480234537545,
      "loss": 3.2879,
      "theoretical_loss": 4.342261445038888,
      "tokens_seen": 204865536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006553314734253831,
      "loss": 3.187,
      "theoretical_loss": 4.342088306573128,
      "tokens_seen": 204931072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006553149233970116,
      "loss": 3.0929,
      "theoretical_loss": 4.341915238965026,
      "tokens_seen": 204996608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006552983733686401,
      "loss": 3.3449,
      "theoretical_loss": 4.34174224216294,
      "tokens_seen": 205062144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006552818233402686,
      "loss": 2.966,
      "theoretical_loss": 4.34156931611529,
      "tokens_seen": 205127680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000655265273311897,
      "loss": 3.0641,
      "theoretical_loss": 4.341396460770547,
      "tokens_seen": 205193216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006552487232835256,
      "loss": 2.9224,
      "theoretical_loss": 4.341223676077232,
      "tokens_seen": 205258752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006552321732551541,
      "loss": 3.0892,
      "theoretical_loss": 4.341050961983926,
      "tokens_seen": 205324288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006552156232267826,
      "loss": 2.9695,
      "theoretical_loss": 4.340878318439261,
      "tokens_seen": 205389824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006551990731984112,
      "loss": 2.9769,
      "theoretical_loss": 4.340705745391922,
      "tokens_seen": 205455360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006551825231700397,
      "loss": 2.9707,
      "theoretical_loss": 4.3405332427906504,
      "tokens_seen": 205520896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006551659731416682,
      "loss": 3.1639,
      "theoretical_loss": 4.340360810584238,
      "tokens_seen": 205586432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006551494231132967,
      "loss": 2.8916,
      "theoretical_loss": 4.340188448721532,
      "tokens_seen": 205651968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006551328730849252,
      "loss": 3.3554,
      "theoretical_loss": 4.3400161571514335,
      "tokens_seen": 205717504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006551163230565538,
      "loss": 3.0346,
      "theoretical_loss": 4.339843935822895,
      "tokens_seen": 205783040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006550997730281823,
      "loss": 3.0053,
      "theoretical_loss": 4.339671784684923,
      "tokens_seen": 205848576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006550832229998108,
      "loss": 3.1945,
      "theoretical_loss": 4.339499703686579,
      "tokens_seen": 205914112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006550666729714394,
      "loss": 3.1294,
      "theoretical_loss": 4.339327692776977,
      "tokens_seen": 205979648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006550501229430678,
      "loss": 3.2701,
      "theoretical_loss": 4.339155751905282,
      "tokens_seen": 206045184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006550335729146964,
      "loss": 3.0935,
      "theoretical_loss": 4.338983881020713,
      "tokens_seen": 206110720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006550170228863249,
      "loss": 3.0336,
      "theoretical_loss": 4.338812080072545,
      "tokens_seen": 206176256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006550004728579534,
      "loss": 2.9011,
      "theoretical_loss": 4.338640349010101,
      "tokens_seen": 206241792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000654983922829582,
      "loss": 3.0379,
      "theoretical_loss": 4.3384686877827585,
      "tokens_seen": 206307328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006549673728012105,
      "loss": 3.1135,
      "theoretical_loss": 4.338297096339951,
      "tokens_seen": 206372864
    },
    {
      "epoch": 0.07,
      "objective/train/avg_token_score": 0.02387400157749653,
      "objective/train/avg_weight": 0.9719002842903137,
      "objective/train/docs_used": 128257,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3289713859558105,
      "objective/train/std_weight": 0.09316538274288177,
      "objective/train/theoretical_loss": 4.33812557463116,
      "objective/train/tokens_used": 226898400,
      "theoretical_loss": 4.33812557463116,
      "tokens_seen": 206438400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006549508227728391,
      "loss": 3.3262,
      "theoretical_loss": 4.33812557463116,
      "tokens_seen": 206438400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006549342727444676,
      "loss": 3.1799,
      "theoretical_loss": 4.3379541226059235,
      "tokens_seen": 206503936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000654917722716096,
      "loss": 3.4479,
      "theoretical_loss": 4.337782740213827,
      "tokens_seen": 206569472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006549011726877246,
      "loss": 3.2047,
      "theoretical_loss": 4.337611427404514,
      "tokens_seen": 206635008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006548846226593531,
      "loss": 3.05,
      "theoretical_loss": 4.337440184127679,
      "tokens_seen": 206700544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006548680726309816,
      "loss": 2.9738,
      "theoretical_loss": 4.337269010333065,
      "tokens_seen": 206766080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006548515226026102,
      "loss": 3.0856,
      "theoretical_loss": 4.337097905970471,
      "tokens_seen": 206831616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006548349725742387,
      "loss": 3.1966,
      "theoretical_loss": 4.336926870989748,
      "tokens_seen": 206897152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006548184225458673,
      "loss": 3.0519,
      "theoretical_loss": 4.336755905340797,
      "tokens_seen": 206962688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006548018725174957,
      "loss": 3.2699,
      "theoretical_loss": 4.336585008973573,
      "tokens_seen": 207028224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006547853224891242,
      "loss": 3.0652,
      "theoretical_loss": 4.336414181838082,
      "tokens_seen": 207093760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006547687724607528,
      "loss": 3.1308,
      "theoretical_loss": 4.336243423884382,
      "tokens_seen": 207159296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006547522224323813,
      "loss": 3.0682,
      "theoretical_loss": 4.336072735062583,
      "tokens_seen": 207224832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006547356724040098,
      "loss": 3.308,
      "theoretical_loss": 4.335902115322847,
      "tokens_seen": 207290368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006547191223756384,
      "loss": 3.2018,
      "theoretical_loss": 4.335731564615387,
      "tokens_seen": 207355904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006547025723472669,
      "loss": 3.0456,
      "theoretical_loss": 4.335561082890468,
      "tokens_seen": 207421440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006546860223188955,
      "loss": 3.2261,
      "theoretical_loss": 4.335390670098407,
      "tokens_seen": 207486976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006546694722905239,
      "loss": 3.3293,
      "theoretical_loss": 4.335220326189571,
      "tokens_seen": 207552512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006546529222621524,
      "loss": 3.0089,
      "theoretical_loss": 4.335050051114379,
      "tokens_seen": 207618048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000654636372233781,
      "loss": 3.3354,
      "theoretical_loss": 4.334879844823304,
      "tokens_seen": 207683584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006546198222054095,
      "loss": 3.0223,
      "theoretical_loss": 4.334709707266865,
      "tokens_seen": 207749120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000654603272177038,
      "loss": 2.9571,
      "theoretical_loss": 4.334539638395636,
      "tokens_seen": 207814656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006545867221486666,
      "loss": 3.1248,
      "theoretical_loss": 4.334369638160242,
      "tokens_seen": 207880192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006545701721202951,
      "loss": 3.1681,
      "theoretical_loss": 4.334199706511358,
      "tokens_seen": 207945728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006545536220919236,
      "loss": 3.1017,
      "theoretical_loss": 4.334029843399709,
      "tokens_seen": 208011264
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006545370720635521,
      "loss": 3.09,
      "theoretical_loss": 4.333860048776074,
      "tokens_seen": 208076800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006545205220351806,
      "loss": 3.0479,
      "theoretical_loss": 4.33369032259128,
      "tokens_seen": 208142336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006545039720068092,
      "loss": 3.0377,
      "theoretical_loss": 4.333520664796206,
      "tokens_seen": 208207872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006544874219784377,
      "loss": 3.115,
      "theoretical_loss": 4.33335107534178,
      "tokens_seen": 208273408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006544708719500662,
      "loss": 3.1345,
      "theoretical_loss": 4.333181554178985,
      "tokens_seen": 208338944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006544543219216948,
      "loss": 3.3008,
      "theoretical_loss": 4.3330121012588485,
      "tokens_seen": 208404480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006544377718933232,
      "loss": 3.0089,
      "theoretical_loss": 4.332842716532454,
      "tokens_seen": 208470016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006544212218649518,
      "loss": 3.1356,
      "theoretical_loss": 4.332673399950932,
      "tokens_seen": 208535552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006544046718365803,
      "loss": 2.9473,
      "theoretical_loss": 4.332504151465464,
      "tokens_seen": 208601088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006543881218082088,
      "loss": 3.202,
      "theoretical_loss": 4.332334971027284,
      "tokens_seen": 208666624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006543715717798374,
      "loss": 3.0429,
      "theoretical_loss": 4.332165858587672,
      "tokens_seen": 208732160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006543550217514659,
      "loss": 3.0859,
      "theoretical_loss": 4.331996814097963,
      "tokens_seen": 208797696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006543384717230944,
      "loss": 3.172,
      "theoretical_loss": 4.331827837509538,
      "tokens_seen": 208863232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000654321921694723,
      "loss": 2.8559,
      "theoretical_loss": 4.331658928773831,
      "tokens_seen": 208928768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006543053716663514,
      "loss": 3.1853,
      "theoretical_loss": 4.331490087842324,
      "tokens_seen": 208994304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00065428882163798,
      "loss": 3.2675,
      "theoretical_loss": 4.33132131466655,
      "tokens_seen": 209059840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006542722716096084,
      "loss": 3.122,
      "theoretical_loss": 4.3311526091980905,
      "tokens_seen": 209125376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000654255721581237,
      "loss": 3.037,
      "theoretical_loss": 4.330983971388578,
      "tokens_seen": 209190912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006542391715528655,
      "loss": 2.8079,
      "theoretical_loss": 4.330815401189695,
      "tokens_seen": 209256448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000654222621524494,
      "loss": 3.1733,
      "theoretical_loss": 4.330646898553173,
      "tokens_seen": 209321984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006542060714961226,
      "loss": 3.1375,
      "theoretical_loss": 4.330478463430792,
      "tokens_seen": 209387520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000654189521467751,
      "loss": 3.188,
      "theoretical_loss": 4.330310095774383,
      "tokens_seen": 209453056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006541729714393795,
      "loss": 2.9683,
      "theoretical_loss": 4.330141795535828,
      "tokens_seen": 209518592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006541564214110081,
      "loss": 3.2492,
      "theoretical_loss": 4.329973562667053,
      "tokens_seen": 209584128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006541398713826366,
      "loss": 3.0438,
      "theoretical_loss": 4.3298053971200385,
      "tokens_seen": 209649664
    },
    {
      "epoch": 0.07,
      "objective/train/avg_token_score": 0.01519742887467146,
      "objective/train/avg_weight": 0.982124388217926,
      "objective/train/docs_used": 130307,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.203678846359253,
      "objective/train/std_weight": 0.09015699476003647,
      "objective/train/theoretical_loss": 4.329637298846812,
      "objective/train/tokens_used": 230175200,
      "theoretical_loss": 4.329637298846812,
      "tokens_seen": 209715200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006541233213542652,
      "loss": 3.0117,
      "theoretical_loss": 4.329637298846812,
      "tokens_seen": 209715200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006541067713258937,
      "loss": 3.1794,
      "theoretical_loss": 4.329469267799451,
      "tokens_seen": 209780736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006540902212975222,
      "loss": 3.0194,
      "theoretical_loss": 4.32930130393008,
      "tokens_seen": 209846272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006540736712691507,
      "loss": 3.2209,
      "theoretical_loss": 4.329133407190876,
      "tokens_seen": 209911808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0006540571212407792,
      "loss": 2.9174,
      "theoretical_loss": 4.3289655775340625,
      "tokens_seen": 209977344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006540405712124077,
      "loss": 3.149,
      "theoretical_loss": 4.328797814911912,
      "tokens_seen": 210042880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006540240211840363,
      "loss": 3.1598,
      "theoretical_loss": 4.328630119276747,
      "tokens_seen": 210108416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006540074711556648,
      "loss": 2.7908,
      "theoretical_loss": 4.328462490580938,
      "tokens_seen": 210173952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006539909211272934,
      "loss": 3.2775,
      "theoretical_loss": 4.328294928776903,
      "tokens_seen": 210239488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006539743710989219,
      "loss": 2.9733,
      "theoretical_loss": 4.328127433817112,
      "tokens_seen": 210305024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006539578210705503,
      "loss": 3.173,
      "theoretical_loss": 4.327960005654081,
      "tokens_seen": 210370560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006539412710421789,
      "loss": 3.28,
      "theoretical_loss": 4.327792644240374,
      "tokens_seen": 210436096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006539247210138074,
      "loss": 3.2135,
      "theoretical_loss": 4.327625349528605,
      "tokens_seen": 210501632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006539081709854359,
      "loss": 3.0227,
      "theoretical_loss": 4.327458121471436,
      "tokens_seen": 210567168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006538916209570645,
      "loss": 3.1773,
      "theoretical_loss": 4.3272909600215765,
      "tokens_seen": 210632704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000653875070928693,
      "loss": 3.1167,
      "theoretical_loss": 4.327123865131786,
      "tokens_seen": 210698240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006538585209003216,
      "loss": 3.2019,
      "theoretical_loss": 4.326956836754871,
      "tokens_seen": 210763776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006538419708719501,
      "loss": 2.9547,
      "theoretical_loss": 4.326789874843685,
      "tokens_seen": 210829312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006538254208435785,
      "loss": 2.9469,
      "theoretical_loss": 4.326622979351132,
      "tokens_seen": 210894848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006538088708152071,
      "loss": 2.9289,
      "theoretical_loss": 4.326456150230163,
      "tokens_seen": 210960384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006537923207868356,
      "loss": 2.8499,
      "theoretical_loss": 4.326289387433776,
      "tokens_seen": 211025920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006537757707584641,
      "loss": 3.0456,
      "theoretical_loss": 4.326122690915017,
      "tokens_seen": 211091456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006537592207300927,
      "loss": 3.1673,
      "theoretical_loss": 4.325956060626982,
      "tokens_seen": 211156992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006537426707017212,
      "loss": 3.1053,
      "theoretical_loss": 4.325789496522812,
      "tokens_seen": 211222528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006537261206733498,
      "loss": 3.1105,
      "theoretical_loss": 4.325622998555697,
      "tokens_seen": 211288064
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006537095706449782,
      "loss": 2.8157,
      "theoretical_loss": 4.3254565666788745,
      "tokens_seen": 211353600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006536930206166067,
      "loss": 2.9349,
      "theoretical_loss": 4.325290200845629,
      "tokens_seen": 211419136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006536764705882353,
      "loss": 3.0843,
      "theoretical_loss": 4.3251239010092934,
      "tokens_seen": 211484672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006536599205598638,
      "loss": 3.0224,
      "theoretical_loss": 4.324957667123249,
      "tokens_seen": 211550208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006536433705314923,
      "loss": 3.2193,
      "theoretical_loss": 4.32479149914092,
      "tokens_seen": 211615744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006536268205031209,
      "loss": 3.1819,
      "theoretical_loss": 4.324625397015783,
      "tokens_seen": 211681280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006536102704747494,
      "loss": 3.0429,
      "theoretical_loss": 4.3244593607013595,
      "tokens_seen": 211746816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000653593720446378,
      "loss": 3.0506,
      "theoretical_loss": 4.324293390151218,
      "tokens_seen": 211812352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006535771704180064,
      "loss": 3.0439,
      "theoretical_loss": 4.324127485318975,
      "tokens_seen": 211877888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006535606203896349,
      "loss": 3.012,
      "theoretical_loss": 4.323961646158294,
      "tokens_seen": 211943424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006535440703612635,
      "loss": 3.307,
      "theoretical_loss": 4.323795872622884,
      "tokens_seen": 212008960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000653527520332892,
      "loss": 3.2534,
      "theoretical_loss": 4.323630164666502,
      "tokens_seen": 212074496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006535109703045205,
      "loss": 2.8642,
      "theoretical_loss": 4.323464522242954,
      "tokens_seen": 212140032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006534944202761491,
      "loss": 3.3584,
      "theoretical_loss": 4.323298945306089,
      "tokens_seen": 212205568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006534778702477776,
      "loss": 3.1101,
      "theoretical_loss": 4.3231334338098035,
      "tokens_seen": 212271104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006534613202194062,
      "loss": 3.0396,
      "theoretical_loss": 4.322967987708043,
      "tokens_seen": 212336640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006534447701910346,
      "loss": 3.111,
      "theoretical_loss": 4.322802606954798,
      "tokens_seen": 212402176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006534282201626631,
      "loss": 3.1864,
      "theoretical_loss": 4.322637291504106,
      "tokens_seen": 212467712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006534116701342917,
      "loss": 2.9817,
      "theoretical_loss": 4.32247204131005,
      "tokens_seen": 212533248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006533951201059202,
      "loss": 2.8881,
      "theoretical_loss": 4.322306856326761,
      "tokens_seen": 212598784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006533785700775488,
      "loss": 3.1577,
      "theoretical_loss": 4.322141736508415,
      "tokens_seen": 212664320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006533620200491773,
      "loss": 3.1716,
      "theoretical_loss": 4.321976681809236,
      "tokens_seen": 212729856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006533454700208058,
      "loss": 2.9229,
      "theoretical_loss": 4.321811692183491,
      "tokens_seen": 212795392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006533289199924343,
      "loss": 3.0771,
      "theoretical_loss": 4.321646767585497,
      "tokens_seen": 212860928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006533123699640628,
      "loss": 3.0866,
      "theoretical_loss": 4.3214819079696145,
      "tokens_seen": 212926464
    },
    {
      "epoch": 0.08,
      "objective/train/avg_token_score": 0.01635868474841118,
      "objective/train/avg_weight": 0.9807456135749817,
      "objective/train/docs_used": 132385,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.277647018432617,
      "objective/train/std_weight": 0.07439122349023819,
      "objective/train/theoretical_loss": 4.321317113290252,
      "objective/train/tokens_used": 233452000,
      "theoretical_loss": 4.321317113290252,
      "tokens_seen": 212992000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006532958199356913,
      "loss": 3.0729,
      "theoretical_loss": 4.321317113290252,
      "tokens_seen": 212992000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006532792699073198,
      "loss": 3.1813,
      "theoretical_loss": 4.321152383501863,
      "tokens_seen": 213057536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006532627198789483,
      "loss": 3.1236,
      "theoretical_loss": 4.320987718558945,
      "tokens_seen": 213123072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006532461698505769,
      "loss": 2.846,
      "theoretical_loss": 4.320823118416046,
      "tokens_seen": 213188608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006532296198222054,
      "loss": 2.9407,
      "theoretical_loss": 4.320658583027755,
      "tokens_seen": 213254144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006532130697938339,
      "loss": 3.0901,
      "theoretical_loss": 4.32049411234871,
      "tokens_seen": 213319680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006531965197654624,
      "loss": 3.0501,
      "theoretical_loss": 4.3203297063335935,
      "tokens_seen": 213385216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006531799697370909,
      "loss": 3.2029,
      "theoretical_loss": 4.320165364937134,
      "tokens_seen": 213450752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006531634197087195,
      "loss": 3.1797,
      "theoretical_loss": 4.320001088114105,
      "tokens_seen": 213516288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000653146869680348,
      "loss": 3.0599,
      "theoretical_loss": 4.319836875819325,
      "tokens_seen": 213581824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006531303196519765,
      "loss": 3.2531,
      "theoretical_loss": 4.31967272800766,
      "tokens_seen": 213647360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006531137696236051,
      "loss": 2.7922,
      "theoretical_loss": 4.319508644634021,
      "tokens_seen": 213712896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006530972195952335,
      "loss": 2.9221,
      "theoretical_loss": 4.319344625653361,
      "tokens_seen": 213778432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006530806695668621,
      "loss": 2.9119,
      "theoretical_loss": 4.319180671020684,
      "tokens_seen": 213843968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006530641195384906,
      "loss": 2.8365,
      "theoretical_loss": 4.319016780691033,
      "tokens_seen": 213909504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006530475695101191,
      "loss": 3.2291,
      "theoretical_loss": 4.318852954619501,
      "tokens_seen": 213975040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006530310194817477,
      "loss": 3.1251,
      "theoretical_loss": 4.318689192761225,
      "tokens_seen": 214040576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006530144694533762,
      "loss": 3.0535,
      "theoretical_loss": 4.318525495071385,
      "tokens_seen": 214106112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006529979194250047,
      "loss": 3.1078,
      "theoretical_loss": 4.318361861505207,
      "tokens_seen": 214171648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006529813693966333,
      "loss": 3.0948,
      "theoretical_loss": 4.318198292017964,
      "tokens_seen": 214237184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006529648193682617,
      "loss": 2.8579,
      "theoretical_loss": 4.318034786564971,
      "tokens_seen": 214302720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006529482693398903,
      "loss": 2.9763,
      "theoretical_loss": 4.31787134510159,
      "tokens_seen": 214368256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006529317193115188,
      "loss": 3.1288,
      "theoretical_loss": 4.3177079675832255,
      "tokens_seen": 214433792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006529151692831473,
      "loss": 3.1702,
      "theoretical_loss": 4.317544653965329,
      "tokens_seen": 214499328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006528986192547759,
      "loss": 3.1652,
      "theoretical_loss": 4.3173814042033944,
      "tokens_seen": 214564864
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006528820692264044,
      "loss": 3.0733,
      "theoretical_loss": 4.317218218252963,
      "tokens_seen": 214630400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006528655191980329,
      "loss": 2.9489,
      "theoretical_loss": 4.317055096069618,
      "tokens_seen": 214695936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006528489691696614,
      "loss": 3.2362,
      "theoretical_loss": 4.316892037608987,
      "tokens_seen": 214761472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006528324191412899,
      "loss": 3.0143,
      "theoretical_loss": 4.316729042826745,
      "tokens_seen": 214827008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006528158691129185,
      "loss": 2.9998,
      "theoretical_loss": 4.316566111678609,
      "tokens_seen": 214892544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000652799319084547,
      "loss": 3.13,
      "theoretical_loss": 4.316403244120339,
      "tokens_seen": 214958080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006527827690561755,
      "loss": 3.0266,
      "theoretical_loss": 4.3162404401077445,
      "tokens_seen": 215023616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006527662190278041,
      "loss": 2.8822,
      "theoretical_loss": 4.316077699596671,
      "tokens_seen": 215089152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006527496689994326,
      "loss": 3.0523,
      "theoretical_loss": 4.315915022543016,
      "tokens_seen": 215154688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000652733118971061,
      "loss": 3.0557,
      "theoretical_loss": 4.315752408902716,
      "tokens_seen": 215220224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006527165689426896,
      "loss": 2.7892,
      "theoretical_loss": 4.315589858631755,
      "tokens_seen": 215285760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006527000189143181,
      "loss": 3.1655,
      "theoretical_loss": 4.315427371686157,
      "tokens_seen": 215351296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006526834688859467,
      "loss": 2.9695,
      "theoretical_loss": 4.315264948021994,
      "tokens_seen": 215416832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006526669188575752,
      "loss": 3.2905,
      "theoretical_loss": 4.315102587595379,
      "tokens_seen": 215482368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006526503688292037,
      "loss": 3.213,
      "theoretical_loss": 4.31494029036247,
      "tokens_seen": 215547904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006526338188008323,
      "loss": 3.1179,
      "theoretical_loss": 4.314778056279468,
      "tokens_seen": 215613440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006526172687724608,
      "loss": 3.1354,
      "theoretical_loss": 4.314615885302619,
      "tokens_seen": 215678976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006526007187440892,
      "loss": 3.0831,
      "theoretical_loss": 4.314453777388209,
      "tokens_seen": 215744512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006525841687157178,
      "loss": 3.0703,
      "theoretical_loss": 4.314291732492573,
      "tokens_seen": 215810048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006525676186873463,
      "loss": 3.1168,
      "theoretical_loss": 4.314129750572087,
      "tokens_seen": 215875584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006525510686589749,
      "loss": 3.055,
      "theoretical_loss": 4.3139678315831675,
      "tokens_seen": 215941120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006525345186306034,
      "loss": 2.8657,
      "theoretical_loss": 4.313805975482278,
      "tokens_seen": 216006656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006525179686022319,
      "loss": 2.8015,
      "theoretical_loss": 4.313644182225926,
      "tokens_seen": 216072192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006525014185738605,
      "loss": 3.188,
      "theoretical_loss": 4.313482451770659,
      "tokens_seen": 216137728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006524848685454889,
      "loss": 2.7971,
      "theoretical_loss": 4.313320784073069,
      "tokens_seen": 216203264
    },
    {
      "epoch": 0.08,
      "objective/train/avg_token_score": 0.016855787485837936,
      "objective/train/avg_weight": 0.9801643490791321,
      "objective/train/docs_used": 133646,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3810300827026367,
      "objective/train/std_weight": 0.07968246936798096,
      "objective/train/theoretical_loss": 4.3131591790897925,
      "objective/train/tokens_used": 236728800,
      "theoretical_loss": 4.3131591790897925,
      "tokens_seen": 216268800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006524683185171174,
      "loss": 3.287,
      "theoretical_loss": 4.3131591790897925,
      "tokens_seen": 216268800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000652451768488746,
      "loss": 2.9808,
      "theoretical_loss": 4.3129976367775065,
      "tokens_seen": 216334336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006524352184603745,
      "loss": 3.0862,
      "theoretical_loss": 4.312836157092934,
      "tokens_seen": 216399872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006524186684320031,
      "loss": 2.8783,
      "theoretical_loss": 4.312674739992839,
      "tokens_seen": 216465408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006524021184036316,
      "loss": 3.1073,
      "theoretical_loss": 4.31251338543403,
      "tokens_seen": 216530944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006523855683752601,
      "loss": 2.9212,
      "theoretical_loss": 4.312352093373354,
      "tokens_seen": 216596480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006523690183468887,
      "loss": 3.0526,
      "theoretical_loss": 4.312190863767708,
      "tokens_seen": 216662016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006523524683185171,
      "loss": 3.0019,
      "theoretical_loss": 4.312029696574027,
      "tokens_seen": 216727552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006523359182901456,
      "loss": 3.2629,
      "theoretical_loss": 4.311868591749287,
      "tokens_seen": 216793088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006523193682617742,
      "loss": 3.2003,
      "theoretical_loss": 4.311707549250514,
      "tokens_seen": 216858624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006523028182334027,
      "loss": 3.2505,
      "theoretical_loss": 4.311546569034767,
      "tokens_seen": 216924160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006522862682050312,
      "loss": 3.1197,
      "theoretical_loss": 4.311385651059155,
      "tokens_seen": 216989696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006522697181766597,
      "loss": 3.1578,
      "theoretical_loss": 4.311224795280825,
      "tokens_seen": 217055232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006522531681482883,
      "loss": 3.2996,
      "theoretical_loss": 4.3110640016569715,
      "tokens_seen": 217120768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006522366181199167,
      "loss": 3.3966,
      "theoretical_loss": 4.310903270144825,
      "tokens_seen": 217186304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006522200680915452,
      "loss": 3.1677,
      "theoretical_loss": 4.310742600701664,
      "tokens_seen": 217251840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006522035180631738,
      "loss": 3.1228,
      "theoretical_loss": 4.310581993284805,
      "tokens_seen": 217317376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006521869680348023,
      "loss": 2.9205,
      "theoretical_loss": 4.310421447851609,
      "tokens_seen": 217382912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006521704180064309,
      "loss": 3.1722,
      "theoretical_loss": 4.310260964359479,
      "tokens_seen": 217448448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006521538679780594,
      "loss": 3.0937,
      "theoretical_loss": 4.310100542765858,
      "tokens_seen": 217513984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006521373179496879,
      "loss": 3.3014,
      "theoretical_loss": 4.309940183028236,
      "tokens_seen": 217579520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006521207679213164,
      "loss": 3.0865,
      "theoretical_loss": 4.309779885104139,
      "tokens_seen": 217645056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006521042178929449,
      "loss": 3.3763,
      "theoretical_loss": 4.309619648951139,
      "tokens_seen": 217710592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006520876678645734,
      "loss": 3.1201,
      "theoretical_loss": 4.3094594745268475,
      "tokens_seen": 217776128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000652071117836202,
      "loss": 3.1628,
      "theoretical_loss": 4.30929936178892,
      "tokens_seen": 217841664
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006520545678078305,
      "loss": 3.0613,
      "theoretical_loss": 4.309139310695053,
      "tokens_seen": 217907200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006520380177794591,
      "loss": 2.9694,
      "theoretical_loss": 4.308979321202983,
      "tokens_seen": 217972736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006520214677510876,
      "loss": 3.3324,
      "theoretical_loss": 4.308819393270491,
      "tokens_seen": 218038272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000652004917722716,
      "loss": 2.9347,
      "theoretical_loss": 4.308659526855396,
      "tokens_seen": 218103808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006519883676943446,
      "loss": 3.1118,
      "theoretical_loss": 4.308499721915563,
      "tokens_seen": 218169344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006519718176659731,
      "loss": 3.181,
      "theoretical_loss": 4.308339978408897,
      "tokens_seen": 218234880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006519552676376016,
      "loss": 3.353,
      "theoretical_loss": 4.308180296293341,
      "tokens_seen": 218300416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006519387176092302,
      "loss": 2.9758,
      "theoretical_loss": 4.308020675526883,
      "tokens_seen": 218365952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006519221675808587,
      "loss": 3.1868,
      "theoretical_loss": 4.307861116067554,
      "tokens_seen": 218431488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006519056175524873,
      "loss": 2.914,
      "theoretical_loss": 4.30770161787342,
      "tokens_seen": 218497024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006518890675241158,
      "loss": 2.9732,
      "theoretical_loss": 4.307542180902594,
      "tokens_seen": 218562560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006518725174957442,
      "loss": 3.1587,
      "theoretical_loss": 4.307382805113228,
      "tokens_seen": 218628096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006518559674673728,
      "loss": 2.9646,
      "theoretical_loss": 4.307223490463516,
      "tokens_seen": 218693632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006518394174390013,
      "loss": 3.1428,
      "theoretical_loss": 4.307064236911692,
      "tokens_seen": 218759168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006518228674106298,
      "loss": 2.9025,
      "theoretical_loss": 4.30690504441603,
      "tokens_seen": 218824704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006518063173822584,
      "loss": 3.1064,
      "theoretical_loss": 4.306745912934849,
      "tokens_seen": 218890240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006517897673538869,
      "loss": 2.9872,
      "theoretical_loss": 4.306586842426504,
      "tokens_seen": 218955776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006517732173255155,
      "loss": 3.3532,
      "theoretical_loss": 4.306427832849394,
      "tokens_seen": 219021312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006517566672971439,
      "loss": 2.9788,
      "theoretical_loss": 4.306268884161959,
      "tokens_seen": 219086848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006517401172687724,
      "loss": 3.1219,
      "theoretical_loss": 4.306109996322679,
      "tokens_seen": 219152384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000651723567240401,
      "loss": 3.0539,
      "theoretical_loss": 4.305951169290073,
      "tokens_seen": 219217920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006517070172120295,
      "loss": 3.0702,
      "theoretical_loss": 4.305792403022703,
      "tokens_seen": 219283456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000651690467183658,
      "loss": 3.0183,
      "theoretical_loss": 4.305633697479171,
      "tokens_seen": 219348992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006516739171552866,
      "loss": 2.8971,
      "theoretical_loss": 4.305475052618119,
      "tokens_seen": 219414528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006516573671269151,
      "loss": 3.106,
      "theoretical_loss": 4.30531646839823,
      "tokens_seen": 219480064
    },
    {
      "epoch": 0.08,
      "objective/train/avg_token_score": 0.014708828181028366,
      "objective/train/avg_weight": 0.9827037453651428,
      "objective/train/docs_used": 136163,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.6228771209716797,
      "objective/train/std_weight": 0.10528632253408432,
      "objective/train/theoretical_loss": 4.305157944778228,
      "objective/train/tokens_used": 240005600,
      "theoretical_loss": 4.305157944778228,
      "tokens_seen": 219545600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006516408170985437,
      "loss": 3.2221,
      "theoretical_loss": 4.305157944778228,
      "tokens_seen": 219545600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006516242670701721,
      "loss": 3.2904,
      "theoretical_loss": 4.304999481716876,
      "tokens_seen": 219611136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006516077170418006,
      "loss": 2.8516,
      "theoretical_loss": 4.304841079172979,
      "tokens_seen": 219676672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006515911670134292,
      "loss": 3.2637,
      "theoretical_loss": 4.30468273710538,
      "tokens_seen": 219742208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006515746169850577,
      "loss": 3.2405,
      "theoretical_loss": 4.304524455472965,
      "tokens_seen": 219807744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006515580669566862,
      "loss": 3.0847,
      "theoretical_loss": 4.304366234234659,
      "tokens_seen": 219873280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006515415169283148,
      "loss": 3.1945,
      "theoretical_loss": 4.304208073349426,
      "tokens_seen": 219938816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006515249668999433,
      "loss": 3.0806,
      "theoretical_loss": 4.304049972776271,
      "tokens_seen": 220004352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006515084168715718,
      "loss": 3.2489,
      "theoretical_loss": 4.30389193247424,
      "tokens_seen": 220069888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006514918668432003,
      "loss": 2.9981,
      "theoretical_loss": 4.303733952402419,
      "tokens_seen": 220135424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006514753168148288,
      "loss": 3.2038,
      "theoretical_loss": 4.303576032519931,
      "tokens_seen": 220200960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006514587667864574,
      "loss": 2.888,
      "theoretical_loss": 4.303418172785943,
      "tokens_seen": 220266496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006514422167580859,
      "loss": 3.0898,
      "theoretical_loss": 4.303260373159659,
      "tokens_seen": 220332032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006514256667297144,
      "loss": 3.1739,
      "theoretical_loss": 4.303102633600322,
      "tokens_seen": 220397568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000651409116701343,
      "loss": 3.1382,
      "theoretical_loss": 4.30294495406722,
      "tokens_seen": 220463104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006513925666729714,
      "loss": 3.1812,
      "theoretical_loss": 4.3027873345196745,
      "tokens_seen": 220528640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006513760166446,
      "loss": 3.2719,
      "theoretical_loss": 4.302629774917049,
      "tokens_seen": 220594176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006513594666162285,
      "loss": 3.0717,
      "theoretical_loss": 4.302472275218748,
      "tokens_seen": 220659712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000651342916587857,
      "loss": 3.2357,
      "theoretical_loss": 4.302314835384214,
      "tokens_seen": 220725248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006513263665594855,
      "loss": 3.0582,
      "theoretical_loss": 4.30215745537293,
      "tokens_seen": 220790784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000651309816531114,
      "loss": 3.2437,
      "theoretical_loss": 4.302000135144416,
      "tokens_seen": 220856320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006512932665027426,
      "loss": 3.0499,
      "theoretical_loss": 4.301842874658235,
      "tokens_seen": 220921856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000651276716474371,
      "loss": 3.1747,
      "theoretical_loss": 4.301685673873987,
      "tokens_seen": 220987392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006512601664459995,
      "loss": 3.1099,
      "theoretical_loss": 4.301528532751312,
      "tokens_seen": 221052928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006512436164176281,
      "loss": 3.0289,
      "theoretical_loss": 4.301371451249888,
      "tokens_seen": 221118464
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006512270663892566,
      "loss": 3.0979,
      "theoretical_loss": 4.301214429329433,
      "tokens_seen": 221184000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006512105163608852,
      "loss": 3.0983,
      "theoretical_loss": 4.301057466949707,
      "tokens_seen": 221249536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006511939663325137,
      "loss": 3.1877,
      "theoretical_loss": 4.300900564070504,
      "tokens_seen": 221315072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006511774163041422,
      "loss": 3.0307,
      "theoretical_loss": 4.30074372065166,
      "tokens_seen": 221380608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006511608662757708,
      "loss": 3.1631,
      "theoretical_loss": 4.300586936653049,
      "tokens_seen": 221446144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006511443162473992,
      "loss": 3.2184,
      "theoretical_loss": 4.300430212034587,
      "tokens_seen": 221511680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006511277662190277,
      "loss": 3.2805,
      "theoretical_loss": 4.300273546756223,
      "tokens_seen": 221577216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006511112161906563,
      "loss": 3.2016,
      "theoretical_loss": 4.300116940777951,
      "tokens_seen": 221642752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006510946661622848,
      "loss": 3.1833,
      "theoretical_loss": 4.299960394059799,
      "tokens_seen": 221708288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006510781161339134,
      "loss": 3.1153,
      "theoretical_loss": 4.299803906561835,
      "tokens_seen": 221773824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006510615661055419,
      "loss": 3.1624,
      "theoretical_loss": 4.29964747824417,
      "tokens_seen": 221839360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006510450160771704,
      "loss": 3.0111,
      "theoretical_loss": 4.299491109066947,
      "tokens_seen": 221904896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000651028466048799,
      "loss": 3.0237,
      "theoretical_loss": 4.299334798990351,
      "tokens_seen": 221970432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006510119160204274,
      "loss": 2.9675,
      "theoretical_loss": 4.2991785479746065,
      "tokens_seen": 222035968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000650995365992056,
      "loss": 3.1999,
      "theoretical_loss": 4.299022355979974,
      "tokens_seen": 222101504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006509788159636845,
      "loss": 3.0093,
      "theoretical_loss": 4.298866222966755,
      "tokens_seen": 222167040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000650962265935313,
      "loss": 3.1553,
      "theoretical_loss": 4.298710148895286,
      "tokens_seen": 222232576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006509457159069416,
      "loss": 3.0062,
      "theoretical_loss": 4.298554133725946,
      "tokens_seen": 222298112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006509291658785701,
      "loss": 3.3132,
      "theoretical_loss": 4.298398177419149,
      "tokens_seen": 222363648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006509126158501985,
      "loss": 3.1581,
      "theoretical_loss": 4.298242279935349,
      "tokens_seen": 222429184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006508960658218271,
      "loss": 3.1596,
      "theoretical_loss": 4.2980864412350375,
      "tokens_seen": 222494720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006508795157934556,
      "loss": 3.0821,
      "theoretical_loss": 4.297930661278745,
      "tokens_seen": 222560256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006508629657650842,
      "loss": 3.0903,
      "theoretical_loss": 4.297774940027038,
      "tokens_seen": 222625792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006508464157367127,
      "loss": 3.2958,
      "theoretical_loss": 4.297619277440523,
      "tokens_seen": 222691328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006508298657083412,
      "loss": 3.0514,
      "theoretical_loss": 4.297463673479846,
      "tokens_seen": 222756864
    },
    {
      "epoch": 0.08,
      "objective/train/avg_token_score": 0.016721993684768677,
      "objective/train/avg_weight": 0.9803293943405151,
      "objective/train/docs_used": 137875,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9991931915283203,
      "objective/train/std_weight": 0.0770498514175415,
      "objective/train/theoretical_loss": 4.297308128105687,
      "objective/train/tokens_used": 243282400,
      "theoretical_loss": 4.297308128105687,
      "tokens_seen": 222822400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006508133156799698,
      "loss": 2.9874,
      "theoretical_loss": 4.297308128105687,
      "tokens_seen": 222822400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006507967656515983,
      "loss": 3.0484,
      "theoretical_loss": 4.297152641278767,
      "tokens_seen": 222887936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006507802156232267,
      "loss": 3.1966,
      "theoretical_loss": 4.296997212959842,
      "tokens_seen": 222953472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006507636655948553,
      "loss": 2.9705,
      "theoretical_loss": 4.296841843109711,
      "tokens_seen": 223019008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006507471155664838,
      "loss": 3.1056,
      "theoretical_loss": 4.296686531689204,
      "tokens_seen": 223084544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006507305655381124,
      "loss": 3.1422,
      "theoretical_loss": 4.296531278659193,
      "tokens_seen": 223150080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006507140155097409,
      "loss": 3.4331,
      "theoretical_loss": 4.296376083980589,
      "tokens_seen": 223215616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006506974654813694,
      "loss": 3.1041,
      "theoretical_loss": 4.296220947614337,
      "tokens_seen": 223281152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000650680915452998,
      "loss": 3.0165,
      "theoretical_loss": 4.296065869521421,
      "tokens_seen": 223346688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006506643654246264,
      "loss": 3.1887,
      "theoretical_loss": 4.295910849662862,
      "tokens_seen": 223412224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006506478153962549,
      "loss": 3.1406,
      "theoretical_loss": 4.2957558879997215,
      "tokens_seen": 223477760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006506312653678835,
      "loss": 3.0783,
      "theoretical_loss": 4.295600984493093,
      "tokens_seen": 223543296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000650614715339512,
      "loss": 3.0413,
      "theoretical_loss": 4.295446139104112,
      "tokens_seen": 223608832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006505981653111406,
      "loss": 3.1615,
      "theoretical_loss": 4.295291351793951,
      "tokens_seen": 223674368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006505816152827691,
      "loss": 3.0256,
      "theoretical_loss": 4.295136622523817,
      "tokens_seen": 223739904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006505650652543976,
      "loss": 3.4189,
      "theoretical_loss": 4.294981951254956,
      "tokens_seen": 223805440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006505485152260262,
      "loss": 2.9753,
      "theoretical_loss": 4.294827337948651,
      "tokens_seen": 223870976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006505319651976546,
      "loss": 2.9203,
      "theoretical_loss": 4.294672782566224,
      "tokens_seen": 223936512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006505154151692831,
      "loss": 3.0538,
      "theoretical_loss": 4.29451828506903,
      "tokens_seen": 224002048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006504988651409117,
      "loss": 3.0477,
      "theoretical_loss": 4.294363845418465,
      "tokens_seen": 224067584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006504823151125402,
      "loss": 3.0337,
      "theoretical_loss": 4.29420946357596,
      "tokens_seen": 224133120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006504657650841688,
      "loss": 3.0791,
      "theoretical_loss": 4.294055139502985,
      "tokens_seen": 224198656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006504492150557973,
      "loss": 2.9725,
      "theoretical_loss": 4.293900873161043,
      "tokens_seen": 224264192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006504326650274258,
      "loss": 3.0228,
      "theoretical_loss": 4.293746664511678,
      "tokens_seen": 224329728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006504161149990543,
      "loss": 3.1276,
      "theoretical_loss": 4.293592513516469,
      "tokens_seen": 224395264
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006503995649706828,
      "loss": 3.1911,
      "theoretical_loss": 4.293438420137031,
      "tokens_seen": 224460800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006503830149423113,
      "loss": 3.4214,
      "theoretical_loss": 4.293284384335017,
      "tokens_seen": 224526336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006503664649139399,
      "loss": 3.1263,
      "theoretical_loss": 4.293130406072118,
      "tokens_seen": 224591872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006503499148855683,
      "loss": 3.0718,
      "theoretical_loss": 4.292976485310057,
      "tokens_seen": 224657408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006503333648571969,
      "loss": 3.1401,
      "theoretical_loss": 4.2928226220106005,
      "tokens_seen": 224722944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006503168148288254,
      "loss": 3.0657,
      "theoretical_loss": 4.292668816135545,
      "tokens_seen": 224788480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000650300264800454,
      "loss": 3.2816,
      "theoretical_loss": 4.292515067646727,
      "tokens_seen": 224854016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006502837147720824,
      "loss": 2.9734,
      "theoretical_loss": 4.29236137650602,
      "tokens_seen": 224919552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006502671647437109,
      "loss": 3.115,
      "theoretical_loss": 4.2922077426753305,
      "tokens_seen": 224985088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006502506147153395,
      "loss": 3.1555,
      "theoretical_loss": 4.292054166116605,
      "tokens_seen": 225050624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000650234064686968,
      "loss": 3.2049,
      "theoretical_loss": 4.291900646791825,
      "tokens_seen": 225116160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006502175146585965,
      "loss": 3.1986,
      "theoretical_loss": 4.2917471846630075,
      "tokens_seen": 225181696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006502009646302251,
      "loss": 3.0152,
      "theoretical_loss": 4.291593779692207,
      "tokens_seen": 225247232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006501844146018535,
      "loss": 3.0025,
      "theoretical_loss": 4.291440431841513,
      "tokens_seen": 225312768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006501678645734821,
      "loss": 3.0182,
      "theoretical_loss": 4.291287141073053,
      "tokens_seen": 225378304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006501513145451106,
      "loss": 3.0317,
      "theoretical_loss": 4.291133907348989,
      "tokens_seen": 225443840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006501347645167391,
      "loss": 3.0934,
      "theoretical_loss": 4.29098073063152,
      "tokens_seen": 225509376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006501182144883677,
      "loss": 2.919,
      "theoretical_loss": 4.29082761088288,
      "tokens_seen": 225574912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006501016644599962,
      "loss": 3.0118,
      "theoretical_loss": 4.290674548065338,
      "tokens_seen": 225640448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006500851144316247,
      "loss": 3.1354,
      "theoretical_loss": 4.290521542141203,
      "tokens_seen": 225705984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006500685644032533,
      "loss": 3.0375,
      "theoretical_loss": 4.290368593072817,
      "tokens_seen": 225771520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006500520143748817,
      "loss": 3.2563,
      "theoretical_loss": 4.290215700822556,
      "tokens_seen": 225837056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006500354643465103,
      "loss": 3.1514,
      "theoretical_loss": 4.290062865352837,
      "tokens_seen": 225902592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006500189143181388,
      "loss": 2.9538,
      "theoretical_loss": 4.289910086626108,
      "tokens_seen": 225968128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006500023642897673,
      "loss": 3.1257,
      "theoretical_loss": 4.289757364604855,
      "tokens_seen": 226033664
    },
    {
      "epoch": 0.08,
      "objective/train/avg_token_score": 0.009182633832097054,
      "objective/train/avg_weight": 0.9891905188560486,
      "objective/train/docs_used": 139547,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3524973392486572,
      "objective/train/std_weight": 0.06434949487447739,
      "objective/train/theoretical_loss": 4.2896046992515995,
      "objective/train/tokens_used": 246559200,
      "theoretical_loss": 4.2896046992515995,
      "tokens_seen": 226099200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006499858142613959,
      "loss": 3.3803,
      "theoretical_loss": 4.2896046992515995,
      "tokens_seen": 226099200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006499692642330244,
      "loss": 3.0914,
      "theoretical_loss": 4.289452090528897,
      "tokens_seen": 226164736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006499527142046529,
      "loss": 3.1281,
      "theoretical_loss": 4.289299538399341,
      "tokens_seen": 226230272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006499361641762814,
      "loss": 3.0291,
      "theoretical_loss": 4.28914704282556,
      "tokens_seen": 226295808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006499196141479099,
      "loss": 3.0392,
      "theoretical_loss": 4.288994603770215,
      "tokens_seen": 226361344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006499030641195385,
      "loss": 3.3667,
      "theoretical_loss": 4.288842221196007,
      "tokens_seen": 226426880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000649886514091167,
      "loss": 3.4251,
      "theoretical_loss": 4.28868989506567,
      "tokens_seen": 226492416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006498699640627955,
      "loss": 3.2872,
      "theoretical_loss": 4.288537625341974,
      "tokens_seen": 226557952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006498534140344241,
      "loss": 3.1393,
      "theoretical_loss": 4.288385411987722,
      "tokens_seen": 226623488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006498368640060526,
      "loss": 3.0528,
      "theoretical_loss": 4.288233254965755,
      "tokens_seen": 226689024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000649820313977681,
      "loss": 3.0929,
      "theoretical_loss": 4.2880811542389505,
      "tokens_seen": 226754560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006498037639493096,
      "loss": 3.1334,
      "theoretical_loss": 4.287929109770217,
      "tokens_seen": 226820096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006497872139209381,
      "loss": 3.3074,
      "theoretical_loss": 4.287777121522501,
      "tokens_seen": 226885632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006497706638925667,
      "loss": 3.0728,
      "theoretical_loss": 4.287625189458781,
      "tokens_seen": 226951168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006497541138641952,
      "loss": 3.0162,
      "theoretical_loss": 4.287473313542077,
      "tokens_seen": 227016704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006497375638358237,
      "loss": 3.2328,
      "theoretical_loss": 4.287321493735438,
      "tokens_seen": 227082240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006497210138074523,
      "loss": 2.9694,
      "theoretical_loss": 4.287169730001949,
      "tokens_seen": 227147776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006497044637790808,
      "loss": 3.1505,
      "theoretical_loss": 4.287018022304733,
      "tokens_seen": 227213312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006496879137507092,
      "loss": 3.1684,
      "theoretical_loss": 4.286866370606943,
      "tokens_seen": 227278848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006496713637223378,
      "loss": 3.1022,
      "theoretical_loss": 4.286714774871772,
      "tokens_seen": 227344384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006496548136939663,
      "loss": 3.17,
      "theoretical_loss": 4.286563235062444,
      "tokens_seen": 227409920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006496382636655949,
      "loss": 3.2313,
      "theoretical_loss": 4.28641175114222,
      "tokens_seen": 227475456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006496217136372234,
      "loss": 2.9878,
      "theoretical_loss": 4.286260323074394,
      "tokens_seen": 227540992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006496051636088519,
      "loss": 3.2522,
      "theoretical_loss": 4.286108950822296,
      "tokens_seen": 227606528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006495886135804805,
      "loss": 3.0452,
      "theoretical_loss": 4.285957634349289,
      "tokens_seen": 227672064
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000649572063552109,
      "loss": 3.2769,
      "theoretical_loss": 4.285806373618774,
      "tokens_seen": 227737600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006495555135237374,
      "loss": 3.2647,
      "theoretical_loss": 4.285655168594182,
      "tokens_seen": 227803136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000649538963495366,
      "loss": 3.2697,
      "theoretical_loss": 4.285504019238982,
      "tokens_seen": 227868672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006495224134669945,
      "loss": 3.1582,
      "theoretical_loss": 4.285352925516676,
      "tokens_seen": 227934208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006495058634386231,
      "loss": 3.0659,
      "theoretical_loss": 4.2852018873908,
      "tokens_seen": 227999744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006494893134102516,
      "loss": 3.1361,
      "theoretical_loss": 4.285050904824925,
      "tokens_seen": 228065280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006494727633818801,
      "loss": 2.9414,
      "theoretical_loss": 4.284899977782658,
      "tokens_seen": 228130816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006494562133535087,
      "loss": 3.2208,
      "theoretical_loss": 4.284749106227636,
      "tokens_seen": 228196352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006494396633251371,
      "loss": 3.0908,
      "theoretical_loss": 4.284598290123535,
      "tokens_seen": 228261888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006494231132967657,
      "loss": 3.2911,
      "theoretical_loss": 4.284447529434061,
      "tokens_seen": 228327424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006494065632683942,
      "loss": 3.298,
      "theoretical_loss": 4.284296824122959,
      "tokens_seen": 228392960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006493900132400227,
      "loss": 3.2193,
      "theoretical_loss": 4.284146174154003,
      "tokens_seen": 228458496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006493734632116513,
      "loss": 3.0973,
      "theoretical_loss": 4.283995579491004,
      "tokens_seen": 228524032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006493569131832797,
      "loss": 3.21,
      "theoretical_loss": 4.283845040097807,
      "tokens_seen": 228589568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006493403631549083,
      "loss": 3.0833,
      "theoretical_loss": 4.28369455593829,
      "tokens_seen": 228655104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006493238131265367,
      "loss": 3.2225,
      "theoretical_loss": 4.2835441269763646,
      "tokens_seen": 228720640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006493072630981652,
      "loss": 3.0687,
      "theoretical_loss": 4.283393753175979,
      "tokens_seen": 228786176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006492907130697938,
      "loss": 3.088,
      "theoretical_loss": 4.283243434501112,
      "tokens_seen": 228851712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006492741630414223,
      "loss": 3.0256,
      "theoretical_loss": 4.283093170915778,
      "tokens_seen": 228917248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006492576130130509,
      "loss": 3.0925,
      "theoretical_loss": 4.282942962384023,
      "tokens_seen": 228982784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006492410629846794,
      "loss": 3.1554,
      "theoretical_loss": 4.282792808869932,
      "tokens_seen": 229048320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006492245129563079,
      "loss": 3.0762,
      "theoretical_loss": 4.282642710337618,
      "tokens_seen": 229113856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006492079629279365,
      "loss": 2.9895,
      "theoretical_loss": 4.28249266675123,
      "tokens_seen": 229179392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006491914128995649,
      "loss": 3.1226,
      "theoretical_loss": 4.282342678074951,
      "tokens_seen": 229244928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006491748628711934,
      "loss": 3.2379,
      "theoretical_loss": 4.2821927442729955,
      "tokens_seen": 229310464
    },
    {
      "epoch": 0.08,
      "objective/train/avg_token_score": 0.012236284092068672,
      "objective/train/avg_weight": 0.9856102466583252,
      "objective/train/docs_used": 141330,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2563705444335938,
      "objective/train/std_weight": 0.0597536675632,
      "objective/train/theoretical_loss": 4.282042865309616,
      "objective/train/tokens_used": 249836000,
      "theoretical_loss": 4.282042865309616,
      "tokens_seen": 229376000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000649158312842822,
      "loss": 3.0458,
      "theoretical_loss": 4.282042865309616,
      "tokens_seen": 229376000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006491417628144505,
      "loss": 3.1639,
      "theoretical_loss": 4.281893041149093,
      "tokens_seen": 229441536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006491252127860791,
      "loss": 3.1049,
      "theoretical_loss": 4.2817432717557455,
      "tokens_seen": 229507072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006491086627577076,
      "loss": 3.0802,
      "theoretical_loss": 4.28159355709392,
      "tokens_seen": 229572608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000649092112729336,
      "loss": 3.2232,
      "theoretical_loss": 4.281443897128004,
      "tokens_seen": 229638144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006490755627009646,
      "loss": 3.2744,
      "theoretical_loss": 4.2812942918224115,
      "tokens_seen": 229703680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006490590126725931,
      "loss": 2.9302,
      "theoretical_loss": 4.281144741141593,
      "tokens_seen": 229769216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006490424626442216,
      "loss": 3.2742,
      "theoretical_loss": 4.280995245050032,
      "tokens_seen": 229834752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006490259126158502,
      "loss": 3.2975,
      "theoretical_loss": 4.2808458035122445,
      "tokens_seen": 229900288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006490093625874787,
      "loss": 3.18,
      "theoretical_loss": 4.2806964164927805,
      "tokens_seen": 229965824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006489928125591073,
      "loss": 2.9765,
      "theoretical_loss": 4.280547083956224,
      "tokens_seen": 230031360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006489762625307358,
      "loss": 3.0253,
      "theoretical_loss": 4.280397805867188,
      "tokens_seen": 230096896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006489597125023642,
      "loss": 3.1168,
      "theoretical_loss": 4.280248582190324,
      "tokens_seen": 230162432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006489431624739928,
      "loss": 3.0641,
      "theoretical_loss": 4.280099412890312,
      "tokens_seen": 230227968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006489266124456213,
      "loss": 2.9004,
      "theoretical_loss": 4.279950297931869,
      "tokens_seen": 230293504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006489100624172498,
      "loss": 3.0196,
      "theoretical_loss": 4.27980123727974,
      "tokens_seen": 230359040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006488935123888784,
      "loss": 3.0446,
      "theoretical_loss": 4.279652230898709,
      "tokens_seen": 230424576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006488769623605069,
      "loss": 3.3457,
      "theoretical_loss": 4.279503278753586,
      "tokens_seen": 230490112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006488604123321355,
      "loss": 3.3637,
      "theoretical_loss": 4.27935438080922,
      "tokens_seen": 230555648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000648843862303764,
      "loss": 2.9002,
      "theoretical_loss": 4.27920553703049,
      "tokens_seen": 230621184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006488273122753924,
      "loss": 2.9388,
      "theoretical_loss": 4.279056747382306,
      "tokens_seen": 230686720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000648810762247021,
      "loss": 3.0034,
      "theoretical_loss": 4.278908011829613,
      "tokens_seen": 230752256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006487942122186495,
      "loss": 3.1291,
      "theoretical_loss": 4.27875933033739,
      "tokens_seen": 230817792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000648777662190278,
      "loss": 3.0247,
      "theoretical_loss": 4.278610702870646,
      "tokens_seen": 230883328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006487611121619066,
      "loss": 3.2503,
      "theoretical_loss": 4.278462129394423,
      "tokens_seen": 230948864
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006487445621335351,
      "loss": 3.1505,
      "theoretical_loss": 4.278313609873795,
      "tokens_seen": 231014400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006487280121051637,
      "loss": 3.0792,
      "theoretical_loss": 4.278165144273871,
      "tokens_seen": 231079936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006487114620767921,
      "loss": 3.1667,
      "theoretical_loss": 4.27801673255979,
      "tokens_seen": 231145472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006486949120484206,
      "loss": 3.2651,
      "theoretical_loss": 4.277868374696725,
      "tokens_seen": 231211008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006486783620200492,
      "loss": 2.939,
      "theoretical_loss": 4.277720070649879,
      "tokens_seen": 231276544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006486618119916777,
      "loss": 3.1286,
      "theoretical_loss": 4.277571820384491,
      "tokens_seen": 231342080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006486452619633062,
      "loss": 3.1566,
      "theoretical_loss": 4.277423623865829,
      "tokens_seen": 231407616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006486287119349348,
      "loss": 2.914,
      "theoretical_loss": 4.277275481059195,
      "tokens_seen": 231473152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006486121619065633,
      "loss": 3.1845,
      "theoretical_loss": 4.2771273919299215,
      "tokens_seen": 231538688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006485956118781919,
      "loss": 2.9896,
      "theoretical_loss": 4.276979356443377,
      "tokens_seen": 231604224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006485790618498203,
      "loss": 3.3497,
      "theoretical_loss": 4.276831374564957,
      "tokens_seen": 231669760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006485625118214488,
      "loss": 3.1355,
      "theoretical_loss": 4.276683446260093,
      "tokens_seen": 231735296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006485459617930774,
      "loss": 3.0614,
      "theoretical_loss": 4.276535571494247,
      "tokens_seen": 231800832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006485294117647059,
      "loss": 3.0868,
      "theoretical_loss": 4.276387750232913,
      "tokens_seen": 231866368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006485128617363344,
      "loss": 3.0658,
      "theoretical_loss": 4.276239982441617,
      "tokens_seen": 231931904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000648496311707963,
      "loss": 3.2786,
      "theoretical_loss": 4.276092268085918,
      "tokens_seen": 231997440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006484797616795915,
      "loss": 2.7872,
      "theoretical_loss": 4.275944607131406,
      "tokens_seen": 232062976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00064846321165122,
      "loss": 3.2606,
      "theoretical_loss": 4.275796999543703,
      "tokens_seen": 232128512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006484466616228485,
      "loss": 3.1313,
      "theoretical_loss": 4.275649445288461,
      "tokens_seen": 232194048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000648430111594477,
      "loss": 3.2228,
      "theoretical_loss": 4.275501944331367,
      "tokens_seen": 232259584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006484135615661056,
      "loss": 2.7997,
      "theoretical_loss": 4.275354496638139,
      "tokens_seen": 232325120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006483970115377341,
      "loss": 3.2128,
      "theoretical_loss": 4.275207102174525,
      "tokens_seen": 232390656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006483804615093626,
      "loss": 3.1179,
      "theoretical_loss": 4.275059760906305,
      "tokens_seen": 232456192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000648363911480991,
      "loss": 3.2108,
      "theoretical_loss": 4.2749124727992935,
      "tokens_seen": 232521728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006483473614526196,
      "loss": 3.1739,
      "theoretical_loss": 4.274765237819333,
      "tokens_seen": 232587264
    },
    {
      "epoch": 0.08,
      "objective/train/avg_token_score": 0.01783275231719017,
      "objective/train/avg_weight": 0.9790226221084595,
      "objective/train/docs_used": 143449,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2052998542785645,
      "objective/train/std_weight": 0.0899391695857048,
      "objective/train/theoretical_loss": 4.274618055932298,
      "objective/train/tokens_used": 253112800,
      "theoretical_loss": 4.274618055932298,
      "tokens_seen": 232652800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006483308114242481,
      "loss": 3.1493,
      "theoretical_loss": 4.274618055932298,
      "tokens_seen": 232652800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006483142613958766,
      "loss": 3.057,
      "theoretical_loss": 4.2744709271040975,
      "tokens_seen": 232718336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006482977113675052,
      "loss": 3.243,
      "theoretical_loss": 4.27432385130067,
      "tokens_seen": 232783872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006482811613391337,
      "loss": 3.1538,
      "theoretical_loss": 4.274176828487984,
      "tokens_seen": 232849408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006482646113107622,
      "loss": 3.1932,
      "theoretical_loss": 4.2740298586320415,
      "tokens_seen": 232914944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006482480612823908,
      "loss": 2.9463,
      "theoretical_loss": 4.273882941698876,
      "tokens_seen": 232980480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006482315112540192,
      "loss": 2.9989,
      "theoretical_loss": 4.27373607765455,
      "tokens_seen": 233046016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006482149612256478,
      "loss": 3.2497,
      "theoretical_loss": 4.2735892664651605,
      "tokens_seen": 233111552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006481984111972763,
      "loss": 3.2491,
      "theoretical_loss": 4.273442508096833,
      "tokens_seen": 233177088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006481818611689048,
      "loss": 3.4084,
      "theoretical_loss": 4.273295802515726,
      "tokens_seen": 233242624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006481653111405334,
      "loss": 3.0896,
      "theoretical_loss": 4.273149149688028,
      "tokens_seen": 233308160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006481487611121619,
      "loss": 3.2572,
      "theoretical_loss": 4.27300254957996,
      "tokens_seen": 233373696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006481322110837904,
      "loss": 3.1042,
      "theoretical_loss": 4.272856002157772,
      "tokens_seen": 233439232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000648115661055419,
      "loss": 3.0126,
      "theoretical_loss": 4.272709507387748,
      "tokens_seen": 233504768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006480991110270474,
      "loss": 3.1749,
      "theoretical_loss": 4.2725630652362,
      "tokens_seen": 233570304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000648082560998676,
      "loss": 3.1098,
      "theoretical_loss": 4.272416675669473,
      "tokens_seen": 233635840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006480660109703045,
      "loss": 3.0887,
      "theoretical_loss": 4.272270338653942,
      "tokens_seen": 233701376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000648049460941933,
      "loss": 3.0789,
      "theoretical_loss": 4.272124054156014,
      "tokens_seen": 233766912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006480329109135616,
      "loss": 3.0856,
      "theoretical_loss": 4.271977822142125,
      "tokens_seen": 233832448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006480163608851901,
      "loss": 3.0715,
      "theoretical_loss": 4.271831642578745,
      "tokens_seen": 233897984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006479998108568186,
      "loss": 3.1309,
      "theoretical_loss": 4.27168551543237,
      "tokens_seen": 233963520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006479832608284471,
      "loss": 3.0125,
      "theoretical_loss": 4.271539440669532,
      "tokens_seen": 234029056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006479667108000756,
      "loss": 2.9394,
      "theoretical_loss": 4.27139341825679,
      "tokens_seen": 234094592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006479501607717042,
      "loss": 3.3449,
      "theoretical_loss": 4.271247448160736,
      "tokens_seen": 234160128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006479336107433327,
      "loss": 3.0884,
      "theoretical_loss": 4.27110153034799,
      "tokens_seen": 234225664
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006479170607149612,
      "loss": 2.9466,
      "theoretical_loss": 4.270955664785207,
      "tokens_seen": 234291200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006479005106865898,
      "loss": 2.8904,
      "theoretical_loss": 4.2708098514390676,
      "tokens_seen": 234356736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006478839606582183,
      "loss": 3.1306,
      "theoretical_loss": 4.270664090276286,
      "tokens_seen": 234422272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006478674106298467,
      "loss": 2.9346,
      "theoretical_loss": 4.2705183812636065,
      "tokens_seen": 234487808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006478508606014753,
      "loss": 3.1502,
      "theoretical_loss": 4.270372724367803,
      "tokens_seen": 234553344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006478343105731038,
      "loss": 2.885,
      "theoretical_loss": 4.270227119555681,
      "tokens_seen": 234618880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006478177605447324,
      "loss": 3.1474,
      "theoretical_loss": 4.270081566794076,
      "tokens_seen": 234684416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006478012105163609,
      "loss": 2.9612,
      "theoretical_loss": 4.269936066049852,
      "tokens_seen": 234749952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006477846604879894,
      "loss": 2.8646,
      "theoretical_loss": 4.269790617289907,
      "tokens_seen": 234815488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000647768110459618,
      "loss": 3.0567,
      "theoretical_loss": 4.269645220481166,
      "tokens_seen": 234881024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006477515604312465,
      "loss": 3.1664,
      "theoretical_loss": 4.269499875590587,
      "tokens_seen": 234946560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006477350104028749,
      "loss": 2.953,
      "theoretical_loss": 4.269354582585156,
      "tokens_seen": 235012096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006477184603745035,
      "loss": 3.2043,
      "theoretical_loss": 4.269209341431889,
      "tokens_seen": 235077632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000647701910346132,
      "loss": 3.0402,
      "theoretical_loss": 4.269064152097835,
      "tokens_seen": 235143168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006476853603177606,
      "loss": 3.1549,
      "theoretical_loss": 4.26891901455007,
      "tokens_seen": 235208704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006476688102893891,
      "loss": 3.325,
      "theoretical_loss": 4.268773928755701,
      "tokens_seen": 235274240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006476522602610176,
      "loss": 3.0928,
      "theoretical_loss": 4.268628894681868,
      "tokens_seen": 235339776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006476357102326462,
      "loss": 3.1008,
      "theoretical_loss": 4.268483912295735,
      "tokens_seen": 235405312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006476191602042746,
      "loss": 2.8228,
      "theoretical_loss": 4.268338981564502,
      "tokens_seen": 235470848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006476026101759031,
      "loss": 3.3596,
      "theoretical_loss": 4.268194102455395,
      "tokens_seen": 235536384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006475860601475317,
      "loss": 2.9365,
      "theoretical_loss": 4.26804927493567,
      "tokens_seen": 235601920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006475695101191602,
      "loss": 3.0681,
      "theoretical_loss": 4.267904498972618,
      "tokens_seen": 235667456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006475529600907888,
      "loss": 3.0589,
      "theoretical_loss": 4.267759774533552,
      "tokens_seen": 235732992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006475364100624173,
      "loss": 2.8506,
      "theoretical_loss": 4.267615101585821,
      "tokens_seen": 235798528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006475198600340458,
      "loss": 3.178,
      "theoretical_loss": 4.267470480096801,
      "tokens_seen": 235864064
    },
    {
      "epoch": 0.08,
      "objective/train/avg_token_score": 0.03510875999927521,
      "objective/train/avg_weight": 0.958698034286499,
      "objective/train/docs_used": 144727,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6956684589385986,
      "objective/train/std_weight": 0.13787387311458588,
      "objective/train/theoretical_loss": 4.267325910033897,
      "objective/train/tokens_used": 256389600,
      "theoretical_loss": 4.267325910033897,
      "tokens_seen": 235929600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006475033100056744,
      "loss": 2.7325,
      "theoretical_loss": 4.267325910033897,
      "tokens_seen": 235929600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006474867599773028,
      "loss": 3.1735,
      "theoretical_loss": 4.267181391364547,
      "tokens_seen": 235995136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006474702099489313,
      "loss": 2.9593,
      "theoretical_loss": 4.267036924056215,
      "tokens_seen": 236060672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006474536599205599,
      "loss": 3.2326,
      "theoretical_loss": 4.266892508076397,
      "tokens_seen": 236126208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006474371098921884,
      "loss": 3.0933,
      "theoretical_loss": 4.266748143392617,
      "tokens_seen": 236191744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000647420559863817,
      "loss": 3.07,
      "theoretical_loss": 4.26660382997243,
      "tokens_seen": 236257280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006474040098354455,
      "loss": 2.9624,
      "theoretical_loss": 4.26645956778342,
      "tokens_seen": 236322816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000647387459807074,
      "loss": 3.1833,
      "theoretical_loss": 4.2663153567932,
      "tokens_seen": 236388352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006473709097787024,
      "loss": 3.0095,
      "theoretical_loss": 4.266171196969412,
      "tokens_seen": 236453888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006473543597503309,
      "loss": 3.141,
      "theoretical_loss": 4.2660270882797295,
      "tokens_seen": 236519424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006473378097219595,
      "loss": 3.0145,
      "theoretical_loss": 4.265883030691853,
      "tokens_seen": 236584960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000647321259693588,
      "loss": 3.099,
      "theoretical_loss": 4.265739024173515,
      "tokens_seen": 236650496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006473047096652165,
      "loss": 3.135,
      "theoretical_loss": 4.265595068692473,
      "tokens_seen": 236716032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006472881596368451,
      "loss": 2.9828,
      "theoretical_loss": 4.26545116421652,
      "tokens_seen": 236781568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006472716096084736,
      "loss": 2.8922,
      "theoretical_loss": 4.265307310713471,
      "tokens_seen": 236847104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006472550595801021,
      "loss": 3.133,
      "theoretical_loss": 4.2651635081511765,
      "tokens_seen": 236912640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006472385095517306,
      "loss": 3.0176,
      "theoretical_loss": 4.265019756497512,
      "tokens_seen": 236978176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006472219595233591,
      "loss": 3.0639,
      "theoretical_loss": 4.264876055720386,
      "tokens_seen": 237043712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006472054094949877,
      "loss": 3.147,
      "theoretical_loss": 4.264732405787731,
      "tokens_seen": 237109248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006471888594666162,
      "loss": 2.9345,
      "theoretical_loss": 4.264588806667513,
      "tokens_seen": 237174784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006471723094382447,
      "loss": 3.2615,
      "theoretical_loss": 4.264445258327724,
      "tokens_seen": 237240320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006471557594098733,
      "loss": 3.253,
      "theoretical_loss": 4.264301760736389,
      "tokens_seen": 237305856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006471392093815017,
      "loss": 3.058,
      "theoretical_loss": 4.264158313861557,
      "tokens_seen": 237371392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006471226593531303,
      "loss": 3.0638,
      "theoretical_loss": 4.264014917671309,
      "tokens_seen": 237436928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006471061093247588,
      "loss": 3.1968,
      "theoretical_loss": 4.2638715721337554,
      "tokens_seen": 237502464
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006470895592963873,
      "loss": 3.0401,
      "theoretical_loss": 4.263728277217032,
      "tokens_seen": 237568000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006470730092680159,
      "loss": 3.0694,
      "theoretical_loss": 4.263585032889306,
      "tokens_seen": 237633536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006470564592396444,
      "loss": 3.2627,
      "theoretical_loss": 4.263441839118776,
      "tokens_seen": 237699072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006470399092112729,
      "loss": 2.7748,
      "theoretical_loss": 4.2632986958736625,
      "tokens_seen": 237764608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006470233591829015,
      "loss": 2.7065,
      "theoretical_loss": 4.263155603122221,
      "tokens_seen": 237830144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006470068091545299,
      "loss": 2.9057,
      "theoretical_loss": 4.263012560832733,
      "tokens_seen": 237895680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0006469902591261585,
      "loss": 3.0541,
      "theoretical_loss": 4.262869568973508,
      "tokens_seen": 237961216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000646973709097787,
      "loss": 2.7371,
      "theoretical_loss": 4.262726627512886,
      "tokens_seen": 238026752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006469571590694155,
      "loss": 2.9943,
      "theoretical_loss": 4.262583736419234,
      "tokens_seen": 238092288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006469406090410441,
      "loss": 2.9808,
      "theoretical_loss": 4.26244089566095,
      "tokens_seen": 238157824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006469240590126726,
      "loss": 3.0542,
      "theoretical_loss": 4.262298105206456,
      "tokens_seen": 238223360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006469075089843012,
      "loss": 2.9994,
      "theoretical_loss": 4.262155365024207,
      "tokens_seen": 238288896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006468909589559296,
      "loss": 3.0027,
      "theoretical_loss": 4.262012675082685,
      "tokens_seen": 238354432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006468744089275581,
      "loss": 3.1746,
      "theoretical_loss": 4.261870035350399,
      "tokens_seen": 238419968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006468578588991867,
      "loss": 2.8992,
      "theoretical_loss": 4.261727445795888,
      "tokens_seen": 238485504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006468413088708152,
      "loss": 2.9528,
      "theoretical_loss": 4.26158490638772,
      "tokens_seen": 238551040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006468247588424437,
      "loss": 3.0733,
      "theoretical_loss": 4.261442417094488,
      "tokens_seen": 238616576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006468082088140723,
      "loss": 2.9785,
      "theoretical_loss": 4.261299977884816,
      "tokens_seen": 238682112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006467916587857008,
      "loss": 2.832,
      "theoretical_loss": 4.2611575887273565,
      "tokens_seen": 238747648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006467751087573294,
      "loss": 3.0761,
      "theoretical_loss": 4.261015249590789,
      "tokens_seen": 238813184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006467585587289578,
      "loss": 3.1062,
      "theoretical_loss": 4.260872960443822,
      "tokens_seen": 238878720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006467420087005863,
      "loss": 2.8798,
      "theoretical_loss": 4.260730721255191,
      "tokens_seen": 238944256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006467254586722149,
      "loss": 3.2458,
      "theoretical_loss": 4.260588531993662,
      "tokens_seen": 239009792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006467089086438434,
      "loss": 3.2474,
      "theoretical_loss": 4.260446392628026,
      "tokens_seen": 239075328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006466923586154719,
      "loss": 2.9991,
      "theoretical_loss": 4.2603043031271035,
      "tokens_seen": 239140864
    },
    {
      "epoch": 0.09,
      "objective/train/avg_token_score": 0.016853267326951027,
      "objective/train/avg_weight": 0.9801644682884216,
      "objective/train/docs_used": 146731,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.224689483642578,
      "objective/train/std_weight": 0.11020276695489883,
      "objective/train/theoretical_loss": 4.260162263459744,
      "objective/train/tokens_used": 259666400,
      "theoretical_loss": 4.260162263459744,
      "tokens_seen": 239206400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006466758085871005,
      "loss": 3.0588,
      "theoretical_loss": 4.260162263459744,
      "tokens_seen": 239206400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000646659258558729,
      "loss": 2.9717,
      "theoretical_loss": 4.260020273594824,
      "tokens_seen": 239271936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006466427085303575,
      "loss": 2.8675,
      "theoretical_loss": 4.259878333501247,
      "tokens_seen": 239337472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000646626158501986,
      "loss": 3.0146,
      "theoretical_loss": 4.259736443147946,
      "tokens_seen": 239403008
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006466096084736145,
      "loss": 2.7853,
      "theoretical_loss": 4.259594602503881,
      "tokens_seen": 239468544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006465930584452431,
      "loss": 3.0161,
      "theoretical_loss": 4.259452811538041,
      "tokens_seen": 239534080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006465765084168716,
      "loss": 2.9322,
      "theoretical_loss": 4.259311070219441,
      "tokens_seen": 239599616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006465599583885001,
      "loss": 3.2121,
      "theoretical_loss": 4.259169378517125,
      "tokens_seen": 239665152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006465434083601287,
      "loss": 3.179,
      "theoretical_loss": 4.259027736400165,
      "tokens_seen": 239730688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006465268583317571,
      "loss": 2.9259,
      "theoretical_loss": 4.258886143837661,
      "tokens_seen": 239796224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006465103083033857,
      "loss": 2.9005,
      "theoretical_loss": 4.258744600798739,
      "tokens_seen": 239861760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006464937582750142,
      "loss": 3.0765,
      "theoretical_loss": 4.2586031072525525,
      "tokens_seen": 239927296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006464772082466427,
      "loss": 2.9983,
      "theoretical_loss": 4.258461663168285,
      "tokens_seen": 239992832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006464606582182713,
      "loss": 3.2922,
      "theoretical_loss": 4.258320268515147,
      "tokens_seen": 240058368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006464441081898998,
      "loss": 3.2208,
      "theoretical_loss": 4.258178923262376,
      "tokens_seen": 240123904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006464275581615283,
      "loss": 3.1519,
      "theoretical_loss": 4.258037627379235,
      "tokens_seen": 240189440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006464110081331567,
      "loss": 3.2225,
      "theoretical_loss": 4.257896380835018,
      "tokens_seen": 240254976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006463944581047852,
      "loss": 3.2263,
      "theoretical_loss": 4.257755183599045,
      "tokens_seen": 240320512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006463779080764138,
      "loss": 3.0618,
      "theoretical_loss": 4.257614035640662,
      "tokens_seen": 240386048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006463613580480423,
      "loss": 2.9514,
      "theoretical_loss": 4.257472936929246,
      "tokens_seen": 240451584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006463448080196709,
      "loss": 3.2317,
      "theoretical_loss": 4.257331887434198,
      "tokens_seen": 240517120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006463282579912994,
      "loss": 3.0928,
      "theoretical_loss": 4.257190887124946,
      "tokens_seen": 240582656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006463117079629279,
      "loss": 3.0154,
      "theoretical_loss": 4.25704993597095,
      "tokens_seen": 240648192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006462951579345565,
      "loss": 2.8017,
      "theoretical_loss": 4.256909033941691,
      "tokens_seen": 240713728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006462786079061849,
      "loss": 3.0918,
      "theoretical_loss": 4.256768181006683,
      "tokens_seen": 240779264
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006462620578778134,
      "loss": 2.9028,
      "theoretical_loss": 4.2566273771354615,
      "tokens_seen": 240844800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000646245507849442,
      "loss": 3.0595,
      "theoretical_loss": 4.256486622297595,
      "tokens_seen": 240910336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006462289578210705,
      "loss": 2.7744,
      "theoretical_loss": 4.256345916462674,
      "tokens_seen": 240975872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006462124077926991,
      "loss": 3.2979,
      "theoretical_loss": 4.256205259600321,
      "tokens_seen": 241041408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006461958577643276,
      "loss": 3.1118,
      "theoretical_loss": 4.256064651680182,
      "tokens_seen": 241106944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006461793077359561,
      "loss": 3.2687,
      "theoretical_loss": 4.255924092671931,
      "tokens_seen": 241172480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006461627577075846,
      "loss": 3.0567,
      "theoretical_loss": 4.255783582545269,
      "tokens_seen": 241238016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006461462076792131,
      "loss": 2.8696,
      "theoretical_loss": 4.255643121269924,
      "tokens_seen": 241303552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006461296576508416,
      "loss": 2.9485,
      "theoretical_loss": 4.255502708815651,
      "tokens_seen": 241369088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006461131076224702,
      "loss": 3.0361,
      "theoretical_loss": 4.255362345152234,
      "tokens_seen": 241434624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006460965575940987,
      "loss": 3.0559,
      "theoretical_loss": 4.255222030249479,
      "tokens_seen": 241500160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006460800075657273,
      "loss": 3.0339,
      "theoretical_loss": 4.255081764077224,
      "tokens_seen": 241565696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006460634575373558,
      "loss": 2.8923,
      "theoretical_loss": 4.25494154660533,
      "tokens_seen": 241631232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006460469075089842,
      "loss": 3.0198,
      "theoretical_loss": 4.254801377803689,
      "tokens_seen": 241696768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006460303574806128,
      "loss": 3.1836,
      "theoretical_loss": 4.254661257642215,
      "tokens_seen": 241762304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006460138074522413,
      "loss": 3.2364,
      "theoretical_loss": 4.254521186090852,
      "tokens_seen": 241827840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006459972574238698,
      "loss": 3.2256,
      "theoretical_loss": 4.254381163119568,
      "tokens_seen": 241893376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006459807073954984,
      "loss": 2.7333,
      "theoretical_loss": 4.254241188698361,
      "tokens_seen": 241958912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006459641573671269,
      "loss": 3.1621,
      "theoretical_loss": 4.2541012627972545,
      "tokens_seen": 242024448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006459476073387555,
      "loss": 3.2954,
      "theoretical_loss": 4.2539613853862965,
      "tokens_seen": 242089984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000645931057310384,
      "loss": 3.232,
      "theoretical_loss": 4.253821556435565,
      "tokens_seen": 242155520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006459145072820124,
      "loss": 3.2008,
      "theoretical_loss": 4.253681775915161,
      "tokens_seen": 242221056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000645897957253641,
      "loss": 3.1458,
      "theoretical_loss": 4.253542043795215,
      "tokens_seen": 242286592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006458814072252695,
      "loss": 3.1888,
      "theoretical_loss": 4.253402360045882,
      "tokens_seen": 242352128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000645864857196898,
      "loss": 3.0382,
      "theoretical_loss": 4.253262724637346,
      "tokens_seen": 242417664
    },
    {
      "epoch": 0.09,
      "objective/train/avg_token_score": 0.011695994064211845,
      "objective/train/avg_weight": 0.9862459301948547,
      "objective/train/docs_used": 148718,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.315828323364258,
      "objective/train/std_weight": 0.07529707252979279,
      "objective/train/theoretical_loss": 4.253123137539814,
      "objective/train/tokens_used": 262943200,
      "theoretical_loss": 4.253123137539814,
      "tokens_seen": 242483200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006458483071685266,
      "loss": 3.1513,
      "theoretical_loss": 4.253123137539814,
      "tokens_seen": 242483200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006458317571401551,
      "loss": 2.9678,
      "theoretical_loss": 4.252983598723521,
      "tokens_seen": 242548736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006458152071117837,
      "loss": 3.21,
      "theoretical_loss": 4.25284410815873,
      "tokens_seen": 242614272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006457986570834121,
      "loss": 3.1246,
      "theoretical_loss": 4.2527046658157275,
      "tokens_seen": 242679808
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006457821070550406,
      "loss": 3.3653,
      "theoretical_loss": 4.252565271664828,
      "tokens_seen": 242745344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006457655570266692,
      "loss": 3.196,
      "theoretical_loss": 4.252425925676373,
      "tokens_seen": 242810880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006457490069982977,
      "loss": 3.0302,
      "theoretical_loss": 4.252286627820727,
      "tokens_seen": 242876416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006457324569699262,
      "loss": 3.4483,
      "theoretical_loss": 4.252147378068285,
      "tokens_seen": 242941952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006457159069415548,
      "loss": 3.165,
      "theoretical_loss": 4.252008176389465,
      "tokens_seen": 243007488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006456993569131833,
      "loss": 2.7241,
      "theoretical_loss": 4.251869022754712,
      "tokens_seen": 243073024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006456828068848119,
      "loss": 2.869,
      "theoretical_loss": 4.251729917134498,
      "tokens_seen": 243138560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006456662568564403,
      "loss": 3.1598,
      "theoretical_loss": 4.251590859499322,
      "tokens_seen": 243204096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006456497068280688,
      "loss": 2.9471,
      "theoretical_loss": 4.251451849819704,
      "tokens_seen": 243269632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006456331567996974,
      "loss": 2.9735,
      "theoretical_loss": 4.251312888066197,
      "tokens_seen": 243335168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006456166067713259,
      "loss": 2.9204,
      "theoretical_loss": 4.251173974209375,
      "tokens_seen": 243400704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006456000567429544,
      "loss": 3.2136,
      "theoretical_loss": 4.251035108219839,
      "tokens_seen": 243466240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000645583506714583,
      "loss": 3.3139,
      "theoretical_loss": 4.250896290068218,
      "tokens_seen": 243531776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006455669566862115,
      "loss": 3.1124,
      "theoretical_loss": 4.250757519725165,
      "tokens_seen": 243597312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00064555040665784,
      "loss": 2.8638,
      "theoretical_loss": 4.25061879716136,
      "tokens_seen": 243662848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006455338566294685,
      "loss": 2.9695,
      "theoretical_loss": 4.250480122347507,
      "tokens_seen": 243728384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000645517306601097,
      "loss": 3.1648,
      "theoretical_loss": 4.250341495254337,
      "tokens_seen": 243793920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006455007565727256,
      "loss": 2.9934,
      "theoretical_loss": 4.250202915852608,
      "tokens_seen": 243859456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006454842065443541,
      "loss": 2.8838,
      "theoretical_loss": 4.250064384113102,
      "tokens_seen": 243924992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006454676565159827,
      "loss": 3.0582,
      "theoretical_loss": 4.249925900006627,
      "tokens_seen": 243990528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006454511064876112,
      "loss": 3.1217,
      "theoretical_loss": 4.249787463504019,
      "tokens_seen": 244056064
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006454345564592396,
      "loss": 3.1354,
      "theoretical_loss": 4.249649074576134,
      "tokens_seen": 244121600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006454180064308681,
      "loss": 2.9946,
      "theoretical_loss": 4.249510733193862,
      "tokens_seen": 244187136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006454014564024966,
      "loss": 3.0448,
      "theoretical_loss": 4.249372439328111,
      "tokens_seen": 244252672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006453849063741252,
      "loss": 2.892,
      "theoretical_loss": 4.249234192949818,
      "tokens_seen": 244318208
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006453683563457537,
      "loss": 3.0661,
      "theoretical_loss": 4.249095994029947,
      "tokens_seen": 244383744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006453518063173822,
      "loss": 3.0633,
      "theoretical_loss": 4.248957842539484,
      "tokens_seen": 244449280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006453352562890108,
      "loss": 3.2605,
      "theoretical_loss": 4.248819738449442,
      "tokens_seen": 244514816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006453187062606392,
      "loss": 3.1697,
      "theoretical_loss": 4.2486816817308615,
      "tokens_seen": 244580352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006453021562322678,
      "loss": 3.2011,
      "theoretical_loss": 4.248543672354805,
      "tokens_seen": 244645888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006452856062038963,
      "loss": 3.0426,
      "theoretical_loss": 4.248405710292364,
      "tokens_seen": 244711424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006452690561755248,
      "loss": 2.8987,
      "theoretical_loss": 4.248267795514652,
      "tokens_seen": 244776960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006452525061471534,
      "loss": 2.9982,
      "theoretical_loss": 4.248129927992808,
      "tokens_seen": 244842496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006452359561187819,
      "loss": 3.1918,
      "theoretical_loss": 4.247992107698002,
      "tokens_seen": 244908032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006452194060904104,
      "loss": 3.1631,
      "theoretical_loss": 4.247854334601421,
      "tokens_seen": 244973568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000645202856062039,
      "loss": 2.8814,
      "theoretical_loss": 4.247716608674283,
      "tokens_seen": 245039104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006451863060336674,
      "loss": 2.9902,
      "theoretical_loss": 4.247578929887829,
      "tokens_seen": 245104640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000645169756005296,
      "loss": 2.8626,
      "theoretical_loss": 4.247441298213326,
      "tokens_seen": 245170176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006451532059769245,
      "loss": 3.232,
      "theoretical_loss": 4.247303713622067,
      "tokens_seen": 245235712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000645136655948553,
      "loss": 3.2602,
      "theoretical_loss": 4.247166176085367,
      "tokens_seen": 245301248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006451201059201816,
      "loss": 3.2327,
      "theoretical_loss": 4.247028685574569,
      "tokens_seen": 245366784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006451035558918101,
      "loss": 3.0843,
      "theoretical_loss": 4.246891242061041,
      "tokens_seen": 245432320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006450870058634386,
      "loss": 3.2161,
      "theoretical_loss": 4.246753845516174,
      "tokens_seen": 245497856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006450704558350672,
      "loss": 3.1705,
      "theoretical_loss": 4.246616495911388,
      "tokens_seen": 245563392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006450539058066956,
      "loss": 2.9168,
      "theoretical_loss": 4.246479193218123,
      "tokens_seen": 245628928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006450373557783242,
      "loss": 3.2379,
      "theoretical_loss": 4.246341937407848,
      "tokens_seen": 245694464
    },
    {
      "epoch": 0.09,
      "objective/train/avg_token_score": 0.01696641743183136,
      "objective/train/avg_weight": 0.9800261855125427,
      "objective/train/docs_used": 150524,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1132922172546387,
      "objective/train/std_weight": 0.1020209789276123,
      "objective/train/theoretical_loss": 4.246204728452055,
      "objective/train/tokens_used": 266220000,
      "theoretical_loss": 4.246204728452055,
      "tokens_seen": 245760000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006450208057499527,
      "loss": 2.9924,
      "theoretical_loss": 4.246204728452055,
      "tokens_seen": 245760000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006450042557215812,
      "loss": 2.9997,
      "theoretical_loss": 4.246067566322259,
      "tokens_seen": 245825536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006449877056932098,
      "loss": 3.1102,
      "theoretical_loss": 4.245930450990007,
      "tokens_seen": 245891072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006449711556648383,
      "loss": 3.222,
      "theoretical_loss": 4.245793382426861,
      "tokens_seen": 245956608
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006449546056364668,
      "loss": 3.0099,
      "theoretical_loss": 4.245656360604417,
      "tokens_seen": 246022144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006449380556080953,
      "loss": 2.9856,
      "theoretical_loss": 4.24551938549429,
      "tokens_seen": 246087680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006449215055797238,
      "loss": 3.303,
      "theoretical_loss": 4.2453824570681205,
      "tokens_seen": 246153216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006449049555513524,
      "loss": 3.0698,
      "theoretical_loss": 4.245245575297577,
      "tokens_seen": 246218752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006448884055229809,
      "loss": 2.9313,
      "theoretical_loss": 4.2451087401543495,
      "tokens_seen": 246284288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006448718554946094,
      "loss": 3.0628,
      "theoretical_loss": 4.244971951610154,
      "tokens_seen": 246349824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000644855305466238,
      "loss": 3.3012,
      "theoretical_loss": 4.24483520963673,
      "tokens_seen": 246415360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006448387554378665,
      "loss": 3.1722,
      "theoretical_loss": 4.244698514205844,
      "tokens_seen": 246480896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006448222054094949,
      "loss": 3.2183,
      "theoretical_loss": 4.244561865289285,
      "tokens_seen": 246546432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006448056553811235,
      "loss": 3.3196,
      "theoretical_loss": 4.244425262858867,
      "tokens_seen": 246611968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000644789105352752,
      "loss": 2.8869,
      "theoretical_loss": 4.2442887068864295,
      "tokens_seen": 246677504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006447725553243806,
      "loss": 3.0632,
      "theoretical_loss": 4.244152197343835,
      "tokens_seen": 246743040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006447560052960091,
      "loss": 2.8352,
      "theoretical_loss": 4.244015734202973,
      "tokens_seen": 246808576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006447394552676376,
      "loss": 3.1696,
      "theoretical_loss": 4.243879317435755,
      "tokens_seen": 246874112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006447229052392662,
      "loss": 3.0234,
      "theoretical_loss": 4.243742947014117,
      "tokens_seen": 246939648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006447063552108947,
      "loss": 3.1598,
      "theoretical_loss": 4.243606622910021,
      "tokens_seen": 247005184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006446898051825231,
      "loss": 2.9983,
      "theoretical_loss": 4.243470345095453,
      "tokens_seen": 247070720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006446732551541517,
      "loss": 3.0202,
      "theoretical_loss": 4.2433341135424225,
      "tokens_seen": 247136256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006446567051257802,
      "loss": 3.1699,
      "theoretical_loss": 4.243197928222964,
      "tokens_seen": 247201792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006446401550974088,
      "loss": 3.1587,
      "theoretical_loss": 4.243061789109136,
      "tokens_seen": 247267328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006446236050690373,
      "loss": 3.1639,
      "theoretical_loss": 4.242925696173021,
      "tokens_seen": 247332864
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006446070550406658,
      "loss": 3.1583,
      "theoretical_loss": 4.2427896493867285,
      "tokens_seen": 247398400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006445905050122944,
      "loss": 3.1068,
      "theoretical_loss": 4.242653648722387,
      "tokens_seen": 247463936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006445739549839228,
      "loss": 3.2815,
      "theoretical_loss": 4.242517694152154,
      "tokens_seen": 247529472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006445574049555513,
      "loss": 3.2243,
      "theoretical_loss": 4.24238178564821,
      "tokens_seen": 247595008
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006445408549271799,
      "loss": 3.2858,
      "theoretical_loss": 4.242245923182756,
      "tokens_seen": 247660544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006445243048988084,
      "loss": 3.3363,
      "theoretical_loss": 4.242110106728022,
      "tokens_seen": 247726080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000644507754870437,
      "loss": 3.094,
      "theoretical_loss": 4.241974336256261,
      "tokens_seen": 247791616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006444912048420655,
      "loss": 3.3022,
      "theoretical_loss": 4.241838611739748,
      "tokens_seen": 247857152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000644474654813694,
      "loss": 3.0832,
      "theoretical_loss": 4.241702933150783,
      "tokens_seen": 247922688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006444581047853226,
      "loss": 3.1986,
      "theoretical_loss": 4.241567300461693,
      "tokens_seen": 247988224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006444415547569509,
      "loss": 2.9595,
      "theoretical_loss": 4.241431713644823,
      "tokens_seen": 248053760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006444250047285795,
      "loss": 2.9374,
      "theoretical_loss": 4.241296172672547,
      "tokens_seen": 248119296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000644408454700208,
      "loss": 2.9124,
      "theoretical_loss": 4.24116067751726,
      "tokens_seen": 248184832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006443919046718366,
      "loss": 2.7975,
      "theoretical_loss": 4.241025228151383,
      "tokens_seen": 248250368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006443753546434651,
      "loss": 3.1083,
      "theoretical_loss": 4.24088982454736,
      "tokens_seen": 248315904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006443588046150936,
      "loss": 3.0997,
      "theoretical_loss": 4.240754466677659,
      "tokens_seen": 248381440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006443422545867222,
      "loss": 3.1033,
      "theoretical_loss": 4.240619154514771,
      "tokens_seen": 248446976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006443257045583506,
      "loss": 3.107,
      "theoretical_loss": 4.240483888031212,
      "tokens_seen": 248512512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006443091545299791,
      "loss": 3.1892,
      "theoretical_loss": 4.240348667199521,
      "tokens_seen": 248578048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006442926045016077,
      "loss": 3.0738,
      "theoretical_loss": 4.240213491992261,
      "tokens_seen": 248643584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006442760544732362,
      "loss": 2.9855,
      "theoretical_loss": 4.240078362382019,
      "tokens_seen": 248709120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006442595044448648,
      "loss": 2.8478,
      "theoretical_loss": 4.239943278341404,
      "tokens_seen": 248774656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006442429544164933,
      "loss": 2.9756,
      "theoretical_loss": 4.239808239843052,
      "tokens_seen": 248840192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006442264043881218,
      "loss": 3.1241,
      "theoretical_loss": 4.239673246859619,
      "tokens_seen": 248905728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006442098543597503,
      "loss": 3.0827,
      "theoretical_loss": 4.239538299363788,
      "tokens_seen": 248971264
    },
    {
      "epoch": 0.09,
      "objective/train/avg_token_score": 0.01935936138033867,
      "objective/train/avg_weight": 0.9772146344184875,
      "objective/train/docs_used": 152460,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.5564610958099365,
      "objective/train/std_weight": 0.10297891497612,
      "objective/train/theoretical_loss": 4.239403397328261,
      "objective/train/tokens_used": 269496800,
      "theoretical_loss": 4.239403397328261,
      "tokens_seen": 249036800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006441933043313788,
      "loss": 3.2026,
      "theoretical_loss": 4.239403397328261,
      "tokens_seen": 249036800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006441767543030073,
      "loss": 3.2142,
      "theoretical_loss": 4.239268540725769,
      "tokens_seen": 249102336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006441602042746359,
      "loss": 3.1834,
      "theoretical_loss": 4.239133729529064,
      "tokens_seen": 249167872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006441436542462644,
      "loss": 3.1452,
      "theoretical_loss": 4.2389989637109196,
      "tokens_seen": 249233408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000644127104217893,
      "loss": 3.0067,
      "theoretical_loss": 4.2388642432441355,
      "tokens_seen": 249298944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006441105541895215,
      "loss": 3.1746,
      "theoretical_loss": 4.238729568101535,
      "tokens_seen": 249364480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006440940041611499,
      "loss": 3.0487,
      "theoretical_loss": 4.238594938255963,
      "tokens_seen": 249430016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006440774541327785,
      "loss": 3.0503,
      "theoretical_loss": 4.2384603536802885,
      "tokens_seen": 249495552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000644060904104407,
      "loss": 3.2652,
      "theoretical_loss": 4.238325814347404,
      "tokens_seen": 249561088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006440443540760355,
      "loss": 2.8821,
      "theoretical_loss": 4.238191320230227,
      "tokens_seen": 249626624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006440278040476641,
      "loss": 2.9749,
      "theoretical_loss": 4.238056871301695,
      "tokens_seen": 249692160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006440112540192926,
      "loss": 2.9557,
      "theoretical_loss": 4.237922467534771,
      "tokens_seen": 249757696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006439947039909212,
      "loss": 3.007,
      "theoretical_loss": 4.237788108902441,
      "tokens_seen": 249823232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006439781539625497,
      "loss": 3.2462,
      "theoretical_loss": 4.237653795377714,
      "tokens_seen": 249888768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006439616039341781,
      "loss": 3.0202,
      "theoretical_loss": 4.237519526933622,
      "tokens_seen": 249954304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006439450539058067,
      "loss": 2.9542,
      "theoretical_loss": 4.2373853035432205,
      "tokens_seen": 250019840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006439285038774352,
      "loss": 3.1444,
      "theoretical_loss": 4.237251125179588,
      "tokens_seen": 250085376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006439119538490637,
      "loss": 2.9416,
      "theoretical_loss": 4.237116991815826,
      "tokens_seen": 250150912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006438954038206923,
      "loss": 3.1806,
      "theoretical_loss": 4.23698290342506,
      "tokens_seen": 250216448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006438788537923208,
      "loss": 3.0485,
      "theoretical_loss": 4.236848859980437,
      "tokens_seen": 250281984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006438623037639494,
      "loss": 3.1513,
      "theoretical_loss": 4.23671486145513,
      "tokens_seen": 250347520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006438457537355778,
      "loss": 2.9611,
      "theoretical_loss": 4.236580907822331,
      "tokens_seen": 250413056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006438292037072063,
      "loss": 3.0424,
      "theoretical_loss": 4.236446999055257,
      "tokens_seen": 250478592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006438126536788349,
      "loss": 3.1592,
      "theoretical_loss": 4.2363131351271495,
      "tokens_seen": 250544128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006437961036504634,
      "loss": 2.8741,
      "theoretical_loss": 4.2361793160112695,
      "tokens_seen": 250609664
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006437795536220919,
      "loss": 3.242,
      "theoretical_loss": 4.236045541680905,
      "tokens_seen": 250675200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006437630035937205,
      "loss": 2.9111,
      "theoretical_loss": 4.235911812109363,
      "tokens_seen": 250740736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000643746453565349,
      "loss": 3.0369,
      "theoretical_loss": 4.235778127269976,
      "tokens_seen": 250806272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006437299035369776,
      "loss": 2.8259,
      "theoretical_loss": 4.235644487136098,
      "tokens_seen": 250871808
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000643713353508606,
      "loss": 2.9256,
      "theoretical_loss": 4.235510891681108,
      "tokens_seen": 250937344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006436968034802345,
      "loss": 3.1587,
      "theoretical_loss": 4.235377340878404,
      "tokens_seen": 251002880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006436802534518631,
      "loss": 3.2056,
      "theoretical_loss": 4.23524383470141,
      "tokens_seen": 251068416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006436637034234916,
      "loss": 3.2686,
      "theoretical_loss": 4.235110373123572,
      "tokens_seen": 251133952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006436471533951201,
      "loss": 3.1144,
      "theoretical_loss": 4.2349769561183574,
      "tokens_seen": 251199488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006436306033667487,
      "loss": 3.0857,
      "theoretical_loss": 4.2348435836592575,
      "tokens_seen": 251265024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006436140533383772,
      "loss": 3.1077,
      "theoretical_loss": 4.234710255719786,
      "tokens_seen": 251330560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006435975033100057,
      "loss": 3.0392,
      "theoretical_loss": 4.234576972273481,
      "tokens_seen": 251396096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006435809532816342,
      "loss": 3.027,
      "theoretical_loss": 4.234443733293899,
      "tokens_seen": 251461632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006435644032532627,
      "loss": 3.2366,
      "theoretical_loss": 4.234310538754624,
      "tokens_seen": 251527168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006435478532248913,
      "loss": 3.0981,
      "theoretical_loss": 4.2341773886292575,
      "tokens_seen": 251592704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006435313031965198,
      "loss": 3.0644,
      "theoretical_loss": 4.234044282891429,
      "tokens_seen": 251658240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006435147531681483,
      "loss": 3.0616,
      "theoretical_loss": 4.233911221514787,
      "tokens_seen": 251723776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006434982031397769,
      "loss": 2.8635,
      "theoretical_loss": 4.233778204473002,
      "tokens_seen": 251789312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006434816531114053,
      "loss": 3.0587,
      "theoretical_loss": 4.23364523173977,
      "tokens_seen": 251854848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006434651030830338,
      "loss": 3.0943,
      "theoretical_loss": 4.233512303288807,
      "tokens_seen": 251920384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006434485530546623,
      "loss": 3.2967,
      "theoretical_loss": 4.233379419093851,
      "tokens_seen": 251985920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006434320030262909,
      "loss": 3.2692,
      "theoretical_loss": 4.233246579128666,
      "tokens_seen": 252051456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006434154529979194,
      "loss": 3.1787,
      "theoretical_loss": 4.233113783367033,
      "tokens_seen": 252116992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006433989029695479,
      "loss": 2.9851,
      "theoretical_loss": 4.232981031782761,
      "tokens_seen": 252182528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006433823529411765,
      "loss": 2.9787,
      "theoretical_loss": 4.232848324349677,
      "tokens_seen": 252248064
    },
    {
      "epoch": 0.09,
      "objective/train/avg_token_score": 0.009345844388008118,
      "objective/train/avg_weight": 0.9890009164810181,
      "objective/train/docs_used": 154232,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9869885444641113,
      "objective/train/std_weight": 0.04104503616690636,
      "objective/train/theoretical_loss": 4.232715661041632,
      "objective/train/tokens_used": 272773600,
      "theoretical_loss": 4.232715661041632,
      "tokens_seen": 252313600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006433658029128049,
      "loss": 2.8989,
      "theoretical_loss": 4.232715661041632,
      "tokens_seen": 252313600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006433492528844334,
      "loss": 3.067,
      "theoretical_loss": 4.232583041832499,
      "tokens_seen": 252379136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000643332702856062,
      "loss": 3.0303,
      "theoretical_loss": 4.232450466696174,
      "tokens_seen": 252444672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006433161528276905,
      "loss": 3.1658,
      "theoretical_loss": 4.2323179356065745,
      "tokens_seen": 252510208
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006432996027993191,
      "loss": 2.9939,
      "theoretical_loss": 4.23218544853764,
      "tokens_seen": 252575744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006432830527709476,
      "loss": 2.93,
      "theoretical_loss": 4.232053005463333,
      "tokens_seen": 252641280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006432665027425761,
      "loss": 2.8313,
      "theoretical_loss": 4.231920606357638,
      "tokens_seen": 252706816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006432499527142047,
      "loss": 3.0473,
      "theoretical_loss": 4.231788251194559,
      "tokens_seen": 252772352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006432334026858331,
      "loss": 2.9785,
      "theoretical_loss": 4.231655939948127,
      "tokens_seen": 252837888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006432168526574616,
      "loss": 3.26,
      "theoretical_loss": 4.231523672592392,
      "tokens_seen": 252903424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006432003026290902,
      "loss": 3.0887,
      "theoretical_loss": 4.231391449101425,
      "tokens_seen": 252968960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006431837526007187,
      "loss": 3.1485,
      "theoretical_loss": 4.231259269449322,
      "tokens_seen": 253034496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006431672025723473,
      "loss": 3.0017,
      "theoretical_loss": 4.231127133610198,
      "tokens_seen": 253100032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006431506525439758,
      "loss": 3.0816,
      "theoretical_loss": 4.230995041558194,
      "tokens_seen": 253165568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006431341025156043,
      "loss": 3.0867,
      "theoretical_loss": 4.230862993267468,
      "tokens_seen": 253231104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006431175524872328,
      "loss": 2.9594,
      "theoretical_loss": 4.230730988712205,
      "tokens_seen": 253296640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006431010024588613,
      "loss": 3.0508,
      "theoretical_loss": 4.230599027866606,
      "tokens_seen": 253362176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006430844524304898,
      "loss": 3.2211,
      "theoretical_loss": 4.2304671107048994,
      "tokens_seen": 253427712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006430679024021184,
      "loss": 3.1315,
      "theoretical_loss": 4.2303352372013325,
      "tokens_seen": 253493248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006430513523737469,
      "loss": 2.978,
      "theoretical_loss": 4.230203407330176,
      "tokens_seen": 253558784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006430348023453755,
      "loss": 2.9681,
      "theoretical_loss": 4.230071621065721,
      "tokens_seen": 253624320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000643018252317004,
      "loss": 2.9598,
      "theoretical_loss": 4.2299398783822815,
      "tokens_seen": 253689856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006430017022886324,
      "loss": 2.9857,
      "theoretical_loss": 4.229808179254192,
      "tokens_seen": 253755392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000642985152260261,
      "loss": 3.0646,
      "theoretical_loss": 4.22967652365581,
      "tokens_seen": 253820928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006429686022318895,
      "loss": 2.9607,
      "theoretical_loss": 4.229544911561513,
      "tokens_seen": 253886464
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006429520522035181,
      "loss": 2.8896,
      "theoretical_loss": 4.229413342945703,
      "tokens_seen": 253952000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006429355021751466,
      "loss": 2.9601,
      "theoretical_loss": 4.229281817782801,
      "tokens_seen": 254017536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006429189521467751,
      "loss": 3.2422,
      "theoretical_loss": 4.229150336047251,
      "tokens_seen": 254083072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006429024021184037,
      "loss": 2.8742,
      "theoretical_loss": 4.229018897713519,
      "tokens_seen": 254148608
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006428858520900322,
      "loss": 3.0193,
      "theoretical_loss": 4.22888750275609,
      "tokens_seen": 254214144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006428693020616606,
      "loss": 2.7981,
      "theoretical_loss": 4.228756151149475,
      "tokens_seen": 254279680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006428527520332892,
      "loss": 2.9684,
      "theoretical_loss": 4.228624842868202,
      "tokens_seen": 254345216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006428362020049177,
      "loss": 2.8007,
      "theoretical_loss": 4.228493577886824,
      "tokens_seen": 254410752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006428196519765463,
      "loss": 3.0419,
      "theoretical_loss": 4.228362356179913,
      "tokens_seen": 254476288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006428031019481748,
      "loss": 3.1551,
      "theoretical_loss": 4.228231177722063,
      "tokens_seen": 254541824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006427865519198033,
      "loss": 3.1656,
      "theoretical_loss": 4.228100042487892,
      "tokens_seen": 254607360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006427700018914319,
      "loss": 3.4394,
      "theoretical_loss": 4.227968950452035,
      "tokens_seen": 254672896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006427534518630603,
      "loss": 3.0446,
      "theoretical_loss": 4.227837901589153,
      "tokens_seen": 254738432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006427369018346888,
      "loss": 3.0995,
      "theoretical_loss": 4.227706895873924,
      "tokens_seen": 254803968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006427203518063174,
      "loss": 3.0393,
      "theoretical_loss": 4.227575933281051,
      "tokens_seen": 254869504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006427038017779459,
      "loss": 2.9437,
      "theoretical_loss": 4.227445013785257,
      "tokens_seen": 254935040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006426872517495745,
      "loss": 3.1661,
      "theoretical_loss": 4.227314137361285,
      "tokens_seen": 255000576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000642670701721203,
      "loss": 3.1664,
      "theoretical_loss": 4.227183303983901,
      "tokens_seen": 255066112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006426541516928315,
      "loss": 3.0627,
      "theoretical_loss": 4.227052513627893,
      "tokens_seen": 255131648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006426376016644601,
      "loss": 3.1446,
      "theoretical_loss": 4.226921766268067,
      "tokens_seen": 255197184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006426210516360885,
      "loss": 3.0157,
      "theoretical_loss": 4.226791061879253,
      "tokens_seen": 255262720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000642604501607717,
      "loss": 3.1547,
      "theoretical_loss": 4.226660400436302,
      "tokens_seen": 255328256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006425879515793456,
      "loss": 3.0709,
      "theoretical_loss": 4.226529781914084,
      "tokens_seen": 255393792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006425714015509741,
      "loss": 3.158,
      "theoretical_loss": 4.226399206287493,
      "tokens_seen": 255459328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006425548515226027,
      "loss": 3.2035,
      "theoretical_loss": 4.226268673531442,
      "tokens_seen": 255524864
    },
    {
      "epoch": 0.09,
      "objective/train/avg_token_score": 0.014296915382146835,
      "objective/train/avg_weight": 0.9831701517105103,
      "objective/train/docs_used": 156109,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0366077423095703,
      "objective/train/std_weight": 0.08283554017543793,
      "objective/train/theoretical_loss": 4.226138183620867,
      "objective/train/tokens_used": 276050400,
      "theoretical_loss": 4.226138183620867,
      "tokens_seen": 255590400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006425383014942312,
      "loss": 3.0463,
      "theoretical_loss": 4.226138183620867,
      "tokens_seen": 255590400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006425217514658597,
      "loss": 3.0246,
      "theoretical_loss": 4.226007736530723,
      "tokens_seen": 255655936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006425052014374882,
      "loss": 3.1768,
      "theoretical_loss": 4.225877332235987,
      "tokens_seen": 255721472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006424886514091167,
      "loss": 2.9601,
      "theoretical_loss": 4.225746970711657,
      "tokens_seen": 255787008
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006424721013807452,
      "loss": 3.3464,
      "theoretical_loss": 4.225616651932753,
      "tokens_seen": 255852544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006424555513523737,
      "loss": 2.8837,
      "theoretical_loss": 4.225486375874315,
      "tokens_seen": 255918080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006424390013240022,
      "loss": 2.9867,
      "theoretical_loss": 4.225356142511402,
      "tokens_seen": 255983616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006424224512956308,
      "loss": 2.9673,
      "theoretical_loss": 4.225225951819099,
      "tokens_seen": 256049152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006424059012672593,
      "loss": 2.7014,
      "theoretical_loss": 4.225095803772507,
      "tokens_seen": 256114688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006423893512388878,
      "loss": 3.0112,
      "theoretical_loss": 4.22496569834675,
      "tokens_seen": 256180224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006423728012105163,
      "loss": 3.0238,
      "theoretical_loss": 4.224835635516973,
      "tokens_seen": 256245760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006423562511821448,
      "loss": 3.0929,
      "theoretical_loss": 4.224705615258341,
      "tokens_seen": 256311296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006423397011537734,
      "loss": 3.0066,
      "theoretical_loss": 4.224575637546041,
      "tokens_seen": 256376832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006423231511254019,
      "loss": 3.1223,
      "theoretical_loss": 4.224445702355279,
      "tokens_seen": 256442368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006423066010970304,
      "loss": 2.7698,
      "theoretical_loss": 4.2243158096612845,
      "tokens_seen": 256507904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000642290051068659,
      "loss": 3.1175,
      "theoretical_loss": 4.224185959439305,
      "tokens_seen": 256573440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006422735010402874,
      "loss": 3.1902,
      "theoretical_loss": 4.22405615166461,
      "tokens_seen": 256638976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000642256951011916,
      "loss": 3.0412,
      "theoretical_loss": 4.22392638631249,
      "tokens_seen": 256704512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006422404009835445,
      "loss": 3.2029,
      "theoretical_loss": 4.223796663358255,
      "tokens_seen": 256770048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000642223850955173,
      "loss": 3.1272,
      "theoretical_loss": 4.223666982777237,
      "tokens_seen": 256835584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006422073009268016,
      "loss": 3.0988,
      "theoretical_loss": 4.223537344544788,
      "tokens_seen": 256901120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006421907508984301,
      "loss": 3.2287,
      "theoretical_loss": 4.223407748636282,
      "tokens_seen": 256966656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006421742008700586,
      "loss": 3.2488,
      "theoretical_loss": 4.22327819502711,
      "tokens_seen": 257032192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006421576508416872,
      "loss": 3.1897,
      "theoretical_loss": 4.223148683692687,
      "tokens_seen": 257097728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006421411008133156,
      "loss": 3.0704,
      "theoretical_loss": 4.223019214608446,
      "tokens_seen": 257163264
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006421245507849442,
      "loss": 3.1891,
      "theoretical_loss": 4.222889787749845,
      "tokens_seen": 257228800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006421080007565727,
      "loss": 2.8782,
      "theoretical_loss": 4.222760403092358,
      "tokens_seen": 257294336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006420914507282012,
      "loss": 3.0895,
      "theoretical_loss": 4.22263106061148,
      "tokens_seen": 257359872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006420749006998298,
      "loss": 3.0545,
      "theoretical_loss": 4.222501760282729,
      "tokens_seen": 257425408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006420583506714583,
      "loss": 2.9101,
      "theoretical_loss": 4.22237250208164,
      "tokens_seen": 257490944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006420418006430868,
      "loss": 2.9291,
      "theoretical_loss": 4.222243285983772,
      "tokens_seen": 257556480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006420252506147153,
      "loss": 3.1174,
      "theoretical_loss": 4.222114111964703,
      "tokens_seen": 257622016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006420087005863438,
      "loss": 2.9791,
      "theoretical_loss": 4.221984980000029,
      "tokens_seen": 257687552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006419921505579724,
      "loss": 2.9495,
      "theoretical_loss": 4.2218558900653695,
      "tokens_seen": 257753088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006419756005296009,
      "loss": 2.8079,
      "theoretical_loss": 4.221726842136364,
      "tokens_seen": 257818624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006419590505012294,
      "loss": 3.2191,
      "theoretical_loss": 4.2215978361886695,
      "tokens_seen": 257884160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000641942500472858,
      "loss": 3.07,
      "theoretical_loss": 4.221468872197967,
      "tokens_seen": 257949696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006419259504444865,
      "loss": 2.8943,
      "theoretical_loss": 4.221339950139956,
      "tokens_seen": 258015232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000641909400416115,
      "loss": 3.0099,
      "theoretical_loss": 4.221211069990357,
      "tokens_seen": 258080768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006418928503877435,
      "loss": 2.6318,
      "theoretical_loss": 4.221082231724908,
      "tokens_seen": 258146304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000641876300359372,
      "loss": 2.9021,
      "theoretical_loss": 4.22095343531937,
      "tokens_seen": 258211840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006418597503310006,
      "loss": 2.8269,
      "theoretical_loss": 4.220824680749525,
      "tokens_seen": 258277376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006418432003026291,
      "loss": 2.9282,
      "theoretical_loss": 4.220695967991171,
      "tokens_seen": 258342912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006418266502742576,
      "loss": 2.9035,
      "theoretical_loss": 4.220567297020131,
      "tokens_seen": 258408448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006418101002458862,
      "loss": 2.8855,
      "theoretical_loss": 4.220438667812244,
      "tokens_seen": 258473984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006417935502175147,
      "loss": 2.9154,
      "theoretical_loss": 4.220310080343373,
      "tokens_seen": 258539520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006417770001891431,
      "loss": 2.7443,
      "theoretical_loss": 4.220181534589398,
      "tokens_seen": 258605056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006417604501607717,
      "loss": 2.8758,
      "theoretical_loss": 4.22005303052622,
      "tokens_seen": 258670592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006417439001324002,
      "loss": 2.9388,
      "theoretical_loss": 4.219924568129759,
      "tokens_seen": 258736128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006417273501040288,
      "loss": 2.9586,
      "theoretical_loss": 4.219796147375957,
      "tokens_seen": 258801664
    },
    {
      "epoch": 0.09,
      "objective/train/avg_token_score": 0.022811155766248703,
      "objective/train/avg_weight": 0.9731929898262024,
      "objective/train/docs_used": 158136,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1077709197998047,
      "objective/train/std_weight": 0.10143781453371048,
      "objective/train/theoretical_loss": 4.219667768240775,
      "objective/train/tokens_used": 279327200,
      "theoretical_loss": 4.219667768240775,
      "tokens_seen": 258867200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006417108000756573,
      "loss": 2.9573,
      "theoretical_loss": 4.219667768240775,
      "tokens_seen": 258867200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006416942500472858,
      "loss": 2.6178,
      "theoretical_loss": 4.219539430700195,
      "tokens_seen": 258932736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006416777000189144,
      "loss": 3.069,
      "theoretical_loss": 4.2194111347302155,
      "tokens_seen": 258998272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006416611499905428,
      "loss": 2.9072,
      "theoretical_loss": 4.219282880306859,
      "tokens_seen": 259063808
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006416445999621713,
      "loss": 3.1441,
      "theoretical_loss": 4.219154667406166,
      "tokens_seen": 259129344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006416280499337999,
      "loss": 3.1607,
      "theoretical_loss": 4.219026496004198,
      "tokens_seen": 259194880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006416114999054284,
      "loss": 3.1408,
      "theoretical_loss": 4.218898366077035,
      "tokens_seen": 259260416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000641594949877057,
      "loss": 3.1729,
      "theoretical_loss": 4.218770277600775,
      "tokens_seen": 259325952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006415783998486855,
      "loss": 2.9757,
      "theoretical_loss": 4.218642230551541,
      "tokens_seen": 259391488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000641561849820314,
      "loss": 2.8137,
      "theoretical_loss": 4.218514224905472,
      "tokens_seen": 259457024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006415452997919426,
      "loss": 3.261,
      "theoretical_loss": 4.218386260638727,
      "tokens_seen": 259522560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000641528749763571,
      "loss": 3.0568,
      "theoretical_loss": 4.2182583377274865,
      "tokens_seen": 259588096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006415121997351995,
      "loss": 3.2007,
      "theoretical_loss": 4.218130456147948,
      "tokens_seen": 259653632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000641495649706828,
      "loss": 2.7772,
      "theoretical_loss": 4.218002615876332,
      "tokens_seen": 259719168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006414790996784566,
      "loss": 3.0297,
      "theoretical_loss": 4.217874816888877,
      "tokens_seen": 259784704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006414625496500851,
      "loss": 3.1714,
      "theoretical_loss": 4.217747059161839,
      "tokens_seen": 259850240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006414459996217136,
      "loss": 3.0361,
      "theoretical_loss": 4.217619342671498,
      "tokens_seen": 259915776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006414294495933422,
      "loss": 3.0565,
      "theoretical_loss": 4.2174916673941505,
      "tokens_seen": 259981312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006414128995649706,
      "loss": 3.2357,
      "theoretical_loss": 4.217364033306113,
      "tokens_seen": 260046848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006413963495365991,
      "loss": 2.9765,
      "theoretical_loss": 4.217236440383724,
      "tokens_seen": 260112384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006413797995082277,
      "loss": 2.9758,
      "theoretical_loss": 4.217108888603337,
      "tokens_seen": 260177920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006413632494798562,
      "loss": 3.207,
      "theoretical_loss": 4.21698137794133,
      "tokens_seen": 260243456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006413466994514848,
      "loss": 3.1994,
      "theoretical_loss": 4.216853908374097,
      "tokens_seen": 260308992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006413301494231133,
      "loss": 2.8866,
      "theoretical_loss": 4.216726479878052,
      "tokens_seen": 260374528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006413135993947418,
      "loss": 3.1672,
      "theoretical_loss": 4.216599092429631,
      "tokens_seen": 260440064
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006412970493663703,
      "loss": 2.8914,
      "theoretical_loss": 4.216471746005286,
      "tokens_seen": 260505600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006412804993379988,
      "loss": 2.9245,
      "theoretical_loss": 4.216344440581491,
      "tokens_seen": 260571136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006412639493096273,
      "loss": 2.9393,
      "theoretical_loss": 4.2162171761347365,
      "tokens_seen": 260636672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006412473992812559,
      "loss": 2.7956,
      "theoretical_loss": 4.2160899526415365,
      "tokens_seen": 260702208
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006412308492528844,
      "loss": 3.2822,
      "theoretical_loss": 4.215962770078422,
      "tokens_seen": 260767744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000641214299224513,
      "loss": 3.2259,
      "theoretical_loss": 4.215835628421942,
      "tokens_seen": 260833280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006411977491961415,
      "loss": 3.1441,
      "theoretical_loss": 4.215708527648667,
      "tokens_seen": 260898816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00064118119916777,
      "loss": 2.6647,
      "theoretical_loss": 4.215581467735187,
      "tokens_seen": 260964352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006411646491393985,
      "loss": 3.0778,
      "theoretical_loss": 4.215454448658109,
      "tokens_seen": 261029888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000641148099111027,
      "loss": 3.1042,
      "theoretical_loss": 4.215327470394062,
      "tokens_seen": 261095424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006411315490826555,
      "loss": 2.6438,
      "theoretical_loss": 4.215200532919691,
      "tokens_seen": 261160960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006411149990542841,
      "loss": 3.0563,
      "theoretical_loss": 4.215073636211664,
      "tokens_seen": 261226496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006410984490259126,
      "loss": 2.7926,
      "theoretical_loss": 4.214946780246666,
      "tokens_seen": 261292032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006410818989975412,
      "loss": 2.933,
      "theoretical_loss": 4.214819965001401,
      "tokens_seen": 261357568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006410653489691697,
      "loss": 2.884,
      "theoretical_loss": 4.214693190452593,
      "tokens_seen": 261423104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006410487989407981,
      "loss": 2.9724,
      "theoretical_loss": 4.214566456576984,
      "tokens_seen": 261488640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006410322489124267,
      "loss": 3.1361,
      "theoretical_loss": 4.214439763351336,
      "tokens_seen": 261554176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006410156988840552,
      "loss": 2.9776,
      "theoretical_loss": 4.214313110752431,
      "tokens_seen": 261619712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006409991488556837,
      "loss": 3.0289,
      "theoretical_loss": 4.214186498757069,
      "tokens_seen": 261685248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006409825988273123,
      "loss": 3.118,
      "theoretical_loss": 4.214059927342068,
      "tokens_seen": 261750784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006409660487989408,
      "loss": 3.0047,
      "theoretical_loss": 4.213933396484267,
      "tokens_seen": 261816320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006409494987705694,
      "loss": 2.9896,
      "theoretical_loss": 4.213806906160523,
      "tokens_seen": 261881856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006409329487421978,
      "loss": 2.8432,
      "theoretical_loss": 4.213680456347712,
      "tokens_seen": 261947392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006409163987138263,
      "loss": 3.0709,
      "theoretical_loss": 4.213554047022729,
      "tokens_seen": 262012928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006408998486854549,
      "loss": 3.1083,
      "theoretical_loss": 4.213427678162489,
      "tokens_seen": 262078464
    },
    {
      "epoch": 0.09,
      "objective/train/avg_token_score": 0.020040113478899002,
      "objective/train/avg_weight": 0.9764354825019836,
      "objective/train/docs_used": 160243,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0845742225646973,
      "objective/train/std_weight": 0.10864706337451935,
      "objective/train/theoretical_loss": 4.213301349743924,
      "objective/train/tokens_used": 282604000,
      "theoretical_loss": 4.213301349743924,
      "tokens_seen": 262144000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006408832986570834,
      "loss": 2.8757,
      "theoretical_loss": 4.213301349743924,
      "tokens_seen": 262144000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006408667486287119,
      "loss": 3.2756,
      "theoretical_loss": 4.2131750617439865,
      "tokens_seen": 262209536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006408501986003405,
      "loss": 3.2409,
      "theoretical_loss": 4.213048814139647,
      "tokens_seen": 262275072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000640833648571969,
      "loss": 3.233,
      "theoretical_loss": 4.212922606907895,
      "tokens_seen": 262340608
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006408170985435976,
      "loss": 3.1548,
      "theoretical_loss": 4.21279644002574,
      "tokens_seen": 262406144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000640800548515226,
      "loss": 2.6978,
      "theoretical_loss": 4.212670313470209,
      "tokens_seen": 262471680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006407839984868545,
      "loss": 2.8689,
      "theoretical_loss": 4.212544227218347,
      "tokens_seen": 262537216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006407674484584831,
      "loss": 2.9821,
      "theoretical_loss": 4.21241818124722,
      "tokens_seen": 262602752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006407508984301116,
      "loss": 3.0794,
      "theoretical_loss": 4.212292175533912,
      "tokens_seen": 262668288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006407343484017401,
      "loss": 2.7098,
      "theoretical_loss": 4.212166210055526,
      "tokens_seen": 262733824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006407177983733687,
      "loss": 2.932,
      "theoretical_loss": 4.212040284789181,
      "tokens_seen": 262799360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006407012483449972,
      "loss": 2.8829,
      "theoretical_loss": 4.211914399712019,
      "tokens_seen": 262864896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006406846983166258,
      "loss": 2.9997,
      "theoretical_loss": 4.211788554801198,
      "tokens_seen": 262930432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006406681482882542,
      "loss": 2.8187,
      "theoretical_loss": 4.211662750033895,
      "tokens_seen": 262995968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006406515982598827,
      "loss": 3.1185,
      "theoretical_loss": 4.211536985387307,
      "tokens_seen": 263061504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006406350482315113,
      "loss": 2.9288,
      "theoretical_loss": 4.211411260838647,
      "tokens_seen": 263127040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006406184982031398,
      "loss": 2.837,
      "theoretical_loss": 4.2112855763651496,
      "tokens_seen": 263192576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006406019481747683,
      "loss": 3.0762,
      "theoretical_loss": 4.211159931944065,
      "tokens_seen": 263258112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006405853981463969,
      "loss": 3.1258,
      "theoretical_loss": 4.211034327552666,
      "tokens_seen": 263323648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006405688481180254,
      "loss": 2.9758,
      "theoretical_loss": 4.210908763168239,
      "tokens_seen": 263389184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006405522980896539,
      "loss": 2.8915,
      "theoretical_loss": 4.210783238768093,
      "tokens_seen": 263454720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006405357480612824,
      "loss": 2.7431,
      "theoretical_loss": 4.210657754329553,
      "tokens_seen": 263520256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006405191980329109,
      "loss": 3.0725,
      "theoretical_loss": 4.210532309829965,
      "tokens_seen": 263585792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006405026480045394,
      "loss": 2.8532,
      "theoretical_loss": 4.21040690524669,
      "tokens_seen": 263651328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006404860979761679,
      "loss": 2.9332,
      "theoretical_loss": 4.21028154055711,
      "tokens_seen": 263716864
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006404695479477965,
      "loss": 3.108,
      "theoretical_loss": 4.2101562157386265,
      "tokens_seen": 263782400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000640452997919425,
      "loss": 3.0901,
      "theoretical_loss": 4.210030930768655,
      "tokens_seen": 263847936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006404364478910535,
      "loss": 3.0199,
      "theoretical_loss": 4.2099056856246335,
      "tokens_seen": 263913472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000640419897862682,
      "loss": 2.8749,
      "theoretical_loss": 4.209780480284017,
      "tokens_seen": 263979008
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006404033478343105,
      "loss": 3.084,
      "theoretical_loss": 4.209655314724279,
      "tokens_seen": 264044544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006403867978059391,
      "loss": 2.8484,
      "theoretical_loss": 4.209530188922911,
      "tokens_seen": 264110080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006403702477775676,
      "loss": 2.8849,
      "theoretical_loss": 4.209405102857422,
      "tokens_seen": 264175616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006403536977491961,
      "loss": 2.9051,
      "theoretical_loss": 4.209280056505342,
      "tokens_seen": 264241152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006403371477208247,
      "loss": 2.8185,
      "theoretical_loss": 4.209155049844217,
      "tokens_seen": 264306688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006403205976924531,
      "loss": 2.946,
      "theoretical_loss": 4.209030082851612,
      "tokens_seen": 264372224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006403040476640817,
      "loss": 2.8219,
      "theoretical_loss": 4.208905155505109,
      "tokens_seen": 264437760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006402874976357102,
      "loss": 2.9162,
      "theoretical_loss": 4.20878026778231,
      "tokens_seen": 264503296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006402709476073387,
      "loss": 3.0735,
      "theoretical_loss": 4.208655419660834,
      "tokens_seen": 264568832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006402543975789673,
      "loss": 3.2119,
      "theoretical_loss": 4.208530611118321,
      "tokens_seen": 264634368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006402378475505958,
      "loss": 2.8759,
      "theoretical_loss": 4.208405842132423,
      "tokens_seen": 264699904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006402212975222243,
      "loss": 2.8682,
      "theoretical_loss": 4.208281112680817,
      "tokens_seen": 264765440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006402047474938529,
      "loss": 3.1851,
      "theoretical_loss": 4.208156422741195,
      "tokens_seen": 264830976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006401881974654813,
      "loss": 3.0899,
      "theoretical_loss": 4.208031772291265,
      "tokens_seen": 264896512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006401716474371099,
      "loss": 2.9197,
      "theoretical_loss": 4.207907161308757,
      "tokens_seen": 264962048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006401550974087384,
      "loss": 3.0122,
      "theoretical_loss": 4.2077825897714165,
      "tokens_seen": 265027584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006401385473803669,
      "loss": 3.0301,
      "theoretical_loss": 4.207658057657008,
      "tokens_seen": 265093120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006401219973519955,
      "loss": 3.0559,
      "theoretical_loss": 4.207533564943316,
      "tokens_seen": 265158656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000640105447323624,
      "loss": 2.9674,
      "theoretical_loss": 4.207409111608138,
      "tokens_seen": 265224192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006400888972952525,
      "loss": 3.028,
      "theoretical_loss": 4.2072846976292935,
      "tokens_seen": 265289728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000640072347266881,
      "loss": 2.7741,
      "theoretical_loss": 4.2071603229846195,
      "tokens_seen": 265355264
    },
    {
      "epoch": 0.09,
      "objective/train/avg_token_score": 0.04353150725364685,
      "objective/train/avg_weight": 0.9487758278846741,
      "objective/train/docs_used": 161560,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.752013683319092,
      "objective/train/std_weight": 0.2160005122423172,
      "objective/train/theoretical_loss": 4.20703598765197,
      "objective/train/tokens_used": 285880800,
      "theoretical_loss": 4.20703598765197,
      "tokens_seen": 265420800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006400557972385095,
      "loss": 2.8258,
      "theoretical_loss": 4.20703598765197,
      "tokens_seen": 265420800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006400392472101381,
      "loss": 2.9767,
      "theoretical_loss": 4.206911691609217,
      "tokens_seen": 265486336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006400226971817666,
      "loss": 2.9537,
      "theoretical_loss": 4.206787434834251,
      "tokens_seen": 265551872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006400061471533951,
      "loss": 2.9716,
      "theoretical_loss": 4.20666321730498,
      "tokens_seen": 265617408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006399895971250237,
      "loss": 2.7785,
      "theoretical_loss": 4.206539038999329,
      "tokens_seen": 265682944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006399730470966522,
      "loss": 2.9584,
      "theoretical_loss": 4.206414899895244,
      "tokens_seen": 265748480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006399564970682806,
      "loss": 2.9071,
      "theoretical_loss": 4.206290799970685,
      "tokens_seen": 265814016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006399399470399092,
      "loss": 3.0797,
      "theoretical_loss": 4.206166739203632,
      "tokens_seen": 265879552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0006399233970115377,
      "loss": 2.9837,
      "theoretical_loss": 4.206042717572082,
      "tokens_seen": 265945088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006399068469831663,
      "loss": 2.9463,
      "theoretical_loss": 4.20591873505405,
      "tokens_seen": 266010624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006398902969547948,
      "loss": 2.8608,
      "theoretical_loss": 4.20579479162757,
      "tokens_seen": 266076160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006398737469264233,
      "loss": 2.9758,
      "theoretical_loss": 4.205670887270691,
      "tokens_seen": 266141696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006398571968980519,
      "loss": 2.9497,
      "theoretical_loss": 4.205547021961482,
      "tokens_seen": 266207232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006398406468696804,
      "loss": 3.1574,
      "theoretical_loss": 4.205423195678029,
      "tokens_seen": 266272768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006398240968413088,
      "loss": 2.9582,
      "theoretical_loss": 4.205299408398435,
      "tokens_seen": 266338304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006398075468129374,
      "loss": 3.067,
      "theoretical_loss": 4.2051756601008226,
      "tokens_seen": 266403840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006397909967845659,
      "loss": 3.0668,
      "theoretical_loss": 4.20505195076333,
      "tokens_seen": 266469376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006397744467561945,
      "loss": 2.9188,
      "theoretical_loss": 4.204928280364115,
      "tokens_seen": 266534912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000639757896727823,
      "loss": 2.8479,
      "theoretical_loss": 4.20480464888135,
      "tokens_seen": 266600448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006397413466994515,
      "loss": 2.8514,
      "theoretical_loss": 4.204681056293228,
      "tokens_seen": 266665984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006397247966710801,
      "loss": 2.8756,
      "theoretical_loss": 4.204557502577957,
      "tokens_seen": 266731520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006397082466427085,
      "loss": 3.0187,
      "theoretical_loss": 4.204433987713767,
      "tokens_seen": 266797056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000639691696614337,
      "loss": 2.9064,
      "theoretical_loss": 4.2043105116789,
      "tokens_seen": 266862592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006396751465859656,
      "loss": 3.0344,
      "theoretical_loss": 4.204187074451617,
      "tokens_seen": 266928128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006396585965575941,
      "loss": 3.0064,
      "theoretical_loss": 4.204063676010202,
      "tokens_seen": 266993664
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006396420465292227,
      "loss": 2.9791,
      "theoretical_loss": 4.203940316332948,
      "tokens_seen": 267059200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006396254965008512,
      "loss": 2.8897,
      "theoretical_loss": 4.203816995398171,
      "tokens_seen": 267124736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006396089464724797,
      "loss": 3.1175,
      "theoretical_loss": 4.203693713184203,
      "tokens_seen": 267190272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006395923964441083,
      "loss": 2.5928,
      "theoretical_loss": 4.203570469669392,
      "tokens_seen": 267255808
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006395758464157367,
      "loss": 2.9245,
      "theoretical_loss": 4.203447264832107,
      "tokens_seen": 267321344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006395592963873652,
      "loss": 2.9822,
      "theoretical_loss": 4.203324098650731,
      "tokens_seen": 267386880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006395427463589938,
      "loss": 3.0397,
      "theoretical_loss": 4.203200971103666,
      "tokens_seen": 267452416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006395261963306222,
      "loss": 2.9347,
      "theoretical_loss": 4.20307788216933,
      "tokens_seen": 267517952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006395096463022508,
      "loss": 3.1927,
      "theoretical_loss": 4.202954831826159,
      "tokens_seen": 267583488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006394930962738793,
      "loss": 2.6679,
      "theoretical_loss": 4.202831820052609,
      "tokens_seen": 267649024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006394765462455079,
      "loss": 2.9388,
      "theoretical_loss": 4.202708846827148,
      "tokens_seen": 267714560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006394599962171363,
      "loss": 2.8728,
      "theoretical_loss": 4.202585912128266,
      "tokens_seen": 267780096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006394434461887648,
      "loss": 2.8964,
      "theoretical_loss": 4.202463015934468,
      "tokens_seen": 267845632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006394268961603934,
      "loss": 2.7936,
      "theoretical_loss": 4.202340158224277,
      "tokens_seen": 267911168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006394103461320219,
      "loss": 2.9908,
      "theoretical_loss": 4.202217338976231,
      "tokens_seen": 267976704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006393937961036504,
      "loss": 2.9314,
      "theoretical_loss": 4.2020945581688895,
      "tokens_seen": 268042240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000639377246075279,
      "loss": 2.9238,
      "theoretical_loss": 4.201971815780826,
      "tokens_seen": 268107776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006393606960469075,
      "loss": 2.751,
      "theoretical_loss": 4.201849111790631,
      "tokens_seen": 268173312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000639344146018536,
      "loss": 2.771,
      "theoretical_loss": 4.201726446176915,
      "tokens_seen": 268238848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006393275959901645,
      "loss": 3.0283,
      "theoretical_loss": 4.201603818918302,
      "tokens_seen": 268304384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000639311045961793,
      "loss": 2.7329,
      "theoretical_loss": 4.201481229993435,
      "tokens_seen": 268369920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006392944959334216,
      "loss": 2.9627,
      "theoretical_loss": 4.201358679380976,
      "tokens_seen": 268435456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006392779459050501,
      "loss": 2.9995,
      "theoretical_loss": 4.201236167059601,
      "tokens_seen": 268500992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006392613958766786,
      "loss": 3.0614,
      "theoretical_loss": 4.201113693008002,
      "tokens_seen": 268566528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006392448458483072,
      "loss": 2.8981,
      "theoretical_loss": 4.200991257204894,
      "tokens_seen": 268632064
    },
    {
      "epoch": 0.1,
      "objective/train/avg_token_score": 0.02670402079820633,
      "objective/train/avg_weight": 0.9686018228530884,
      "objective/train/docs_used": 163521,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0134329795837402,
      "objective/train/std_weight": 0.10621526092290878,
      "objective/train/theoretical_loss": 4.2008688596290025,
      "objective/train/tokens_used": 289157600,
      "theoretical_loss": 4.2008688596290025,
      "tokens_seen": 268697600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006392282958199356,
      "loss": 2.8193,
      "theoretical_loss": 4.2008688596290025,
      "tokens_seen": 268697600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006392117457915642,
      "loss": 2.8075,
      "theoretical_loss": 4.200746500259073,
      "tokens_seen": 268763136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006391951957631927,
      "loss": 2.9058,
      "theoretical_loss": 4.200624179073869,
      "tokens_seen": 268828672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006391786457348212,
      "loss": 2.7341,
      "theoretical_loss": 4.2005018960521685,
      "tokens_seen": 268894208
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006391620957064498,
      "loss": 3.112,
      "theoretical_loss": 4.200379651172769,
      "tokens_seen": 268959744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006391455456780783,
      "loss": 2.9955,
      "theoretical_loss": 4.200257444414483,
      "tokens_seen": 269025280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006391289956497068,
      "loss": 2.7253,
      "theoretical_loss": 4.200135275756139,
      "tokens_seen": 269090816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006391124456213354,
      "loss": 2.8001,
      "theoretical_loss": 4.200013145176587,
      "tokens_seen": 269156352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006390958955929638,
      "loss": 2.8136,
      "theoretical_loss": 4.199891052654689,
      "tokens_seen": 269221888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006390793455645924,
      "loss": 3.1372,
      "theoretical_loss": 4.199768998169326,
      "tokens_seen": 269287424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006390627955362209,
      "loss": 2.9767,
      "theoretical_loss": 4.199646981699395,
      "tokens_seen": 269352960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006390462455078494,
      "loss": 2.8506,
      "theoretical_loss": 4.199525003223812,
      "tokens_seen": 269418496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000639029695479478,
      "loss": 2.8127,
      "theoretical_loss": 4.199403062721506,
      "tokens_seen": 269484032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006390131454511065,
      "loss": 3.0163,
      "theoretical_loss": 4.199281160171427,
      "tokens_seen": 269549568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006389965954227351,
      "loss": 2.8901,
      "theoretical_loss": 4.1991592955525405,
      "tokens_seen": 269615104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006389800453943635,
      "loss": 3.0618,
      "theoretical_loss": 4.199037468843825,
      "tokens_seen": 269680640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000638963495365992,
      "loss": 2.9665,
      "theoretical_loss": 4.198915680024282,
      "tokens_seen": 269746176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006389469453376206,
      "loss": 2.7124,
      "theoretical_loss": 4.198793929072925,
      "tokens_seen": 269811712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006389303953092491,
      "loss": 3.0083,
      "theoretical_loss": 4.198672215968785,
      "tokens_seen": 269877248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006389138452808776,
      "loss": 2.9891,
      "theoretical_loss": 4.198550540690912,
      "tokens_seen": 269942784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006388972952525062,
      "loss": 2.9166,
      "theoretical_loss": 4.198428903218371,
      "tokens_seen": 270008320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006388807452241347,
      "loss": 2.6742,
      "theoretical_loss": 4.198307303530243,
      "tokens_seen": 270073856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006388641951957633,
      "loss": 2.7331,
      "theoretical_loss": 4.198185741605628,
      "tokens_seen": 270139392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006388476451673917,
      "loss": 2.8592,
      "theoretical_loss": 4.19806421742364,
      "tokens_seen": 270204928
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006388310951390202,
      "loss": 3.0184,
      "theoretical_loss": 4.197942730963412,
      "tokens_seen": 270270464
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006388145451106488,
      "loss": 3.1099,
      "theoretical_loss": 4.19782128220409,
      "tokens_seen": 270336000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006387979950822773,
      "loss": 3.0008,
      "theoretical_loss": 4.19769987112484,
      "tokens_seen": 270401536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006387814450539058,
      "loss": 2.7353,
      "theoretical_loss": 4.1975784977048445,
      "tokens_seen": 270467072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006387648950255344,
      "loss": 3.0457,
      "theoretical_loss": 4.1974571619233,
      "tokens_seen": 270532608
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006387483449971629,
      "loss": 2.871,
      "theoretical_loss": 4.197335863759422,
      "tokens_seen": 270598144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006387317949687914,
      "loss": 2.784,
      "theoretical_loss": 4.1972146031924416,
      "tokens_seen": 270663680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006387152449404199,
      "loss": 2.926,
      "theoretical_loss": 4.197093380201606,
      "tokens_seen": 270729216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006386986949120484,
      "loss": 2.7175,
      "theoretical_loss": 4.196972194766179,
      "tokens_seen": 270794752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000638682144883677,
      "loss": 2.8045,
      "theoretical_loss": 4.196851046865442,
      "tokens_seen": 270860288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006386655948553055,
      "loss": 2.9926,
      "theoretical_loss": 4.1967299364786905,
      "tokens_seen": 270925824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000638649044826934,
      "loss": 3.0259,
      "theoretical_loss": 4.196608863585239,
      "tokens_seen": 270991360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006386324947985626,
      "loss": 2.7649,
      "theoretical_loss": 4.1964878281644165,
      "tokens_seen": 271056896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000638615944770191,
      "loss": 2.9275,
      "theoretical_loss": 4.19636683019557,
      "tokens_seen": 271122432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006385993947418196,
      "loss": 2.9146,
      "theoretical_loss": 4.196245869658061,
      "tokens_seen": 271187968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006385828447134481,
      "loss": 2.797,
      "theoretical_loss": 4.1961249465312696,
      "tokens_seen": 271253504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006385662946850766,
      "loss": 2.7824,
      "theoretical_loss": 4.196004060794589,
      "tokens_seen": 271319040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006385497446567051,
      "loss": 3.0146,
      "theoretical_loss": 4.195883212427433,
      "tokens_seen": 271384576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006385331946283336,
      "loss": 2.8547,
      "theoretical_loss": 4.195762401409229,
      "tokens_seen": 271450112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006385166445999622,
      "loss": 2.8978,
      "theoretical_loss": 4.19564162771942,
      "tokens_seen": 271515648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006385000945715906,
      "loss": 2.8583,
      "theoretical_loss": 4.195520891337466,
      "tokens_seen": 271581184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006384835445432191,
      "loss": 2.657,
      "theoretical_loss": 4.195400192242845,
      "tokens_seen": 271646720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006384669945148477,
      "loss": 2.7609,
      "theoretical_loss": 4.19527953041505,
      "tokens_seen": 271712256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006384504444864762,
      "loss": 2.9111,
      "theoretical_loss": 4.19515890583359,
      "tokens_seen": 271777792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006384338944581048,
      "loss": 2.9683,
      "theoretical_loss": 4.195038318477989,
      "tokens_seen": 271843328
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006384173444297333,
      "loss": 3.0182,
      "theoretical_loss": 4.194917768327789,
      "tokens_seen": 271908864
    },
    {
      "epoch": 0.1,
      "objective/train/avg_token_score": 0.026347417384386063,
      "objective/train/avg_weight": 0.9689865112304688,
      "objective/train/docs_used": 165284,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.022684097290039,
      "objective/train/std_weight": 0.08971311151981354,
      "objective/train/theoretical_loss": 4.194797255362549,
      "objective/train/tokens_used": 292434400,
      "theoretical_loss": 4.194797255362549,
      "tokens_seen": 271974400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006384007944013618,
      "loss": 2.9241,
      "theoretical_loss": 4.194797255362549,
      "tokens_seen": 271974400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006383842443729904,
      "loss": 3.0414,
      "theoretical_loss": 4.194676779561841,
      "tokens_seen": 272039936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006383676943446188,
      "loss": 2.7914,
      "theoretical_loss": 4.194556340905256,
      "tokens_seen": 272105472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006383511443162473,
      "loss": 2.9955,
      "theoretical_loss": 4.194435939372401,
      "tokens_seen": 272171008
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006383345942878759,
      "loss": 2.8373,
      "theoretical_loss": 4.194315574942896,
      "tokens_seen": 272236544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006383180442595044,
      "loss": 2.5561,
      "theoretical_loss": 4.194195247596381,
      "tokens_seen": 272302080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000638301494231133,
      "loss": 3.0066,
      "theoretical_loss": 4.19407495731251,
      "tokens_seen": 272367616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006382849442027615,
      "loss": 2.7357,
      "theoretical_loss": 4.193954704070952,
      "tokens_seen": 272433152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00063826839417439,
      "loss": 2.9691,
      "theoretical_loss": 4.193834487851396,
      "tokens_seen": 272498688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006382518441460185,
      "loss": 2.9193,
      "theoretical_loss": 4.193714308633542,
      "tokens_seen": 272564224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000638235294117647,
      "loss": 2.6239,
      "theoretical_loss": 4.1935941663971095,
      "tokens_seen": 272629760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006382187440892755,
      "loss": 2.7777,
      "theoretical_loss": 4.193474061121833,
      "tokens_seen": 272695296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006382021940609041,
      "loss": 3.1458,
      "theoretical_loss": 4.193353992787463,
      "tokens_seen": 272760832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006381856440325326,
      "loss": 3.1655,
      "theoretical_loss": 4.193233961373766,
      "tokens_seen": 272826368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006381690940041612,
      "loss": 3.1594,
      "theoretical_loss": 4.1931139668605235,
      "tokens_seen": 272891904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006381525439757897,
      "loss": 3.0286,
      "theoretical_loss": 4.192994009227535,
      "tokens_seen": 272957440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006381359939474181,
      "loss": 2.9442,
      "theoretical_loss": 4.192874088454613,
      "tokens_seen": 273022976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006381194439190467,
      "loss": 2.7822,
      "theoretical_loss": 4.19275420452159,
      "tokens_seen": 273088512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006381028938906752,
      "loss": 2.93,
      "theoretical_loss": 4.192634357408309,
      "tokens_seen": 273154048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006380863438623037,
      "loss": 2.9907,
      "theoretical_loss": 4.192514547094634,
      "tokens_seen": 273219584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006380697938339323,
      "loss": 2.9068,
      "theoretical_loss": 4.192394773560441,
      "tokens_seen": 273285120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006380532438055608,
      "loss": 3.079,
      "theoretical_loss": 4.192275036785625,
      "tokens_seen": 273350656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006380366937771894,
      "loss": 3.0449,
      "theoretical_loss": 4.192155336750094,
      "tokens_seen": 273416192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006380201437488179,
      "loss": 2.8671,
      "theoretical_loss": 4.192035673433773,
      "tokens_seen": 273481728
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006380035937204463,
      "loss": 3.06,
      "theoretical_loss": 4.191916046816605,
      "tokens_seen": 273547264
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006379870436920749,
      "loss": 3.1046,
      "theoretical_loss": 4.191796456878544,
      "tokens_seen": 273612800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006379704936637034,
      "loss": 3.0691,
      "theoretical_loss": 4.191676903599563,
      "tokens_seen": 273678336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006379539436353319,
      "loss": 3.0123,
      "theoretical_loss": 4.191557386959651,
      "tokens_seen": 273743872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006379373936069605,
      "loss": 2.8048,
      "theoretical_loss": 4.191437906938811,
      "tokens_seen": 273809408
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000637920843578589,
      "loss": 2.7867,
      "theoretical_loss": 4.191318463517062,
      "tokens_seen": 273874944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006379042935502176,
      "loss": 3.2419,
      "theoretical_loss": 4.19119905667444,
      "tokens_seen": 273940480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000637887743521846,
      "loss": 2.7533,
      "theoretical_loss": 4.191079686390996,
      "tokens_seen": 274006016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006378711934934745,
      "loss": 3.1083,
      "theoretical_loss": 4.190960352646796,
      "tokens_seen": 274071552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006378546434651031,
      "loss": 2.9966,
      "theoretical_loss": 4.190841055421921,
      "tokens_seen": 274137088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006378380934367316,
      "loss": 2.7345,
      "theoretical_loss": 4.19072179469647,
      "tokens_seen": 274202624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006378215434083601,
      "loss": 3.0109,
      "theoretical_loss": 4.190602570450556,
      "tokens_seen": 274268160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006378049933799887,
      "loss": 3.0941,
      "theoretical_loss": 4.190483382664308,
      "tokens_seen": 274333696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006377884433516172,
      "loss": 3.1449,
      "theoretical_loss": 4.19036423131787,
      "tokens_seen": 274399232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006377718933232458,
      "loss": 2.8597,
      "theoretical_loss": 4.190245116391403,
      "tokens_seen": 274464768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006377553432948742,
      "loss": 3.0847,
      "theoretical_loss": 4.190126037865082,
      "tokens_seen": 274530304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006377387932665027,
      "loss": 2.7819,
      "theoretical_loss": 4.190006995719098,
      "tokens_seen": 274595840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006377222432381313,
      "loss": 2.9618,
      "theoretical_loss": 4.1898879899336565,
      "tokens_seen": 274661376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006377056932097598,
      "loss": 3.0757,
      "theoretical_loss": 4.189769020488981,
      "tokens_seen": 274726912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006376891431813883,
      "loss": 2.9824,
      "theoretical_loss": 4.189650087365309,
      "tokens_seen": 274792448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006376725931530169,
      "loss": 3.1248,
      "theoretical_loss": 4.189531190542893,
      "tokens_seen": 274857984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006376560431246454,
      "loss": 3.1678,
      "theoretical_loss": 4.189412330002001,
      "tokens_seen": 274923520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000637639493096274,
      "loss": 2.9761,
      "theoretical_loss": 4.189293505722918,
      "tokens_seen": 274989056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006376229430679024,
      "loss": 3.0682,
      "theoretical_loss": 4.189174717685942,
      "tokens_seen": 275054592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006376063930395309,
      "loss": 3.1699,
      "theoretical_loss": 4.189055965871389,
      "tokens_seen": 275120128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006375898430111595,
      "loss": 3.0784,
      "theoretical_loss": 4.188937250259587,
      "tokens_seen": 275185664
    },
    {
      "epoch": 0.1,
      "objective/train/avg_token_score": 0.037516407668590546,
      "objective/train/avg_weight": 0.9558591246604919,
      "objective/train/docs_used": 166979,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7970197200775146,
      "objective/train/std_weight": 0.20613829791545868,
      "objective/train/theoretical_loss": 4.188818570830883,
      "objective/train/tokens_used": 295711200,
      "theoretical_loss": 4.188818570830883,
      "tokens_seen": 275251200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000637573292982788,
      "loss": 2.8519,
      "theoretical_loss": 4.188818570830883,
      "tokens_seen": 275251200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006375567429544165,
      "loss": 3.0023,
      "theoretical_loss": 4.188699927565638,
      "tokens_seen": 275316736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000637540192926045,
      "loss": 2.8393,
      "theoretical_loss": 4.188581320444228,
      "tokens_seen": 275382272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006375236428976735,
      "loss": 2.9094,
      "theoretical_loss": 4.1884627494470426,
      "tokens_seen": 275447808
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000637507092869302,
      "loss": 3.0975,
      "theoretical_loss": 4.1883442145544905,
      "tokens_seen": 275513344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006374905428409305,
      "loss": 2.9467,
      "theoretical_loss": 4.188225715746992,
      "tokens_seen": 275578880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006374739928125591,
      "loss": 3.0486,
      "theoretical_loss": 4.188107253004986,
      "tokens_seen": 275644416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006374574427841876,
      "loss": 2.8647,
      "theoretical_loss": 4.187988826308925,
      "tokens_seen": 275709952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006374408927558161,
      "loss": 3.1136,
      "theoretical_loss": 4.187870435639275,
      "tokens_seen": 275775488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006374243427274447,
      "loss": 3.0937,
      "theoretical_loss": 4.18775208097652,
      "tokens_seen": 275841024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006374077926990731,
      "loss": 2.8682,
      "theoretical_loss": 4.187633762301159,
      "tokens_seen": 275906560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006373912426707017,
      "loss": 2.8317,
      "theoretical_loss": 4.187515479593704,
      "tokens_seen": 275972096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006373746926423302,
      "loss": 3.0655,
      "theoretical_loss": 4.187397232834683,
      "tokens_seen": 276037632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006373581426139587,
      "loss": 2.8156,
      "theoretical_loss": 4.187279022004642,
      "tokens_seen": 276103168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006373415925855873,
      "loss": 3.1159,
      "theoretical_loss": 4.1871608470841375,
      "tokens_seen": 276168704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006373250425572158,
      "loss": 3.1252,
      "theoretical_loss": 4.1870427080537445,
      "tokens_seen": 276234240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006373084925288443,
      "loss": 2.836,
      "theoretical_loss": 4.1869246048940525,
      "tokens_seen": 276299776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006372919425004729,
      "loss": 3.0579,
      "theoretical_loss": 4.186806537585666,
      "tokens_seen": 276365312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006372753924721013,
      "loss": 2.9231,
      "theoretical_loss": 4.186688506109202,
      "tokens_seen": 276430848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006372588424437299,
      "loss": 2.9232,
      "theoretical_loss": 4.186570510445296,
      "tokens_seen": 276496384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006372422924153584,
      "loss": 2.9724,
      "theoretical_loss": 4.186452550574599,
      "tokens_seen": 276561920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006372257423869869,
      "loss": 2.8267,
      "theoretical_loss": 4.186334626477774,
      "tokens_seen": 276627456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006372091923586155,
      "loss": 2.8641,
      "theoretical_loss": 4.186216738135501,
      "tokens_seen": 276692992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000637192642330244,
      "loss": 2.8471,
      "theoretical_loss": 4.186098885528473,
      "tokens_seen": 276758528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006371760923018725,
      "loss": 2.905,
      "theoretical_loss": 4.185981068637401,
      "tokens_seen": 276824064
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000637159542273501,
      "loss": 2.9946,
      "theoretical_loss": 4.185863287443008,
      "tokens_seen": 276889600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006371429922451295,
      "loss": 3.1546,
      "theoretical_loss": 4.185745541926035,
      "tokens_seen": 276955136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006371264422167581,
      "loss": 3.179,
      "theoretical_loss": 4.185627832067237,
      "tokens_seen": 277020672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006371098921883866,
      "loss": 3.1328,
      "theoretical_loss": 4.1855101578473795,
      "tokens_seen": 277086208
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006370933421600151,
      "loss": 3.1449,
      "theoretical_loss": 4.18539251924725,
      "tokens_seen": 277151744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006370767921316437,
      "loss": 2.9361,
      "theoretical_loss": 4.185274916247646,
      "tokens_seen": 277217280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006370602421032722,
      "loss": 3.0016,
      "theoretical_loss": 4.185157348829383,
      "tokens_seen": 277282816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006370436920749006,
      "loss": 3.057,
      "theoretical_loss": 4.185039816973289,
      "tokens_seen": 277348352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006370271420465292,
      "loss": 3.4111,
      "theoretical_loss": 4.184922320660207,
      "tokens_seen": 277413888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006370105920181577,
      "loss": 3.179,
      "theoretical_loss": 4.184804859870997,
      "tokens_seen": 277479424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006369940419897863,
      "loss": 3.1052,
      "theoretical_loss": 4.184687434586531,
      "tokens_seen": 277544960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006369774919614148,
      "loss": 3.1965,
      "theoretical_loss": 4.184570044787698,
      "tokens_seen": 277610496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006369609419330433,
      "loss": 3.1831,
      "theoretical_loss": 4.1844526904554,
      "tokens_seen": 277676032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006369443919046719,
      "loss": 3.0241,
      "theoretical_loss": 4.184335371570556,
      "tokens_seen": 277741568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006369278418763004,
      "loss": 2.9742,
      "theoretical_loss": 4.184218088114097,
      "tokens_seen": 277807104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006369112918479288,
      "loss": 2.8972,
      "theoretical_loss": 4.1841008400669715,
      "tokens_seen": 277872640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006368947418195574,
      "loss": 3.021,
      "theoretical_loss": 4.183983627410142,
      "tokens_seen": 277938176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006368781917911859,
      "loss": 2.769,
      "theoretical_loss": 4.183866450124584,
      "tokens_seen": 278003712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006368616417628145,
      "loss": 3.1005,
      "theoretical_loss": 4.18374930819129,
      "tokens_seen": 278069248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000636845091734443,
      "loss": 3.2051,
      "theoretical_loss": 4.183632201591264,
      "tokens_seen": 278134784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006368285417060715,
      "loss": 2.9994,
      "theoretical_loss": 4.18351513030553,
      "tokens_seen": 278200320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006368119916777001,
      "loss": 2.9105,
      "theoretical_loss": 4.1833980943151206,
      "tokens_seen": 278265856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006367954416493285,
      "loss": 2.9369,
      "theoretical_loss": 4.183281093601087,
      "tokens_seen": 278331392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000636778891620957,
      "loss": 3.0172,
      "theoretical_loss": 4.183164128144495,
      "tokens_seen": 278396928
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006367623415925856,
      "loss": 2.9857,
      "theoretical_loss": 4.183047197926422,
      "tokens_seen": 278462464
    },
    {
      "epoch": 0.1,
      "objective/train/avg_token_score": 0.019835636019706726,
      "objective/train/avg_weight": 0.9766846895217896,
      "objective/train/docs_used": 168865,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.542299747467041,
      "objective/train/std_weight": 0.10501302778720856,
      "objective/train/theoretical_loss": 4.182930302927963,
      "objective/train/tokens_used": 298988000,
      "theoretical_loss": 4.182930302927963,
      "tokens_seen": 278528000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006367457915642141,
      "loss": 3.1176,
      "theoretical_loss": 4.182930302927963,
      "tokens_seen": 278528000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006367292415358427,
      "loss": 3.1431,
      "theoretical_loss": 4.182813443130227,
      "tokens_seen": 278593536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006367126915074712,
      "loss": 2.8705,
      "theoretical_loss": 4.182696618514337,
      "tokens_seen": 278659072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006366961414790997,
      "loss": 2.7937,
      "theoretical_loss": 4.18257982906143,
      "tokens_seen": 278724608
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006366795914507283,
      "loss": 3.0432,
      "theoretical_loss": 4.1824630747526585,
      "tokens_seen": 278790144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006366630414223567,
      "loss": 2.9542,
      "theoretical_loss": 4.182346355569189,
      "tokens_seen": 278855680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006366464913939852,
      "loss": 2.7875,
      "theoretical_loss": 4.182229671492204,
      "tokens_seen": 278921216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006366299413656138,
      "loss": 3.0543,
      "theoretical_loss": 4.1821130225028975,
      "tokens_seen": 278986752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006366133913372423,
      "loss": 2.9394,
      "theoretical_loss": 4.1819964085824815,
      "tokens_seen": 279052288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006365968413088709,
      "loss": 2.8168,
      "theoretical_loss": 4.181879829712178,
      "tokens_seen": 279117824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006365802912804993,
      "loss": 2.9303,
      "theoretical_loss": 4.181763285873231,
      "tokens_seen": 279183360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006365637412521279,
      "loss": 2.8531,
      "theoretical_loss": 4.181646777046889,
      "tokens_seen": 279248896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006365471912237563,
      "loss": 2.9505,
      "theoretical_loss": 4.181530303214423,
      "tokens_seen": 279314432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006365306411953848,
      "loss": 2.8987,
      "theoretical_loss": 4.181413864357115,
      "tokens_seen": 279379968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006365140911670134,
      "loss": 2.9186,
      "theoretical_loss": 4.181297460456262,
      "tokens_seen": 279445504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006364975411386419,
      "loss": 3.2116,
      "theoretical_loss": 4.181181091493174,
      "tokens_seen": 279511040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006364809911102705,
      "loss": 2.7583,
      "theoretical_loss": 4.181064757449178,
      "tokens_seen": 279576576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000636464441081899,
      "loss": 3.0119,
      "theoretical_loss": 4.180948458305615,
      "tokens_seen": 279642112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006364478910535275,
      "loss": 3.0192,
      "theoretical_loss": 4.180832194043836,
      "tokens_seen": 279707648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000636431341025156,
      "loss": 3.0937,
      "theoretical_loss": 4.180715964645213,
      "tokens_seen": 279773184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006364147909967845,
      "loss": 2.9823,
      "theoretical_loss": 4.180599770091126,
      "tokens_seen": 279838720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000636398240968413,
      "loss": 3.0296,
      "theoretical_loss": 4.180483610362975,
      "tokens_seen": 279904256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006363816909400416,
      "loss": 3.0328,
      "theoretical_loss": 4.18036748544217,
      "tokens_seen": 279969792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006363651409116701,
      "loss": 2.95,
      "theoretical_loss": 4.180251395310137,
      "tokens_seen": 280035328
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006363485908832987,
      "loss": 3.1671,
      "theoretical_loss": 4.1801353399483165,
      "tokens_seen": 280100864
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006363320408549272,
      "loss": 2.9244,
      "theoretical_loss": 4.180019319338163,
      "tokens_seen": 280166400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006363154908265557,
      "loss": 3.1574,
      "theoretical_loss": 4.179903333461144,
      "tokens_seen": 280231936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006362989407981842,
      "loss": 3.039,
      "theoretical_loss": 4.179787382298744,
      "tokens_seen": 280297472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006362823907698127,
      "loss": 2.9779,
      "theoretical_loss": 4.179671465832458,
      "tokens_seen": 280363008
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006362658407414412,
      "loss": 3.08,
      "theoretical_loss": 4.179555584043799,
      "tokens_seen": 280428544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006362492907130698,
      "loss": 2.7667,
      "theoretical_loss": 4.17943973691429,
      "tokens_seen": 280494080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006362327406846983,
      "loss": 3.0073,
      "theoretical_loss": 4.179323924425472,
      "tokens_seen": 280559616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006362161906563269,
      "loss": 2.8563,
      "theoretical_loss": 4.179208146558899,
      "tokens_seen": 280625152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006361996406279554,
      "loss": 2.7344,
      "theoretical_loss": 4.1790924032961385,
      "tokens_seen": 280690688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006361830905995838,
      "loss": 2.8469,
      "theoretical_loss": 4.178976694618772,
      "tokens_seen": 280756224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006361665405712124,
      "loss": 3.2086,
      "theoretical_loss": 4.178861020508395,
      "tokens_seen": 280821760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006361499905428409,
      "loss": 3.0331,
      "theoretical_loss": 4.178745380946619,
      "tokens_seen": 280887296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006361334405144694,
      "loss": 3.0668,
      "theoretical_loss": 4.178629775915066,
      "tokens_seen": 280952832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000636116890486098,
      "loss": 2.9938,
      "theoretical_loss": 4.178514205395376,
      "tokens_seen": 281018368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006361003404577265,
      "loss": 2.8992,
      "theoretical_loss": 4.178398669369201,
      "tokens_seen": 281083904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006360837904293551,
      "loss": 3.0637,
      "theoretical_loss": 4.178283167818206,
      "tokens_seen": 281149440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006360672404009836,
      "loss": 2.7598,
      "theoretical_loss": 4.178167700724073,
      "tokens_seen": 281214976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000636050690372612,
      "loss": 3.014,
      "theoretical_loss": 4.178052268068494,
      "tokens_seen": 281280512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006360341403442406,
      "loss": 2.8609,
      "theoretical_loss": 4.177936869833179,
      "tokens_seen": 281346048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006360175903158691,
      "loss": 2.8812,
      "theoretical_loss": 4.17782150599985,
      "tokens_seen": 281411584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006360010402874976,
      "loss": 2.9687,
      "theoretical_loss": 4.1777061765502435,
      "tokens_seen": 281477120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006359844902591262,
      "loss": 2.9588,
      "theoretical_loss": 4.1775908814661085,
      "tokens_seen": 281542656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006359679402307547,
      "loss": 2.8776,
      "theoretical_loss": 4.17747562072921,
      "tokens_seen": 281608192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006359513902023833,
      "loss": 3.195,
      "theoretical_loss": 4.177360394321325,
      "tokens_seen": 281673728
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006359348401740117,
      "loss": 2.9009,
      "theoretical_loss": 4.177245202224246,
      "tokens_seen": 281739264
    },
    {
      "epoch": 0.1,
      "objective/train/avg_token_score": 0.016931619495153427,
      "objective/train/avg_weight": 0.9800909757614136,
      "objective/train/docs_used": 170536,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.203178882598877,
      "objective/train/std_weight": 0.10828635841608047,
      "objective/train/theoretical_loss": 4.17713004441978,
      "objective/train/tokens_used": 302264800,
      "theoretical_loss": 4.17713004441978,
      "tokens_seen": 281804800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006359182901456402,
      "loss": 2.9706,
      "theoretical_loss": 4.17713004441978,
      "tokens_seen": 281804800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006359017401172688,
      "loss": 3.1611,
      "theoretical_loss": 4.177014920889745,
      "tokens_seen": 281870336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006358851900888973,
      "loss": 3.0565,
      "theoretical_loss": 4.176899831615974,
      "tokens_seen": 281935872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006358686400605258,
      "loss": 2.9352,
      "theoretical_loss": 4.176784776580316,
      "tokens_seen": 282001408
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006358520900321544,
      "loss": 3.0625,
      "theoretical_loss": 4.176669755764632,
      "tokens_seen": 282066944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006358355400037829,
      "loss": 3.1504,
      "theoretical_loss": 4.176554769150796,
      "tokens_seen": 282132480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006358189899754115,
      "loss": 3.0592,
      "theoretical_loss": 4.176439816720697,
      "tokens_seen": 282198016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006358024399470399,
      "loss": 2.9379,
      "theoretical_loss": 4.1763248984562376,
      "tokens_seen": 282263552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006357858899186684,
      "loss": 2.8641,
      "theoretical_loss": 4.176210014339335,
      "tokens_seen": 282329088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000635769339890297,
      "loss": 3.0126,
      "theoretical_loss": 4.17609516435192,
      "tokens_seen": 282394624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006357527898619255,
      "loss": 2.7685,
      "theoretical_loss": 4.1759803484759335,
      "tokens_seen": 282460160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000635736239833554,
      "loss": 2.9765,
      "theoretical_loss": 4.175865566693336,
      "tokens_seen": 282525696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006357196898051826,
      "loss": 2.909,
      "theoretical_loss": 4.175750818986098,
      "tokens_seen": 282591232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000635703139776811,
      "loss": 3.0467,
      "theoretical_loss": 4.1756361053362046,
      "tokens_seen": 282656768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006356865897484396,
      "loss": 3.0062,
      "theoretical_loss": 4.1755214257256545,
      "tokens_seen": 282722304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006356700397200681,
      "loss": 3.0777,
      "theoretical_loss": 4.17540678013646,
      "tokens_seen": 282787840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006356534896916966,
      "loss": 2.8759,
      "theoretical_loss": 4.175292168550648,
      "tokens_seen": 282853376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006356369396633252,
      "loss": 2.922,
      "theoretical_loss": 4.175177590950257,
      "tokens_seen": 282918912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006356203896349537,
      "loss": 3.0646,
      "theoretical_loss": 4.175063047317342,
      "tokens_seen": 282984448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006356038396065822,
      "loss": 3.1201,
      "theoretical_loss": 4.174948537633968,
      "tokens_seen": 283049984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006355872895782107,
      "loss": 2.8971,
      "theoretical_loss": 4.174834061882218,
      "tokens_seen": 283115520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006355707395498391,
      "loss": 3.0867,
      "theoretical_loss": 4.1747196200441845,
      "tokens_seen": 283181056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006355541895214677,
      "loss": 2.7788,
      "theoretical_loss": 4.174605212101977,
      "tokens_seen": 283246592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006355376394930962,
      "loss": 3.0352,
      "theoretical_loss": 4.174490838037716,
      "tokens_seen": 283312128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006355210894647248,
      "loss": 3.1896,
      "theoretical_loss": 4.174376497833537,
      "tokens_seen": 283377664
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006355045394363533,
      "loss": 2.943,
      "theoretical_loss": 4.174262191471587,
      "tokens_seen": 283443200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006354879894079818,
      "loss": 3.0801,
      "theoretical_loss": 4.17414791893403,
      "tokens_seen": 283508736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006354714393796104,
      "loss": 2.8049,
      "theoretical_loss": 4.17403368020304,
      "tokens_seen": 283574272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006354548893512388,
      "loss": 3.1551,
      "theoretical_loss": 4.173919475260808,
      "tokens_seen": 283639808
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006354383393228673,
      "loss": 2.9584,
      "theoretical_loss": 4.173805304089536,
      "tokens_seen": 283705344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006354217892944959,
      "loss": 3.1921,
      "theoretical_loss": 4.173691166671439,
      "tokens_seen": 283770880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006354052392661244,
      "loss": 2.9897,
      "theoretical_loss": 4.173577062988748,
      "tokens_seen": 283836416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000635388689237753,
      "loss": 2.9368,
      "theoretical_loss": 4.173462993023706,
      "tokens_seen": 283901952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006353721392093815,
      "loss": 2.8565,
      "theoretical_loss": 4.173348956758568,
      "tokens_seen": 283967488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00063535558918101,
      "loss": 3.0626,
      "theoretical_loss": 4.173234954175605,
      "tokens_seen": 284033024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006353390391526386,
      "loss": 3.0424,
      "theoretical_loss": 4.173120985257102,
      "tokens_seen": 284098560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000635322489124267,
      "loss": 3.1132,
      "theoretical_loss": 4.173007049985352,
      "tokens_seen": 284164096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006353059390958955,
      "loss": 2.9565,
      "theoretical_loss": 4.172893148342667,
      "tokens_seen": 284229632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006352893890675241,
      "loss": 3.0962,
      "theoretical_loss": 4.172779280311372,
      "tokens_seen": 284295168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006352728390391526,
      "loss": 3.1541,
      "theoretical_loss": 4.172665445873801,
      "tokens_seen": 284360704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006352562890107812,
      "loss": 3.129,
      "theoretical_loss": 4.172551645012307,
      "tokens_seen": 284426240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006352397389824097,
      "loss": 2.521,
      "theoretical_loss": 4.1724378777092515,
      "tokens_seen": 284491776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006352231889540382,
      "loss": 2.9811,
      "theoretical_loss": 4.172324143947012,
      "tokens_seen": 284557312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006352066389256667,
      "loss": 2.9814,
      "theoretical_loss": 4.172210443707979,
      "tokens_seen": 284622848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006351900888972952,
      "loss": 2.9334,
      "theoretical_loss": 4.1720967769745565,
      "tokens_seen": 284688384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006351735388689237,
      "loss": 2.8548,
      "theoretical_loss": 4.171983143729159,
      "tokens_seen": 284753920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006351569888405523,
      "loss": 2.8333,
      "theoretical_loss": 4.1718695439542195,
      "tokens_seen": 284819456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006351404388121808,
      "loss": 3.3152,
      "theoretical_loss": 4.17175597763218,
      "tokens_seen": 284884992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006351238887838094,
      "loss": 3.0293,
      "theoretical_loss": 4.171642444745497,
      "tokens_seen": 284950528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006351073387554379,
      "loss": 2.8804,
      "theoretical_loss": 4.1715289452766395,
      "tokens_seen": 285016064
    },
    {
      "epoch": 0.1,
      "objective/train/avg_token_score": 0.007366314064711332,
      "objective/train/avg_weight": 0.991330623626709,
      "objective/train/docs_used": 172212,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1810994148254395,
      "objective/train/std_weight": 0.055957991629838943,
      "objective/train/theoretical_loss": 4.1714154792080915,
      "objective/train/tokens_used": 305541600,
      "theoretical_loss": 4.1714154792080915,
      "tokens_seen": 285081600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006350907887270663,
      "loss": 2.9729,
      "theoretical_loss": 4.1714154792080915,
      "tokens_seen": 285081600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006350742386986949,
      "loss": 3.0319,
      "theoretical_loss": 4.171302046522349,
      "tokens_seen": 285147136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006350576886703234,
      "loss": 3.1945,
      "theoretical_loss": 4.171188647201921,
      "tokens_seen": 285212672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006350411386419519,
      "loss": 3.0985,
      "theoretical_loss": 4.1710752812293315,
      "tokens_seen": 285278208
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006350245886135805,
      "loss": 3.0272,
      "theoretical_loss": 4.170961948587115,
      "tokens_seen": 285343744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000635008038585209,
      "loss": 2.9622,
      "theoretical_loss": 4.17084864925782,
      "tokens_seen": 285409280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006349914885568376,
      "loss": 3.0703,
      "theoretical_loss": 4.1707353832240095,
      "tokens_seen": 285474816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000634974938528466,
      "loss": 3.0925,
      "theoretical_loss": 4.170622150468258,
      "tokens_seen": 285540352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006349583885000945,
      "loss": 2.9003,
      "theoretical_loss": 4.170508950973154,
      "tokens_seen": 285605888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006349418384717231,
      "loss": 2.9662,
      "theoretical_loss": 4.1703957847213,
      "tokens_seen": 285671424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006349252884433516,
      "loss": 3.0732,
      "theoretical_loss": 4.170282651695308,
      "tokens_seen": 285736960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006349087384149802,
      "loss": 2.9117,
      "theoretical_loss": 4.170169551877808,
      "tokens_seen": 285802496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006348921883866087,
      "loss": 3.197,
      "theoretical_loss": 4.170056485251439,
      "tokens_seen": 285868032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006348756383582372,
      "loss": 3.1661,
      "theoretical_loss": 4.169943451798856,
      "tokens_seen": 285933568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006348590883298658,
      "loss": 2.807,
      "theoretical_loss": 4.169830451502724,
      "tokens_seen": 285999104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006348425383014942,
      "loss": 2.6778,
      "theoretical_loss": 4.169717484345725,
      "tokens_seen": 286064640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006348259882731227,
      "loss": 2.8852,
      "theoretical_loss": 4.1696045503105506,
      "tokens_seen": 286130176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006348094382447513,
      "loss": 3.2044,
      "theoretical_loss": 4.169491649379905,
      "tokens_seen": 286195712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006347928882163798,
      "loss": 3.0204,
      "theoretical_loss": 4.169378781536509,
      "tokens_seen": 286261248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006347763381880084,
      "loss": 2.9219,
      "theoretical_loss": 4.169265946763095,
      "tokens_seen": 286326784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006347597881596369,
      "loss": 2.7451,
      "theoretical_loss": 4.169153145042405,
      "tokens_seen": 286392320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006347432381312654,
      "loss": 3.0713,
      "theoretical_loss": 4.169040376357199,
      "tokens_seen": 286457856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000634726688102894,
      "loss": 3.1268,
      "theoretical_loss": 4.168927640690246,
      "tokens_seen": 286523392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006347101380745224,
      "loss": 3.2598,
      "theoretical_loss": 4.16881493802433,
      "tokens_seen": 286588928
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006346935880461509,
      "loss": 2.9384,
      "theoretical_loss": 4.168702268342248,
      "tokens_seen": 286654464
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006346770380177795,
      "loss": 2.8524,
      "theoretical_loss": 4.168589631626808,
      "tokens_seen": 286720000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000634660487989408,
      "loss": 3.2622,
      "theoretical_loss": 4.168477027860833,
      "tokens_seen": 286785536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006346439379610366,
      "loss": 2.9277,
      "theoretical_loss": 4.168364457027158,
      "tokens_seen": 286851072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006346273879326651,
      "loss": 2.974,
      "theoretical_loss": 4.168251919108632,
      "tokens_seen": 286916608
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006346108379042936,
      "loss": 2.7533,
      "theoretical_loss": 4.168139414088113,
      "tokens_seen": 286982144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000634594287875922,
      "loss": 3.0569,
      "theoretical_loss": 4.168026941948478,
      "tokens_seen": 287047680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006345777378475505,
      "loss": 3.187,
      "theoretical_loss": 4.167914502672611,
      "tokens_seen": 287113216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006345611878191791,
      "loss": 3.0635,
      "theoretical_loss": 4.1678020962434115,
      "tokens_seen": 287178752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006345446377908076,
      "loss": 2.9791,
      "theoretical_loss": 4.167689722643792,
      "tokens_seen": 287244288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006345280877624361,
      "loss": 3.1326,
      "theoretical_loss": 4.1675773818566775,
      "tokens_seen": 287309824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006345115377340647,
      "loss": 2.9617,
      "theoretical_loss": 4.167465073865006,
      "tokens_seen": 287375360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006344949877056932,
      "loss": 3.0207,
      "theoretical_loss": 4.167352798651726,
      "tokens_seen": 287440896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006344784376773217,
      "loss": 2.9403,
      "theoretical_loss": 4.167240556199802,
      "tokens_seen": 287506432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006344618876489502,
      "loss": 2.9474,
      "theoretical_loss": 4.167128346492211,
      "tokens_seen": 287571968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006344453376205787,
      "loss": 3.0457,
      "theoretical_loss": 4.16701616951194,
      "tokens_seen": 287637504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006344287875922073,
      "loss": 3.0834,
      "theoretical_loss": 4.1669040252419896,
      "tokens_seen": 287703040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006344122375638358,
      "loss": 2.9007,
      "theoretical_loss": 4.166791913665375,
      "tokens_seen": 287768576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006343956875354643,
      "loss": 3.1569,
      "theoretical_loss": 4.166679834765123,
      "tokens_seen": 287834112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006343791375070929,
      "loss": 3.0703,
      "theoretical_loss": 4.166567788524272,
      "tokens_seen": 287899648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006343625874787213,
      "loss": 2.9765,
      "theoretical_loss": 4.166455774925875,
      "tokens_seen": 287965184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006343460374503499,
      "loss": 2.9319,
      "theoretical_loss": 4.166343793952995,
      "tokens_seen": 288030720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006343294874219784,
      "loss": 3.0432,
      "theoretical_loss": 4.166231845588712,
      "tokens_seen": 288096256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006343129373936069,
      "loss": 3.2125,
      "theoretical_loss": 4.166119929816113,
      "tokens_seen": 288161792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006342963873652355,
      "loss": 2.8947,
      "theoretical_loss": 4.166008046618303,
      "tokens_seen": 288227328
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000634279837336864,
      "loss": 2.7974,
      "theoretical_loss": 4.1658961959783944,
      "tokens_seen": 288292864
    },
    {
      "epoch": 0.1,
      "objective/train/avg_token_score": 0.007860920391976833,
      "objective/train/avg_weight": 0.9907665848731995,
      "objective/train/docs_used": 173965,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.190627098083496,
      "objective/train/std_weight": 0.05251345783472061,
      "objective/train/theoretical_loss": 4.165784377879517,
      "objective/train/tokens_used": 308818400,
      "theoretical_loss": 4.165784377879517,
      "tokens_seen": 288358400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006342632873084925,
      "loss": 2.9741,
      "theoretical_loss": 4.165784377879517,
      "tokens_seen": 288358400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006342467372801211,
      "loss": 3.0224,
      "theoretical_loss": 4.165672592304811,
      "tokens_seen": 288423936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006342301872517495,
      "loss": 2.9392,
      "theoretical_loss": 4.165560839237429,
      "tokens_seen": 288489472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006342136372233781,
      "loss": 2.8423,
      "theoretical_loss": 4.165449118660536,
      "tokens_seen": 288555008
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006341970871950066,
      "loss": 3.1182,
      "theoretical_loss": 4.16533743055731,
      "tokens_seen": 288620544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006341805371666351,
      "loss": 3.1705,
      "theoretical_loss": 4.165225774910941,
      "tokens_seen": 288686080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006341639871382637,
      "loss": 2.7672,
      "theoretical_loss": 4.165114151704634,
      "tokens_seen": 288751616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006341474371098922,
      "loss": 2.9541,
      "theoretical_loss": 4.165002560921601,
      "tokens_seen": 288817152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006341308870815207,
      "loss": 2.97,
      "theoretical_loss": 4.164891002545073,
      "tokens_seen": 288882688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006341143370531492,
      "loss": 2.8309,
      "theoretical_loss": 4.16477947655829,
      "tokens_seen": 288948224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006340977870247777,
      "loss": 3.206,
      "theoretical_loss": 4.164667982944504,
      "tokens_seen": 289013760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006340812369964063,
      "loss": 3.1287,
      "theoretical_loss": 4.164556521686981,
      "tokens_seen": 289079296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006340646869680348,
      "loss": 3.1504,
      "theoretical_loss": 4.1644450927689975,
      "tokens_seen": 289144832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006340481369396633,
      "loss": 2.9001,
      "theoretical_loss": 4.164333696173846,
      "tokens_seen": 289210368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006340315869112919,
      "loss": 2.8709,
      "theoretical_loss": 4.164222331884827,
      "tokens_seen": 289275904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006340150368829204,
      "loss": 2.7433,
      "theoretical_loss": 4.164110999885256,
      "tokens_seen": 289341440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006339984868545488,
      "loss": 2.9315,
      "theoretical_loss": 4.163999700158462,
      "tokens_seen": 289406976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006339819368261774,
      "loss": 2.8943,
      "theoretical_loss": 4.163888432687784,
      "tokens_seen": 289472512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006339653867978059,
      "loss": 3.3278,
      "theoretical_loss": 4.163777197456573,
      "tokens_seen": 289538048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006339488367694345,
      "loss": 2.7403,
      "theoretical_loss": 4.163665994448197,
      "tokens_seen": 289603584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000633932286741063,
      "loss": 2.8423,
      "theoretical_loss": 4.163554823646027,
      "tokens_seen": 289669120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006339157367126915,
      "loss": 3.0333,
      "theoretical_loss": 4.163443685033458,
      "tokens_seen": 289734656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006338991866843201,
      "loss": 2.9428,
      "theoretical_loss": 4.163332578593889,
      "tokens_seen": 289800192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006338826366559486,
      "loss": 3.086,
      "theoretical_loss": 4.163221504310734,
      "tokens_seen": 289865728
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000633866086627577,
      "loss": 2.8916,
      "theoretical_loss": 4.1631104621674195,
      "tokens_seen": 289931264
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006338495365992056,
      "loss": 2.9221,
      "theoretical_loss": 4.162999452147384,
      "tokens_seen": 289996800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006338329865708341,
      "loss": 2.9745,
      "theoretical_loss": 4.1628884742340775,
      "tokens_seen": 290062336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006338164365424627,
      "loss": 2.7603,
      "theoretical_loss": 4.162777528410963,
      "tokens_seen": 290127872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006337998865140912,
      "loss": 2.9545,
      "theoretical_loss": 4.162666614661518,
      "tokens_seen": 290193408
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006337833364857197,
      "loss": 2.7561,
      "theoretical_loss": 4.162555732969227,
      "tokens_seen": 290258944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006337667864573483,
      "loss": 2.7046,
      "theoretical_loss": 4.162444883317591,
      "tokens_seen": 290324480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006337502364289767,
      "loss": 3.0028,
      "theoretical_loss": 4.162334065690123,
      "tokens_seen": 290390016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006337336864006052,
      "loss": 3.2178,
      "theoretical_loss": 4.162223280070345,
      "tokens_seen": 290455552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006337171363722338,
      "loss": 2.7209,
      "theoretical_loss": 4.1621125264417955,
      "tokens_seen": 290521088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006337005863438623,
      "loss": 2.9559,
      "theoretical_loss": 4.162001804788021,
      "tokens_seen": 290586624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006336840363154909,
      "loss": 3.1625,
      "theoretical_loss": 4.161891115092583,
      "tokens_seen": 290652160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006336674862871194,
      "loss": 2.9053,
      "theoretical_loss": 4.161780457339055,
      "tokens_seen": 290717696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006336509362587479,
      "loss": 3.1764,
      "theoretical_loss": 4.161669831511022,
      "tokens_seen": 290783232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006336343862303763,
      "loss": 3.0822,
      "theoretical_loss": 4.16155923759208,
      "tokens_seen": 290848768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006336178362020048,
      "loss": 2.9363,
      "theoretical_loss": 4.161448675565838,
      "tokens_seen": 290914304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006336012861736334,
      "loss": 2.7644,
      "theoretical_loss": 4.161338145415918,
      "tokens_seen": 290979840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006335847361452619,
      "loss": 2.7638,
      "theoretical_loss": 4.161227647125955,
      "tokens_seen": 291045376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006335681861168905,
      "loss": 3.1342,
      "theoretical_loss": 4.161117180679591,
      "tokens_seen": 291110912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000633551636088519,
      "loss": 2.7969,
      "theoretical_loss": 4.161006746060488,
      "tokens_seen": 291176448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006335350860601475,
      "loss": 3.0723,
      "theoretical_loss": 4.160896343252311,
      "tokens_seen": 291241984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006335185360317761,
      "loss": 3.1953,
      "theoretical_loss": 4.160785972238745,
      "tokens_seen": 291307520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006335019860034045,
      "loss": 3.1007,
      "theoretical_loss": 4.160675633003484,
      "tokens_seen": 291373056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000633485435975033,
      "loss": 3.1575,
      "theoretical_loss": 4.16056532553023,
      "tokens_seen": 291438592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006334688859466616,
      "loss": 2.8404,
      "theoretical_loss": 4.160455049802706,
      "tokens_seen": 291504128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006334523359182901,
      "loss": 3.087,
      "theoretical_loss": 4.1603448058046375,
      "tokens_seen": 291569664
    },
    {
      "epoch": 0.1,
      "objective/train/avg_token_score": 0.014044761657714844,
      "objective/train/avg_weight": 0.9834749698638916,
      "objective/train/docs_used": 176091,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1842668056488037,
      "objective/train/std_weight": 0.07885220646858215,
      "objective/train/theoretical_loss": 4.160234593519768,
      "objective/train/tokens_used": 312095200,
      "theoretical_loss": 4.160234593519768,
      "tokens_seen": 291635200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006334357858899187,
      "loss": 3.1145,
      "theoretical_loss": 4.160234593519768,
      "tokens_seen": 291635200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006334192358615472,
      "loss": 3.2571,
      "theoretical_loss": 4.160124412931852,
      "tokens_seen": 291700736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006334026858331757,
      "loss": 3.1836,
      "theoretical_loss": 4.160014264024654,
      "tokens_seen": 291766272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006333861358048042,
      "loss": 2.7299,
      "theoretical_loss": 4.159904146781952,
      "tokens_seen": 291831808
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006333695857764327,
      "loss": 2.794,
      "theoretical_loss": 4.159794061187536,
      "tokens_seen": 291897344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006333530357480612,
      "loss": 3.2368,
      "theoretical_loss": 4.1596840072252075,
      "tokens_seen": 291962880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006333364857196898,
      "loss": 2.9652,
      "theoretical_loss": 4.159573984878779,
      "tokens_seen": 292028416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006333199356913183,
      "loss": 2.9841,
      "theoretical_loss": 4.159463994132079,
      "tokens_seen": 292093952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006333033856629469,
      "loss": 3.0366,
      "theoretical_loss": 4.15935403496894,
      "tokens_seen": 292159488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006332868356345754,
      "loss": 3.0816,
      "theoretical_loss": 4.159244107373215,
      "tokens_seen": 292225024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006332702856062038,
      "loss": 3.0643,
      "theoretical_loss": 4.159134211328765,
      "tokens_seen": 292290560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006332537355778324,
      "loss": 2.9952,
      "theoretical_loss": 4.159024346819461,
      "tokens_seen": 292356096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006332371855494609,
      "loss": 3.123,
      "theoretical_loss": 4.158914513829189,
      "tokens_seen": 292421632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006332206355210894,
      "loss": 3.0764,
      "theoretical_loss": 4.158804712341845,
      "tokens_seen": 292487168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000633204085492718,
      "loss": 2.9089,
      "theoretical_loss": 4.158694942341338,
      "tokens_seen": 292552704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006331875354643465,
      "loss": 3.0635,
      "theoretical_loss": 4.1585852038115885,
      "tokens_seen": 292618240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006331709854359751,
      "loss": 2.6435,
      "theoretical_loss": 4.1584754967365285,
      "tokens_seen": 292683776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006331544354076036,
      "loss": 3.3446,
      "theoretical_loss": 4.1583658211001016,
      "tokens_seen": 292749312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000633137885379232,
      "loss": 3.1104,
      "theoretical_loss": 4.158256176886264,
      "tokens_seen": 292814848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006331213353508606,
      "loss": 2.9893,
      "theoretical_loss": 4.158146564078982,
      "tokens_seen": 292880384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006331047853224891,
      "loss": 2.9911,
      "theoretical_loss": 4.158036982662237,
      "tokens_seen": 292945920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006330882352941176,
      "loss": 3.1734,
      "theoretical_loss": 4.157927432620018,
      "tokens_seen": 293011456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006330716852657462,
      "loss": 3.0761,
      "theoretical_loss": 4.157817913936329,
      "tokens_seen": 293076992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006330551352373747,
      "loss": 2.8638,
      "theoretical_loss": 4.157708426595184,
      "tokens_seen": 293142528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006330385852090033,
      "loss": 3.0032,
      "theoretical_loss": 4.157598970580608,
      "tokens_seen": 293208064
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006330220351806317,
      "loss": 2.7573,
      "theoretical_loss": 4.157489545876642,
      "tokens_seen": 293273600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006330054851522602,
      "loss": 3.1682,
      "theoretical_loss": 4.157380152467333,
      "tokens_seen": 293339136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006329889351238888,
      "loss": 3.1287,
      "theoretical_loss": 4.157270790336742,
      "tokens_seen": 293404672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006329723850955173,
      "loss": 3.1321,
      "theoretical_loss": 4.157161459468944,
      "tokens_seen": 293470208
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006329558350671458,
      "loss": 2.8061,
      "theoretical_loss": 4.157052159848023,
      "tokens_seen": 293535744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006329392850387744,
      "loss": 2.8451,
      "theoretical_loss": 4.156942891458074,
      "tokens_seen": 293601280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006329227350104029,
      "loss": 2.7698,
      "theoretical_loss": 4.156833654283207,
      "tokens_seen": 293666816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006329061849820315,
      "loss": 2.8143,
      "theoretical_loss": 4.15672444830754,
      "tokens_seen": 293732352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006328896349536599,
      "loss": 3.0111,
      "theoretical_loss": 4.156615273515205,
      "tokens_seen": 293797888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006328730849252884,
      "loss": 3.0893,
      "theoretical_loss": 4.156506129890344,
      "tokens_seen": 293863424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000632856534896917,
      "loss": 2.7498,
      "theoretical_loss": 4.156397017417111,
      "tokens_seen": 293928960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0006328399848685455,
      "loss": 3.0828,
      "theoretical_loss": 4.156287936079675,
      "tokens_seen": 293994496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000632823434840174,
      "loss": 2.8662,
      "theoretical_loss": 4.156178885862209,
      "tokens_seen": 294060032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006328068848118026,
      "loss": 3.0874,
      "theoretical_loss": 4.156069866748906,
      "tokens_seen": 294125568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006327903347834311,
      "loss": 3.0512,
      "theoretical_loss": 4.155960878723965,
      "tokens_seen": 294191104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006327737847550596,
      "loss": 3.189,
      "theoretical_loss": 4.155851921771598,
      "tokens_seen": 294256640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006327572347266881,
      "loss": 3.0354,
      "theoretical_loss": 4.155742995876029,
      "tokens_seen": 294322176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006327406846983166,
      "loss": 2.8944,
      "theoretical_loss": 4.155634101021494,
      "tokens_seen": 294387712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006327241346699452,
      "loss": 2.9253,
      "theoretical_loss": 4.155525237192238,
      "tokens_seen": 294453248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006327075846415737,
      "loss": 2.9474,
      "theoretical_loss": 4.155416404372522,
      "tokens_seen": 294518784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006326910346132022,
      "loss": 2.9908,
      "theoretical_loss": 4.155307602546614,
      "tokens_seen": 294584320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006326744845848308,
      "loss": 2.9238,
      "theoretical_loss": 4.155198831698795,
      "tokens_seen": 294649856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006326579345564592,
      "loss": 2.9602,
      "theoretical_loss": 4.155090091813358,
      "tokens_seen": 294715392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006326413845280877,
      "loss": 2.7735,
      "theoretical_loss": 4.154981382874608,
      "tokens_seen": 294780928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006326248344997162,
      "loss": 3.0825,
      "theoretical_loss": 4.154872704866859,
      "tokens_seen": 294846464
    },
    {
      "epoch": 0.11,
      "objective/train/avg_token_score": 0.008944209665060043,
      "objective/train/avg_weight": 0.9894801378250122,
      "objective/train/docs_used": 178137,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.963225841522217,
      "objective/train/std_weight": 0.054859012365341187,
      "objective/train/theoretical_loss": 4.15476405777444,
      "objective/train/tokens_used": 315372000,
      "theoretical_loss": 4.15476405777444,
      "tokens_seen": 294912000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006326082844713448,
      "loss": 2.9265,
      "theoretical_loss": 4.15476405777444,
      "tokens_seen": 294912000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006325917344429733,
      "loss": 3.094,
      "theoretical_loss": 4.154655441581687,
      "tokens_seen": 294977536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006325751844146018,
      "loss": 2.6984,
      "theoretical_loss": 4.154546856272952,
      "tokens_seen": 295043072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006325586343862304,
      "loss": 3.0442,
      "theoretical_loss": 4.154438301832596,
      "tokens_seen": 295108608
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006325420843578588,
      "loss": 2.7001,
      "theoretical_loss": 4.154329778244991,
      "tokens_seen": 295174144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006325255343294873,
      "loss": 2.9914,
      "theoretical_loss": 4.154221285494521,
      "tokens_seen": 295239680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006325089843011159,
      "loss": 2.9701,
      "theoretical_loss": 4.154112823565582,
      "tokens_seen": 295305216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006324924342727444,
      "loss": 3.028,
      "theoretical_loss": 4.15400439244258,
      "tokens_seen": 295370752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000632475884244373,
      "loss": 2.9858,
      "theoretical_loss": 4.153895992109935,
      "tokens_seen": 295436288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006324593342160015,
      "loss": 2.8457,
      "theoretical_loss": 4.153787622552073,
      "tokens_seen": 295501824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00063244278418763,
      "loss": 3.1612,
      "theoretical_loss": 4.153679283753439,
      "tokens_seen": 295567360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006324262341592586,
      "loss": 3.0203,
      "theoretical_loss": 4.15357097569848,
      "tokens_seen": 295632896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000632409684130887,
      "loss": 3.0853,
      "theoretical_loss": 4.153462698371665,
      "tokens_seen": 295698432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006323931341025156,
      "loss": 2.9932,
      "theoretical_loss": 4.1533544517574645,
      "tokens_seen": 295763968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006323765840741441,
      "loss": 3.0949,
      "theoretical_loss": 4.153246235840367,
      "tokens_seen": 295829504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006323600340457726,
      "loss": 2.9819,
      "theoretical_loss": 4.153138050604868,
      "tokens_seen": 295895040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006323434840174012,
      "loss": 2.9602,
      "theoretical_loss": 4.153029896035476,
      "tokens_seen": 295960576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006323269339890297,
      "loss": 2.948,
      "theoretical_loss": 4.152921772116712,
      "tokens_seen": 296026112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006323103839606582,
      "loss": 2.9623,
      "theoretical_loss": 4.152813678833106,
      "tokens_seen": 296091648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006322938339322867,
      "loss": 2.825,
      "theoretical_loss": 4.152705616169202,
      "tokens_seen": 296157184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006322772839039152,
      "loss": 3.1309,
      "theoretical_loss": 4.15259758410955,
      "tokens_seen": 296222720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006322607338755438,
      "loss": 3.0093,
      "theoretical_loss": 4.152489582638719,
      "tokens_seen": 296288256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006322441838471723,
      "loss": 3.0254,
      "theoretical_loss": 4.152381611741281,
      "tokens_seen": 296353792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006322276338188008,
      "loss": 3.1127,
      "theoretical_loss": 4.152273671401824,
      "tokens_seen": 296419328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006322110837904294,
      "loss": 2.8707,
      "theoretical_loss": 4.152165761604948,
      "tokens_seen": 296484864
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006321945337620579,
      "loss": 2.9257,
      "theoretical_loss": 4.152057882335261,
      "tokens_seen": 296550400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006321779837336864,
      "loss": 2.9009,
      "theoretical_loss": 4.151950033577383,
      "tokens_seen": 296615936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006321614337053149,
      "loss": 3.0589,
      "theoretical_loss": 4.151842215315947,
      "tokens_seen": 296681472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006321448836769434,
      "loss": 2.9613,
      "theoretical_loss": 4.151734427535594,
      "tokens_seen": 296747008
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000632128333648572,
      "loss": 2.7652,
      "theoretical_loss": 4.151626670220979,
      "tokens_seen": 296812544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006321117836202005,
      "loss": 3.1652,
      "theoretical_loss": 4.151518943356768,
      "tokens_seen": 296878080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000632095233591829,
      "loss": 2.8519,
      "theoretical_loss": 4.151411246927636,
      "tokens_seen": 296943616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006320786835634576,
      "loss": 3.0106,
      "theoretical_loss": 4.15130358091827,
      "tokens_seen": 297009152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006320621335350861,
      "loss": 2.8805,
      "theoretical_loss": 4.151195945313369,
      "tokens_seen": 297074688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006320455835067145,
      "loss": 3.0284,
      "theoretical_loss": 4.151088340097642,
      "tokens_seen": 297140224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006320290334783431,
      "loss": 3.072,
      "theoretical_loss": 4.15098076525581,
      "tokens_seen": 297205760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006320124834499716,
      "loss": 2.9996,
      "theoretical_loss": 4.150873220772604,
      "tokens_seen": 297271296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006319959334216002,
      "loss": 2.9728,
      "theoretical_loss": 4.150765706632766,
      "tokens_seen": 297336832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006319793833932287,
      "loss": 2.8555,
      "theoretical_loss": 4.1506582228210505,
      "tokens_seen": 297402368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006319628333648572,
      "loss": 2.9493,
      "theoretical_loss": 4.150550769322221,
      "tokens_seen": 297467904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006319462833364858,
      "loss": 2.9091,
      "theoretical_loss": 4.150443346121054,
      "tokens_seen": 297533440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006319297333081143,
      "loss": 3.0988,
      "theoretical_loss": 4.150335953202336,
      "tokens_seen": 297598976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006319131832797427,
      "loss": 2.9347,
      "theoretical_loss": 4.150228590550864,
      "tokens_seen": 297664512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006318966332513713,
      "loss": 2.9117,
      "theoretical_loss": 4.150121258151447,
      "tokens_seen": 297730048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006318800832229998,
      "loss": 2.962,
      "theoretical_loss": 4.150013955988905,
      "tokens_seen": 297795584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006318635331946284,
      "loss": 2.8372,
      "theoretical_loss": 4.149906684048068,
      "tokens_seen": 297861120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006318469831662569,
      "loss": 3.0417,
      "theoretical_loss": 4.1497994423137765,
      "tokens_seen": 297926656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006318304331378854,
      "loss": 2.8536,
      "theoretical_loss": 4.149692230770884,
      "tokens_seen": 297992192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000631813883109514,
      "loss": 2.9618,
      "theoretical_loss": 4.149585049404253,
      "tokens_seen": 298057728
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006317973330811424,
      "loss": 2.9654,
      "theoretical_loss": 4.149477898198759,
      "tokens_seen": 298123264
    },
    {
      "epoch": 0.11,
      "objective/train/avg_token_score": 0.01585209183394909,
      "objective/train/avg_weight": 0.9813624024391174,
      "objective/train/docs_used": 180121,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1811931133270264,
      "objective/train/std_weight": 0.10468541830778122,
      "objective/train/theoretical_loss": 4.149370777139286,
      "objective/train/tokens_used": 318648800,
      "theoretical_loss": 4.149370777139286,
      "tokens_seen": 298188800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006317807830527709,
      "loss": 3.075,
      "theoretical_loss": 4.149370777139286,
      "tokens_seen": 298188800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006317642330243995,
      "loss": 2.9869,
      "theoretical_loss": 4.14926368621073,
      "tokens_seen": 298254336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000631747682996028,
      "loss": 2.9203,
      "theoretical_loss": 4.149156625397998,
      "tokens_seen": 298319872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006317311329676566,
      "loss": 3.063,
      "theoretical_loss": 4.149049594686008,
      "tokens_seen": 298385408
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006317145829392851,
      "loss": 3.0243,
      "theoretical_loss": 4.1489425940596885,
      "tokens_seen": 298450944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006316980329109136,
      "loss": 2.6825,
      "theoretical_loss": 4.148835623503978,
      "tokens_seen": 298516480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006316814828825422,
      "loss": 2.9904,
      "theoretical_loss": 4.148728683003829,
      "tokens_seen": 298582016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006316649328541705,
      "loss": 3.0981,
      "theoretical_loss": 4.1486217725442005,
      "tokens_seen": 298647552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006316483828257991,
      "loss": 3.0891,
      "theoretical_loss": 4.148514892110065,
      "tokens_seen": 298713088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006316318327974276,
      "loss": 2.9372,
      "theoretical_loss": 4.148408041686406,
      "tokens_seen": 298778624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006316152827690561,
      "loss": 3.112,
      "theoretical_loss": 4.148301221258217,
      "tokens_seen": 298844160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006315987327406847,
      "loss": 2.9832,
      "theoretical_loss": 4.148194430810502,
      "tokens_seen": 298909696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006315821827123132,
      "loss": 3.0267,
      "theoretical_loss": 4.148087670328276,
      "tokens_seen": 298975232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006315656326839418,
      "loss": 3.105,
      "theoretical_loss": 4.147980939796565,
      "tokens_seen": 299040768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006315490826555702,
      "loss": 2.9993,
      "theoretical_loss": 4.147874239200405,
      "tokens_seen": 299106304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006315325326271987,
      "loss": 2.8588,
      "theoretical_loss": 4.147767568524845,
      "tokens_seen": 299171840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006315159825988273,
      "loss": 2.9511,
      "theoretical_loss": 4.147660927754942,
      "tokens_seen": 299237376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006314994325704558,
      "loss": 2.7855,
      "theoretical_loss": 4.147554316875766,
      "tokens_seen": 299302912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006314828825420843,
      "loss": 2.7728,
      "theoretical_loss": 4.147447735872396,
      "tokens_seen": 299368448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006314663325137129,
      "loss": 2.8035,
      "theoretical_loss": 4.147341184729921,
      "tokens_seen": 299433984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006314497824853414,
      "loss": 2.9767,
      "theoretical_loss": 4.147234663433444,
      "tokens_seen": 299499520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006314332324569699,
      "loss": 2.9234,
      "theoretical_loss": 4.147128171968077,
      "tokens_seen": 299565056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006314166824285984,
      "loss": 2.8025,
      "theoretical_loss": 4.14702171031894,
      "tokens_seen": 299630592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006314001324002269,
      "loss": 2.9135,
      "theoretical_loss": 4.146915278471169,
      "tokens_seen": 299696128
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006313835823718555,
      "loss": 3.0552,
      "theoretical_loss": 4.146808876409906,
      "tokens_seen": 299761664
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000631367032343484,
      "loss": 2.9085,
      "theoretical_loss": 4.146702504120305,
      "tokens_seen": 299827200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006313504823151125,
      "loss": 2.9191,
      "theoretical_loss": 4.146596161587532,
      "tokens_seen": 299892736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006313339322867411,
      "loss": 2.9342,
      "theoretical_loss": 4.146489848796763,
      "tokens_seen": 299958272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006313173822583695,
      "loss": 2.7548,
      "theoretical_loss": 4.146383565733184,
      "tokens_seen": 300023808
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006313008322299981,
      "loss": 3.1,
      "theoretical_loss": 4.146277312381991,
      "tokens_seen": 300089344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006312842822016266,
      "loss": 3.0072,
      "theoretical_loss": 4.1461710887283925,
      "tokens_seen": 300154880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006312677321732551,
      "loss": 3.1337,
      "theoretical_loss": 4.146064894757606,
      "tokens_seen": 300220416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006312511821448837,
      "loss": 3.0185,
      "theoretical_loss": 4.145958730454861,
      "tokens_seen": 300285952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006312346321165122,
      "loss": 3.064,
      "theoretical_loss": 4.145852595805396,
      "tokens_seen": 300351488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006312180820881407,
      "loss": 3.0406,
      "theoretical_loss": 4.145746490794461,
      "tokens_seen": 300417024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006312015320597693,
      "loss": 2.8979,
      "theoretical_loss": 4.145640415407317,
      "tokens_seen": 300482560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006311849820313977,
      "loss": 3.1235,
      "theoretical_loss": 4.145534369629234,
      "tokens_seen": 300548096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006311684320030263,
      "loss": 2.8838,
      "theoretical_loss": 4.145428353445494,
      "tokens_seen": 300613632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006311518819746548,
      "loss": 2.9453,
      "theoretical_loss": 4.145322366841389,
      "tokens_seen": 300679168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006311353319462833,
      "loss": 2.7989,
      "theoretical_loss": 4.145216409802221,
      "tokens_seen": 300744704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006311187819179119,
      "loss": 3.1166,
      "theoretical_loss": 4.145110482313304,
      "tokens_seen": 300810240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006311022318895404,
      "loss": 2.9317,
      "theoretical_loss": 4.1450045843599606,
      "tokens_seen": 300875776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006310856818611689,
      "loss": 3.015,
      "theoretical_loss": 4.144898715927525,
      "tokens_seen": 300941312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006310691318327974,
      "loss": 3.0441,
      "theoretical_loss": 4.144792877001342,
      "tokens_seen": 301006848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006310525818044259,
      "loss": 3.0764,
      "theoretical_loss": 4.144687067566765,
      "tokens_seen": 301072384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006310360317760545,
      "loss": 3.0131,
      "theoretical_loss": 4.144581287609161,
      "tokens_seen": 301137920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000631019481747683,
      "loss": 2.8792,
      "theoretical_loss": 4.144475537113905,
      "tokens_seen": 301203456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006310029317193115,
      "loss": 3.0519,
      "theoretical_loss": 4.144369816066385,
      "tokens_seen": 301268992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006309863816909401,
      "loss": 2.8746,
      "theoretical_loss": 4.144264124451995,
      "tokens_seen": 301334528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006309698316625686,
      "loss": 2.9514,
      "theoretical_loss": 4.1441584622561445,
      "tokens_seen": 301400064
    },
    {
      "epoch": 0.11,
      "objective/train/avg_token_score": 0.011133592575788498,
      "objective/train/avg_weight": 0.986901044845581,
      "objective/train/docs_used": 182298,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0074894428253174,
      "objective/train/std_weight": 0.08645734190940857,
      "objective/train/theoretical_loss": 4.144052829464249,
      "objective/train/tokens_used": 321925600,
      "theoretical_loss": 4.144052829464249,
      "tokens_seen": 301465600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006309532816341972,
      "loss": 2.8586,
      "theoretical_loss": 4.144052829464249,
      "tokens_seen": 301465600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006309367316058256,
      "loss": 2.851,
      "theoretical_loss": 4.143947226061737,
      "tokens_seen": 301531136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006309201815774541,
      "loss": 2.6221,
      "theoretical_loss": 4.143841652034048,
      "tokens_seen": 301596672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006309036315490827,
      "loss": 3.1702,
      "theoretical_loss": 4.143736107366629,
      "tokens_seen": 301662208
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006308870815207112,
      "loss": 2.756,
      "theoretical_loss": 4.14363059204494,
      "tokens_seen": 301727744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006308705314923397,
      "loss": 3.0586,
      "theoretical_loss": 4.14352510605445,
      "tokens_seen": 301793280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006308539814639683,
      "loss": 2.7128,
      "theoretical_loss": 4.143419649380639,
      "tokens_seen": 301858816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006308374314355968,
      "loss": 3.2144,
      "theoretical_loss": 4.143314222008997,
      "tokens_seen": 301924352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006308208814072253,
      "loss": 3.0389,
      "theoretical_loss": 4.143208823925024,
      "tokens_seen": 301989888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006308043313788538,
      "loss": 2.9003,
      "theoretical_loss": 4.143103455114231,
      "tokens_seen": 302055424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006307877813504823,
      "loss": 3.0506,
      "theoretical_loss": 4.142998115562139,
      "tokens_seen": 302120960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006307712313221109,
      "loss": 2.8995,
      "theoretical_loss": 4.14289280525428,
      "tokens_seen": 302186496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006307546812937394,
      "loss": 2.9156,
      "theoretical_loss": 4.142787524176194,
      "tokens_seen": 302252032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006307381312653679,
      "loss": 2.9099,
      "theoretical_loss": 4.142682272313435,
      "tokens_seen": 302317568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006307215812369965,
      "loss": 2.75,
      "theoretical_loss": 4.142577049651563,
      "tokens_seen": 302383104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006307050312086249,
      "loss": 2.8082,
      "theoretical_loss": 4.142471856176152,
      "tokens_seen": 302448640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006306884811802535,
      "loss": 2.8852,
      "theoretical_loss": 4.142366691872784,
      "tokens_seen": 302514176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006306719311518819,
      "loss": 2.9948,
      "theoretical_loss": 4.142261556727052,
      "tokens_seen": 302579712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006306553811235105,
      "loss": 2.9718,
      "theoretical_loss": 4.14215645072456,
      "tokens_seen": 302645248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000630638831095139,
      "loss": 2.9761,
      "theoretical_loss": 4.14205137385092,
      "tokens_seen": 302710784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006306222810667675,
      "loss": 2.6816,
      "theoretical_loss": 4.141946326091756,
      "tokens_seen": 302776320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006306057310383961,
      "loss": 2.9381,
      "theoretical_loss": 4.141841307432703,
      "tokens_seen": 302841856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006305891810100245,
      "loss": 2.9061,
      "theoretical_loss": 4.1417363178594035,
      "tokens_seen": 302907392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000630572630981653,
      "loss": 2.9579,
      "theoretical_loss": 4.141631357357513,
      "tokens_seen": 302972928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006305560809532816,
      "loss": 2.9148,
      "theoretical_loss": 4.141526425912694,
      "tokens_seen": 303038464
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006305395309249101,
      "loss": 3.0591,
      "theoretical_loss": 4.141421523510623,
      "tokens_seen": 303104000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006305229808965387,
      "loss": 2.7599,
      "theoretical_loss": 4.141316650136983,
      "tokens_seen": 303169536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006305064308681672,
      "loss": 2.8385,
      "theoretical_loss": 4.14121180577747,
      "tokens_seen": 303235072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006304898808397957,
      "loss": 3.0386,
      "theoretical_loss": 4.141106990417789,
      "tokens_seen": 303300608
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006304733308114243,
      "loss": 2.925,
      "theoretical_loss": 4.141002204043654,
      "tokens_seen": 303366144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006304567807830527,
      "loss": 3.0755,
      "theoretical_loss": 4.140897446640793,
      "tokens_seen": 303431680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006304402307546812,
      "loss": 3.0273,
      "theoretical_loss": 4.1407927181949375,
      "tokens_seen": 303497216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006304236807263098,
      "loss": 2.9718,
      "theoretical_loss": 4.140688018691835,
      "tokens_seen": 303562752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006304071306979383,
      "loss": 2.8941,
      "theoretical_loss": 4.140583348117241,
      "tokens_seen": 303628288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006303905806695669,
      "loss": 2.8848,
      "theoretical_loss": 4.140478706456921,
      "tokens_seen": 303693824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006303740306411954,
      "loss": 2.6684,
      "theoretical_loss": 4.140374093696651,
      "tokens_seen": 303759360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006303574806128239,
      "loss": 2.9055,
      "theoretical_loss": 4.1402695098222155,
      "tokens_seen": 303824896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006303409305844524,
      "loss": 2.9479,
      "theoretical_loss": 4.140164954819412,
      "tokens_seen": 303890432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006303243805560809,
      "loss": 3.1559,
      "theoretical_loss": 4.140060428674046,
      "tokens_seen": 303955968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006303078305277094,
      "loss": 3.0499,
      "theoretical_loss": 4.139955931371932,
      "tokens_seen": 304021504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000630291280499338,
      "loss": 3.038,
      "theoretical_loss": 4.139851462898897,
      "tokens_seen": 304087040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006302747304709665,
      "loss": 2.7309,
      "theoretical_loss": 4.139747023240777,
      "tokens_seen": 304152576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006302581804425951,
      "loss": 2.883,
      "theoretical_loss": 4.139642612383418,
      "tokens_seen": 304218112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006302416304142236,
      "loss": 3.1251,
      "theoretical_loss": 4.1395382303126755,
      "tokens_seen": 304283648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000630225080385852,
      "loss": 3.0238,
      "theoretical_loss": 4.139433877014415,
      "tokens_seen": 304349184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006302085303574806,
      "loss": 3.0178,
      "theoretical_loss": 4.139329552474514,
      "tokens_seen": 304414720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006301919803291091,
      "loss": 2.6784,
      "theoretical_loss": 4.139225256678857,
      "tokens_seen": 304480256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006301754303007376,
      "loss": 2.9758,
      "theoretical_loss": 4.139120989613341,
      "tokens_seen": 304545792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006301588802723662,
      "loss": 2.7868,
      "theoretical_loss": 4.1390167512638705,
      "tokens_seen": 304611328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006301423302439947,
      "loss": 2.9567,
      "theoretical_loss": 4.138912541616363,
      "tokens_seen": 304676864
    },
    {
      "epoch": 0.11,
      "objective/train/avg_token_score": 0.012638669461011887,
      "objective/train/avg_weight": 0.9851237535476685,
      "objective/train/docs_used": 183436,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.202432155609131,
      "objective/train/std_weight": 0.07855448871850967,
      "objective/train/theoretical_loss": 4.138808360656742,
      "objective/train/tokens_used": 325202400,
      "theoretical_loss": 4.138808360656742,
      "tokens_seen": 304742400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006301257802156233,
      "loss": 2.9231,
      "theoretical_loss": 4.138808360656742,
      "tokens_seen": 304742400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006301092301872518,
      "loss": 2.8528,
      "theoretical_loss": 4.138704208370944,
      "tokens_seen": 304807936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006300926801588802,
      "loss": 3.0493,
      "theoretical_loss": 4.138600084744915,
      "tokens_seen": 304873472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006300761301305088,
      "loss": 2.7771,
      "theoretical_loss": 4.1384959897646105,
      "tokens_seen": 304939008
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006300595801021373,
      "loss": 3.0609,
      "theoretical_loss": 4.138391923415996,
      "tokens_seen": 305004544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006300430300737658,
      "loss": 3.0906,
      "theoretical_loss": 4.138287885685045,
      "tokens_seen": 305070080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006300264800453944,
      "loss": 2.9482,
      "theoretical_loss": 4.138183876557745,
      "tokens_seen": 305135616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006300099300170229,
      "loss": 2.865,
      "theoretical_loss": 4.1380798960200895,
      "tokens_seen": 305201152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006299933799886515,
      "loss": 2.8263,
      "theoretical_loss": 4.137975944058083,
      "tokens_seen": 305266688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00062997682996028,
      "loss": 2.9431,
      "theoretical_loss": 4.137872020657742,
      "tokens_seen": 305332224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006299602799319084,
      "loss": 2.6944,
      "theoretical_loss": 4.1377681258050885,
      "tokens_seen": 305397760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000629943729903537,
      "loss": 2.7965,
      "theoretical_loss": 4.13766425948616,
      "tokens_seen": 305463296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006299271798751655,
      "loss": 3.0236,
      "theoretical_loss": 4.137560421686998,
      "tokens_seen": 305528832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000629910629846794,
      "loss": 2.92,
      "theoretical_loss": 4.137456612393658,
      "tokens_seen": 305594368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006298940798184226,
      "loss": 2.9466,
      "theoretical_loss": 4.137352831592203,
      "tokens_seen": 305659904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006298775297900511,
      "loss": 2.9624,
      "theoretical_loss": 4.137249079268707,
      "tokens_seen": 305725440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006298609797616797,
      "loss": 2.8241,
      "theoretical_loss": 4.137145355409253,
      "tokens_seen": 305790976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006298444297333081,
      "loss": 2.9991,
      "theoretical_loss": 4.137041659999936,
      "tokens_seen": 305856512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006298278797049366,
      "loss": 3.1769,
      "theoretical_loss": 4.136937993026857,
      "tokens_seen": 305922048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006298113296765652,
      "loss": 2.9186,
      "theoretical_loss": 4.136834354476129,
      "tokens_seen": 305987584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006297947796481937,
      "loss": 2.8932,
      "theoretical_loss": 4.1367307443338746,
      "tokens_seen": 306053120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006297782296198222,
      "loss": 2.8561,
      "theoretical_loss": 4.136627162586226,
      "tokens_seen": 306118656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006297616795914508,
      "loss": 2.8321,
      "theoretical_loss": 4.136523609219327,
      "tokens_seen": 306184192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006297451295630793,
      "loss": 3.0244,
      "theoretical_loss": 4.136420084219327,
      "tokens_seen": 306249728
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006297285795347078,
      "loss": 2.9489,
      "theoretical_loss": 4.136316587572388,
      "tokens_seen": 306315264
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006297120295063363,
      "loss": 2.8454,
      "theoretical_loss": 4.136213119264681,
      "tokens_seen": 306380800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006296954794779648,
      "loss": 2.8897,
      "theoretical_loss": 4.136109679282388,
      "tokens_seen": 306446336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006296789294495933,
      "loss": 2.9974,
      "theoretical_loss": 4.136006267611697,
      "tokens_seen": 306511872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006296623794212218,
      "loss": 3.0724,
      "theoretical_loss": 4.135902884238812,
      "tokens_seen": 306577408
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006296458293928504,
      "loss": 2.9839,
      "theoretical_loss": 4.135799529149939,
      "tokens_seen": 306642944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006296292793644789,
      "loss": 2.7623,
      "theoretical_loss": 4.1356962023313,
      "tokens_seen": 306708480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006296127293361074,
      "loss": 2.5293,
      "theoretical_loss": 4.135592903769124,
      "tokens_seen": 306774016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006295961793077359,
      "loss": 3.0962,
      "theoretical_loss": 4.135489633449649,
      "tokens_seen": 306839552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006295796292793644,
      "loss": 2.7656,
      "theoretical_loss": 4.135386391359123,
      "tokens_seen": 306905088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000629563079250993,
      "loss": 2.9181,
      "theoretical_loss": 4.135283177483807,
      "tokens_seen": 306970624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006295465292226215,
      "loss": 2.6723,
      "theoretical_loss": 4.135179991809965,
      "tokens_seen": 307036160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00062952997919425,
      "loss": 2.8999,
      "theoretical_loss": 4.135076834323876,
      "tokens_seen": 307101696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006295134291658786,
      "loss": 2.9264,
      "theoretical_loss": 4.134973705011828,
      "tokens_seen": 307167232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000629496879137507,
      "loss": 2.9816,
      "theoretical_loss": 4.134870603860117,
      "tokens_seen": 307232768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006294803291091356,
      "loss": 2.7217,
      "theoretical_loss": 4.134767530855047,
      "tokens_seen": 307298304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006294637790807641,
      "loss": 2.8832,
      "theoretical_loss": 4.1346644859829365,
      "tokens_seen": 307363840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006294472290523926,
      "loss": 3.1779,
      "theoretical_loss": 4.1345614692301105,
      "tokens_seen": 307429376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006294306790240212,
      "loss": 3.1238,
      "theoretical_loss": 4.134458480582902,
      "tokens_seen": 307494912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006294141289956497,
      "loss": 2.9418,
      "theoretical_loss": 4.134355520027657,
      "tokens_seen": 307560448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006293975789672782,
      "loss": 3.0734,
      "theoretical_loss": 4.134252587550728,
      "tokens_seen": 307625984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006293810289389068,
      "loss": 2.9896,
      "theoretical_loss": 4.134149683138481,
      "tokens_seen": 307691520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006293644789105352,
      "loss": 3.163,
      "theoretical_loss": 4.134046806777286,
      "tokens_seen": 307757056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006293479288821638,
      "loss": 2.8223,
      "theoretical_loss": 4.133943958453528,
      "tokens_seen": 307822592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006293313788537923,
      "loss": 2.9766,
      "theoretical_loss": 4.133841138153597,
      "tokens_seen": 307888128
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006293148288254208,
      "loss": 2.941,
      "theoretical_loss": 4.133738345863896,
      "tokens_seen": 307953664
    },
    {
      "epoch": 0.11,
      "objective/train/avg_token_score": 0.004725649952888489,
      "objective/train/avg_weight": 0.994437038898468,
      "objective/train/docs_used": 185431,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1532816886901855,
      "objective/train/std_weight": 0.033813364803791046,
      "objective/train/theoretical_loss": 4.133635581570836,
      "objective/train/tokens_used": 328479200,
      "theoretical_loss": 4.133635581570836,
      "tokens_seen": 308019200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006292982787970494,
      "loss": 2.9334,
      "theoretical_loss": 4.133635581570836,
      "tokens_seen": 308019200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006292817287686779,
      "loss": 2.9269,
      "theoretical_loss": 4.133532845260836,
      "tokens_seen": 308084736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006292651787403064,
      "loss": 2.9317,
      "theoretical_loss": 4.133430136920327,
      "tokens_seen": 308150272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000629248628711935,
      "loss": 2.8326,
      "theoretical_loss": 4.133327456535749,
      "tokens_seen": 308215808
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006292320786835634,
      "loss": 2.995,
      "theoretical_loss": 4.13322480409355,
      "tokens_seen": 308281344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000629215528655192,
      "loss": 3.1477,
      "theoretical_loss": 4.133122179580189,
      "tokens_seen": 308346880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006291989786268205,
      "loss": 2.7055,
      "theoretical_loss": 4.133019582982134,
      "tokens_seen": 308412416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000629182428598449,
      "loss": 2.9894,
      "theoretical_loss": 4.1329170142858604,
      "tokens_seen": 308477952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006291658785700776,
      "loss": 2.9278,
      "theoretical_loss": 4.132814473477857,
      "tokens_seen": 308543488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006291493285417061,
      "loss": 2.8395,
      "theoretical_loss": 4.1327119605446185,
      "tokens_seen": 308609024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006291327785133345,
      "loss": 2.9515,
      "theoretical_loss": 4.132609475472651,
      "tokens_seen": 308674560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006291162284849631,
      "loss": 3.2247,
      "theoretical_loss": 4.132507018248469,
      "tokens_seen": 308740096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006290996784565916,
      "loss": 2.9345,
      "theoretical_loss": 4.132404588858597,
      "tokens_seen": 308805632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006290831284282202,
      "loss": 2.9397,
      "theoretical_loss": 4.132302187289568,
      "tokens_seen": 308871168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006290665783998487,
      "loss": 2.8182,
      "theoretical_loss": 4.132199813527926,
      "tokens_seen": 308936704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006290500283714772,
      "loss": 2.9067,
      "theoretical_loss": 4.132097467560223,
      "tokens_seen": 309002240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006290334783431058,
      "loss": 2.8568,
      "theoretical_loss": 4.1319951493730205,
      "tokens_seen": 309067776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006290169283147343,
      "loss": 2.8611,
      "theoretical_loss": 4.131892858952889,
      "tokens_seen": 309133312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006290003782863627,
      "loss": 2.8981,
      "theoretical_loss": 4.131790596286409,
      "tokens_seen": 309198848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006289838282579913,
      "loss": 3.0376,
      "theoretical_loss": 4.1316883613601725,
      "tokens_seen": 309264384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006289672782296198,
      "loss": 2.8282,
      "theoretical_loss": 4.131586154160775,
      "tokens_seen": 309329920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006289507282012484,
      "loss": 2.952,
      "theoretical_loss": 4.131483974674827,
      "tokens_seen": 309395456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006289341781728769,
      "loss": 2.8053,
      "theoretical_loss": 4.131381822888946,
      "tokens_seen": 309460992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006289176281445054,
      "loss": 2.7996,
      "theoretical_loss": 4.131279698789759,
      "tokens_seen": 309526528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000628901078116134,
      "loss": 3.0475,
      "theoretical_loss": 4.1311776023639,
      "tokens_seen": 309592064
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006288845280877624,
      "loss": 2.9759,
      "theoretical_loss": 4.131075533598018,
      "tokens_seen": 309657600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006288679780593909,
      "loss": 2.9608,
      "theoretical_loss": 4.130973492478766,
      "tokens_seen": 309723136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006288514280310195,
      "loss": 2.9006,
      "theoretical_loss": 4.130871478992807,
      "tokens_seen": 309788672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000628834878002648,
      "loss": 2.8849,
      "theoretical_loss": 4.130769493126817,
      "tokens_seen": 309854208
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006288183279742766,
      "loss": 3.0694,
      "theoretical_loss": 4.130667534867476,
      "tokens_seen": 309919744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006288017779459051,
      "loss": 2.6658,
      "theoretical_loss": 4.130565604201477,
      "tokens_seen": 309985280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006287852279175336,
      "loss": 2.9713,
      "theoretical_loss": 4.130463701115521,
      "tokens_seen": 310050816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006287686778891622,
      "loss": 2.9067,
      "theoretical_loss": 4.130361825596317,
      "tokens_seen": 310116352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006287521278607906,
      "loss": 2.7695,
      "theoretical_loss": 4.130259977630586,
      "tokens_seen": 310181888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006287355778324191,
      "loss": 2.9114,
      "theoretical_loss": 4.130158157205056,
      "tokens_seen": 310247424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006287190278040476,
      "loss": 2.714,
      "theoretical_loss": 4.130056364306465,
      "tokens_seen": 310312960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006287024777756761,
      "loss": 3.0097,
      "theoretical_loss": 4.129954598921559,
      "tokens_seen": 310378496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006286859277473047,
      "loss": 2.8046,
      "theoretical_loss": 4.1298528610370955,
      "tokens_seen": 310444032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006286693777189332,
      "loss": 2.6582,
      "theoretical_loss": 4.12975115063984,
      "tokens_seen": 310509568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006286528276905618,
      "loss": 2.7761,
      "theoretical_loss": 4.129649467716565,
      "tokens_seen": 310575104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006286362776621902,
      "loss": 2.9868,
      "theoretical_loss": 4.1295478122540565,
      "tokens_seen": 310640640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006286197276338187,
      "loss": 3.097,
      "theoretical_loss": 4.1294461842391055,
      "tokens_seen": 310706176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006286031776054473,
      "loss": 2.7629,
      "theoretical_loss": 4.129344583658516,
      "tokens_seen": 310771712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006285866275770758,
      "loss": 3.0451,
      "theoretical_loss": 4.1292430104990965,
      "tokens_seen": 310837248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006285700775487043,
      "loss": 2.6791,
      "theoretical_loss": 4.1291414647476685,
      "tokens_seen": 310902784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006285535275203329,
      "loss": 2.4823,
      "theoretical_loss": 4.129039946391062,
      "tokens_seen": 310968320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006285369774919614,
      "loss": 2.9809,
      "theoretical_loss": 4.128938455416115,
      "tokens_seen": 311033856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00062852042746359,
      "loss": 2.9014,
      "theoretical_loss": 4.128836991809674,
      "tokens_seen": 311099392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006285038774352184,
      "loss": 3.2385,
      "theoretical_loss": 4.128735555558597,
      "tokens_seen": 311164928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006284873274068469,
      "loss": 3.0113,
      "theoretical_loss": 4.128634146649748,
      "tokens_seen": 311230464
    },
    {
      "epoch": 0.11,
      "objective/train/avg_token_score": 0.006613967940211296,
      "objective/train/avg_weight": 0.992220401763916,
      "objective/train/docs_used": 187155,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0075864791870117,
      "objective/train/std_weight": 0.041239481419324875,
      "objective/train/theoretical_loss": 4.128532765070004,
      "objective/train/tokens_used": 331756000,
      "theoretical_loss": 4.128532765070004,
      "tokens_seen": 311296000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006284707773784755,
      "loss": 2.8094,
      "theoretical_loss": 4.128532765070004,
      "tokens_seen": 311296000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000628454227350104,
      "loss": 2.9528,
      "theoretical_loss": 4.128431410806247,
      "tokens_seen": 311361536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006284376773217326,
      "loss": 3.0265,
      "theoretical_loss": 4.12833008384537,
      "tokens_seen": 311427072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006284211272933611,
      "loss": 2.9558,
      "theoretical_loss": 4.128228784174275,
      "tokens_seen": 311492608
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006284045772649895,
      "loss": 2.9124,
      "theoretical_loss": 4.128127511779873,
      "tokens_seen": 311558144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006283880272366181,
      "loss": 3.0,
      "theoretical_loss": 4.128026266649085,
      "tokens_seen": 311623680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006283714772082466,
      "loss": 3.1755,
      "theoretical_loss": 4.127925048768839,
      "tokens_seen": 311689216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006283549271798751,
      "loss": 2.9111,
      "theoretical_loss": 4.127823858126073,
      "tokens_seen": 311754752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006283383771515037,
      "loss": 2.7325,
      "theoretical_loss": 4.1277226947077335,
      "tokens_seen": 311820288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006283218271231322,
      "loss": 2.9665,
      "theoretical_loss": 4.127621558500778,
      "tokens_seen": 311885824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006283052770947608,
      "loss": 2.8619,
      "theoretical_loss": 4.12752044949217,
      "tokens_seen": 311951360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006282887270663893,
      "loss": 2.7279,
      "theoretical_loss": 4.127419367668884,
      "tokens_seen": 312016896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006282721770380177,
      "loss": 2.8925,
      "theoretical_loss": 4.127318313017904,
      "tokens_seen": 312082432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006282556270096463,
      "loss": 2.9724,
      "theoretical_loss": 4.12721728552622,
      "tokens_seen": 312147968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006282390769812748,
      "loss": 2.8649,
      "theoretical_loss": 4.1271162851808345,
      "tokens_seen": 312213504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006282225269529033,
      "loss": 3.0347,
      "theoretical_loss": 4.127015311968757,
      "tokens_seen": 312279040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006282059769245319,
      "loss": 2.8733,
      "theoretical_loss": 4.126914365877004,
      "tokens_seen": 312344576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006281894268961604,
      "loss": 2.8318,
      "theoretical_loss": 4.126813446892607,
      "tokens_seen": 312410112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000628172876867789,
      "loss": 2.8045,
      "theoretical_loss": 4.1267125550026,
      "tokens_seen": 312475648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006281563268394174,
      "loss": 2.7229,
      "theoretical_loss": 4.1266116901940295,
      "tokens_seen": 312541184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006281397768110459,
      "loss": 2.9938,
      "theoretical_loss": 4.126510852453949,
      "tokens_seen": 312606720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006281232267826745,
      "loss": 3.0035,
      "theoretical_loss": 4.126410041769423,
      "tokens_seen": 312672256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000628106676754303,
      "loss": 2.8292,
      "theoretical_loss": 4.126309258127524,
      "tokens_seen": 312737792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006280901267259315,
      "loss": 3.0273,
      "theoretical_loss": 4.126208501515331,
      "tokens_seen": 312803328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006280735766975601,
      "loss": 2.9696,
      "theoretical_loss": 4.126107771919935,
      "tokens_seen": 312868864
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006280570266691886,
      "loss": 3.2137,
      "theoretical_loss": 4.126007069328436,
      "tokens_seen": 312934400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006280404766408172,
      "loss": 2.7599,
      "theoretical_loss": 4.125906393727941,
      "tokens_seen": 312999936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006280239266124456,
      "loss": 2.8076,
      "theoretical_loss": 4.125805745105566,
      "tokens_seen": 313065472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006280073765840741,
      "loss": 2.9557,
      "theoretical_loss": 4.125705123448437,
      "tokens_seen": 313131008
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006279908265557027,
      "loss": 3.1212,
      "theoretical_loss": 4.125604528743689,
      "tokens_seen": 313196544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006279742765273312,
      "loss": 2.7492,
      "theoretical_loss": 4.125503960978464,
      "tokens_seen": 313262080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006279577264989597,
      "loss": 3.0423,
      "theoretical_loss": 4.1254034201399135,
      "tokens_seen": 313327616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006279411764705883,
      "loss": 2.8286,
      "theoretical_loss": 4.125302906215199,
      "tokens_seen": 313393152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006279246264422168,
      "loss": 2.913,
      "theoretical_loss": 4.12520241919149,
      "tokens_seen": 313458688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006279080764138454,
      "loss": 2.6727,
      "theoretical_loss": 4.125101959055965,
      "tokens_seen": 313524224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006278915263854738,
      "loss": 2.8305,
      "theoretical_loss": 4.125001525795811,
      "tokens_seen": 313589760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006278749763571023,
      "loss": 2.8528,
      "theoretical_loss": 4.124901119398222,
      "tokens_seen": 313655296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006278584263287309,
      "loss": 3.1415,
      "theoretical_loss": 4.124800739850406,
      "tokens_seen": 313720832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006278418763003594,
      "loss": 3.165,
      "theoretical_loss": 4.124700387139574,
      "tokens_seen": 313786368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006278253262719879,
      "loss": 2.895,
      "theoretical_loss": 4.12460006125295,
      "tokens_seen": 313851904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006278087762436165,
      "loss": 2.958,
      "theoretical_loss": 4.124499762177764,
      "tokens_seen": 313917440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000627792226215245,
      "loss": 3.1353,
      "theoretical_loss": 4.124399489901254,
      "tokens_seen": 313982976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006277756761868735,
      "loss": 2.9796,
      "theoretical_loss": 4.124299244410672,
      "tokens_seen": 314048512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000627759126158502,
      "loss": 2.761,
      "theoretical_loss": 4.124199025693272,
      "tokens_seen": 314114048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006277425761301305,
      "loss": 2.8584,
      "theoretical_loss": 4.124098833736321,
      "tokens_seen": 314179584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000627726026101759,
      "loss": 2.9824,
      "theoretical_loss": 4.123998668527094,
      "tokens_seen": 314245120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006277094760733875,
      "loss": 2.7338,
      "theoretical_loss": 4.123898530052874,
      "tokens_seen": 314310656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006276929260450161,
      "loss": 3.0223,
      "theoretical_loss": 4.123798418300953,
      "tokens_seen": 314376192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006276763760166446,
      "loss": 3.0328,
      "theoretical_loss": 4.123698333258631,
      "tokens_seen": 314441728
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000627659825988273,
      "loss": 2.8945,
      "theoretical_loss": 4.123598274913219,
      "tokens_seen": 314507264
    },
    {
      "epoch": 0.11,
      "objective/train/avg_token_score": 0.020839013159275055,
      "objective/train/avg_weight": 0.9754979014396667,
      "objective/train/docs_used": 189293,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.5371997356414795,
      "objective/train/std_weight": 0.13766953349113464,
      "objective/train/theoretical_loss": 4.123498243252032,
      "objective/train/tokens_used": 335032800,
      "theoretical_loss": 4.123498243252032,
      "tokens_seen": 314572800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006276432759599016,
      "loss": 3.126,
      "theoretical_loss": 4.123498243252032,
      "tokens_seen": 314572800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006276267259315301,
      "loss": 3.0883,
      "theoretical_loss": 4.1233982382624,
      "tokens_seen": 314638336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006276101759031587,
      "loss": 2.8896,
      "theoretical_loss": 4.123298259931657,
      "tokens_seen": 314703872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006275936258747872,
      "loss": 2.9567,
      "theoretical_loss": 4.123198308247146,
      "tokens_seen": 314769408
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006275770758464157,
      "loss": 2.9338,
      "theoretical_loss": 4.123098383196222,
      "tokens_seen": 314834944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006275605258180443,
      "loss": 2.9691,
      "theoretical_loss": 4.122998484766244,
      "tokens_seen": 314900480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006275439757896727,
      "loss": 2.8651,
      "theoretical_loss": 4.122898612944582,
      "tokens_seen": 314966016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006275274257613012,
      "loss": 2.6425,
      "theoretical_loss": 4.122798767718616,
      "tokens_seen": 315031552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006275108757329298,
      "loss": 2.8313,
      "theoretical_loss": 4.122698949075732,
      "tokens_seen": 315097088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006274943257045583,
      "loss": 3.0334,
      "theoretical_loss": 4.122599157003327,
      "tokens_seen": 315162624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006274777756761869,
      "loss": 2.8168,
      "theoretical_loss": 4.1224993914888035,
      "tokens_seen": 315228160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006274612256478154,
      "loss": 3.1466,
      "theoretical_loss": 4.122399652519576,
      "tokens_seen": 315293696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006274446756194439,
      "loss": 3.0737,
      "theoretical_loss": 4.122299940083065,
      "tokens_seen": 315359232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006274281255910725,
      "loss": 3.054,
      "theoretical_loss": 4.1222002541667,
      "tokens_seen": 315424768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006274115755627009,
      "loss": 2.8736,
      "theoretical_loss": 4.122100594757921,
      "tokens_seen": 315490304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006273950255343294,
      "loss": 2.7789,
      "theoretical_loss": 4.122000961844175,
      "tokens_seen": 315555840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000627378475505958,
      "loss": 2.8428,
      "theoretical_loss": 4.121901355412917,
      "tokens_seen": 315621376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006273619254775865,
      "loss": 2.7751,
      "theoretical_loss": 4.121801775451612,
      "tokens_seen": 315686912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006273453754492151,
      "loss": 2.7749,
      "theoretical_loss": 4.121702221947732,
      "tokens_seen": 315752448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006273288254208436,
      "loss": 3.1385,
      "theoretical_loss": 4.121602694888759,
      "tokens_seen": 315817984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000627312275392472,
      "loss": 2.8472,
      "theoretical_loss": 4.121503194262183,
      "tokens_seen": 315883520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006272957253641006,
      "loss": 2.8763,
      "theoretical_loss": 4.121403720055502,
      "tokens_seen": 315949056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006272791753357291,
      "loss": 2.954,
      "theoretical_loss": 4.121304272256222,
      "tokens_seen": 316014592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006272626253073576,
      "loss": 3.127,
      "theoretical_loss": 4.121204850851861,
      "tokens_seen": 316080128
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006272460752789862,
      "loss": 3.0192,
      "theoretical_loss": 4.121105455829939,
      "tokens_seen": 316145664
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006272295252506147,
      "loss": 2.926,
      "theoretical_loss": 4.121006087177992,
      "tokens_seen": 316211200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006272129752222433,
      "loss": 2.7413,
      "theoretical_loss": 4.120906744883559,
      "tokens_seen": 316276736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006271964251938718,
      "loss": 3.0267,
      "theoretical_loss": 4.120807428934189,
      "tokens_seen": 316342272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006271798751655002,
      "loss": 2.7284,
      "theoretical_loss": 4.120708139317441,
      "tokens_seen": 316407808
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006271633251371288,
      "loss": 2.8412,
      "theoretical_loss": 4.12060887602088,
      "tokens_seen": 316473344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006271467751087573,
      "loss": 2.8317,
      "theoretical_loss": 4.120509639032081,
      "tokens_seen": 316538880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006271302250803858,
      "loss": 2.846,
      "theoretical_loss": 4.120410428338628,
      "tokens_seen": 316604416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006271136750520144,
      "loss": 3.1419,
      "theoretical_loss": 4.120311243928111,
      "tokens_seen": 316669952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006270971250236429,
      "loss": 3.055,
      "theoretical_loss": 4.120212085788131,
      "tokens_seen": 316735488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006270805749952715,
      "loss": 3.1046,
      "theoretical_loss": 4.120112953906296,
      "tokens_seen": 316801024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006270640249669,
      "loss": 2.8645,
      "theoretical_loss": 4.120013848270222,
      "tokens_seen": 316866560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006270474749385284,
      "loss": 3.0394,
      "theoretical_loss": 4.119914768867536,
      "tokens_seen": 316932096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000627030924910157,
      "loss": 2.8479,
      "theoretical_loss": 4.11981571568587,
      "tokens_seen": 316997632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006270143748817855,
      "loss": 3.0257,
      "theoretical_loss": 4.119716688712866,
      "tokens_seen": 317063168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000626997824853414,
      "loss": 3.0285,
      "theoretical_loss": 4.119617687936175,
      "tokens_seen": 317128704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006269812748250426,
      "loss": 2.9389,
      "theoretical_loss": 4.119518713343455,
      "tokens_seen": 317194240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006269647247966711,
      "loss": 2.8104,
      "theoretical_loss": 4.119419764922374,
      "tokens_seen": 317259776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006269481747682997,
      "loss": 2.7709,
      "theoretical_loss": 4.119320842660606,
      "tokens_seen": 317325312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006269316247399281,
      "loss": 3.1257,
      "theoretical_loss": 4.119221946545836,
      "tokens_seen": 317390848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006269150747115566,
      "loss": 2.8922,
      "theoretical_loss": 4.119123076565755,
      "tokens_seen": 317456384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006268985246831852,
      "loss": 2.7928,
      "theoretical_loss": 4.119024232708064,
      "tokens_seen": 317521920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006268819746548137,
      "loss": 2.7517,
      "theoretical_loss": 4.118925414960472,
      "tokens_seen": 317587456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006268654246264423,
      "loss": 3.0078,
      "theoretical_loss": 4.118826623310696,
      "tokens_seen": 317652992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006268488745980708,
      "loss": 2.8073,
      "theoretical_loss": 4.11872785774646,
      "tokens_seen": 317718528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006268323245696993,
      "loss": 2.7108,
      "theoretical_loss": 4.1186291182555,
      "tokens_seen": 317784064
    },
    {
      "epoch": 0.11,
      "objective/train/avg_token_score": 0.011040781624615192,
      "objective/train/avg_weight": 0.9870314598083496,
      "objective/train/docs_used": 191279,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.704138994216919,
      "objective/train/std_weight": 0.07322721928358078,
      "objective/train/theoretical_loss": 4.118530404825556,
      "objective/train/tokens_used": 338309600,
      "theoretical_loss": 4.118530404825556,
      "tokens_seen": 317849600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006268157745413279,
      "loss": 2.9503,
      "theoretical_loss": 4.118530404825556,
      "tokens_seen": 317849600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006267992245129563,
      "loss": 2.8879,
      "theoretical_loss": 4.11843171744438,
      "tokens_seen": 317915136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006267826744845848,
      "loss": 2.8744,
      "theoretical_loss": 4.118333056099728,
      "tokens_seen": 317980672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006267661244562134,
      "loss": 2.9328,
      "theoretical_loss": 4.11823442077937,
      "tokens_seen": 318046208
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006267495744278418,
      "loss": 2.9916,
      "theoretical_loss": 4.1181358114710775,
      "tokens_seen": 318111744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006267330243994704,
      "loss": 2.8472,
      "theoretical_loss": 4.1180372281626365,
      "tokens_seen": 318177280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006267164743710989,
      "loss": 2.7711,
      "theoretical_loss": 4.117938670841838,
      "tokens_seen": 318242816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006266999243427275,
      "loss": 2.8285,
      "theoretical_loss": 4.117840139496482,
      "tokens_seen": 318308352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006266833743143559,
      "loss": 2.9542,
      "theoretical_loss": 4.117741634114376,
      "tokens_seen": 318373888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006266668242859844,
      "loss": 2.833,
      "theoretical_loss": 4.1176431546833365,
      "tokens_seen": 318439424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000626650274257613,
      "loss": 3.1472,
      "theoretical_loss": 4.117544701191187,
      "tokens_seen": 318504960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006266337242292415,
      "loss": 3.0736,
      "theoretical_loss": 4.117446273625763,
      "tokens_seen": 318570496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00062661717420087,
      "loss": 2.8657,
      "theoretical_loss": 4.117347871974903,
      "tokens_seen": 318636032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006266006241724986,
      "loss": 2.7442,
      "theoretical_loss": 4.1172494962264565,
      "tokens_seen": 318701568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000626584074144127,
      "loss": 2.9438,
      "theoretical_loss": 4.117151146368282,
      "tokens_seen": 318767104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006265675241157556,
      "loss": 3.0049,
      "theoretical_loss": 4.117052822388243,
      "tokens_seen": 318832640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006265509740873841,
      "loss": 2.9318,
      "theoretical_loss": 4.116954524274216,
      "tokens_seen": 318898176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006265344240590126,
      "loss": 3.0232,
      "theoretical_loss": 4.11685625201408,
      "tokens_seen": 318963712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006265178740306412,
      "loss": 2.7519,
      "theoretical_loss": 4.116758005595727,
      "tokens_seen": 319029248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006265013240022697,
      "loss": 3.1679,
      "theoretical_loss": 4.116659785007055,
      "tokens_seen": 319094784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006264847739738982,
      "loss": 3.0065,
      "theoretical_loss": 4.116561590235969,
      "tokens_seen": 319160320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006264682239455268,
      "loss": 2.9886,
      "theoretical_loss": 4.116463421270385,
      "tokens_seen": 319225856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006264516739171552,
      "loss": 2.9347,
      "theoretical_loss": 4.116365278098225,
      "tokens_seen": 319291392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006264351238887838,
      "loss": 2.7641,
      "theoretical_loss": 4.116267160707421,
      "tokens_seen": 319356928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006264185738604123,
      "loss": 3.0541,
      "theoretical_loss": 4.11616906908591,
      "tokens_seen": 319422464
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006264020238320408,
      "loss": 2.8242,
      "theoretical_loss": 4.11607100322164,
      "tokens_seen": 319488000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006263854738036694,
      "loss": 2.9401,
      "theoretical_loss": 4.115972963102565,
      "tokens_seen": 319553536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006263689237752979,
      "loss": 2.9213,
      "theoretical_loss": 4.11587494871665,
      "tokens_seen": 319619072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006263523737469264,
      "loss": 2.8158,
      "theoretical_loss": 4.115776960051864,
      "tokens_seen": 319684608
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000626335823718555,
      "loss": 3.0289,
      "theoretical_loss": 4.11567899709619,
      "tokens_seen": 319750144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006263192736901834,
      "loss": 2.8515,
      "theoretical_loss": 4.115581059837612,
      "tokens_seen": 319815680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000626302723661812,
      "loss": 3.0111,
      "theoretical_loss": 4.115483148264127,
      "tokens_seen": 319881216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006262861736334405,
      "loss": 3.0828,
      "theoretical_loss": 4.115385262363739,
      "tokens_seen": 319946752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000626269623605069,
      "loss": 3.075,
      "theoretical_loss": 4.1152874021244585,
      "tokens_seen": 320012288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006262530735766976,
      "loss": 3.1988,
      "theoretical_loss": 4.115189567534307,
      "tokens_seen": 320077824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006262365235483261,
      "loss": 2.8606,
      "theoretical_loss": 4.115091758581309,
      "tokens_seen": 320143360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006262199735199546,
      "loss": 2.9184,
      "theoretical_loss": 4.114993975253505,
      "tokens_seen": 320208896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006262034234915831,
      "loss": 3.0731,
      "theoretical_loss": 4.114896217538935,
      "tokens_seen": 320274432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006261868734632116,
      "loss": 2.9892,
      "theoretical_loss": 4.114798485425652,
      "tokens_seen": 320339968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006261703234348402,
      "loss": 2.7993,
      "theoretical_loss": 4.114700778901717,
      "tokens_seen": 320405504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006261537734064687,
      "loss": 3.0074,
      "theoretical_loss": 4.114603097955197,
      "tokens_seen": 320471040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006261372233780972,
      "loss": 3.1152,
      "theoretical_loss": 4.114505442574167,
      "tokens_seen": 320536576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006261206733497258,
      "loss": 2.9307,
      "theoretical_loss": 4.1144078127467125,
      "tokens_seen": 320602112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006261041233213543,
      "loss": 2.8885,
      "theoretical_loss": 4.114310208460924,
      "tokens_seen": 320667648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006260875732929827,
      "loss": 2.9778,
      "theoretical_loss": 4.114212629704902,
      "tokens_seen": 320733184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006260710232646113,
      "loss": 2.9715,
      "theoretical_loss": 4.114115076466755,
      "tokens_seen": 320798720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006260544732362398,
      "loss": 3.0885,
      "theoretical_loss": 4.114017548734598,
      "tokens_seen": 320864256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006260379232078684,
      "loss": 3.0231,
      "theoretical_loss": 4.113920046496554,
      "tokens_seen": 320929792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006260213731794969,
      "loss": 2.8667,
      "theoretical_loss": 4.113822569740757,
      "tokens_seen": 320995328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006260048231511254,
      "loss": 2.8442,
      "theoretical_loss": 4.113725118455344,
      "tokens_seen": 321060864
    },
    {
      "epoch": 0.11,
      "objective/train/avg_token_score": 0.010336566716432571,
      "objective/train/avg_weight": 0.9878377914428711,
      "objective/train/docs_used": 193378,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.2650017738342285,
      "objective/train/std_weight": 0.033733878284692764,
      "objective/train/theoretical_loss": 4.113627692628464,
      "objective/train/tokens_used": 341586400,
      "theoretical_loss": 4.113627692628464,
      "tokens_seen": 321126400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000625988273122754,
      "loss": 2.4646,
      "theoretical_loss": 4.113627692628464,
      "tokens_seen": 321126400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006259717230943825,
      "loss": 2.7276,
      "theoretical_loss": 4.113530292248273,
      "tokens_seen": 321191936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006259551730660109,
      "loss": 3.0546,
      "theoretical_loss": 4.113432917302934,
      "tokens_seen": 321257472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006259386230376395,
      "loss": 2.9609,
      "theoretical_loss": 4.113335567780618,
      "tokens_seen": 321323008
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000625922073009268,
      "loss": 2.5804,
      "theoretical_loss": 4.113238243669504,
      "tokens_seen": 321388544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006259055229808966,
      "loss": 2.9552,
      "theoretical_loss": 4.113140944957781,
      "tokens_seen": 321454080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006258889729525251,
      "loss": 2.8574,
      "theoretical_loss": 4.113043671633641,
      "tokens_seen": 321519616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006258724229241536,
      "loss": 2.9314,
      "theoretical_loss": 4.11294642368529,
      "tokens_seen": 321585152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006258558728957822,
      "loss": 2.9265,
      "theoretical_loss": 4.112849201100938,
      "tokens_seen": 321650688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006258393228674106,
      "loss": 3.1498,
      "theoretical_loss": 4.1127520038688035,
      "tokens_seen": 321716224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006258227728390391,
      "loss": 2.9271,
      "theoretical_loss": 4.112654831977112,
      "tokens_seen": 321781760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006258062228106677,
      "loss": 3.0688,
      "theoretical_loss": 4.1125576854141,
      "tokens_seen": 321847296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006257896727822962,
      "loss": 2.7692,
      "theoretical_loss": 4.112460564168009,
      "tokens_seen": 321912832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0006257731227539248,
      "loss": 2.7883,
      "theoretical_loss": 4.112363468227088,
      "tokens_seen": 321978368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006257565727255532,
      "loss": 3.2427,
      "theoretical_loss": 4.112266397579598,
      "tokens_seen": 322043904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006257400226971818,
      "loss": 3.0669,
      "theoretical_loss": 4.112169352213801,
      "tokens_seen": 322109440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006257234726688102,
      "loss": 3.2112,
      "theoretical_loss": 4.1120723321179735,
      "tokens_seen": 322174976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006257069226404387,
      "loss": 3.0334,
      "theoretical_loss": 4.111975337280397,
      "tokens_seen": 322240512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006256903726120673,
      "loss": 3.0903,
      "theoretical_loss": 4.111878367689359,
      "tokens_seen": 322306048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006256738225836958,
      "loss": 2.8164,
      "theoretical_loss": 4.1117814233331575,
      "tokens_seen": 322371584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006256572725553244,
      "loss": 2.8107,
      "theoretical_loss": 4.111684504200099,
      "tokens_seen": 322437120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006256407225269529,
      "loss": 3.1427,
      "theoretical_loss": 4.111587610278494,
      "tokens_seen": 322502656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006256241724985814,
      "loss": 3.0374,
      "theoretical_loss": 4.111490741556663,
      "tokens_seen": 322568192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00062560762247021,
      "loss": 3.0059,
      "theoretical_loss": 4.1113938980229365,
      "tokens_seen": 322633728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006255910724418384,
      "loss": 3.1915,
      "theoretical_loss": 4.11129707966565,
      "tokens_seen": 322699264
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006255745224134669,
      "loss": 2.8086,
      "theoretical_loss": 4.111200286473145,
      "tokens_seen": 322764800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006255579723850955,
      "loss": 2.8692,
      "theoretical_loss": 4.111103518433776,
      "tokens_seen": 322830336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000625541422356724,
      "loss": 3.0725,
      "theoretical_loss": 4.111006775535901,
      "tokens_seen": 322895872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006255248723283526,
      "loss": 2.843,
      "theoretical_loss": 4.110910057767887,
      "tokens_seen": 322961408
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006255083222999811,
      "loss": 3.0631,
      "theoretical_loss": 4.110813365118109,
      "tokens_seen": 323026944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006254917722716096,
      "loss": 2.8239,
      "theoretical_loss": 4.110716697574951,
      "tokens_seen": 323092480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006254752222432381,
      "loss": 2.7597,
      "theoretical_loss": 4.110620055126802,
      "tokens_seen": 323158016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006254586722148666,
      "loss": 3.1503,
      "theoretical_loss": 4.110523437762059,
      "tokens_seen": 323223552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006254421221864951,
      "loss": 2.9786,
      "theoretical_loss": 4.11042684546913,
      "tokens_seen": 323289088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006254255721581237,
      "loss": 2.8863,
      "theoretical_loss": 4.110330278236427,
      "tokens_seen": 323354624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006254090221297522,
      "loss": 2.7978,
      "theoretical_loss": 4.110233736052372,
      "tokens_seen": 323420160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006253924721013808,
      "loss": 2.9608,
      "theoretical_loss": 4.110137218905393,
      "tokens_seen": 323485696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006253759220730093,
      "loss": 2.9716,
      "theoretical_loss": 4.110040726783927,
      "tokens_seen": 323551232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006253593720446377,
      "loss": 2.9936,
      "theoretical_loss": 4.109944259676419,
      "tokens_seen": 323616768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006253428220162663,
      "loss": 2.9303,
      "theoretical_loss": 4.109847817571319,
      "tokens_seen": 323682304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006253262719878948,
      "loss": 2.7123,
      "theoretical_loss": 4.109751400457089,
      "tokens_seen": 323747840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006253097219595233,
      "loss": 2.8092,
      "theoretical_loss": 4.109655008322195,
      "tokens_seen": 323813376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006252931719311519,
      "loss": 2.9603,
      "theoretical_loss": 4.109558641155112,
      "tokens_seen": 323878912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006252766219027804,
      "loss": 3.015,
      "theoretical_loss": 4.109462298944322,
      "tokens_seen": 323944448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000625260071874409,
      "loss": 2.9541,
      "theoretical_loss": 4.109365981678316,
      "tokens_seen": 324009984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006252435218460375,
      "loss": 3.0019,
      "theoretical_loss": 4.109269689345592,
      "tokens_seen": 324075520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006252269718176659,
      "loss": 2.8225,
      "theoretical_loss": 4.109173421934654,
      "tokens_seen": 324141056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006252104217892945,
      "loss": 2.7836,
      "theoretical_loss": 4.109077179434016,
      "tokens_seen": 324206592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000625193871760923,
      "loss": 3.0257,
      "theoretical_loss": 4.1089809618321995,
      "tokens_seen": 324272128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006251773217325515,
      "loss": 2.8189,
      "theoretical_loss": 4.108884769117731,
      "tokens_seen": 324337664
    },
    {
      "epoch": 0.12,
      "objective/train/avg_token_score": 0.005501428619027138,
      "objective/train/avg_weight": 0.9935237169265747,
      "objective/train/docs_used": 195243,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.95985746383667,
      "objective/train/std_weight": 0.05099454149603844,
      "objective/train/theoretical_loss": 4.108788601279149,
      "objective/train/tokens_used": 344863200,
      "theoretical_loss": 4.108788601279149,
      "tokens_seen": 324403200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006251607717041801,
      "loss": 2.9111,
      "theoretical_loss": 4.108788601279149,
      "tokens_seen": 324403200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006251442216758086,
      "loss": 2.7511,
      "theoretical_loss": 4.108692458304994,
      "tokens_seen": 324468736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006251276716474372,
      "loss": 2.9881,
      "theoretical_loss": 4.108596340183819,
      "tokens_seen": 324534272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006251111216190656,
      "loss": 2.7785,
      "theoretical_loss": 4.108500246904184,
      "tokens_seen": 324599808
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006250945715906941,
      "loss": 2.9735,
      "theoretical_loss": 4.108404178454651,
      "tokens_seen": 324665344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006250780215623227,
      "loss": 2.9233,
      "theoretical_loss": 4.1083081348237975,
      "tokens_seen": 324730880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006250614715339512,
      "loss": 3.0283,
      "theoretical_loss": 4.108212116000203,
      "tokens_seen": 324796416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006250449215055797,
      "loss": 2.8803,
      "theoretical_loss": 4.108116121972457,
      "tokens_seen": 324861952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006250283714772083,
      "loss": 3.0285,
      "theoretical_loss": 4.108020152729157,
      "tokens_seen": 324927488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006250118214488368,
      "loss": 3.0317,
      "theoretical_loss": 4.107924208258905,
      "tokens_seen": 324993024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006249952714204654,
      "loss": 3.1522,
      "theoretical_loss": 4.107828288550314,
      "tokens_seen": 325058560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006249787213920938,
      "loss": 3.1245,
      "theoretical_loss": 4.107732393592003,
      "tokens_seen": 325124096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006249621713637223,
      "loss": 2.9407,
      "theoretical_loss": 4.107636523372598,
      "tokens_seen": 325189632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006249456213353509,
      "loss": 2.9732,
      "theoretical_loss": 4.107540677880734,
      "tokens_seen": 325255168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006249290713069794,
      "loss": 3.1152,
      "theoretical_loss": 4.107444857105052,
      "tokens_seen": 325320704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006249125212786079,
      "loss": 2.7934,
      "theoretical_loss": 4.107349061034201,
      "tokens_seen": 325386240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006248959712502365,
      "loss": 3.0021,
      "theoretical_loss": 4.107253289656838,
      "tokens_seen": 325451776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000624879421221865,
      "loss": 2.9856,
      "theoretical_loss": 4.107157542961628,
      "tokens_seen": 325517312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006248628711934935,
      "loss": 3.0816,
      "theoretical_loss": 4.10706182093724,
      "tokens_seen": 325582848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000624846321165122,
      "loss": 3.0125,
      "theoretical_loss": 4.106966123572356,
      "tokens_seen": 325648384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006248297711367505,
      "loss": 3.112,
      "theoretical_loss": 4.106870450855661,
      "tokens_seen": 325713920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006248132211083791,
      "loss": 2.9481,
      "theoretical_loss": 4.106774802775849,
      "tokens_seen": 325779456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006247966710800076,
      "loss": 3.2498,
      "theoretical_loss": 4.106679179321622,
      "tokens_seen": 325844992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006247801210516361,
      "loss": 2.8639,
      "theoretical_loss": 4.106583580481689,
      "tokens_seen": 325910528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006247635710232646,
      "loss": 2.9633,
      "theoretical_loss": 4.106488006244767,
      "tokens_seen": 325976064
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000624747020994893,
      "loss": 2.9204,
      "theoretical_loss": 4.106392456599577,
      "tokens_seen": 326041600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006247304709665216,
      "loss": 3.1653,
      "theoretical_loss": 4.106296931534854,
      "tokens_seen": 326107136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006247139209381501,
      "loss": 2.9708,
      "theoretical_loss": 4.106201431039335,
      "tokens_seen": 326172672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006246973709097787,
      "loss": 2.984,
      "theoretical_loss": 4.106105955101766,
      "tokens_seen": 326238208
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006246808208814072,
      "loss": 2.8431,
      "theoretical_loss": 4.1060105037109,
      "tokens_seen": 326303744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006246642708530357,
      "loss": 2.9453,
      "theoretical_loss": 4.105915076855499,
      "tokens_seen": 326369280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006246477208246643,
      "loss": 2.8175,
      "theoretical_loss": 4.105819674524332,
      "tokens_seen": 326434816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006246311707962927,
      "loss": 3.1987,
      "theoretical_loss": 4.105724296706172,
      "tokens_seen": 326500352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006246146207679212,
      "loss": 2.8046,
      "theoretical_loss": 4.105628943389805,
      "tokens_seen": 326565888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006245980707395498,
      "loss": 3.0885,
      "theoretical_loss": 4.1055336145640196,
      "tokens_seen": 326631424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006245815207111783,
      "loss": 2.97,
      "theoretical_loss": 4.105438310217615,
      "tokens_seen": 326696960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006245649706828069,
      "loss": 3.062,
      "theoretical_loss": 4.105343030339395,
      "tokens_seen": 326762496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006245484206544354,
      "loss": 2.9439,
      "theoretical_loss": 4.1052477749181735,
      "tokens_seen": 326828032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006245318706260639,
      "loss": 3.2554,
      "theoretical_loss": 4.10515254394277,
      "tokens_seen": 326893568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006245153205976925,
      "loss": 3.0478,
      "theoretical_loss": 4.1050573374020125,
      "tokens_seen": 326959104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006244987705693209,
      "loss": 2.9134,
      "theoretical_loss": 4.104962155284734,
      "tokens_seen": 327024640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006244822205409495,
      "loss": 3.1202,
      "theoretical_loss": 4.104866997579778,
      "tokens_seen": 327090176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000624465670512578,
      "loss": 2.8685,
      "theoretical_loss": 4.104771864275993,
      "tokens_seen": 327155712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006244491204842065,
      "loss": 2.6471,
      "theoretical_loss": 4.104676755362237,
      "tokens_seen": 327221248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006244325704558351,
      "loss": 2.6296,
      "theoretical_loss": 4.104581670827372,
      "tokens_seen": 327286784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006244160204274636,
      "loss": 2.6568,
      "theoretical_loss": 4.10448661066027,
      "tokens_seen": 327352320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006243994703990921,
      "loss": 3.2321,
      "theoretical_loss": 4.104391574849812,
      "tokens_seen": 327417856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006243829203707206,
      "loss": 3.1438,
      "theoretical_loss": 4.10429656338488,
      "tokens_seen": 327483392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006243663703423491,
      "loss": 2.8592,
      "theoretical_loss": 4.104201576254369,
      "tokens_seen": 327548928
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006243498203139777,
      "loss": 2.7813,
      "theoretical_loss": 4.10410661344718,
      "tokens_seen": 327614464
    },
    {
      "debugging/Self-BLEU-5": 0.5365128506817183,
      "debugging/distinct-1-grams": 0.7612814402327299,
      "debugging/distinct-2-grams": 0.9694583753853511,
      "debugging/entropy-1-grams": 6.003629944255698,
      "debugging/entropy-2-grams": 7.054987089269872,
      "debugging/length": 495.25,
      "debugging/num_segments": 16,
      "epoch": 0.12,
      "objective/train/avg_token_score": 0.04385810345411301,
      "objective/train/avg_weight": 0.9484303593635559,
      "objective/train/docs_used": 197327,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.5824131965637207,
      "objective/train/std_weight": 0.18450337648391724,
      "objective/train/theoretical_loss": 4.10401167495222,
      "objective/train/tokens_used": 348140000,
      "theoretical_loss": 4.10401167495222,
      "tokens_seen": 327680000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006243332702856062,
      "loss": 2.9128,
      "theoretical_loss": 4.10401167495222,
      "tokens_seen": 327680000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006243167202572347,
      "loss": 3.0387,
      "theoretical_loss": 4.103916760758405,
      "tokens_seen": 327745536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006243001702288633,
      "loss": 2.9298,
      "theoretical_loss": 4.103821870854656,
      "tokens_seen": 327811072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006242836202004918,
      "loss": 2.9698,
      "theoretical_loss": 4.103727005229903,
      "tokens_seen": 327876608
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006242670701721202,
      "loss": 3.0984,
      "theoretical_loss": 4.103632163873083,
      "tokens_seen": 327942144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006242505201437488,
      "loss": 3.1104,
      "theoretical_loss": 4.10353734677314,
      "tokens_seen": 328007680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006242339701153773,
      "loss": 3.0129,
      "theoretical_loss": 4.103442553919026,
      "tokens_seen": 328073216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006242174200870059,
      "loss": 2.6853,
      "theoretical_loss": 4.1033477852996985,
      "tokens_seen": 328138752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006242008700586344,
      "loss": 2.9647,
      "theoretical_loss": 4.103253040904124,
      "tokens_seen": 328204288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006241843200302629,
      "loss": 3.1966,
      "theoretical_loss": 4.103158320721276,
      "tokens_seen": 328269824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006241677700018915,
      "loss": 3.0337,
      "theoretical_loss": 4.103063624740133,
      "tokens_seen": 328335360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00062415121997352,
      "loss": 2.9106,
      "theoretical_loss": 4.102968952949684,
      "tokens_seen": 328400896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006241346699451484,
      "loss": 2.6802,
      "theoretical_loss": 4.102874305338923,
      "tokens_seen": 328466432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000624118119916777,
      "loss": 2.9556,
      "theoretical_loss": 4.102779681896852,
      "tokens_seen": 328531968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006241015698884055,
      "loss": 3.1563,
      "theoretical_loss": 4.10268508261248,
      "tokens_seen": 328597504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006240850198600341,
      "loss": 3.0841,
      "theoretical_loss": 4.102590507474824,
      "tokens_seen": 328663040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006240684698316626,
      "loss": 2.8995,
      "theoretical_loss": 4.1024959564729055,
      "tokens_seen": 328728576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006240519198032911,
      "loss": 2.9284,
      "theoretical_loss": 4.102401429595758,
      "tokens_seen": 328794112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006240353697749197,
      "loss": 3.2442,
      "theoretical_loss": 4.102306926832417,
      "tokens_seen": 328859648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006240188197465481,
      "loss": 2.942,
      "theoretical_loss": 4.102212448171928,
      "tokens_seen": 328925184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006240022697181766,
      "loss": 3.1145,
      "theoretical_loss": 4.1021179936033425,
      "tokens_seen": 328990720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006239857196898052,
      "loss": 2.8813,
      "theoretical_loss": 4.102023563115721,
      "tokens_seen": 329056256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006239691696614337,
      "loss": 3.0621,
      "theoretical_loss": 4.10192915669813,
      "tokens_seen": 329121792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006239526196330623,
      "loss": 2.9934,
      "theoretical_loss": 4.1018347743396415,
      "tokens_seen": 329187328
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006239360696046908,
      "loss": 3.0086,
      "theoretical_loss": 4.101740416029338,
      "tokens_seen": 329252864
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006239195195763193,
      "loss": 3.0286,
      "theoretical_loss": 4.101646081756305,
      "tokens_seen": 329318400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006239029695479479,
      "loss": 2.7866,
      "theoretical_loss": 4.101551771509641,
      "tokens_seen": 329383936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006238864195195763,
      "loss": 3.0686,
      "theoretical_loss": 4.101457485278444,
      "tokens_seen": 329449472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006238698694912048,
      "loss": 2.9936,
      "theoretical_loss": 4.101363223051826,
      "tokens_seen": 329515008
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006238533194628334,
      "loss": 2.8082,
      "theoretical_loss": 4.101268984818901,
      "tokens_seen": 329580544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006238367694344619,
      "loss": 2.9779,
      "theoretical_loss": 4.101174770568795,
      "tokens_seen": 329646080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006238202194060905,
      "loss": 2.9739,
      "theoretical_loss": 4.1010805802906365,
      "tokens_seen": 329711616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006238036693777189,
      "loss": 2.8683,
      "theoretical_loss": 4.100986413973564,
      "tokens_seen": 329777152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006237871193493475,
      "loss": 2.9597,
      "theoretical_loss": 4.100892271606721,
      "tokens_seen": 329842688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006237705693209759,
      "loss": 3.2958,
      "theoretical_loss": 4.1007981531792606,
      "tokens_seen": 329908224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006237540192926044,
      "loss": 3.0085,
      "theoretical_loss": 4.100704058680341,
      "tokens_seen": 329973760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000623737469264233,
      "loss": 2.8581,
      "theoretical_loss": 4.1006099880991265,
      "tokens_seen": 330039296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006237209192358615,
      "loss": 2.8333,
      "theoretical_loss": 4.100515941424792,
      "tokens_seen": 330104832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00062370436920749,
      "loss": 2.9824,
      "theoretical_loss": 4.100421918646517,
      "tokens_seen": 330170368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006236878191791186,
      "loss": 3.0022,
      "theoretical_loss": 4.1003279197534885,
      "tokens_seen": 330235904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006236712691507471,
      "loss": 2.9395,
      "theoretical_loss": 4.100233944734899,
      "tokens_seen": 330301440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006236547191223757,
      "loss": 2.9247,
      "theoretical_loss": 4.100139993579952,
      "tokens_seen": 330366976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006236381690940041,
      "loss": 2.948,
      "theoretical_loss": 4.100046066277853,
      "tokens_seen": 330432512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006236216190656326,
      "loss": 2.7531,
      "theoretical_loss": 4.09995216281782,
      "tokens_seen": 330498048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006236050690372612,
      "loss": 2.8791,
      "theoretical_loss": 4.0998582831890715,
      "tokens_seen": 330563584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006235885190088897,
      "loss": 3.0832,
      "theoretical_loss": 4.0997644273808405,
      "tokens_seen": 330629120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006235719689805182,
      "loss": 2.9443,
      "theoretical_loss": 4.09967059538236,
      "tokens_seen": 330694656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006235554189521468,
      "loss": 2.761,
      "theoretical_loss": 4.099576787182874,
      "tokens_seen": 330760192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006235388689237753,
      "loss": 3.1144,
      "theoretical_loss": 4.099483002771633,
      "tokens_seen": 330825728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006235223188954038,
      "loss": 2.8395,
      "theoretical_loss": 4.099389242137894,
      "tokens_seen": 330891264
    },
    {
      "epoch": 0.12,
      "objective/train/avg_token_score": 0.014471501111984253,
      "objective/train/avg_weight": 0.9829629063606262,
      "objective/train/docs_used": 199093,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2153048515319824,
      "objective/train/std_weight": 0.10116875171661377,
      "objective/train/theoretical_loss": 4.099295505270921,
      "objective/train/tokens_used": 351416800,
      "theoretical_loss": 4.099295505270921,
      "tokens_seen": 330956800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006235057688670323,
      "loss": 3.0416,
      "theoretical_loss": 4.099295505270921,
      "tokens_seen": 330956800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006234892188386608,
      "loss": 3.3357,
      "theoretical_loss": 4.099201792159985,
      "tokens_seen": 331022336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006234726688102894,
      "loss": 3.0358,
      "theoretical_loss": 4.099108102794363,
      "tokens_seen": 331087872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006234561187819179,
      "loss": 2.97,
      "theoretical_loss": 4.099014437163342,
      "tokens_seen": 331153408
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006234395687535464,
      "loss": 3.0715,
      "theoretical_loss": 4.098920795256213,
      "tokens_seen": 331218944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000623423018725175,
      "loss": 2.9947,
      "theoretical_loss": 4.098827177062273,
      "tokens_seen": 331284480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006234064686968034,
      "loss": 2.7144,
      "theoretical_loss": 4.098733582570831,
      "tokens_seen": 331350016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000623389918668432,
      "loss": 2.8131,
      "theoretical_loss": 4.098640011771198,
      "tokens_seen": 331415552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006233733686400605,
      "loss": 2.8694,
      "theoretical_loss": 4.098546464652693,
      "tokens_seen": 331481088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000623356818611689,
      "loss": 3.0237,
      "theoretical_loss": 4.098452941204643,
      "tokens_seen": 331546624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006233402685833176,
      "loss": 3.1084,
      "theoretical_loss": 4.098359441416383,
      "tokens_seen": 331612160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006233237185549461,
      "loss": 2.874,
      "theoretical_loss": 4.0982659652772515,
      "tokens_seen": 331677696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006233071685265746,
      "loss": 3.0433,
      "theoretical_loss": 4.098172512776597,
      "tokens_seen": 331743232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006232906184982032,
      "loss": 2.8514,
      "theoretical_loss": 4.098079083903773,
      "tokens_seen": 331808768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006232740684698316,
      "loss": 3.0414,
      "theoretical_loss": 4.097985678648142,
      "tokens_seen": 331874304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006232575184414602,
      "loss": 2.7637,
      "theoretical_loss": 4.09789229699907,
      "tokens_seen": 331939840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006232409684130887,
      "loss": 2.9237,
      "theoretical_loss": 4.097798938945933,
      "tokens_seen": 332005376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006232244183847172,
      "loss": 2.9382,
      "theoretical_loss": 4.097705604478112,
      "tokens_seen": 332070912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006232078683563458,
      "loss": 2.8933,
      "theoretical_loss": 4.097612293584998,
      "tokens_seen": 332136448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006231913183279743,
      "loss": 2.9539,
      "theoretical_loss": 4.0975190062559825,
      "tokens_seen": 332201984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006231747682996028,
      "loss": 2.6617,
      "theoretical_loss": 4.097425742480472,
      "tokens_seen": 332267520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006231582182712313,
      "loss": 2.9928,
      "theoretical_loss": 4.097332502247873,
      "tokens_seen": 332333056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006231416682428598,
      "loss": 2.8866,
      "theoretical_loss": 4.0972392855476025,
      "tokens_seen": 332398592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006231251182144884,
      "loss": 3.0534,
      "theoretical_loss": 4.097146092369084,
      "tokens_seen": 332464128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006231085681861169,
      "loss": 2.8049,
      "theoretical_loss": 4.097052922701746,
      "tokens_seen": 332529664
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006230920181577454,
      "loss": 2.8718,
      "theoretical_loss": 4.0969597765350265,
      "tokens_seen": 332595200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000623075468129374,
      "loss": 2.8454,
      "theoretical_loss": 4.096866653858368,
      "tokens_seen": 332660736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006230589181010025,
      "loss": 3.1193,
      "theoretical_loss": 4.09677355466122,
      "tokens_seen": 332726272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006230423680726309,
      "loss": 2.5917,
      "theoretical_loss": 4.096680478933042,
      "tokens_seen": 332791808
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006230258180442595,
      "loss": 2.9411,
      "theoretical_loss": 4.0965874266632945,
      "tokens_seen": 332857344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000623009268015888,
      "loss": 2.8542,
      "theoretical_loss": 4.09649439784145,
      "tokens_seen": 332922880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006229927179875166,
      "loss": 2.7163,
      "theoretical_loss": 4.096401392456988,
      "tokens_seen": 332988416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006229761679591451,
      "loss": 3.0043,
      "theoretical_loss": 4.09630841049939,
      "tokens_seen": 333053952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006229596179307736,
      "loss": 2.8241,
      "theoretical_loss": 4.096215451958146,
      "tokens_seen": 333119488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006229430679024022,
      "loss": 3.0319,
      "theoretical_loss": 4.096122516822757,
      "tokens_seen": 333185024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006229265178740307,
      "loss": 2.9808,
      "theoretical_loss": 4.096029605082726,
      "tokens_seen": 333250560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006229099678456592,
      "loss": 2.9831,
      "theoretical_loss": 4.095936716727564,
      "tokens_seen": 333316096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006228934178172877,
      "loss": 2.8993,
      "theoretical_loss": 4.095843851746791,
      "tokens_seen": 333381632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006228768677889162,
      "loss": 2.878,
      "theoretical_loss": 4.095751010129929,
      "tokens_seen": 333447168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006228603177605448,
      "loss": 3.1412,
      "theoretical_loss": 4.095658191866512,
      "tokens_seen": 333512704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006228437677321733,
      "loss": 2.6921,
      "theoretical_loss": 4.0955653969460775,
      "tokens_seen": 333578240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006228272177038018,
      "loss": 2.868,
      "theoretical_loss": 4.095472625358171,
      "tokens_seen": 333643776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006228106676754303,
      "loss": 3.0725,
      "theoretical_loss": 4.095379877092343,
      "tokens_seen": 333709312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006227941176470587,
      "loss": 3.0687,
      "theoretical_loss": 4.095287152138154,
      "tokens_seen": 333774848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006227775676186873,
      "loss": 2.6029,
      "theoretical_loss": 4.0951944504851685,
      "tokens_seen": 333840384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006227610175903158,
      "loss": 2.898,
      "theoretical_loss": 4.095101772122959,
      "tokens_seen": 333905920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006227444675619444,
      "loss": 2.8089,
      "theoretical_loss": 4.095009117041102,
      "tokens_seen": 333971456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006227279175335729,
      "loss": 2.7669,
      "theoretical_loss": 4.094916485229186,
      "tokens_seen": 334036992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006227113675052014,
      "loss": 3.0128,
      "theoretical_loss": 4.094823876676802,
      "tokens_seen": 334102528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00062269481747683,
      "loss": 2.6615,
      "theoretical_loss": 4.094731291373548,
      "tokens_seen": 334168064
    },
    {
      "epoch": 0.12,
      "objective/train/avg_token_score": 0.03409600257873535,
      "objective/train/avg_weight": 0.9599053859710693,
      "objective/train/docs_used": 200392,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.317330837249756,
      "objective/train/std_weight": 0.13070474565029144,
      "objective/train/theoretical_loss": 4.094638729309031,
      "objective/train/tokens_used": 354693600,
      "theoretical_loss": 4.094638729309031,
      "tokens_seen": 334233600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006226782674484584,
      "loss": 3.0537,
      "theoretical_loss": 4.094638729309031,
      "tokens_seen": 334233600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006226617174200869,
      "loss": 3.059,
      "theoretical_loss": 4.094546190472862,
      "tokens_seen": 334299136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006226451673917155,
      "loss": 2.7782,
      "theoretical_loss": 4.09445367485466,
      "tokens_seen": 334364672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000622628617363344,
      "loss": 2.7705,
      "theoretical_loss": 4.09436118244405,
      "tokens_seen": 334430208
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006226120673349726,
      "loss": 3.1397,
      "theoretical_loss": 4.094268713230667,
      "tokens_seen": 334495744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006225955173066011,
      "loss": 3.169,
      "theoretical_loss": 4.094176267204148,
      "tokens_seen": 334561280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006225789672782296,
      "loss": 3.1457,
      "theoretical_loss": 4.094083844354137,
      "tokens_seen": 334626816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006225624172498582,
      "loss": 2.9801,
      "theoretical_loss": 4.093991444670289,
      "tokens_seen": 334692352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006225458672214866,
      "loss": 2.9206,
      "theoretical_loss": 4.093899068142262,
      "tokens_seen": 334757888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006225293171931151,
      "loss": 3.0979,
      "theoretical_loss": 4.093806714759721,
      "tokens_seen": 334823424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006225127671647437,
      "loss": 3.0771,
      "theoretical_loss": 4.093714384512337,
      "tokens_seen": 334888960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006224962171363722,
      "loss": 2.9765,
      "theoretical_loss": 4.093622077389791,
      "tokens_seen": 334954496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006224796671080008,
      "loss": 3.0487,
      "theoretical_loss": 4.093529793381768,
      "tokens_seen": 335020032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006224631170796293,
      "loss": 3.0337,
      "theoretical_loss": 4.093437532477958,
      "tokens_seen": 335085568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006224465670512578,
      "loss": 2.7349,
      "theoretical_loss": 4.093345294668063,
      "tokens_seen": 335151104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006224300170228863,
      "loss": 3.1678,
      "theoretical_loss": 4.0932530799417846,
      "tokens_seen": 335216640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006224134669945148,
      "loss": 2.8741,
      "theoretical_loss": 4.093160888288837,
      "tokens_seen": 335282176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006223969169661433,
      "loss": 2.9864,
      "theoretical_loss": 4.0930687196989375,
      "tokens_seen": 335347712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006223803669377719,
      "loss": 3.1267,
      "theoretical_loss": 4.0929765741618125,
      "tokens_seen": 335413248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006223638169094004,
      "loss": 2.7119,
      "theoretical_loss": 4.092884451667191,
      "tokens_seen": 335478784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000622347266881029,
      "loss": 2.9003,
      "theoretical_loss": 4.092792352204814,
      "tokens_seen": 335544320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006223307168526575,
      "loss": 2.8534,
      "theoretical_loss": 4.092700275764424,
      "tokens_seen": 335609856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006223141668242859,
      "loss": 3.0541,
      "theoretical_loss": 4.092608222335774,
      "tokens_seen": 335675392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006222976167959145,
      "loss": 2.867,
      "theoretical_loss": 4.092516191908621,
      "tokens_seen": 335740928
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000622281066767543,
      "loss": 2.8893,
      "theoretical_loss": 4.09242418447273,
      "tokens_seen": 335806464
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006222645167391715,
      "loss": 2.7417,
      "theoretical_loss": 4.092332200017871,
      "tokens_seen": 335872000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006222479667108001,
      "loss": 2.565,
      "theoretical_loss": 4.092240238533822,
      "tokens_seen": 335937536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006222314166824286,
      "loss": 2.7427,
      "theoretical_loss": 4.092148300010367,
      "tokens_seen": 336003072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006222148666540572,
      "loss": 3.211,
      "theoretical_loss": 4.092056384437297,
      "tokens_seen": 336068608
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006221983166256857,
      "loss": 2.9683,
      "theoretical_loss": 4.091964491804409,
      "tokens_seen": 336134144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006221817665973141,
      "loss": 2.9521,
      "theoretical_loss": 4.091872622101506,
      "tokens_seen": 336199680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006221652165689427,
      "loss": 2.9541,
      "theoretical_loss": 4.091780775318399,
      "tokens_seen": 336265216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006221486665405712,
      "loss": 3.0361,
      "theoretical_loss": 4.091688951444904,
      "tokens_seen": 336330752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006221321165121997,
      "loss": 2.8226,
      "theoretical_loss": 4.091597150470845,
      "tokens_seen": 336396288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006221155664838283,
      "loss": 2.7323,
      "theoretical_loss": 4.091505372386051,
      "tokens_seen": 336461824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006220990164554568,
      "loss": 2.9485,
      "theoretical_loss": 4.091413617180358,
      "tokens_seen": 336527360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006220824664270854,
      "loss": 2.8317,
      "theoretical_loss": 4.091321884843609,
      "tokens_seen": 336592896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006220659163987138,
      "loss": 2.7097,
      "theoretical_loss": 4.091230175365653,
      "tokens_seen": 336658432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006220493663703423,
      "loss": 3.0608,
      "theoretical_loss": 4.0911384887363464,
      "tokens_seen": 336723968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006220328163419709,
      "loss": 2.8485,
      "theoretical_loss": 4.091046824945551,
      "tokens_seen": 336789504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006220162663135994,
      "loss": 2.9937,
      "theoretical_loss": 4.0909551839831355,
      "tokens_seen": 336855040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006219997162852279,
      "loss": 2.9027,
      "theoretical_loss": 4.090863565838974,
      "tokens_seen": 336920576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006219831662568565,
      "loss": 2.6566,
      "theoretical_loss": 4.090771970502948,
      "tokens_seen": 336986112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000621966616228485,
      "loss": 3.1184,
      "theoretical_loss": 4.090680397964947,
      "tokens_seen": 337051648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006219500662001136,
      "loss": 2.8832,
      "theoretical_loss": 4.090588848214865,
      "tokens_seen": 337117184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000621933516171742,
      "loss": 2.9178,
      "theoretical_loss": 4.0904973212426015,
      "tokens_seen": 337182720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006219169661433705,
      "loss": 2.7881,
      "theoretical_loss": 4.090405817038065,
      "tokens_seen": 337248256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006219004161149991,
      "loss": 2.8194,
      "theoretical_loss": 4.090314335591169,
      "tokens_seen": 337313792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006218838660866276,
      "loss": 2.8505,
      "theoretical_loss": 4.090222876891834,
      "tokens_seen": 337379328
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006218673160582561,
      "loss": 3.1798,
      "theoretical_loss": 4.090131440929985,
      "tokens_seen": 337444864
    },
    {
      "epoch": 0.12,
      "objective/train/avg_token_score": 0.012404380366206169,
      "objective/train/avg_weight": 0.9854112267494202,
      "objective/train/docs_used": 202247,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.00565242767334,
      "objective/train/std_weight": 0.0718371719121933,
      "objective/train/theoretical_loss": 4.090040027695556,
      "objective/train/tokens_used": 357970400,
      "theoretical_loss": 4.090040027695556,
      "tokens_seen": 337510400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006218507660298847,
      "loss": 2.9469,
      "theoretical_loss": 4.090040027695556,
      "tokens_seen": 337510400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006218342160015132,
      "loss": 2.8854,
      "theoretical_loss": 4.0899486371784874,
      "tokens_seen": 337575936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006218176659731416,
      "loss": 2.8337,
      "theoretical_loss": 4.089857269368725,
      "tokens_seen": 337641472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006218011159447701,
      "loss": 2.8903,
      "theoretical_loss": 4.089765924256218,
      "tokens_seen": 337707008
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006217845659163987,
      "loss": 3.1098,
      "theoretical_loss": 4.089674601830929,
      "tokens_seen": 337772544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006217680158880272,
      "loss": 3.061,
      "theoretical_loss": 4.08958330208282,
      "tokens_seen": 337838080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006217514658596557,
      "loss": 2.7098,
      "theoretical_loss": 4.089492025001864,
      "tokens_seen": 337903616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006217349158312843,
      "loss": 2.7637,
      "theoretical_loss": 4.089400770578038,
      "tokens_seen": 337969152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006217183658029128,
      "loss": 2.8285,
      "theoretical_loss": 4.089309538801327,
      "tokens_seen": 338034688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006217018157745413,
      "loss": 2.8845,
      "theoretical_loss": 4.08921832966172,
      "tokens_seen": 338100224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006216852657461698,
      "loss": 2.5483,
      "theoretical_loss": 4.0891271431492155,
      "tokens_seen": 338165760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006216687157177983,
      "loss": 2.841,
      "theoretical_loss": 4.089035979253816,
      "tokens_seen": 338231296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006216521656894269,
      "loss": 2.9034,
      "theoretical_loss": 4.08894483796553,
      "tokens_seen": 338296832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006216356156610554,
      "loss": 3.0122,
      "theoretical_loss": 4.0888537192743755,
      "tokens_seen": 338362368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006216190656326839,
      "loss": 2.9066,
      "theoretical_loss": 4.088762623170373,
      "tokens_seen": 338427904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006216025156043125,
      "loss": 2.8105,
      "theoretical_loss": 4.088671549643553,
      "tokens_seen": 338493440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000621585965575941,
      "loss": 2.8429,
      "theoretical_loss": 4.088580498683948,
      "tokens_seen": 338558976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006215694155475695,
      "loss": 3.1848,
      "theoretical_loss": 4.088489470281601,
      "tokens_seen": 338624512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000621552865519198,
      "loss": 3.0331,
      "theoretical_loss": 4.088398464426559,
      "tokens_seen": 338690048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006215363154908265,
      "loss": 2.9359,
      "theoretical_loss": 4.088307481108876,
      "tokens_seen": 338755584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006215197654624551,
      "loss": 2.9517,
      "theoretical_loss": 4.088216520318612,
      "tokens_seen": 338821120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006215032154340836,
      "loss": 2.8285,
      "theoretical_loss": 4.0881255820458335,
      "tokens_seen": 338886656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006214866654057121,
      "loss": 2.9872,
      "theoretical_loss": 4.088034666280614,
      "tokens_seen": 338952192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006214701153773407,
      "loss": 2.7347,
      "theoretical_loss": 4.087943773013032,
      "tokens_seen": 339017728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006214535653489691,
      "loss": 2.8961,
      "theoretical_loss": 4.087852902233173,
      "tokens_seen": 339083264
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006214370153205977,
      "loss": 2.9426,
      "theoretical_loss": 4.087762053931129,
      "tokens_seen": 339148800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006214204652922262,
      "loss": 3.0664,
      "theoretical_loss": 4.087671228096997,
      "tokens_seen": 339214336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006214039152638547,
      "loss": 3.045,
      "theoretical_loss": 4.087580424720882,
      "tokens_seen": 339279872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006213873652354833,
      "loss": 2.9053,
      "theoretical_loss": 4.087489643792894,
      "tokens_seen": 339345408
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006213708152071118,
      "loss": 2.9456,
      "theoretical_loss": 4.08739888530315,
      "tokens_seen": 339410944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006213542651787403,
      "loss": 2.7596,
      "theoretical_loss": 4.087308149241774,
      "tokens_seen": 339476480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006213377151503688,
      "loss": 2.9711,
      "theoretical_loss": 4.087217435598894,
      "tokens_seen": 339542016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006213211651219973,
      "loss": 2.8222,
      "theoretical_loss": 4.087126744364646,
      "tokens_seen": 339607552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006213046150936259,
      "loss": 2.8543,
      "theoretical_loss": 4.087036075529172,
      "tokens_seen": 339673088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006212880650652544,
      "loss": 2.8581,
      "theoretical_loss": 4.086945429082618,
      "tokens_seen": 339738624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006212715150368829,
      "loss": 2.7524,
      "theoretical_loss": 4.086854805015141,
      "tokens_seen": 339804160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006212549650085115,
      "loss": 2.9431,
      "theoretical_loss": 4.086764203316902,
      "tokens_seen": 339869696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00062123841498014,
      "loss": 3.0339,
      "theoretical_loss": 4.086673623978064,
      "tokens_seen": 339935232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006212218649517684,
      "loss": 2.7962,
      "theoretical_loss": 4.086583066988802,
      "tokens_seen": 340000768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000621205314923397,
      "loss": 2.8104,
      "theoretical_loss": 4.086492532339296,
      "tokens_seen": 340066304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006211887648950255,
      "loss": 2.9034,
      "theoretical_loss": 4.0864020200197295,
      "tokens_seen": 340131840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006211722148666541,
      "loss": 2.9256,
      "theoretical_loss": 4.086311530020296,
      "tokens_seen": 340197376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006211556648382826,
      "loss": 2.8804,
      "theoretical_loss": 4.086221062331192,
      "tokens_seen": 340262912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006211391148099111,
      "loss": 2.9864,
      "theoretical_loss": 4.086130616942621,
      "tokens_seen": 340328448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006211225647815397,
      "loss": 3.0284,
      "theoretical_loss": 4.086040193844794,
      "tokens_seen": 340393984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006211060147531682,
      "loss": 2.9509,
      "theoretical_loss": 4.085949793027927,
      "tokens_seen": 340459520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006210894647247966,
      "loss": 2.8027,
      "theoretical_loss": 4.0858594144822415,
      "tokens_seen": 340525056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006210729146964252,
      "loss": 2.9748,
      "theoretical_loss": 4.085769058197968,
      "tokens_seen": 340590592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006210563646680537,
      "loss": 2.8681,
      "theoretical_loss": 4.085678724165341,
      "tokens_seen": 340656128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006210398146396823,
      "loss": 2.919,
      "theoretical_loss": 4.0855884123746,
      "tokens_seen": 340721664
    },
    {
      "epoch": 0.12,
      "objective/train/avg_token_score": 0.010883872397243977,
      "objective/train/avg_weight": 0.9871997833251953,
      "objective/train/docs_used": 204260,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7689108848571777,
      "objective/train/std_weight": 0.04935688525438309,
      "objective/train/theoretical_loss": 4.085498122815992,
      "objective/train/tokens_used": 361247200,
      "theoretical_loss": 4.085498122815992,
      "tokens_seen": 340787200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006210232646113108,
      "loss": 2.8958,
      "theoretical_loss": 4.085498122815992,
      "tokens_seen": 340787200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006210067145829393,
      "loss": 2.8346,
      "theoretical_loss": 4.085407855479772,
      "tokens_seen": 340852736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006209901645545679,
      "loss": 2.8644,
      "theoretical_loss": 4.085317610356199,
      "tokens_seen": 340918272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006209736145261963,
      "loss": 3.1555,
      "theoretical_loss": 4.085227387435538,
      "tokens_seen": 340983808
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006209570644978248,
      "loss": 2.8466,
      "theoretical_loss": 4.08513718670806,
      "tokens_seen": 341049344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006209405144694534,
      "loss": 3.0176,
      "theoretical_loss": 4.085047008164044,
      "tokens_seen": 341114880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006209239644410819,
      "loss": 3.007,
      "theoretical_loss": 4.084956851793773,
      "tokens_seen": 341180416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006209074144127105,
      "loss": 3.098,
      "theoretical_loss": 4.0848667175875395,
      "tokens_seen": 341245952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000620890864384339,
      "loss": 2.7945,
      "theoretical_loss": 4.0847766055356365,
      "tokens_seen": 341311488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006208743143559675,
      "loss": 3.1571,
      "theoretical_loss": 4.084686515628368,
      "tokens_seen": 341377024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006208577643275961,
      "loss": 2.7735,
      "theoretical_loss": 4.0845964478560415,
      "tokens_seen": 341442560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006208412142992244,
      "loss": 2.6419,
      "theoretical_loss": 4.084506402208972,
      "tokens_seen": 341508096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000620824664270853,
      "loss": 2.9243,
      "theoretical_loss": 4.0844163786774805,
      "tokens_seen": 341573632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006208081142424815,
      "loss": 2.756,
      "theoretical_loss": 4.084326377251894,
      "tokens_seen": 341639168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00062079156421411,
      "loss": 2.871,
      "theoretical_loss": 4.084236397922544,
      "tokens_seen": 341704704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006207750141857386,
      "loss": 2.8271,
      "theoretical_loss": 4.084146440679769,
      "tokens_seen": 341770240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006207584641573671,
      "loss": 2.8199,
      "theoretical_loss": 4.0840565055139155,
      "tokens_seen": 341835776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006207419141289957,
      "loss": 2.9091,
      "theoretical_loss": 4.083966592415333,
      "tokens_seen": 341901312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006207253641006241,
      "loss": 2.8459,
      "theoretical_loss": 4.083876701374381,
      "tokens_seen": 341966848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006207088140722526,
      "loss": 3.0665,
      "theoretical_loss": 4.08378683238142,
      "tokens_seen": 342032384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006206922640438812,
      "loss": 2.8628,
      "theoretical_loss": 4.08369698542682,
      "tokens_seen": 342097920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006206757140155097,
      "loss": 2.8708,
      "theoretical_loss": 4.083607160500957,
      "tokens_seen": 342163456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006206591639871382,
      "loss": 2.9352,
      "theoretical_loss": 4.083517357594212,
      "tokens_seen": 342228992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006206426139587668,
      "loss": 2.8609,
      "theoretical_loss": 4.0834275766969705,
      "tokens_seen": 342294528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006206260639303953,
      "loss": 2.8524,
      "theoretical_loss": 4.083337817799629,
      "tokens_seen": 342360064
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006206095139020238,
      "loss": 2.8316,
      "theoretical_loss": 4.083248080892584,
      "tokens_seen": 342425600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006205929638736523,
      "loss": 3.0618,
      "theoretical_loss": 4.083158365966243,
      "tokens_seen": 342491136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006205764138452808,
      "loss": 3.0973,
      "theoretical_loss": 4.083068673011017,
      "tokens_seen": 342556672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006205598638169094,
      "loss": 2.7091,
      "theoretical_loss": 4.082979002017321,
      "tokens_seen": 342622208
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006205433137885379,
      "loss": 2.9529,
      "theoretical_loss": 4.082889352975583,
      "tokens_seen": 342687744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006205267637601664,
      "loss": 3.1131,
      "theoretical_loss": 4.082799725876228,
      "tokens_seen": 342753280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000620510213731795,
      "loss": 2.7445,
      "theoretical_loss": 4.082710120709694,
      "tokens_seen": 342818816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006204936637034234,
      "loss": 2.9459,
      "theoretical_loss": 4.082620537466422,
      "tokens_seen": 342884352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000620477113675052,
      "loss": 2.9801,
      "theoretical_loss": 4.08253097613686,
      "tokens_seen": 342949888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006204605636466805,
      "loss": 3.1722,
      "theoretical_loss": 4.082441436711459,
      "tokens_seen": 343015424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000620444013618309,
      "loss": 3.0992,
      "theoretical_loss": 4.082351919180682,
      "tokens_seen": 343080960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006204274635899376,
      "loss": 2.7557,
      "theoretical_loss": 4.082262423534992,
      "tokens_seen": 343146496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006204109135615661,
      "loss": 3.1453,
      "theoretical_loss": 4.0821729497648604,
      "tokens_seen": 343212032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006203943635331947,
      "loss": 2.9178,
      "theoretical_loss": 4.082083497860766,
      "tokens_seen": 343277568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006203778135048232,
      "loss": 2.6981,
      "theoretical_loss": 4.0819940678131905,
      "tokens_seen": 343343104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006203612634764516,
      "loss": 2.9642,
      "theoretical_loss": 4.081904659612624,
      "tokens_seen": 343408640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006203447134480802,
      "loss": 3.0785,
      "theoretical_loss": 4.081815273249561,
      "tokens_seen": 343474176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006203281634197087,
      "loss": 2.9083,
      "theoretical_loss": 4.081725908714505,
      "tokens_seen": 343539712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006203116133913372,
      "loss": 2.7995,
      "theoretical_loss": 4.081636565997959,
      "tokens_seen": 343605248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006202950633629658,
      "loss": 2.8841,
      "theoretical_loss": 4.08154724509044,
      "tokens_seen": 343670784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006202785133345943,
      "loss": 2.8559,
      "theoretical_loss": 4.081457945982465,
      "tokens_seen": 343736320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006202619633062229,
      "loss": 2.6048,
      "theoretical_loss": 4.08136866866456,
      "tokens_seen": 343801856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006202454132778513,
      "loss": 3.0393,
      "theoretical_loss": 4.081279413127254,
      "tokens_seen": 343867392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006202288632494798,
      "loss": 2.9387,
      "theoretical_loss": 4.0811901793610845,
      "tokens_seen": 343932928
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006202123132211084,
      "loss": 3.0181,
      "theoretical_loss": 4.081100967356594,
      "tokens_seen": 343998464
    },
    {
      "epoch": 0.12,
      "objective/train/avg_token_score": 0.023849839344620705,
      "objective/train/avg_weight": 0.9719239473342896,
      "objective/train/docs_used": 206338,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0161733627319336,
      "objective/train/std_weight": 0.10900367796421051,
      "objective/train/theoretical_loss": 4.081011777104333,
      "objective/train/tokens_used": 364524000,
      "theoretical_loss": 4.081011777104333,
      "tokens_seen": 344064000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006201957631927369,
      "loss": 2.856,
      "theoretical_loss": 4.081011777104333,
      "tokens_seen": 344064000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006201792131643654,
      "loss": 2.9409,
      "theoretical_loss": 4.080922608594854,
      "tokens_seen": 344129536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000620162663135994,
      "loss": 2.7991,
      "theoretical_loss": 4.080833461818718,
      "tokens_seen": 344195072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006201461131076225,
      "loss": 2.7057,
      "theoretical_loss": 4.08074433676649,
      "tokens_seen": 344260608
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006201295630792511,
      "loss": 2.7689,
      "theoretical_loss": 4.080655233428745,
      "tokens_seen": 344326144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006201130130508795,
      "loss": 3.0976,
      "theoretical_loss": 4.0805661517960585,
      "tokens_seen": 344391680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000620096463022508,
      "loss": 2.9808,
      "theoretical_loss": 4.080477091859015,
      "tokens_seen": 344457216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006200799129941366,
      "loss": 2.996,
      "theoretical_loss": 4.080388053608205,
      "tokens_seen": 344522752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006200633629657651,
      "loss": 2.8651,
      "theoretical_loss": 4.080299037034224,
      "tokens_seen": 344588288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006200468129373936,
      "loss": 2.9404,
      "theoretical_loss": 4.080210042127672,
      "tokens_seen": 344653824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006200302629090222,
      "loss": 2.8364,
      "theoretical_loss": 4.0801210688791585,
      "tokens_seen": 344719360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006200137128806507,
      "loss": 2.7594,
      "theoretical_loss": 4.080032117279296,
      "tokens_seen": 344784896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006199971628522792,
      "loss": 2.9024,
      "theoretical_loss": 4.079943187318703,
      "tokens_seen": 344850432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006199806128239077,
      "loss": 3.0268,
      "theoretical_loss": 4.079854278988005,
      "tokens_seen": 344915968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006199640627955362,
      "loss": 2.9575,
      "theoretical_loss": 4.0797653922778325,
      "tokens_seen": 344981504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006199475127671648,
      "loss": 2.8425,
      "theoretical_loss": 4.0796765271788225,
      "tokens_seen": 345047040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006199309627387933,
      "loss": 3.1776,
      "theoretical_loss": 4.079587683681617,
      "tokens_seen": 345112576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006199144127104218,
      "loss": 3.1336,
      "theoretical_loss": 4.079498861776865,
      "tokens_seen": 345178112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006198978626820504,
      "loss": 2.9752,
      "theoretical_loss": 4.07941006145522,
      "tokens_seen": 345243648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006198813126536788,
      "loss": 2.6648,
      "theoretical_loss": 4.079321282707342,
      "tokens_seen": 345309184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006198647626253073,
      "loss": 3.0262,
      "theoretical_loss": 4.0792325255238975,
      "tokens_seen": 345374720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006198482125969358,
      "loss": 3.0196,
      "theoretical_loss": 4.079143789895556,
      "tokens_seen": 345440256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006198316625685644,
      "loss": 3.1887,
      "theoretical_loss": 4.079055075812997,
      "tokens_seen": 345505792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006198151125401929,
      "loss": 2.8298,
      "theoretical_loss": 4.078966383266904,
      "tokens_seen": 345571328
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006197985625118214,
      "loss": 2.931,
      "theoretical_loss": 4.078877712247963,
      "tokens_seen": 345636864
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00061978201248345,
      "loss": 2.5876,
      "theoretical_loss": 4.078789062746872,
      "tokens_seen": 345702400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006197654624550784,
      "loss": 2.87,
      "theoretical_loss": 4.07870043475433,
      "tokens_seen": 345767936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006197489124267069,
      "loss": 2.9265,
      "theoretical_loss": 4.078611828261042,
      "tokens_seen": 345833472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006197323623983355,
      "loss": 2.8643,
      "theoretical_loss": 4.078523243257723,
      "tokens_seen": 345899008
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000619715812369964,
      "loss": 2.8906,
      "theoretical_loss": 4.078434679735088,
      "tokens_seen": 345964544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006196992623415926,
      "loss": 2.8789,
      "theoretical_loss": 4.078346137683862,
      "tokens_seen": 346030080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006196827123132211,
      "loss": 2.9891,
      "theoretical_loss": 4.078257617094776,
      "tokens_seen": 346095616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006196661622848496,
      "loss": 2.9737,
      "theoretical_loss": 4.078169117958562,
      "tokens_seen": 346161152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006196496122564782,
      "loss": 2.8059,
      "theoretical_loss": 4.078080640265963,
      "tokens_seen": 346226688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006196330622281066,
      "loss": 2.9851,
      "theoretical_loss": 4.077992184007725,
      "tokens_seen": 346292224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006196165121997351,
      "loss": 3.0125,
      "theoretical_loss": 4.0779037491746,
      "tokens_seen": 346357760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006195999621713637,
      "loss": 2.9083,
      "theoretical_loss": 4.077815335757347,
      "tokens_seen": 346423296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006195834121429922,
      "loss": 2.9062,
      "theoretical_loss": 4.077726943746729,
      "tokens_seen": 346488832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006195668621146208,
      "loss": 3.1738,
      "theoretical_loss": 4.077638573133515,
      "tokens_seen": 346554368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006195503120862493,
      "loss": 3.1174,
      "theoretical_loss": 4.077550223908482,
      "tokens_seen": 346619904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006195337620578778,
      "loss": 2.772,
      "theoretical_loss": 4.0774618960624105,
      "tokens_seen": 346685440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006195172120295063,
      "loss": 2.9291,
      "theoretical_loss": 4.077373589586087,
      "tokens_seen": 346750976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006195006620011348,
      "loss": 2.9774,
      "theoretical_loss": 4.077285304470304,
      "tokens_seen": 346816512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006194841119727633,
      "loss": 2.9153,
      "theoretical_loss": 4.077197040705859,
      "tokens_seen": 346882048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006194675619443919,
      "loss": 2.6834,
      "theoretical_loss": 4.077108798283557,
      "tokens_seen": 346947584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006194510119160204,
      "loss": 2.9243,
      "theoretical_loss": 4.077020577194207,
      "tokens_seen": 347013120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000619434461887649,
      "loss": 3.1362,
      "theoretical_loss": 4.076932377428625,
      "tokens_seen": 347078656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006194179118592775,
      "loss": 2.7985,
      "theoretical_loss": 4.076844198977629,
      "tokens_seen": 347144192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000619401361830906,
      "loss": 3.043,
      "theoretical_loss": 4.076756041832049,
      "tokens_seen": 347209728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006193848118025345,
      "loss": 2.895,
      "theoretical_loss": 4.076667905982717,
      "tokens_seen": 347275264
    },
    {
      "epoch": 0.12,
      "objective/train/avg_token_score": 0.03382844105362892,
      "objective/train/avg_weight": 0.9601943492889404,
      "objective/train/docs_used": 208654,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0445556640625,
      "objective/train/std_weight": 0.14660023152828217,
      "objective/train/theoretical_loss": 4.076579791420469,
      "objective/train/tokens_used": 367800800,
      "theoretical_loss": 4.076579791420469,
      "tokens_seen": 347340800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000619368261774163,
      "loss": 2.8931,
      "theoretical_loss": 4.076579791420469,
      "tokens_seen": 347340800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006193517117457915,
      "loss": 2.8893,
      "theoretical_loss": 4.0764916981361505,
      "tokens_seen": 347406336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006193351617174201,
      "loss": 2.7194,
      "theoretical_loss": 4.07640362612061,
      "tokens_seen": 347471872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006193186116890486,
      "loss": 2.7994,
      "theoretical_loss": 4.076315575364702,
      "tokens_seen": 347537408
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006193020616606772,
      "loss": 3.0175,
      "theoretical_loss": 4.076227545859288,
      "tokens_seen": 347602944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006192855116323057,
      "loss": 2.9104,
      "theoretical_loss": 4.076139537595234,
      "tokens_seen": 347668480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006192689616039341,
      "loss": 3.1039,
      "theoretical_loss": 4.076051550563411,
      "tokens_seen": 347734016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006192524115755627,
      "loss": 3.0343,
      "theoretical_loss": 4.075963584754698,
      "tokens_seen": 347799552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006192358615471912,
      "loss": 3.1525,
      "theoretical_loss": 4.075875640159978,
      "tokens_seen": 347865088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006192193115188197,
      "loss": 2.8902,
      "theoretical_loss": 4.075787716770139,
      "tokens_seen": 347930624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006192027614904483,
      "loss": 2.896,
      "theoretical_loss": 4.075699814576076,
      "tokens_seen": 347996160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006191862114620768,
      "loss": 2.7954,
      "theoretical_loss": 4.075611933568689,
      "tokens_seen": 348061696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006191696614337054,
      "loss": 3.0413,
      "theoretical_loss": 4.075524073738883,
      "tokens_seen": 348127232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006191531114053339,
      "loss": 3.0904,
      "theoretical_loss": 4.0754362350775715,
      "tokens_seen": 348192768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006191365613769623,
      "loss": 2.9179,
      "theoretical_loss": 4.075348417575669,
      "tokens_seen": 348258304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006191200113485909,
      "loss": 2.9931,
      "theoretical_loss": 4.075260621224099,
      "tokens_seen": 348323840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006191034613202194,
      "loss": 2.7803,
      "theoretical_loss": 4.075172846013789,
      "tokens_seen": 348389376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006190869112918479,
      "loss": 2.7226,
      "theoretical_loss": 4.075085091935675,
      "tokens_seen": 348454912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006190703612634765,
      "loss": 2.8195,
      "theoretical_loss": 4.074997358980694,
      "tokens_seen": 348520448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000619053811235105,
      "loss": 2.742,
      "theoretical_loss": 4.0749096471397905,
      "tokens_seen": 348585984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006190372612067336,
      "loss": 2.9389,
      "theoretical_loss": 4.074821956403917,
      "tokens_seen": 348651520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000619020711178362,
      "loss": 2.9588,
      "theoretical_loss": 4.074734286764029,
      "tokens_seen": 348717056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006190041611499905,
      "loss": 2.9387,
      "theoretical_loss": 4.074646638211087,
      "tokens_seen": 348782592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006189876111216191,
      "loss": 2.9559,
      "theoretical_loss": 4.074559010736059,
      "tokens_seen": 348848128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006189710610932476,
      "loss": 3.0074,
      "theoretical_loss": 4.074471404329918,
      "tokens_seen": 348913664
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006189545110648762,
      "loss": 3.2139,
      "theoretical_loss": 4.074383818983643,
      "tokens_seen": 348979200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006189379610365047,
      "loss": 2.8301,
      "theoretical_loss": 4.074296254688216,
      "tokens_seen": 349044736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006189214110081332,
      "loss": 2.9183,
      "theoretical_loss": 4.074208711434628,
      "tokens_seen": 349110272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006189048609797618,
      "loss": 3.0517,
      "theoretical_loss": 4.074121189213873,
      "tokens_seen": 349175808
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006188883109513901,
      "loss": 3.0024,
      "theoretical_loss": 4.074033688016953,
      "tokens_seen": 349241344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006188717609230187,
      "loss": 2.8264,
      "theoretical_loss": 4.073946207834872,
      "tokens_seen": 349306880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006188552108946472,
      "loss": 2.9747,
      "theoretical_loss": 4.073858748658644,
      "tokens_seen": 349372416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006188386608662757,
      "loss": 2.9924,
      "theoretical_loss": 4.073771310479284,
      "tokens_seen": 349437952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006188221108379043,
      "loss": 2.8755,
      "theoretical_loss": 4.073683893287816,
      "tokens_seen": 349503488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006188055608095328,
      "loss": 2.8213,
      "theoretical_loss": 4.073596497075267,
      "tokens_seen": 349569024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006187890107811614,
      "loss": 2.975,
      "theoretical_loss": 4.073509121832672,
      "tokens_seen": 349634560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006187724607527898,
      "loss": 2.6503,
      "theoretical_loss": 4.07342176755107,
      "tokens_seen": 349700096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006187559107244183,
      "loss": 2.9423,
      "theoretical_loss": 4.073334434221506,
      "tokens_seen": 349765632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006187393606960469,
      "loss": 3.0028,
      "theoretical_loss": 4.073247121835028,
      "tokens_seen": 349831168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006187228106676754,
      "loss": 2.9729,
      "theoretical_loss": 4.073159830382695,
      "tokens_seen": 349896704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0006187062606393039,
      "loss": 2.875,
      "theoretical_loss": 4.073072559855565,
      "tokens_seen": 349962240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006186897106109325,
      "loss": 2.9957,
      "theoretical_loss": 4.072985310244707,
      "tokens_seen": 350027776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000618673160582561,
      "loss": 2.8692,
      "theoretical_loss": 4.072898081541193,
      "tokens_seen": 350093312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006186566105541895,
      "loss": 2.9282,
      "theoretical_loss": 4.0728108737361,
      "tokens_seen": 350158848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000618640060525818,
      "loss": 2.8707,
      "theoretical_loss": 4.0727236868205114,
      "tokens_seen": 350224384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006186235104974465,
      "loss": 3.0036,
      "theoretical_loss": 4.072636520785515,
      "tokens_seen": 350289920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006186069604690751,
      "loss": 3.011,
      "theoretical_loss": 4.072549375622207,
      "tokens_seen": 350355456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006185904104407036,
      "loss": 2.9916,
      "theoretical_loss": 4.072462251321684,
      "tokens_seen": 350420992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006185738604123321,
      "loss": 2.9395,
      "theoretical_loss": 4.072375147875054,
      "tokens_seen": 350486528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006185573103839607,
      "loss": 3.0515,
      "theoretical_loss": 4.072288065273426,
      "tokens_seen": 350552064
    },
    {
      "epoch": 0.13,
      "objective/train/avg_token_score": 0.023719556629657745,
      "objective/train/avg_weight": 0.9721058011054993,
      "objective/train/docs_used": 210524,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1200644969940186,
      "objective/train/std_weight": 0.1370306760072708,
      "objective/train/theoretical_loss": 4.0722010035079155,
      "objective/train/tokens_used": 371077600,
      "theoretical_loss": 4.0722010035079155,
      "tokens_seen": 350617600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006185407603555891,
      "loss": 2.711,
      "theoretical_loss": 4.0722010035079155,
      "tokens_seen": 350617600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006185242103272177,
      "loss": 2.9968,
      "theoretical_loss": 4.072113962569645,
      "tokens_seen": 350683136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006185076602988462,
      "loss": 3.0746,
      "theoretical_loss": 4.07202694244974,
      "tokens_seen": 350748672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006184911102704747,
      "loss": 2.9386,
      "theoretical_loss": 4.071939943139334,
      "tokens_seen": 350814208
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006184745602421033,
      "loss": 2.6445,
      "theoretical_loss": 4.0718529646295645,
      "tokens_seen": 350879744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006184580102137318,
      "loss": 2.9857,
      "theoretical_loss": 4.071766006911574,
      "tokens_seen": 350945280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006184414601853603,
      "loss": 2.8581,
      "theoretical_loss": 4.071679069976511,
      "tokens_seen": 351010816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006184249101569889,
      "loss": 2.8057,
      "theoretical_loss": 4.07159215381553,
      "tokens_seen": 351076352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006184083601286173,
      "loss": 3.1321,
      "theoretical_loss": 4.07150525841979,
      "tokens_seen": 351141888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006183918101002459,
      "loss": 2.7122,
      "theoretical_loss": 4.071418383780456,
      "tokens_seen": 351207424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006183752600718744,
      "loss": 2.9182,
      "theoretical_loss": 4.071331529888698,
      "tokens_seen": 351272960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006183587100435029,
      "loss": 2.6686,
      "theoretical_loss": 4.071244696735692,
      "tokens_seen": 351338496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006183421600151315,
      "loss": 3.0463,
      "theoretical_loss": 4.071157884312619,
      "tokens_seen": 351404032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00061832560998676,
      "loss": 2.9287,
      "theoretical_loss": 4.071071092610664,
      "tokens_seen": 351469568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006183090599583885,
      "loss": 2.952,
      "theoretical_loss": 4.070984321621021,
      "tokens_seen": 351535104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000618292509930017,
      "loss": 3.0184,
      "theoretical_loss": 4.070897571334886,
      "tokens_seen": 351600640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006182759599016455,
      "loss": 3.1669,
      "theoretical_loss": 4.0708108417434605,
      "tokens_seen": 351666176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006182594098732741,
      "loss": 2.9408,
      "theoretical_loss": 4.070724132837954,
      "tokens_seen": 351731712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006182428598449026,
      "loss": 2.6937,
      "theoretical_loss": 4.07063744460958,
      "tokens_seen": 351797248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006182263098165311,
      "loss": 3.0755,
      "theoretical_loss": 4.070550777049555,
      "tokens_seen": 351862784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006182097597881597,
      "loss": 2.9414,
      "theoretical_loss": 4.070464130149106,
      "tokens_seen": 351928320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006181932097597882,
      "loss": 2.7585,
      "theoretical_loss": 4.07037750389946,
      "tokens_seen": 351993856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006181766597314166,
      "loss": 3.0359,
      "theoretical_loss": 4.070290898291852,
      "tokens_seen": 352059392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006181601097030452,
      "loss": 3.1169,
      "theoretical_loss": 4.070204313317523,
      "tokens_seen": 352124928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006181435596746737,
      "loss": 2.9417,
      "theoretical_loss": 4.07011774896772,
      "tokens_seen": 352190464
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006181270096463023,
      "loss": 3.2522,
      "theoretical_loss": 4.07003120523369,
      "tokens_seen": 352256000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006181104596179308,
      "loss": 2.8611,
      "theoretical_loss": 4.069944682106691,
      "tokens_seen": 352321536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006180939095895593,
      "loss": 2.7865,
      "theoretical_loss": 4.0698581795779845,
      "tokens_seen": 352387072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006180773595611879,
      "loss": 2.882,
      "theoretical_loss": 4.069771697638838,
      "tokens_seen": 352452608
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006180608095328164,
      "loss": 2.6856,
      "theoretical_loss": 4.069685236280522,
      "tokens_seen": 352518144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006180442595044448,
      "loss": 2.8825,
      "theoretical_loss": 4.069598795494315,
      "tokens_seen": 352583680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006180277094760734,
      "loss": 3.0027,
      "theoretical_loss": 4.0695123752715,
      "tokens_seen": 352649216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006180111594477019,
      "loss": 2.8475,
      "theoretical_loss": 4.069425975603364,
      "tokens_seen": 352714752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006179946094193305,
      "loss": 2.8278,
      "theoretical_loss": 4.069339596481201,
      "tokens_seen": 352780288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000617978059390959,
      "loss": 2.8779,
      "theoretical_loss": 4.0692532378963095,
      "tokens_seen": 352845824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006179615093625875,
      "loss": 2.9307,
      "theoretical_loss": 4.069166899839994,
      "tokens_seen": 352911360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006179449593342161,
      "loss": 2.721,
      "theoretical_loss": 4.069080582303563,
      "tokens_seen": 352976896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006179284093058445,
      "loss": 3.0611,
      "theoretical_loss": 4.068994285278332,
      "tokens_seen": 353042432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000617911859277473,
      "loss": 3.0472,
      "theoretical_loss": 4.068908008755619,
      "tokens_seen": 353107968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006178953092491015,
      "loss": 2.8473,
      "theoretical_loss": 4.068821752726751,
      "tokens_seen": 353173504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006178787592207301,
      "loss": 2.9486,
      "theoretical_loss": 4.0687355171830575,
      "tokens_seen": 353239040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006178622091923586,
      "loss": 2.7418,
      "theoretical_loss": 4.068649302115875,
      "tokens_seen": 353304576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006178456591639871,
      "loss": 3.0651,
      "theoretical_loss": 4.068563107516544,
      "tokens_seen": 353370112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006178291091356157,
      "loss": 3.072,
      "theoretical_loss": 4.068476933376411,
      "tokens_seen": 353435648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006178125591072441,
      "loss": 2.6907,
      "theoretical_loss": 4.068390779686826,
      "tokens_seen": 353501184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006177960090788726,
      "loss": 3.1106,
      "theoretical_loss": 4.068304646439148,
      "tokens_seen": 353566720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006177794590505012,
      "loss": 2.741,
      "theoretical_loss": 4.068218533624737,
      "tokens_seen": 353632256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006177629090221297,
      "loss": 2.8484,
      "theoretical_loss": 4.068132441234962,
      "tokens_seen": 353697792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006177463589937583,
      "loss": 3.1077,
      "theoretical_loss": 4.068046369261194,
      "tokens_seen": 353763328
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006177298089653868,
      "loss": 2.855,
      "theoretical_loss": 4.067960317694812,
      "tokens_seen": 353828864
    },
    {
      "epoch": 0.13,
      "objective/train/avg_token_score": 0.01628197357058525,
      "objective/train/avg_weight": 0.9808387160301208,
      "objective/train/docs_used": 211739,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2036495208740234,
      "objective/train/std_weight": 0.0901326909661293,
      "objective/train/theoretical_loss": 4.067874286527197,
      "objective/train/tokens_used": 374354400,
      "theoretical_loss": 4.067874286527197,
      "tokens_seen": 353894400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006177132589370153,
      "loss": 2.9786,
      "theoretical_loss": 4.067874286527197,
      "tokens_seen": 353894400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006176967089086439,
      "loss": 3.0932,
      "theoretical_loss": 4.06778827574974,
      "tokens_seen": 353959936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006176801588802723,
      "loss": 3.2189,
      "theoretical_loss": 4.067702285353833,
      "tokens_seen": 354025472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006176636088519008,
      "loss": 2.8618,
      "theoretical_loss": 4.0676163153308735,
      "tokens_seen": 354091008
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006176470588235294,
      "loss": 3.0996,
      "theoretical_loss": 4.067530365672267,
      "tokens_seen": 354156544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006176305087951579,
      "loss": 2.9975,
      "theoretical_loss": 4.067444436369422,
      "tokens_seen": 354222080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006176139587667865,
      "loss": 2.8404,
      "theoretical_loss": 4.067358527413753,
      "tokens_seen": 354287616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000617597408738415,
      "loss": 2.7468,
      "theoretical_loss": 4.06727263879668,
      "tokens_seen": 354353152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006175808587100435,
      "loss": 3.0427,
      "theoretical_loss": 4.067186770509627,
      "tokens_seen": 354418688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000617564308681672,
      "loss": 2.975,
      "theoretical_loss": 4.067100922544023,
      "tokens_seen": 354484224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006175477586533005,
      "loss": 2.8142,
      "theoretical_loss": 4.067015094891306,
      "tokens_seen": 354549760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000617531208624929,
      "loss": 2.7654,
      "theoretical_loss": 4.066929287542914,
      "tokens_seen": 354615296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006175146585965576,
      "loss": 2.8528,
      "theoretical_loss": 4.066843500490292,
      "tokens_seen": 354680832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006174981085681861,
      "loss": 2.9847,
      "theoretical_loss": 4.066757733724893,
      "tokens_seen": 354746368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006174815585398147,
      "loss": 3.0207,
      "theoretical_loss": 4.066671987238171,
      "tokens_seen": 354811904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006174650085114432,
      "loss": 2.8963,
      "theoretical_loss": 4.0665862610215875,
      "tokens_seen": 354877440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006174484584830716,
      "loss": 2.8873,
      "theoretical_loss": 4.066500555066609,
      "tokens_seen": 354942976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006174319084547002,
      "loss": 2.8729,
      "theoretical_loss": 4.066414869364707,
      "tokens_seen": 355008512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006174153584263287,
      "loss": 2.8813,
      "theoretical_loss": 4.066329203907358,
      "tokens_seen": 355074048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006173988083979572,
      "loss": 2.8852,
      "theoretical_loss": 4.066243558686043,
      "tokens_seen": 355139584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006173822583695858,
      "loss": 2.8815,
      "theoretical_loss": 4.066157933692249,
      "tokens_seen": 355205120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006173657083412143,
      "loss": 2.8374,
      "theoretical_loss": 4.066072328917469,
      "tokens_seen": 355270656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006173491583128429,
      "loss": 2.6957,
      "theoretical_loss": 4.0659867443532,
      "tokens_seen": 355336192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006173326082844714,
      "loss": 2.7301,
      "theoretical_loss": 4.065901179990943,
      "tokens_seen": 355401728
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006173160582560998,
      "loss": 3.014,
      "theoretical_loss": 4.065815635822206,
      "tokens_seen": 355467264
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006172995082277284,
      "loss": 2.8348,
      "theoretical_loss": 4.065730111838502,
      "tokens_seen": 355532800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006172829581993569,
      "loss": 2.8401,
      "theoretical_loss": 4.0656446080313495,
      "tokens_seen": 355598336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006172664081709854,
      "loss": 3.0022,
      "theoretical_loss": 4.065559124392269,
      "tokens_seen": 355663872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000617249858142614,
      "loss": 2.8854,
      "theoretical_loss": 4.065473660912791,
      "tokens_seen": 355729408
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006172333081142425,
      "loss": 3.0903,
      "theoretical_loss": 4.065388217584447,
      "tokens_seen": 355794944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006172167580858711,
      "loss": 3.0647,
      "theoretical_loss": 4.065302794398775,
      "tokens_seen": 355860480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006172002080574995,
      "loss": 3.0596,
      "theoretical_loss": 4.0652173913473195,
      "tokens_seen": 355926016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000617183658029128,
      "loss": 2.833,
      "theoretical_loss": 4.065132008421628,
      "tokens_seen": 355991552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006171671080007566,
      "loss": 2.7072,
      "theoretical_loss": 4.065046645613254,
      "tokens_seen": 356057088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006171505579723851,
      "loss": 2.8749,
      "theoretical_loss": 4.0649613029137575,
      "tokens_seen": 356122624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006171340079440136,
      "loss": 2.8984,
      "theoretical_loss": 4.0648759803147,
      "tokens_seen": 356188160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006171174579156422,
      "loss": 2.7478,
      "theoretical_loss": 4.064790677807652,
      "tokens_seen": 356253696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006171009078872707,
      "loss": 2.8818,
      "theoretical_loss": 4.064705395384187,
      "tokens_seen": 356319232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006170843578588993,
      "loss": 2.9733,
      "theoretical_loss": 4.064620133035883,
      "tokens_seen": 356384768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006170678078305277,
      "loss": 2.989,
      "theoretical_loss": 4.0645348907543255,
      "tokens_seen": 356450304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006170512578021562,
      "loss": 2.9495,
      "theoretical_loss": 4.064449668531103,
      "tokens_seen": 356515840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006170347077737848,
      "loss": 2.8543,
      "theoretical_loss": 4.064364466357809,
      "tokens_seen": 356581376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006170181577454133,
      "loss": 2.9483,
      "theoretical_loss": 4.064279284226044,
      "tokens_seen": 356646912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006170016077170418,
      "loss": 2.8246,
      "theoretical_loss": 4.064194122127411,
      "tokens_seen": 356712448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006169850576886704,
      "loss": 2.8645,
      "theoretical_loss": 4.06410898005352,
      "tokens_seen": 356777984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006169685076602989,
      "loss": 2.9775,
      "theoretical_loss": 4.064023857995986,
      "tokens_seen": 356843520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006169519576319274,
      "loss": 3.1454,
      "theoretical_loss": 4.0639387559464275,
      "tokens_seen": 356909056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006169354076035559,
      "loss": 3.0565,
      "theoretical_loss": 4.063853673896469,
      "tokens_seen": 356974592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006169188575751844,
      "loss": 3.0403,
      "theoretical_loss": 4.063768611837742,
      "tokens_seen": 357040128
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006169023075468129,
      "loss": 2.8544,
      "theoretical_loss": 4.063683569761879,
      "tokens_seen": 357105664
    },
    {
      "epoch": 0.13,
      "objective/train/avg_token_score": 0.007791871204972267,
      "objective/train/avg_weight": 0.9908267855644226,
      "objective/train/docs_used": 213839,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6770308017730713,
      "objective/train/std_weight": 0.06348834186792374,
      "objective/train/theoretical_loss": 4.063598547660519,
      "objective/train/tokens_used": 377631200,
      "theoretical_loss": 4.063598547660519,
      "tokens_seen": 357171200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006168857575184414,
      "loss": 2.733,
      "theoretical_loss": 4.063598547660519,
      "tokens_seen": 357171200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00061686920749007,
      "loss": 2.8529,
      "theoretical_loss": 4.06351354552531,
      "tokens_seen": 357236736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006168526574616985,
      "loss": 2.8399,
      "theoretical_loss": 4.063428563347898,
      "tokens_seen": 357302272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006168361074333269,
      "loss": 3.0472,
      "theoretical_loss": 4.063343601119941,
      "tokens_seen": 357367808
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006168195574049555,
      "loss": 2.9812,
      "theoretical_loss": 4.063258658833096,
      "tokens_seen": 357433344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000616803007376584,
      "loss": 2.8839,
      "theoretical_loss": 4.06317373647903,
      "tokens_seen": 357498880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006167864573482126,
      "loss": 2.8104,
      "theoretical_loss": 4.0630888340494105,
      "tokens_seen": 357564416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006167699073198411,
      "loss": 2.9536,
      "theoretical_loss": 4.063003951535913,
      "tokens_seen": 357629952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006167533572914696,
      "loss": 3.1242,
      "theoretical_loss": 4.06291908893022,
      "tokens_seen": 357695488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006167368072630982,
      "loss": 2.9756,
      "theoretical_loss": 4.062834246224012,
      "tokens_seen": 357761024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006167202572347266,
      "loss": 3.0253,
      "theoretical_loss": 4.062749423408981,
      "tokens_seen": 357826560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006167037072063551,
      "loss": 2.8265,
      "theoretical_loss": 4.062664620476822,
      "tokens_seen": 357892096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006166871571779837,
      "loss": 3.0953,
      "theoretical_loss": 4.062579837419235,
      "tokens_seen": 357957632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006166706071496122,
      "loss": 2.8888,
      "theoretical_loss": 4.062495074227923,
      "tokens_seen": 358023168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006166540571212408,
      "loss": 2.8223,
      "theoretical_loss": 4.062410330894597,
      "tokens_seen": 358088704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006166375070928693,
      "loss": 3.0507,
      "theoretical_loss": 4.062325607410972,
      "tokens_seen": 358154240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006166209570644978,
      "loss": 2.8219,
      "theoretical_loss": 4.062240903768767,
      "tokens_seen": 358219776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006166044070361264,
      "loss": 3.0103,
      "theoretical_loss": 4.062156219959705,
      "tokens_seen": 358285312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006165878570077548,
      "loss": 3.0025,
      "theoretical_loss": 4.062071555975519,
      "tokens_seen": 358350848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006165713069793833,
      "loss": 3.1968,
      "theoretical_loss": 4.061986911807942,
      "tokens_seen": 358416384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006165547569510119,
      "loss": 2.9794,
      "theoretical_loss": 4.061902287448714,
      "tokens_seen": 358481920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006165382069226404,
      "loss": 3.0388,
      "theoretical_loss": 4.061817682889577,
      "tokens_seen": 358547456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000616521656894269,
      "loss": 2.9156,
      "theoretical_loss": 4.061733098122284,
      "tokens_seen": 358612992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006165051068658975,
      "loss": 3.0486,
      "theoretical_loss": 4.061648533138587,
      "tokens_seen": 358678528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000616488556837526,
      "loss": 3.0424,
      "theoretical_loss": 4.0615639879302465,
      "tokens_seen": 358744064
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006164720068091545,
      "loss": 2.9747,
      "theoretical_loss": 4.061479462489026,
      "tokens_seen": 358809600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000616455456780783,
      "loss": 2.8795,
      "theoretical_loss": 4.061394956806694,
      "tokens_seen": 358875136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006164389067524116,
      "loss": 3.0494,
      "theoretical_loss": 4.061310470875027,
      "tokens_seen": 358940672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006164223567240401,
      "loss": 2.8636,
      "theoretical_loss": 4.0612260046858015,
      "tokens_seen": 359006208
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006164058066956686,
      "loss": 2.785,
      "theoretical_loss": 4.061141558230803,
      "tokens_seen": 359071744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006163892566672972,
      "loss": 2.9051,
      "theoretical_loss": 4.06105713150182,
      "tokens_seen": 359137280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006163727066389257,
      "loss": 2.9354,
      "theoretical_loss": 4.060972724490646,
      "tokens_seen": 359202816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006163561566105541,
      "loss": 3.128,
      "theoretical_loss": 4.0608883371890805,
      "tokens_seen": 359268352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006163396065821827,
      "loss": 2.6105,
      "theoretical_loss": 4.060803969588925,
      "tokens_seen": 359333888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006163230565538112,
      "loss": 2.8588,
      "theoretical_loss": 4.06071962168199,
      "tokens_seen": 359399424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006163065065254398,
      "loss": 3.17,
      "theoretical_loss": 4.060635293460089,
      "tokens_seen": 359464960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006162899564970683,
      "loss": 2.8177,
      "theoretical_loss": 4.060550984915039,
      "tokens_seen": 359530496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006162734064686968,
      "loss": 2.9553,
      "theoretical_loss": 4.060466696038664,
      "tokens_seen": 359596032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006162568564403254,
      "loss": 3.1269,
      "theoretical_loss": 4.060382426822792,
      "tokens_seen": 359661568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006162403064119539,
      "loss": 2.6659,
      "theoretical_loss": 4.060298177259256,
      "tokens_seen": 359727104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006162237563835823,
      "loss": 3.0039,
      "theoretical_loss": 4.060213947339894,
      "tokens_seen": 359792640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006162072063552109,
      "loss": 3.0088,
      "theoretical_loss": 4.060129737056547,
      "tokens_seen": 359858176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006161906563268394,
      "loss": 2.7505,
      "theoretical_loss": 4.060045546401065,
      "tokens_seen": 359923712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000616174106298468,
      "loss": 2.987,
      "theoretical_loss": 4.0599613753653,
      "tokens_seen": 359989248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006161575562700965,
      "loss": 2.8119,
      "theoretical_loss": 4.0598772239411085,
      "tokens_seen": 360054784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000616141006241725,
      "loss": 3.0074,
      "theoretical_loss": 4.059793092120353,
      "tokens_seen": 360120320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006161244562133536,
      "loss": 2.9664,
      "theoretical_loss": 4.0597089798949,
      "tokens_seen": 360185856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000616107906184982,
      "loss": 3.0599,
      "theoretical_loss": 4.059624887256622,
      "tokens_seen": 360251392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006160913561566105,
      "loss": 2.9448,
      "theoretical_loss": 4.059540814197396,
      "tokens_seen": 360316928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006160748061282391,
      "loss": 2.9914,
      "theoretical_loss": 4.059456760709104,
      "tokens_seen": 360382464
    },
    {
      "epoch": 0.13,
      "objective/train/avg_token_score": 0.015419309958815575,
      "objective/train/avg_weight": 0.9818492531776428,
      "objective/train/docs_used": 215849,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.869628429412842,
      "objective/train/std_weight": 0.09784294664859772,
      "objective/train/theoretical_loss": 4.05937272678363,
      "objective/train/tokens_used": 380908000,
      "theoretical_loss": 4.05937272678363,
      "tokens_seen": 360448000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006160582560998676,
      "loss": 2.9915,
      "theoretical_loss": 4.05937272678363,
      "tokens_seen": 360448000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006160417060714962,
      "loss": 2.7699,
      "theoretical_loss": 4.059288712412868,
      "tokens_seen": 360513536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006160251560431247,
      "loss": 3.0565,
      "theoretical_loss": 4.059204717588712,
      "tokens_seen": 360579072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006160086060147532,
      "loss": 2.912,
      "theoretical_loss": 4.059120742303065,
      "tokens_seen": 360644608
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006159920559863818,
      "loss": 2.6967,
      "theoretical_loss": 4.059036786547831,
      "tokens_seen": 360710144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006159755059580102,
      "loss": 2.9639,
      "theoretical_loss": 4.058952850314922,
      "tokens_seen": 360775680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006159589559296387,
      "loss": 2.9119,
      "theoretical_loss": 4.058868933596252,
      "tokens_seen": 360841216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006159424059012673,
      "loss": 3.0569,
      "theoretical_loss": 4.0587850363837426,
      "tokens_seen": 360906752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006159258558728957,
      "loss": 2.7222,
      "theoretical_loss": 4.058701158669318,
      "tokens_seen": 360972288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006159093058445243,
      "loss": 2.9952,
      "theoretical_loss": 4.058617300444908,
      "tokens_seen": 361037824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006158927558161528,
      "loss": 2.8335,
      "theoretical_loss": 4.0585334617024476,
      "tokens_seen": 361103360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006158762057877814,
      "loss": 3.2127,
      "theoretical_loss": 4.058449642433875,
      "tokens_seen": 361168896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006158596557594098,
      "loss": 2.7841,
      "theoretical_loss": 4.058365842631137,
      "tokens_seen": 361234432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006158431057310383,
      "loss": 3.0753,
      "theoretical_loss": 4.05828206228618,
      "tokens_seen": 361299968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006158265557026669,
      "loss": 2.869,
      "theoretical_loss": 4.058198301390959,
      "tokens_seen": 361365504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006158100056742954,
      "loss": 2.9628,
      "theoretical_loss": 4.058114559937433,
      "tokens_seen": 361431040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006157934556459239,
      "loss": 2.902,
      "theoretical_loss": 4.058030837917565,
      "tokens_seen": 361496576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006157769056175525,
      "loss": 3.1816,
      "theoretical_loss": 4.057947135323323,
      "tokens_seen": 361562112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000615760355589181,
      "loss": 3.0783,
      "theoretical_loss": 4.057863452146679,
      "tokens_seen": 361627648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006157438055608095,
      "loss": 2.8034,
      "theoretical_loss": 4.057779788379613,
      "tokens_seen": 361693184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000615727255532438,
      "loss": 2.9349,
      "theoretical_loss": 4.057696144014106,
      "tokens_seen": 361758720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006157107055040665,
      "loss": 3.0719,
      "theoretical_loss": 4.057612519042145,
      "tokens_seen": 361824256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006156941554756951,
      "loss": 2.7926,
      "theoretical_loss": 4.057528913455723,
      "tokens_seen": 361889792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006156776054473236,
      "loss": 2.9898,
      "theoretical_loss": 4.057445327246836,
      "tokens_seen": 361955328
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006156610554189521,
      "loss": 2.7688,
      "theoretical_loss": 4.0573617604074865,
      "tokens_seen": 362020864
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006156445053905807,
      "loss": 3.1064,
      "theoretical_loss": 4.0572782129296785,
      "tokens_seen": 362086400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006156279553622091,
      "loss": 3.1065,
      "theoretical_loss": 4.057194684805426,
      "tokens_seen": 362151936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006156114053338377,
      "loss": 2.8998,
      "theoretical_loss": 4.057111176026742,
      "tokens_seen": 362217472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006155948553054662,
      "loss": 2.9904,
      "theoretical_loss": 4.0570276865856485,
      "tokens_seen": 362283008
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006155783052770947,
      "loss": 3.0736,
      "theoretical_loss": 4.056944216474171,
      "tokens_seen": 362348544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006155617552487233,
      "loss": 2.9726,
      "theoretical_loss": 4.056860765684338,
      "tokens_seen": 362414080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006155452052203518,
      "loss": 2.7407,
      "theoretical_loss": 4.056777334208186,
      "tokens_seen": 362479616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006155286551919803,
      "loss": 2.8925,
      "theoretical_loss": 4.056693922037753,
      "tokens_seen": 362545152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006155121051636089,
      "loss": 3.0585,
      "theoretical_loss": 4.056610529165083,
      "tokens_seen": 362610688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006154955551352373,
      "loss": 2.7556,
      "theoretical_loss": 4.056527155582226,
      "tokens_seen": 362676224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006154790051068659,
      "loss": 2.9511,
      "theoretical_loss": 4.056443801281235,
      "tokens_seen": 362741760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006154624550784944,
      "loss": 2.926,
      "theoretical_loss": 4.056360466254168,
      "tokens_seen": 362807296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006154459050501229,
      "loss": 3.0382,
      "theoretical_loss": 4.056277150493088,
      "tokens_seen": 362872832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006154293550217515,
      "loss": 2.9447,
      "theoretical_loss": 4.056193853990063,
      "tokens_seen": 362938368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00061541280499338,
      "loss": 2.8544,
      "theoretical_loss": 4.056110576737166,
      "tokens_seen": 363003904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006153962549650085,
      "loss": 3.0026,
      "theoretical_loss": 4.056027318726472,
      "tokens_seen": 363069440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000615379704936637,
      "loss": 3.0533,
      "theoretical_loss": 4.055944079950065,
      "tokens_seen": 363134976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006153631549082655,
      "loss": 2.7591,
      "theoretical_loss": 4.0558608604000295,
      "tokens_seen": 363200512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006153466048798941,
      "loss": 3.0994,
      "theoretical_loss": 4.055777660068458,
      "tokens_seen": 363266048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006153300548515226,
      "loss": 2.8444,
      "theoretical_loss": 4.055694478947445,
      "tokens_seen": 363331584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006153135048231511,
      "loss": 2.7773,
      "theoretical_loss": 4.055611317029093,
      "tokens_seen": 363397120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006152969547947797,
      "loss": 3.0295,
      "theoretical_loss": 4.055528174305505,
      "tokens_seen": 363462656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006152804047664082,
      "loss": 2.7628,
      "theoretical_loss": 4.055445050768792,
      "tokens_seen": 363528192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006152638547380366,
      "loss": 2.678,
      "theoretical_loss": 4.055361946411068,
      "tokens_seen": 363593728
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006152473047096652,
      "loss": 2.9856,
      "theoretical_loss": 4.055278861224452,
      "tokens_seen": 363659264
    },
    {
      "epoch": 0.13,
      "objective/train/avg_token_score": 0.03244902566075325,
      "objective/train/avg_weight": 0.9618296027183533,
      "objective/train/docs_used": 217416,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.267622232437134,
      "objective/train/std_weight": 0.1261177510023117,
      "objective/train/theoretical_loss": 4.055195795201069,
      "objective/train/tokens_used": 384184800,
      "theoretical_loss": 4.055195795201069,
      "tokens_seen": 363724800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006152307546812937,
      "loss": 2.9655,
      "theoretical_loss": 4.055195795201069,
      "tokens_seen": 363724800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006152142046529223,
      "loss": 2.9696,
      "theoretical_loss": 4.055112748333046,
      "tokens_seen": 363790336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006151976546245508,
      "loss": 2.9974,
      "theoretical_loss": 4.0550297206125165,
      "tokens_seen": 363855872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006151811045961793,
      "loss": 2.9921,
      "theoretical_loss": 4.054946712031619,
      "tokens_seen": 363921408
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006151645545678079,
      "loss": 2.8795,
      "theoretical_loss": 4.054863722582494,
      "tokens_seen": 363986944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006151480045394364,
      "loss": 2.8352,
      "theoretical_loss": 4.054780752257291,
      "tokens_seen": 364052480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006151314545110648,
      "loss": 2.8331,
      "theoretical_loss": 4.05469780104816,
      "tokens_seen": 364118016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006151149044826934,
      "loss": 3.0073,
      "theoretical_loss": 4.054614868947258,
      "tokens_seen": 364183552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006150983544543219,
      "loss": 2.8915,
      "theoretical_loss": 4.054531955946745,
      "tokens_seen": 364249088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006150818044259505,
      "loss": 3.0262,
      "theoretical_loss": 4.054449062038787,
      "tokens_seen": 364314624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000615065254397579,
      "loss": 2.8599,
      "theoretical_loss": 4.054366187215556,
      "tokens_seen": 364380160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006150487043692075,
      "loss": 2.9226,
      "theoretical_loss": 4.0542833314692235,
      "tokens_seen": 364445696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006150321543408361,
      "loss": 2.7202,
      "theoretical_loss": 4.054200494791972,
      "tokens_seen": 364511232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006150156043124646,
      "loss": 2.9587,
      "theoretical_loss": 4.054117677175983,
      "tokens_seen": 364576768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000614999054284093,
      "loss": 3.0439,
      "theoretical_loss": 4.054034878613447,
      "tokens_seen": 364642304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006149825042557216,
      "loss": 2.9418,
      "theoretical_loss": 4.053952099096556,
      "tokens_seen": 364707840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006149659542273501,
      "loss": 2.89,
      "theoretical_loss": 4.053869338617508,
      "tokens_seen": 364773376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006149494041989786,
      "loss": 2.9068,
      "theoretical_loss": 4.053786597168506,
      "tokens_seen": 364838912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006149328541706071,
      "loss": 2.7496,
      "theoretical_loss": 4.053703874741757,
      "tokens_seen": 364904448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006149163041422357,
      "loss": 3.1701,
      "theoretical_loss": 4.053621171329472,
      "tokens_seen": 364969984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006148997541138642,
      "loss": 2.615,
      "theoretical_loss": 4.053538486923868,
      "tokens_seen": 365035520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006148832040854926,
      "loss": 2.7956,
      "theoretical_loss": 4.053455821517165,
      "tokens_seen": 365101056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006148666540571212,
      "loss": 2.75,
      "theoretical_loss": 4.053373175101589,
      "tokens_seen": 365166592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006148501040287497,
      "loss": 2.9542,
      "theoretical_loss": 4.053290547669369,
      "tokens_seen": 365232128
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006148335540003783,
      "loss": 2.9275,
      "theoretical_loss": 4.05320793921274,
      "tokens_seen": 365297664
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006148170039720068,
      "loss": 3.0149,
      "theoretical_loss": 4.053125349723942,
      "tokens_seen": 365363200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006148004539436353,
      "loss": 3.1628,
      "theoretical_loss": 4.053042779195217,
      "tokens_seen": 365428736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006147839039152639,
      "loss": 2.8581,
      "theoretical_loss": 4.052960227618815,
      "tokens_seen": 365494272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006147673538868923,
      "loss": 2.7924,
      "theoretical_loss": 4.052877694986987,
      "tokens_seen": 365559808
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006147508038585208,
      "loss": 2.9382,
      "theoretical_loss": 4.052795181291992,
      "tokens_seen": 365625344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006147342538301494,
      "loss": 2.787,
      "theoretical_loss": 4.05271268652609,
      "tokens_seen": 365690880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006147177038017779,
      "loss": 2.8592,
      "theoretical_loss": 4.052630210681549,
      "tokens_seen": 365756416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006147011537734065,
      "loss": 2.7821,
      "theoretical_loss": 4.052547753750639,
      "tokens_seen": 365821952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000614684603745035,
      "loss": 2.6533,
      "theoretical_loss": 4.0524653157256365,
      "tokens_seen": 365887488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006146680537166635,
      "loss": 2.9545,
      "theoretical_loss": 4.05238289659882,
      "tokens_seen": 365953024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000614651503688292,
      "loss": 2.8376,
      "theoretical_loss": 4.052300496362475,
      "tokens_seen": 366018560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006146349536599205,
      "loss": 3.0094,
      "theoretical_loss": 4.052218115008892,
      "tokens_seen": 366084096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000614618403631549,
      "loss": 3.1254,
      "theoretical_loss": 4.052135752530362,
      "tokens_seen": 366149632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006146018536031776,
      "loss": 2.7955,
      "theoretical_loss": 4.0520534089191855,
      "tokens_seen": 366215168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006145853035748061,
      "loss": 2.918,
      "theoretical_loss": 4.051971084167663,
      "tokens_seen": 366280704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006145687535464347,
      "loss": 2.9366,
      "theoretical_loss": 4.051888778268103,
      "tokens_seen": 366346240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006145522035180632,
      "loss": 2.9,
      "theoretical_loss": 4.051806491212817,
      "tokens_seen": 366411776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006145356534896917,
      "loss": 2.983,
      "theoretical_loss": 4.051724222994122,
      "tokens_seen": 366477312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006145191034613202,
      "loss": 2.6773,
      "theoretical_loss": 4.051641973604337,
      "tokens_seen": 366542848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006145025534329487,
      "loss": 2.9195,
      "theoretical_loss": 4.051559743035789,
      "tokens_seen": 366608384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006144860034045772,
      "loss": 2.8545,
      "theoretical_loss": 4.051477531280806,
      "tokens_seen": 366673920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006144694533762058,
      "loss": 2.8086,
      "theoretical_loss": 4.051395338331723,
      "tokens_seen": 366739456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006144529033478343,
      "loss": 2.7364,
      "theoretical_loss": 4.05131316418088,
      "tokens_seen": 366804992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006144363533194629,
      "loss": 2.5921,
      "theoretical_loss": 4.051231008820618,
      "tokens_seen": 366870528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006144198032910914,
      "loss": 2.623,
      "theoretical_loss": 4.051148872243286,
      "tokens_seen": 366936064
    },
    {
      "epoch": 0.13,
      "objective/train/avg_token_score": 0.011045483872294426,
      "objective/train/avg_weight": 0.9870135188102722,
      "objective/train/docs_used": 219277,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1950573921203613,
      "objective/train/std_weight": 0.0713377520442009,
      "objective/train/theoretical_loss": 4.051066754441235,
      "objective/train/tokens_used": 387461600,
      "theoretical_loss": 4.051066754441235,
      "tokens_seen": 367001600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006144032532627198,
      "loss": 2.9759,
      "theoretical_loss": 4.051066754441235,
      "tokens_seen": 367001600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006143867032343484,
      "loss": 2.6768,
      "theoretical_loss": 4.050984655406824,
      "tokens_seen": 367067136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006143701532059769,
      "loss": 2.8894,
      "theoretical_loss": 4.050902575132412,
      "tokens_seen": 367132672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006143536031776054,
      "loss": 3.1317,
      "theoretical_loss": 4.050820513610366,
      "tokens_seen": 367198208
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000614337053149234,
      "loss": 2.8885,
      "theoretical_loss": 4.050738470833054,
      "tokens_seen": 367263744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006143205031208625,
      "loss": 2.9459,
      "theoretical_loss": 4.0506564467928525,
      "tokens_seen": 367329280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006143039530924911,
      "loss": 2.8077,
      "theoretical_loss": 4.05057444148214,
      "tokens_seen": 367394816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006142874030641196,
      "loss": 2.9918,
      "theoretical_loss": 4.0504924548933,
      "tokens_seen": 367460352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000614270853035748,
      "loss": 2.7564,
      "theoretical_loss": 4.05041048701872,
      "tokens_seen": 367525888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006142543030073766,
      "loss": 2.806,
      "theoretical_loss": 4.050328537850792,
      "tokens_seen": 367591424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006142377529790051,
      "loss": 2.985,
      "theoretical_loss": 4.050246607381913,
      "tokens_seen": 367656960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006142212029506336,
      "loss": 2.7275,
      "theoretical_loss": 4.0501646956044866,
      "tokens_seen": 367722496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006142046529222622,
      "loss": 3.0474,
      "theoretical_loss": 4.050082802510915,
      "tokens_seen": 367788032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006141881028938907,
      "loss": 2.7624,
      "theoretical_loss": 4.050000928093611,
      "tokens_seen": 367853568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006141715528655193,
      "loss": 2.9636,
      "theoretical_loss": 4.049919072344987,
      "tokens_seen": 367919104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006141550028371477,
      "loss": 2.8107,
      "theoretical_loss": 4.049837235257463,
      "tokens_seen": 367984640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006141384528087762,
      "loss": 2.9543,
      "theoretical_loss": 4.049755416823463,
      "tokens_seen": 368050176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006141219027804048,
      "loss": 2.765,
      "theoretical_loss": 4.049673617035414,
      "tokens_seen": 368115712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006141053527520333,
      "loss": 2.8627,
      "theoretical_loss": 4.0495918358857494,
      "tokens_seen": 368181248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006140888027236618,
      "loss": 2.7941,
      "theoretical_loss": 4.049510073366903,
      "tokens_seen": 368246784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006140722526952904,
      "loss": 2.8329,
      "theoretical_loss": 4.04942832947132,
      "tokens_seen": 368312320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006140557026669189,
      "loss": 2.8983,
      "theoretical_loss": 4.049346604191443,
      "tokens_seen": 368377856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006140391526385475,
      "loss": 2.9721,
      "theoretical_loss": 4.049264897519723,
      "tokens_seen": 368443392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006140226026101759,
      "loss": 2.855,
      "theoretical_loss": 4.049183209448615,
      "tokens_seen": 368508928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006140060525818044,
      "loss": 3.2592,
      "theoretical_loss": 4.049101539970576,
      "tokens_seen": 368574464
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000613989502553433,
      "loss": 2.7909,
      "theoretical_loss": 4.04901988907807,
      "tokens_seen": 368640000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006139729525250614,
      "loss": 2.8937,
      "theoretical_loss": 4.048938256763565,
      "tokens_seen": 368705536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00061395640249669,
      "loss": 2.8215,
      "theoretical_loss": 4.048856643019532,
      "tokens_seen": 368771072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006139398524683185,
      "loss": 2.7512,
      "theoretical_loss": 4.048775047838447,
      "tokens_seen": 368836608
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000613923302439947,
      "loss": 3.127,
      "theoretical_loss": 4.048693471212792,
      "tokens_seen": 368902144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006139067524115755,
      "loss": 2.9025,
      "theoretical_loss": 4.048611913135052,
      "tokens_seen": 368967680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000613890202383204,
      "loss": 3.2139,
      "theoretical_loss": 4.048530373597716,
      "tokens_seen": 369033216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006138736523548326,
      "loss": 2.8988,
      "theoretical_loss": 4.048448852593277,
      "tokens_seen": 369098752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006138571023264611,
      "loss": 2.8764,
      "theoretical_loss": 4.048367350114235,
      "tokens_seen": 369164288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006138405522980896,
      "loss": 2.8443,
      "theoretical_loss": 4.0482858661530905,
      "tokens_seen": 369229824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006138240022697182,
      "loss": 3.0677,
      "theoretical_loss": 4.048204400702351,
      "tokens_seen": 369295360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006138074522413467,
      "loss": 2.9121,
      "theoretical_loss": 4.04812295375453,
      "tokens_seen": 369360896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006137909022129752,
      "loss": 2.7336,
      "theoretical_loss": 4.04804152530214,
      "tokens_seen": 369426432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006137743521846037,
      "loss": 2.8535,
      "theoretical_loss": 4.047960115337702,
      "tokens_seen": 369491968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006137578021562322,
      "loss": 2.7839,
      "theoretical_loss": 4.047878723853741,
      "tokens_seen": 369557504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006137412521278608,
      "loss": 3.0827,
      "theoretical_loss": 4.047797350842786,
      "tokens_seen": 369623040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006137247020994893,
      "loss": 3.0986,
      "theoretical_loss": 4.047715996297369,
      "tokens_seen": 369688576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006137081520711178,
      "loss": 2.7075,
      "theoretical_loss": 4.0476346602100275,
      "tokens_seen": 369754112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006136916020427464,
      "loss": 2.8222,
      "theoretical_loss": 4.047553342573304,
      "tokens_seen": 369819648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006136750520143748,
      "loss": 2.8878,
      "theoretical_loss": 4.047472043379743,
      "tokens_seen": 369885184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006136585019860034,
      "loss": 2.8454,
      "theoretical_loss": 4.047390762621897,
      "tokens_seen": 369950720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006136419519576319,
      "loss": 2.8306,
      "theoretical_loss": 4.04730950029232,
      "tokens_seen": 370016256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006136254019292604,
      "loss": 2.867,
      "theoretical_loss": 4.047228256383569,
      "tokens_seen": 370081792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000613608851900889,
      "loss": 2.8679,
      "theoretical_loss": 4.04714703088821,
      "tokens_seen": 370147328
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006135923018725175,
      "loss": 2.9109,
      "theoretical_loss": 4.0470658237988095,
      "tokens_seen": 370212864
    },
    {
      "epoch": 0.13,
      "objective/train/avg_token_score": 0.013134578242897987,
      "objective/train/avg_weight": 0.9845654964447021,
      "objective/train/docs_used": 221307,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.147231340408325,
      "objective/train/std_weight": 0.07831548899412155,
      "objective/train/theoretical_loss": 4.04698463510794,
      "objective/train/tokens_used": 390738400,
      "theoretical_loss": 4.04698463510794,
      "tokens_seen": 370278400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000613575751844146,
      "loss": 2.8589,
      "theoretical_loss": 4.04698463510794,
      "tokens_seen": 370278400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006135592018157746,
      "loss": 3.1672,
      "theoretical_loss": 4.046903464808177,
      "tokens_seen": 370343936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000613542651787403,
      "loss": 2.8571,
      "theoretical_loss": 4.046822312892102,
      "tokens_seen": 370409472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006135261017590316,
      "loss": 2.9972,
      "theoretical_loss": 4.046741179352299,
      "tokens_seen": 370475008
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006135095517306601,
      "loss": 2.7261,
      "theoretical_loss": 4.046660064181359,
      "tokens_seen": 370540544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006134930017022886,
      "loss": 2.8906,
      "theoretical_loss": 4.0465789673718735,
      "tokens_seen": 370606080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006134764516739172,
      "loss": 2.8364,
      "theoretical_loss": 4.0464978889164405,
      "tokens_seen": 370671616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006134599016455457,
      "loss": 2.936,
      "theoretical_loss": 4.046416828807663,
      "tokens_seen": 370737152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006134433516171742,
      "loss": 2.9876,
      "theoretical_loss": 4.046335787038148,
      "tokens_seen": 370802688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006134268015888027,
      "loss": 2.8147,
      "theoretical_loss": 4.046254763600503,
      "tokens_seen": 370868224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006134102515604312,
      "loss": 2.8468,
      "theoretical_loss": 4.046173758487347,
      "tokens_seen": 370933760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006133937015320598,
      "loss": 2.8261,
      "theoretical_loss": 4.046092771691297,
      "tokens_seen": 370999296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006133771515036883,
      "loss": 2.8064,
      "theoretical_loss": 4.046011803204976,
      "tokens_seen": 371064832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006133606014753168,
      "loss": 2.9385,
      "theoretical_loss": 4.045930853021014,
      "tokens_seen": 371130368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006133440514469454,
      "loss": 2.6248,
      "theoretical_loss": 4.045849921132042,
      "tokens_seen": 371195904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006133275014185739,
      "loss": 2.7923,
      "theoretical_loss": 4.045769007530695,
      "tokens_seen": 371261440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006133109513902023,
      "loss": 2.8974,
      "theoretical_loss": 4.045688112209615,
      "tokens_seen": 371326976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006132944013618309,
      "loss": 3.2413,
      "theoretical_loss": 4.045607235161447,
      "tokens_seen": 371392512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006132778513334594,
      "loss": 2.8673,
      "theoretical_loss": 4.04552637637884,
      "tokens_seen": 371458048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000613261301305088,
      "loss": 2.9053,
      "theoretical_loss": 4.045445535854446,
      "tokens_seen": 371523584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006132447512767165,
      "loss": 2.9911,
      "theoretical_loss": 4.045364713580923,
      "tokens_seen": 371589120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000613228201248345,
      "loss": 2.8242,
      "theoretical_loss": 4.045283909550935,
      "tokens_seen": 371654656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006132116512199736,
      "loss": 3.003,
      "theoretical_loss": 4.045203123757146,
      "tokens_seen": 371720192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000613195101191602,
      "loss": 3.0689,
      "theoretical_loss": 4.045122356192227,
      "tokens_seen": 371785728
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006131785511632305,
      "loss": 3.067,
      "theoretical_loss": 4.0450416068488515,
      "tokens_seen": 371851264
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006131620011348591,
      "loss": 3.0161,
      "theoretical_loss": 4.044960875719701,
      "tokens_seen": 371916800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006131454511064876,
      "loss": 2.806,
      "theoretical_loss": 4.0448801627974555,
      "tokens_seen": 371982336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006131289010781162,
      "loss": 2.8028,
      "theoretical_loss": 4.044799468074804,
      "tokens_seen": 372047872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006131123510497447,
      "loss": 2.6984,
      "theoretical_loss": 4.044718791544438,
      "tokens_seen": 372113408
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006130958010213732,
      "loss": 3.0639,
      "theoretical_loss": 4.044638133199052,
      "tokens_seen": 372178944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006130792509930018,
      "loss": 2.8914,
      "theoretical_loss": 4.044557493031347,
      "tokens_seen": 372244480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006130627009646302,
      "loss": 2.9005,
      "theoretical_loss": 4.044476871034027,
      "tokens_seen": 372310016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006130461509362587,
      "loss": 2.4987,
      "theoretical_loss": 4.0443962671998,
      "tokens_seen": 372375552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006130296009078873,
      "loss": 2.9107,
      "theoretical_loss": 4.044315681521379,
      "tokens_seen": 372441088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006130130508795158,
      "loss": 3.2756,
      "theoretical_loss": 4.04423511399148,
      "tokens_seen": 372506624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006129965008511444,
      "loss": 2.8977,
      "theoretical_loss": 4.044154564602826,
      "tokens_seen": 372572160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006129799508227728,
      "loss": 2.6506,
      "theoretical_loss": 4.044074033348139,
      "tokens_seen": 372637696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006129634007944014,
      "loss": 3.0538,
      "theoretical_loss": 4.043993520220151,
      "tokens_seen": 372703232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006129468507660298,
      "loss": 2.9348,
      "theoretical_loss": 4.0439130252115945,
      "tokens_seen": 372768768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006129303007376583,
      "loss": 2.8748,
      "theoretical_loss": 4.043832548315207,
      "tokens_seen": 372834304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006129137507092869,
      "loss": 2.8105,
      "theoretical_loss": 4.043752089523732,
      "tokens_seen": 372899840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006128972006809154,
      "loss": 2.8007,
      "theoretical_loss": 4.043671648829913,
      "tokens_seen": 372965376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006128806506525439,
      "loss": 2.8664,
      "theoretical_loss": 4.043591226226503,
      "tokens_seen": 373030912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006128641006241725,
      "loss": 3.0817,
      "theoretical_loss": 4.0435108217062545,
      "tokens_seen": 373096448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000612847550595801,
      "loss": 2.7472,
      "theoretical_loss": 4.043430435261927,
      "tokens_seen": 373161984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006128310005674296,
      "loss": 2.6841,
      "theoretical_loss": 4.043350066886283,
      "tokens_seen": 373227520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000612814450539058,
      "loss": 3.0322,
      "theoretical_loss": 4.043269716572091,
      "tokens_seen": 373293056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006127979005106865,
      "loss": 2.6099,
      "theoretical_loss": 4.04318938431212,
      "tokens_seen": 373358592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006127813504823151,
      "loss": 2.9373,
      "theoretical_loss": 4.043109070099146,
      "tokens_seen": 373424128
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006127648004539436,
      "loss": 2.9802,
      "theoretical_loss": 4.043028773925949,
      "tokens_seen": 373489664
    },
    {
      "epoch": 0.13,
      "objective/train/avg_token_score": 0.017580652609467506,
      "objective/train/avg_weight": 0.9793002009391785,
      "objective/train/docs_used": 223312,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.183922290802002,
      "objective/train/std_weight": 0.10959427803754807,
      "objective/train/theoretical_loss": 4.042948495785312,
      "objective/train/tokens_used": 394015200,
      "theoretical_loss": 4.042948495785312,
      "tokens_seen": 373555200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006127482504255721,
      "loss": 2.9827,
      "theoretical_loss": 4.042948495785312,
      "tokens_seen": 373555200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006127317003972007,
      "loss": 2.9422,
      "theoretical_loss": 4.042868235670023,
      "tokens_seen": 373620736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006127151503688292,
      "loss": 3.0622,
      "theoretical_loss": 4.042787993572875,
      "tokens_seen": 373686272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006126986003404577,
      "loss": 2.7493,
      "theoretical_loss": 4.042707769486663,
      "tokens_seen": 373751808
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006126820503120862,
      "loss": 2.6954,
      "theoretical_loss": 4.042627563404187,
      "tokens_seen": 373817344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006126655002837147,
      "loss": 2.9236,
      "theoretical_loss": 4.042547375318251,
      "tokens_seen": 373882880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006126489502553433,
      "loss": 3.0439,
      "theoretical_loss": 4.042467205221665,
      "tokens_seen": 373948416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006126324002269718,
      "loss": 2.8436,
      "theoretical_loss": 4.042387053107241,
      "tokens_seen": 374013952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006126158501986003,
      "loss": 2.893,
      "theoretical_loss": 4.042306918967795,
      "tokens_seen": 374079488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006125993001702289,
      "loss": 2.8541,
      "theoretical_loss": 4.042226802796149,
      "tokens_seen": 374145024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006125827501418573,
      "loss": 2.9338,
      "theoretical_loss": 4.042146704585127,
      "tokens_seen": 374210560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006125662001134859,
      "loss": 2.9826,
      "theoretical_loss": 4.042066624327559,
      "tokens_seen": 374276096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006125496500851144,
      "loss": 3.1085,
      "theoretical_loss": 4.041986562016278,
      "tokens_seen": 374341632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006125331000567429,
      "loss": 2.9082,
      "theoretical_loss": 4.041906517644121,
      "tokens_seen": 374407168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006125165500283715,
      "loss": 2.8645,
      "theoretical_loss": 4.0418264912039295,
      "tokens_seen": 374472704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006125,
      "loss": 2.886,
      "theoretical_loss": 4.04174648268855,
      "tokens_seen": 374538240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006124834499716286,
      "loss": 2.8042,
      "theoretical_loss": 4.041666492090831,
      "tokens_seen": 374603776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006124668999432571,
      "loss": 2.9901,
      "theoretical_loss": 4.041586519403627,
      "tokens_seen": 374669312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006124503499148855,
      "loss": 2.7557,
      "theoretical_loss": 4.041506564619796,
      "tokens_seen": 374734848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006124337998865141,
      "loss": 2.9729,
      "theoretical_loss": 4.0414266277322,
      "tokens_seen": 374800384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006124172498581426,
      "loss": 2.7396,
      "theoretical_loss": 4.041346708733704,
      "tokens_seen": 374865920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006124006998297711,
      "loss": 2.8081,
      "theoretical_loss": 4.041266807617179,
      "tokens_seen": 374931456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006123841498013997,
      "loss": 2.933,
      "theoretical_loss": 4.041186924375499,
      "tokens_seen": 374996992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006123675997730282,
      "loss": 2.6217,
      "theoretical_loss": 4.041107059001544,
      "tokens_seen": 375062528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006123510497446568,
      "loss": 2.8935,
      "theoretical_loss": 4.041027211488194,
      "tokens_seen": 375128064
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006123344997162852,
      "loss": 2.9124,
      "theoretical_loss": 4.040947381828337,
      "tokens_seen": 375193600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006123179496879137,
      "loss": 2.8301,
      "theoretical_loss": 4.040867570014862,
      "tokens_seen": 375259136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006123013996595423,
      "loss": 2.74,
      "theoretical_loss": 4.040787776040666,
      "tokens_seen": 375324672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006122848496311708,
      "loss": 2.8485,
      "theoretical_loss": 4.040707999898647,
      "tokens_seen": 375390208
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006122682996027993,
      "loss": 2.8611,
      "theoretical_loss": 4.040628241581706,
      "tokens_seen": 375455744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006122517495744279,
      "loss": 2.9283,
      "theoretical_loss": 4.0405485010827515,
      "tokens_seen": 375521280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006122351995460564,
      "loss": 2.8711,
      "theoretical_loss": 4.040468778394694,
      "tokens_seen": 375586816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000612218649517685,
      "loss": 2.7628,
      "theoretical_loss": 4.040389073510449,
      "tokens_seen": 375652352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006122020994893134,
      "loss": 2.9323,
      "theoretical_loss": 4.040309386422934,
      "tokens_seen": 375717888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006121855494609419,
      "loss": 2.6567,
      "theoretical_loss": 4.040229717125074,
      "tokens_seen": 375783424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006121689994325705,
      "loss": 3.0078,
      "theoretical_loss": 4.0401500656097955,
      "tokens_seen": 375848960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000612152449404199,
      "loss": 2.9116,
      "theoretical_loss": 4.040070431870029,
      "tokens_seen": 375914496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006121358993758275,
      "loss": 2.5213,
      "theoretical_loss": 4.03999081589871,
      "tokens_seen": 375980032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006121193493474561,
      "loss": 2.7807,
      "theoretical_loss": 4.039911217688777,
      "tokens_seen": 376045568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006121027993190846,
      "loss": 2.6774,
      "theoretical_loss": 4.039831637233176,
      "tokens_seen": 376111104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006120862492907131,
      "loss": 3.0818,
      "theoretical_loss": 4.039752074524851,
      "tokens_seen": 376176640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006120696992623416,
      "loss": 2.8564,
      "theoretical_loss": 4.039672529556755,
      "tokens_seen": 376242176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006120531492339701,
      "loss": 2.9123,
      "theoretical_loss": 4.039593002321844,
      "tokens_seen": 376307712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006120531492339701,
      "loss": 3.1245,
      "theoretical_loss": 4.039513492813076,
      "tokens_seen": 376373248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006120365992055987,
      "loss": 2.8261,
      "theoretical_loss": 4.039434001023415,
      "tokens_seen": 376438784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006120200491772272,
      "loss": 2.7749,
      "theoretical_loss": 4.039354526945829,
      "tokens_seen": 376504320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006120034991488557,
      "loss": 2.7321,
      "theoretical_loss": 4.039275070573288,
      "tokens_seen": 376569856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006119869491204842,
      "loss": 2.9999,
      "theoretical_loss": 4.039195631898768,
      "tokens_seen": 376635392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006119703990921126,
      "loss": 3.027,
      "theoretical_loss": 4.03911621091525,
      "tokens_seen": 376700928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006119538490637412,
      "loss": 2.6856,
      "theoretical_loss": 4.039036807615716,
      "tokens_seen": 376766464
    },
    {
      "epoch": 0.13,
      "objective/train/avg_token_score": 0.027676671743392944,
      "objective/train/avg_weight": 0.9674200415611267,
      "objective/train/docs_used": 225129,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.870966911315918,
      "objective/train/std_weight": 0.12293899059295654,
      "objective/train/theoretical_loss": 4.038957421993153,
      "objective/train/tokens_used": 397292000,
      "theoretical_loss": 4.038957421993153,
      "tokens_seen": 376832000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006119372990353697,
      "loss": 2.8671,
      "theoretical_loss": 4.038957421993153,
      "tokens_seen": 376832000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006119207490069983,
      "loss": 2.7618,
      "theoretical_loss": 4.038878054040554,
      "tokens_seen": 376897536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006119041989786268,
      "loss": 2.7203,
      "theoretical_loss": 4.038798703750913,
      "tokens_seen": 376963072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006118876489502553,
      "loss": 3.0215,
      "theoretical_loss": 4.038719371117231,
      "tokens_seen": 377028608
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006118710989218839,
      "loss": 2.9999,
      "theoretical_loss": 4.038640056132509,
      "tokens_seen": 377094144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006118545488935123,
      "loss": 3.0243,
      "theoretical_loss": 4.038560758789758,
      "tokens_seen": 377159680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006118379988651408,
      "loss": 2.7674,
      "theoretical_loss": 4.0384814790819865,
      "tokens_seen": 377225216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006118214488367694,
      "loss": 3.236,
      "theoretical_loss": 4.038402217002211,
      "tokens_seen": 377290752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006118048988083979,
      "loss": 2.7531,
      "theoretical_loss": 4.038322972543451,
      "tokens_seen": 377356288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006117883487800265,
      "loss": 2.7817,
      "theoretical_loss": 4.03824374569873,
      "tokens_seen": 377421824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000611771798751655,
      "loss": 2.9659,
      "theoretical_loss": 4.038164536461075,
      "tokens_seen": 377487360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006117552487232835,
      "loss": 3.1223,
      "theoretical_loss": 4.038085344823516,
      "tokens_seen": 377552896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006117386986949121,
      "loss": 2.9081,
      "theoretical_loss": 4.038006170779091,
      "tokens_seen": 377618432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006117221486665405,
      "loss": 2.9698,
      "theoretical_loss": 4.037927014320838,
      "tokens_seen": 377683968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000611705598638169,
      "loss": 2.7442,
      "theoretical_loss": 4.037847875441798,
      "tokens_seen": 377749504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006116890486097976,
      "loss": 2.8357,
      "theoretical_loss": 4.037768754135023,
      "tokens_seen": 377815040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006116724985814261,
      "loss": 2.8293,
      "theoretical_loss": 4.037689650393559,
      "tokens_seen": 377880576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0006116559485530547,
      "loss": 2.7858,
      "theoretical_loss": 4.037610564210464,
      "tokens_seen": 377946112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006116393985246832,
      "loss": 2.9777,
      "theoretical_loss": 4.037531495578797,
      "tokens_seen": 378011648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006116228484963117,
      "loss": 2.6891,
      "theoretical_loss": 4.03745244449162,
      "tokens_seen": 378077184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006116062984679402,
      "loss": 3.009,
      "theoretical_loss": 4.037373410942001,
      "tokens_seen": 378142720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006115897484395687,
      "loss": 3.0119,
      "theoretical_loss": 4.037294394923009,
      "tokens_seen": 378208256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006115731984111972,
      "loss": 2.8353,
      "theoretical_loss": 4.037215396427721,
      "tokens_seen": 378273792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006115566483828258,
      "loss": 2.9615,
      "theoretical_loss": 4.037136415449214,
      "tokens_seen": 378339328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006115400983544543,
      "loss": 3.0614,
      "theoretical_loss": 4.037057451980571,
      "tokens_seen": 378404864
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006115235483260829,
      "loss": 2.9079,
      "theoretical_loss": 4.036978506014879,
      "tokens_seen": 378470400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006115069982977114,
      "loss": 2.9648,
      "theoretical_loss": 4.036899577545229,
      "tokens_seen": 378535936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006114904482693398,
      "loss": 2.6282,
      "theoretical_loss": 4.036820666564714,
      "tokens_seen": 378601472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006114738982409684,
      "loss": 2.9474,
      "theoretical_loss": 4.036741773066434,
      "tokens_seen": 378667008
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006114573482125969,
      "loss": 3.0916,
      "theoretical_loss": 4.03666289704349,
      "tokens_seen": 378732544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006114407981842254,
      "loss": 3.1245,
      "theoretical_loss": 4.036584038488988,
      "tokens_seen": 378798080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000611424248155854,
      "loss": 2.8777,
      "theoretical_loss": 4.03650519739604,
      "tokens_seen": 378863616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006114076981274825,
      "loss": 2.8863,
      "theoretical_loss": 4.036426373757759,
      "tokens_seen": 378929152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006113911480991111,
      "loss": 3.0048,
      "theoretical_loss": 4.036347567567262,
      "tokens_seen": 378994688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006113745980707396,
      "loss": 2.8856,
      "theoretical_loss": 4.0362687788176705,
      "tokens_seen": 379060224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000611358048042368,
      "loss": 3.0584,
      "theoretical_loss": 4.036190007502112,
      "tokens_seen": 379125760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006113414980139966,
      "loss": 2.8722,
      "theoretical_loss": 4.036111253613715,
      "tokens_seen": 379191296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006113249479856251,
      "loss": 2.8916,
      "theoretical_loss": 4.036032517145614,
      "tokens_seen": 379256832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006113083979572536,
      "loss": 2.8461,
      "theoretical_loss": 4.0359537980909455,
      "tokens_seen": 379322368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006112918479288822,
      "loss": 2.8538,
      "theoretical_loss": 4.03587509644285,
      "tokens_seen": 379387904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006112752979005107,
      "loss": 2.8666,
      "theoretical_loss": 4.0357964121944745,
      "tokens_seen": 379453440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006112587478721393,
      "loss": 2.8775,
      "theoretical_loss": 4.035717745338967,
      "tokens_seen": 379518976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006112421978437677,
      "loss": 3.1078,
      "theoretical_loss": 4.035639095869481,
      "tokens_seen": 379584512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006112256478153962,
      "loss": 3.0668,
      "theoretical_loss": 4.035560463779173,
      "tokens_seen": 379650048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006112090977870248,
      "loss": 2.9113,
      "theoretical_loss": 4.0354818490612026,
      "tokens_seen": 379715584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006111925477586533,
      "loss": 2.9079,
      "theoretical_loss": 4.035403251708736,
      "tokens_seen": 379781120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006111759977302818,
      "loss": 2.7886,
      "theoretical_loss": 4.035324671714942,
      "tokens_seen": 379846656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006111594477019104,
      "loss": 3.0289,
      "theoretical_loss": 4.0352461090729905,
      "tokens_seen": 379912192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006111428976735389,
      "loss": 2.9163,
      "theoretical_loss": 4.03516756377606,
      "tokens_seen": 379977728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006111263476451675,
      "loss": 2.8516,
      "theoretical_loss": 4.035089035817329,
      "tokens_seen": 380043264
    },
    {
      "epoch": 0.14,
      "objective/train/avg_token_score": 0.01097310334444046,
      "objective/train/avg_weight": 0.9871113896369934,
      "objective/train/docs_used": 226852,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0914549827575684,
      "objective/train/std_weight": 0.07874582707881927,
      "objective/train/theoretical_loss": 4.035010525189982,
      "objective/train/tokens_used": 400568800,
      "theoretical_loss": 4.035010525189982,
      "tokens_seen": 380108800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006111097976167959,
      "loss": 3.033,
      "theoretical_loss": 4.035010525189982,
      "tokens_seen": 380108800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006110932475884244,
      "loss": 3.1307,
      "theoretical_loss": 4.034932031887207,
      "tokens_seen": 380174336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000611076697560053,
      "loss": 2.673,
      "theoretical_loss": 4.0348535559021945,
      "tokens_seen": 380239872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006110601475316815,
      "loss": 3.0195,
      "theoretical_loss": 4.034775097228142,
      "tokens_seen": 380305408
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00061104359750331,
      "loss": 2.8171,
      "theoretical_loss": 4.034696655858246,
      "tokens_seen": 380370944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006110270474749386,
      "loss": 3.1038,
      "theoretical_loss": 4.0346182317857116,
      "tokens_seen": 380436480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006110104974465671,
      "loss": 2.7892,
      "theoretical_loss": 4.034539825003746,
      "tokens_seen": 380502016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006109939474181955,
      "loss": 3.0341,
      "theoretical_loss": 4.034461435505557,
      "tokens_seen": 380567552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000610977397389824,
      "loss": 2.8783,
      "theoretical_loss": 4.034383063284364,
      "tokens_seen": 380633088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006109608473614526,
      "loss": 2.9607,
      "theoretical_loss": 4.034304708333382,
      "tokens_seen": 380698624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006109442973330811,
      "loss": 3.0192,
      "theoretical_loss": 4.0342263706458334,
      "tokens_seen": 380764160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006109277473047096,
      "loss": 2.9712,
      "theoretical_loss": 4.034148050214946,
      "tokens_seen": 380829696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006109111972763382,
      "loss": 3.1724,
      "theoretical_loss": 4.034069747033948,
      "tokens_seen": 380895232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006108946472479667,
      "loss": 2.7757,
      "theoretical_loss": 4.0339914610960745,
      "tokens_seen": 380960768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006108780972195953,
      "loss": 2.7183,
      "theoretical_loss": 4.033913192394563,
      "tokens_seen": 381026304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006108615471912237,
      "loss": 2.841,
      "theoretical_loss": 4.033834940922653,
      "tokens_seen": 381091840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006108449971628522,
      "loss": 2.8629,
      "theoretical_loss": 4.033756706673593,
      "tokens_seen": 381157376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006108284471344808,
      "loss": 2.8172,
      "theoretical_loss": 4.033678489640629,
      "tokens_seen": 381222912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006108118971061093,
      "loss": 2.6956,
      "theoretical_loss": 4.033600289817015,
      "tokens_seen": 381288448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006107953470777378,
      "loss": 2.8222,
      "theoretical_loss": 4.033522107196007,
      "tokens_seen": 381353984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006107787970493664,
      "loss": 2.7612,
      "theoretical_loss": 4.033443941770867,
      "tokens_seen": 381419520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006107622470209949,
      "loss": 3.1515,
      "theoretical_loss": 4.033365793534857,
      "tokens_seen": 381485056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006107456969926234,
      "loss": 2.9953,
      "theoretical_loss": 4.033287662481247,
      "tokens_seen": 381550592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006107291469642519,
      "loss": 2.999,
      "theoretical_loss": 4.033209548603308,
      "tokens_seen": 381616128
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006107125969358804,
      "loss": 2.7149,
      "theoretical_loss": 4.033131451894317,
      "tokens_seen": 381681664
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000610696046907509,
      "loss": 3.1153,
      "theoretical_loss": 4.03305337234755,
      "tokens_seen": 381747200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006106794968791375,
      "loss": 2.8385,
      "theoretical_loss": 4.032975309956292,
      "tokens_seen": 381812736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000610662946850766,
      "loss": 3.0799,
      "theoretical_loss": 4.0328972647138315,
      "tokens_seen": 381878272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006106463968223946,
      "loss": 2.9007,
      "theoretical_loss": 4.032819236613457,
      "tokens_seen": 381943808
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000610629846794023,
      "loss": 3.0488,
      "theoretical_loss": 4.032741225648463,
      "tokens_seen": 382009344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006106132967656516,
      "loss": 2.9517,
      "theoretical_loss": 4.03266323181215,
      "tokens_seen": 382074880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006105967467372801,
      "loss": 3.0398,
      "theoretical_loss": 4.0325852550978185,
      "tokens_seen": 382140416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006105801967089086,
      "loss": 2.9329,
      "theoretical_loss": 4.032507295498774,
      "tokens_seen": 382205952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006105636466805372,
      "loss": 2.8467,
      "theoretical_loss": 4.032429353008326,
      "tokens_seen": 382271488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006105470966521657,
      "loss": 3.0799,
      "theoretical_loss": 4.032351427619789,
      "tokens_seen": 382337024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006105305466237942,
      "loss": 2.82,
      "theoretical_loss": 4.03227351932648,
      "tokens_seen": 382402560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006105139965954228,
      "loss": 2.9072,
      "theoretical_loss": 4.032195628121718,
      "tokens_seen": 382468096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006104974465670512,
      "loss": 3.0571,
      "theoretical_loss": 4.032117753998829,
      "tokens_seen": 382533632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006104808965386798,
      "loss": 2.9345,
      "theoretical_loss": 4.032039896951142,
      "tokens_seen": 382599168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006104643465103083,
      "loss": 2.6122,
      "theoretical_loss": 4.031962056971988,
      "tokens_seen": 382664704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006104477964819368,
      "loss": 2.9767,
      "theoretical_loss": 4.0318842340547025,
      "tokens_seen": 382730240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006104312464535654,
      "loss": 2.8994,
      "theoretical_loss": 4.031806428192627,
      "tokens_seen": 382795776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006104146964251939,
      "loss": 2.8802,
      "theoretical_loss": 4.031728639379103,
      "tokens_seen": 382861312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006103981463968224,
      "loss": 2.9892,
      "theoretical_loss": 4.031650867607478,
      "tokens_seen": 382926848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006103815963684509,
      "loss": 2.895,
      "theoretical_loss": 4.031573112871103,
      "tokens_seen": 382992384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006103650463400794,
      "loss": 2.9208,
      "theoretical_loss": 4.031495375163333,
      "tokens_seen": 383057920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000610348496311708,
      "loss": 2.9415,
      "theoretical_loss": 4.031417654477526,
      "tokens_seen": 383123456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006103319462833365,
      "loss": 2.9962,
      "theoretical_loss": 4.031339950807043,
      "tokens_seen": 383188992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000610315396254965,
      "loss": 3.0598,
      "theoretical_loss": 4.03126226414525,
      "tokens_seen": 383254528
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006102988462265936,
      "loss": 3.0418,
      "theoretical_loss": 4.031184594485517,
      "tokens_seen": 383320064
    },
    {
      "epoch": 0.14,
      "objective/train/avg_token_score": 0.020832430571317673,
      "objective/train/avg_weight": 0.9754883050918579,
      "objective/train/docs_used": 228664,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.838676929473877,
      "objective/train/std_weight": 0.08737167716026306,
      "objective/train/theoretical_loss": 4.031106941821218,
      "objective/train/tokens_used": 403845600,
      "theoretical_loss": 4.031106941821218,
      "tokens_seen": 383385600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006102822961982221,
      "loss": 2.8508,
      "theoretical_loss": 4.031106941821218,
      "tokens_seen": 383385600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006102657461698505,
      "loss": 3.2098,
      "theoretical_loss": 4.031029306145728,
      "tokens_seen": 383451136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006102491961414791,
      "loss": 2.8855,
      "theoretical_loss": 4.030951687452429,
      "tokens_seen": 383516672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006102326461131076,
      "loss": 2.9341,
      "theoretical_loss": 4.0308740857347045,
      "tokens_seen": 383582208
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006102160960847362,
      "loss": 2.8868,
      "theoretical_loss": 4.030796500985943,
      "tokens_seen": 383647744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006101995460563647,
      "loss": 2.9738,
      "theoretical_loss": 4.030718933199536,
      "tokens_seen": 383713280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006101829960279932,
      "loss": 2.7718,
      "theoretical_loss": 4.030641382368878,
      "tokens_seen": 383778816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006101664459996218,
      "loss": 2.8346,
      "theoretical_loss": 4.03056384848737,
      "tokens_seen": 383844352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006101498959712503,
      "loss": 2.7969,
      "theoretical_loss": 4.030486331548413,
      "tokens_seen": 383909888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006101333459428787,
      "loss": 2.8257,
      "theoretical_loss": 4.030408831545414,
      "tokens_seen": 383975424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006101167959145073,
      "loss": 2.8073,
      "theoretical_loss": 4.030331348471784,
      "tokens_seen": 384040960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006101002458861358,
      "loss": 2.8207,
      "theoretical_loss": 4.030253882320936,
      "tokens_seen": 384106496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006100836958577644,
      "loss": 2.6528,
      "theoretical_loss": 4.030176433086289,
      "tokens_seen": 384172032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006100671458293929,
      "loss": 2.6167,
      "theoretical_loss": 4.030099000761261,
      "tokens_seen": 384237568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006100505958010214,
      "loss": 2.7699,
      "theoretical_loss": 4.030021585339281,
      "tokens_seen": 384303104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006100340457726499,
      "loss": 2.9442,
      "theoretical_loss": 4.029944186813774,
      "tokens_seen": 384368640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006100174957442783,
      "loss": 2.6335,
      "theoretical_loss": 4.029866805178175,
      "tokens_seen": 384434176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006100009457159069,
      "loss": 3.0509,
      "theoretical_loss": 4.029789440425917,
      "tokens_seen": 384499712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006099843956875354,
      "loss": 2.9875,
      "theoretical_loss": 4.029712092550443,
      "tokens_seen": 384565248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000609967845659164,
      "loss": 2.683,
      "theoretical_loss": 4.029634761545195,
      "tokens_seen": 384630784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006099512956307925,
      "loss": 3.0679,
      "theoretical_loss": 4.029557447403618,
      "tokens_seen": 384696320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000609934745602421,
      "loss": 2.9204,
      "theoretical_loss": 4.029480150119165,
      "tokens_seen": 384761856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006099181955740496,
      "loss": 2.8029,
      "theoretical_loss": 4.029402869685288,
      "tokens_seen": 384827392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000609901645545678,
      "loss": 2.8512,
      "theoretical_loss": 4.0293256060954485,
      "tokens_seen": 384892928
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006098850955173065,
      "loss": 2.9973,
      "theoretical_loss": 4.029248359343105,
      "tokens_seen": 384958464
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006098685454889351,
      "loss": 2.9027,
      "theoretical_loss": 4.029171129421724,
      "tokens_seen": 385024000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006098519954605636,
      "loss": 2.8383,
      "theoretical_loss": 4.029093916324774,
      "tokens_seen": 385089536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006098354454321922,
      "loss": 2.9094,
      "theoretical_loss": 4.0290167200457265,
      "tokens_seen": 385155072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006098188954038207,
      "loss": 3.0602,
      "theoretical_loss": 4.02893954057806,
      "tokens_seen": 385220608
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006098023453754492,
      "loss": 2.5698,
      "theoretical_loss": 4.028862377915252,
      "tokens_seen": 385286144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006097857953470778,
      "loss": 3.0832,
      "theoretical_loss": 4.0287852320507875,
      "tokens_seen": 385351680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006097692453187062,
      "loss": 2.8836,
      "theoretical_loss": 4.028708102978152,
      "tokens_seen": 385417216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006097526952903347,
      "loss": 2.7284,
      "theoretical_loss": 4.028630990690838,
      "tokens_seen": 385482752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006097361452619633,
      "loss": 3.0602,
      "theoretical_loss": 4.028553895182339,
      "tokens_seen": 385548288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006097195952335918,
      "loss": 2.9172,
      "theoretical_loss": 4.028476816446153,
      "tokens_seen": 385613824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006097030452052204,
      "loss": 2.8724,
      "theoretical_loss": 4.028399754475781,
      "tokens_seen": 385679360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006096864951768489,
      "loss": 2.903,
      "theoretical_loss": 4.028322709264729,
      "tokens_seen": 385744896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006096699451484774,
      "loss": 2.8965,
      "theoretical_loss": 4.0282456808065055,
      "tokens_seen": 385810432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006096533951201059,
      "loss": 3.0837,
      "theoretical_loss": 4.028168669094623,
      "tokens_seen": 385875968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006096368450917344,
      "loss": 2.9836,
      "theoretical_loss": 4.028091674122598,
      "tokens_seen": 385941504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006096202950633629,
      "loss": 2.7181,
      "theoretical_loss": 4.0280146958839484,
      "tokens_seen": 386007040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006096037450349915,
      "loss": 2.9563,
      "theoretical_loss": 4.0279377343722,
      "tokens_seen": 386072576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00060958719500662,
      "loss": 3.0068,
      "theoretical_loss": 4.027860789580878,
      "tokens_seen": 386138112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006095706449782486,
      "loss": 2.8303,
      "theoretical_loss": 4.027783861503513,
      "tokens_seen": 386203648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006095540949498771,
      "loss": 2.8307,
      "theoretical_loss": 4.0277069501336396,
      "tokens_seen": 386269184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006095375449215055,
      "loss": 2.9814,
      "theoretical_loss": 4.027630055464795,
      "tokens_seen": 386334720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006095209948931341,
      "loss": 2.8997,
      "theoretical_loss": 4.027553177490521,
      "tokens_seen": 386400256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006095044448647626,
      "loss": 2.7692,
      "theoretical_loss": 4.027476316204362,
      "tokens_seen": 386465792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006094878948363911,
      "loss": 2.642,
      "theoretical_loss": 4.027399471599867,
      "tokens_seen": 386531328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006094713448080197,
      "loss": 2.8851,
      "theoretical_loss": 4.027322643670587,
      "tokens_seen": 386596864
    },
    {
      "epoch": 0.14,
      "objective/train/avg_token_score": 0.0145777128636837,
      "objective/train/avg_weight": 0.9828572869300842,
      "objective/train/docs_used": 230356,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6820449829101562,
      "objective/train/std_weight": 0.10390425473451614,
      "objective/train/theoretical_loss": 4.027245832410079,
      "objective/train/tokens_used": 407122400,
      "theoretical_loss": 4.027245832410079,
      "tokens_seen": 386662400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006094547947796482,
      "loss": 2.596,
      "theoretical_loss": 4.027245832410079,
      "tokens_seen": 386662400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006094382447512768,
      "loss": 2.7722,
      "theoretical_loss": 4.027169037811901,
      "tokens_seen": 386727936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006094216947229053,
      "loss": 2.8779,
      "theoretical_loss": 4.027092259869616,
      "tokens_seen": 386793472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006094051446945337,
      "loss": 2.9127,
      "theoretical_loss": 4.027015498576791,
      "tokens_seen": 386859008
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006093885946661623,
      "loss": 2.9343,
      "theoretical_loss": 4.026938753926997,
      "tokens_seen": 386924544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006093720446377908,
      "loss": 2.8971,
      "theoretical_loss": 4.026862025913804,
      "tokens_seen": 386990080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006093554946094193,
      "loss": 2.7715,
      "theoretical_loss": 4.026785314530792,
      "tokens_seen": 387055616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006093389445810479,
      "loss": 3.0395,
      "theoretical_loss": 4.026708619771541,
      "tokens_seen": 387121152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006093223945526764,
      "loss": 2.8059,
      "theoretical_loss": 4.026631941629635,
      "tokens_seen": 387186688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000609305844524305,
      "loss": 2.8484,
      "theoretical_loss": 4.026555280098661,
      "tokens_seen": 387252224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006092892944959334,
      "loss": 2.8025,
      "theoretical_loss": 4.026478635172212,
      "tokens_seen": 387317760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006092727444675619,
      "loss": 2.8223,
      "theoretical_loss": 4.026402006843882,
      "tokens_seen": 387383296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006092561944391905,
      "loss": 2.7547,
      "theoretical_loss": 4.0263253951072695,
      "tokens_seen": 387448832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000609239644410819,
      "loss": 2.5897,
      "theoretical_loss": 4.026248799955976,
      "tokens_seen": 387514368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006092230943824475,
      "loss": 2.7241,
      "theoretical_loss": 4.026172221383607,
      "tokens_seen": 387579904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006092065443540761,
      "loss": 2.6182,
      "theoretical_loss": 4.026095659383773,
      "tokens_seen": 387645440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006091899943257046,
      "loss": 3.1049,
      "theoretical_loss": 4.026019113950086,
      "tokens_seen": 387710976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006091734442973332,
      "loss": 2.7468,
      "theoretical_loss": 4.025942585076161,
      "tokens_seen": 387776512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006091568942689616,
      "loss": 2.8489,
      "theoretical_loss": 4.025866072755619,
      "tokens_seen": 387842048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006091403442405901,
      "loss": 2.7242,
      "theoretical_loss": 4.025789576982083,
      "tokens_seen": 387907584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006091237942122187,
      "loss": 2.8588,
      "theoretical_loss": 4.02571309774918,
      "tokens_seen": 387973120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006091072441838472,
      "loss": 2.7276,
      "theoretical_loss": 4.025636635050539,
      "tokens_seen": 388038656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006090906941554757,
      "loss": 2.8737,
      "theoretical_loss": 4.025560188879796,
      "tokens_seen": 388104192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006090741441271043,
      "loss": 2.9522,
      "theoretical_loss": 4.025483759230585,
      "tokens_seen": 388169728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006090575940987326,
      "loss": 2.8891,
      "theoretical_loss": 4.025407346096551,
      "tokens_seen": 388235264
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006090410440703612,
      "loss": 2.8699,
      "theoretical_loss": 4.025330949471336,
      "tokens_seen": 388300800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006090244940419897,
      "loss": 2.8242,
      "theoretical_loss": 4.025254569348588,
      "tokens_seen": 388366336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006090079440136183,
      "loss": 3.1283,
      "theoretical_loss": 4.025178205721959,
      "tokens_seen": 388431872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006089913939852468,
      "loss": 2.8572,
      "theoretical_loss": 4.025101858585104,
      "tokens_seen": 388497408
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006089748439568753,
      "loss": 2.6129,
      "theoretical_loss": 4.025025527931681,
      "tokens_seen": 388562944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006089582939285039,
      "loss": 2.8986,
      "theoretical_loss": 4.0249492137553515,
      "tokens_seen": 388628480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006089417439001324,
      "loss": 2.6005,
      "theoretical_loss": 4.024872916049782,
      "tokens_seen": 388694016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006089251938717608,
      "loss": 2.8404,
      "theoretical_loss": 4.0247966348086415,
      "tokens_seen": 388759552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006089086438433894,
      "loss": 2.8698,
      "theoretical_loss": 4.0247203700256025,
      "tokens_seen": 388825088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006088920938150179,
      "loss": 2.7596,
      "theoretical_loss": 4.02464412169434,
      "tokens_seen": 388890624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006088755437866465,
      "loss": 2.8876,
      "theoretical_loss": 4.024567889808534,
      "tokens_seen": 388956160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000608858993758275,
      "loss": 2.8813,
      "theoretical_loss": 4.024491674361868,
      "tokens_seen": 389021696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006088424437299035,
      "loss": 2.8359,
      "theoretical_loss": 4.024415475348027,
      "tokens_seen": 389087232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006088258937015321,
      "loss": 2.8002,
      "theoretical_loss": 4.024339292760703,
      "tokens_seen": 389152768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006088093436731605,
      "loss": 3.1659,
      "theoretical_loss": 4.024263126593588,
      "tokens_seen": 389218304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000608792793644789,
      "loss": 2.8935,
      "theoretical_loss": 4.0241869768403795,
      "tokens_seen": 389283840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006087762436164176,
      "loss": 2.9625,
      "theoretical_loss": 4.024110843494777,
      "tokens_seen": 389349376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006087596935880461,
      "loss": 2.9675,
      "theoretical_loss": 4.024034726550486,
      "tokens_seen": 389414912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006087431435596747,
      "loss": 2.8467,
      "theoretical_loss": 4.023958626001212,
      "tokens_seen": 389480448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006087265935313032,
      "loss": 2.8154,
      "theoretical_loss": 4.023882541840667,
      "tokens_seen": 389545984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006087100435029317,
      "loss": 2.5479,
      "theoretical_loss": 4.023806474062566,
      "tokens_seen": 389611520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006086934934745603,
      "loss": 2.9682,
      "theoretical_loss": 4.023730422660624,
      "tokens_seen": 389677056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006086769434461887,
      "loss": 2.8169,
      "theoretical_loss": 4.023654387628565,
      "tokens_seen": 389742592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006086603934178172,
      "loss": 2.7533,
      "theoretical_loss": 4.023578368960113,
      "tokens_seen": 389808128
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006086438433894458,
      "loss": 2.8822,
      "theoretical_loss": 4.023502366648994,
      "tokens_seen": 389873664
    },
    {
      "epoch": 0.14,
      "objective/train/avg_token_score": 0.028794823214411736,
      "objective/train/avg_weight": 0.9661146998405457,
      "objective/train/docs_used": 231398,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8994593620300293,
      "objective/train/std_weight": 0.11174099147319794,
      "objective/train/theoretical_loss": 4.023426380688943,
      "objective/train/tokens_used": 410399200,
      "theoretical_loss": 4.023426380688943,
      "tokens_seen": 389939200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006086272933610743,
      "loss": 2.8711,
      "theoretical_loss": 4.023426380688943,
      "tokens_seen": 389939200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006086107433327029,
      "loss": 2.7694,
      "theoretical_loss": 4.023350411073693,
      "tokens_seen": 390004736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006085941933043314,
      "loss": 2.9701,
      "theoretical_loss": 4.023274457796983,
      "tokens_seen": 390070272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006085776432759599,
      "loss": 2.7053,
      "theoretical_loss": 4.023198520852555,
      "tokens_seen": 390135808
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006085610932475884,
      "loss": 2.9174,
      "theoretical_loss": 4.023122600234155,
      "tokens_seen": 390201344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006085445432192169,
      "loss": 2.8678,
      "theoretical_loss": 4.0230466959355295,
      "tokens_seen": 390266880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006085279931908454,
      "loss": 2.882,
      "theoretical_loss": 4.022970807950434,
      "tokens_seen": 390332416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000608511443162474,
      "loss": 2.9443,
      "theoretical_loss": 4.022894936272621,
      "tokens_seen": 390397952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006084948931341025,
      "loss": 2.8701,
      "theoretical_loss": 4.022819080895853,
      "tokens_seen": 390463488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006084783431057311,
      "loss": 3.0012,
      "theoretical_loss": 4.02274324181389,
      "tokens_seen": 390529024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006084617930773596,
      "loss": 2.8405,
      "theoretical_loss": 4.022667419020499,
      "tokens_seen": 390594560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000608445243048988,
      "loss": 3.0145,
      "theoretical_loss": 4.0225916125094505,
      "tokens_seen": 390660096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006084286930206166,
      "loss": 2.8628,
      "theoretical_loss": 4.022515822274515,
      "tokens_seen": 390725632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006084121429922451,
      "loss": 2.6189,
      "theoretical_loss": 4.0224400483094715,
      "tokens_seen": 390791168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006083955929638737,
      "loss": 2.659,
      "theoretical_loss": 4.022364290608098,
      "tokens_seen": 390856704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006083790429355022,
      "loss": 3.2258,
      "theoretical_loss": 4.022288549164178,
      "tokens_seen": 390922240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006083624929071307,
      "loss": 2.7317,
      "theoretical_loss": 4.022212823971498,
      "tokens_seen": 390987776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006083459428787593,
      "loss": 2.7659,
      "theoretical_loss": 4.022137115023848,
      "tokens_seen": 391053312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006083293928503878,
      "loss": 2.8942,
      "theoretical_loss": 4.022061422315023,
      "tokens_seen": 391118848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006083128428220162,
      "loss": 2.8767,
      "theoretical_loss": 4.021985745838817,
      "tokens_seen": 391184384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006082962927936448,
      "loss": 2.8233,
      "theoretical_loss": 4.021910085589032,
      "tokens_seen": 391249920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006082797427652733,
      "loss": 2.8344,
      "theoretical_loss": 4.021834441559472,
      "tokens_seen": 391315456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006082631927369019,
      "loss": 3.0792,
      "theoretical_loss": 4.021758813743942,
      "tokens_seen": 391380992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006082466427085304,
      "loss": 2.6322,
      "theoretical_loss": 4.021683202136255,
      "tokens_seen": 391446528
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006082300926801589,
      "loss": 3.1472,
      "theoretical_loss": 4.021607606730223,
      "tokens_seen": 391512064
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006082135426517875,
      "loss": 2.5931,
      "theoretical_loss": 4.0215320275196635,
      "tokens_seen": 391577600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000608196992623416,
      "loss": 2.6266,
      "theoretical_loss": 4.0214564644983986,
      "tokens_seen": 391643136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006081804425950444,
      "loss": 3.0004,
      "theoretical_loss": 4.021380917660251,
      "tokens_seen": 391708672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000608163892566673,
      "loss": 2.8123,
      "theoretical_loss": 4.021305386999048,
      "tokens_seen": 391774208
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006081473425383015,
      "loss": 2.8416,
      "theoretical_loss": 4.021229872508621,
      "tokens_seen": 391839744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006081307925099301,
      "loss": 2.9128,
      "theoretical_loss": 4.021154374182803,
      "tokens_seen": 391905280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006081142424815586,
      "loss": 2.8295,
      "theoretical_loss": 4.021078892015433,
      "tokens_seen": 391970816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006080976924531871,
      "loss": 2.8271,
      "theoretical_loss": 4.021003426000353,
      "tokens_seen": 392036352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006080811424248157,
      "loss": 3.1352,
      "theoretical_loss": 4.020927976131404,
      "tokens_seen": 392101888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000608064592396444,
      "loss": 2.9951,
      "theoretical_loss": 4.0208525424024355,
      "tokens_seen": 392167424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006080480423680726,
      "loss": 2.7905,
      "theoretical_loss": 4.020777124807299,
      "tokens_seen": 392232960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006080314923397011,
      "loss": 2.795,
      "theoretical_loss": 4.020701723339848,
      "tokens_seen": 392298496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006080149423113296,
      "loss": 3.0855,
      "theoretical_loss": 4.020626337993941,
      "tokens_seen": 392364032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006079983922829582,
      "loss": 2.9453,
      "theoretical_loss": 4.020550968763439,
      "tokens_seen": 392429568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006079818422545867,
      "loss": 2.8741,
      "theoretical_loss": 4.020475615642207,
      "tokens_seen": 392495104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006079652922262153,
      "loss": 2.891,
      "theoretical_loss": 4.020400278624112,
      "tokens_seen": 392560640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006079487421978437,
      "loss": 2.9253,
      "theoretical_loss": 4.020324957703025,
      "tokens_seen": 392626176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006079321921694722,
      "loss": 2.8031,
      "theoretical_loss": 4.020249652872822,
      "tokens_seen": 392691712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006079156421411008,
      "loss": 2.8167,
      "theoretical_loss": 4.02017436412738,
      "tokens_seen": 392757248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006078990921127293,
      "loss": 3.0089,
      "theoretical_loss": 4.02009909146058,
      "tokens_seen": 392822784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006078825420843578,
      "loss": 2.9614,
      "theoretical_loss": 4.020023834866308,
      "tokens_seen": 392888320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006078659920559864,
      "loss": 2.7057,
      "theoretical_loss": 4.01994859433845,
      "tokens_seen": 392953856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006078494420276149,
      "loss": 2.9831,
      "theoretical_loss": 4.019873369870899,
      "tokens_seen": 393019392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006078328919992434,
      "loss": 2.8707,
      "theoretical_loss": 4.019798161457548,
      "tokens_seen": 393084928
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006078163419708719,
      "loss": 3.0247,
      "theoretical_loss": 4.019722969092298,
      "tokens_seen": 393150464
    },
    {
      "epoch": 0.14,
      "objective/train/avg_token_score": 0.02045375108718872,
      "objective/train/avg_weight": 0.9759305715560913,
      "objective/train/docs_used": 233309,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.4347057342529297,
      "objective/train/std_weight": 0.11442361772060394,
      "objective/train/theoretical_loss": 4.019647792769048,
      "objective/train/tokens_used": 413676000,
      "theoretical_loss": 4.019647792769048,
      "tokens_seen": 393216000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006077997919425004,
      "loss": 3.0097,
      "theoretical_loss": 4.019647792769048,
      "tokens_seen": 393216000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000607783241914129,
      "loss": 2.9276,
      "theoretical_loss": 4.019572632481702,
      "tokens_seen": 393281536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006077666918857575,
      "loss": 2.9666,
      "theoretical_loss": 4.01949748822417,
      "tokens_seen": 393347072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000607750141857386,
      "loss": 3.1389,
      "theoretical_loss": 4.0194223599903625,
      "tokens_seen": 393412608
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006077335918290146,
      "loss": 2.9292,
      "theoretical_loss": 4.019347247774195,
      "tokens_seen": 393478144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000607717041800643,
      "loss": 2.7207,
      "theoretical_loss": 4.0192721515695835,
      "tokens_seen": 393543680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006077004917722716,
      "loss": 3.1004,
      "theoretical_loss": 4.019197071370451,
      "tokens_seen": 393609216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006076839417439001,
      "loss": 2.7909,
      "theoretical_loss": 4.019122007170722,
      "tokens_seen": 393674752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006076673917155286,
      "loss": 3.0767,
      "theoretical_loss": 4.019046958964324,
      "tokens_seen": 393740288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006076508416871572,
      "loss": 2.9481,
      "theoretical_loss": 4.018971926745189,
      "tokens_seen": 393805824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006076342916587857,
      "loss": 2.8303,
      "theoretical_loss": 4.01889691050725,
      "tokens_seen": 393871360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006076177416304142,
      "loss": 3.0176,
      "theoretical_loss": 4.018821910244446,
      "tokens_seen": 393936896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006076011916020428,
      "loss": 2.8679,
      "theoretical_loss": 4.018746925950719,
      "tokens_seen": 394002432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006075846415736712,
      "loss": 2.8328,
      "theoretical_loss": 4.018671957620012,
      "tokens_seen": 394067968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006075680915452998,
      "loss": 3.256,
      "theoretical_loss": 4.018597005246273,
      "tokens_seen": 394133504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006075515415169283,
      "loss": 2.8728,
      "theoretical_loss": 4.018522068823454,
      "tokens_seen": 394199040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006075349914885568,
      "loss": 2.8718,
      "theoretical_loss": 4.01844714834551,
      "tokens_seen": 394264576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006075184414601854,
      "loss": 2.8945,
      "theoretical_loss": 4.018372243806397,
      "tokens_seen": 394330112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006075018914318139,
      "loss": 2.954,
      "theoretical_loss": 4.018297355200076,
      "tokens_seen": 394395648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006074853414034424,
      "loss": 2.7901,
      "theoretical_loss": 4.018222482520512,
      "tokens_seen": 394461184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000607468791375071,
      "loss": 3.0602,
      "theoretical_loss": 4.018147625761673,
      "tokens_seen": 394526720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006074522413466994,
      "loss": 2.9088,
      "theoretical_loss": 4.018072784917529,
      "tokens_seen": 394592256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000607435691318328,
      "loss": 2.7592,
      "theoretical_loss": 4.017997959982055,
      "tokens_seen": 394657792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006074191412899565,
      "loss": 2.9235,
      "theoretical_loss": 4.017923150949228,
      "tokens_seen": 394723328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000607402591261585,
      "loss": 3.0469,
      "theoretical_loss": 4.017848357813028,
      "tokens_seen": 394788864
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006073860412332136,
      "loss": 2.8564,
      "theoretical_loss": 4.01777358056744,
      "tokens_seen": 394854400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006073694912048421,
      "loss": 3.0139,
      "theoretical_loss": 4.017698819206451,
      "tokens_seen": 394919936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006073529411764705,
      "loss": 2.9637,
      "theoretical_loss": 4.017624073724052,
      "tokens_seen": 394985472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006073363911480991,
      "loss": 3.0169,
      "theoretical_loss": 4.017549344114236,
      "tokens_seen": 395051008
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006073198411197276,
      "loss": 3.022,
      "theoretical_loss": 4.017474630371,
      "tokens_seen": 395116544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006073032910913562,
      "loss": 2.8219,
      "theoretical_loss": 4.0173999324883445,
      "tokens_seen": 395182080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006072867410629847,
      "loss": 2.8244,
      "theoretical_loss": 4.0173252504602734,
      "tokens_seen": 395247616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006072701910346132,
      "loss": 2.8445,
      "theoretical_loss": 4.017250584280794,
      "tokens_seen": 395313152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006072536410062418,
      "loss": 3.265,
      "theoretical_loss": 4.017175933943916,
      "tokens_seen": 395378688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006072370909778703,
      "loss": 2.9567,
      "theoretical_loss": 4.0171012994436515,
      "tokens_seen": 395444224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006072205409494987,
      "loss": 2.9326,
      "theoretical_loss": 4.017026680774019,
      "tokens_seen": 395509760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006072039909211273,
      "loss": 2.8822,
      "theoretical_loss": 4.016952077929037,
      "tokens_seen": 395575296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006071874408927558,
      "loss": 2.997,
      "theoretical_loss": 4.016877490902731,
      "tokens_seen": 395640832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006071708908643844,
      "loss": 3.0658,
      "theoretical_loss": 4.016802919689124,
      "tokens_seen": 395706368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006071543408360129,
      "loss": 2.9392,
      "theoretical_loss": 4.0167283642822476,
      "tokens_seen": 395771904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006071377908076414,
      "loss": 2.756,
      "theoretical_loss": 4.016653824676135,
      "tokens_seen": 395837440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00060712124077927,
      "loss": 2.9441,
      "theoretical_loss": 4.016579300864821,
      "tokens_seen": 395902976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006071046907508984,
      "loss": 2.6893,
      "theoretical_loss": 4.016504792842346,
      "tokens_seen": 395968512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006070881407225269,
      "loss": 2.9766,
      "theoretical_loss": 4.016430300602751,
      "tokens_seen": 396034048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006070715906941554,
      "loss": 2.8986,
      "theoretical_loss": 4.016355824140085,
      "tokens_seen": 396099584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000607055040665784,
      "loss": 2.9585,
      "theoretical_loss": 4.016281363448394,
      "tokens_seen": 396165120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006070384906374125,
      "loss": 3.1278,
      "theoretical_loss": 4.016206918521731,
      "tokens_seen": 396230656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000607021940609041,
      "loss": 2.9051,
      "theoretical_loss": 4.016132489354153,
      "tokens_seen": 396296192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006070053905806696,
      "loss": 2.8831,
      "theoretical_loss": 4.016058075939716,
      "tokens_seen": 396361728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000606988840552298,
      "loss": 2.8956,
      "theoretical_loss": 4.015983678272484,
      "tokens_seen": 396427264
    },
    {
      "epoch": 0.14,
      "objective/train/avg_token_score": 0.011101280339062214,
      "objective/train/avg_weight": 0.9869309067726135,
      "objective/train/docs_used": 234734,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8041229248046875,
      "objective/train/std_weight": 0.0521705336868763,
      "objective/train/theoretical_loss": 4.015909296346521,
      "objective/train/tokens_used": 416952800,
      "theoretical_loss": 4.015909296346521,
      "tokens_seen": 396492800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006069722905239265,
      "loss": 2.9332,
      "theoretical_loss": 4.015909296346521,
      "tokens_seen": 396492800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006069557404955551,
      "loss": 2.7598,
      "theoretical_loss": 4.015834930155897,
      "tokens_seen": 396558336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006069391904671836,
      "loss": 2.9583,
      "theoretical_loss": 4.015760579694683,
      "tokens_seen": 396623872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006069226404388122,
      "loss": 2.9241,
      "theoretical_loss": 4.015686244956953,
      "tokens_seen": 396689408
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006069060904104407,
      "loss": 3.0451,
      "theoretical_loss": 4.015611925936785,
      "tokens_seen": 396754944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006068895403820692,
      "loss": 2.8518,
      "theoretical_loss": 4.01553762262826,
      "tokens_seen": 396820480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006068729903536978,
      "loss": 3.0111,
      "theoretical_loss": 4.015463335025465,
      "tokens_seen": 396886016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006068564403253262,
      "loss": 2.8094,
      "theoretical_loss": 4.015389063122484,
      "tokens_seen": 396951552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006068398902969547,
      "loss": 3.0523,
      "theoretical_loss": 4.015314806913411,
      "tokens_seen": 397017088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006068233402685833,
      "loss": 2.9283,
      "theoretical_loss": 4.015240566392338,
      "tokens_seen": 397082624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006068067902402118,
      "loss": 2.8241,
      "theoretical_loss": 4.015166341553362,
      "tokens_seen": 397148160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006067902402118404,
      "loss": 2.875,
      "theoretical_loss": 4.015092132390585,
      "tokens_seen": 397213696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006067736901834689,
      "loss": 3.077,
      "theoretical_loss": 4.01501793889811,
      "tokens_seen": 397279232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006067571401550974,
      "loss": 2.6291,
      "theoretical_loss": 4.014943761070043,
      "tokens_seen": 397344768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000606740590126726,
      "loss": 2.8493,
      "theoretical_loss": 4.014869598900496,
      "tokens_seen": 397410304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006067240400983544,
      "loss": 2.7417,
      "theoretical_loss": 4.01479545238358,
      "tokens_seen": 397475840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006067074900699829,
      "loss": 2.691,
      "theoretical_loss": 4.014721321513412,
      "tokens_seen": 397541376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006066909400416115,
      "loss": 2.9152,
      "theoretical_loss": 4.0146472062841125,
      "tokens_seen": 397606912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00060667439001324,
      "loss": 2.7423,
      "theoretical_loss": 4.014573106689803,
      "tokens_seen": 397672448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006066578399848686,
      "loss": 2.729,
      "theoretical_loss": 4.0144990227246105,
      "tokens_seen": 397737984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006066412899564971,
      "loss": 2.9441,
      "theoretical_loss": 4.014424954382663,
      "tokens_seen": 397803520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006066247399281256,
      "loss": 2.7935,
      "theoretical_loss": 4.014350901658093,
      "tokens_seen": 397869056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006066081898997541,
      "loss": 2.9494,
      "theoretical_loss": 4.014276864545036,
      "tokens_seen": 397934592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006065916398713826,
      "loss": 2.5468,
      "theoretical_loss": 4.014202843037632,
      "tokens_seen": 398000128
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006065750898430111,
      "loss": 2.553,
      "theoretical_loss": 4.014128837130021,
      "tokens_seen": 398065664
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006065585398146397,
      "loss": 2.6829,
      "theoretical_loss": 4.0140548468163475,
      "tokens_seen": 398131200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006065419897862682,
      "loss": 2.845,
      "theoretical_loss": 4.0139808720907615,
      "tokens_seen": 398196736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006065254397578968,
      "loss": 2.5868,
      "theoretical_loss": 4.013906912947412,
      "tokens_seen": 398262272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006065088897295253,
      "loss": 2.9817,
      "theoretical_loss": 4.013832969380456,
      "tokens_seen": 398327808
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006064923397011537,
      "loss": 2.8758,
      "theoretical_loss": 4.013759041384049,
      "tokens_seen": 398393344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006064757896727823,
      "loss": 2.7579,
      "theoretical_loss": 4.013685128952353,
      "tokens_seen": 398458880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006064592396444108,
      "loss": 2.9897,
      "theoretical_loss": 4.013611232079531,
      "tokens_seen": 398524416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006064426896160393,
      "loss": 2.9643,
      "theoretical_loss": 4.01353735075975,
      "tokens_seen": 398589952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006064261395876679,
      "loss": 2.9759,
      "theoretical_loss": 4.01346348498718,
      "tokens_seen": 398655488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006064095895592964,
      "loss": 2.9255,
      "theoretical_loss": 4.013389634755995,
      "tokens_seen": 398721024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000606393039530925,
      "loss": 2.9803,
      "theoretical_loss": 4.013315800060371,
      "tokens_seen": 398786560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006063764895025535,
      "loss": 2.9653,
      "theoretical_loss": 4.013241980894488,
      "tokens_seen": 398852096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006063599394741819,
      "loss": 2.9774,
      "theoretical_loss": 4.013168177252528,
      "tokens_seen": 398917632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006063433894458105,
      "loss": 2.7289,
      "theoretical_loss": 4.013094389128677,
      "tokens_seen": 398983168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000606326839417439,
      "loss": 2.7901,
      "theoretical_loss": 4.0130206165171245,
      "tokens_seen": 399048704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006063102893890675,
      "loss": 2.9467,
      "theoretical_loss": 4.012946859412062,
      "tokens_seen": 399114240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006062937393606961,
      "loss": 2.9804,
      "theoretical_loss": 4.012873117807685,
      "tokens_seen": 399179776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006062771893323246,
      "loss": 2.8653,
      "theoretical_loss": 4.012799391698191,
      "tokens_seen": 399245312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006062606393039532,
      "loss": 2.781,
      "theoretical_loss": 4.012725681077783,
      "tokens_seen": 399310848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006062440892755816,
      "loss": 2.8014,
      "theoretical_loss": 4.012651985940666,
      "tokens_seen": 399376384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006062275392472101,
      "loss": 2.9609,
      "theoretical_loss": 4.012578306281045,
      "tokens_seen": 399441920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006062109892188387,
      "loss": 2.8147,
      "theoretical_loss": 4.012504642093133,
      "tokens_seen": 399507456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006061944391904672,
      "loss": 2.9392,
      "theoretical_loss": 4.012430993371144,
      "tokens_seen": 399572992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006061778891620957,
      "loss": 2.951,
      "theoretical_loss": 4.012357360109293,
      "tokens_seen": 399638528
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006061613391337243,
      "loss": 2.8902,
      "theoretical_loss": 4.012283742301802,
      "tokens_seen": 399704064
    },
    {
      "epoch": 0.14,
      "objective/train/avg_token_score": 0.023300614207983017,
      "objective/train/avg_weight": 0.9725866913795471,
      "objective/train/docs_used": 236668,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.953869342803955,
      "objective/train/std_weight": 0.07471281290054321,
      "objective/train/theoretical_loss": 4.012210139942894,
      "objective/train/tokens_used": 420229600,
      "theoretical_loss": 4.012210139942894,
      "tokens_seen": 399769600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006061447891053528,
      "loss": 3.0069,
      "theoretical_loss": 4.012210139942894,
      "tokens_seen": 399769600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006061282390769814,
      "loss": 2.7458,
      "theoretical_loss": 4.012136553026796,
      "tokens_seen": 399835136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006061116890486098,
      "loss": 2.7273,
      "theoretical_loss": 4.012062981547736,
      "tokens_seen": 399900672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006060951390202383,
      "loss": 2.7362,
      "theoretical_loss": 4.011989425499947,
      "tokens_seen": 399966208
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006060785889918668,
      "loss": 2.8784,
      "theoretical_loss": 4.0119158848776655,
      "tokens_seen": 400031744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006060620389634953,
      "loss": 2.9168,
      "theoretical_loss": 4.011842359675129,
      "tokens_seen": 400097280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006060454889351239,
      "loss": 2.7578,
      "theoretical_loss": 4.011768849886582,
      "tokens_seen": 400162816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006060289389067524,
      "loss": 2.8685,
      "theoretical_loss": 4.011695355506266,
      "tokens_seen": 400228352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000606012388878381,
      "loss": 2.7612,
      "theoretical_loss": 4.011621876528431,
      "tokens_seen": 400293888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006059958388500094,
      "loss": 3.1645,
      "theoretical_loss": 4.011548412947327,
      "tokens_seen": 400359424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006059792888216379,
      "loss": 3.1043,
      "theoretical_loss": 4.011474964757211,
      "tokens_seen": 400424960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006059627387932665,
      "loss": 3.2017,
      "theoretical_loss": 4.0114015319523375,
      "tokens_seen": 400490496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000605946188764895,
      "loss": 2.8571,
      "theoretical_loss": 4.011328114526968,
      "tokens_seen": 400556032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006059296387365235,
      "loss": 2.882,
      "theoretical_loss": 4.011254712475365,
      "tokens_seen": 400621568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006059130887081521,
      "loss": 2.9872,
      "theoretical_loss": 4.011181325791797,
      "tokens_seen": 400687104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006058965386797806,
      "loss": 2.8275,
      "theoretical_loss": 4.011107954470531,
      "tokens_seen": 400752640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006058799886514091,
      "loss": 3.0488,
      "theoretical_loss": 4.011034598505843,
      "tokens_seen": 400818176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006058634386230376,
      "loss": 2.8313,
      "theoretical_loss": 4.010961257892005,
      "tokens_seen": 400883712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006058468885946661,
      "loss": 2.9518,
      "theoretical_loss": 4.010887932623299,
      "tokens_seen": 400949248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006058303385662947,
      "loss": 3.1025,
      "theoretical_loss": 4.010814622694006,
      "tokens_seen": 401014784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006058137885379232,
      "loss": 2.7905,
      "theoretical_loss": 4.01074132809841,
      "tokens_seen": 401080320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006057972385095517,
      "loss": 2.8149,
      "theoretical_loss": 4.0106680488308,
      "tokens_seen": 401145856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006057806884811803,
      "loss": 2.8737,
      "theoretical_loss": 4.010594784885466,
      "tokens_seen": 401211392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006057641384528087,
      "loss": 3.0078,
      "theoretical_loss": 4.010521536256705,
      "tokens_seen": 401276928
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006057475884244373,
      "loss": 2.8469,
      "theoretical_loss": 4.01044830293881,
      "tokens_seen": 401342464
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006057310383960658,
      "loss": 2.9485,
      "theoretical_loss": 4.0103750849260855,
      "tokens_seen": 401408000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006057144883676943,
      "loss": 2.7356,
      "theoretical_loss": 4.010301882212831,
      "tokens_seen": 401473536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006056979383393229,
      "loss": 2.9722,
      "theoretical_loss": 4.010228694793356,
      "tokens_seen": 401539072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006056813883109514,
      "loss": 3.0585,
      "theoretical_loss": 4.010155522661968,
      "tokens_seen": 401604608
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006056648382825799,
      "loss": 2.7693,
      "theoretical_loss": 4.0100823658129805,
      "tokens_seen": 401670144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006056482882542085,
      "loss": 2.7891,
      "theoretical_loss": 4.010009224240708,
      "tokens_seen": 401735680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006056317382258369,
      "loss": 2.64,
      "theoretical_loss": 4.0099360979394705,
      "tokens_seen": 401801216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006056151881974655,
      "loss": 3.0611,
      "theoretical_loss": 4.009862986903588,
      "tokens_seen": 401866752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000605598638169094,
      "loss": 2.9479,
      "theoretical_loss": 4.009789891127387,
      "tokens_seen": 401932288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006055820881407225,
      "loss": 2.9181,
      "theoretical_loss": 4.009716810605193,
      "tokens_seen": 401997824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006055655381123511,
      "loss": 3.0689,
      "theoretical_loss": 4.009643745331338,
      "tokens_seen": 402063360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006055489880839796,
      "loss": 2.9284,
      "theoretical_loss": 4.009570695300155,
      "tokens_seen": 402128896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000605532438055608,
      "loss": 2.7433,
      "theoretical_loss": 4.009497660505982,
      "tokens_seen": 402194432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006055158880272366,
      "loss": 2.7185,
      "theoretical_loss": 4.009424640943157,
      "tokens_seen": 402259968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006054993379988651,
      "loss": 3.0463,
      "theoretical_loss": 4.009351636606025,
      "tokens_seen": 402325504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006054827879704937,
      "loss": 2.9727,
      "theoretical_loss": 4.00927864748893,
      "tokens_seen": 402391040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006054662379421222,
      "loss": 2.9264,
      "theoretical_loss": 4.009205673586222,
      "tokens_seen": 402456576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006054496879137507,
      "loss": 2.7651,
      "theoretical_loss": 4.009132714892251,
      "tokens_seen": 402522112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006054331378853793,
      "loss": 3.0406,
      "theoretical_loss": 4.009059771401375,
      "tokens_seen": 402587648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006054165878570078,
      "loss": 2.8691,
      "theoretical_loss": 4.008986843107949,
      "tokens_seen": 402653184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006054000378286362,
      "loss": 2.6457,
      "theoretical_loss": 4.008913930006336,
      "tokens_seen": 402718720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006053834878002648,
      "loss": 2.9319,
      "theoretical_loss": 4.008841032090898,
      "tokens_seen": 402784256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006053669377718933,
      "loss": 3.008,
      "theoretical_loss": 4.008768149356003,
      "tokens_seen": 402849792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006053503877435219,
      "loss": 2.8642,
      "theoretical_loss": 4.0086952817960215,
      "tokens_seen": 402915328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006053338377151504,
      "loss": 2.9952,
      "theoretical_loss": 4.0086224294053245,
      "tokens_seen": 402980864
    },
    {
      "epoch": 0.14,
      "objective/train/avg_token_score": 0.027103539556264877,
      "objective/train/avg_weight": 0.9681052565574646,
      "objective/train/docs_used": 238677,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.831923484802246,
      "objective/train/std_weight": 0.11309883743524551,
      "objective/train/theoretical_loss": 4.008549592178291,
      "objective/train/tokens_used": 423506400,
      "theoretical_loss": 4.008549592178291,
      "tokens_seen": 403046400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006053172876867789,
      "loss": 2.7298,
      "theoretical_loss": 4.008549592178291,
      "tokens_seen": 403046400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006053007376584075,
      "loss": 2.8827,
      "theoretical_loss": 4.008476770109296,
      "tokens_seen": 403111936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000605284187630036,
      "loss": 2.9666,
      "theoretical_loss": 4.008403963192725,
      "tokens_seen": 403177472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006052676376016644,
      "loss": 3.0018,
      "theoretical_loss": 4.008331171422961,
      "tokens_seen": 403243008
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000605251087573293,
      "loss": 3.1806,
      "theoretical_loss": 4.008258394794392,
      "tokens_seen": 403308544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006052345375449215,
      "loss": 3.0075,
      "theoretical_loss": 4.0081856333014105,
      "tokens_seen": 403374080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006052179875165501,
      "loss": 2.8074,
      "theoretical_loss": 4.008112886938409,
      "tokens_seen": 403439616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006052014374881786,
      "loss": 2.7422,
      "theoretical_loss": 4.008040155699783,
      "tokens_seen": 403505152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006051848874598071,
      "loss": 2.8677,
      "theoretical_loss": 4.007967439579936,
      "tokens_seen": 403570688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006051683374314357,
      "loss": 2.8997,
      "theoretical_loss": 4.007894738573267,
      "tokens_seen": 403636224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006051517874030641,
      "loss": 3.1318,
      "theoretical_loss": 4.007822052674185,
      "tokens_seen": 403701760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006051352373746926,
      "loss": 2.8656,
      "theoretical_loss": 4.007749381877097,
      "tokens_seen": 403767296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006051186873463211,
      "loss": 2.7652,
      "theoretical_loss": 4.007676726176415,
      "tokens_seen": 403832832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006051021373179496,
      "loss": 2.8809,
      "theoretical_loss": 4.0076040855665545,
      "tokens_seen": 403898368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006050855872895782,
      "loss": 2.9537,
      "theoretical_loss": 4.007531460041933,
      "tokens_seen": 403963904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006050690372612067,
      "loss": 2.8345,
      "theoretical_loss": 4.00745884959697,
      "tokens_seen": 404029440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006050524872328353,
      "loss": 2.9955,
      "theoretical_loss": 4.00738625422609,
      "tokens_seen": 404094976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006050359372044637,
      "loss": 3.0293,
      "theoretical_loss": 4.00731367392372,
      "tokens_seen": 404160512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006050193871760922,
      "loss": 2.8323,
      "theoretical_loss": 4.0072411086842905,
      "tokens_seen": 404226048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006050028371477208,
      "loss": 2.9201,
      "theoretical_loss": 4.007168558502232,
      "tokens_seen": 404291584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006049862871193493,
      "loss": 2.888,
      "theoretical_loss": 4.007096023371981,
      "tokens_seen": 404357120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006049697370909778,
      "loss": 2.6901,
      "theoretical_loss": 4.007023503287976,
      "tokens_seen": 404422656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006049531870626064,
      "loss": 2.8074,
      "theoretical_loss": 4.006950998244658,
      "tokens_seen": 404488192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006049366370342349,
      "loss": 2.8603,
      "theoretical_loss": 4.006878508236473,
      "tokens_seen": 404553728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006049200870058635,
      "loss": 2.7214,
      "theoretical_loss": 4.006806033257866,
      "tokens_seen": 404619264
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006049035369774919,
      "loss": 2.8753,
      "theoretical_loss": 4.006733573303289,
      "tokens_seen": 404684800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006048869869491204,
      "loss": 3.069,
      "theoretical_loss": 4.0066611283671945,
      "tokens_seen": 404750336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000604870436920749,
      "loss": 2.9189,
      "theoretical_loss": 4.0065886984440375,
      "tokens_seen": 404815872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006048538868923775,
      "loss": 2.9973,
      "theoretical_loss": 4.00651628352828,
      "tokens_seen": 404881408
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000604837336864006,
      "loss": 2.9959,
      "theoretical_loss": 4.0064438836143825,
      "tokens_seen": 404946944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006048207868356346,
      "loss": 2.7651,
      "theoretical_loss": 4.00637149869681,
      "tokens_seen": 405012480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000604804236807263,
      "loss": 2.9381,
      "theoretical_loss": 4.0062991287700305,
      "tokens_seen": 405078016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006047876867788916,
      "loss": 2.9106,
      "theoretical_loss": 4.006226773828514,
      "tokens_seen": 405143552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006047711367505201,
      "loss": 2.6835,
      "theoretical_loss": 4.006154433866736,
      "tokens_seen": 405209088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006047545867221486,
      "loss": 3.0429,
      "theoretical_loss": 4.0060821088791725,
      "tokens_seen": 405274624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006047380366937772,
      "loss": 2.777,
      "theoretical_loss": 4.006009798860303,
      "tokens_seen": 405340160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006047214866654057,
      "loss": 2.7659,
      "theoretical_loss": 4.00593750380461,
      "tokens_seen": 405405696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006047049366370342,
      "loss": 2.9427,
      "theoretical_loss": 4.005865223706579,
      "tokens_seen": 405471232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006046883866086628,
      "loss": 2.9827,
      "theoretical_loss": 4.005792958560699,
      "tokens_seen": 405536768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006046718365802912,
      "loss": 2.6279,
      "theoretical_loss": 4.005720708361461,
      "tokens_seen": 405602304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006046552865519198,
      "loss": 2.9744,
      "theoretical_loss": 4.00564847310336,
      "tokens_seen": 405667840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006046387365235483,
      "loss": 3.1372,
      "theoretical_loss": 4.005576252780892,
      "tokens_seen": 405733376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006046221864951768,
      "loss": 2.6852,
      "theoretical_loss": 4.005504047388558,
      "tokens_seen": 405798912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006046056364668054,
      "loss": 2.736,
      "theoretical_loss": 4.00543185692086,
      "tokens_seen": 405864448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0006045890864384339,
      "loss": 2.8883,
      "theoretical_loss": 4.005359681372306,
      "tokens_seen": 405929984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006045725364100624,
      "loss": 2.6355,
      "theoretical_loss": 4.005287520737403,
      "tokens_seen": 405995520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000604555986381691,
      "loss": 2.8031,
      "theoretical_loss": 4.005215375010662,
      "tokens_seen": 406061056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006045394363533194,
      "loss": 2.8634,
      "theoretical_loss": 4.0051432441866,
      "tokens_seen": 406126592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000604522886324948,
      "loss": 2.5351,
      "theoretical_loss": 4.005071128259734,
      "tokens_seen": 406192128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006045063362965765,
      "loss": 2.9981,
      "theoretical_loss": 4.004999027224584,
      "tokens_seen": 406257664
    },
    {
      "epoch": 0.15,
      "objective/train/avg_token_score": 0.018662843853235245,
      "objective/train/avg_weight": 0.9780441522598267,
      "objective/train/docs_used": 240483,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0720653533935547,
      "objective/train/std_weight": 0.06829021126031876,
      "objective/train/theoretical_loss": 4.004926941075674,
      "objective/train/tokens_used": 426783200,
      "theoretical_loss": 4.004926941075674,
      "tokens_seen": 406323200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000604489786268205,
      "loss": 2.9059,
      "theoretical_loss": 4.004926941075674,
      "tokens_seen": 406323200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006044732362398336,
      "loss": 2.9516,
      "theoretical_loss": 4.004854869807528,
      "tokens_seen": 406388736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006044566862114621,
      "loss": 2.8326,
      "theoretical_loss": 4.004782813414678,
      "tokens_seen": 406454272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006044401361830907,
      "loss": 2.9458,
      "theoretical_loss": 4.004710771891654,
      "tokens_seen": 406519808
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006044235861547191,
      "loss": 2.7286,
      "theoretical_loss": 4.0046387452329935,
      "tokens_seen": 406585344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006044070361263476,
      "loss": 2.8757,
      "theoretical_loss": 4.004566733433231,
      "tokens_seen": 406650880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006043904860979762,
      "loss": 2.9712,
      "theoretical_loss": 4.00449473648691,
      "tokens_seen": 406716416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006043739360696047,
      "loss": 2.9292,
      "theoretical_loss": 4.004422754388573,
      "tokens_seen": 406781952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006043573860412332,
      "loss": 2.9344,
      "theoretical_loss": 4.004350787132766,
      "tokens_seen": 406847488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006043408360128618,
      "loss": 3.0945,
      "theoretical_loss": 4.004278834714039,
      "tokens_seen": 406913024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006043242859844903,
      "loss": 2.8423,
      "theoretical_loss": 4.004206897126943,
      "tokens_seen": 406978560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006043077359561189,
      "loss": 2.9779,
      "theoretical_loss": 4.004134974366036,
      "tokens_seen": 407044096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006042911859277473,
      "loss": 3.018,
      "theoretical_loss": 4.004063066425874,
      "tokens_seen": 407109632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006042746358993758,
      "loss": 2.7139,
      "theoretical_loss": 4.003991173301018,
      "tokens_seen": 407175168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006042580858710044,
      "loss": 2.8537,
      "theoretical_loss": 4.003919294986032,
      "tokens_seen": 407240704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006042415358426329,
      "loss": 3.0143,
      "theoretical_loss": 4.003847431475482,
      "tokens_seen": 407306240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006042249858142614,
      "loss": 2.8837,
      "theoretical_loss": 4.003775582763939,
      "tokens_seen": 407371776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00060420843578589,
      "loss": 2.8314,
      "theoretical_loss": 4.0037037488459735,
      "tokens_seen": 407437312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006041918857575185,
      "loss": 2.858,
      "theoretical_loss": 4.003631929716162,
      "tokens_seen": 407502848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000604175335729147,
      "loss": 3.0748,
      "theoretical_loss": 4.0035601253690825,
      "tokens_seen": 407568384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006041587857007755,
      "loss": 2.9391,
      "theoretical_loss": 4.0034883357993145,
      "tokens_seen": 407633920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000604142235672404,
      "loss": 2.94,
      "theoretical_loss": 4.003416561001444,
      "tokens_seen": 407699456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006041256856440325,
      "loss": 2.8855,
      "theoretical_loss": 4.003344800970056,
      "tokens_seen": 407764992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000604109135615661,
      "loss": 2.978,
      "theoretical_loss": 4.003273055699741,
      "tokens_seen": 407830528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006040925855872896,
      "loss": 2.7622,
      "theoretical_loss": 4.00320132518509,
      "tokens_seen": 407896064
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006040760355589181,
      "loss": 2.9469,
      "theoretical_loss": 4.0031296094207,
      "tokens_seen": 407961600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006040594855305465,
      "loss": 2.9325,
      "theoretical_loss": 4.003057908401168,
      "tokens_seen": 408027136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006040429355021751,
      "loss": 2.9811,
      "theoretical_loss": 4.002986222121095,
      "tokens_seen": 408092672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006040263854738036,
      "loss": 3.0231,
      "theoretical_loss": 4.002914550575085,
      "tokens_seen": 408158208
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006040098354454322,
      "loss": 2.8103,
      "theoretical_loss": 4.002842893757744,
      "tokens_seen": 408223744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006039932854170607,
      "loss": 2.8799,
      "theoretical_loss": 4.002771251663681,
      "tokens_seen": 408289280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006039767353886892,
      "loss": 2.9665,
      "theoretical_loss": 4.00269962428751,
      "tokens_seen": 408354816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006039601853603178,
      "loss": 2.8069,
      "theoretical_loss": 4.0026280116238455,
      "tokens_seen": 408420352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006039436353319462,
      "loss": 3.1032,
      "theoretical_loss": 4.002556413667305,
      "tokens_seen": 408485888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006039270853035747,
      "loss": 2.9818,
      "theoretical_loss": 4.0024848304125085,
      "tokens_seen": 408551424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006039105352752033,
      "loss": 2.9639,
      "theoretical_loss": 4.002413261854081,
      "tokens_seen": 408616960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006038939852468318,
      "loss": 3.0342,
      "theoretical_loss": 4.002341707986648,
      "tokens_seen": 408682496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006038774352184604,
      "loss": 2.9798,
      "theoretical_loss": 4.00227016880484,
      "tokens_seen": 408748032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006038608851900889,
      "loss": 2.9591,
      "theoretical_loss": 4.002198644303288,
      "tokens_seen": 408813568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006038443351617174,
      "loss": 2.7756,
      "theoretical_loss": 4.002127134476626,
      "tokens_seen": 408879104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000603827785133346,
      "loss": 3.0931,
      "theoretical_loss": 4.002055639319494,
      "tokens_seen": 408944640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006038112351049744,
      "loss": 3.174,
      "theoretical_loss": 4.00198415882653,
      "tokens_seen": 409010176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006037946850766029,
      "loss": 2.8227,
      "theoretical_loss": 4.001912692992379,
      "tokens_seen": 409075712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006037781350482315,
      "loss": 2.9207,
      "theoretical_loss": 4.001841241811688,
      "tokens_seen": 409141248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00060376158501986,
      "loss": 2.7679,
      "theoretical_loss": 4.001769805279103,
      "tokens_seen": 409206784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006037450349914886,
      "loss": 2.9633,
      "theoretical_loss": 4.001698383389279,
      "tokens_seen": 409272320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006037284849631171,
      "loss": 2.6494,
      "theoretical_loss": 4.001626976136868,
      "tokens_seen": 409337856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006037119349347456,
      "loss": 3.2612,
      "theoretical_loss": 4.001555583516529,
      "tokens_seen": 409403392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006036953849063741,
      "loss": 2.8935,
      "theoretical_loss": 4.001484205522922,
      "tokens_seen": 409468928
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006036788348780026,
      "loss": 3.0187,
      "theoretical_loss": 4.00141284215071,
      "tokens_seen": 409534464
    },
    {
      "epoch": 0.15,
      "objective/train/avg_token_score": 0.019540540874004364,
      "objective/train/avg_weight": 0.97700434923172,
      "objective/train/docs_used": 242341,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2874226570129395,
      "objective/train/std_weight": 0.1036716029047966,
      "objective/train/theoretical_loss": 4.001341493394558,
      "objective/train/tokens_used": 430060000,
      "theoretical_loss": 4.001341493394558,
      "tokens_seen": 409600000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006036622848496311,
      "loss": 3.1489,
      "theoretical_loss": 4.001341493394558,
      "tokens_seen": 409600000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006036457348212597,
      "loss": 2.9013,
      "theoretical_loss": 4.0012701592491355,
      "tokens_seen": 409665536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006036291847928882,
      "loss": 2.9259,
      "theoretical_loss": 4.001198839709113,
      "tokens_seen": 409731072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006036126347645168,
      "loss": 3.0363,
      "theoretical_loss": 4.001127534769166,
      "tokens_seen": 409796608
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006035960847361453,
      "loss": 3.0913,
      "theoretical_loss": 4.001056244423971,
      "tokens_seen": 409862144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006035795347077737,
      "loss": 2.8218,
      "theoretical_loss": 4.000984968668206,
      "tokens_seen": 409927680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006035629846794023,
      "loss": 2.7808,
      "theoretical_loss": 4.000913707496556,
      "tokens_seen": 409993216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006035464346510308,
      "loss": 2.8761,
      "theoretical_loss": 4.000842460903706,
      "tokens_seen": 410058752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006035298846226593,
      "loss": 2.9348,
      "theoretical_loss": 4.000771228884344,
      "tokens_seen": 410124288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006035133345942879,
      "loss": 2.8063,
      "theoretical_loss": 4.0007000114331595,
      "tokens_seen": 410189824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006034967845659164,
      "loss": 2.8414,
      "theoretical_loss": 4.000628808544848,
      "tokens_seen": 410255360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000603480234537545,
      "loss": 2.9596,
      "theoretical_loss": 4.000557620214105,
      "tokens_seen": 410320896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006034636845091735,
      "loss": 3.2222,
      "theoretical_loss": 4.000486446435629,
      "tokens_seen": 410386432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006034471344808019,
      "loss": 2.8812,
      "theoretical_loss": 4.000415287204124,
      "tokens_seen": 410451968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006034305844524305,
      "loss": 2.888,
      "theoretical_loss": 4.000344142514294,
      "tokens_seen": 410517504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000603414034424059,
      "loss": 2.8851,
      "theoretical_loss": 4.0002730123608465,
      "tokens_seen": 410583040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006033974843956875,
      "loss": 2.9571,
      "theoretical_loss": 4.000201896738492,
      "tokens_seen": 410648576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006033809343673161,
      "loss": 3.0713,
      "theoretical_loss": 4.000130795641944,
      "tokens_seen": 410714112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006033643843389446,
      "loss": 2.9244,
      "theoretical_loss": 4.0000597090659165,
      "tokens_seen": 410779648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006033478343105732,
      "loss": 3.1038,
      "theoretical_loss": 3.9999886370051305,
      "tokens_seen": 410845184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006033312842822016,
      "loss": 2.8676,
      "theoretical_loss": 3.9999175794543063,
      "tokens_seen": 410910720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006033147342538301,
      "loss": 2.8773,
      "theoretical_loss": 3.9998465364081683,
      "tokens_seen": 410976256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006032981842254587,
      "loss": 2.7587,
      "theoretical_loss": 3.999775507861443,
      "tokens_seen": 411041792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006032816341970872,
      "loss": 2.9289,
      "theoretical_loss": 3.999704493808861,
      "tokens_seen": 411107328
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006032650841687157,
      "loss": 2.8067,
      "theoretical_loss": 3.9996334942451544,
      "tokens_seen": 411172864
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006032485341403443,
      "loss": 3.0616,
      "theoretical_loss": 3.999562509165058,
      "tokens_seen": 411238400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006032319841119728,
      "loss": 2.9315,
      "theoretical_loss": 3.9994915385633103,
      "tokens_seen": 411303936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006032154340836014,
      "loss": 2.9422,
      "theoretical_loss": 3.9994205824346514,
      "tokens_seen": 411369472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006031988840552298,
      "loss": 2.846,
      "theoretical_loss": 3.9993496407738256,
      "tokens_seen": 411435008
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006031823340268583,
      "loss": 2.8205,
      "theoretical_loss": 3.9992787135755785,
      "tokens_seen": 411500544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006031657839984869,
      "loss": 3.1717,
      "theoretical_loss": 3.999207800834659,
      "tokens_seen": 411566080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006031492339701153,
      "loss": 2.8367,
      "theoretical_loss": 3.9991369025458194,
      "tokens_seen": 411631616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006031326839417439,
      "loss": 2.9074,
      "theoretical_loss": 3.999066018703814,
      "tokens_seen": 411697152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006031161339133724,
      "loss": 3.0979,
      "theoretical_loss": 3.9989951493033997,
      "tokens_seen": 411762688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000603099583885001,
      "loss": 2.905,
      "theoretical_loss": 3.9989242943393357,
      "tokens_seen": 411828224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006030830338566294,
      "loss": 3.0004,
      "theoretical_loss": 3.998853453806386,
      "tokens_seen": 411893760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006030664838282579,
      "loss": 2.9747,
      "theoretical_loss": 3.998782627699315,
      "tokens_seen": 411959296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006030499337998865,
      "loss": 2.9215,
      "theoretical_loss": 3.998711816012891,
      "tokens_seen": 412024832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000603033383771515,
      "loss": 2.8628,
      "theoretical_loss": 3.9986410187418855,
      "tokens_seen": 412090368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006030168337431435,
      "loss": 2.9075,
      "theoretical_loss": 3.9985702358810715,
      "tokens_seen": 412155904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006030002837147721,
      "loss": 3.0602,
      "theoretical_loss": 3.9984994674252254,
      "tokens_seen": 412221440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006029837336864006,
      "loss": 3.0922,
      "theoretical_loss": 3.998428713369126,
      "tokens_seen": 412286976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006029671836580291,
      "loss": 2.917,
      "theoretical_loss": 3.9983579737075554,
      "tokens_seen": 412352512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006029506336296576,
      "loss": 2.9473,
      "theoretical_loss": 3.9982872484352976,
      "tokens_seen": 412418048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006029340836012861,
      "loss": 2.8945,
      "theoretical_loss": 3.9982165375471403,
      "tokens_seen": 412483584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006029175335729147,
      "loss": 2.9238,
      "theoretical_loss": 3.998145841037873,
      "tokens_seen": 412549120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006029009835445432,
      "loss": 3.0687,
      "theoretical_loss": 3.9980751589022887,
      "tokens_seen": 412614656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006028844335161717,
      "loss": 2.9225,
      "theoretical_loss": 3.9980044911351826,
      "tokens_seen": 412680192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006028678834878003,
      "loss": 3.0446,
      "theoretical_loss": 3.9979338377313525,
      "tokens_seen": 412745728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006028513334594287,
      "loss": 2.9375,
      "theoretical_loss": 3.997863198685599,
      "tokens_seen": 412811264
    },
    {
      "epoch": 0.15,
      "objective/train/avg_token_score": 0.019250638782978058,
      "objective/train/avg_weight": 0.9773463010787964,
      "objective/train/docs_used": 244062,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0527615547180176,
      "objective/train/std_weight": 0.09655958414077759,
      "objective/train/theoretical_loss": 3.997792573992726,
      "objective/train/tokens_used": 433336800,
      "theoretical_loss": 3.997792573992726,
      "tokens_seen": 412876800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006028347834310573,
      "loss": 3.0389,
      "theoretical_loss": 3.997792573992726,
      "tokens_seen": 412876800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006028182334026858,
      "loss": 3.1346,
      "theoretical_loss": 3.997721963647539,
      "tokens_seen": 412942336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006028016833743143,
      "loss": 2.8934,
      "theoretical_loss": 3.997651367644848,
      "tokens_seen": 413007872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006027851333459429,
      "loss": 2.8238,
      "theoretical_loss": 3.997580785979464,
      "tokens_seen": 413073408
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006027685833175714,
      "loss": 2.9822,
      "theoretical_loss": 3.9975102186462004,
      "tokens_seen": 413138944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006027520332891999,
      "loss": 2.86,
      "theoretical_loss": 3.997439665639875,
      "tokens_seen": 413204480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006027354832608285,
      "loss": 2.7938,
      "theoretical_loss": 3.9973691269553076,
      "tokens_seen": 413270016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006027189332324569,
      "loss": 2.5687,
      "theoretical_loss": 3.9972986025873203,
      "tokens_seen": 413335552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006027023832040855,
      "loss": 2.7916,
      "theoretical_loss": 3.9972280925307384,
      "tokens_seen": 413401088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000602685833175714,
      "loss": 2.8267,
      "theoretical_loss": 3.9971575967803887,
      "tokens_seen": 413466624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006026692831473425,
      "loss": 2.8364,
      "theoretical_loss": 3.9970871153311025,
      "tokens_seen": 413532160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006026527331189711,
      "loss": 2.8728,
      "theoretical_loss": 3.9970166481777127,
      "tokens_seen": 413597696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006026361830905996,
      "loss": 2.7794,
      "theoretical_loss": 3.996946195315055,
      "tokens_seen": 413663232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006026196330622281,
      "loss": 3.2335,
      "theoretical_loss": 3.996875756737968,
      "tokens_seen": 413728768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006026030830338566,
      "loss": 3.0036,
      "theoretical_loss": 3.996805332441293,
      "tokens_seen": 413794304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006025865330054851,
      "loss": 2.9018,
      "theoretical_loss": 3.996734922419873,
      "tokens_seen": 413859840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006025699829771137,
      "loss": 2.9107,
      "theoretical_loss": 3.996664526668555,
      "tokens_seen": 413925376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006025534329487422,
      "loss": 3.0185,
      "theoretical_loss": 3.9965941451821894,
      "tokens_seen": 413990912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006025368829203707,
      "loss": 2.9568,
      "theoretical_loss": 3.996523777955626,
      "tokens_seen": 414056448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006025203328919993,
      "loss": 2.8617,
      "theoretical_loss": 3.996453424983721,
      "tokens_seen": 414121984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006025037828636278,
      "loss": 2.9728,
      "theoretical_loss": 3.9963830862613303,
      "tokens_seen": 414187520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006024872328352562,
      "loss": 2.7585,
      "theoretical_loss": 3.9963127617833143,
      "tokens_seen": 414253056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006024706828068848,
      "loss": 3.045,
      "theoretical_loss": 3.9962424515445356,
      "tokens_seen": 414318592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006024541327785133,
      "loss": 3.0163,
      "theoretical_loss": 3.99617215553986,
      "tokens_seen": 414384128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006024375827501419,
      "loss": 2.7367,
      "theoretical_loss": 3.9961018737641547,
      "tokens_seen": 414449664
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006024210327217704,
      "loss": 2.9487,
      "theoretical_loss": 3.99603160621229,
      "tokens_seen": 414515200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006024044826933989,
      "loss": 2.8197,
      "theoretical_loss": 3.9959613528791396,
      "tokens_seen": 414580736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006023879326650275,
      "loss": 3.088,
      "theoretical_loss": 3.995891113759579,
      "tokens_seen": 414646272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000602371382636656,
      "loss": 2.8906,
      "theoretical_loss": 3.9958208888484865,
      "tokens_seen": 414711808
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006023548326082844,
      "loss": 3.0442,
      "theoretical_loss": 3.9957506781407446,
      "tokens_seen": 414777344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000602338282579913,
      "loss": 2.8999,
      "theoretical_loss": 3.9956804816312355,
      "tokens_seen": 414842880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006023217325515415,
      "loss": 2.7277,
      "theoretical_loss": 3.995610299314846,
      "tokens_seen": 414908416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006023051825231701,
      "loss": 3.0588,
      "theoretical_loss": 3.9955401311864662,
      "tokens_seen": 414973952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006022886324947986,
      "loss": 3.1365,
      "theoretical_loss": 3.9954699772409867,
      "tokens_seen": 415039488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006022720824664271,
      "loss": 3.0951,
      "theoretical_loss": 3.995399837473302,
      "tokens_seen": 415105024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006022555324380557,
      "loss": 2.8913,
      "theoretical_loss": 3.9953297118783104,
      "tokens_seen": 415170560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006022389824096842,
      "loss": 2.7535,
      "theoretical_loss": 3.99525960045091,
      "tokens_seen": 415236096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006022224323813126,
      "loss": 2.924,
      "theoretical_loss": 3.9951895031860043,
      "tokens_seen": 415301632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006022058823529412,
      "loss": 2.9323,
      "theoretical_loss": 3.9951194200784976,
      "tokens_seen": 415367168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006021893323245697,
      "loss": 3.1,
      "theoretical_loss": 3.9950493511232974,
      "tokens_seen": 415432704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006021727822961982,
      "loss": 3.0412,
      "theoretical_loss": 3.9949792963153143,
      "tokens_seen": 415498240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006021562322678267,
      "loss": 2.7998,
      "theoretical_loss": 3.9949092556494614,
      "tokens_seen": 415563776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006021396822394553,
      "loss": 2.6608,
      "theoretical_loss": 3.994839229120654,
      "tokens_seen": 415629312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006021231322110838,
      "loss": 2.5011,
      "theoretical_loss": 3.99476921672381,
      "tokens_seen": 415694848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006021065821827122,
      "loss": 2.9408,
      "theoretical_loss": 3.99469921845385,
      "tokens_seen": 415760384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006020900321543408,
      "loss": 2.8742,
      "theoretical_loss": 3.994629234305698,
      "tokens_seen": 415825920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006020734821259693,
      "loss": 2.8661,
      "theoretical_loss": 3.9945592642742795,
      "tokens_seen": 415891456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006020569320975978,
      "loss": 2.7917,
      "theoretical_loss": 3.994489308354523,
      "tokens_seen": 415956992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006020403820692264,
      "loss": 2.8565,
      "theoretical_loss": 3.9944193665413605,
      "tokens_seen": 416022528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006020238320408549,
      "loss": 2.8506,
      "theoretical_loss": 3.9943494388297256,
      "tokens_seen": 416088064
    },
    {
      "epoch": 0.15,
      "objective/train/avg_token_score": 0.01576858013868332,
      "objective/train/avg_weight": 0.9814376831054688,
      "objective/train/docs_used": 245710,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.23366117477417,
      "objective/train/std_weight": 0.07803238928318024,
      "objective/train/theoretical_loss": 3.994279525214554,
      "objective/train/tokens_used": 436613600,
      "theoretical_loss": 3.994279525214554,
      "tokens_seen": 416153600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006020072820124835,
      "loss": 2.9133,
      "theoretical_loss": 3.994279525214554,
      "tokens_seen": 416153600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006019907319841119,
      "loss": 2.8809,
      "theoretical_loss": 3.9942096256907864,
      "tokens_seen": 416219136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006019741819557404,
      "loss": 2.8438,
      "theoretical_loss": 3.994139740253363,
      "tokens_seen": 416284672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000601957631927369,
      "loss": 2.9436,
      "theoretical_loss": 3.9940698688972285,
      "tokens_seen": 416350208
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006019410818989975,
      "loss": 2.7825,
      "theoretical_loss": 3.99400001161733,
      "tokens_seen": 416415744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006019245318706261,
      "loss": 2.8202,
      "theoretical_loss": 3.9939301684086175,
      "tokens_seen": 416481280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006019079818422546,
      "loss": 3.0151,
      "theoretical_loss": 3.9938603392660426,
      "tokens_seen": 416546816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006018914318138831,
      "loss": 2.8155,
      "theoretical_loss": 3.9937905241845595,
      "tokens_seen": 416612352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006018748817855117,
      "loss": 2.9385,
      "theoretical_loss": 3.993720723159127,
      "tokens_seen": 416677888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006018583317571401,
      "loss": 3.071,
      "theoretical_loss": 3.9936509361847037,
      "tokens_seen": 416743424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006018417817287686,
      "loss": 3.0581,
      "theoretical_loss": 3.9935811632562532,
      "tokens_seen": 416808960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006018252317003972,
      "loss": 2.7961,
      "theoretical_loss": 3.99351140436874,
      "tokens_seen": 416874496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006018086816720257,
      "loss": 3.0128,
      "theoretical_loss": 3.9934416595171323,
      "tokens_seen": 416940032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006017921316436543,
      "loss": 2.8181,
      "theoretical_loss": 3.9933719286963996,
      "tokens_seen": 417005568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006017755816152828,
      "loss": 2.8732,
      "theoretical_loss": 3.993302211901516,
      "tokens_seen": 417071104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006017590315869113,
      "loss": 2.9054,
      "theoretical_loss": 3.9932325091274565,
      "tokens_seen": 417136640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006017424815585398,
      "loss": 2.787,
      "theoretical_loss": 3.993162820369199,
      "tokens_seen": 417202176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006017259315301683,
      "loss": 2.85,
      "theoretical_loss": 3.9930931456217245,
      "tokens_seen": 417267712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006017093815017968,
      "loss": 2.748,
      "theoretical_loss": 3.993023484880016,
      "tokens_seen": 417333248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006016928314734254,
      "loss": 2.9354,
      "theoretical_loss": 3.9929538381390595,
      "tokens_seen": 417398784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006016762814450539,
      "loss": 2.8928,
      "theoretical_loss": 3.9928842053938443,
      "tokens_seen": 417464320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006016597314166825,
      "loss": 2.9578,
      "theoretical_loss": 3.9928145866393594,
      "tokens_seen": 417529856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000601643181388311,
      "loss": 2.9157,
      "theoretical_loss": 3.9927449818706005,
      "tokens_seen": 417595392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006016266313599394,
      "loss": 3.0945,
      "theoretical_loss": 3.992675391082563,
      "tokens_seen": 417660928
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000601610081331568,
      "loss": 2.9656,
      "theoretical_loss": 3.9926058142702456,
      "tokens_seen": 417726464
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006015935313031965,
      "loss": 2.9943,
      "theoretical_loss": 3.9925362514286498,
      "tokens_seen": 417792000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000601576981274825,
      "loss": 2.8964,
      "theoretical_loss": 3.992466702552779,
      "tokens_seen": 417857536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006015604312464536,
      "loss": 2.9031,
      "theoretical_loss": 3.99239716763764,
      "tokens_seen": 417923072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006015438812180821,
      "loss": 2.9679,
      "theoretical_loss": 3.9923276466782425,
      "tokens_seen": 417988608
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006015273311897107,
      "loss": 3.0948,
      "theoretical_loss": 3.9922581396695973,
      "tokens_seen": 418054144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006015107811613392,
      "loss": 2.7882,
      "theoretical_loss": 3.992188646606719,
      "tokens_seen": 418119680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006014942311329676,
      "loss": 3.0483,
      "theoretical_loss": 3.992119167484624,
      "tokens_seen": 418185216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006014776811045962,
      "loss": 3.2243,
      "theoretical_loss": 3.992049702298332,
      "tokens_seen": 418250752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006014611310762247,
      "loss": 2.9877,
      "theoretical_loss": 3.9919802510428646,
      "tokens_seen": 418316288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006014445810478532,
      "loss": 2.9311,
      "theoretical_loss": 3.991910813713246,
      "tokens_seen": 418381824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006014280310194818,
      "loss": 3.0654,
      "theoretical_loss": 3.991841390304504,
      "tokens_seen": 418447360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006014114809911103,
      "loss": 3.2646,
      "theoretical_loss": 3.991771980811667,
      "tokens_seen": 418512896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006013949309627389,
      "loss": 2.9017,
      "theoretical_loss": 3.991702585229768,
      "tokens_seen": 418578432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006013783809343673,
      "loss": 2.8493,
      "theoretical_loss": 3.9916332035538415,
      "tokens_seen": 418643968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006013618309059958,
      "loss": 2.804,
      "theoretical_loss": 3.9915638357789245,
      "tokens_seen": 418709504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006013452808776244,
      "loss": 2.8971,
      "theoretical_loss": 3.991494481900056,
      "tokens_seen": 418775040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006013287308492529,
      "loss": 3.2956,
      "theoretical_loss": 3.9914251419122797,
      "tokens_seen": 418840576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006013121808208814,
      "loss": 3.085,
      "theoretical_loss": 3.99135581581064,
      "tokens_seen": 418906112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00060129563079251,
      "loss": 3.0165,
      "theoretical_loss": 3.991286503590184,
      "tokens_seen": 418971648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006012790807641385,
      "loss": 2.9246,
      "theoretical_loss": 3.9912172052459614,
      "tokens_seen": 419037184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000601262530735767,
      "loss": 2.91,
      "theoretical_loss": 3.9911479207730247,
      "tokens_seen": 419102720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006012459807073955,
      "loss": 2.8413,
      "theoretical_loss": 3.991078650166429,
      "tokens_seen": 419168256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000601229430679024,
      "loss": 3.0239,
      "theoretical_loss": 3.9910093934212325,
      "tokens_seen": 419233792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006012128806506526,
      "loss": 2.9822,
      "theoretical_loss": 3.990940150532494,
      "tokens_seen": 419299328
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006011963306222811,
      "loss": 2.8648,
      "theoretical_loss": 3.990870921495277,
      "tokens_seen": 419364864
    },
    {
      "epoch": 0.15,
      "objective/train/avg_token_score": 0.026410864666104317,
      "objective/train/avg_weight": 0.9689216017723083,
      "objective/train/docs_used": 247828,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0506021976470947,
      "objective/train/std_weight": 0.12447349727153778,
      "objective/train/theoretical_loss": 3.990801706304647,
      "objective/train/tokens_used": 439890400,
      "theoretical_loss": 3.990801706304647,
      "tokens_seen": 419430400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006011797805939096,
      "loss": 3.0278,
      "theoretical_loss": 3.990801706304647,
      "tokens_seen": 419430400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006011632305655381,
      "loss": 3.0526,
      "theoretical_loss": 3.99073250495567,
      "tokens_seen": 419495936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006011466805371665,
      "loss": 3.0269,
      "theoretical_loss": 3.990663317443418,
      "tokens_seen": 419561472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006011301305087951,
      "loss": 2.9892,
      "theoretical_loss": 3.990594143762962,
      "tokens_seen": 419627008
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006011135804804236,
      "loss": 2.8384,
      "theoretical_loss": 3.9905249839093786,
      "tokens_seen": 419692544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006010970304520522,
      "loss": 2.7901,
      "theoretical_loss": 3.9904558378777453,
      "tokens_seen": 419758080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006010804804236807,
      "loss": 2.8295,
      "theoretical_loss": 3.990386705663142,
      "tokens_seen": 419823616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006010639303953092,
      "loss": 2.8307,
      "theoretical_loss": 3.9903175872606518,
      "tokens_seen": 419889152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006010473803669378,
      "loss": 3.0681,
      "theoretical_loss": 3.990248482665359,
      "tokens_seen": 419954688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006010308303385663,
      "loss": 2.9657,
      "theoretical_loss": 3.9901793918723527,
      "tokens_seen": 420020224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006010142803101947,
      "loss": 2.8852,
      "theoretical_loss": 3.9901103148767225,
      "tokens_seen": 420085760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006009977302818233,
      "loss": 2.7275,
      "theoretical_loss": 3.9900412516735617,
      "tokens_seen": 420151296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006009811802534518,
      "loss": 3.1769,
      "theoretical_loss": 3.989972202257966,
      "tokens_seen": 420216832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006009646302250804,
      "loss": 2.7542,
      "theoretical_loss": 3.989903166625032,
      "tokens_seen": 420282368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006009480801967089,
      "loss": 2.9263,
      "theoretical_loss": 3.989834144769861,
      "tokens_seen": 420347904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006009315301683374,
      "loss": 2.8963,
      "theoretical_loss": 3.9897651366875557,
      "tokens_seen": 420413440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000600914980139966,
      "loss": 2.9346,
      "theoretical_loss": 3.989696142373221,
      "tokens_seen": 420478976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006008984301115944,
      "loss": 2.9667,
      "theoretical_loss": 3.9896271618219656,
      "tokens_seen": 420544512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006008818800832229,
      "loss": 2.9233,
      "theoretical_loss": 3.9895581950288994,
      "tokens_seen": 420610048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006008653300548515,
      "loss": 3.0638,
      "theoretical_loss": 3.9894892419891352,
      "tokens_seen": 420675584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00060084878002648,
      "loss": 2.596,
      "theoretical_loss": 3.989420302697789,
      "tokens_seen": 420741120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006008322299981086,
      "loss": 2.799,
      "theoretical_loss": 3.989351377149978,
      "tokens_seen": 420806656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006008156799697371,
      "loss": 3.0725,
      "theoretical_loss": 3.989282465340823,
      "tokens_seen": 420872192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006007991299413656,
      "loss": 2.957,
      "theoretical_loss": 3.9892135672654465,
      "tokens_seen": 420937728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006007825799129942,
      "loss": 3.1583,
      "theoretical_loss": 3.9891446829189734,
      "tokens_seen": 421003264
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006007660298846226,
      "loss": 3.0023,
      "theoretical_loss": 3.989075812296533,
      "tokens_seen": 421068800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006007494798562511,
      "loss": 3.0324,
      "theoretical_loss": 3.989006955393254,
      "tokens_seen": 421134336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006007329298278797,
      "loss": 2.9037,
      "theoretical_loss": 3.98893811220427,
      "tokens_seen": 421199872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006007163797995082,
      "loss": 2.8968,
      "theoretical_loss": 3.9888692827247167,
      "tokens_seen": 421265408
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006006998297711368,
      "loss": 3.06,
      "theoretical_loss": 3.9888004669497317,
      "tokens_seen": 421330944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006006832797427653,
      "loss": 2.8874,
      "theoretical_loss": 3.988731664874454,
      "tokens_seen": 421396480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006006667297143938,
      "loss": 2.9268,
      "theoretical_loss": 3.9886628764940277,
      "tokens_seen": 421462016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006006501796860223,
      "loss": 2.8386,
      "theoretical_loss": 3.988594101803598,
      "tokens_seen": 421527552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006006336296576508,
      "loss": 2.6843,
      "theoretical_loss": 3.9885253407983114,
      "tokens_seen": 421593088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006006170796292793,
      "loss": 3.0944,
      "theoretical_loss": 3.9884565934733196,
      "tokens_seen": 421658624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006006005296009079,
      "loss": 2.8298,
      "theoretical_loss": 3.9883878598237743,
      "tokens_seen": 421724160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006005839795725364,
      "loss": 2.7653,
      "theoretical_loss": 3.9883191398448306,
      "tokens_seen": 421789696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000600567429544165,
      "loss": 2.6758,
      "theoretical_loss": 3.9882504335316464,
      "tokens_seen": 421855232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006005508795157935,
      "loss": 2.7807,
      "theoretical_loss": 3.988181740879382,
      "tokens_seen": 421920768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006005343294874219,
      "loss": 2.9442,
      "theoretical_loss": 3.9881130618831984,
      "tokens_seen": 421986304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006005177794590505,
      "loss": 2.8146,
      "theoretical_loss": 3.988044396538263,
      "tokens_seen": 422051840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000600501229430679,
      "loss": 2.7209,
      "theoretical_loss": 3.9879757448397406,
      "tokens_seen": 422117376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006004846794023076,
      "loss": 2.8777,
      "theoretical_loss": 3.9879071067828034,
      "tokens_seen": 422182912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006004681293739361,
      "loss": 2.8835,
      "theoretical_loss": 3.9878384823626227,
      "tokens_seen": 422248448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006004515793455646,
      "loss": 2.8242,
      "theoretical_loss": 3.987769871574373,
      "tokens_seen": 422313984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006004350293171932,
      "loss": 3.1026,
      "theoretical_loss": 3.9877012744132316,
      "tokens_seen": 422379520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006004184792888217,
      "loss": 2.9139,
      "theoretical_loss": 3.987632690874379,
      "tokens_seen": 422445056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006004019292604501,
      "loss": 2.9013,
      "theoretical_loss": 3.987564120952997,
      "tokens_seen": 422510592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006003853792320787,
      "loss": 2.686,
      "theoretical_loss": 3.98749556464427,
      "tokens_seen": 422576128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006003688292037072,
      "loss": 2.8089,
      "theoretical_loss": 3.987427021943385,
      "tokens_seen": 422641664
    },
    {
      "epoch": 0.15,
      "objective/train/avg_token_score": 0.023017197847366333,
      "objective/train/avg_weight": 0.9729694724082947,
      "objective/train/docs_used": 249572,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.6010639667510986,
      "objective/train/std_weight": 0.13741151988506317,
      "objective/train/theoretical_loss": 3.987358492845532,
      "objective/train/tokens_used": 443167200,
      "theoretical_loss": 3.987358492845532,
      "tokens_seen": 422707200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006003522791753358,
      "loss": 2.9603,
      "theoretical_loss": 3.987358492845532,
      "tokens_seen": 422707200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006003357291469643,
      "loss": 2.9444,
      "theoretical_loss": 3.9872899773459025,
      "tokens_seen": 422772736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006003191791185928,
      "loss": 3.0402,
      "theoretical_loss": 3.987221475439691,
      "tokens_seen": 422838272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006003026290902214,
      "loss": 3.085,
      "theoretical_loss": 3.9871529871220943,
      "tokens_seen": 422903808
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006002860790618498,
      "loss": 3.1126,
      "theoretical_loss": 3.9870845123883116,
      "tokens_seen": 422969344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006002695290334783,
      "loss": 3.2228,
      "theoretical_loss": 3.987016051233546,
      "tokens_seen": 423034880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006002529790051069,
      "loss": 2.8522,
      "theoretical_loss": 3.9869476036529994,
      "tokens_seen": 423100416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006002364289767354,
      "loss": 3.008,
      "theoretical_loss": 3.98687916964188,
      "tokens_seen": 423165952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000600219878948364,
      "loss": 2.8949,
      "theoretical_loss": 3.986810749195396,
      "tokens_seen": 423231488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006002033289199924,
      "loss": 3.0243,
      "theoretical_loss": 3.9867423423087596,
      "tokens_seen": 423297024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000600186778891621,
      "loss": 3.067,
      "theoretical_loss": 3.9866739489771845,
      "tokens_seen": 423362560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006001702288632494,
      "loss": 3.2078,
      "theoretical_loss": 3.986605569195887,
      "tokens_seen": 423428096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006001536788348779,
      "loss": 3.049,
      "theoretical_loss": 3.986537202960085,
      "tokens_seen": 423493632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006001371288065065,
      "loss": 2.9201,
      "theoretical_loss": 3.986468850265001,
      "tokens_seen": 423559168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000600120578778135,
      "loss": 2.9303,
      "theoretical_loss": 3.986400511105858,
      "tokens_seen": 423624704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006001040287497635,
      "loss": 2.9767,
      "theoretical_loss": 3.9863321854778815,
      "tokens_seen": 423690240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006000874787213921,
      "loss": 2.8814,
      "theoretical_loss": 3.986263873376301,
      "tokens_seen": 423755776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006000709286930206,
      "loss": 2.9389,
      "theoretical_loss": 3.986195574796347,
      "tokens_seen": 423821312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006000543786646492,
      "loss": 2.8298,
      "theoretical_loss": 3.9861272897332527,
      "tokens_seen": 423886848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006000378286362776,
      "loss": 2.5864,
      "theoretical_loss": 3.986059018182254,
      "tokens_seen": 423952384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006000212786079061,
      "loss": 2.852,
      "theoretical_loss": 3.9859907601385887,
      "tokens_seen": 424017920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0006000047285795347,
      "loss": 2.8718,
      "theoretical_loss": 3.9859225155974976,
      "tokens_seen": 424083456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005999881785511632,
      "loss": 2.984,
      "theoretical_loss": 3.985854284554223,
      "tokens_seen": 424148992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005999716285227917,
      "loss": 2.8675,
      "theoretical_loss": 3.9857860670040113,
      "tokens_seen": 424214528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005999550784944203,
      "loss": 2.9553,
      "theoretical_loss": 3.98571786294211,
      "tokens_seen": 424280064
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005999385284660488,
      "loss": 3.0688,
      "theoretical_loss": 3.9856496723637687,
      "tokens_seen": 424345600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005999219784376773,
      "loss": 3.0196,
      "theoretical_loss": 3.985581495264241,
      "tokens_seen": 424411136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005999054284093058,
      "loss": 3.04,
      "theoretical_loss": 3.9855133316387805,
      "tokens_seen": 424476672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005998888783809343,
      "loss": 3.1181,
      "theoretical_loss": 3.9854451814826457,
      "tokens_seen": 424542208
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005998723283525629,
      "loss": 2.9238,
      "theoretical_loss": 3.985377044791096,
      "tokens_seen": 424607744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005998557783241914,
      "loss": 2.9551,
      "theoretical_loss": 3.985308921559394,
      "tokens_seen": 424673280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005998392282958199,
      "loss": 2.8932,
      "theoretical_loss": 3.9852408117828038,
      "tokens_seen": 424738816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005998226782674485,
      "loss": 2.7572,
      "theoretical_loss": 3.9851727154565926,
      "tokens_seen": 424804352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000599806128239077,
      "loss": 2.7198,
      "theoretical_loss": 3.98510463257603,
      "tokens_seen": 424869888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005997895782107055,
      "loss": 2.7956,
      "theoretical_loss": 3.9850365631363873,
      "tokens_seen": 424935424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000599773028182334,
      "loss": 2.546,
      "theoretical_loss": 3.9849685071329386,
      "tokens_seen": 425000960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005997564781539625,
      "loss": 2.9586,
      "theoretical_loss": 3.984900464560961,
      "tokens_seen": 425066496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005997399281255911,
      "loss": 3.0618,
      "theoretical_loss": 3.984832435415733,
      "tokens_seen": 425132032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005997233780972196,
      "loss": 2.8011,
      "theoretical_loss": 3.984764419692537,
      "tokens_seen": 425197568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005997068280688481,
      "loss": 2.8648,
      "theoretical_loss": 3.984696417386655,
      "tokens_seen": 425263104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005996902780404767,
      "loss": 3.0058,
      "theoretical_loss": 3.984628428493374,
      "tokens_seen": 425328640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005996737280121051,
      "loss": 3.0637,
      "theoretical_loss": 3.9845604530079832,
      "tokens_seen": 425394176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005996571779837337,
      "loss": 2.9116,
      "theoretical_loss": 3.984492490925772,
      "tokens_seen": 425459712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005996406279553622,
      "loss": 2.882,
      "theoretical_loss": 3.984424542242034,
      "tokens_seen": 425525248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005996240779269907,
      "loss": 2.9388,
      "theoretical_loss": 3.9843566069520664,
      "tokens_seen": 425590784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005996075278986193,
      "loss": 3.1039,
      "theoretical_loss": 3.9842886850511654,
      "tokens_seen": 425656320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005995909778702478,
      "loss": 2.959,
      "theoretical_loss": 3.984220776534632,
      "tokens_seen": 425721856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005995744278418763,
      "loss": 2.6236,
      "theoretical_loss": 3.984152881397769,
      "tokens_seen": 425787392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005995578778135048,
      "loss": 3.0044,
      "theoretical_loss": 3.9840849996358814,
      "tokens_seen": 425852928
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005995413277851333,
      "loss": 3.0222,
      "theoretical_loss": 3.9840171312442765,
      "tokens_seen": 425918464
    },
    {
      "epoch": 0.15,
      "objective/train/avg_token_score": 0.040628284215927124,
      "objective/train/avg_weight": 0.9522133469581604,
      "objective/train/docs_used": 251147,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8920133113861084,
      "objective/train/std_weight": 0.16872364282608032,
      "objective/train/theoretical_loss": 3.9839492762182647,
      "objective/train/tokens_used": 446444000,
      "theoretical_loss": 3.9839492762182647,
      "tokens_seen": 425984000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005995247777567619,
      "loss": 2.8212,
      "theoretical_loss": 3.9839492762182647,
      "tokens_seen": 425984000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005995082277283904,
      "loss": 3.1619,
      "theoretical_loss": 3.9838814345531577,
      "tokens_seen": 426049536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005994916777000189,
      "loss": 3.0366,
      "theoretical_loss": 3.9838136062442704,
      "tokens_seen": 426115072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005994751276716475,
      "loss": 2.9686,
      "theoretical_loss": 3.98374579128692,
      "tokens_seen": 426180608
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000599458577643276,
      "loss": 2.8459,
      "theoretical_loss": 3.9836779896764254,
      "tokens_seen": 426246144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005994420276149044,
      "loss": 2.917,
      "theoretical_loss": 3.9836102014081076,
      "tokens_seen": 426311680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000599425477586533,
      "loss": 2.9091,
      "theoretical_loss": 3.983542426477292,
      "tokens_seen": 426377216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005994089275581615,
      "loss": 3.0331,
      "theoretical_loss": 3.9834746648793042,
      "tokens_seen": 426442752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005993923775297901,
      "loss": 2.9891,
      "theoretical_loss": 3.983406916609473,
      "tokens_seen": 426508288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005993758275014186,
      "loss": 2.9391,
      "theoretical_loss": 3.98333918166313,
      "tokens_seen": 426573824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005993592774730471,
      "loss": 2.9862,
      "theoretical_loss": 3.983271460035608,
      "tokens_seen": 426639360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005993427274446757,
      "loss": 3.1689,
      "theoretical_loss": 3.983203751722243,
      "tokens_seen": 426704896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005993261774163042,
      "loss": 2.7477,
      "theoretical_loss": 3.983136056718373,
      "tokens_seen": 426770432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005993096273879326,
      "loss": 3.296,
      "theoretical_loss": 3.9830683750193394,
      "tokens_seen": 426835968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005992930773595612,
      "loss": 3.0071,
      "theoretical_loss": 3.9830007066204836,
      "tokens_seen": 426901504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005992765273311897,
      "loss": 2.8864,
      "theoretical_loss": 3.9829330515171524,
      "tokens_seen": 426967040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005992599773028183,
      "loss": 2.8448,
      "theoretical_loss": 3.982865409704692,
      "tokens_seen": 427032576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005992434272744468,
      "loss": 2.9289,
      "theoretical_loss": 3.9827977811784523,
      "tokens_seen": 427098112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005992268772460753,
      "loss": 2.9241,
      "theoretical_loss": 3.9827301659337864,
      "tokens_seen": 427163648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005992103272177038,
      "loss": 2.5205,
      "theoretical_loss": 3.9826625639660485,
      "tokens_seen": 427229184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005991937771893322,
      "loss": 3.0867,
      "theoretical_loss": 3.982594975270595,
      "tokens_seen": 427294720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005991772271609608,
      "loss": 3.0042,
      "theoretical_loss": 3.982527399842786,
      "tokens_seen": 427360256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005991606771325893,
      "loss": 2.8397,
      "theoretical_loss": 3.9824598376779825,
      "tokens_seen": 427425792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005991441271042179,
      "loss": 2.7715,
      "theoretical_loss": 3.982392288771549,
      "tokens_seen": 427491328
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005991275770758464,
      "loss": 3.2547,
      "theoretical_loss": 3.98232475311885,
      "tokens_seen": 427556864
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005991110270474749,
      "loss": 2.8631,
      "theoretical_loss": 3.9822572307152564,
      "tokens_seen": 427622400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005990944770191035,
      "loss": 2.7508,
      "theoretical_loss": 3.9821897215561375,
      "tokens_seen": 427687936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000599077926990732,
      "loss": 2.9708,
      "theoretical_loss": 3.982122225636867,
      "tokens_seen": 427753472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005990613769623604,
      "loss": 2.8303,
      "theoretical_loss": 3.98205474295282,
      "tokens_seen": 427819008
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000599044826933989,
      "loss": 2.9242,
      "theoretical_loss": 3.9819872734993753,
      "tokens_seen": 427884544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005990282769056175,
      "loss": 2.867,
      "theoretical_loss": 3.9819198172719124,
      "tokens_seen": 427950080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005990117268772461,
      "loss": 2.9029,
      "theoretical_loss": 3.981852374265814,
      "tokens_seen": 428015616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005989951768488746,
      "loss": 2.6671,
      "theoretical_loss": 3.9817849444764644,
      "tokens_seen": 428081152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005989786268205031,
      "loss": 2.9076,
      "theoretical_loss": 3.981717527899252,
      "tokens_seen": 428146688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005989620767921317,
      "loss": 2.591,
      "theoretical_loss": 3.9816501245295646,
      "tokens_seen": 428212224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005989455267637601,
      "loss": 2.5125,
      "theoretical_loss": 3.9815827343627954,
      "tokens_seen": 428277760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005989289767353886,
      "loss": 3.1044,
      "theoretical_loss": 3.981515357394338,
      "tokens_seen": 428343296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005989124267070172,
      "loss": 2.922,
      "theoretical_loss": 3.9814479936195886,
      "tokens_seen": 428408832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005988958766786457,
      "loss": 2.8629,
      "theoretical_loss": 3.981380643033945,
      "tokens_seen": 428474368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005988793266502743,
      "loss": 2.8372,
      "theoretical_loss": 3.9813133056328103,
      "tokens_seen": 428539904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005988627766219028,
      "loss": 2.918,
      "theoretical_loss": 3.981245981411586,
      "tokens_seen": 428605440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005988462265935313,
      "loss": 2.7094,
      "theoretical_loss": 3.981178670365679,
      "tokens_seen": 428670976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005988296765651598,
      "loss": 2.9871,
      "theoretical_loss": 3.981111372490496,
      "tokens_seen": 428736512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005988131265367883,
      "loss": 2.6648,
      "theoretical_loss": 3.9810440877814486,
      "tokens_seen": 428802048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005987965765084168,
      "loss": 2.9006,
      "theoretical_loss": 3.9809768162339476,
      "tokens_seen": 428867584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005987800264800454,
      "loss": 2.9982,
      "theoretical_loss": 3.9809095578434093,
      "tokens_seen": 428933120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005987634764516739,
      "loss": 2.8992,
      "theoretical_loss": 3.9808423126052506,
      "tokens_seen": 428998656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005987469264233025,
      "loss": 2.8837,
      "theoretical_loss": 3.9807750805148903,
      "tokens_seen": 429064192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000598730376394931,
      "loss": 2.7883,
      "theoretical_loss": 3.9807078615677502,
      "tokens_seen": 429129728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005987138263665594,
      "loss": 3.0446,
      "theoretical_loss": 3.980640655759255,
      "tokens_seen": 429195264
    },
    {
      "epoch": 0.15,
      "objective/train/avg_token_score": 0.020728271454572678,
      "objective/train/avg_weight": 0.9756053686141968,
      "objective/train/docs_used": 253081,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6988675594329834,
      "objective/train/std_weight": 0.09366560727357864,
      "objective/train/theoretical_loss": 3.9805734630848306,
      "objective/train/tokens_used": 449720800,
      "theoretical_loss": 3.9805734630848306,
      "tokens_seen": 429260800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000598697276338188,
      "loss": 2.7976,
      "theoretical_loss": 3.9805734630848306,
      "tokens_seen": 429260800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005986807263098165,
      "loss": 3.0865,
      "theoretical_loss": 3.9805062835399054,
      "tokens_seen": 429326336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000598664176281445,
      "loss": 3.0104,
      "theoretical_loss": 3.9804391171199103,
      "tokens_seen": 429391872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005986476262530736,
      "loss": 2.8694,
      "theoretical_loss": 3.9803719638202786,
      "tokens_seen": 429457408
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005986310762247021,
      "loss": 3.0011,
      "theoretical_loss": 3.9803048236364456,
      "tokens_seen": 429522944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005986145261963307,
      "loss": 3.0107,
      "theoretical_loss": 3.9802376965638495,
      "tokens_seen": 429588480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005985979761679592,
      "loss": 2.7347,
      "theoretical_loss": 3.98017058259793,
      "tokens_seen": 429654016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005985814261395876,
      "loss": 3.0011,
      "theoretical_loss": 3.980103481734129,
      "tokens_seen": 429719552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005985648761112162,
      "loss": 2.8838,
      "theoretical_loss": 3.9800363939678918,
      "tokens_seen": 429785088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005985483260828447,
      "loss": 2.7859,
      "theoretical_loss": 3.9799693192946646,
      "tokens_seen": 429850624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005985317760544732,
      "loss": 2.9392,
      "theoretical_loss": 3.979902257709897,
      "tokens_seen": 429916160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005985152260261018,
      "loss": 2.9009,
      "theoretical_loss": 3.97983520920904,
      "tokens_seen": 429981696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005984986759977303,
      "loss": 2.7979,
      "theoretical_loss": 3.979768173787548,
      "tokens_seen": 430047232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005984821259693589,
      "loss": 2.9119,
      "theoretical_loss": 3.9797011514408758,
      "tokens_seen": 430112768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005984655759409873,
      "loss": 2.8733,
      "theoretical_loss": 3.9796341421644827,
      "tokens_seen": 430178304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005984490259126158,
      "loss": 2.8439,
      "theoretical_loss": 3.979567145953829,
      "tokens_seen": 430243840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005984324758842444,
      "loss": 2.6958,
      "theoretical_loss": 3.979500162804377,
      "tokens_seen": 430309376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005984159258558729,
      "loss": 2.9318,
      "theoretical_loss": 3.979433192711592,
      "tokens_seen": 430374912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005983993758275014,
      "loss": 2.828,
      "theoretical_loss": 3.9793662356709407,
      "tokens_seen": 430440448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00059838282579913,
      "loss": 2.9509,
      "theoretical_loss": 3.979299291677894,
      "tokens_seen": 430505984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005983662757707585,
      "loss": 2.6387,
      "theoretical_loss": 3.9792323607279223,
      "tokens_seen": 430571520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005983497257423871,
      "loss": 2.7875,
      "theoretical_loss": 3.979165442816501,
      "tokens_seen": 430637056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005983331757140155,
      "loss": 3.0429,
      "theoretical_loss": 3.9790985379391053,
      "tokens_seen": 430702592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000598316625685644,
      "loss": 3.1599,
      "theoretical_loss": 3.9790316460912143,
      "tokens_seen": 430768128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005983000756572726,
      "loss": 3.0717,
      "theoretical_loss": 3.978964767268309,
      "tokens_seen": 430833664
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005982835256289011,
      "loss": 2.7389,
      "theoretical_loss": 3.9788979014658716,
      "tokens_seen": 430899200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005982669756005296,
      "loss": 2.9193,
      "theoretical_loss": 3.978831048679389,
      "tokens_seen": 430964736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005982504255721582,
      "loss": 2.7545,
      "theoretical_loss": 3.978764208904347,
      "tokens_seen": 431030272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005982338755437865,
      "loss": 2.9502,
      "theoretical_loss": 3.9786973821362377,
      "tokens_seen": 431095808
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005982173255154151,
      "loss": 2.9719,
      "theoretical_loss": 3.9786305683705505,
      "tokens_seen": 431161344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005982007754870436,
      "loss": 2.863,
      "theoretical_loss": 3.978563767602782,
      "tokens_seen": 431226880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005981842254586722,
      "loss": 2.811,
      "theoretical_loss": 3.978496979828428,
      "tokens_seen": 431292416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005981676754303007,
      "loss": 2.7029,
      "theoretical_loss": 3.9784302050429874,
      "tokens_seen": 431357952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005981511254019292,
      "loss": 3.0713,
      "theoretical_loss": 3.978363443241961,
      "tokens_seen": 431423488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005981345753735578,
      "loss": 2.6564,
      "theoretical_loss": 3.9782966944208518,
      "tokens_seen": 431489024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005981180253451863,
      "loss": 2.946,
      "theoretical_loss": 3.9782299585751666,
      "tokens_seen": 431554560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005981014753168147,
      "loss": 2.9428,
      "theoretical_loss": 3.9781632357004124,
      "tokens_seen": 431620096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005980849252884433,
      "loss": 2.8767,
      "theoretical_loss": 3.9780965257920995,
      "tokens_seen": 431685632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005980683752600718,
      "loss": 2.7804,
      "theoretical_loss": 3.9780298288457394,
      "tokens_seen": 431751168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005980518252317004,
      "loss": 2.9149,
      "theoretical_loss": 3.9779631448568473,
      "tokens_seen": 431816704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005980352752033289,
      "loss": 3.0636,
      "theoretical_loss": 3.977896473820941,
      "tokens_seen": 431882240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005980187251749574,
      "loss": 2.8904,
      "theoretical_loss": 3.9778298157335374,
      "tokens_seen": 431947776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000598002175146586,
      "loss": 2.9381,
      "theoretical_loss": 3.977763170590159,
      "tokens_seen": 432013312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005979856251182145,
      "loss": 3.0062,
      "theoretical_loss": 3.9776965383863288,
      "tokens_seen": 432078848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000597969075089843,
      "loss": 2.9734,
      "theoretical_loss": 3.9776299191175726,
      "tokens_seen": 432144384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005979525250614715,
      "loss": 2.8586,
      "theoretical_loss": 3.977563312779418,
      "tokens_seen": 432209920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005979359750331,
      "loss": 2.9421,
      "theoretical_loss": 3.977496719367396,
      "tokens_seen": 432275456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005979194250047286,
      "loss": 2.9256,
      "theoretical_loss": 3.9774301388770374,
      "tokens_seen": 432340992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005979028749763571,
      "loss": 2.7283,
      "theoretical_loss": 3.9773635713038784,
      "tokens_seen": 432406528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005978863249479856,
      "loss": 2.7138,
      "theoretical_loss": 3.9772970166434547,
      "tokens_seen": 432472064
    },
    {
      "epoch": 0.15,
      "objective/train/avg_token_score": 0.01623384840786457,
      "objective/train/avg_weight": 0.9809018969535828,
      "objective/train/docs_used": 255033,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.765838384628296,
      "objective/train/std_weight": 0.07280940562486649,
      "objective/train/theoretical_loss": 3.9772304748913054,
      "objective/train/tokens_used": 452997600,
      "theoretical_loss": 3.9772304748913054,
      "tokens_seen": 432537600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005978697749196142,
      "loss": 2.7686,
      "theoretical_loss": 3.9772304748913054,
      "tokens_seen": 432537600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005978532248912426,
      "loss": 2.9449,
      "theoretical_loss": 3.977163946042973,
      "tokens_seen": 432603136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005978366748628712,
      "loss": 2.799,
      "theoretical_loss": 3.977097430093999,
      "tokens_seen": 432668672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005978201248344997,
      "loss": 2.8278,
      "theoretical_loss": 3.97703092703993,
      "tokens_seen": 432734208
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005978035748061282,
      "loss": 2.973,
      "theoretical_loss": 3.9769644368763135,
      "tokens_seen": 432799744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005977870247777568,
      "loss": 2.9891,
      "theoretical_loss": 3.9768979595987,
      "tokens_seen": 432865280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005977704747493853,
      "loss": 2.9739,
      "theoretical_loss": 3.9768314952026422,
      "tokens_seen": 432930816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005977539247210138,
      "loss": 2.976,
      "theoretical_loss": 3.9767650436836934,
      "tokens_seen": 432996352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005977373746926424,
      "loss": 2.8853,
      "theoretical_loss": 3.976698605037411,
      "tokens_seen": 433061888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005977208246642708,
      "loss": 2.9744,
      "theoretical_loss": 3.976632179259354,
      "tokens_seen": 433127424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005977042746358994,
      "loss": 2.9472,
      "theoretical_loss": 3.976565766345083,
      "tokens_seen": 433192960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005976877246075279,
      "loss": 2.817,
      "theoretical_loss": 3.9764993662901613,
      "tokens_seen": 433258496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005976711745791564,
      "loss": 2.9588,
      "theoretical_loss": 3.976432979090155,
      "tokens_seen": 433324032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000597654624550785,
      "loss": 2.7673,
      "theoretical_loss": 3.9763666047406314,
      "tokens_seen": 433389568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005976380745224135,
      "loss": 2.7027,
      "theoretical_loss": 3.9763002432371604,
      "tokens_seen": 433455104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000597621524494042,
      "loss": 2.8141,
      "theoretical_loss": 3.9762338945753144,
      "tokens_seen": 433520640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005976049744656705,
      "loss": 2.9937,
      "theoretical_loss": 3.9761675587506673,
      "tokens_seen": 433586176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000597588424437299,
      "loss": 3.1433,
      "theoretical_loss": 3.9761012357587955,
      "tokens_seen": 433651712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005975718744089276,
      "loss": 2.7436,
      "theoretical_loss": 3.976034925595278,
      "tokens_seen": 433717248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005975553243805561,
      "loss": 2.8251,
      "theoretical_loss": 3.975968628255696,
      "tokens_seen": 433782784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005975387743521846,
      "loss": 2.8759,
      "theoretical_loss": 3.9759023437356316,
      "tokens_seen": 433848320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005975222243238132,
      "loss": 2.9544,
      "theoretical_loss": 3.9758360720306705,
      "tokens_seen": 433913856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0005975056742954417,
      "loss": 2.8187,
      "theoretical_loss": 3.9757698131364005,
      "tokens_seen": 433979392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005974891242670701,
      "loss": 2.9856,
      "theoretical_loss": 3.9757035670484107,
      "tokens_seen": 434044928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005974725742386987,
      "loss": 2.753,
      "theoretical_loss": 3.9756373337622932,
      "tokens_seen": 434110464
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005974560242103272,
      "loss": 2.9893,
      "theoretical_loss": 3.975571113273642,
      "tokens_seen": 434176000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005974394741819558,
      "loss": 2.8591,
      "theoretical_loss": 3.975504905578053,
      "tokens_seen": 434241536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005974229241535843,
      "loss": 2.8356,
      "theoretical_loss": 3.975438710671125,
      "tokens_seen": 434307072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005974063741252128,
      "loss": 2.8205,
      "theoretical_loss": 3.9753725285484576,
      "tokens_seen": 434372608
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005973898240968414,
      "loss": 2.9093,
      "theoretical_loss": 3.9753063592056552,
      "tokens_seen": 434438144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005973732740684699,
      "loss": 2.9999,
      "theoretical_loss": 3.9752402026383207,
      "tokens_seen": 434503680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005973567240400983,
      "loss": 3.0068,
      "theoretical_loss": 3.9751740588420623,
      "tokens_seen": 434569216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005973401740117269,
      "loss": 2.8022,
      "theoretical_loss": 3.9751079278124886,
      "tokens_seen": 434634752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005973236239833554,
      "loss": 2.9165,
      "theoretical_loss": 3.9750418095452122,
      "tokens_seen": 434700288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000597307073954984,
      "loss": 2.7526,
      "theoretical_loss": 3.9749757040358453,
      "tokens_seen": 434765824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005972905239266125,
      "loss": 2.931,
      "theoretical_loss": 3.9749096112800046,
      "tokens_seen": 434831360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000597273973898241,
      "loss": 2.8398,
      "theoretical_loss": 3.974843531273307,
      "tokens_seen": 434896896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005972574238698695,
      "loss": 2.7739,
      "theoretical_loss": 3.9747774640113738,
      "tokens_seen": 434962432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005972408738414979,
      "loss": 2.9621,
      "theoretical_loss": 3.974711409489826,
      "tokens_seen": 435027968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005972243238131265,
      "loss": 3.052,
      "theoretical_loss": 3.974645367704289,
      "tokens_seen": 435093504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000597207773784755,
      "loss": 2.7843,
      "theoretical_loss": 3.9745793386503885,
      "tokens_seen": 435159040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005971912237563835,
      "loss": 3.0077,
      "theoretical_loss": 3.974513322323754,
      "tokens_seen": 435224576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005971746737280121,
      "loss": 2.6607,
      "theoretical_loss": 3.9744473187200158,
      "tokens_seen": 435290112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005971581236996406,
      "loss": 2.8484,
      "theoretical_loss": 3.974381327834807,
      "tokens_seen": 435355648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005971415736712692,
      "loss": 2.6859,
      "theoretical_loss": 3.974315349663763,
      "tokens_seen": 435421184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005971250236428976,
      "loss": 2.9791,
      "theoretical_loss": 3.9742493842025213,
      "tokens_seen": 435486720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005971084736145261,
      "loss": 2.9377,
      "theoretical_loss": 3.974183431446721,
      "tokens_seen": 435552256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005970919235861547,
      "loss": 2.8383,
      "theoretical_loss": 3.974117491392004,
      "tokens_seen": 435617792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005970753735577832,
      "loss": 2.9146,
      "theoretical_loss": 3.9740515640340135,
      "tokens_seen": 435683328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005970588235294117,
      "loss": 2.7434,
      "theoretical_loss": 3.9739856493683967,
      "tokens_seen": 435748864
    },
    {
      "epoch": 0.16,
      "objective/train/avg_token_score": 0.030142229050397873,
      "objective/train/avg_weight": 0.9645326137542725,
      "objective/train/docs_used": 256972,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7301976680755615,
      "objective/train/std_weight": 0.10640638321638107,
      "objective/train/theoretical_loss": 3.973919747390801,
      "objective/train/tokens_used": 456274400,
      "theoretical_loss": 3.973919747390801,
      "tokens_seen": 435814400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005970422735010403,
      "loss": 2.7515,
      "theoretical_loss": 3.973919747390801,
      "tokens_seen": 435814400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005970257234726688,
      "loss": 3.0389,
      "theoretical_loss": 3.973853858096876,
      "tokens_seen": 435879936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005970091734442974,
      "loss": 2.4266,
      "theoretical_loss": 3.9737879814822747,
      "tokens_seen": 435945472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005969926234159258,
      "loss": 2.953,
      "theoretical_loss": 3.973722117542652,
      "tokens_seen": 436011008
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005969760733875543,
      "loss": 3.1201,
      "theoretical_loss": 3.9736562662736636,
      "tokens_seen": 436076544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005969595233591829,
      "loss": 2.8267,
      "theoretical_loss": 3.9735904276709695,
      "tokens_seen": 436142080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005969429733308114,
      "loss": 2.97,
      "theoretical_loss": 3.9735246017302295,
      "tokens_seen": 436207616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005969264233024399,
      "loss": 2.9539,
      "theoretical_loss": 3.9734587884471075,
      "tokens_seen": 436273152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005969098732740685,
      "loss": 2.7701,
      "theoretical_loss": 3.9733929878172685,
      "tokens_seen": 436338688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000596893323245697,
      "loss": 2.9028,
      "theoretical_loss": 3.9733271998363797,
      "tokens_seen": 436404224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005968767732173255,
      "loss": 2.9409,
      "theoretical_loss": 3.9732614245001106,
      "tokens_seen": 436469760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000596860223188954,
      "loss": 2.8345,
      "theoretical_loss": 3.9731956618041333,
      "tokens_seen": 436535296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005968436731605825,
      "loss": 3.1136,
      "theoretical_loss": 3.9731299117441203,
      "tokens_seen": 436600832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005968271231322111,
      "loss": 3.0388,
      "theoretical_loss": 3.9730641743157493,
      "tokens_seen": 436666368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005968105731038396,
      "loss": 2.8131,
      "theoretical_loss": 3.972998449514697,
      "tokens_seen": 436731904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005967940230754681,
      "loss": 3.101,
      "theoretical_loss": 3.9729327373366443,
      "tokens_seen": 436797440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005967774730470967,
      "loss": 2.8791,
      "theoretical_loss": 3.972867037777273,
      "tokens_seen": 436862976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005967609230187251,
      "loss": 2.9559,
      "theoretical_loss": 3.9728013508322673,
      "tokens_seen": 436928512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005967443729903537,
      "loss": 2.8421,
      "theoretical_loss": 3.9727356764973143,
      "tokens_seen": 436994048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005967278229619822,
      "loss": 2.7694,
      "theoretical_loss": 3.972670014768102,
      "tokens_seen": 437059584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005967112729336107,
      "loss": 2.9202,
      "theoretical_loss": 3.972604365640322,
      "tokens_seen": 437125120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005966947229052393,
      "loss": 3.0522,
      "theoretical_loss": 3.972538729109666,
      "tokens_seen": 437190656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005966781728768678,
      "loss": 2.8729,
      "theoretical_loss": 3.97247310517183,
      "tokens_seen": 437256192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005966616228484963,
      "loss": 2.8464,
      "theoretical_loss": 3.972407493822511,
      "tokens_seen": 437321728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005966450728201249,
      "loss": 3.0395,
      "theoretical_loss": 3.972341895057408,
      "tokens_seen": 437387264
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005966285227917533,
      "loss": 3.0256,
      "theoretical_loss": 3.972276308872222,
      "tokens_seen": 437452800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005966119727633819,
      "loss": 2.7981,
      "theoretical_loss": 3.972210735262657,
      "tokens_seen": 437518336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005965954227350104,
      "loss": 2.6883,
      "theoretical_loss": 3.9721451742244183,
      "tokens_seen": 437583872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005965788727066389,
      "loss": 2.9244,
      "theoretical_loss": 3.9720796257532136,
      "tokens_seen": 437649408
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005965623226782675,
      "loss": 2.8167,
      "theoretical_loss": 3.9720140898447527,
      "tokens_seen": 437714944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000596545772649896,
      "loss": 2.9535,
      "theoretical_loss": 3.9719485664947483,
      "tokens_seen": 437780480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005965292226215245,
      "loss": 3.0767,
      "theoretical_loss": 3.971883055698913,
      "tokens_seen": 437846016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000596512672593153,
      "loss": 2.8826,
      "theoretical_loss": 3.9718175574529635,
      "tokens_seen": 437911552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005964961225647815,
      "loss": 2.8073,
      "theoretical_loss": 3.9717520717526176,
      "tokens_seen": 437977088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005964795725364101,
      "loss": 2.9294,
      "theoretical_loss": 3.971686598593596,
      "tokens_seen": 438042624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005964630225080386,
      "loss": 3.0365,
      "theoretical_loss": 3.971621137971622,
      "tokens_seen": 438108160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005964464724796671,
      "loss": 2.8973,
      "theoretical_loss": 3.9715556898824182,
      "tokens_seen": 438173696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005964299224512957,
      "loss": 2.8181,
      "theoretical_loss": 3.971490254321713,
      "tokens_seen": 438239232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005964133724229242,
      "loss": 2.831,
      "theoretical_loss": 3.9714248312852343,
      "tokens_seen": 438304768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005963968223945528,
      "loss": 2.7864,
      "theoretical_loss": 3.971359420768712,
      "tokens_seen": 438370304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005963802723661812,
      "loss": 2.9482,
      "theoretical_loss": 3.9712940227678803,
      "tokens_seen": 438435840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005963637223378097,
      "loss": 2.8833,
      "theoretical_loss": 3.9712286372784735,
      "tokens_seen": 438501376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005963471723094383,
      "loss": 2.6633,
      "theoretical_loss": 3.971163264296229,
      "tokens_seen": 438566912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005963306222810668,
      "loss": 2.8698,
      "theoretical_loss": 3.971097903816886,
      "tokens_seen": 438632448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005963140722526953,
      "loss": 2.9173,
      "theoretical_loss": 3.9710325558361856,
      "tokens_seen": 438697984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005962975222243239,
      "loss": 2.7923,
      "theoretical_loss": 3.9709672203498707,
      "tokens_seen": 438763520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005962809721959524,
      "loss": 2.6109,
      "theoretical_loss": 3.970901897353688,
      "tokens_seen": 438829056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005962644221675808,
      "loss": 2.8022,
      "theoretical_loss": 3.970836586843383,
      "tokens_seen": 438894592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005962478721392093,
      "loss": 2.9811,
      "theoretical_loss": 3.9707712888147073,
      "tokens_seen": 438960128
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005962313221108379,
      "loss": 2.8281,
      "theoretical_loss": 3.9707060032634107,
      "tokens_seen": 439025664
    },
    {
      "epoch": 0.16,
      "objective/train/avg_token_score": 0.022712869569659233,
      "objective/train/avg_weight": 0.9732867479324341,
      "objective/train/docs_used": 259026,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.206941843032837,
      "objective/train/std_weight": 0.1314157396554947,
      "objective/train/theoretical_loss": 3.9706407301852487,
      "objective/train/tokens_used": 459551200,
      "theoretical_loss": 3.9706407301852487,
      "tokens_seen": 439091200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005962147720824664,
      "loss": 2.9483,
      "theoretical_loss": 3.9706407301852487,
      "tokens_seen": 439091200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005961982220540949,
      "loss": 3.0456,
      "theoretical_loss": 3.9705754695759756,
      "tokens_seen": 439156736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005961816720257235,
      "loss": 2.9079,
      "theoretical_loss": 3.9705102214313506,
      "tokens_seen": 439222272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000596165121997352,
      "loss": 2.9538,
      "theoretical_loss": 3.9704449857471324,
      "tokens_seen": 439287808
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005961485719689804,
      "loss": 2.8903,
      "theoretical_loss": 3.970379762519084,
      "tokens_seen": 439353344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000596132021940609,
      "loss": 2.7144,
      "theoretical_loss": 3.9703145517429697,
      "tokens_seen": 439418880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005961154719122375,
      "loss": 2.8398,
      "theoretical_loss": 3.9702493534145544,
      "tokens_seen": 439484416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005960989218838661,
      "loss": 2.6166,
      "theoretical_loss": 3.9701841675296077,
      "tokens_seen": 439549952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005960823718554946,
      "loss": 3.0249,
      "theoretical_loss": 3.970118994083899,
      "tokens_seen": 439615488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005960658218271231,
      "loss": 2.8073,
      "theoretical_loss": 3.9700538330732016,
      "tokens_seen": 439681024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005960492717987517,
      "loss": 3.0068,
      "theoretical_loss": 3.969988684493289,
      "tokens_seen": 439746560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005960327217703801,
      "loss": 2.7064,
      "theoretical_loss": 3.9699235483399384,
      "tokens_seen": 439812096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005960161717420086,
      "loss": 2.8544,
      "theoretical_loss": 3.969858424608928,
      "tokens_seen": 439877632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005959996217136372,
      "loss": 2.94,
      "theoretical_loss": 3.969793313296039,
      "tokens_seen": 439943168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005959830716852657,
      "loss": 2.7213,
      "theoretical_loss": 3.969728214397053,
      "tokens_seen": 440008704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005959665216568943,
      "loss": 2.8363,
      "theoretical_loss": 3.969663127907756,
      "tokens_seen": 440074240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005959499716285228,
      "loss": 3.0672,
      "theoretical_loss": 3.969598053823934,
      "tokens_seen": 440139776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005959334216001513,
      "loss": 2.6353,
      "theoretical_loss": 3.969532992141377,
      "tokens_seen": 440205312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005959168715717799,
      "loss": 2.8384,
      "theoretical_loss": 3.969467942855875,
      "tokens_seen": 440270848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005959003215434083,
      "loss": 2.8356,
      "theoretical_loss": 3.969402905963221,
      "tokens_seen": 440336384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005958837715150368,
      "loss": 3.0639,
      "theoretical_loss": 3.9693378814592104,
      "tokens_seen": 440401920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005958672214866654,
      "loss": 3.0393,
      "theoretical_loss": 3.96927286933964,
      "tokens_seen": 440467456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005958506714582939,
      "loss": 2.8277,
      "theoretical_loss": 3.969207869600309,
      "tokens_seen": 440532992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005958341214299225,
      "loss": 2.925,
      "theoretical_loss": 3.969142882237019,
      "tokens_seen": 440598528
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000595817571401551,
      "loss": 3.0382,
      "theoretical_loss": 3.9690779072455733,
      "tokens_seen": 440664064
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005958010213731795,
      "loss": 2.9116,
      "theoretical_loss": 3.969012944621776,
      "tokens_seen": 440729600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000595784471344808,
      "loss": 2.795,
      "theoretical_loss": 3.968947994361436,
      "tokens_seen": 440795136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005957679213164365,
      "loss": 2.9502,
      "theoretical_loss": 3.9688830564603625,
      "tokens_seen": 440860672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000595751371288065,
      "loss": 2.8072,
      "theoretical_loss": 3.9688181309143657,
      "tokens_seen": 440926208
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005957348212596936,
      "loss": 2.6244,
      "theoretical_loss": 3.9687532177192604,
      "tokens_seen": 440991744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005957182712313221,
      "loss": 3.032,
      "theoretical_loss": 3.9686883168708613,
      "tokens_seen": 441057280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005957017212029507,
      "loss": 2.9293,
      "theoretical_loss": 3.968623428364986,
      "tokens_seen": 441122816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005956851711745792,
      "loss": 2.7413,
      "theoretical_loss": 3.9685585521974556,
      "tokens_seen": 441188352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005956686211462076,
      "loss": 3.0101,
      "theoretical_loss": 3.968493688364089,
      "tokens_seen": 441253888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005956520711178362,
      "loss": 2.965,
      "theoretical_loss": 3.968428836860712,
      "tokens_seen": 441319424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005956355210894647,
      "loss": 2.68,
      "theoretical_loss": 3.96836399768315,
      "tokens_seen": 441384960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005956189710610932,
      "loss": 2.8551,
      "theoretical_loss": 3.96829917082723,
      "tokens_seen": 441450496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005956024210327218,
      "loss": 2.7868,
      "theoretical_loss": 3.9682343562887823,
      "tokens_seen": 441516032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005955858710043503,
      "loss": 2.787,
      "theoretical_loss": 3.968169554063639,
      "tokens_seen": 441581568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005955693209759789,
      "loss": 2.9267,
      "theoretical_loss": 3.9681047641476326,
      "tokens_seen": 441647104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005955527709476074,
      "loss": 2.7888,
      "theoretical_loss": 3.968039986536601,
      "tokens_seen": 441712640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005955362209192358,
      "loss": 2.7906,
      "theoretical_loss": 3.9679752212263795,
      "tokens_seen": 441778176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005955196708908644,
      "loss": 2.8161,
      "theoretical_loss": 3.9679104682128106,
      "tokens_seen": 441843712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005955031208624929,
      "loss": 3.0235,
      "theoretical_loss": 3.967845727491735,
      "tokens_seen": 441909248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005954865708341214,
      "loss": 2.7906,
      "theoretical_loss": 3.9677809990589967,
      "tokens_seen": 441974784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00059547002080575,
      "loss": 3.0485,
      "theoretical_loss": 3.967716282910441,
      "tokens_seen": 442040320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005954534707773785,
      "loss": 2.8059,
      "theoretical_loss": 3.9676515790419176,
      "tokens_seen": 442105856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005954369207490071,
      "loss": 2.8351,
      "theoretical_loss": 3.967586887449275,
      "tokens_seen": 442171392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005954203707206355,
      "loss": 2.7272,
      "theoretical_loss": 3.967522208128366,
      "tokens_seen": 442236928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000595403820692264,
      "loss": 3.034,
      "theoretical_loss": 3.967457541075044,
      "tokens_seen": 442302464
    },
    {
      "epoch": 0.16,
      "objective/train/avg_token_score": 0.0159809198230505,
      "objective/train/avg_weight": 0.9811975955963135,
      "objective/train/docs_used": 261126,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.955383062362671,
      "objective/train/std_weight": 0.08982670307159424,
      "objective/train/theoretical_loss": 3.9673928862851655,
      "objective/train/tokens_used": 462828000,
      "theoretical_loss": 3.9673928862851655,
      "tokens_seen": 442368000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005953872706638926,
      "loss": 2.9599,
      "theoretical_loss": 3.9673928862851655,
      "tokens_seen": 442368000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005953707206355211,
      "loss": 2.7128,
      "theoretical_loss": 3.9673282437545883,
      "tokens_seen": 442433536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005953541706071496,
      "loss": 2.87,
      "theoretical_loss": 3.9672636134791737,
      "tokens_seen": 442499072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005953376205787782,
      "loss": 2.8224,
      "theoretical_loss": 3.967198995454782,
      "tokens_seen": 442564608
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005953210705504067,
      "loss": 2.88,
      "theoretical_loss": 3.9671343896772786,
      "tokens_seen": 442630144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005953045205220353,
      "loss": 3.001,
      "theoretical_loss": 3.9670697961425287,
      "tokens_seen": 442695680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005952879704936636,
      "loss": 3.1948,
      "theoretical_loss": 3.967005214846401,
      "tokens_seen": 442761216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005952714204652922,
      "loss": 2.6385,
      "theoretical_loss": 3.9669406457847654,
      "tokens_seen": 442826752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005952548704369207,
      "loss": 3.131,
      "theoretical_loss": 3.9668760889534944,
      "tokens_seen": 442892288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005952383204085492,
      "loss": 2.7579,
      "theoretical_loss": 3.9668115443484617,
      "tokens_seen": 442957824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005952217703801778,
      "loss": 2.8468,
      "theoretical_loss": 3.966747011965544,
      "tokens_seen": 443023360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005952052203518063,
      "loss": 2.9812,
      "theoretical_loss": 3.9666824918006185,
      "tokens_seen": 443088896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005951886703234349,
      "loss": 2.8564,
      "theoretical_loss": 3.9666179838495657,
      "tokens_seen": 443154432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005951721202950633,
      "loss": 2.7007,
      "theoretical_loss": 3.9665534881082687,
      "tokens_seen": 443219968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005951555702666918,
      "loss": 2.7779,
      "theoretical_loss": 3.96648900457261,
      "tokens_seen": 443285504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005951390202383204,
      "loss": 2.8957,
      "theoretical_loss": 3.966424533238477,
      "tokens_seen": 443351040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005951224702099489,
      "loss": 2.7122,
      "theoretical_loss": 3.9663600741017575,
      "tokens_seen": 443416576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005951059201815774,
      "loss": 3.063,
      "theoretical_loss": 3.966295627158341,
      "tokens_seen": 443482112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000595089370153206,
      "loss": 2.8865,
      "theoretical_loss": 3.9662311924041203,
      "tokens_seen": 443547648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005950728201248345,
      "loss": 2.7109,
      "theoretical_loss": 3.96616676983499,
      "tokens_seen": 443613184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000595056270096463,
      "loss": 2.9006,
      "theoretical_loss": 3.966102359446845,
      "tokens_seen": 443678720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005950397200680915,
      "loss": 2.6849,
      "theoretical_loss": 3.9660379612355836,
      "tokens_seen": 443744256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00059502317003972,
      "loss": 2.8658,
      "theoretical_loss": 3.9659735751971064,
      "tokens_seen": 443809792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005950066200113486,
      "loss": 2.6275,
      "theoretical_loss": 3.965909201327315,
      "tokens_seen": 443875328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005949900699829771,
      "loss": 2.7268,
      "theoretical_loss": 3.9658448396221138,
      "tokens_seen": 443940864
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005949735199546056,
      "loss": 2.7204,
      "theoretical_loss": 3.9657804900774085,
      "tokens_seen": 444006400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005949569699262342,
      "loss": 2.9495,
      "theoretical_loss": 3.9657161526891076,
      "tokens_seen": 444071936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005949404198978626,
      "loss": 2.9614,
      "theoretical_loss": 3.9656518274531205,
      "tokens_seen": 444137472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005949238698694912,
      "loss": 2.9083,
      "theoretical_loss": 3.9655875143653594,
      "tokens_seen": 444203008
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005949073198411197,
      "loss": 2.7426,
      "theoretical_loss": 3.965523213421738,
      "tokens_seen": 444268544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005948907698127482,
      "loss": 2.6435,
      "theoretical_loss": 3.965458924618173,
      "tokens_seen": 444334080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005948742197843768,
      "loss": 2.8797,
      "theoretical_loss": 3.965394647950581,
      "tokens_seen": 444399616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005948576697560053,
      "loss": 2.849,
      "theoretical_loss": 3.9653303834148828,
      "tokens_seen": 444465152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005948411197276338,
      "loss": 2.9596,
      "theoretical_loss": 3.9652661310070005,
      "tokens_seen": 444530688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005948245696992624,
      "loss": 2.9313,
      "theoretical_loss": 3.965201890722857,
      "tokens_seen": 444596224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005948080196708908,
      "loss": 2.9424,
      "theoretical_loss": 3.9651376625583783,
      "tokens_seen": 444661760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005947914696425194,
      "loss": 2.8132,
      "theoretical_loss": 3.965073446509493,
      "tokens_seen": 444727296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005947749196141479,
      "loss": 2.7148,
      "theoretical_loss": 3.9650092425721297,
      "tokens_seen": 444792832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005947583695857764,
      "loss": 3.1161,
      "theoretical_loss": 3.9649450507422204,
      "tokens_seen": 444858368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000594741819557405,
      "loss": 2.9333,
      "theoretical_loss": 3.9648808710156986,
      "tokens_seen": 444923904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005947252695290335,
      "loss": 2.7705,
      "theoretical_loss": 3.964816703388501,
      "tokens_seen": 444989440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000594708719500662,
      "loss": 2.7809,
      "theoretical_loss": 3.9647525478565644,
      "tokens_seen": 445054976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005946921694722905,
      "loss": 2.8653,
      "theoretical_loss": 3.9646884044158277,
      "tokens_seen": 445120512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000594675619443919,
      "loss": 2.9247,
      "theoretical_loss": 3.964624273062234,
      "tokens_seen": 445186048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005946590694155476,
      "loss": 2.8876,
      "theoretical_loss": 3.9645601537917248,
      "tokens_seen": 445251584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005946425193871761,
      "loss": 2.8615,
      "theoretical_loss": 3.964496046600247,
      "tokens_seen": 445317120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005946259693588046,
      "loss": 3.0311,
      "theoretical_loss": 3.9644319514837467,
      "tokens_seen": 445382656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005946094193304332,
      "loss": 3.0382,
      "theoretical_loss": 3.964367868438175,
      "tokens_seen": 445448192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005945928693020617,
      "loss": 2.9562,
      "theoretical_loss": 3.9643037974594817,
      "tokens_seen": 445513728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005945763192736901,
      "loss": 2.7153,
      "theoretical_loss": 3.96423973854362,
      "tokens_seen": 445579264
    },
    {
      "epoch": 0.16,
      "objective/train/avg_token_score": 0.01842174492776394,
      "objective/train/avg_weight": 0.9783101677894592,
      "objective/train/docs_used": 262792,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.111844301223755,
      "objective/train/std_weight": 0.11798816919326782,
      "objective/train/theoretical_loss": 3.9641756916865463,
      "objective/train/tokens_used": 466104800,
      "theoretical_loss": 3.9641756916865463,
      "tokens_seen": 445644800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005945597692453187,
      "loss": 2.71,
      "theoretical_loss": 3.9641756916865463,
      "tokens_seen": 445644800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005945432192169472,
      "loss": 3.0071,
      "theoretical_loss": 3.964111656884217,
      "tokens_seen": 445710336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005945266691885758,
      "loss": 2.8286,
      "theoretical_loss": 3.9640476341325916,
      "tokens_seen": 445775872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005945101191602043,
      "loss": 2.8786,
      "theoretical_loss": 3.96398362342763,
      "tokens_seen": 445841408
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005944935691318328,
      "loss": 2.8182,
      "theoretical_loss": 3.9639196247652966,
      "tokens_seen": 445906944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005944770191034614,
      "loss": 3.0645,
      "theoretical_loss": 3.9638556381415553,
      "tokens_seen": 445972480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005944604690750899,
      "loss": 2.9063,
      "theoretical_loss": 3.9637916635523736,
      "tokens_seen": 446038016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005944439190467183,
      "loss": 2.6978,
      "theoretical_loss": 3.9637277009937195,
      "tokens_seen": 446103552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005944273690183469,
      "loss": 3.0494,
      "theoretical_loss": 3.9636637504615653,
      "tokens_seen": 446169088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005944108189899754,
      "loss": 2.8455,
      "theoretical_loss": 3.963599811951882,
      "tokens_seen": 446234624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000594394268961604,
      "loss": 2.8168,
      "theoretical_loss": 3.9635358854606455,
      "tokens_seen": 446300160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005943777189332325,
      "loss": 2.8378,
      "theoretical_loss": 3.9634719709838317,
      "tokens_seen": 446365696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000594361168904861,
      "loss": 2.9558,
      "theoretical_loss": 3.963408068517419,
      "tokens_seen": 446431232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005943446188764896,
      "loss": 3.1906,
      "theoretical_loss": 3.9633441780573886,
      "tokens_seen": 446496768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000594328068848118,
      "loss": 2.8481,
      "theoretical_loss": 3.963280299599722,
      "tokens_seen": 446562304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005943115188197465,
      "loss": 2.6798,
      "theoretical_loss": 3.9632164331404045,
      "tokens_seen": 446627840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000594294968791375,
      "loss": 2.6403,
      "theoretical_loss": 3.9631525786754214,
      "tokens_seen": 446693376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005942784187630035,
      "loss": 3.0439,
      "theoretical_loss": 3.9630887362007616,
      "tokens_seen": 446758912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005942618687346321,
      "loss": 2.7697,
      "theoretical_loss": 3.9630249057124147,
      "tokens_seen": 446824448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005942453187062606,
      "loss": 2.9471,
      "theoretical_loss": 3.9629610872063727,
      "tokens_seen": 446889984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005942287686778892,
      "loss": 2.9436,
      "theoretical_loss": 3.9628972806786305,
      "tokens_seen": 446955520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005942122186495176,
      "loss": 2.7669,
      "theoretical_loss": 3.9628334861251826,
      "tokens_seen": 447021056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005941956686211461,
      "loss": 2.8177,
      "theoretical_loss": 3.9627697035420284,
      "tokens_seen": 447086592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005941791185927747,
      "loss": 3.0322,
      "theoretical_loss": 3.9627059329251666,
      "tokens_seen": 447152128
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005941625685644032,
      "loss": 2.7658,
      "theoretical_loss": 3.9626421742705986,
      "tokens_seen": 447217664
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005941460185360317,
      "loss": 2.9635,
      "theoretical_loss": 3.9625784275743294,
      "tokens_seen": 447283200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005941294685076603,
      "loss": 2.8122,
      "theoretical_loss": 3.9625146928323627,
      "tokens_seen": 447348736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005941129184792888,
      "loss": 2.8115,
      "theoretical_loss": 3.962450970040707,
      "tokens_seen": 447414272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005940963684509174,
      "loss": 2.8728,
      "theoretical_loss": 3.9623872591953715,
      "tokens_seen": 447479808
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005940798184225458,
      "loss": 2.919,
      "theoretical_loss": 3.9623235602923677,
      "tokens_seen": 447545344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005940632683941743,
      "loss": 2.9188,
      "theoretical_loss": 3.9622598733277083,
      "tokens_seen": 447610880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005940467183658029,
      "loss": 2.8738,
      "theoretical_loss": 3.9621961982974088,
      "tokens_seen": 447676416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005940301683374314,
      "loss": 2.9639,
      "theoretical_loss": 3.962132535197486,
      "tokens_seen": 447741952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005940136183090599,
      "loss": 3.0726,
      "theoretical_loss": 3.9620688840239593,
      "tokens_seen": 447807488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005939970682806885,
      "loss": 3.0395,
      "theoretical_loss": 3.9620052447728478,
      "tokens_seen": 447873024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000593980518252317,
      "loss": 2.9369,
      "theoretical_loss": 3.9619416174401767,
      "tokens_seen": 447938560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005939639682239455,
      "loss": 2.6279,
      "theoretical_loss": 3.9618780020219693,
      "tokens_seen": 448004096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000593947418195574,
      "loss": 2.6823,
      "theoretical_loss": 3.961814398514252,
      "tokens_seen": 448069632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005939308681672025,
      "loss": 2.9069,
      "theoretical_loss": 3.961750806913054,
      "tokens_seen": 448135168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005939143181388311,
      "loss": 3.0,
      "theoretical_loss": 3.961687227214406,
      "tokens_seen": 448200704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005938977681104596,
      "loss": 2.9655,
      "theoretical_loss": 3.961623659414338,
      "tokens_seen": 448266240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005938812180820882,
      "loss": 2.9965,
      "theoretical_loss": 3.961560103508887,
      "tokens_seen": 448331776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005938646680537167,
      "loss": 2.7414,
      "theoretical_loss": 3.961496559494087,
      "tokens_seen": 448397312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005938481180253451,
      "loss": 2.6633,
      "theoretical_loss": 3.9614330273659775,
      "tokens_seen": 448462848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005938315679969737,
      "loss": 2.8958,
      "theoretical_loss": 3.9613695071205974,
      "tokens_seen": 448528384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005938150179686022,
      "loss": 2.8716,
      "theoretical_loss": 3.9613059987539887,
      "tokens_seen": 448593920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005937984679402307,
      "loss": 2.7208,
      "theoretical_loss": 3.9612425022621953,
      "tokens_seen": 448659456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005937819179118593,
      "loss": 2.5546,
      "theoretical_loss": 3.9611790176412622,
      "tokens_seen": 448724992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005937653678834878,
      "loss": 2.8941,
      "theoretical_loss": 3.961115544887238,
      "tokens_seen": 448790528
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005937488178551164,
      "loss": 2.9305,
      "theoretical_loss": 3.961052083996171,
      "tokens_seen": 448856064
    },
    {
      "epoch": 0.16,
      "objective/train/avg_token_score": 0.018095117062330246,
      "objective/train/avg_weight": 0.9787071347236633,
      "objective/train/docs_used": 264102,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.862161159515381,
      "objective/train/std_weight": 0.1140826940536499,
      "objective/train/theoretical_loss": 3.960988634964113,
      "objective/train/tokens_used": 469381600,
      "theoretical_loss": 3.960988634964113,
      "tokens_seen": 448921600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005937322678267449,
      "loss": 2.7882,
      "theoretical_loss": 3.960988634964113,
      "tokens_seen": 448921600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005937157177983733,
      "loss": 2.7615,
      "theoretical_loss": 3.960925197787116,
      "tokens_seen": 448987136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005936991677700019,
      "loss": 2.931,
      "theoretical_loss": 3.9608617724612367,
      "tokens_seen": 449052672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005936826177416304,
      "loss": 2.8264,
      "theoretical_loss": 3.960798358982531,
      "tokens_seen": 449118208
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005936660677132589,
      "loss": 2.7277,
      "theoretical_loss": 3.960734957347058,
      "tokens_seen": 449183744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005936495176848875,
      "loss": 2.8335,
      "theoretical_loss": 3.960671567550878,
      "tokens_seen": 449249280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000593632967656516,
      "loss": 2.9898,
      "theoretical_loss": 3.9606081895900545,
      "tokens_seen": 449314816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005936164176281446,
      "loss": 3.0274,
      "theoretical_loss": 3.9605448234606513,
      "tokens_seen": 449380352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000593599867599773,
      "loss": 2.9395,
      "theoretical_loss": 3.9604814691587347,
      "tokens_seen": 449445888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005935833175714015,
      "loss": 2.9249,
      "theoretical_loss": 3.9604181266803726,
      "tokens_seen": 449511424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005935667675430301,
      "loss": 2.8832,
      "theoretical_loss": 3.960354796021636,
      "tokens_seen": 449576960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005935502175146586,
      "loss": 3.0239,
      "theoretical_loss": 3.960291477178596,
      "tokens_seen": 449642496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005935336674862871,
      "loss": 2.747,
      "theoretical_loss": 3.960228170147327,
      "tokens_seen": 449708032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005935171174579157,
      "loss": 2.6281,
      "theoretical_loss": 3.9601648749239047,
      "tokens_seen": 449773568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005935005674295442,
      "loss": 2.9212,
      "theoretical_loss": 3.960101591504406,
      "tokens_seen": 449839104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005934840174011728,
      "loss": 2.9671,
      "theoretical_loss": 3.960038319884912,
      "tokens_seen": 449904640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005934674673728012,
      "loss": 3.0116,
      "theoretical_loss": 3.9599750600615025,
      "tokens_seen": 449970176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005934509173444297,
      "loss": 2.9404,
      "theoretical_loss": 3.9599118120302608,
      "tokens_seen": 450035712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005934343673160583,
      "loss": 2.6917,
      "theoretical_loss": 3.959848575787273,
      "tokens_seen": 450101248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005934178172876868,
      "loss": 2.8644,
      "theoretical_loss": 3.959785351328626,
      "tokens_seen": 450166784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005934012672593153,
      "loss": 2.8684,
      "theoretical_loss": 3.9597221386504073,
      "tokens_seen": 450232320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005933847172309439,
      "loss": 2.7379,
      "theoretical_loss": 3.9596589377487086,
      "tokens_seen": 450297856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005933681672025724,
      "loss": 2.7614,
      "theoretical_loss": 3.9595957486196225,
      "tokens_seen": 450363392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000593351617174201,
      "loss": 2.6893,
      "theoretical_loss": 3.9595325712592433,
      "tokens_seen": 450428928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005933350671458294,
      "loss": 3.0287,
      "theoretical_loss": 3.9594694056636675,
      "tokens_seen": 450494464
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005933185171174579,
      "loss": 3.034,
      "theoretical_loss": 3.9594062518289923,
      "tokens_seen": 450560000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005933019670890864,
      "loss": 2.8586,
      "theoretical_loss": 3.9593431097513188,
      "tokens_seen": 450625536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005932854170607149,
      "loss": 2.8096,
      "theoretical_loss": 3.9592799794267486,
      "tokens_seen": 450691072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005932688670323435,
      "loss": 2.7253,
      "theoretical_loss": 3.9592168608513854,
      "tokens_seen": 450756608
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000593252317003972,
      "loss": 2.7961,
      "theoretical_loss": 3.959153754021334,
      "tokens_seen": 450822144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005932357669756004,
      "loss": 2.8781,
      "theoretical_loss": 3.9590906589327037,
      "tokens_seen": 450887680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000593219216947229,
      "loss": 2.7916,
      "theoretical_loss": 3.9590275755816022,
      "tokens_seen": 450953216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005932026669188575,
      "loss": 2.8924,
      "theoretical_loss": 3.958964503964141,
      "tokens_seen": 451018752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005931861168904861,
      "loss": 2.5351,
      "theoretical_loss": 3.9589014440764334,
      "tokens_seen": 451084288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005931695668621146,
      "loss": 2.8578,
      "theoretical_loss": 3.9588383959145945,
      "tokens_seen": 451149824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005931530168337431,
      "loss": 2.7553,
      "theoretical_loss": 3.9587753594747404,
      "tokens_seen": 451215360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005931364668053717,
      "loss": 2.8722,
      "theoretical_loss": 3.95871233475299,
      "tokens_seen": 451280896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005931199167770002,
      "loss": 2.7568,
      "theoretical_loss": 3.9586493217454635,
      "tokens_seen": 451346432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005931033667486286,
      "loss": 2.686,
      "theoretical_loss": 3.9585863204482834,
      "tokens_seen": 451411968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005930868167202572,
      "loss": 2.8762,
      "theoretical_loss": 3.958523330857574,
      "tokens_seen": 451477504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005930702666918857,
      "loss": 2.833,
      "theoretical_loss": 3.9584603529694604,
      "tokens_seen": 451543040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005930537166635143,
      "loss": 2.734,
      "theoretical_loss": 3.9583973867800717,
      "tokens_seen": 451608576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005930371666351428,
      "loss": 2.8574,
      "theoretical_loss": 3.958334432285537,
      "tokens_seen": 451674112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005930206166067713,
      "loss": 2.6669,
      "theoretical_loss": 3.9582714894819873,
      "tokens_seen": 451739648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005930040665783999,
      "loss": 2.7463,
      "theoretical_loss": 3.9582085583655564,
      "tokens_seen": 451805184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005929875165500283,
      "loss": 2.9005,
      "theoretical_loss": 3.958145638932379,
      "tokens_seen": 451870720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005929709665216568,
      "loss": 2.7702,
      "theoretical_loss": 3.9580827311785924,
      "tokens_seen": 451936256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005929544164932854,
      "loss": 2.9149,
      "theoretical_loss": 3.9580198351003357,
      "tokens_seen": 452001792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005929378664649139,
      "loss": 3.0265,
      "theoretical_loss": 3.9579569506937498,
      "tokens_seen": 452067328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005929213164365425,
      "loss": 2.714,
      "theoretical_loss": 3.9578940779549763,
      "tokens_seen": 452132864
    },
    {
      "epoch": 0.16,
      "objective/train/avg_token_score": 0.00964416190981865,
      "objective/train/avg_weight": 0.9886701107025146,
      "objective/train/docs_used": 265980,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.065774440765381,
      "objective/train/std_weight": 0.06799417734146118,
      "objective/train/theoretical_loss": 3.9578312168801597,
      "objective/train/tokens_used": 472658400,
      "theoretical_loss": 3.9578312168801597,
      "tokens_seen": 452198400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000592904766408171,
      "loss": 3.0046,
      "theoretical_loss": 3.9578312168801597,
      "tokens_seen": 452198400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005928882163797995,
      "loss": 3.0556,
      "theoretical_loss": 3.957768367465447,
      "tokens_seen": 452263936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000592871666351428,
      "loss": 2.8572,
      "theoretical_loss": 3.9577055297069856,
      "tokens_seen": 452329472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005928551163230565,
      "loss": 2.7614,
      "theoretical_loss": 3.957642703600926,
      "tokens_seen": 452395008
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000592838566294685,
      "loss": 2.6831,
      "theoretical_loss": 3.9575798891434184,
      "tokens_seen": 452460544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005928220162663136,
      "loss": 2.7821,
      "theoretical_loss": 3.957517086330618,
      "tokens_seen": 452526080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005928054662379421,
      "loss": 2.8772,
      "theoretical_loss": 3.957454295158679,
      "tokens_seen": 452591616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005927889162095707,
      "loss": 2.8846,
      "theoretical_loss": 3.9573915156237587,
      "tokens_seen": 452657152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005927723661811992,
      "loss": 2.6451,
      "theoretical_loss": 3.9573287477220167,
      "tokens_seen": 452722688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005927558161528277,
      "loss": 2.598,
      "theoretical_loss": 3.957265991449613,
      "tokens_seen": 452788224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005927392661244562,
      "loss": 2.7434,
      "theoretical_loss": 3.957203246802711,
      "tokens_seen": 452853760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005927227160960847,
      "loss": 2.9557,
      "theoretical_loss": 3.9571405137774747,
      "tokens_seen": 452919296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005927061660677132,
      "loss": 2.6772,
      "theoretical_loss": 3.9570777923700704,
      "tokens_seen": 452984832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005926896160393418,
      "loss": 2.8539,
      "theoretical_loss": 3.9570150825766657,
      "tokens_seen": 453050368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005926730660109703,
      "loss": 3.0365,
      "theoretical_loss": 3.956952384393432,
      "tokens_seen": 453115904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005926565159825989,
      "loss": 2.762,
      "theoretical_loss": 3.956889697816539,
      "tokens_seen": 453181440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005926399659542274,
      "loss": 2.96,
      "theoretical_loss": 3.9568270228421616,
      "tokens_seen": 453246976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005926234159258558,
      "loss": 3.0041,
      "theoretical_loss": 3.956764359466475,
      "tokens_seen": 453312512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005926068658974844,
      "loss": 2.9862,
      "theoretical_loss": 3.9567017076856565,
      "tokens_seen": 453378048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005925903158691129,
      "loss": 3.0197,
      "theoretical_loss": 3.9566390674958845,
      "tokens_seen": 453443584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005925737658407414,
      "loss": 2.9882,
      "theoretical_loss": 3.9565764388933395,
      "tokens_seen": 453509120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00059255721581237,
      "loss": 2.835,
      "theoretical_loss": 3.9565138218742053,
      "tokens_seen": 453574656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005925406657839985,
      "loss": 2.9749,
      "theoretical_loss": 3.9564512164346652,
      "tokens_seen": 453640192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005925241157556271,
      "loss": 2.8877,
      "theoretical_loss": 3.956388622570906,
      "tokens_seen": 453705728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005925075657272556,
      "loss": 2.8342,
      "theoretical_loss": 3.956326040279116,
      "tokens_seen": 453771264
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000592491015698884,
      "loss": 2.7916,
      "theoretical_loss": 3.956263469555484,
      "tokens_seen": 453836800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005924744656705126,
      "loss": 2.7918,
      "theoretical_loss": 3.956200910396202,
      "tokens_seen": 453902336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005924579156421411,
      "loss": 2.7591,
      "theoretical_loss": 3.9561383627974642,
      "tokens_seen": 453967872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005924413656137697,
      "loss": 3.0838,
      "theoretical_loss": 3.956075826755465,
      "tokens_seen": 454033408
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005924248155853982,
      "loss": 2.6186,
      "theoretical_loss": 3.9560133022664026,
      "tokens_seen": 454098944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005924082655570267,
      "loss": 2.6742,
      "theoretical_loss": 3.955950789326474,
      "tokens_seen": 454164480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005923917155286553,
      "loss": 2.6505,
      "theoretical_loss": 3.955888287931881,
      "tokens_seen": 454230016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005923751655002837,
      "loss": 2.7772,
      "theoretical_loss": 3.955825798078826,
      "tokens_seen": 454295552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005923586154719122,
      "loss": 2.8163,
      "theoretical_loss": 3.9557633197635127,
      "tokens_seen": 454361088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005923420654435407,
      "loss": 2.8269,
      "theoretical_loss": 3.9557008529821474,
      "tokens_seen": 454426624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005923255154151692,
      "loss": 2.6376,
      "theoretical_loss": 3.955638397730939,
      "tokens_seen": 454492160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005923089653867978,
      "loss": 2.8799,
      "theoretical_loss": 3.955575954006095,
      "tokens_seen": 454557696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005922924153584263,
      "loss": 3.0885,
      "theoretical_loss": 3.9555135218038284,
      "tokens_seen": 454623232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005922758653300549,
      "loss": 2.9899,
      "theoretical_loss": 3.955451101120352,
      "tokens_seen": 454688768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005922593153016833,
      "loss": 2.8704,
      "theoretical_loss": 3.955388691951881,
      "tokens_seen": 454754304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005922427652733118,
      "loss": 2.8133,
      "theoretical_loss": 3.9553262942946312,
      "tokens_seen": 454819840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005922262152449404,
      "loss": 2.8123,
      "theoretical_loss": 3.9552639081448224,
      "tokens_seen": 454885376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005922096652165689,
      "loss": 2.861,
      "theoretical_loss": 3.9552015334986743,
      "tokens_seen": 454950912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005921931151881974,
      "loss": 2.5776,
      "theoretical_loss": 3.9551391703524086,
      "tokens_seen": 455016448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000592176565159826,
      "loss": 2.6713,
      "theoretical_loss": 3.95507681870225,
      "tokens_seen": 455081984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005921600151314545,
      "loss": 2.7942,
      "theoretical_loss": 3.955014478544425,
      "tokens_seen": 455147520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000592143465103083,
      "loss": 2.8517,
      "theoretical_loss": 3.954952149875159,
      "tokens_seen": 455213056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005921269150747115,
      "loss": 3.1382,
      "theoretical_loss": 3.9548898326906823,
      "tokens_seen": 455278592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00059211036504634,
      "loss": 2.8537,
      "theoretical_loss": 3.954827526987226,
      "tokens_seen": 455344128
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005920938150179686,
      "loss": 2.8854,
      "theoretical_loss": 3.954765232761023,
      "tokens_seen": 455409664
    },
    {
      "epoch": 0.16,
      "objective/train/avg_token_score": 0.024321366101503372,
      "objective/train/avg_weight": 0.9713907837867737,
      "objective/train/docs_used": 267725,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.015338182449341,
      "objective/train/std_weight": 0.12523865699768066,
      "objective/train/theoretical_loss": 3.954702950008308,
      "objective/train/tokens_used": 475935200,
      "theoretical_loss": 3.954702950008308,
      "tokens_seen": 455475200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005920772649895971,
      "loss": 2.8071,
      "theoretical_loss": 3.954702950008308,
      "tokens_seen": 455475200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005920607149612256,
      "loss": 2.7536,
      "theoretical_loss": 3.954640678725317,
      "tokens_seen": 455540736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005920441649328542,
      "loss": 2.813,
      "theoretical_loss": 3.9545784189082887,
      "tokens_seen": 455606272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005920276149044827,
      "loss": 2.6639,
      "theoretical_loss": 3.9545161705534624,
      "tokens_seen": 455671808
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005920110648761112,
      "loss": 2.8372,
      "theoretical_loss": 3.9544539336570796,
      "tokens_seen": 455737344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005919945148477397,
      "loss": 2.879,
      "theoretical_loss": 3.954391708215385,
      "tokens_seen": 455802880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005919779648193682,
      "loss": 2.7569,
      "theoretical_loss": 3.954329494224623,
      "tokens_seen": 455868416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005919614147909968,
      "loss": 2.8359,
      "theoretical_loss": 3.9542672916810404,
      "tokens_seen": 455933952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005919448647626253,
      "loss": 2.9365,
      "theoretical_loss": 3.9542051005808867,
      "tokens_seen": 455999488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005919283147342538,
      "loss": 2.6771,
      "theoretical_loss": 3.954142920920412,
      "tokens_seen": 456065024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005919117647058824,
      "loss": 2.5938,
      "theoretical_loss": 3.954080752695868,
      "tokens_seen": 456130560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005918952146775108,
      "loss": 2.9532,
      "theoretical_loss": 3.95401859590351,
      "tokens_seen": 456196096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005918786646491394,
      "loss": 2.7489,
      "theoretical_loss": 3.953956450539594,
      "tokens_seen": 456261632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005918621146207679,
      "loss": 2.7361,
      "theoretical_loss": 3.953894316600376,
      "tokens_seen": 456327168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005918455645923964,
      "loss": 3.0372,
      "theoretical_loss": 3.953832194082117,
      "tokens_seen": 456392704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000591829014564025,
      "loss": 2.9546,
      "theoretical_loss": 3.9537700829810767,
      "tokens_seen": 456458240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005918124645356535,
      "loss": 2.665,
      "theoretical_loss": 3.95370798329352,
      "tokens_seen": 456523776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000591795914507282,
      "loss": 2.8212,
      "theoretical_loss": 3.9536458950157094,
      "tokens_seen": 456589312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005917793644789106,
      "loss": 2.9279,
      "theoretical_loss": 3.9535838181439127,
      "tokens_seen": 456654848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000591762814450539,
      "loss": 2.978,
      "theoretical_loss": 3.9535217526743978,
      "tokens_seen": 456720384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005917462644221676,
      "loss": 2.8663,
      "theoretical_loss": 3.953459698603434,
      "tokens_seen": 456785920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005917297143937961,
      "loss": 2.813,
      "theoretical_loss": 3.9533976559272936,
      "tokens_seen": 456851456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005917131643654246,
      "loss": 3.0083,
      "theoretical_loss": 3.95333562464225,
      "tokens_seen": 456916992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005916966143370532,
      "loss": 2.8688,
      "theoretical_loss": 3.9532736047445787,
      "tokens_seen": 456982528
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005916800643086817,
      "loss": 2.7333,
      "theoretical_loss": 3.953211596230556,
      "tokens_seen": 457048064
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005916635142803102,
      "loss": 2.9331,
      "theoretical_loss": 3.953149599096461,
      "tokens_seen": 457113600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005916469642519387,
      "loss": 2.938,
      "theoretical_loss": 3.9530876133385737,
      "tokens_seen": 457179136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005916304142235672,
      "loss": 2.8987,
      "theoretical_loss": 3.953025638953177,
      "tokens_seen": 457244672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005916138641951958,
      "loss": 2.8423,
      "theoretical_loss": 3.9529636759365543,
      "tokens_seen": 457310208
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005915973141668243,
      "loss": 2.9171,
      "theoretical_loss": 3.9529017242849918,
      "tokens_seen": 457375744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005915807641384528,
      "loss": 2.9134,
      "theoretical_loss": 3.952839783994776,
      "tokens_seen": 457441280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005915642141100814,
      "loss": 2.8608,
      "theoretical_loss": 3.9527778550621973,
      "tokens_seen": 457506816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005915476640817099,
      "loss": 2.7593,
      "theoretical_loss": 3.9527159374835463,
      "tokens_seen": 457572352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005915311140533383,
      "loss": 2.7284,
      "theoretical_loss": 3.952654031255115,
      "tokens_seen": 457637888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005915145640249669,
      "loss": 2.8356,
      "theoretical_loss": 3.9525921363731977,
      "tokens_seen": 457703424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005914980139965954,
      "loss": 2.6958,
      "theoretical_loss": 3.9525302528340918,
      "tokens_seen": 457768960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000591481463968224,
      "loss": 2.8301,
      "theoretical_loss": 3.9524683806340937,
      "tokens_seen": 457834496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005914649139398525,
      "loss": 2.6371,
      "theoretical_loss": 3.9524065197695046,
      "tokens_seen": 457900032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000591448363911481,
      "loss": 2.7942,
      "theoretical_loss": 3.9523446702366245,
      "tokens_seen": 457965568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005914318138831096,
      "loss": 3.0647,
      "theoretical_loss": 3.9522828320317576,
      "tokens_seen": 458031104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005914152638547381,
      "loss": 2.689,
      "theoretical_loss": 3.9522210051512077,
      "tokens_seen": 458096640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005913987138263665,
      "loss": 2.9516,
      "theoretical_loss": 3.9521591895912818,
      "tokens_seen": 458162176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005913821637979951,
      "loss": 2.7936,
      "theoretical_loss": 3.9520973853482886,
      "tokens_seen": 458227712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005913656137696236,
      "loss": 3.0209,
      "theoretical_loss": 3.9520355924185377,
      "tokens_seen": 458293248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005913490637412521,
      "loss": 2.6882,
      "theoretical_loss": 3.951973810798341,
      "tokens_seen": 458358784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005913325137128806,
      "loss": 2.7228,
      "theoretical_loss": 3.951912040484012,
      "tokens_seen": 458424320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005913159636845092,
      "loss": 2.6692,
      "theoretical_loss": 3.9518502814718652,
      "tokens_seen": 458489856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005912994136561377,
      "loss": 2.7163,
      "theoretical_loss": 3.9517885337582186,
      "tokens_seen": 458555392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005912828636277661,
      "loss": 2.8611,
      "theoretical_loss": 3.951726797339391,
      "tokens_seen": 458620928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005912663135993947,
      "loss": 2.8354,
      "theoretical_loss": 3.9516650722117017,
      "tokens_seen": 458686464
    },
    {
      "epoch": 0.16,
      "objective/train/avg_token_score": 0.026224780827760696,
      "objective/train/avg_weight": 0.9691897034645081,
      "objective/train/docs_used": 269663,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1817445755004883,
      "objective/train/std_weight": 0.12314725667238235,
      "objective/train/theoretical_loss": 3.9516033583714734,
      "objective/train/tokens_used": 479212000,
      "theoretical_loss": 3.9516033583714734,
      "tokens_seen": 458752000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005912497635710232,
      "loss": 2.9552,
      "theoretical_loss": 3.9516033583714734,
      "tokens_seen": 458752000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005912332135426518,
      "loss": 2.6895,
      "theoretical_loss": 3.9515416558150305,
      "tokens_seen": 458817536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005912166635142803,
      "loss": 2.5203,
      "theoretical_loss": 3.9514799645386973,
      "tokens_seen": 458883072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005912001134859088,
      "loss": 2.7762,
      "theoretical_loss": 3.9514182845388026,
      "tokens_seen": 458948608
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005911835634575374,
      "loss": 2.698,
      "theoretical_loss": 3.951356615811674,
      "tokens_seen": 459014144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005911670134291658,
      "loss": 2.746,
      "theoretical_loss": 3.9512949583536434,
      "tokens_seen": 459079680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005911504634007943,
      "loss": 3.0089,
      "theoretical_loss": 3.951233312161042,
      "tokens_seen": 459145216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005911339133724229,
      "loss": 2.8486,
      "theoretical_loss": 3.9511716772302057,
      "tokens_seen": 459210752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005911173633440514,
      "loss": 2.9939,
      "theoretical_loss": 3.9511100535574686,
      "tokens_seen": 459276288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00059110081331568,
      "loss": 2.9078,
      "theoretical_loss": 3.9510484411391698,
      "tokens_seen": 459341824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005910842632873085,
      "loss": 2.8707,
      "theoretical_loss": 3.9509868399716472,
      "tokens_seen": 459407360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000591067713258937,
      "loss": 2.778,
      "theoretical_loss": 3.9509252500512426,
      "tokens_seen": 459472896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005910511632305656,
      "loss": 2.9129,
      "theoretical_loss": 3.950863671374299,
      "tokens_seen": 459538432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000591034613202194,
      "loss": 2.8598,
      "theoretical_loss": 3.9508021039371597,
      "tokens_seen": 459603968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005910180631738225,
      "loss": 2.9421,
      "theoretical_loss": 3.9507405477361726,
      "tokens_seen": 459669504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005910015131454511,
      "loss": 2.8657,
      "theoretical_loss": 3.9506790027676844,
      "tokens_seen": 459735040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005909849631170796,
      "loss": 2.865,
      "theoretical_loss": 3.9506174690280442,
      "tokens_seen": 459800576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005909684130887082,
      "loss": 2.7729,
      "theoretical_loss": 3.950555946513605,
      "tokens_seen": 459866112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005909518630603367,
      "loss": 2.8573,
      "theoretical_loss": 3.9504944352207176,
      "tokens_seen": 459931648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005909353130319652,
      "loss": 3.2062,
      "theoretical_loss": 3.9504329351457383,
      "tokens_seen": 459997184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005909187630035937,
      "loss": 2.7795,
      "theoretical_loss": 3.9503714462850237,
      "tokens_seen": 460062720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005909022129752222,
      "loss": 3.0428,
      "theoretical_loss": 3.9503099686349303,
      "tokens_seen": 460128256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005908856629468507,
      "loss": 2.9199,
      "theoretical_loss": 3.950248502191819,
      "tokens_seen": 460193792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005908691129184793,
      "loss": 2.7935,
      "theoretical_loss": 3.950187046952051,
      "tokens_seen": 460259328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005908525628901078,
      "loss": 2.7383,
      "theoretical_loss": 3.9501256029119896,
      "tokens_seen": 460324864
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005908360128617364,
      "loss": 2.81,
      "theoretical_loss": 3.950064170068,
      "tokens_seen": 460390400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005908194628333649,
      "loss": 3.0563,
      "theoretical_loss": 3.9500027484164484,
      "tokens_seen": 460455936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005908029128049933,
      "loss": 2.9766,
      "theoretical_loss": 3.9499413379537027,
      "tokens_seen": 460521472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005907863627766219,
      "loss": 2.8829,
      "theoretical_loss": 3.9498799386761334,
      "tokens_seen": 460587008
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005907698127482504,
      "loss": 2.9156,
      "theoretical_loss": 3.949818550580112,
      "tokens_seen": 460652544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005907532627198789,
      "loss": 3.0578,
      "theoretical_loss": 3.9497571736620127,
      "tokens_seen": 460718080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005907367126915075,
      "loss": 2.8467,
      "theoretical_loss": 3.949695807918209,
      "tokens_seen": 460783616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000590720162663136,
      "loss": 2.7921,
      "theoretical_loss": 3.9496344533450793,
      "tokens_seen": 460849152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005907036126347646,
      "loss": 2.732,
      "theoretical_loss": 3.949573109939001,
      "tokens_seen": 460914688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005906870626063931,
      "loss": 2.7504,
      "theoretical_loss": 3.9495117776963538,
      "tokens_seen": 460980224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005906705125780215,
      "loss": 2.7284,
      "theoretical_loss": 3.949450456613521,
      "tokens_seen": 461045760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005906539625496501,
      "loss": 3.0186,
      "theoretical_loss": 3.9493891466868845,
      "tokens_seen": 461111296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005906374125212786,
      "loss": 2.9699,
      "theoretical_loss": 3.949327847912831,
      "tokens_seen": 461176832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005906208624929071,
      "loss": 2.664,
      "theoretical_loss": 3.9492665602877457,
      "tokens_seen": 461242368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005906043124645357,
      "loss": 3.0202,
      "theoretical_loss": 3.9492052838080185,
      "tokens_seen": 461307904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005905877624361642,
      "loss": 2.7841,
      "theoretical_loss": 3.9491440184700393,
      "tokens_seen": 461373440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005905712124077928,
      "loss": 2.7024,
      "theoretical_loss": 3.9490827642701998,
      "tokens_seen": 461438976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005905546623794212,
      "loss": 2.9765,
      "theoretical_loss": 3.9490215212048936,
      "tokens_seen": 461504512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005905381123510497,
      "loss": 2.7665,
      "theoretical_loss": 3.9489602892705165,
      "tokens_seen": 461570048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005905215623226783,
      "loss": 3.0235,
      "theoretical_loss": 3.948899068463465,
      "tokens_seen": 461635584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005905050122943068,
      "loss": 3.05,
      "theoretical_loss": 3.948837858780138,
      "tokens_seen": 461701120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005904884622659353,
      "loss": 2.7842,
      "theoretical_loss": 3.948776660216935,
      "tokens_seen": 461766656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005904719122375639,
      "loss": 2.8235,
      "theoretical_loss": 3.948715472770259,
      "tokens_seen": 461832192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0005904553622091924,
      "loss": 2.7432,
      "theoretical_loss": 3.948654296436513,
      "tokens_seen": 461897728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000590438812180821,
      "loss": 2.7577,
      "theoretical_loss": 3.9485931312121028,
      "tokens_seen": 461963264
    },
    {
      "epoch": 0.16,
      "objective/train/avg_token_score": 0.013129767030477524,
      "objective/train/avg_weight": 0.9845472574234009,
      "objective/train/docs_used": 271706,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9124603271484375,
      "objective/train/std_weight": 0.0559675469994545,
      "objective/train/theoretical_loss": 3.9485319770934355,
      "objective/train/tokens_used": 482488800,
      "theoretical_loss": 3.9485319770934355,
      "tokens_seen": 462028800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005904222621524494,
      "loss": 2.8477,
      "theoretical_loss": 3.9485319770934355,
      "tokens_seen": 462028800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005904057121240779,
      "loss": 3.2762,
      "theoretical_loss": 3.948470834076919,
      "tokens_seen": 462094336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005903891620957065,
      "loss": 2.5986,
      "theoretical_loss": 3.9484097021589646,
      "tokens_seen": 462159872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005903726120673349,
      "loss": 2.8666,
      "theoretical_loss": 3.948348581335984,
      "tokens_seen": 462225408
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005903560620389635,
      "loss": 2.6919,
      "theoretical_loss": 3.94828747160439,
      "tokens_seen": 462290944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000590339512010592,
      "loss": 2.8212,
      "theoretical_loss": 3.9482263729605993,
      "tokens_seen": 462356480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005903229619822204,
      "loss": 2.9027,
      "theoretical_loss": 3.9481652854010285,
      "tokens_seen": 462422016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000590306411953849,
      "loss": 2.6612,
      "theoretical_loss": 3.948104208922096,
      "tokens_seen": 462487552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005902898619254775,
      "loss": 2.7817,
      "theoretical_loss": 3.9480431435202226,
      "tokens_seen": 462553088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005902733118971061,
      "loss": 2.8793,
      "theoretical_loss": 3.947982089191829,
      "tokens_seen": 462618624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005902567618687346,
      "loss": 2.9851,
      "theoretical_loss": 3.9479210459333407,
      "tokens_seen": 462684160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005902402118403631,
      "loss": 2.8617,
      "theoretical_loss": 3.9478600137411823,
      "tokens_seen": 462749696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005902236618119917,
      "loss": 2.8605,
      "theoretical_loss": 3.94779899261178,
      "tokens_seen": 462815232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005902071117836202,
      "loss": 2.9082,
      "theoretical_loss": 3.9477379825415637,
      "tokens_seen": 462880768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005901905617552486,
      "loss": 2.8931,
      "theoretical_loss": 3.9476769835269634,
      "tokens_seen": 462946304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005901740117268772,
      "loss": 2.7554,
      "theoretical_loss": 3.9476159955644112,
      "tokens_seen": 463011840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005901574616985057,
      "loss": 3.0695,
      "theoretical_loss": 3.94755501865034,
      "tokens_seen": 463077376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005901409116701343,
      "loss": 2.9035,
      "theoretical_loss": 3.947494052781185,
      "tokens_seen": 463142912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005901243616417628,
      "loss": 2.9619,
      "theoretical_loss": 3.9474330979533843,
      "tokens_seen": 463208448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005901078116133913,
      "loss": 2.7395,
      "theoretical_loss": 3.947372154163375,
      "tokens_seen": 463273984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005900912615850199,
      "loss": 2.7899,
      "theoretical_loss": 3.947311221407599,
      "tokens_seen": 463339520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005900747115566483,
      "loss": 2.7856,
      "theoretical_loss": 3.947250299682498,
      "tokens_seen": 463405056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005900581615282768,
      "loss": 2.7593,
      "theoretical_loss": 3.9471893889845138,
      "tokens_seen": 463470592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005900416114999054,
      "loss": 2.6871,
      "theoretical_loss": 3.947128489310093,
      "tokens_seen": 463536128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005900250614715339,
      "loss": 3.0032,
      "theoretical_loss": 3.947067600655682,
      "tokens_seen": 463601664
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005900085114431625,
      "loss": 2.7121,
      "theoretical_loss": 3.94700672301773,
      "tokens_seen": 463667200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000589991961414791,
      "loss": 2.6619,
      "theoretical_loss": 3.946945856392687,
      "tokens_seen": 463732736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005899754113864195,
      "loss": 2.8619,
      "theoretical_loss": 3.9468850007770033,
      "tokens_seen": 463798272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005899588613580481,
      "loss": 2.8108,
      "theoretical_loss": 3.9468241561671342,
      "tokens_seen": 463863808
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005899423113296765,
      "loss": 2.8591,
      "theoretical_loss": 3.946763322559534,
      "tokens_seen": 463929344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005899257613013051,
      "loss": 2.736,
      "theoretical_loss": 3.946702499950659,
      "tokens_seen": 463994880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005899092112729336,
      "loss": 2.6993,
      "theoretical_loss": 3.9466416883369684,
      "tokens_seen": 464060416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005898926612445621,
      "loss": 2.9211,
      "theoretical_loss": 3.946580887714921,
      "tokens_seen": 464125952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005898761112161907,
      "loss": 2.8076,
      "theoretical_loss": 3.9465200980809803,
      "tokens_seen": 464191488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005898595611878192,
      "loss": 2.9371,
      "theoretical_loss": 3.946459319431608,
      "tokens_seen": 464257024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005898430111594477,
      "loss": 2.7876,
      "theoretical_loss": 3.9463985517632696,
      "tokens_seen": 464322560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005898264611310762,
      "loss": 2.9828,
      "theoretical_loss": 3.9463377950724317,
      "tokens_seen": 464388096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005898099111027047,
      "loss": 2.8842,
      "theoretical_loss": 3.9462770493555617,
      "tokens_seen": 464453632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005897933610743333,
      "loss": 3.0288,
      "theoretical_loss": 3.9462163146091305,
      "tokens_seen": 464519168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005897768110459618,
      "loss": 2.9702,
      "theoretical_loss": 3.9461555908296093,
      "tokens_seen": 464584704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005897602610175903,
      "loss": 2.6108,
      "theoretical_loss": 3.9460948780134713,
      "tokens_seen": 464650240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005897437109892189,
      "loss": 2.9946,
      "theoretical_loss": 3.94603417615719,
      "tokens_seen": 464715776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005897271609608474,
      "loss": 2.9387,
      "theoretical_loss": 3.9459734852572437,
      "tokens_seen": 464781312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005897106109324758,
      "loss": 2.7861,
      "theoretical_loss": 3.9459128053101082,
      "tokens_seen": 464846848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005896940609041044,
      "loss": 2.681,
      "theoretical_loss": 3.945852136312265,
      "tokens_seen": 464912384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005896775108757329,
      "loss": 2.7148,
      "theoretical_loss": 3.945791478260195,
      "tokens_seen": 464977920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005896609608473615,
      "loss": 3.027,
      "theoretical_loss": 3.94573083115038,
      "tokens_seen": 465043456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00058964441081899,
      "loss": 2.7709,
      "theoretical_loss": 3.9456701949793054,
      "tokens_seen": 465108992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005896278607906185,
      "loss": 2.7332,
      "theoretical_loss": 3.9456095697434566,
      "tokens_seen": 465174528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005896113107622471,
      "loss": 3.0073,
      "theoretical_loss": 3.945548955439322,
      "tokens_seen": 465240064
    },
    {
      "epoch": 0.17,
      "objective/train/avg_token_score": 0.017837002873420715,
      "objective/train/avg_weight": 0.9790071249008179,
      "objective/train/docs_used": 273436,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7174651622772217,
      "objective/train/std_weight": 0.08553119003772736,
      "objective/train/theoretical_loss": 3.945488352063391,
      "objective/train/tokens_used": 485765600,
      "theoretical_loss": 3.945488352063391,
      "tokens_seen": 465305600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005895947607338756,
      "loss": 2.7631,
      "theoretical_loss": 3.945488352063391,
      "tokens_seen": 465305600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000589578210705504,
      "loss": 2.7317,
      "theoretical_loss": 3.945427759612154,
      "tokens_seen": 465371136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005895616606771326,
      "loss": 2.9723,
      "theoretical_loss": 3.945367178082104,
      "tokens_seen": 465436672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005895451106487611,
      "loss": 2.8575,
      "theoretical_loss": 3.945306607469735,
      "tokens_seen": 465502208
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005895285606203897,
      "loss": 2.8615,
      "theoretical_loss": 3.945246047771543,
      "tokens_seen": 465567744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005895120105920182,
      "loss": 2.924,
      "theoretical_loss": 3.9451854989840256,
      "tokens_seen": 465633280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005894954605636467,
      "loss": 2.9516,
      "theoretical_loss": 3.945124961103682,
      "tokens_seen": 465698816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005894789105352753,
      "loss": 2.5361,
      "theoretical_loss": 3.9450644341270116,
      "tokens_seen": 465764352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005894623605069038,
      "loss": 2.5778,
      "theoretical_loss": 3.945003918050518,
      "tokens_seen": 465829888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005894458104785322,
      "loss": 2.8606,
      "theoretical_loss": 3.9449434128707055,
      "tokens_seen": 465895424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005894292604501608,
      "loss": 2.8415,
      "theoretical_loss": 3.9448829185840784,
      "tokens_seen": 465960960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005894127104217893,
      "loss": 2.7219,
      "theoretical_loss": 3.9448224351871444,
      "tokens_seen": 466026496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005893961603934179,
      "loss": 2.6733,
      "theoretical_loss": 3.944761962676412,
      "tokens_seen": 466092032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005893796103650463,
      "loss": 2.9218,
      "theoretical_loss": 3.9447015010483923,
      "tokens_seen": 466157568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005893630603366749,
      "loss": 2.8183,
      "theoretical_loss": 3.9446410502995963,
      "tokens_seen": 466223104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005893465103083034,
      "loss": 2.9255,
      "theoretical_loss": 3.9445806104265384,
      "tokens_seen": 466288640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005893299602799318,
      "loss": 2.642,
      "theoretical_loss": 3.9445201814257334,
      "tokens_seen": 466354176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005893134102515604,
      "loss": 2.7826,
      "theoretical_loss": 3.9444597632936977,
      "tokens_seen": 466419712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005892968602231889,
      "loss": 2.7918,
      "theoretical_loss": 3.9443993560269504,
      "tokens_seen": 466485248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005892803101948174,
      "loss": 3.0742,
      "theoretical_loss": 3.9443389596220113,
      "tokens_seen": 466550784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000589263760166446,
      "loss": 2.8927,
      "theoretical_loss": 3.944278574075402,
      "tokens_seen": 466616320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005892472101380745,
      "loss": 2.858,
      "theoretical_loss": 3.944218199383645,
      "tokens_seen": 466681856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005892306601097031,
      "loss": 2.8029,
      "theoretical_loss": 3.9441578355432663,
      "tokens_seen": 466747392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005892141100813315,
      "loss": 3.0168,
      "theoretical_loss": 3.9440974825507915,
      "tokens_seen": 466812928
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00058919756005296,
      "loss": 2.631,
      "theoretical_loss": 3.9440371404027488,
      "tokens_seen": 466878464
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005891810100245886,
      "loss": 2.7308,
      "theoretical_loss": 3.943976809095668,
      "tokens_seen": 466944000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005891644599962171,
      "loss": 2.8276,
      "theoretical_loss": 3.94391648862608,
      "tokens_seen": 467009536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005891479099678456,
      "loss": 2.8228,
      "theoretical_loss": 3.9438561789905178,
      "tokens_seen": 467075072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005891313599394742,
      "loss": 2.8473,
      "theoretical_loss": 3.9437958801855157,
      "tokens_seen": 467140608
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005891148099111027,
      "loss": 2.7318,
      "theoretical_loss": 3.9437355922076094,
      "tokens_seen": 467206144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005890982598827313,
      "loss": 3.0146,
      "theoretical_loss": 3.9436753150533375,
      "tokens_seen": 467271680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005890817098543597,
      "loss": 2.8108,
      "theoretical_loss": 3.943615048719238,
      "tokens_seen": 467337216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005890651598259882,
      "loss": 3.0556,
      "theoretical_loss": 3.9435547932018524,
      "tokens_seen": 467402752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005890486097976168,
      "loss": 2.526,
      "theoretical_loss": 3.943494548497723,
      "tokens_seen": 467468288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005890320597692453,
      "loss": 2.8565,
      "theoretical_loss": 3.943434314603394,
      "tokens_seen": 467533824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005890155097408738,
      "loss": 2.7363,
      "theoretical_loss": 3.9433740915154094,
      "tokens_seen": 467599360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005889989597125024,
      "loss": 2.9755,
      "theoretical_loss": 3.9433138792303177,
      "tokens_seen": 467664896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005889824096841309,
      "loss": 2.8948,
      "theoretical_loss": 3.943253677744668,
      "tokens_seen": 467730432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005889658596557594,
      "loss": 2.681,
      "theoretical_loss": 3.943193487055009,
      "tokens_seen": 467795968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005889493096273879,
      "loss": 2.7932,
      "theoretical_loss": 3.9431333071578942,
      "tokens_seen": 467861504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005889327595990164,
      "loss": 2.8599,
      "theoretical_loss": 3.9430731380498765,
      "tokens_seen": 467927040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000588916209570645,
      "loss": 2.861,
      "theoretical_loss": 3.94301297972751,
      "tokens_seen": 467992576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005888996595422735,
      "loss": 2.4497,
      "theoretical_loss": 3.9429528321873537,
      "tokens_seen": 468058112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000588883109513902,
      "loss": 2.689,
      "theoretical_loss": 3.942892695425963,
      "tokens_seen": 468123648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005888665594855306,
      "loss": 2.9959,
      "theoretical_loss": 3.9428325694398993,
      "tokens_seen": 468189184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000588850009457159,
      "loss": 2.79,
      "theoretical_loss": 3.942772454225724,
      "tokens_seen": 468254720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005888334594287876,
      "loss": 2.7949,
      "theoretical_loss": 3.94271234978,
      "tokens_seen": 468320256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005888169094004161,
      "loss": 2.7898,
      "theoretical_loss": 3.9426522560992905,
      "tokens_seen": 468385792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005888003593720446,
      "loss": 3.0595,
      "theoretical_loss": 3.942592173180163,
      "tokens_seen": 468451328
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005887838093436732,
      "loss": 2.8542,
      "theoretical_loss": 3.9425321010191854,
      "tokens_seen": 468516864
    },
    {
      "epoch": 0.17,
      "objective/train/avg_token_score": 0.02065173164010048,
      "objective/train/avg_weight": 0.9756922721862793,
      "objective/train/docs_used": 275552,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.913214683532715,
      "objective/train/std_weight": 0.1083492636680603,
      "objective/train/theoretical_loss": 3.942472039612926,
      "objective/train/tokens_used": 489042400,
      "theoretical_loss": 3.942472039612926,
      "tokens_seen": 468582400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005887672593153017,
      "loss": 2.9026,
      "theoretical_loss": 3.942472039612926,
      "tokens_seen": 468582400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005887507092869302,
      "loss": 2.8303,
      "theoretical_loss": 3.9424119889579563,
      "tokens_seen": 468647936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005887341592585588,
      "loss": 2.773,
      "theoretical_loss": 3.9423519490508476,
      "tokens_seen": 468713472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005887176092301872,
      "loss": 2.7199,
      "theoretical_loss": 3.9422919198881754,
      "tokens_seen": 468779008
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005887010592018158,
      "loss": 2.7767,
      "theoretical_loss": 3.9422319014665144,
      "tokens_seen": 468844544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005886845091734443,
      "loss": 2.6581,
      "theoretical_loss": 3.942171893782442,
      "tokens_seen": 468910080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005886679591450728,
      "loss": 2.8944,
      "theoretical_loss": 3.942111896832537,
      "tokens_seen": 468975616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005886514091167014,
      "loss": 2.7545,
      "theoretical_loss": 3.9420519106133787,
      "tokens_seen": 469041152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005886348590883299,
      "loss": 2.5439,
      "theoretical_loss": 3.9419919351215498,
      "tokens_seen": 469106688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005886183090599584,
      "loss": 2.7955,
      "theoretical_loss": 3.9419319703536337,
      "tokens_seen": 469172224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005886017590315869,
      "loss": 2.846,
      "theoretical_loss": 3.941872016306215,
      "tokens_seen": 469237760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005885852090032154,
      "loss": 2.4137,
      "theoretical_loss": 3.941812072975881,
      "tokens_seen": 469303296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000588568658974844,
      "loss": 2.848,
      "theoretical_loss": 3.9417521403592186,
      "tokens_seen": 469368832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005885521089464725,
      "loss": 2.859,
      "theoretical_loss": 3.9416922184528183,
      "tokens_seen": 469434368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000588535558918101,
      "loss": 2.5237,
      "theoretical_loss": 3.94163230725327,
      "tokens_seen": 469499904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005885190088897296,
      "loss": 2.7811,
      "theoretical_loss": 3.9415724067571687,
      "tokens_seen": 469565440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005885024588613581,
      "loss": 2.8722,
      "theoretical_loss": 3.941512516961107,
      "tokens_seen": 469630976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005884859088329867,
      "loss": 2.6774,
      "theoretical_loss": 3.941452637861681,
      "tokens_seen": 469696512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005884693588046151,
      "loss": 2.9086,
      "theoretical_loss": 3.9413927694554887,
      "tokens_seen": 469762048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005884528087762436,
      "loss": 2.7403,
      "theoretical_loss": 3.9413329117391287,
      "tokens_seen": 469827584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005884362587478722,
      "loss": 2.9882,
      "theoretical_loss": 3.9412730647092022,
      "tokens_seen": 469893120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005884197087195007,
      "loss": 2.8522,
      "theoretical_loss": 3.94121322836231,
      "tokens_seen": 469958656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005884031586911292,
      "loss": 2.9836,
      "theoretical_loss": 3.9411534026950568,
      "tokens_seen": 470024192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005883866086627577,
      "loss": 2.7326,
      "theoretical_loss": 3.9410935877040476,
      "tokens_seen": 470089728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005883700586343861,
      "loss": 2.7034,
      "theoretical_loss": 3.9410337833858886,
      "tokens_seen": 470155264
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005883535086060147,
      "loss": 2.9504,
      "theoretical_loss": 3.9409739897371896,
      "tokens_seen": 470220800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005883369585776432,
      "loss": 2.7242,
      "theoretical_loss": 3.9409142067545586,
      "tokens_seen": 470286336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005883204085492718,
      "loss": 2.9122,
      "theoretical_loss": 3.940854434434608,
      "tokens_seen": 470351872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005883038585209003,
      "loss": 2.9249,
      "theoretical_loss": 3.940794672773951,
      "tokens_seen": 470417408
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005882873084925288,
      "loss": 2.8958,
      "theoretical_loss": 3.9407349217692023,
      "tokens_seen": 470482944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005882707584641574,
      "loss": 2.7903,
      "theoretical_loss": 3.9406751814169763,
      "tokens_seen": 470548480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005882542084357859,
      "loss": 2.7491,
      "theoretical_loss": 3.9406154517138923,
      "tokens_seen": 470614016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005882376584074143,
      "loss": 2.8142,
      "theoretical_loss": 3.940555732656569,
      "tokens_seen": 470679552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005882211083790429,
      "loss": 3.099,
      "theoretical_loss": 3.940496024241627,
      "tokens_seen": 470745088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005882045583506714,
      "loss": 3.0605,
      "theoretical_loss": 3.9404363264656883,
      "tokens_seen": 470810624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005881880083223,
      "loss": 2.8662,
      "theoretical_loss": 3.9403766393253767,
      "tokens_seen": 470876160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005881714582939285,
      "loss": 2.6241,
      "theoretical_loss": 3.9403169628173185,
      "tokens_seen": 470941696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000588154908265557,
      "loss": 2.8378,
      "theoretical_loss": 3.9402572969381393,
      "tokens_seen": 471007232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005881383582371856,
      "loss": 3.0425,
      "theoretical_loss": 3.9401976416844677,
      "tokens_seen": 471072768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000588121808208814,
      "loss": 3.1904,
      "theoretical_loss": 3.9401379970529344,
      "tokens_seen": 471138304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005881052581804425,
      "loss": 2.842,
      "theoretical_loss": 3.94007836304017,
      "tokens_seen": 471203840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005880887081520711,
      "loss": 2.8817,
      "theoretical_loss": 3.9400187396428086,
      "tokens_seen": 471269376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005880721581236996,
      "loss": 2.6682,
      "theoretical_loss": 3.9399591268574836,
      "tokens_seen": 471334912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005880556080953282,
      "loss": 2.7494,
      "theoretical_loss": 3.9398995246808317,
      "tokens_seen": 471400448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005880390580669567,
      "loss": 2.8169,
      "theoretical_loss": 3.93983993310949,
      "tokens_seen": 471465984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005880225080385852,
      "loss": 2.7968,
      "theoretical_loss": 3.9397803521400983,
      "tokens_seen": 471531520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005880059580102138,
      "loss": 2.7547,
      "theoretical_loss": 3.9397207817692967,
      "tokens_seen": 471597056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005879894079818422,
      "loss": 2.7119,
      "theoretical_loss": 3.9396612219937284,
      "tokens_seen": 471662592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005879728579534707,
      "loss": 3.1484,
      "theoretical_loss": 3.9396016728100363,
      "tokens_seen": 471728128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005879563079250993,
      "loss": 2.7997,
      "theoretical_loss": 3.9395421342148653,
      "tokens_seen": 471793664
    },
    {
      "epoch": 0.17,
      "objective/train/avg_token_score": 0.014187118969857693,
      "objective/train/avg_weight": 0.9832996726036072,
      "objective/train/docs_used": 277578,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.892911434173584,
      "objective/train/std_weight": 0.08063876628875732,
      "objective/train/theoretical_loss": 3.939482606204863,
      "objective/train/tokens_used": 492319200,
      "theoretical_loss": 3.939482606204863,
      "tokens_seen": 471859200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005879397578967278,
      "loss": 2.9057,
      "theoretical_loss": 3.939482606204863,
      "tokens_seen": 471859200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005879232078683564,
      "loss": 3.0567,
      "theoretical_loss": 3.9394230887766772,
      "tokens_seen": 471924736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005879066578399849,
      "loss": 2.8036,
      "theoretical_loss": 3.939363581926959,
      "tokens_seen": 471990272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005878901078116134,
      "loss": 2.6622,
      "theoretical_loss": 3.939304085652358,
      "tokens_seen": 472055808
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005878735577832419,
      "loss": 2.9636,
      "theoretical_loss": 3.939244599949528,
      "tokens_seen": 472121344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005878570077548704,
      "loss": 2.9982,
      "theoretical_loss": 3.939185124815123,
      "tokens_seen": 472186880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005878404577264989,
      "loss": 2.8001,
      "theoretical_loss": 3.9391256602457996,
      "tokens_seen": 472252416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005878239076981275,
      "loss": 2.7789,
      "theoretical_loss": 3.9390662062382145,
      "tokens_seen": 472317952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000587807357669756,
      "loss": 2.66,
      "theoretical_loss": 3.9390067627890275,
      "tokens_seen": 472383488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005877908076413846,
      "loss": 2.5747,
      "theoretical_loss": 3.938947329894899,
      "tokens_seen": 472449024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005877742576130131,
      "loss": 2.9578,
      "theoretical_loss": 3.9388879075524907,
      "tokens_seen": 472514560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005877577075846415,
      "loss": 2.9083,
      "theoretical_loss": 3.9388284957584663,
      "tokens_seen": 472580096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005877411575562701,
      "loss": 2.5816,
      "theoretical_loss": 3.9387690945094906,
      "tokens_seen": 472645632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005877246075278986,
      "loss": 2.8311,
      "theoretical_loss": 3.93870970380223,
      "tokens_seen": 472711168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005877080574995271,
      "loss": 3.0393,
      "theoretical_loss": 3.938650323633354,
      "tokens_seen": 472776704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005876915074711557,
      "loss": 2.7137,
      "theoretical_loss": 3.9385909539995305,
      "tokens_seen": 472842240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005876749574427842,
      "loss": 2.6118,
      "theoretical_loss": 3.9385315948974315,
      "tokens_seen": 472907776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005876584074144128,
      "loss": 2.9234,
      "theoretical_loss": 3.9384722463237294,
      "tokens_seen": 472973312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005876418573860413,
      "loss": 2.8829,
      "theoretical_loss": 3.938412908275099,
      "tokens_seen": 473038848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005876253073576697,
      "loss": 2.8656,
      "theoretical_loss": 3.938353580748215,
      "tokens_seen": 473104384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005876087573292983,
      "loss": 3.0001,
      "theoretical_loss": 3.938294263739755,
      "tokens_seen": 473169920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005875922073009268,
      "loss": 2.8491,
      "theoretical_loss": 3.9382349572463973,
      "tokens_seen": 473235456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005875756572725553,
      "loss": 2.8272,
      "theoretical_loss": 3.938175661264823,
      "tokens_seen": 473300992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005875591072441839,
      "loss": 2.8706,
      "theoretical_loss": 3.9381163757917133,
      "tokens_seen": 473366528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005875425572158124,
      "loss": 2.8487,
      "theoretical_loss": 3.938057100823751,
      "tokens_seen": 473432064
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000587526007187441,
      "loss": 2.8462,
      "theoretical_loss": 3.9379978363576216,
      "tokens_seen": 473497600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005875094571590694,
      "loss": 2.8801,
      "theoretical_loss": 3.937938582390011,
      "tokens_seen": 473563136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005874929071306979,
      "loss": 2.7948,
      "theoretical_loss": 3.9378793389176066,
      "tokens_seen": 473628672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005874763571023265,
      "loss": 2.868,
      "theoretical_loss": 3.9378201059370976,
      "tokens_seen": 473694208
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000587459807073955,
      "loss": 3.0631,
      "theoretical_loss": 3.9377608834451756,
      "tokens_seen": 473759744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005874432570455835,
      "loss": 2.5867,
      "theoretical_loss": 3.9377016714385316,
      "tokens_seen": 473825280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000587426707017212,
      "loss": 2.6671,
      "theoretical_loss": 3.9376424699138606,
      "tokens_seen": 473890816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005874101569888406,
      "loss": 3.2022,
      "theoretical_loss": 3.937583278867857,
      "tokens_seen": 473956352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000587393606960469,
      "loss": 2.8888,
      "theoretical_loss": 3.9375240982972173,
      "tokens_seen": 474021888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005873770569320975,
      "loss": 2.6517,
      "theoretical_loss": 3.9374649281986405,
      "tokens_seen": 474087424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005873605069037261,
      "loss": 2.9324,
      "theoretical_loss": 3.937405768568826,
      "tokens_seen": 474152960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005873439568753546,
      "loss": 2.8151,
      "theoretical_loss": 3.937346619404475,
      "tokens_seen": 474218496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005873274068469831,
      "loss": 2.6648,
      "theoretical_loss": 3.93728748070229,
      "tokens_seen": 474284032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005873108568186117,
      "loss": 2.5639,
      "theoretical_loss": 3.937228352458976,
      "tokens_seen": 474349568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005872943067902402,
      "loss": 2.9581,
      "theoretical_loss": 3.9371692346712375,
      "tokens_seen": 474415104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005872777567618688,
      "loss": 2.7086,
      "theoretical_loss": 3.9371101273357825,
      "tokens_seen": 474480640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005872612067334972,
      "loss": 2.8275,
      "theoretical_loss": 3.93705103044932,
      "tokens_seen": 474546176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005872446567051257,
      "loss": 2.8133,
      "theoretical_loss": 3.9369919440085592,
      "tokens_seen": 474611712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005872281066767543,
      "loss": 2.9612,
      "theoretical_loss": 3.9369328680102122,
      "tokens_seen": 474677248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005872115566483828,
      "loss": 2.9543,
      "theoretical_loss": 3.9368738024509926,
      "tokens_seen": 474742784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005871950066200113,
      "loss": 2.8104,
      "theoretical_loss": 3.9368147473276145,
      "tokens_seen": 474808320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005871784565916399,
      "loss": 2.9162,
      "theoretical_loss": 3.936755702636794,
      "tokens_seen": 474873856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005871619065632684,
      "loss": 2.8655,
      "theoretical_loss": 3.9366966683752493,
      "tokens_seen": 474939392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000587145356534897,
      "loss": 2.6806,
      "theoretical_loss": 3.936637644539699,
      "tokens_seen": 475004928
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005871288065065254,
      "loss": 2.6666,
      "theoretical_loss": 3.9365786311268636,
      "tokens_seen": 475070464
    },
    {
      "epoch": 0.17,
      "objective/train/avg_token_score": 0.018377993255853653,
      "objective/train/avg_weight": 0.9783634543418884,
      "objective/train/docs_used": 279357,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3731799125671387,
      "objective/train/std_weight": 0.12130886316299438,
      "objective/train/theoretical_loss": 3.936519628133466,
      "objective/train/tokens_used": 495596000,
      "theoretical_loss": 3.936519628133466,
      "tokens_seen": 475136000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005871122564781539,
      "loss": 3.139,
      "theoretical_loss": 3.936519628133466,
      "tokens_seen": 475136000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005870957064497825,
      "loss": 2.9362,
      "theoretical_loss": 3.9364606355562284,
      "tokens_seen": 475201536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000587079156421411,
      "loss": 2.8323,
      "theoretical_loss": 3.9364016533918766,
      "tokens_seen": 475267072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005870626063930395,
      "loss": 2.5546,
      "theoretical_loss": 3.9363426816371376,
      "tokens_seen": 475332608
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005870460563646681,
      "loss": 2.6278,
      "theoretical_loss": 3.9362837202887384,
      "tokens_seen": 475398144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005870295063362965,
      "loss": 2.7451,
      "theoretical_loss": 3.9362247693434096,
      "tokens_seen": 475463680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005870129563079251,
      "loss": 2.6794,
      "theoretical_loss": 3.9361658287978813,
      "tokens_seen": 475529216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005869964062795536,
      "loss": 2.9223,
      "theoretical_loss": 3.9361068986488856,
      "tokens_seen": 475594752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005869798562511821,
      "loss": 2.9281,
      "theoretical_loss": 3.936047978893158,
      "tokens_seen": 475660288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005869633062228107,
      "loss": 3.0529,
      "theoretical_loss": 3.935989069527432,
      "tokens_seen": 475725824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005869467561944392,
      "loss": 2.9629,
      "theoretical_loss": 3.9359301705484455,
      "tokens_seen": 475791360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005869302061660677,
      "loss": 2.7445,
      "theoretical_loss": 3.935871281952937,
      "tokens_seen": 475856896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005869136561376963,
      "loss": 2.6704,
      "theoretical_loss": 3.935812403737646,
      "tokens_seen": 475922432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005868971061093247,
      "loss": 3.0548,
      "theoretical_loss": 3.9357535358993134,
      "tokens_seen": 475987968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005868805560809533,
      "loss": 2.7731,
      "theoretical_loss": 3.9356946784346825,
      "tokens_seen": 476053504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005868640060525818,
      "loss": 2.6225,
      "theoretical_loss": 3.9356358313404973,
      "tokens_seen": 476119040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005868474560242103,
      "loss": 2.818,
      "theoretical_loss": 3.9355769946135037,
      "tokens_seen": 476184576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005868309059958389,
      "loss": 3.0434,
      "theoretical_loss": 3.9355181682504483,
      "tokens_seen": 476250112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005868143559674674,
      "loss": 2.8921,
      "theoretical_loss": 3.9354593522480803,
      "tokens_seen": 476315648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005867978059390959,
      "loss": 2.995,
      "theoretical_loss": 3.93540054660315,
      "tokens_seen": 476381184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005867812559107244,
      "loss": 2.603,
      "theoretical_loss": 3.935341751312408,
      "tokens_seen": 476446720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005867647058823529,
      "loss": 2.7762,
      "theoretical_loss": 3.9352829663726085,
      "tokens_seen": 476512256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005867481558539815,
      "loss": 2.7404,
      "theoretical_loss": 3.935224191780505,
      "tokens_seen": 476577792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00058673160582561,
      "loss": 3.034,
      "theoretical_loss": 3.9351654275328536,
      "tokens_seen": 476643328
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005867150557972385,
      "loss": 2.8801,
      "theoretical_loss": 3.9351066736264126,
      "tokens_seen": 476708864
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005866985057688671,
      "loss": 2.8489,
      "theoretical_loss": 3.93504793005794,
      "tokens_seen": 476774400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005866819557404956,
      "loss": 3.0136,
      "theoretical_loss": 3.9349891968241963,
      "tokens_seen": 476839936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000586665405712124,
      "loss": 2.9106,
      "theoretical_loss": 3.934930473921944,
      "tokens_seen": 476905472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005866488556837526,
      "loss": 2.7834,
      "theoretical_loss": 3.934871761347945,
      "tokens_seen": 476971008
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005866323056553811,
      "loss": 2.8861,
      "theoretical_loss": 3.9348130590989654,
      "tokens_seen": 477036544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005866157556270097,
      "loss": 2.931,
      "theoretical_loss": 3.93475436717177,
      "tokens_seen": 477102080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005865992055986382,
      "loss": 2.8461,
      "theoretical_loss": 3.9346956855631285,
      "tokens_seen": 477167616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005865826555702667,
      "loss": 2.7998,
      "theoretical_loss": 3.934637014269808,
      "tokens_seen": 477233152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005865661055418953,
      "loss": 3.1009,
      "theoretical_loss": 3.93457835328858,
      "tokens_seen": 477298688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005865495555135238,
      "loss": 2.8704,
      "theoretical_loss": 3.9345197026162158,
      "tokens_seen": 477364224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005865330054851522,
      "loss": 3.3008,
      "theoretical_loss": 3.9344610622494893,
      "tokens_seen": 477429760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005865164554567808,
      "loss": 2.8969,
      "theoretical_loss": 3.934402432185176,
      "tokens_seen": 477495296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005864999054284093,
      "loss": 2.8261,
      "theoretical_loss": 3.9343438124200514,
      "tokens_seen": 477560832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005864833554000379,
      "loss": 2.7495,
      "theoretical_loss": 3.9342852029508935,
      "tokens_seen": 477626368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005864668053716664,
      "loss": 3.0327,
      "theoretical_loss": 3.9342266037744817,
      "tokens_seen": 477691904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005864502553432949,
      "loss": 2.6191,
      "theoretical_loss": 3.934168014887597,
      "tokens_seen": 477757440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005864337053149234,
      "loss": 2.8983,
      "theoretical_loss": 3.934109436287021,
      "tokens_seen": 477822976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005864171552865518,
      "loss": 3.1039,
      "theoretical_loss": 3.934050867969537,
      "tokens_seen": 477888512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005864006052581804,
      "loss": 2.9226,
      "theoretical_loss": 3.9339923099319316,
      "tokens_seen": 477954048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005863840552298089,
      "loss": 2.7389,
      "theoretical_loss": 3.93393376217099,
      "tokens_seen": 478019584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005863675052014374,
      "loss": 2.8641,
      "theoretical_loss": 3.9338752246835,
      "tokens_seen": 478085120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000586350955173066,
      "loss": 2.9595,
      "theoretical_loss": 3.933816697466252,
      "tokens_seen": 478150656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005863344051446945,
      "loss": 2.954,
      "theoretical_loss": 3.933758180516036,
      "tokens_seen": 478216192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005863178551163231,
      "loss": 2.8713,
      "theoretical_loss": 3.9336996738296444,
      "tokens_seen": 478281728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005863013050879515,
      "loss": 2.9057,
      "theoretical_loss": 3.9336411774038713,
      "tokens_seen": 478347264
    },
    {
      "epoch": 0.17,
      "objective/train/avg_token_score": 0.022203948348760605,
      "objective/train/avg_weight": 0.9738686084747314,
      "objective/train/docs_used": 280385,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.16548752784729,
      "objective/train/std_weight": 0.08596497029066086,
      "objective/train/theoretical_loss": 3.9335826912355114,
      "objective/train/tokens_used": 498872800,
      "theoretical_loss": 3.9335826912355114,
      "tokens_seen": 478412800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00058628475505958,
      "loss": 2.8366,
      "theoretical_loss": 3.9335826912355114,
      "tokens_seen": 478412800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005862682050312086,
      "loss": 2.9231,
      "theoretical_loss": 3.9335242153213614,
      "tokens_seen": 478478336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005862516550028371,
      "loss": 2.9236,
      "theoretical_loss": 3.9334657496582195,
      "tokens_seen": 478543872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005862351049744656,
      "loss": 3.1238,
      "theoretical_loss": 3.933407294242885,
      "tokens_seen": 478609408
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005862185549460942,
      "loss": 2.9562,
      "theoretical_loss": 3.9333488490721598,
      "tokens_seen": 478674944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005862020049177227,
      "loss": 2.9031,
      "theoretical_loss": 3.933290414142845,
      "tokens_seen": 478740480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005861854548893513,
      "loss": 3.0765,
      "theoretical_loss": 3.933231989451744,
      "tokens_seen": 478806016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005861689048609797,
      "loss": 2.8061,
      "theoretical_loss": 3.9331735749956627,
      "tokens_seen": 478871552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005861523548326082,
      "loss": 2.7041,
      "theoretical_loss": 3.933115170771409,
      "tokens_seen": 478937088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005861358048042368,
      "loss": 2.9911,
      "theoretical_loss": 3.9330567767757887,
      "tokens_seen": 479002624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005861192547758653,
      "loss": 2.8291,
      "theoretical_loss": 3.932998393005613,
      "tokens_seen": 479068160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005861027047474938,
      "loss": 2.9792,
      "theoretical_loss": 3.932940019457692,
      "tokens_seen": 479133696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005860861547191224,
      "loss": 2.8471,
      "theoretical_loss": 3.9328816561288384,
      "tokens_seen": 479199232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005860696046907509,
      "loss": 2.8919,
      "theoretical_loss": 3.932823303015866,
      "tokens_seen": 479264768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005860530546623794,
      "loss": 2.9483,
      "theoretical_loss": 3.93276496011559,
      "tokens_seen": 479330304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005860365046340079,
      "loss": 2.8098,
      "theoretical_loss": 3.9327066274248272,
      "tokens_seen": 479395840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005860199546056364,
      "loss": 3.0146,
      "theoretical_loss": 3.932648304940395,
      "tokens_seen": 479461376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000586003404577265,
      "loss": 2.9061,
      "theoretical_loss": 3.932589992659114,
      "tokens_seen": 479526912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005859868545488935,
      "loss": 2.7998,
      "theoretical_loss": 3.9325316905778047,
      "tokens_seen": 479592448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005859703045205221,
      "loss": 2.9001,
      "theoretical_loss": 3.932473398693289,
      "tokens_seen": 479657984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005859537544921506,
      "loss": 2.9925,
      "theoretical_loss": 3.932415117002391,
      "tokens_seen": 479723520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000585937204463779,
      "loss": 3.0013,
      "theoretical_loss": 3.9323568455019364,
      "tokens_seen": 479789056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005859206544354076,
      "loss": 2.9901,
      "theoretical_loss": 3.932298584188751,
      "tokens_seen": 479854592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005859041044070361,
      "loss": 3.112,
      "theoretical_loss": 3.9322403330596636,
      "tokens_seen": 479920128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005858875543786646,
      "loss": 2.9521,
      "theoretical_loss": 3.9321820921115034,
      "tokens_seen": 479985664
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005858710043502932,
      "loss": 2.7432,
      "theoretical_loss": 3.932123861341101,
      "tokens_seen": 480051200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005858544543219217,
      "loss": 2.7354,
      "theoretical_loss": 3.932065640745289,
      "tokens_seen": 480116736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005858379042935503,
      "loss": 2.9494,
      "theoretical_loss": 3.932007430320901,
      "tokens_seen": 480182272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005858213542651788,
      "loss": 3.0016,
      "theoretical_loss": 3.9319492300647725,
      "tokens_seen": 480247808
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005858048042368072,
      "loss": 2.9472,
      "theoretical_loss": 3.93189103997374,
      "tokens_seen": 480313344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005857882542084358,
      "loss": 2.9741,
      "theoretical_loss": 3.9318328600446413,
      "tokens_seen": 480378880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005857717041800643,
      "loss": 2.8156,
      "theoretical_loss": 3.931774690274316,
      "tokens_seen": 480444416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005857551541516928,
      "loss": 2.9143,
      "theoretical_loss": 3.931716530659604,
      "tokens_seen": 480509952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005857386041233214,
      "loss": 2.9739,
      "theoretical_loss": 3.931658381197349,
      "tokens_seen": 480575488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005857220540949499,
      "loss": 2.9914,
      "theoretical_loss": 3.9316002418843934,
      "tokens_seen": 480641024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005857055040665785,
      "loss": 2.97,
      "theoretical_loss": 3.931542112717583,
      "tokens_seen": 480706560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000585688954038207,
      "loss": 2.7659,
      "theoretical_loss": 3.931483993693764,
      "tokens_seen": 480772096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005856724040098354,
      "loss": 2.8279,
      "theoretical_loss": 3.9314258848097845,
      "tokens_seen": 480837632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000585655853981464,
      "loss": 3.0356,
      "theoretical_loss": 3.9313677860624936,
      "tokens_seen": 480903168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005856393039530925,
      "loss": 2.9678,
      "theoretical_loss": 3.931309697448742,
      "tokens_seen": 480968704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000585622753924721,
      "loss": 3.2585,
      "theoretical_loss": 3.9312516189653817,
      "tokens_seen": 481034240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005856062038963496,
      "loss": 2.828,
      "theoretical_loss": 3.931193550609266,
      "tokens_seen": 481099776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005855896538679781,
      "loss": 2.742,
      "theoretical_loss": 3.9311354923772504,
      "tokens_seen": 481165312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005855731038396067,
      "loss": 2.9021,
      "theoretical_loss": 3.9310774442661907,
      "tokens_seen": 481230848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005855565538112351,
      "loss": 2.7526,
      "theoretical_loss": 3.931019406272945,
      "tokens_seen": 481296384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005855400037828636,
      "loss": 2.6995,
      "theoretical_loss": 3.930961378394372,
      "tokens_seen": 481361920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005855234537544922,
      "loss": 2.8533,
      "theoretical_loss": 3.9309033606273327,
      "tokens_seen": 481427456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005855069037261207,
      "loss": 2.9409,
      "theoretical_loss": 3.9308453529686886,
      "tokens_seen": 481492992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005854903536977492,
      "loss": 2.7137,
      "theoretical_loss": 3.930787355415304,
      "tokens_seen": 481558528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005854738036693778,
      "loss": 2.9611,
      "theoretical_loss": 3.930729367964042,
      "tokens_seen": 481624064
    },
    {
      "epoch": 0.17,
      "objective/train/avg_token_score": 0.019138995558023453,
      "objective/train/avg_weight": 0.9774834513664246,
      "objective/train/docs_used": 282096,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8073654174804688,
      "objective/train/std_weight": 0.09075847268104553,
      "objective/train/theoretical_loss": 3.93067139061177,
      "objective/train/tokens_used": 502149600,
      "theoretical_loss": 3.93067139061177,
      "tokens_seen": 481689600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005854572536410061,
      "loss": 2.7699,
      "theoretical_loss": 3.93067139061177,
      "tokens_seen": 481689600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005854407036126347,
      "loss": 2.8624,
      "theoretical_loss": 3.930613423355355,
      "tokens_seen": 481755136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005854241535842632,
      "loss": 2.8646,
      "theoretical_loss": 3.9305554661916657,
      "tokens_seen": 481820672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005854076035558918,
      "loss": 2.9068,
      "theoretical_loss": 3.9304975191175733,
      "tokens_seen": 481886208
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005853910535275203,
      "loss": 2.8591,
      "theoretical_loss": 3.9304395821299485,
      "tokens_seen": 481951744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005853745034991488,
      "loss": 2.6914,
      "theoretical_loss": 3.9303816552256654,
      "tokens_seen": 482017280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005853579534707774,
      "loss": 2.7216,
      "theoretical_loss": 3.930323738401597,
      "tokens_seen": 482082816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005853414034424059,
      "loss": 3.1355,
      "theoretical_loss": 3.930265831654621,
      "tokens_seen": 482148352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005853248534140343,
      "loss": 2.6937,
      "theoretical_loss": 3.930207934981614,
      "tokens_seen": 482213888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005853083033856629,
      "loss": 2.941,
      "theoretical_loss": 3.930150048379454,
      "tokens_seen": 482279424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005852917533572914,
      "loss": 2.6687,
      "theoretical_loss": 3.930092171845022,
      "tokens_seen": 482344960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00058527520332892,
      "loss": 2.9074,
      "theoretical_loss": 3.930034305375199,
      "tokens_seen": 482410496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005852586533005485,
      "loss": 2.9622,
      "theoretical_loss": 3.929976448966868,
      "tokens_seen": 482476032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000585242103272177,
      "loss": 2.9417,
      "theoretical_loss": 3.929918602616913,
      "tokens_seen": 482541568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005852255532438056,
      "loss": 2.7001,
      "theoretical_loss": 3.92986076632222,
      "tokens_seen": 482607104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000585209003215434,
      "loss": 2.8235,
      "theoretical_loss": 3.929802940079676,
      "tokens_seen": 482672640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005851924531870625,
      "loss": 2.8267,
      "theoretical_loss": 3.9297451238861685,
      "tokens_seen": 482738176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005851759031586911,
      "loss": 2.7169,
      "theoretical_loss": 3.9296873177385887,
      "tokens_seen": 482803712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005851593531303196,
      "loss": 2.827,
      "theoretical_loss": 3.929629521633827,
      "tokens_seen": 482869248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005851428031019482,
      "loss": 2.9422,
      "theoretical_loss": 3.929571735568776,
      "tokens_seen": 482934784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005851262530735767,
      "loss": 2.9433,
      "theoretical_loss": 3.92951395954033,
      "tokens_seen": 483000320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005851097030452052,
      "loss": 2.9217,
      "theoretical_loss": 3.9294561935453833,
      "tokens_seen": 483065856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005850931530168338,
      "loss": 2.9649,
      "theoretical_loss": 3.9293984375808337,
      "tokens_seen": 483131392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005850766029884622,
      "loss": 2.9658,
      "theoretical_loss": 3.929340691643579,
      "tokens_seen": 483196928
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005850600529600907,
      "loss": 2.7892,
      "theoretical_loss": 3.929282955730519,
      "tokens_seen": 483262464
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005850435029317193,
      "loss": 2.8229,
      "theoretical_loss": 3.9292252298385533,
      "tokens_seen": 483328000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005850269529033478,
      "loss": 2.7191,
      "theoretical_loss": 3.929167513964585,
      "tokens_seen": 483393536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005850104028749764,
      "loss": 2.7963,
      "theoretical_loss": 3.929109808105518,
      "tokens_seen": 483459072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005849938528466049,
      "loss": 2.7705,
      "theoretical_loss": 3.9290521122582573,
      "tokens_seen": 483524608
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005849773028182334,
      "loss": 2.8741,
      "theoretical_loss": 3.928994426419708,
      "tokens_seen": 483590144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000584960752789862,
      "loss": 2.7239,
      "theoretical_loss": 3.928936750586779,
      "tokens_seen": 483655680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005849442027614904,
      "loss": 2.9012,
      "theoretical_loss": 3.9288790847563795,
      "tokens_seen": 483721216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005849276527331189,
      "loss": 3.0932,
      "theoretical_loss": 3.9288214289254193,
      "tokens_seen": 483786752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005849111027047475,
      "loss": 3.0357,
      "theoretical_loss": 3.9287637830908104,
      "tokens_seen": 483852288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000584894552676376,
      "loss": 2.8078,
      "theoretical_loss": 3.928706147249466,
      "tokens_seen": 483917824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005848780026480046,
      "loss": 2.651,
      "theoretical_loss": 3.9286485213983013,
      "tokens_seen": 483983360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005848614526196331,
      "loss": 2.8959,
      "theoretical_loss": 3.9285909055342314,
      "tokens_seen": 484048896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005848449025912616,
      "loss": 3.0959,
      "theoretical_loss": 3.928533299654174,
      "tokens_seen": 484114432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005848283525628901,
      "loss": 2.7468,
      "theoretical_loss": 3.928475703755048,
      "tokens_seen": 484179968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005848118025345186,
      "loss": 2.7558,
      "theoretical_loss": 3.9284181178337736,
      "tokens_seen": 484245504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005847952525061471,
      "loss": 2.9181,
      "theoretical_loss": 3.9283605418872716,
      "tokens_seen": 484311040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005847787024777757,
      "loss": 2.5269,
      "theoretical_loss": 3.9283029759124646,
      "tokens_seen": 484376576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005847621524494042,
      "loss": 2.7148,
      "theoretical_loss": 3.9282454199062777,
      "tokens_seen": 484442112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005847456024210328,
      "loss": 2.6675,
      "theoretical_loss": 3.928187873865636,
      "tokens_seen": 484507648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005847290523926613,
      "loss": 2.9003,
      "theoretical_loss": 3.928130337787467,
      "tokens_seen": 484573184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005847125023642897,
      "loss": 2.8356,
      "theoretical_loss": 3.9280728116686974,
      "tokens_seen": 484638720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005846959523359183,
      "loss": 2.7521,
      "theoretical_loss": 3.9280152955062584,
      "tokens_seen": 484704256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005846794023075468,
      "loss": 2.8284,
      "theoretical_loss": 3.92795778929708,
      "tokens_seen": 484769792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005846628522791753,
      "loss": 2.7513,
      "theoretical_loss": 3.9279002930380953,
      "tokens_seen": 484835328
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005846463022508039,
      "loss": 2.6308,
      "theoretical_loss": 3.927842806726237,
      "tokens_seen": 484900864
    },
    {
      "epoch": 0.17,
      "objective/train/avg_token_score": 0.016102910041809082,
      "objective/train/avg_weight": 0.9810641407966614,
      "objective/train/docs_used": 284179,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.4117207527160645,
      "objective/train/std_weight": 0.09006743133068085,
      "objective/train/theoretical_loss": 3.927785330358441,
      "objective/train/tokens_used": 505426400,
      "theoretical_loss": 3.927785330358441,
      "tokens_seen": 484966400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005846297522224324,
      "loss": 3.0818,
      "theoretical_loss": 3.927785330358441,
      "tokens_seen": 484966400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000584613202194061,
      "loss": 2.8776,
      "theoretical_loss": 3.9277278639316435,
      "tokens_seen": 485031936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005845966521656895,
      "loss": 2.98,
      "theoretical_loss": 3.9276704074427826,
      "tokens_seen": 485097472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005845801021373179,
      "loss": 2.9537,
      "theoretical_loss": 3.927612960888797,
      "tokens_seen": 485163008
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005845635521089465,
      "loss": 2.9753,
      "theoretical_loss": 3.9275555242666265,
      "tokens_seen": 485228544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000584547002080575,
      "loss": 2.8305,
      "theoretical_loss": 3.9274980975732143,
      "tokens_seen": 485294080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005845304520522035,
      "loss": 2.8089,
      "theoretical_loss": 3.927440680805503,
      "tokens_seen": 485359616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005845139020238321,
      "loss": 2.9686,
      "theoretical_loss": 3.9273832739604373,
      "tokens_seen": 485425152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005844973519954606,
      "loss": 2.5078,
      "theoretical_loss": 3.9273258770349626,
      "tokens_seen": 485490688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005844808019670892,
      "loss": 2.9003,
      "theoretical_loss": 3.9272684900260266,
      "tokens_seen": 485556224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005844642519387175,
      "loss": 2.5632,
      "theoretical_loss": 3.9272111129305785,
      "tokens_seen": 485621760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005844477019103461,
      "loss": 2.8634,
      "theoretical_loss": 3.927153745745567,
      "tokens_seen": 485687296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005844311518819746,
      "loss": 2.8894,
      "theoretical_loss": 3.9270963884679437,
      "tokens_seen": 485752832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005844146018536031,
      "loss": 2.6852,
      "theoretical_loss": 3.9270390410946625,
      "tokens_seen": 485818368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005843980518252317,
      "loss": 2.7696,
      "theoretical_loss": 3.9269817036226757,
      "tokens_seen": 485883904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005843815017968602,
      "loss": 2.8204,
      "theoretical_loss": 3.92692437604894,
      "tokens_seen": 485949440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005843649517684888,
      "loss": 2.7603,
      "theoretical_loss": 3.9268670583704113,
      "tokens_seen": 486014976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005843484017401172,
      "loss": 2.6127,
      "theoretical_loss": 3.926809750584048,
      "tokens_seen": 486080512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005843318517117457,
      "loss": 2.8764,
      "theoretical_loss": 3.9267524526868085,
      "tokens_seen": 486146048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005843153016833743,
      "loss": 2.9764,
      "theoretical_loss": 3.9266951646756554,
      "tokens_seen": 486211584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005842987516550028,
      "loss": 2.8118,
      "theoretical_loss": 3.9266378865475495,
      "tokens_seen": 486277120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005842822016266313,
      "loss": 2.7757,
      "theoretical_loss": 3.9265806182994543,
      "tokens_seen": 486342656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005842656515982599,
      "loss": 2.8629,
      "theoretical_loss": 3.926523359928335,
      "tokens_seen": 486408192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005842491015698884,
      "loss": 2.9414,
      "theoretical_loss": 3.9264661114311576,
      "tokens_seen": 486473728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000584232551541517,
      "loss": 2.8755,
      "theoretical_loss": 3.926408872804889,
      "tokens_seen": 486539264
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005842160015131454,
      "loss": 2.9433,
      "theoretical_loss": 3.9263516440464987,
      "tokens_seen": 486604800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005841994514847739,
      "loss": 2.842,
      "theoretical_loss": 3.926294425152957,
      "tokens_seen": 486670336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005841829014564025,
      "loss": 2.6871,
      "theoretical_loss": 3.926237216121234,
      "tokens_seen": 486735872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000584166351428031,
      "loss": 2.8704,
      "theoretical_loss": 3.9261800169483037,
      "tokens_seen": 486801408
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005841498013996595,
      "loss": 2.8838,
      "theoretical_loss": 3.9261228276311404,
      "tokens_seen": 486866944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005841332513712881,
      "loss": 2.9771,
      "theoretical_loss": 3.926065648166719,
      "tokens_seen": 486932480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005841167013429166,
      "loss": 2.9547,
      "theoretical_loss": 3.926008478552016,
      "tokens_seen": 486998016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005841001513145451,
      "loss": 2.556,
      "theoretical_loss": 3.9259513187840103,
      "tokens_seen": 487063552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005840836012861736,
      "loss": 2.7264,
      "theoretical_loss": 3.9258941688596805,
      "tokens_seen": 487129088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005840670512578021,
      "loss": 2.7851,
      "theoretical_loss": 3.9258370287760087,
      "tokens_seen": 487194624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005840505012294307,
      "loss": 2.8917,
      "theoretical_loss": 3.9257798985299757,
      "tokens_seen": 487260160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005840339512010592,
      "loss": 2.6772,
      "theoretical_loss": 3.9257227781185655,
      "tokens_seen": 487325696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005840174011726877,
      "loss": 2.7153,
      "theoretical_loss": 3.9256656675387633,
      "tokens_seen": 487391232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005840008511443163,
      "loss": 2.7485,
      "theoretical_loss": 3.925608566787555,
      "tokens_seen": 487456768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005839843011159447,
      "loss": 2.8529,
      "theoretical_loss": 3.925551475861928,
      "tokens_seen": 487522304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005839677510875733,
      "loss": 2.9812,
      "theoretical_loss": 3.9254943947588705,
      "tokens_seen": 487587840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005839512010592018,
      "loss": 2.7244,
      "theoretical_loss": 3.9254373234753737,
      "tokens_seen": 487653376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005839346510308303,
      "loss": 2.7451,
      "theoretical_loss": 3.925380262008428,
      "tokens_seen": 487718912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005839181010024589,
      "loss": 2.706,
      "theoretical_loss": 3.9253232103550273,
      "tokens_seen": 487784448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005839015509740874,
      "loss": 3.0523,
      "theoretical_loss": 3.925266168512165,
      "tokens_seen": 487849984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005838850009457159,
      "loss": 2.8369,
      "theoretical_loss": 3.925209136476836,
      "tokens_seen": 487915520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005838684509173445,
      "loss": 2.8163,
      "theoretical_loss": 3.9251521142460377,
      "tokens_seen": 487981056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005838519008889729,
      "loss": 2.7369,
      "theoretical_loss": 3.925095101816768,
      "tokens_seen": 488046592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005838353508606015,
      "loss": 3.0812,
      "theoretical_loss": 3.925038099186027,
      "tokens_seen": 488112128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00058381880083223,
      "loss": 2.8409,
      "theoretical_loss": 3.924981106350815,
      "tokens_seen": 488177664
    },
    {
      "epoch": 0.17,
      "objective/train/avg_token_score": 0.01588856801390648,
      "objective/train/avg_weight": 0.9812967777252197,
      "objective/train/docs_used": 285838,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9200539588928223,
      "objective/train/std_weight": 0.09067492187023163,
      "objective/train/theoretical_loss": 3.9249241233081333,
      "objective/train/tokens_used": 508703200,
      "theoretical_loss": 3.9249241233081333,
      "tokens_seen": 488243200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005838022508038585,
      "loss": 2.8374,
      "theoretical_loss": 3.9249241233081333,
      "tokens_seen": 488243200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005837857007754871,
      "loss": 2.6168,
      "theoretical_loss": 3.924867150054985,
      "tokens_seen": 488308736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005837691507471156,
      "loss": 2.875,
      "theoretical_loss": 3.9248101865883767,
      "tokens_seen": 488374272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000583752600718744,
      "loss": 2.8229,
      "theoretical_loss": 3.9247532329053127,
      "tokens_seen": 488439808
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005837360506903726,
      "loss": 2.6734,
      "theoretical_loss": 3.924696289002801,
      "tokens_seen": 488505344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005837195006620011,
      "loss": 2.7294,
      "theoretical_loss": 3.9246393548778498,
      "tokens_seen": 488570880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005837029506336297,
      "loss": 2.9183,
      "theoretical_loss": 3.9245824305274697,
      "tokens_seen": 488636416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005836864006052582,
      "loss": 3.1046,
      "theoretical_loss": 3.9245255159486714,
      "tokens_seen": 488701952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005836698505768867,
      "loss": 2.729,
      "theoretical_loss": 3.924468611138467,
      "tokens_seen": 488767488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005836533005485153,
      "loss": 2.857,
      "theoretical_loss": 3.9244117160938714,
      "tokens_seen": 488833024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005836367505201438,
      "loss": 2.8198,
      "theoretical_loss": 3.9243548308118994,
      "tokens_seen": 488898560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005836202004917722,
      "loss": 2.7855,
      "theoretical_loss": 3.9242979552895676,
      "tokens_seen": 488964096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005836036504634008,
      "loss": 2.9778,
      "theoretical_loss": 3.9242410895238935,
      "tokens_seen": 489029632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005835871004350293,
      "loss": 2.7921,
      "theoretical_loss": 3.9241842335118964,
      "tokens_seen": 489095168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005835705504066579,
      "loss": 2.7832,
      "theoretical_loss": 3.9241273872505973,
      "tokens_seen": 489160704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005835540003782864,
      "loss": 2.7653,
      "theoretical_loss": 3.924070550737017,
      "tokens_seen": 489226240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005835374503499149,
      "loss": 2.7587,
      "theoretical_loss": 3.9240137239681787,
      "tokens_seen": 489291776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005835209003215435,
      "loss": 2.6643,
      "theoretical_loss": 3.9239569069411067,
      "tokens_seen": 489357312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000583504350293172,
      "loss": 2.9842,
      "theoretical_loss": 3.923900099652828,
      "tokens_seen": 489422848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005834878002648004,
      "loss": 3.0359,
      "theoretical_loss": 3.923843302100368,
      "tokens_seen": 489488384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005834712502364289,
      "loss": 2.9194,
      "theoretical_loss": 3.9237865142807555,
      "tokens_seen": 489553920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005834547002080575,
      "loss": 2.705,
      "theoretical_loss": 3.92372973619102,
      "tokens_seen": 489619456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000583438150179686,
      "loss": 2.5936,
      "theoretical_loss": 3.9236729678281925,
      "tokens_seen": 489684992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005834216001513145,
      "loss": 2.7201,
      "theoretical_loss": 3.923616209189306,
      "tokens_seen": 489750528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005834050501229431,
      "loss": 2.9459,
      "theoretical_loss": 3.9235594602713926,
      "tokens_seen": 489816064
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005833885000945716,
      "loss": 2.9982,
      "theoretical_loss": 3.923502721071488,
      "tokens_seen": 489881600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0005833719500662,
      "loss": 2.7863,
      "theoretical_loss": 3.9234459915866275,
      "tokens_seen": 489947136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005833554000378286,
      "loss": 2.7996,
      "theoretical_loss": 3.923389271813849,
      "tokens_seen": 490012672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005833388500094571,
      "loss": 2.4696,
      "theoretical_loss": 3.923332561750192,
      "tokens_seen": 490078208
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005833222999810857,
      "loss": 2.9982,
      "theoretical_loss": 3.9232758613926952,
      "tokens_seen": 490143744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005833057499527142,
      "loss": 2.6378,
      "theoretical_loss": 3.9232191707384,
      "tokens_seen": 490209280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005832891999243427,
      "loss": 2.8631,
      "theoretical_loss": 3.9231624897843496,
      "tokens_seen": 490274816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005832726498959713,
      "loss": 2.6556,
      "theoretical_loss": 3.9231058185275884,
      "tokens_seen": 490340352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005832560998675997,
      "loss": 2.7173,
      "theoretical_loss": 3.9230491569651598,
      "tokens_seen": 490405888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005832395498392282,
      "loss": 2.8693,
      "theoretical_loss": 3.922992505094112,
      "tokens_seen": 490471424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005832229998108568,
      "loss": 2.8456,
      "theoretical_loss": 3.9229358629114923,
      "tokens_seen": 490536960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005832064497824853,
      "loss": 2.852,
      "theoretical_loss": 3.922879230414349,
      "tokens_seen": 490602496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005831898997541139,
      "loss": 2.872,
      "theoretical_loss": 3.9228226075997332,
      "tokens_seen": 490668032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005831733497257424,
      "loss": 2.7774,
      "theoretical_loss": 3.9227659944646964,
      "tokens_seen": 490733568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005831567996973709,
      "loss": 2.8358,
      "theoretical_loss": 3.9227093910062916,
      "tokens_seen": 490799104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005831402496689995,
      "loss": 2.7289,
      "theoretical_loss": 3.922652797221573,
      "tokens_seen": 490864640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005831236996406279,
      "loss": 2.6452,
      "theoretical_loss": 3.922596213107596,
      "tokens_seen": 490930176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005831071496122564,
      "loss": 2.843,
      "theoretical_loss": 3.922539638661418,
      "tokens_seen": 490995712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000583090599583885,
      "loss": 2.7328,
      "theoretical_loss": 3.9224830738800964,
      "tokens_seen": 491061248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005830740495555135,
      "loss": 2.8626,
      "theoretical_loss": 3.92242651876069,
      "tokens_seen": 491126784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005830574995271421,
      "loss": 2.923,
      "theoretical_loss": 3.9223699733002615,
      "tokens_seen": 491192320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005830409494987706,
      "loss": 2.8484,
      "theoretical_loss": 3.922313437495871,
      "tokens_seen": 491257856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005830243994703991,
      "loss": 2.7417,
      "theoretical_loss": 3.9222569113445824,
      "tokens_seen": 491323392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005830078494420276,
      "loss": 2.8703,
      "theoretical_loss": 3.9222003948434603,
      "tokens_seen": 491388928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005829912994136561,
      "loss": 2.7047,
      "theoretical_loss": 3.92214388798957,
      "tokens_seen": 491454464
    },
    {
      "epoch": 0.18,
      "objective/train/avg_token_score": 0.004644907079637051,
      "objective/train/avg_weight": 0.9945394396781921,
      "objective/train/docs_used": 287192,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0166656970977783,
      "objective/train/std_weight": 0.027169661596417427,
      "objective/train/theoretical_loss": 3.92208739077998,
      "objective/train/tokens_used": 511980000,
      "theoretical_loss": 3.92208739077998,
      "tokens_seen": 491520000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005829747493852846,
      "loss": 2.9308,
      "theoretical_loss": 3.92208739077998,
      "tokens_seen": 491520000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005829581993569132,
      "loss": 2.7765,
      "theoretical_loss": 3.9220309032117573,
      "tokens_seen": 491585536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005829416493285417,
      "loss": 2.9237,
      "theoretical_loss": 3.921974425281972,
      "tokens_seen": 491651072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005829250993001703,
      "loss": 2.8181,
      "theoretical_loss": 3.921917956987695,
      "tokens_seen": 491716608
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005829085492717988,
      "loss": 2.8043,
      "theoretical_loss": 3.921861498325998,
      "tokens_seen": 491782144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005828919992434272,
      "loss": 2.8876,
      "theoretical_loss": 3.921805049293956,
      "tokens_seen": 491847680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005828754492150558,
      "loss": 2.5969,
      "theoretical_loss": 3.9217486098886427,
      "tokens_seen": 491913216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005828588991866843,
      "loss": 2.8604,
      "theoretical_loss": 3.921692180107134,
      "tokens_seen": 491978752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005828423491583128,
      "loss": 3.0166,
      "theoretical_loss": 3.9216357599465073,
      "tokens_seen": 492044288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005828257991299414,
      "loss": 2.8975,
      "theoretical_loss": 3.921579349403842,
      "tokens_seen": 492109824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005828092491015699,
      "loss": 2.8395,
      "theoretical_loss": 3.921522948476217,
      "tokens_seen": 492175360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005827926990731985,
      "loss": 2.7104,
      "theoretical_loss": 3.9214665571607146,
      "tokens_seen": 492240896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000582776149044827,
      "loss": 2.8161,
      "theoretical_loss": 3.921410175454416,
      "tokens_seen": 492306432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005827595990164554,
      "loss": 2.662,
      "theoretical_loss": 3.921353803354406,
      "tokens_seen": 492371968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000582743048988084,
      "loss": 3.0829,
      "theoretical_loss": 3.9212974408577685,
      "tokens_seen": 492437504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005827264989597125,
      "loss": 2.4802,
      "theoretical_loss": 3.921241087961591,
      "tokens_seen": 492503040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000582709948931341,
      "loss": 2.9131,
      "theoretical_loss": 3.92118474466296,
      "tokens_seen": 492568576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005826933989029696,
      "loss": 2.7943,
      "theoretical_loss": 3.9211284109589646,
      "tokens_seen": 492634112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005826768488745981,
      "loss": 2.7648,
      "theoretical_loss": 3.9210720868466953,
      "tokens_seen": 492699648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005826602988462267,
      "loss": 2.8344,
      "theoretical_loss": 3.9210157723232424,
      "tokens_seen": 492765184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005826437488178551,
      "loss": 2.88,
      "theoretical_loss": 3.9209594673857,
      "tokens_seen": 492830720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005826271987894836,
      "loss": 2.8447,
      "theoretical_loss": 3.920903172031161,
      "tokens_seen": 492896256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005826106487611122,
      "loss": 3.1862,
      "theoretical_loss": 3.9208468862567205,
      "tokens_seen": 492961792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005825940987327407,
      "loss": 2.9104,
      "theoretical_loss": 3.9207906100594756,
      "tokens_seen": 493027328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005825775487043692,
      "loss": 2.975,
      "theoretical_loss": 3.920734343436523,
      "tokens_seen": 493092864
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005825609986759978,
      "loss": 2.9234,
      "theoretical_loss": 3.920678086384963,
      "tokens_seen": 493158400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005825444486476263,
      "loss": 2.8708,
      "theoretical_loss": 3.9206218389018943,
      "tokens_seen": 493223936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005825278986192549,
      "loss": 2.7347,
      "theoretical_loss": 3.9205656009844194,
      "tokens_seen": 493289472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005825113485908832,
      "loss": 2.576,
      "theoretical_loss": 3.9205093726296405,
      "tokens_seen": 493355008
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005824947985625118,
      "loss": 2.8919,
      "theoretical_loss": 3.920453153834662,
      "tokens_seen": 493420544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005824782485341403,
      "loss": 2.8776,
      "theoretical_loss": 3.9203969445965887,
      "tokens_seen": 493486080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005824616985057688,
      "loss": 2.6035,
      "theoretical_loss": 3.9203407449125276,
      "tokens_seen": 493551616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005824451484773974,
      "loss": 2.5581,
      "theoretical_loss": 3.9202845547795864,
      "tokens_seen": 493617152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005824285984490259,
      "loss": 2.8556,
      "theoretical_loss": 3.920228374194874,
      "tokens_seen": 493682688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005824120484206543,
      "loss": 2.8478,
      "theoretical_loss": 3.9201722031555004,
      "tokens_seen": 493748224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005823954983922829,
      "loss": 2.6948,
      "theoretical_loss": 3.920116041658578,
      "tokens_seen": 493813760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005823789483639114,
      "loss": 2.9782,
      "theoretical_loss": 3.9200598897012187,
      "tokens_seen": 493879296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00058236239833554,
      "loss": 2.9393,
      "theoretical_loss": 3.9200037472805374,
      "tokens_seen": 493944832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005823458483071685,
      "loss": 2.8859,
      "theoretical_loss": 3.9199476143936485,
      "tokens_seen": 494010368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000582329298278797,
      "loss": 2.7926,
      "theoretical_loss": 3.919891491037669,
      "tokens_seen": 494075904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005823127482504256,
      "loss": 2.8616,
      "theoretical_loss": 3.9198353772097168,
      "tokens_seen": 494141440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005822961982220541,
      "loss": 2.8596,
      "theoretical_loss": 3.9197792729069114,
      "tokens_seen": 494206976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005822796481936825,
      "loss": 2.8402,
      "theoretical_loss": 3.9197231781263726,
      "tokens_seen": 494272512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005822630981653111,
      "loss": 2.685,
      "theoretical_loss": 3.919667092865222,
      "tokens_seen": 494338048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005822465481369396,
      "loss": 2.7017,
      "theoretical_loss": 3.919611017120583,
      "tokens_seen": 494403584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005822299981085682,
      "loss": 2.6409,
      "theoretical_loss": 3.919554950889579,
      "tokens_seen": 494469120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005822134480801967,
      "loss": 2.7553,
      "theoretical_loss": 3.9194988941693354,
      "tokens_seen": 494534656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005821968980518252,
      "loss": 2.6436,
      "theoretical_loss": 3.9194428469569793,
      "tokens_seen": 494600192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005821803480234538,
      "loss": 2.8503,
      "theoretical_loss": 3.9193868092496382,
      "tokens_seen": 494665728
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005821637979950822,
      "loss": 2.8945,
      "theoretical_loss": 3.919330781044441,
      "tokens_seen": 494731264
    },
    {
      "epoch": 0.18,
      "objective/train/avg_token_score": 0.012342417612671852,
      "objective/train/avg_weight": 0.9854872226715088,
      "objective/train/docs_used": 289259,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.809492588043213,
      "objective/train/std_weight": 0.07280265539884567,
      "objective/train/theoretical_loss": 3.919274762338519,
      "objective/train/tokens_used": 515256800,
      "theoretical_loss": 3.919274762338519,
      "tokens_seen": 494796800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005821472479667107,
      "loss": 2.8749,
      "theoretical_loss": 3.919274762338519,
      "tokens_seen": 494796800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005821306979383393,
      "loss": 2.9715,
      "theoretical_loss": 3.9192187531290026,
      "tokens_seen": 494862336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005821141479099678,
      "loss": 2.9817,
      "theoretical_loss": 3.9191627534130253,
      "tokens_seen": 494927872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005820975978815964,
      "loss": 2.955,
      "theoretical_loss": 3.9191067631877208,
      "tokens_seen": 494993408
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005820810478532249,
      "loss": 2.7097,
      "theoretical_loss": 3.919050782450225,
      "tokens_seen": 495058944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005820644978248534,
      "loss": 2.9256,
      "theoretical_loss": 3.9189948111976736,
      "tokens_seen": 495124480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000582047947796482,
      "loss": 2.5492,
      "theoretical_loss": 3.9189388494272057,
      "tokens_seen": 495190016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005820313977681104,
      "loss": 3.0048,
      "theoretical_loss": 3.9188828971359584,
      "tokens_seen": 495255552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005820148477397389,
      "loss": 2.8745,
      "theoretical_loss": 3.918826954321074,
      "tokens_seen": 495321088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005819982977113675,
      "loss": 2.6035,
      "theoretical_loss": 3.918771020979693,
      "tokens_seen": 495386624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000581981747682996,
      "loss": 2.722,
      "theoretical_loss": 3.9187150971089584,
      "tokens_seen": 495452160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005819651976546246,
      "loss": 2.8587,
      "theoretical_loss": 3.918659182706014,
      "tokens_seen": 495517696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005819486476262531,
      "loss": 2.9175,
      "theoretical_loss": 3.9186032777680055,
      "tokens_seen": 495583232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005819320975978816,
      "loss": 2.835,
      "theoretical_loss": 3.918547382292079,
      "tokens_seen": 495648768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005819155475695101,
      "loss": 2.8171,
      "theoretical_loss": 3.918491496275383,
      "tokens_seen": 495714304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005818989975411386,
      "loss": 2.9693,
      "theoretical_loss": 3.9184356197150647,
      "tokens_seen": 495779840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005818824475127672,
      "loss": 2.8739,
      "theoretical_loss": 3.918379752608276,
      "tokens_seen": 495845376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005818658974843957,
      "loss": 2.7976,
      "theoretical_loss": 3.918323894952168,
      "tokens_seen": 495910912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005818493474560242,
      "loss": 2.9086,
      "theoretical_loss": 3.9182680467438926,
      "tokens_seen": 495976448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005818327974276528,
      "loss": 2.789,
      "theoretical_loss": 3.9182122079806048,
      "tokens_seen": 496041984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005818162473992813,
      "loss": 2.7804,
      "theoretical_loss": 3.918156378659459,
      "tokens_seen": 496107520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005817996973709097,
      "loss": 2.9527,
      "theoretical_loss": 3.9181005587776117,
      "tokens_seen": 496173056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005817831473425383,
      "loss": 2.8098,
      "theoretical_loss": 3.9180447483322203,
      "tokens_seen": 496238592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005817665973141668,
      "loss": 2.8583,
      "theoretical_loss": 3.9179889473204446,
      "tokens_seen": 496304128
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005817500472857954,
      "loss": 2.7978,
      "theoretical_loss": 3.9179331557394432,
      "tokens_seen": 496369664
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005817334972574239,
      "loss": 2.529,
      "theoretical_loss": 3.9178773735863786,
      "tokens_seen": 496435200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005817169472290524,
      "loss": 2.9483,
      "theoretical_loss": 3.917821600858413,
      "tokens_seen": 496500736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000581700397200681,
      "loss": 2.7395,
      "theoretical_loss": 3.917765837552709,
      "tokens_seen": 496566272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005816838471723095,
      "loss": 2.7977,
      "theoretical_loss": 3.917710083666434,
      "tokens_seen": 496631808
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005816672971439379,
      "loss": 3.1654,
      "theoretical_loss": 3.917654339196752,
      "tokens_seen": 496697344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005816507471155665,
      "loss": 2.6849,
      "theoretical_loss": 3.9175986041408315,
      "tokens_seen": 496762880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000581634197087195,
      "loss": 2.8224,
      "theoretical_loss": 3.9175428784958406,
      "tokens_seen": 496828416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005816176470588236,
      "loss": 2.8701,
      "theoretical_loss": 3.9174871622589498,
      "tokens_seen": 496893952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005816010970304521,
      "loss": 2.9361,
      "theoretical_loss": 3.91743145542733,
      "tokens_seen": 496959488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005815845470020806,
      "loss": 2.8718,
      "theoretical_loss": 3.9173757579981534,
      "tokens_seen": 497025024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005815679969737092,
      "loss": 2.7895,
      "theoretical_loss": 3.9173200699685937,
      "tokens_seen": 497090560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005815514469453376,
      "loss": 2.7573,
      "theoretical_loss": 3.9172643913358254,
      "tokens_seen": 497156096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005815348969169661,
      "loss": 2.7315,
      "theoretical_loss": 3.9172087220970244,
      "tokens_seen": 497221632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005815183468885946,
      "loss": 2.7399,
      "theoretical_loss": 3.917153062249368,
      "tokens_seen": 497287168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005815017968602231,
      "loss": 2.9821,
      "theoretical_loss": 3.917097411790036,
      "tokens_seen": 497352704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005814852468318517,
      "loss": 2.9179,
      "theoretical_loss": 3.9170417707162057,
      "tokens_seen": 497418240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005814686968034802,
      "loss": 2.8077,
      "theoretical_loss": 3.9169861390250595,
      "tokens_seen": 497483776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005814521467751088,
      "loss": 3.0129,
      "theoretical_loss": 3.9169305167137796,
      "tokens_seen": 497549312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005814355967467372,
      "loss": 3.0542,
      "theoretical_loss": 3.916874903779549,
      "tokens_seen": 497614848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005814190467183657,
      "loss": 2.9225,
      "theoretical_loss": 3.9168193002195513,
      "tokens_seen": 497680384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005814024966899943,
      "loss": 2.712,
      "theoretical_loss": 3.9167637060309732,
      "tokens_seen": 497745920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005813859466616228,
      "loss": 2.936,
      "theoretical_loss": 3.916708121211002,
      "tokens_seen": 497811456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005813693966332513,
      "loss": 2.7438,
      "theoretical_loss": 3.9166525457568246,
      "tokens_seen": 497876992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005813528466048799,
      "loss": 2.6225,
      "theoretical_loss": 3.916596979665632,
      "tokens_seen": 497942528
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005813362965765084,
      "loss": 2.9487,
      "theoretical_loss": 3.9165414229346136,
      "tokens_seen": 498008064
    },
    {
      "epoch": 0.18,
      "objective/train/avg_token_score": 0.03051290474832058,
      "objective/train/avg_weight": 0.9641326665878296,
      "objective/train/docs_used": 291398,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1227211952209473,
      "objective/train/std_weight": 0.11961238831281662,
      "objective/train/theoretical_loss": 3.9164858755609613,
      "objective/train/tokens_used": 518533600,
      "theoretical_loss": 3.9164858755609613,
      "tokens_seen": 498073600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000581319746548137,
      "loss": 2.7502,
      "theoretical_loss": 3.9164858755609613,
      "tokens_seen": 498073600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005813031965197654,
      "loss": 2.8675,
      "theoretical_loss": 3.916430337541869,
      "tokens_seen": 498139136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005812866464913939,
      "loss": 2.7059,
      "theoretical_loss": 3.9163748088745303,
      "tokens_seen": 498204672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005812700964630225,
      "loss": 2.697,
      "theoretical_loss": 3.9163192895561405,
      "tokens_seen": 498270208
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000581253546434651,
      "loss": 2.9153,
      "theoretical_loss": 3.916263779583896,
      "tokens_seen": 498335744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005812369964062795,
      "loss": 2.8957,
      "theoretical_loss": 3.9162082789549952,
      "tokens_seen": 498401280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005812204463779081,
      "loss": 2.704,
      "theoretical_loss": 3.916152787666638,
      "tokens_seen": 498466816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005812038963495366,
      "loss": 2.7382,
      "theoretical_loss": 3.916097305716023,
      "tokens_seen": 498532352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005811873463211651,
      "loss": 2.9831,
      "theoretical_loss": 3.9160418331003526,
      "tokens_seen": 498597888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005811707962927936,
      "loss": 2.8332,
      "theoretical_loss": 3.91598636981683,
      "tokens_seen": 498663424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005811542462644221,
      "loss": 2.7552,
      "theoretical_loss": 3.9159309158626576,
      "tokens_seen": 498728960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005811376962360507,
      "loss": 2.6391,
      "theoretical_loss": 3.9158754712350414,
      "tokens_seen": 498794496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005811211462076792,
      "loss": 2.7392,
      "theoretical_loss": 3.915820035931188,
      "tokens_seen": 498860032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005811045961793077,
      "loss": 2.9097,
      "theoretical_loss": 3.9157646099483046,
      "tokens_seen": 498925568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005810880461509363,
      "loss": 2.7184,
      "theoretical_loss": 3.9157091932836,
      "tokens_seen": 498991104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005810714961225647,
      "loss": 2.8525,
      "theoretical_loss": 3.9156537859342837,
      "tokens_seen": 499056640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005810549460941933,
      "loss": 2.8617,
      "theoretical_loss": 3.9155983878975675,
      "tokens_seen": 499122176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005810383960658218,
      "loss": 3.0386,
      "theoretical_loss": 3.9155429991706634,
      "tokens_seen": 499187712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005810218460374503,
      "loss": 3.0977,
      "theoretical_loss": 3.9154876197507846,
      "tokens_seen": 499253248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005810052960090789,
      "loss": 2.6522,
      "theoretical_loss": 3.915432249635147,
      "tokens_seen": 499318784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005809887459807074,
      "loss": 3.0653,
      "theoretical_loss": 3.9153768888209646,
      "tokens_seen": 499384320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005809721959523359,
      "loss": 2.7475,
      "theoretical_loss": 3.915321537305456,
      "tokens_seen": 499449856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005809556459239645,
      "loss": 3.0389,
      "theoretical_loss": 3.915266195085839,
      "tokens_seen": 499515392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005809390958955929,
      "loss": 2.8245,
      "theoretical_loss": 3.9152108621593333,
      "tokens_seen": 499580928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005809225458672215,
      "loss": 2.8743,
      "theoretical_loss": 3.91515553852316,
      "tokens_seen": 499646464
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00058090599583885,
      "loss": 2.6459,
      "theoretical_loss": 3.91510022417454,
      "tokens_seen": 499712000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005808894458104785,
      "loss": 2.8952,
      "theoretical_loss": 3.915044919110697,
      "tokens_seen": 499777536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005808728957821071,
      "loss": 2.9836,
      "theoretical_loss": 3.9149896233288555,
      "tokens_seen": 499843072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005808563457537356,
      "loss": 2.9794,
      "theoretical_loss": 3.9149343368262404,
      "tokens_seen": 499908608
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005808397957253641,
      "loss": 2.7914,
      "theoretical_loss": 3.9148790596000786,
      "tokens_seen": 499974144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005808232456969927,
      "loss": 2.747,
      "theoretical_loss": 3.9148237916475983,
      "tokens_seen": 500039680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005808066956686211,
      "loss": 2.7698,
      "theoretical_loss": 3.9147685329660282,
      "tokens_seen": 500105216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005807901456402497,
      "loss": 2.7886,
      "theoretical_loss": 3.914713283552599,
      "tokens_seen": 500170752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005807735956118782,
      "loss": 2.7427,
      "theoretical_loss": 3.9146580434045415,
      "tokens_seen": 500236288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005807570455835067,
      "loss": 2.8811,
      "theoretical_loss": 3.914602812519089,
      "tokens_seen": 500301824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005807404955551353,
      "loss": 2.7477,
      "theoretical_loss": 3.9145475908934753,
      "tokens_seen": 500367360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005807239455267638,
      "loss": 2.9889,
      "theoretical_loss": 3.914492378524934,
      "tokens_seen": 500432896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005807073954983923,
      "loss": 2.7148,
      "theoretical_loss": 3.9144371754107032,
      "tokens_seen": 500498432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005806908454700208,
      "loss": 2.4607,
      "theoretical_loss": 3.9143819815480194,
      "tokens_seen": 500563968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005806742954416493,
      "loss": 2.8423,
      "theoretical_loss": 3.9143267969341218,
      "tokens_seen": 500629504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005806577454132779,
      "loss": 2.8392,
      "theoretical_loss": 3.9142716215662494,
      "tokens_seen": 500695040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005806411953849064,
      "loss": 2.9545,
      "theoretical_loss": 3.9142164554416428,
      "tokens_seen": 500760576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005806246453565349,
      "loss": 2.7856,
      "theoretical_loss": 3.9141612985575454,
      "tokens_seen": 500826112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005806080953281635,
      "loss": 2.9663,
      "theoretical_loss": 3.9141061509111994,
      "tokens_seen": 500891648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000580591545299792,
      "loss": 3.04,
      "theoretical_loss": 3.91405101249985,
      "tokens_seen": 500957184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005805749952714204,
      "loss": 2.6985,
      "theoretical_loss": 3.9139958833207427,
      "tokens_seen": 501022720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000580558445243049,
      "loss": 2.9126,
      "theoretical_loss": 3.9139407633711243,
      "tokens_seen": 501088256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005805418952146775,
      "loss": 2.842,
      "theoretical_loss": 3.9138856526482426,
      "tokens_seen": 501153792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000580525345186306,
      "loss": 2.9264,
      "theoretical_loss": 3.9138305511493474,
      "tokens_seen": 501219328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005805087951579345,
      "loss": 2.4606,
      "theoretical_loss": 3.9137754588716884,
      "tokens_seen": 501284864
    },
    {
      "epoch": 0.18,
      "objective/train/avg_token_score": 0.033610619604587555,
      "objective/train/avg_weight": 0.9604612588882446,
      "objective/train/docs_used": 293236,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7664332389831543,
      "objective/train/std_weight": 0.1465364694595337,
      "objective/train/theoretical_loss": 3.9137203758125176,
      "objective/train/tokens_used": 521810400,
      "theoretical_loss": 3.9137203758125176,
      "tokens_seen": 501350400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005804922451295631,
      "loss": 2.5537,
      "theoretical_loss": 3.9137203758125176,
      "tokens_seen": 501350400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005804756951011916,
      "loss": 2.6653,
      "theoretical_loss": 3.913665301969088,
      "tokens_seen": 501415936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00058045914507282,
      "loss": 2.931,
      "theoretical_loss": 3.913610237338653,
      "tokens_seen": 501481472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005804425950444486,
      "loss": 3.097,
      "theoretical_loss": 3.9135551819184675,
      "tokens_seen": 501547008
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005804260450160771,
      "loss": 2.6894,
      "theoretical_loss": 3.913500135705789,
      "tokens_seen": 501612544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005804094949877057,
      "loss": 2.6764,
      "theoretical_loss": 3.9134450986978737,
      "tokens_seen": 501678080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005803929449593342,
      "loss": 2.9632,
      "theoretical_loss": 3.913390070891981,
      "tokens_seen": 501743616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005803763949309627,
      "loss": 2.6631,
      "theoretical_loss": 3.9133350522853703,
      "tokens_seen": 501809152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005803598449025913,
      "loss": 2.7191,
      "theoretical_loss": 3.9132800428753027,
      "tokens_seen": 501874688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005803432948742198,
      "loss": 2.7657,
      "theoretical_loss": 3.9132250426590405,
      "tokens_seen": 501940224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005803267448458482,
      "loss": 2.8536,
      "theoretical_loss": 3.9131700516338475,
      "tokens_seen": 502005760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005803101948174768,
      "loss": 3.087,
      "theoretical_loss": 3.9131150697969868,
      "tokens_seen": 502071296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005802936447891053,
      "loss": 2.71,
      "theoretical_loss": 3.913060097145725,
      "tokens_seen": 502136832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005802770947607339,
      "loss": 2.6979,
      "theoretical_loss": 3.913005133677329,
      "tokens_seen": 502202368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005802605447323624,
      "loss": 2.7452,
      "theoretical_loss": 3.9129501793890666,
      "tokens_seen": 502267904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005802439947039909,
      "loss": 2.6876,
      "theoretical_loss": 3.9128952342782073,
      "tokens_seen": 502333440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005802274446756195,
      "loss": 2.9632,
      "theoretical_loss": 3.9128402983420214,
      "tokens_seen": 502398976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005802108946472479,
      "loss": 2.9438,
      "theoretical_loss": 3.91278537157778,
      "tokens_seen": 502464512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005801943446188764,
      "loss": 2.8683,
      "theoretical_loss": 3.912730453982756,
      "tokens_seen": 502530048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000580177794590505,
      "loss": 2.8909,
      "theoretical_loss": 3.9126755455542233,
      "tokens_seen": 502595584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005801612445621335,
      "loss": 2.3967,
      "theoretical_loss": 3.912620646289457,
      "tokens_seen": 502661120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005801446945337621,
      "loss": 2.8187,
      "theoretical_loss": 3.9125657561857334,
      "tokens_seen": 502726656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005801281445053906,
      "loss": 2.8924,
      "theoretical_loss": 3.9125108752403293,
      "tokens_seen": 502792192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005801115944770191,
      "loss": 2.4629,
      "theoretical_loss": 3.9124560034505236,
      "tokens_seen": 502857728
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005800950444486477,
      "loss": 2.9451,
      "theoretical_loss": 3.9124011408135964,
      "tokens_seen": 502923264
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005800784944202761,
      "loss": 2.5594,
      "theoretical_loss": 3.9123462873268275,
      "tokens_seen": 502988800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005800619443919046,
      "loss": 2.7981,
      "theoretical_loss": 3.9122914429874998,
      "tokens_seen": 503054336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005800453943635332,
      "loss": 2.8984,
      "theoretical_loss": 3.912236607792896,
      "tokens_seen": 503119872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005800288443351617,
      "loss": 2.9132,
      "theoretical_loss": 3.9121817817403004,
      "tokens_seen": 503185408
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005800122943067903,
      "loss": 2.7013,
      "theoretical_loss": 3.912126964826999,
      "tokens_seen": 503250944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005799957442784188,
      "loss": 2.6288,
      "theoretical_loss": 3.912072157050278,
      "tokens_seen": 503316480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005799791942500473,
      "loss": 2.966,
      "theoretical_loss": 3.9120173584074256,
      "tokens_seen": 503382016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005799626442216758,
      "loss": 2.8063,
      "theoretical_loss": 3.91196256889573,
      "tokens_seen": 503447552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005799460941933043,
      "loss": 2.5069,
      "theoretical_loss": 3.911907788512482,
      "tokens_seen": 503513088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005799295441649328,
      "loss": 2.9357,
      "theoretical_loss": 3.911853017254973,
      "tokens_seen": 503578624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005799129941365614,
      "loss": 2.7822,
      "theoretical_loss": 3.9117982551204946,
      "tokens_seen": 503644160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005798964441081899,
      "loss": 2.7334,
      "theoretical_loss": 3.911743502106341,
      "tokens_seen": 503709696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005798798940798185,
      "loss": 2.8099,
      "theoretical_loss": 3.911688758209807,
      "tokens_seen": 503775232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000579863344051447,
      "loss": 2.4698,
      "theoretical_loss": 3.9116340234281886,
      "tokens_seen": 503840768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005798467940230754,
      "loss": 2.8112,
      "theoretical_loss": 3.911579297758782,
      "tokens_seen": 503906304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000579830243994704,
      "loss": 2.6349,
      "theoretical_loss": 3.9115245811988864,
      "tokens_seen": 503971840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005798136939663325,
      "loss": 2.6672,
      "theoretical_loss": 3.9114698737458005,
      "tokens_seen": 504037376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000579797143937961,
      "loss": 2.9185,
      "theoretical_loss": 3.911415175396825,
      "tokens_seen": 504102912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005797805939095896,
      "loss": 2.9901,
      "theoretical_loss": 3.911360486149262,
      "tokens_seen": 504168448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005797640438812181,
      "loss": 2.6167,
      "theoretical_loss": 3.9113058060004136,
      "tokens_seen": 504233984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005797474938528467,
      "loss": 3.0373,
      "theoretical_loss": 3.9112511349475843,
      "tokens_seen": 504299520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005797309438244752,
      "loss": 2.8614,
      "theoretical_loss": 3.911196472988079,
      "tokens_seen": 504365056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005797143937961036,
      "loss": 2.7792,
      "theoretical_loss": 3.9111418201192034,
      "tokens_seen": 504430592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005796978437677322,
      "loss": 2.7488,
      "theoretical_loss": 3.911087176338266,
      "tokens_seen": 504496128
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005796812937393607,
      "loss": 2.8627,
      "theoretical_loss": 3.911032541642575,
      "tokens_seen": 504561664
    },
    {
      "epoch": 0.18,
      "objective/train/avg_token_score": 0.03031834214925766,
      "objective/train/avg_weight": 0.9643309116363525,
      "objective/train/docs_used": 295292,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.767930030822754,
      "objective/train/std_weight": 0.13676677644252777,
      "objective/train/theoretical_loss": 3.910977916029439,
      "objective/train/tokens_used": 525087200,
      "theoretical_loss": 3.910977916029439,
      "tokens_seen": 504627200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005796647437109892,
      "loss": 2.7652,
      "theoretical_loss": 3.910977916029439,
      "tokens_seen": 504627200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005796481936826178,
      "loss": 2.8694,
      "theoretical_loss": 3.9109232994961705,
      "tokens_seen": 504692736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005796316436542463,
      "loss": 2.8872,
      "theoretical_loss": 3.91086869204008,
      "tokens_seen": 504758272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005796150936258749,
      "loss": 2.6634,
      "theoretical_loss": 3.910814093658482,
      "tokens_seen": 504823808
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005795985435975033,
      "loss": 2.8003,
      "theoretical_loss": 3.9107595043486905,
      "tokens_seen": 504889344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005795819935691318,
      "loss": 2.6323,
      "theoretical_loss": 3.91070492410802,
      "tokens_seen": 504954880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005795654435407604,
      "loss": 2.944,
      "theoretical_loss": 3.9106503529337875,
      "tokens_seen": 505020416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005795488935123888,
      "loss": 2.6875,
      "theoretical_loss": 3.9105957908233107,
      "tokens_seen": 505085952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005795323434840174,
      "loss": 2.8184,
      "theoretical_loss": 3.910541237773909,
      "tokens_seen": 505151488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005795157934556459,
      "loss": 2.8294,
      "theoretical_loss": 3.910486693782902,
      "tokens_seen": 505217024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005794992434272745,
      "loss": 2.7087,
      "theoretical_loss": 3.9104321588476108,
      "tokens_seen": 505282560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005794826933989029,
      "loss": 3.0108,
      "theoretical_loss": 3.910377632965357,
      "tokens_seen": 505348096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005794661433705314,
      "loss": 2.6596,
      "theoretical_loss": 3.9103231161334655,
      "tokens_seen": 505413632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00057944959334216,
      "loss": 2.7729,
      "theoretical_loss": 3.9102686083492593,
      "tokens_seen": 505479168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005794330433137885,
      "loss": 2.7308,
      "theoretical_loss": 3.910214109610065,
      "tokens_seen": 505544704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000579416493285417,
      "loss": 2.6026,
      "theoretical_loss": 3.910159619913209,
      "tokens_seen": 505610240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005793999432570456,
      "loss": 2.9632,
      "theoretical_loss": 3.9101051392560198,
      "tokens_seen": 505675776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005793833932286741,
      "loss": 2.631,
      "theoretical_loss": 3.9100506676358266,
      "tokens_seen": 505741312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005793668432003027,
      "loss": 2.7145,
      "theoretical_loss": 3.9099962050499584,
      "tokens_seen": 505806848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005793502931719311,
      "loss": 2.8627,
      "theoretical_loss": 3.9099417514957473,
      "tokens_seen": 505872384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005793337431435596,
      "loss": 2.8017,
      "theoretical_loss": 3.9098873069705267,
      "tokens_seen": 505937920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005793171931151882,
      "loss": 2.9176,
      "theoretical_loss": 3.9098328714716284,
      "tokens_seen": 506003456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005793006430868167,
      "loss": 2.7036,
      "theoretical_loss": 3.909778444996389,
      "tokens_seen": 506068992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005792840930584452,
      "loss": 2.6892,
      "theoretical_loss": 3.9097240275421425,
      "tokens_seen": 506134528
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005792675430300738,
      "loss": 2.9341,
      "theoretical_loss": 3.9096696191062277,
      "tokens_seen": 506200064
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005792509930017023,
      "loss": 2.9712,
      "theoretical_loss": 3.909615219685982,
      "tokens_seen": 506265600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005792344429733308,
      "loss": 2.927,
      "theoretical_loss": 3.9095608292787443,
      "tokens_seen": 506331136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005792178929449593,
      "loss": 2.84,
      "theoretical_loss": 3.9095064478818555,
      "tokens_seen": 506396672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005792013429165878,
      "loss": 2.6812,
      "theoretical_loss": 3.9094520754926574,
      "tokens_seen": 506462208
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005791847928882164,
      "loss": 2.751,
      "theoretical_loss": 3.9093977121084915,
      "tokens_seen": 506527744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005791682428598449,
      "loss": 2.8553,
      "theoretical_loss": 3.909343357726703,
      "tokens_seen": 506593280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005791516928314734,
      "loss": 2.939,
      "theoretical_loss": 3.9092890123446367,
      "tokens_seen": 506658816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000579135142803102,
      "loss": 2.9267,
      "theoretical_loss": 3.9092346759596373,
      "tokens_seen": 506724352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005791185927747304,
      "loss": 2.512,
      "theoretical_loss": 3.9091803485690537,
      "tokens_seen": 506789888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000579102042746359,
      "loss": 2.9833,
      "theoretical_loss": 3.9091260301702326,
      "tokens_seen": 506855424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005790854927179875,
      "loss": 2.9228,
      "theoretical_loss": 3.9090717207605246,
      "tokens_seen": 506920960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000579068942689616,
      "loss": 2.7082,
      "theoretical_loss": 3.90901742033728,
      "tokens_seen": 506986496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005790523926612446,
      "loss": 2.8369,
      "theoretical_loss": 3.9089631288978497,
      "tokens_seen": 507052032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005790358426328731,
      "loss": 2.9071,
      "theoretical_loss": 3.9089088464395876,
      "tokens_seen": 507117568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005790192926045016,
      "loss": 2.6849,
      "theoretical_loss": 3.908854572959847,
      "tokens_seen": 507183104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005790027425761302,
      "loss": 2.9706,
      "theoretical_loss": 3.9088003084559833,
      "tokens_seen": 507248640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005789861925477586,
      "loss": 2.6893,
      "theoretical_loss": 3.908746052925352,
      "tokens_seen": 507314176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005789696425193872,
      "loss": 2.5844,
      "theoretical_loss": 3.9086918063653115,
      "tokens_seen": 507379712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005789530924910157,
      "loss": 2.643,
      "theoretical_loss": 3.908637568773219,
      "tokens_seen": 507445248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005789365424626442,
      "loss": 2.7871,
      "theoretical_loss": 3.9085833401464347,
      "tokens_seen": 507510784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005789199924342728,
      "loss": 2.82,
      "theoretical_loss": 3.908529120482319,
      "tokens_seen": 507576320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005789034424059013,
      "loss": 2.6854,
      "theoretical_loss": 3.908474909778234,
      "tokens_seen": 507641856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005788868923775298,
      "loss": 2.724,
      "theoretical_loss": 3.908420708031542,
      "tokens_seen": 507707392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005788703423491583,
      "loss": 2.789,
      "theoretical_loss": 3.9083665152396074,
      "tokens_seen": 507772928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005788537923207868,
      "loss": 2.9543,
      "theoretical_loss": 3.908312331399795,
      "tokens_seen": 507838464
    },
    {
      "epoch": 0.18,
      "objective/train/avg_token_score": 0.012849971652030945,
      "objective/train/avg_weight": 0.9848799109458923,
      "objective/train/docs_used": 297338,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9367456436157227,
      "objective/train/std_weight": 0.057525064796209335,
      "objective/train/theoretical_loss": 3.908258156509472,
      "objective/train/tokens_used": 528364000,
      "theoretical_loss": 3.908258156509472,
      "tokens_seen": 507904000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005788372422924154,
      "loss": 2.823,
      "theoretical_loss": 3.908258156509472,
      "tokens_seen": 507904000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005788206922640439,
      "loss": 2.7956,
      "theoretical_loss": 3.9082039905660046,
      "tokens_seen": 507969536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005788041422356724,
      "loss": 2.7049,
      "theoretical_loss": 3.9081498335667613,
      "tokens_seen": 508035072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000578787592207301,
      "loss": 2.8,
      "theoretical_loss": 3.908095685509112,
      "tokens_seen": 508100608
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005787710421789295,
      "loss": 2.7821,
      "theoretical_loss": 3.908041546390428,
      "tokens_seen": 508166144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000578754492150558,
      "loss": 3.0248,
      "theoretical_loss": 3.9079874162080794,
      "tokens_seen": 508231680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005787379421221865,
      "loss": 2.7902,
      "theoretical_loss": 3.9079332949594408,
      "tokens_seen": 508297216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000578721392093815,
      "loss": 2.7964,
      "theoretical_loss": 3.907879182641886,
      "tokens_seen": 508362752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005787048420654436,
      "loss": 2.9479,
      "theoretical_loss": 3.907825079252789,
      "tokens_seen": 508428288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005786882920370721,
      "loss": 2.6633,
      "theoretical_loss": 3.9077709847895266,
      "tokens_seen": 508493824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005786717420087006,
      "loss": 2.6177,
      "theoretical_loss": 3.907716899249477,
      "tokens_seen": 508559360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005786551919803292,
      "loss": 2.6801,
      "theoretical_loss": 3.9076628226300167,
      "tokens_seen": 508624896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005786386419519577,
      "loss": 2.9231,
      "theoretical_loss": 3.9076087549285274,
      "tokens_seen": 508690432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005786220919235861,
      "loss": 2.9888,
      "theoretical_loss": 3.907554696142389,
      "tokens_seen": 508755968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005786055418952147,
      "loss": 2.7088,
      "theoretical_loss": 3.9075006462689825,
      "tokens_seen": 508821504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005785889918668432,
      "loss": 2.7953,
      "theoretical_loss": 3.9074466053056915,
      "tokens_seen": 508887040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005785724418384717,
      "loss": 2.9275,
      "theoretical_loss": 3.9073925732498997,
      "tokens_seen": 508952576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005785558918101002,
      "loss": 2.8256,
      "theoretical_loss": 3.907338550098993,
      "tokens_seen": 509018112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005785393417817288,
      "loss": 2.7605,
      "theoretical_loss": 3.907284535850356,
      "tokens_seen": 509083648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005785227917533573,
      "loss": 2.8958,
      "theoretical_loss": 3.9072305305013773,
      "tokens_seen": 509149184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005785062417249857,
      "loss": 2.7276,
      "theoretical_loss": 3.9071765340494453,
      "tokens_seen": 509214720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005784896916966143,
      "loss": 2.9723,
      "theoretical_loss": 3.9071225464919483,
      "tokens_seen": 509280256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005784731416682428,
      "loss": 2.6182,
      "theoretical_loss": 3.907068567826278,
      "tokens_seen": 509345792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005784565916398713,
      "loss": 2.7009,
      "theoretical_loss": 3.9070145980498263,
      "tokens_seen": 509411328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005784400416114999,
      "loss": 2.9164,
      "theoretical_loss": 3.906960637159985,
      "tokens_seen": 509476864
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005784234915831284,
      "loss": 2.8719,
      "theoretical_loss": 3.9069066851541487,
      "tokens_seen": 509542400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000578406941554757,
      "loss": 2.6692,
      "theoretical_loss": 3.906852742029712,
      "tokens_seen": 509607936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005783903915263854,
      "loss": 2.6401,
      "theoretical_loss": 3.906798807784072,
      "tokens_seen": 509673472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005783738414980139,
      "loss": 2.8311,
      "theoretical_loss": 3.9067448824146243,
      "tokens_seen": 509739008
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005783572914696425,
      "loss": 2.7646,
      "theoretical_loss": 3.9066909659187683,
      "tokens_seen": 509804544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000578340741441271,
      "loss": 2.5074,
      "theoretical_loss": 3.906637058293903,
      "tokens_seen": 509870080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005783241914128995,
      "loss": 2.8395,
      "theoretical_loss": 3.906583159537429,
      "tokens_seen": 509935616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005783076413845281,
      "loss": 2.9071,
      "theoretical_loss": 3.9065292696467475,
      "tokens_seen": 510001152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005782910913561566,
      "loss": 2.7119,
      "theoretical_loss": 3.906475388619262,
      "tokens_seen": 510066688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005782745413277852,
      "loss": 2.8652,
      "theoretical_loss": 3.906421516452376,
      "tokens_seen": 510132224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005782579912994136,
      "loss": 2.8052,
      "theoretical_loss": 3.9063676531434934,
      "tokens_seen": 510197760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005782414412710421,
      "loss": 2.9053,
      "theoretical_loss": 3.9063137986900216,
      "tokens_seen": 510263296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005782248912426707,
      "loss": 2.7732,
      "theoretical_loss": 3.906259953089367,
      "tokens_seen": 510328832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005782083412142992,
      "loss": 2.7681,
      "theoretical_loss": 3.9062061163389377,
      "tokens_seen": 510394368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005781917911859277,
      "loss": 2.9286,
      "theoretical_loss": 3.906152288436143,
      "tokens_seen": 510459904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005781752411575563,
      "loss": 2.8339,
      "theoretical_loss": 3.906098469378393,
      "tokens_seen": 510525440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005781586911291848,
      "loss": 2.724,
      "theoretical_loss": 3.9060446591630997,
      "tokens_seen": 510590976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005781421411008133,
      "loss": 2.7055,
      "theoretical_loss": 3.905990857787675,
      "tokens_seen": 510656512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005781255910724418,
      "loss": 2.9272,
      "theoretical_loss": 3.9059370652495327,
      "tokens_seen": 510722048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005781090410440703,
      "loss": 2.8501,
      "theoretical_loss": 3.905883281546088,
      "tokens_seen": 510787584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005780924910156989,
      "loss": 2.6832,
      "theoretical_loss": 3.9058295066747553,
      "tokens_seen": 510853120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005780759409873274,
      "loss": 2.8755,
      "theoretical_loss": 3.9057757406329534,
      "tokens_seen": 510918656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005780593909589559,
      "loss": 2.7774,
      "theoretical_loss": 3.9057219834180987,
      "tokens_seen": 510984192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005780428409305845,
      "loss": 2.7164,
      "theoretical_loss": 3.9056682350276106,
      "tokens_seen": 511049728
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000578026290902213,
      "loss": 2.8218,
      "theoretical_loss": 3.90561449545891,
      "tokens_seen": 511115264
    },
    {
      "epoch": 0.18,
      "objective/train/avg_token_score": 0.03070048987865448,
      "objective/train/avg_weight": 0.9638665914535522,
      "objective/train/docs_used": 299351,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.804051399230957,
      "objective/train/std_weight": 0.12230520695447922,
      "objective/train/theoretical_loss": 3.905560764709417,
      "objective/train/tokens_used": 531640800,
      "theoretical_loss": 3.905560764709417,
      "tokens_seen": 511180800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000578026290902213,
      "loss": 2.4895,
      "theoretical_loss": 3.905560764709417,
      "tokens_seen": 511180800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005780097408738415,
      "loss": 2.9015,
      "theoretical_loss": 3.905507042776554,
      "tokens_seen": 511246336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00057799319084547,
      "loss": 2.9555,
      "theoretical_loss": 3.905453329657745,
      "tokens_seen": 511311872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005779766408170985,
      "loss": 2.959,
      "theoretical_loss": 3.905399625350414,
      "tokens_seen": 511377408
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005779600907887271,
      "loss": 2.9414,
      "theoretical_loss": 3.9053459298519875,
      "tokens_seen": 511442944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005779435407603556,
      "loss": 2.6219,
      "theoretical_loss": 3.905292243159891,
      "tokens_seen": 511508480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005779269907319842,
      "loss": 2.6984,
      "theoretical_loss": 3.905238565271553,
      "tokens_seen": 511574016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005779104407036127,
      "loss": 3.1077,
      "theoretical_loss": 3.9051848961844007,
      "tokens_seen": 511639552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005778938906752411,
      "loss": 2.7671,
      "theoretical_loss": 3.9051312358958663,
      "tokens_seen": 511705088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005778773406468697,
      "loss": 2.8795,
      "theoretical_loss": 3.9050775844033785,
      "tokens_seen": 511770624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005778607906184982,
      "loss": 2.7583,
      "theoretical_loss": 3.905023941704371,
      "tokens_seen": 511836160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005778442405901267,
      "loss": 2.8272,
      "theoretical_loss": 3.904970307796276,
      "tokens_seen": 511901696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005778276905617553,
      "loss": 2.7322,
      "theoretical_loss": 3.9049166826765283,
      "tokens_seen": 511967232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005778111405333838,
      "loss": 2.6321,
      "theoretical_loss": 3.9048630663425623,
      "tokens_seen": 512032768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005777945905050124,
      "loss": 2.6234,
      "theoretical_loss": 3.9048094587918154,
      "tokens_seen": 512098304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005777780404766408,
      "loss": 2.834,
      "theoretical_loss": 3.904755860021724,
      "tokens_seen": 512163840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005777614904482693,
      "loss": 2.8534,
      "theoretical_loss": 3.904702270029727,
      "tokens_seen": 512229376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005777449404198979,
      "loss": 2.935,
      "theoretical_loss": 3.904648688813264,
      "tokens_seen": 512294912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005777283903915264,
      "loss": 2.6973,
      "theoretical_loss": 3.9045951163697756,
      "tokens_seen": 512360448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005777118403631549,
      "loss": 2.56,
      "theoretical_loss": 3.9045415526967036,
      "tokens_seen": 512425984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005776952903347835,
      "loss": 2.9867,
      "theoretical_loss": 3.904487997791491,
      "tokens_seen": 512491520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000577678740306412,
      "loss": 2.6288,
      "theoretical_loss": 3.904434451651581,
      "tokens_seen": 512557056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005776621902780406,
      "loss": 2.7662,
      "theoretical_loss": 3.904380914274419,
      "tokens_seen": 512622592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000577645640249669,
      "loss": 2.7755,
      "theoretical_loss": 3.9043273856574507,
      "tokens_seen": 512688128
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005776290902212975,
      "loss": 2.8357,
      "theoretical_loss": 3.9042738657981237,
      "tokens_seen": 512753664
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005776125401929261,
      "loss": 2.723,
      "theoretical_loss": 3.9042203546938854,
      "tokens_seen": 512819200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005775959901645545,
      "loss": 3.0607,
      "theoretical_loss": 3.9041668523421853,
      "tokens_seen": 512884736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005775794401361831,
      "loss": 2.7822,
      "theoretical_loss": 3.9041133587404744,
      "tokens_seen": 512950272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005775628901078116,
      "loss": 3.2053,
      "theoretical_loss": 3.9040598738862027,
      "tokens_seen": 513015808
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00057754634007944,
      "loss": 2.6492,
      "theoretical_loss": 3.904006397776823,
      "tokens_seen": 513081344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005775297900510686,
      "loss": 2.5182,
      "theoretical_loss": 3.90395293040979,
      "tokens_seen": 513146880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005775132400226971,
      "loss": 2.9389,
      "theoretical_loss": 3.903899471782557,
      "tokens_seen": 513212416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005774966899943257,
      "loss": 2.8802,
      "theoretical_loss": 3.90384602189258,
      "tokens_seen": 513277952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005774801399659542,
      "loss": 2.8655,
      "theoretical_loss": 3.9037925807373153,
      "tokens_seen": 513343488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005774635899375827,
      "loss": 2.8618,
      "theoretical_loss": 3.9037391483142208,
      "tokens_seen": 513409024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005774470399092113,
      "loss": 3.0095,
      "theoretical_loss": 3.9036857246207557,
      "tokens_seen": 513474560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005774304898808398,
      "loss": 2.6902,
      "theoretical_loss": 3.9036323096543795,
      "tokens_seen": 513540096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005774139398524682,
      "loss": 2.6859,
      "theoretical_loss": 3.903578903412553,
      "tokens_seen": 513605632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005773973898240968,
      "loss": 2.8986,
      "theoretical_loss": 3.9035255058927394,
      "tokens_seen": 513671168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005773808397957253,
      "loss": 2.6741,
      "theoretical_loss": 3.9034721170924005,
      "tokens_seen": 513736704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005773642897673539,
      "loss": 2.6872,
      "theoretical_loss": 3.903418737009,
      "tokens_seen": 513802240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005773477397389824,
      "loss": 2.7315,
      "theoretical_loss": 3.9033653656400045,
      "tokens_seen": 513867776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005773311897106109,
      "loss": 2.993,
      "theoretical_loss": 3.903312002982879,
      "tokens_seen": 513933312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005773146396822395,
      "loss": 2.9706,
      "theoretical_loss": 3.9032586490350916,
      "tokens_seen": 513998848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000577298089653868,
      "loss": 2.7469,
      "theoretical_loss": 3.9032053037941106,
      "tokens_seen": 514064384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005772815396254964,
      "loss": 2.9513,
      "theoretical_loss": 3.9031519672574047,
      "tokens_seen": 514129920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000577264989597125,
      "loss": 2.6608,
      "theoretical_loss": 3.903098639422445,
      "tokens_seen": 514195456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005772484395687535,
      "loss": 2.7036,
      "theoretical_loss": 3.9030453202867035,
      "tokens_seen": 514260992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005772318895403821,
      "loss": 2.7508,
      "theoretical_loss": 3.902992009847652,
      "tokens_seen": 514326528
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005772153395120106,
      "loss": 2.9422,
      "theoretical_loss": 3.9029387081027638,
      "tokens_seen": 514392064
    },
    {
      "epoch": 0.18,
      "objective/train/avg_token_score": 0.006388853769749403,
      "objective/train/avg_weight": 0.9924840927124023,
      "objective/train/docs_used": 301377,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8189711570739746,
      "objective/train/std_weight": 0.032740019261837006,
      "objective/train/theoretical_loss": 3.9028854150495143,
      "objective/train/tokens_used": 534917600,
      "theoretical_loss": 3.9028854150495143,
      "tokens_seen": 514457600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005771987894836391,
      "loss": 2.9181,
      "theoretical_loss": 3.9028854150495143,
      "tokens_seen": 514457600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005771822394552677,
      "loss": 2.6512,
      "theoretical_loss": 3.9028321306853795,
      "tokens_seen": 514523136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005771656894268961,
      "loss": 2.8715,
      "theoretical_loss": 3.9027788550078357,
      "tokens_seen": 514588672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005771491393985246,
      "loss": 2.5765,
      "theoretical_loss": 3.902725588014361,
      "tokens_seen": 514654208
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005771325893701532,
      "loss": 2.9101,
      "theoretical_loss": 3.9026723297024337,
      "tokens_seen": 514719744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005771160393417817,
      "loss": 2.9529,
      "theoretical_loss": 3.9026190800695346,
      "tokens_seen": 514785280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005770994893134103,
      "loss": 2.8383,
      "theoretical_loss": 3.9025658391131444,
      "tokens_seen": 514850816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005770829392850388,
      "loss": 2.8548,
      "theoretical_loss": 3.902512606830745,
      "tokens_seen": 514916352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005770663892566673,
      "loss": 2.7156,
      "theoretical_loss": 3.9024593832198198,
      "tokens_seen": 514981888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005770498392282958,
      "loss": 2.7533,
      "theoretical_loss": 3.902406168277853,
      "tokens_seen": 515047424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005770332891999243,
      "loss": 2.7707,
      "theoretical_loss": 3.9023529620023294,
      "tokens_seen": 515112960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005770167391715528,
      "loss": 2.8518,
      "theoretical_loss": 3.902299764390736,
      "tokens_seen": 515178496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005770001891431814,
      "loss": 2.7765,
      "theoretical_loss": 3.902246575440559,
      "tokens_seen": 515244032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005769836391148099,
      "loss": 2.9582,
      "theoretical_loss": 3.902193395149288,
      "tokens_seen": 515309568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005769670890864385,
      "loss": 2.8394,
      "theoretical_loss": 3.9021402235144116,
      "tokens_seen": 515375104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000576950539058067,
      "loss": 2.7961,
      "theoretical_loss": 3.9020870605334204,
      "tokens_seen": 515440640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005769339890296954,
      "loss": 2.7967,
      "theoretical_loss": 3.9020339062038065,
      "tokens_seen": 515506176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000576917439001324,
      "loss": 2.7869,
      "theoretical_loss": 3.9019807605230614,
      "tokens_seen": 515571712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005769008889729525,
      "loss": 2.7768,
      "theoretical_loss": 3.90192762348868,
      "tokens_seen": 515637248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000576884338944581,
      "loss": 2.6999,
      "theoretical_loss": 3.9018744950981556,
      "tokens_seen": 515702784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005768677889162096,
      "loss": 2.8505,
      "theoretical_loss": 3.901821375348985,
      "tokens_seen": 515768320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005768512388878381,
      "loss": 2.5052,
      "theoretical_loss": 3.9017682642386644,
      "tokens_seen": 515833856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005768346888594667,
      "loss": 3.065,
      "theoretical_loss": 3.9017151617646917,
      "tokens_seen": 515899392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005768181388310952,
      "loss": 2.9436,
      "theoretical_loss": 3.901662067924566,
      "tokens_seen": 515964928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005768015888027236,
      "loss": 2.7676,
      "theoretical_loss": 3.9016089827157865,
      "tokens_seen": 516030464
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005767850387743522,
      "loss": 2.8197,
      "theoretical_loss": 3.9015559061358545,
      "tokens_seen": 516096000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005767684887459807,
      "loss": 2.5636,
      "theoretical_loss": 3.901502838182272,
      "tokens_seen": 516161536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005767519387176092,
      "loss": 2.8389,
      "theoretical_loss": 3.901449778852542,
      "tokens_seen": 516227072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005767353886892378,
      "loss": 3.2529,
      "theoretical_loss": 3.9013967281441677,
      "tokens_seen": 516292608
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005767188386608663,
      "loss": 3.176,
      "theoretical_loss": 3.9013436860546555,
      "tokens_seen": 516358144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005767022886324949,
      "loss": 2.6535,
      "theoretical_loss": 3.901290652581511,
      "tokens_seen": 516423680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005766857386041234,
      "loss": 3.0913,
      "theoretical_loss": 3.901237627722241,
      "tokens_seen": 516489216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005766691885757518,
      "loss": 2.8262,
      "theoretical_loss": 3.9011846114743536,
      "tokens_seen": 516554752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005766526385473804,
      "loss": 2.7999,
      "theoretical_loss": 3.901131603835359,
      "tokens_seen": 516620288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005766360885190089,
      "loss": 2.6254,
      "theoretical_loss": 3.9010786048027666,
      "tokens_seen": 516685824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005766195384906374,
      "loss": 2.9237,
      "theoretical_loss": 3.9010256143740873,
      "tokens_seen": 516751360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005766029884622659,
      "loss": 2.7463,
      "theoretical_loss": 3.900972632546835,
      "tokens_seen": 516816896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005765864384338945,
      "loss": 2.6678,
      "theoretical_loss": 3.900919659318521,
      "tokens_seen": 516882432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000576569888405523,
      "loss": 2.5868,
      "theoretical_loss": 3.9008666946866613,
      "tokens_seen": 516947968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005765533383771514,
      "loss": 2.8933,
      "theoretical_loss": 3.900813738648771,
      "tokens_seen": 517013504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00057653678834878,
      "loss": 2.9732,
      "theoretical_loss": 3.900760791202366,
      "tokens_seen": 517079040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005765202383204085,
      "loss": 3.0545,
      "theoretical_loss": 3.900707852344964,
      "tokens_seen": 517144576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000576503688292037,
      "loss": 2.745,
      "theoretical_loss": 3.9006549220740836,
      "tokens_seen": 517210112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005764871382636656,
      "loss": 3.0,
      "theoretical_loss": 3.9006020003872446,
      "tokens_seen": 517275648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005764705882352941,
      "loss": 2.6456,
      "theoretical_loss": 3.9005490872819673,
      "tokens_seen": 517341184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005764540382069227,
      "loss": 2.7336,
      "theoretical_loss": 3.900496182755773,
      "tokens_seen": 517406720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005764374881785511,
      "loss": 2.7306,
      "theoretical_loss": 3.9004432868061856,
      "tokens_seen": 517472256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005764209381501796,
      "loss": 2.9386,
      "theoretical_loss": 3.9003903994307274,
      "tokens_seen": 517537792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005764043881218082,
      "loss": 2.8147,
      "theoretical_loss": 3.9003375206269233,
      "tokens_seen": 517603328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005763878380934367,
      "loss": 2.9487,
      "theoretical_loss": 3.9002846503923,
      "tokens_seen": 517668864
    },
    {
      "epoch": 0.18,
      "objective/train/avg_token_score": 0.011262793093919754,
      "objective/train/avg_weight": 0.9867475032806396,
      "objective/train/docs_used": 303002,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0381741523742676,
      "objective/train/std_weight": 0.07756645232439041,
      "objective/train/theoretical_loss": 3.9002317887243834,
      "objective/train/tokens_used": 538194400,
      "theoretical_loss": 3.9002317887243834,
      "tokens_seen": 517734400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005763712880650652,
      "loss": 2.8187,
      "theoretical_loss": 3.9002317887243834,
      "tokens_seen": 517734400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005763547380366938,
      "loss": 2.5722,
      "theoretical_loss": 3.9001789356207013,
      "tokens_seen": 517799936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005763381880083223,
      "loss": 2.7116,
      "theoretical_loss": 3.900126091078783,
      "tokens_seen": 517865472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0005763216379799509,
      "loss": 2.7702,
      "theoretical_loss": 3.9000732550961574,
      "tokens_seen": 517931008
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005763050879515793,
      "loss": 2.7642,
      "theoretical_loss": 3.9000204276703565,
      "tokens_seen": 517996544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005762885379232078,
      "loss": 3.0452,
      "theoretical_loss": 3.8999676087989115,
      "tokens_seen": 518062080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005762719878948364,
      "loss": 2.7221,
      "theoretical_loss": 3.8999147984793554,
      "tokens_seen": 518127616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005762554378664649,
      "loss": 2.8949,
      "theoretical_loss": 3.8998619967092223,
      "tokens_seen": 518193152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005762388878380934,
      "loss": 2.7754,
      "theoretical_loss": 3.8998092034860474,
      "tokens_seen": 518258688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000576222337809722,
      "loss": 3.0485,
      "theoretical_loss": 3.8997564188073657,
      "tokens_seen": 518324224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005762057877813505,
      "loss": 2.9017,
      "theoretical_loss": 3.8997036426707155,
      "tokens_seen": 518389760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000576189237752979,
      "loss": 2.8847,
      "theoretical_loss": 3.8996508750736343,
      "tokens_seen": 518455296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005761726877246075,
      "loss": 2.795,
      "theoretical_loss": 3.89959811601366,
      "tokens_seen": 518520832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000576156137696236,
      "loss": 2.6613,
      "theoretical_loss": 3.8995453654883345,
      "tokens_seen": 518586368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005761395876678646,
      "loss": 3.0174,
      "theoretical_loss": 3.899492623495198,
      "tokens_seen": 518651904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005761230376394931,
      "loss": 2.5481,
      "theoretical_loss": 3.8994398900317924,
      "tokens_seen": 518717440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005761064876111216,
      "loss": 2.793,
      "theoretical_loss": 3.8993871650956606,
      "tokens_seen": 518782976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005760899375827502,
      "loss": 2.7198,
      "theoretical_loss": 3.899334448684348,
      "tokens_seen": 518848512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005760733875543786,
      "loss": 2.5824,
      "theoretical_loss": 3.8992817407953986,
      "tokens_seen": 518914048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005760568375260072,
      "loss": 2.6805,
      "theoretical_loss": 3.8992290414263593,
      "tokens_seen": 518979584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005760402874976357,
      "loss": 2.658,
      "theoretical_loss": 3.8991763505747765,
      "tokens_seen": 519045120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005760237374692642,
      "loss": 2.7431,
      "theoretical_loss": 3.8991236682381993,
      "tokens_seen": 519110656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005760071874408928,
      "loss": 2.7174,
      "theoretical_loss": 3.899070994414176,
      "tokens_seen": 519176192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005759906374125213,
      "loss": 2.8743,
      "theoretical_loss": 3.899018329100257,
      "tokens_seen": 519241728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005759740873841498,
      "loss": 2.9839,
      "theoretical_loss": 3.8989656722939943,
      "tokens_seen": 519307264
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005759575373557784,
      "loss": 2.8165,
      "theoretical_loss": 3.8989130239929395,
      "tokens_seen": 519372800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005759409873274068,
      "loss": 2.7342,
      "theoretical_loss": 3.8988603841946463,
      "tokens_seen": 519438336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005759244372990354,
      "loss": 2.9678,
      "theoretical_loss": 3.8988077528966687,
      "tokens_seen": 519503872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005759078872706639,
      "loss": 2.8501,
      "theoretical_loss": 3.898755130096562,
      "tokens_seen": 519569408
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005758913372422924,
      "loss": 2.763,
      "theoretical_loss": 3.898702515791882,
      "tokens_seen": 519634944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000575874787213921,
      "loss": 2.9064,
      "theoretical_loss": 3.8986499099801875,
      "tokens_seen": 519700480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005758582371855495,
      "loss": 2.6002,
      "theoretical_loss": 3.898597312659035,
      "tokens_seen": 519766016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000575841687157178,
      "loss": 2.9811,
      "theoretical_loss": 3.8985447238259856,
      "tokens_seen": 519831552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005758251371288065,
      "loss": 3.0781,
      "theoretical_loss": 3.8984921434785984,
      "tokens_seen": 519897088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000575808587100435,
      "loss": 2.7932,
      "theoretical_loss": 3.8984395716144347,
      "tokens_seen": 519962624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005757920370720636,
      "loss": 2.5671,
      "theoretical_loss": 3.898387008231058,
      "tokens_seen": 520028160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005757754870436921,
      "loss": 2.8034,
      "theoretical_loss": 3.8983344533260307,
      "tokens_seen": 520093696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005757589370153206,
      "loss": 2.8565,
      "theoretical_loss": 3.8982819068969174,
      "tokens_seen": 520159232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005757423869869492,
      "loss": 2.8608,
      "theoretical_loss": 3.8982293689412835,
      "tokens_seen": 520224768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005757258369585777,
      "loss": 2.5992,
      "theoretical_loss": 3.8981768394566956,
      "tokens_seen": 520290304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005757092869302061,
      "loss": 2.7438,
      "theoretical_loss": 3.898124318440721,
      "tokens_seen": 520355840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005756927369018347,
      "loss": 2.8231,
      "theoretical_loss": 3.8980718058909276,
      "tokens_seen": 520421376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005756761868734632,
      "loss": 2.7445,
      "theoretical_loss": 3.8980193018048857,
      "tokens_seen": 520486912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005756596368450918,
      "loss": 2.8507,
      "theoretical_loss": 3.8979668061801656,
      "tokens_seen": 520552448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005756430868167203,
      "loss": 2.712,
      "theoretical_loss": 3.8979143190143377,
      "tokens_seen": 520617984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005756265367883488,
      "loss": 2.8139,
      "theoretical_loss": 3.8978618403049747,
      "tokens_seen": 520683520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005756099867599773,
      "loss": 2.7141,
      "theoretical_loss": 3.8978093700496514,
      "tokens_seen": 520749056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005755934367316057,
      "loss": 2.9467,
      "theoretical_loss": 3.897756908245941,
      "tokens_seen": 520814592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005755768867032343,
      "loss": 2.7988,
      "theoretical_loss": 3.897704454891419,
      "tokens_seen": 520880128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005755603366748628,
      "loss": 2.9762,
      "theoretical_loss": 3.897652009983662,
      "tokens_seen": 520945664
    },
    {
      "epoch": 0.19,
      "objective/train/avg_token_score": 0.03213544189929962,
      "objective/train/avg_weight": 0.962203323841095,
      "objective/train/docs_used": 304925,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9299445152282715,
      "objective/train/std_weight": 0.11483699828386307,
      "objective/train/theoretical_loss": 3.897599573520247,
      "objective/train/tokens_used": 541471200,
      "theoretical_loss": 3.897599573520247,
      "tokens_seen": 521011200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005755437866464913,
      "loss": 2.6588,
      "theoretical_loss": 3.897599573520247,
      "tokens_seen": 521011200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005755272366181199,
      "loss": 2.7099,
      "theoretical_loss": 3.897547145498753,
      "tokens_seen": 521076736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005755106865897484,
      "loss": 2.7238,
      "theoretical_loss": 3.8974947259167596,
      "tokens_seen": 521142272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000575494136561377,
      "loss": 2.7899,
      "theoretical_loss": 3.8974423147718467,
      "tokens_seen": 521207808
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005754775865330055,
      "loss": 2.9586,
      "theoretical_loss": 3.8973899120615956,
      "tokens_seen": 521273344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005754610365046339,
      "loss": 2.6981,
      "theoretical_loss": 3.8973375177835896,
      "tokens_seen": 521338880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005754444864762625,
      "loss": 2.7749,
      "theoretical_loss": 3.897285131935411,
      "tokens_seen": 521404416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000575427936447891,
      "loss": 3.0378,
      "theoretical_loss": 3.8972327545146443,
      "tokens_seen": 521469952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005754113864195196,
      "loss": 3.0326,
      "theoretical_loss": 3.897180385518876,
      "tokens_seen": 521535488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005753948363911481,
      "loss": 3.0351,
      "theoretical_loss": 3.897128024945692,
      "tokens_seen": 521601024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005753782863627766,
      "loss": 2.6832,
      "theoretical_loss": 3.8970756727926785,
      "tokens_seen": 521666560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005753617363344052,
      "loss": 2.9403,
      "theoretical_loss": 3.8970233290574257,
      "tokens_seen": 521732096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005753451863060336,
      "loss": 2.7694,
      "theoretical_loss": 3.896970993737522,
      "tokens_seen": 521797632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005753286362776621,
      "loss": 2.7756,
      "theoretical_loss": 3.896918666830558,
      "tokens_seen": 521863168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005753120862492907,
      "loss": 2.8862,
      "theoretical_loss": 3.8968663483341253,
      "tokens_seen": 521928704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005752955362209192,
      "loss": 2.6969,
      "theoretical_loss": 3.8968140382458154,
      "tokens_seen": 521994240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005752789861925478,
      "loss": 2.7336,
      "theoretical_loss": 3.8967617365632217,
      "tokens_seen": 522059776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005752624361641763,
      "loss": 2.8169,
      "theoretical_loss": 3.8967094432839398,
      "tokens_seen": 522125312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005752458861358048,
      "loss": 2.8022,
      "theoretical_loss": 3.8966571584055645,
      "tokens_seen": 522190848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005752293361074334,
      "loss": 2.9676,
      "theoretical_loss": 3.896604881925691,
      "tokens_seen": 522256384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005752127860790618,
      "loss": 2.7965,
      "theoretical_loss": 3.8965526138419184,
      "tokens_seen": 522321920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005751962360506903,
      "loss": 2.8403,
      "theoretical_loss": 3.8965003541518435,
      "tokens_seen": 522387456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005751796860223189,
      "loss": 2.6079,
      "theoretical_loss": 3.896448102853067,
      "tokens_seen": 522452992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005751631359939474,
      "loss": 2.7575,
      "theoretical_loss": 3.8963958599431874,
      "tokens_seen": 522518528
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000575146585965576,
      "loss": 3.1594,
      "theoretical_loss": 3.8963436254198074,
      "tokens_seen": 522584064
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005751300359372045,
      "loss": 2.9052,
      "theoretical_loss": 3.8962913992805284,
      "tokens_seen": 522649600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000575113485908833,
      "loss": 2.7212,
      "theoretical_loss": 3.896239181522954,
      "tokens_seen": 522715136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005750969358804615,
      "loss": 2.8347,
      "theoretical_loss": 3.8961869721446885,
      "tokens_seen": 522780672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00057508038585209,
      "loss": 2.5281,
      "theoretical_loss": 3.8961347711433376,
      "tokens_seen": 522846208
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005750638358237185,
      "loss": 2.6295,
      "theoretical_loss": 3.896082578516506,
      "tokens_seen": 522911744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005750472857953471,
      "loss": 2.8568,
      "theoretical_loss": 3.8960303942618024,
      "tokens_seen": 522977280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005750307357669756,
      "loss": 2.7089,
      "theoretical_loss": 3.8959782183768343,
      "tokens_seen": 523042816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005750141857386042,
      "loss": 2.9986,
      "theoretical_loss": 3.8959260508592104,
      "tokens_seen": 523108352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005749976357102327,
      "loss": 2.8017,
      "theoretical_loss": 3.895873891706542,
      "tokens_seen": 523173888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005749810856818611,
      "loss": 3.0268,
      "theoretical_loss": 3.895821740916439,
      "tokens_seen": 523239424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005749645356534897,
      "loss": 2.8959,
      "theoretical_loss": 3.8957695984865137,
      "tokens_seen": 523304960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005749479856251182,
      "loss": 2.8082,
      "theoretical_loss": 3.89571746441438,
      "tokens_seen": 523370496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005749314355967467,
      "loss": 2.858,
      "theoretical_loss": 3.8956653386976514,
      "tokens_seen": 523436032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005749148855683753,
      "loss": 2.7193,
      "theoretical_loss": 3.8956132213339423,
      "tokens_seen": 523501568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005748983355400038,
      "loss": 2.7242,
      "theoretical_loss": 3.8955611123208698,
      "tokens_seen": 523567104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005748817855116324,
      "loss": 2.8435,
      "theoretical_loss": 3.8955090116560505,
      "tokens_seen": 523632640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005748652354832609,
      "loss": 2.8081,
      "theoretical_loss": 3.895456919337102,
      "tokens_seen": 523698176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005748486854548893,
      "loss": 2.8208,
      "theoretical_loss": 3.8954048353616435,
      "tokens_seen": 523763712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005748321354265179,
      "loss": 2.8052,
      "theoretical_loss": 3.8953527597272952,
      "tokens_seen": 523829248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005748155853981464,
      "loss": 2.7859,
      "theoretical_loss": 3.895300692431677,
      "tokens_seen": 523894784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005747990353697749,
      "loss": 2.7142,
      "theoretical_loss": 3.8952486334724123,
      "tokens_seen": 523960320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005747824853414035,
      "loss": 2.7809,
      "theoretical_loss": 3.895196582847123,
      "tokens_seen": 524025856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000574765935313032,
      "loss": 2.9196,
      "theoretical_loss": 3.8951445405534324,
      "tokens_seen": 524091392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005747493852846606,
      "loss": 2.9574,
      "theoretical_loss": 3.8950925065889663,
      "tokens_seen": 524156928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000574732835256289,
      "loss": 2.9844,
      "theoretical_loss": 3.8950404809513506,
      "tokens_seen": 524222464
    },
    {
      "epoch": 0.19,
      "objective/train/avg_token_score": 0.012975646182894707,
      "objective/train/avg_weight": 0.984740138053894,
      "objective/train/docs_used": 306825,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.61712646484375,
      "objective/train/std_weight": 0.057298749685287476,
      "objective/train/theoretical_loss": 3.8949884636382106,
      "objective/train/tokens_used": 544748000,
      "theoretical_loss": 3.8949884636382106,
      "tokens_seen": 524288000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005747162852279175,
      "loss": 2.6464,
      "theoretical_loss": 3.8949884636382106,
      "tokens_seen": 524288000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005746997351995461,
      "loss": 2.6837,
      "theoretical_loss": 3.8949364546471754,
      "tokens_seen": 524353536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005746831851711746,
      "loss": 2.7933,
      "theoretical_loss": 3.894884453975873,
      "tokens_seen": 524419072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005746666351428031,
      "loss": 2.982,
      "theoretical_loss": 3.894832461621934,
      "tokens_seen": 524484608
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005746500851144317,
      "loss": 2.4842,
      "theoretical_loss": 3.894780477582988,
      "tokens_seen": 524550144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00057463353508606,
      "loss": 2.6027,
      "theoretical_loss": 3.894728501856667,
      "tokens_seen": 524615680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005746169850576886,
      "loss": 2.7871,
      "theoretical_loss": 3.8946765344406034,
      "tokens_seen": 524681216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005746004350293171,
      "loss": 2.7726,
      "theoretical_loss": 3.8946245753324304,
      "tokens_seen": 524746752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005745838850009457,
      "loss": 2.933,
      "theoretical_loss": 3.8945726245297836,
      "tokens_seen": 524812288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005745673349725742,
      "loss": 2.9378,
      "theoretical_loss": 3.894520682030297,
      "tokens_seen": 524877824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005745507849442027,
      "loss": 2.8873,
      "theoretical_loss": 3.894468747831609,
      "tokens_seen": 524943360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005745342349158313,
      "loss": 2.7973,
      "theoretical_loss": 3.8944168219313546,
      "tokens_seen": 525008896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005745176848874598,
      "loss": 3.0243,
      "theoretical_loss": 3.8943649043271744,
      "tokens_seen": 525074432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005745011348590882,
      "loss": 2.9433,
      "theoretical_loss": 3.8943129950167066,
      "tokens_seen": 525139968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005744845848307168,
      "loss": 3.0011,
      "theoretical_loss": 3.894261093997592,
      "tokens_seen": 525205504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005744680348023453,
      "loss": 2.8745,
      "theoretical_loss": 3.894209201267471,
      "tokens_seen": 525271040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005744514847739739,
      "loss": 2.5923,
      "theoretical_loss": 3.8941573168239874,
      "tokens_seen": 525336576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005744349347456024,
      "loss": 2.7545,
      "theoretical_loss": 3.8941054406647826,
      "tokens_seen": 525402112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005744183847172309,
      "loss": 2.7766,
      "theoretical_loss": 3.894053572787502,
      "tokens_seen": 525467648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005744018346888595,
      "loss": 2.7797,
      "theoretical_loss": 3.89400171318979,
      "tokens_seen": 525533184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000574385284660488,
      "loss": 2.7343,
      "theoretical_loss": 3.8939498618692934,
      "tokens_seen": 525598720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005743687346321164,
      "loss": 3.0411,
      "theoretical_loss": 3.893898018823659,
      "tokens_seen": 525664256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000574352184603745,
      "loss": 2.792,
      "theoretical_loss": 3.8938461840505347,
      "tokens_seen": 525729792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005743356345753735,
      "loss": 2.9731,
      "theoretical_loss": 3.8937943575475695,
      "tokens_seen": 525795328
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005743190845470021,
      "loss": 2.7847,
      "theoretical_loss": 3.893742539312414,
      "tokens_seen": 525860864
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005743025345186306,
      "loss": 2.7544,
      "theoretical_loss": 3.8936907293427176,
      "tokens_seen": 525926400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005742859844902591,
      "loss": 2.7773,
      "theoretical_loss": 3.893638927636134,
      "tokens_seen": 525991936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005742694344618877,
      "loss": 2.9972,
      "theoretical_loss": 3.893587134190314,
      "tokens_seen": 526057472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005742528844335161,
      "loss": 2.8771,
      "theoretical_loss": 3.8935353490029136,
      "tokens_seen": 526123008
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005742363344051446,
      "loss": 2.7461,
      "theoretical_loss": 3.893483572071586,
      "tokens_seen": 526188544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005742197843767732,
      "loss": 2.7642,
      "theoretical_loss": 3.8934318033939874,
      "tokens_seen": 526254080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005742032343484017,
      "loss": 2.7606,
      "theoretical_loss": 3.8933800429677743,
      "tokens_seen": 526319616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005741866843200303,
      "loss": 2.9927,
      "theoretical_loss": 3.8933282907906044,
      "tokens_seen": 526385152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005741701342916588,
      "loss": 2.8801,
      "theoretical_loss": 3.8932765468601365,
      "tokens_seen": 526450688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005741535842632873,
      "loss": 3.0399,
      "theoretical_loss": 3.89322481117403,
      "tokens_seen": 526516224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005741370342349159,
      "loss": 2.7306,
      "theoretical_loss": 3.8931730837299456,
      "tokens_seen": 526581760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005741204842065443,
      "loss": 2.7056,
      "theoretical_loss": 3.893121364525544,
      "tokens_seen": 526647296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005741039341781728,
      "loss": 2.8461,
      "theoretical_loss": 3.893069653558488,
      "tokens_seen": 526712832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005740873841498014,
      "loss": 2.8953,
      "theoretical_loss": 3.893017950826441,
      "tokens_seen": 526778368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005740708341214299,
      "loss": 2.7337,
      "theoretical_loss": 3.892966256327068,
      "tokens_seen": 526843904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005740542840930585,
      "loss": 2.671,
      "theoretical_loss": 3.892914570058033,
      "tokens_seen": 526909440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000574037734064687,
      "loss": 2.4271,
      "theoretical_loss": 3.8928628920170034,
      "tokens_seen": 526974976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005740211840363155,
      "loss": 2.6455,
      "theoretical_loss": 3.8928112222016455,
      "tokens_seen": 527040512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000574004634007944,
      "loss": 2.919,
      "theoretical_loss": 3.892759560609627,
      "tokens_seen": 527106048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005739880839795725,
      "loss": 2.7471,
      "theoretical_loss": 3.8927079072386186,
      "tokens_seen": 527171584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005739715339512011,
      "loss": 2.7863,
      "theoretical_loss": 3.892656262086289,
      "tokens_seen": 527237120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005739549839228296,
      "loss": 2.8709,
      "theoretical_loss": 3.8926046251503092,
      "tokens_seen": 527302656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005739384338944581,
      "loss": 2.7204,
      "theoretical_loss": 3.8925529964283516,
      "tokens_seen": 527368192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005739218838660867,
      "loss": 2.6811,
      "theoretical_loss": 3.892501375918089,
      "tokens_seen": 527433728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005739053338377152,
      "loss": 2.9162,
      "theoretical_loss": 3.8924497636171953,
      "tokens_seen": 527499264
    },
    {
      "epoch": 0.19,
      "objective/train/avg_token_score": 0.012383617460727692,
      "objective/train/avg_weight": 0.9854322075843811,
      "objective/train/docs_used": 308042,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9607300758361816,
      "objective/train/std_weight": 0.06975019723176956,
      "objective/train/theoretical_loss": 3.892398159523345,
      "objective/train/tokens_used": 548024800,
      "theoretical_loss": 3.892398159523345,
      "tokens_seen": 527564800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005738887838093436,
      "loss": 2.8811,
      "theoretical_loss": 3.892398159523345,
      "tokens_seen": 527564800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005738722337809722,
      "loss": 2.9312,
      "theoretical_loss": 3.892346563634213,
      "tokens_seen": 527630336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005738556837526007,
      "loss": 2.952,
      "theoretical_loss": 3.892294975947477,
      "tokens_seen": 527695872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005738391337242293,
      "loss": 2.6789,
      "theoretical_loss": 3.892243396460815,
      "tokens_seen": 527761408
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005738225836958578,
      "loss": 2.9304,
      "theoretical_loss": 3.8921918251719054,
      "tokens_seen": 527826944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005738060336674863,
      "loss": 2.7242,
      "theoretical_loss": 3.892140262078427,
      "tokens_seen": 527892480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005737894836391149,
      "loss": 2.7423,
      "theoretical_loss": 3.8920887071780594,
      "tokens_seen": 527958016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005737729336107434,
      "loss": 2.7639,
      "theoretical_loss": 3.8920371604684862,
      "tokens_seen": 528023552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005737563835823718,
      "loss": 2.6657,
      "theoretical_loss": 3.891985621947388,
      "tokens_seen": 528089088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005737398335540004,
      "loss": 2.6433,
      "theoretical_loss": 3.891934091612449,
      "tokens_seen": 528154624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005737232835256289,
      "loss": 3.0461,
      "theoretical_loss": 3.891882569461353,
      "tokens_seen": 528220160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005737067334972575,
      "loss": 2.9103,
      "theoretical_loss": 3.891831055491785,
      "tokens_seen": 528285696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000573690183468886,
      "loss": 2.9235,
      "theoretical_loss": 3.8917795497014316,
      "tokens_seen": 528351232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005736736334405145,
      "loss": 2.913,
      "theoretical_loss": 3.891728052087979,
      "tokens_seen": 528416768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000573657083412143,
      "loss": 2.9626,
      "theoretical_loss": 3.891676562649116,
      "tokens_seen": 528482304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005736405333837714,
      "loss": 3.0447,
      "theoretical_loss": 3.8916250813825313,
      "tokens_seen": 528547840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005736239833554,
      "loss": 3.0486,
      "theoretical_loss": 3.891573608285915,
      "tokens_seen": 528613376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005736074333270285,
      "loss": 2.9593,
      "theoretical_loss": 3.891522143356957,
      "tokens_seen": 528678912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000573590883298657,
      "loss": 2.8593,
      "theoretical_loss": 3.89147068659335,
      "tokens_seen": 528744448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005735743332702856,
      "loss": 2.834,
      "theoretical_loss": 3.8914192379927863,
      "tokens_seen": 528809984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005735577832419141,
      "loss": 2.8912,
      "theoretical_loss": 3.891367797552959,
      "tokens_seen": 528875520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005735412332135427,
      "loss": 2.8681,
      "theoretical_loss": 3.891316365271564,
      "tokens_seen": 528941056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005735246831851711,
      "loss": 2.9204,
      "theoretical_loss": 3.891264941146295,
      "tokens_seen": 529006592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005735081331567996,
      "loss": 2.8608,
      "theoretical_loss": 3.89121352517485,
      "tokens_seen": 529072128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005734915831284282,
      "loss": 2.7171,
      "theoretical_loss": 3.8911621173549253,
      "tokens_seen": 529137664
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005734750331000567,
      "loss": 2.7986,
      "theoretical_loss": 3.89111071768422,
      "tokens_seen": 529203200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005734584830716852,
      "loss": 2.9611,
      "theoretical_loss": 3.8910593261604323,
      "tokens_seen": 529268736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005734419330433138,
      "loss": 2.7719,
      "theoretical_loss": 3.891007942781264,
      "tokens_seen": 529334272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005734253830149423,
      "loss": 2.7587,
      "theoretical_loss": 3.8909565675444138,
      "tokens_seen": 529399808
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005734088329865709,
      "loss": 2.7678,
      "theoretical_loss": 3.8909052004475857,
      "tokens_seen": 529465344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005733922829581993,
      "loss": 2.8367,
      "theoretical_loss": 3.8908538414884823,
      "tokens_seen": 529530880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005733757329298278,
      "loss": 2.6511,
      "theoretical_loss": 3.8908024906648073,
      "tokens_seen": 529596416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005733591829014564,
      "loss": 2.8864,
      "theoretical_loss": 3.890751147974265,
      "tokens_seen": 529661952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005733426328730849,
      "loss": 2.6957,
      "theoretical_loss": 3.890699813414562,
      "tokens_seen": 529727488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005733260828447134,
      "loss": 2.9248,
      "theoretical_loss": 3.890648486983405,
      "tokens_seen": 529793024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000573309532816342,
      "loss": 2.9603,
      "theoretical_loss": 3.8905971686785006,
      "tokens_seen": 529858560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005732929827879705,
      "loss": 3.0183,
      "theoretical_loss": 3.8905458584975583,
      "tokens_seen": 529924096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000573276432759599,
      "loss": 3.0248,
      "theoretical_loss": 3.890494556438287,
      "tokens_seen": 529989632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005732598827312275,
      "loss": 2.8915,
      "theoretical_loss": 3.8904432624983976,
      "tokens_seen": 530055168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000573243332702856,
      "loss": 2.7128,
      "theoretical_loss": 3.890391976675601,
      "tokens_seen": 530120704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005732267826744846,
      "loss": 2.7939,
      "theoretical_loss": 3.89034069896761,
      "tokens_seen": 530186240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005732102326461131,
      "loss": 2.7278,
      "theoretical_loss": 3.890289429372137,
      "tokens_seen": 530251776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005731936826177416,
      "loss": 3.0418,
      "theoretical_loss": 3.8902381678868974,
      "tokens_seen": 530317312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005731771325893702,
      "loss": 3.0757,
      "theoretical_loss": 3.890186914509605,
      "tokens_seen": 530382848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005731605825609986,
      "loss": 2.9038,
      "theoretical_loss": 3.890135669237976,
      "tokens_seen": 530448384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005731440325326272,
      "loss": 2.9684,
      "theoretical_loss": 3.8900844320697274,
      "tokens_seen": 530513920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005731274825042557,
      "loss": 2.8479,
      "theoretical_loss": 3.8900332030025773,
      "tokens_seen": 530579456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005731109324758842,
      "loss": 2.8234,
      "theoretical_loss": 3.8899819820342443,
      "tokens_seen": 530644992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005730943824475128,
      "loss": 2.8111,
      "theoretical_loss": 3.889930769162448,
      "tokens_seen": 530710528
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005730778324191413,
      "loss": 2.9689,
      "theoretical_loss": 3.8898795643849087,
      "tokens_seen": 530776064
    },
    {
      "epoch": 0.19,
      "objective/train/avg_token_score": 0.02212177775800228,
      "objective/train/avg_weight": 0.973977267742157,
      "objective/train/docs_used": 310103,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.950751781463623,
      "objective/train/std_weight": 0.09997051954269409,
      "objective/train/theoretical_loss": 3.889828367699349,
      "objective/train/tokens_used": 551301600,
      "theoretical_loss": 3.889828367699349,
      "tokens_seen": 530841600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005730612823907698,
      "loss": 2.6976,
      "theoretical_loss": 3.889828367699349,
      "tokens_seen": 530841600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005730447323623984,
      "loss": 2.7745,
      "theoretical_loss": 3.88977717910349,
      "tokens_seen": 530907136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005730281823340268,
      "loss": 2.773,
      "theoretical_loss": 3.8897259985950554,
      "tokens_seen": 530972672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005730116323056554,
      "loss": 2.9232,
      "theoretical_loss": 3.8896748261717695,
      "tokens_seen": 531038208
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005729950822772839,
      "loss": 2.9012,
      "theoretical_loss": 3.889623661831359,
      "tokens_seen": 531103744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005729785322489124,
      "loss": 2.7998,
      "theoretical_loss": 3.8895725055715475,
      "tokens_seen": 531169280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000572961982220541,
      "loss": 3.0787,
      "theoretical_loss": 3.8895213573900635,
      "tokens_seen": 531234816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005729454321921695,
      "loss": 2.7584,
      "theoretical_loss": 3.889470217284635,
      "tokens_seen": 531300352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000572928882163798,
      "loss": 2.9632,
      "theoretical_loss": 3.8894190852529906,
      "tokens_seen": 531365888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005729123321354265,
      "loss": 2.845,
      "theoretical_loss": 3.8893679612928604,
      "tokens_seen": 531431424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000572895782107055,
      "loss": 2.7513,
      "theoretical_loss": 3.889316845401974,
      "tokens_seen": 531496960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005728792320786836,
      "loss": 2.7715,
      "theoretical_loss": 3.8892657375780644,
      "tokens_seen": 531562496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005728626820503121,
      "loss": 2.9048,
      "theoretical_loss": 3.8892146378188643,
      "tokens_seen": 531628032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005728461320219406,
      "loss": 2.6059,
      "theoretical_loss": 3.8891635461221057,
      "tokens_seen": 531693568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005728295819935692,
      "loss": 2.7038,
      "theoretical_loss": 3.8891124624855236,
      "tokens_seen": 531759104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005728130319651977,
      "loss": 2.6834,
      "theoretical_loss": 3.889061386906854,
      "tokens_seen": 531824640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005727964819368261,
      "loss": 2.6946,
      "theoretical_loss": 3.889010319383832,
      "tokens_seen": 531890176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005727799319084547,
      "loss": 2.8478,
      "theoretical_loss": 3.8889592599141958,
      "tokens_seen": 531955712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005727633818800832,
      "loss": 2.7295,
      "theoretical_loss": 3.888908208495683,
      "tokens_seen": 532021248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005727468318517118,
      "loss": 2.8303,
      "theoretical_loss": 3.8888571651260326,
      "tokens_seen": 532086784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005727302818233403,
      "loss": 2.7104,
      "theoretical_loss": 3.888806129802984,
      "tokens_seen": 532152320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005727137317949688,
      "loss": 2.8065,
      "theoretical_loss": 3.888755102524279,
      "tokens_seen": 532217856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005726971817665974,
      "loss": 2.7272,
      "theoretical_loss": 3.8887040832876587,
      "tokens_seen": 532283392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005726806317382257,
      "loss": 2.8558,
      "theoretical_loss": 3.8886530720908654,
      "tokens_seen": 532348928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005726640817098543,
      "loss": 2.6542,
      "theoretical_loss": 3.888602068931643,
      "tokens_seen": 532414464
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005726475316814828,
      "loss": 2.7932,
      "theoretical_loss": 3.888551073807736,
      "tokens_seen": 532480000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005726309816531114,
      "loss": 2.9845,
      "theoretical_loss": 3.8885000867168893,
      "tokens_seen": 532545536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005726144316247399,
      "loss": 2.7003,
      "theoretical_loss": 3.88844910765685,
      "tokens_seen": 532611072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005725978815963684,
      "loss": 2.671,
      "theoretical_loss": 3.888398136625364,
      "tokens_seen": 532676608
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000572581331567997,
      "loss": 3.0253,
      "theoretical_loss": 3.888347173620181,
      "tokens_seen": 532742144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005725647815396255,
      "loss": 2.8418,
      "theoretical_loss": 3.8882962186390486,
      "tokens_seen": 532807680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005725482315112539,
      "loss": 2.6181,
      "theoretical_loss": 3.8882452716797173,
      "tokens_seen": 532873216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005725316814828825,
      "loss": 2.8147,
      "theoretical_loss": 3.8881943327399378,
      "tokens_seen": 532938752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000572515131454511,
      "loss": 2.8606,
      "theoretical_loss": 3.8881434018174614,
      "tokens_seen": 533004288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005724985814261396,
      "loss": 2.6589,
      "theoretical_loss": 3.8880924789100417,
      "tokens_seen": 533069824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005724820313977681,
      "loss": 2.6798,
      "theoretical_loss": 3.888041564015431,
      "tokens_seen": 533135360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005724654813693966,
      "loss": 2.8673,
      "theoretical_loss": 3.8879906571313847,
      "tokens_seen": 533200896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005724489313410252,
      "loss": 2.6427,
      "theoretical_loss": 3.887939758255658,
      "tokens_seen": 533266432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005724323813126537,
      "loss": 2.9647,
      "theoretical_loss": 3.887888867386007,
      "tokens_seen": 533331968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005724158312842821,
      "loss": 2.8864,
      "theoretical_loss": 3.8878379845201883,
      "tokens_seen": 533397504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005723992812559107,
      "loss": 2.7664,
      "theoretical_loss": 3.88778710965596,
      "tokens_seen": 533463040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005723827312275392,
      "loss": 2.6642,
      "theoretical_loss": 3.887736242791082,
      "tokens_seen": 533528576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005723661811991678,
      "loss": 2.9179,
      "theoretical_loss": 3.887685383923314,
      "tokens_seen": 533594112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005723496311707963,
      "loss": 2.4603,
      "theoretical_loss": 3.8876345330504156,
      "tokens_seen": 533659648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005723330811424248,
      "loss": 2.9416,
      "theoretical_loss": 3.887583690170149,
      "tokens_seen": 533725184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005723165311140534,
      "loss": 2.825,
      "theoretical_loss": 3.887532855280277,
      "tokens_seen": 533790720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005722999810856818,
      "loss": 2.9545,
      "theoretical_loss": 3.887482028378564,
      "tokens_seen": 533856256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005722834310573103,
      "loss": 3.1257,
      "theoretical_loss": 3.8874312094627723,
      "tokens_seen": 533921792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005722668810289389,
      "loss": 2.7097,
      "theoretical_loss": 3.8873803985306683,
      "tokens_seen": 533987328
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005722503310005674,
      "loss": 2.852,
      "theoretical_loss": 3.8873295955800184,
      "tokens_seen": 534052864
    },
    {
      "epoch": 0.19,
      "objective/train/avg_token_score": 0.014837130904197693,
      "objective/train/avg_weight": 0.9825793504714966,
      "objective/train/docs_used": 311341,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.5029468536376953,
      "objective/train/std_weight": 0.10442639142274857,
      "objective/train/theoretical_loss": 3.8872788006085894,
      "objective/train/tokens_used": 554578400,
      "theoretical_loss": 3.8872788006085894,
      "tokens_seen": 534118400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000572233780972196,
      "loss": 2.7474,
      "theoretical_loss": 3.8872788006085894,
      "tokens_seen": 534118400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005722172309438245,
      "loss": 2.798,
      "theoretical_loss": 3.8872280136141484,
      "tokens_seen": 534183936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000572200680915453,
      "loss": 2.8465,
      "theoretical_loss": 3.887177234594466,
      "tokens_seen": 534249472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005721841308870816,
      "loss": 2.6773,
      "theoretical_loss": 3.88712646354731,
      "tokens_seen": 534315008
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00057216758085871,
      "loss": 2.7299,
      "theoretical_loss": 3.8870757004704526,
      "tokens_seen": 534380544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005721510308303385,
      "loss": 2.8673,
      "theoretical_loss": 3.8870249453616648,
      "tokens_seen": 534446080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005721344808019671,
      "loss": 2.5908,
      "theoretical_loss": 3.8869741982187183,
      "tokens_seen": 534511616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005721179307735956,
      "loss": 2.9122,
      "theoretical_loss": 3.8869234590393873,
      "tokens_seen": 534577152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005721013807452242,
      "loss": 3.0313,
      "theoretical_loss": 3.886872727821446,
      "tokens_seen": 534642688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005720848307168527,
      "loss": 2.9147,
      "theoretical_loss": 3.88682200456267,
      "tokens_seen": 534708224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005720682806884812,
      "loss": 2.5443,
      "theoretical_loss": 3.886771289260834,
      "tokens_seen": 534773760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005720517306601097,
      "loss": 2.6923,
      "theoretical_loss": 3.886720581913716,
      "tokens_seen": 534839296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005720351806317382,
      "loss": 2.6076,
      "theoretical_loss": 3.886669882519093,
      "tokens_seen": 534904832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005720186306033667,
      "loss": 2.8422,
      "theoretical_loss": 3.886619191074744,
      "tokens_seen": 534970368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005720020805749953,
      "loss": 2.7541,
      "theoretical_loss": 3.886568507578448,
      "tokens_seen": 535035904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005719855305466238,
      "loss": 2.745,
      "theoretical_loss": 3.8865178320279874,
      "tokens_seen": 535101440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005719689805182524,
      "loss": 2.7564,
      "theoretical_loss": 3.886467164421142,
      "tokens_seen": 535166976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005719524304898809,
      "loss": 2.8414,
      "theoretical_loss": 3.8864165047556942,
      "tokens_seen": 535232512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005719358804615093,
      "loss": 2.9261,
      "theoretical_loss": 3.886365853029427,
      "tokens_seen": 535298048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005719193304331379,
      "loss": 2.9331,
      "theoretical_loss": 3.8863152092401254,
      "tokens_seen": 535363584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005719027804047664,
      "loss": 2.6778,
      "theoretical_loss": 3.886264573385574,
      "tokens_seen": 535429120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005718862303763949,
      "loss": 2.8286,
      "theoretical_loss": 3.886213945463558,
      "tokens_seen": 535494656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005718696803480235,
      "loss": 2.7127,
      "theoretical_loss": 3.886163325471864,
      "tokens_seen": 535560192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000571853130319652,
      "loss": 2.9369,
      "theoretical_loss": 3.8861127134082807,
      "tokens_seen": 535625728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005718365802912806,
      "loss": 2.589,
      "theoretical_loss": 3.8860621092705956,
      "tokens_seen": 535691264
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000571820030262909,
      "loss": 2.9244,
      "theoretical_loss": 3.8860115130565984,
      "tokens_seen": 535756800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005718034802345375,
      "loss": 2.8512,
      "theoretical_loss": 3.8859609247640794,
      "tokens_seen": 535822336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005717869302061661,
      "loss": 2.8479,
      "theoretical_loss": 3.88591034439083,
      "tokens_seen": 535887872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005717703801777946,
      "loss": 2.9085,
      "theoretical_loss": 3.885859771934642,
      "tokens_seen": 535953408
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005717538301494231,
      "loss": 2.9567,
      "theoretical_loss": 3.885809207393308,
      "tokens_seen": 536018944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005717372801210517,
      "loss": 3.0564,
      "theoretical_loss": 3.8857586507646227,
      "tokens_seen": 536084480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005717207300926802,
      "loss": 2.7423,
      "theoretical_loss": 3.88570810204638,
      "tokens_seen": 536150016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005717041800643088,
      "loss": 2.7628,
      "theoretical_loss": 3.8856575612363753,
      "tokens_seen": 536215552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005716876300359371,
      "loss": 2.7431,
      "theoretical_loss": 3.8856070283324056,
      "tokens_seen": 536281088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005716710800075657,
      "loss": 2.9932,
      "theoretical_loss": 3.885556503332268,
      "tokens_seen": 536346624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005716545299791942,
      "loss": 2.8197,
      "theoretical_loss": 3.8855059862337606,
      "tokens_seen": 536412160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005716379799508227,
      "loss": 3.1035,
      "theoretical_loss": 3.885455477034683,
      "tokens_seen": 536477696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005716214299224513,
      "loss": 2.8957,
      "theoretical_loss": 3.8854049757328353,
      "tokens_seen": 536543232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005716048798940798,
      "loss": 2.9376,
      "theoretical_loss": 3.8853544823260178,
      "tokens_seen": 536608768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005715883298657083,
      "loss": 2.7023,
      "theoretical_loss": 3.8853039968120324,
      "tokens_seen": 536674304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005715717798373368,
      "loss": 2.8349,
      "theoretical_loss": 3.885253519188682,
      "tokens_seen": 536739840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005715552298089653,
      "loss": 2.8329,
      "theoretical_loss": 3.885203049453769,
      "tokens_seen": 536805376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005715386797805939,
      "loss": 2.8502,
      "theoretical_loss": 3.8851525876050994,
      "tokens_seen": 536870912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005715221297522224,
      "loss": 2.7717,
      "theoretical_loss": 3.885102133640477,
      "tokens_seen": 536936448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005715055797238509,
      "loss": 2.7182,
      "theoretical_loss": 3.8850516875577092,
      "tokens_seen": 537001984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005714890296954795,
      "loss": 2.9726,
      "theoretical_loss": 3.8850012493546027,
      "tokens_seen": 537067520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000571472479667108,
      "loss": 2.7881,
      "theoretical_loss": 3.884950819028965,
      "tokens_seen": 537133056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005714559296387366,
      "loss": 2.7417,
      "theoretical_loss": 3.884900396578605,
      "tokens_seen": 537198592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000571439379610365,
      "loss": 2.8588,
      "theoretical_loss": 3.8848499820013327,
      "tokens_seen": 537264128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005714228295819935,
      "loss": 2.8753,
      "theoretical_loss": 3.884799575294958,
      "tokens_seen": 537329664
    },
    {
      "epoch": 0.19,
      "objective/train/avg_token_score": 0.022589176893234253,
      "objective/train/avg_weight": 0.9734197854995728,
      "objective/train/docs_used": 313016,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.971989631652832,
      "objective/train/std_weight": 0.07564128190279007,
      "objective/train/theoretical_loss": 3.8847491764572926,
      "objective/train/tokens_used": 557855200,
      "theoretical_loss": 3.8847491764572926,
      "tokens_seen": 537395200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005714062795536221,
      "loss": 2.7626,
      "theoretical_loss": 3.8847491764572926,
      "tokens_seen": 537395200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005713897295252506,
      "loss": 2.5813,
      "theoretical_loss": 3.884698785486149,
      "tokens_seen": 537460736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005713731794968791,
      "loss": 2.9542,
      "theoretical_loss": 3.884648402379341,
      "tokens_seen": 537526272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005713566294685077,
      "loss": 3.0083,
      "theoretical_loss": 3.8845980271346816,
      "tokens_seen": 537591808
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005713400794401362,
      "loss": 3.0087,
      "theoretical_loss": 3.8845476597499857,
      "tokens_seen": 537657344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005713235294117647,
      "loss": 2.8084,
      "theoretical_loss": 3.884497300223069,
      "tokens_seen": 537722880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005713069793833932,
      "loss": 2.8209,
      "theoretical_loss": 3.8844469485517488,
      "tokens_seen": 537788416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005712904293550217,
      "loss": 2.7946,
      "theoretical_loss": 3.8843966047338423,
      "tokens_seen": 537853952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005712738793266503,
      "loss": 2.8504,
      "theoretical_loss": 3.8843462687671684,
      "tokens_seen": 537919488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005712573292982788,
      "loss": 2.7248,
      "theoretical_loss": 3.8842959406495456,
      "tokens_seen": 537985024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005712407792699073,
      "loss": 2.8841,
      "theoretical_loss": 3.884245620378794,
      "tokens_seen": 538050560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005712242292415359,
      "loss": 2.635,
      "theoretical_loss": 3.8841953079527354,
      "tokens_seen": 538116096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005712076792131643,
      "loss": 2.8294,
      "theoretical_loss": 3.8841450033691913,
      "tokens_seen": 538181632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005711911291847929,
      "loss": 2.724,
      "theoretical_loss": 3.8840947066259845,
      "tokens_seen": 538247168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005711745791564214,
      "loss": 2.8007,
      "theoretical_loss": 3.884044417720938,
      "tokens_seen": 538312704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005711580291280499,
      "loss": 2.8145,
      "theoretical_loss": 3.8839941366518778,
      "tokens_seen": 538378240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005711414790996785,
      "loss": 3.0492,
      "theoretical_loss": 3.8839438634166275,
      "tokens_seen": 538443776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000571124929071307,
      "loss": 2.9279,
      "theoretical_loss": 3.883893598013014,
      "tokens_seen": 538509312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005711083790429355,
      "loss": 2.6667,
      "theoretical_loss": 3.883843340438865,
      "tokens_seen": 538574848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000571091829014564,
      "loss": 2.8915,
      "theoretical_loss": 3.8837930906920075,
      "tokens_seen": 538640384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005710752789861925,
      "loss": 2.8325,
      "theoretical_loss": 3.8837428487702716,
      "tokens_seen": 538705920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005710587289578211,
      "loss": 3.1202,
      "theoretical_loss": 3.8836926146714856,
      "tokens_seen": 538771456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005710421789294496,
      "loss": 2.8324,
      "theoretical_loss": 3.883642388393481,
      "tokens_seen": 538836992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005710256289010781,
      "loss": 2.6852,
      "theoretical_loss": 3.8835921699340887,
      "tokens_seen": 538902528
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005710090788727067,
      "loss": 2.7128,
      "theoretical_loss": 3.8835419592911413,
      "tokens_seen": 538968064
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005709925288443352,
      "loss": 2.8906,
      "theoretical_loss": 3.8834917564624716,
      "tokens_seen": 539033600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005709759788159637,
      "loss": 2.5986,
      "theoretical_loss": 3.883441561445914,
      "tokens_seen": 539099136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005709594287875922,
      "loss": 2.8722,
      "theoretical_loss": 3.8833913742393036,
      "tokens_seen": 539164672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005709428787592207,
      "loss": 2.7361,
      "theoretical_loss": 3.8833411948404755,
      "tokens_seen": 539230208
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005709263287308493,
      "loss": 2.8665,
      "theoretical_loss": 3.883291023247267,
      "tokens_seen": 539295744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005709097787024778,
      "loss": 2.8493,
      "theoretical_loss": 3.8832408594575147,
      "tokens_seen": 539361280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005708932286741063,
      "loss": 2.5903,
      "theoretical_loss": 3.883190703469058,
      "tokens_seen": 539426816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005708766786457349,
      "loss": 2.6987,
      "theoretical_loss": 3.883140555279735,
      "tokens_seen": 539492352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005708601286173634,
      "loss": 2.6378,
      "theoretical_loss": 3.883090414887387,
      "tokens_seen": 539557888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005708435785889918,
      "loss": 2.9505,
      "theoretical_loss": 3.8830402822898535,
      "tokens_seen": 539623424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005708270285606204,
      "loss": 2.8729,
      "theoretical_loss": 3.8829901574849774,
      "tokens_seen": 539688960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005708104785322489,
      "loss": 2.7588,
      "theoretical_loss": 3.8829400404706003,
      "tokens_seen": 539754496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005707939285038775,
      "loss": 3.009,
      "theoretical_loss": 3.8828899312445673,
      "tokens_seen": 539820032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000570777378475506,
      "loss": 2.7819,
      "theoretical_loss": 3.8828398298047206,
      "tokens_seen": 539885568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005707608284471345,
      "loss": 2.7904,
      "theoretical_loss": 3.8827897361489074,
      "tokens_seen": 539951104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005707442784187631,
      "loss": 2.905,
      "theoretical_loss": 3.8827396502749734,
      "tokens_seen": 540016640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005707277283903916,
      "loss": 2.9784,
      "theoretical_loss": 3.882689572180764,
      "tokens_seen": 540082176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00057071117836202,
      "loss": 2.6865,
      "theoretical_loss": 3.882639501864129,
      "tokens_seen": 540147712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005706946283336485,
      "loss": 2.7438,
      "theoretical_loss": 3.8825894393229152,
      "tokens_seen": 540213248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000570678078305277,
      "loss": 2.7986,
      "theoretical_loss": 3.882539384554974,
      "tokens_seen": 540278784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005706615282769056,
      "loss": 2.8093,
      "theoretical_loss": 3.8824893375581544,
      "tokens_seen": 540344320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005706449782485341,
      "loss": 2.881,
      "theoretical_loss": 3.8824392983303078,
      "tokens_seen": 540409856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005706284282201627,
      "loss": 2.8219,
      "theoretical_loss": 3.8823892668692865,
      "tokens_seen": 540475392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005706118781917912,
      "loss": 2.4158,
      "theoretical_loss": 3.8823392431729435,
      "tokens_seen": 540540928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005705953281634196,
      "loss": 2.7421,
      "theoretical_loss": 3.8822892272391325,
      "tokens_seen": 540606464
    },
    {
      "epoch": 0.19,
      "objective/train/avg_token_score": 0.01827181503176689,
      "objective/train/avg_weight": 0.9784961938858032,
      "objective/train/docs_used": 314826,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9091579914093018,
      "objective/train/std_weight": 0.08282213658094406,
      "objective/train/theoretical_loss": 3.882239219065708,
      "objective/train/tokens_used": 561132000,
      "theoretical_loss": 3.882239219065708,
      "tokens_seen": 540672000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005705787781350482,
      "loss": 2.6336,
      "theoretical_loss": 3.882239219065708,
      "tokens_seen": 540672000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005705622281066767,
      "loss": 2.9292,
      "theoretical_loss": 3.882189218650525,
      "tokens_seen": 540737536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005705456780783052,
      "loss": 2.6643,
      "theoretical_loss": 3.882139225991441,
      "tokens_seen": 540803072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005705291280499338,
      "loss": 2.8516,
      "theoretical_loss": 3.8820892410863124,
      "tokens_seen": 540868608
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005705125780215623,
      "loss": 2.7506,
      "theoretical_loss": 3.882039263932997,
      "tokens_seen": 540934144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005704960279931909,
      "loss": 2.6635,
      "theoretical_loss": 3.8819892945293546,
      "tokens_seen": 540999680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005704794779648193,
      "loss": 2.9092,
      "theoretical_loss": 3.8819393328732446,
      "tokens_seen": 541065216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005704629279364478,
      "loss": 2.7795,
      "theoretical_loss": 3.8818893789625273,
      "tokens_seen": 541130752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005704463779080764,
      "loss": 2.9071,
      "theoretical_loss": 3.881839432795064,
      "tokens_seen": 541196288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005704298278797049,
      "loss": 2.6226,
      "theoretical_loss": 3.881789494368717,
      "tokens_seen": 541261824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005704132778513334,
      "loss": 2.8125,
      "theoretical_loss": 3.88173956368135,
      "tokens_seen": 541327360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000570396727822962,
      "loss": 2.7087,
      "theoretical_loss": 3.8816896407308263,
      "tokens_seen": 541392896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005703801777945905,
      "loss": 2.7265,
      "theoretical_loss": 3.8816397255150115,
      "tokens_seen": 541458432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000570363627766219,
      "loss": 2.7346,
      "theoretical_loss": 3.8815898180317703,
      "tokens_seen": 541523968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005703470777378475,
      "loss": 3.0147,
      "theoretical_loss": 3.8815399182789703,
      "tokens_seen": 541589504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000570330527709476,
      "loss": 2.5523,
      "theoretical_loss": 3.881490026254478,
      "tokens_seen": 541655040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005703139776811046,
      "loss": 2.9439,
      "theoretical_loss": 3.881440141956162,
      "tokens_seen": 541720576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005702974276527331,
      "loss": 2.916,
      "theoretical_loss": 3.8813902653818912,
      "tokens_seen": 541786112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005702808776243616,
      "loss": 2.7966,
      "theoretical_loss": 3.8813403965295357,
      "tokens_seen": 541851648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005702643275959902,
      "loss": 2.7749,
      "theoretical_loss": 3.8812905353969667,
      "tokens_seen": 541917184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005702477775676187,
      "loss": 2.8118,
      "theoretical_loss": 3.881240681982054,
      "tokens_seen": 541982720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005702312275392472,
      "loss": 2.9163,
      "theoretical_loss": 3.8811908362826726,
      "tokens_seen": 542048256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005702146775108757,
      "loss": 2.7805,
      "theoretical_loss": 3.881140998296694,
      "tokens_seen": 542113792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005701981274825042,
      "loss": 2.7505,
      "theoretical_loss": 3.8810911680219933,
      "tokens_seen": 542179328
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005701815774541328,
      "loss": 2.6807,
      "theoretical_loss": 3.881041345456444,
      "tokens_seen": 542244864
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005701650274257613,
      "loss": 3.139,
      "theoretical_loss": 3.8809915305979237,
      "tokens_seen": 542310400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005701484773973898,
      "loss": 2.7936,
      "theoretical_loss": 3.880941723444308,
      "tokens_seen": 542375936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005701319273690184,
      "loss": 2.8321,
      "theoretical_loss": 3.8808919239934747,
      "tokens_seen": 542441472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005701153773406468,
      "loss": 2.8112,
      "theoretical_loss": 3.8808421322433024,
      "tokens_seen": 542507008
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005700988273122754,
      "loss": 2.743,
      "theoretical_loss": 3.8807923481916697,
      "tokens_seen": 542572544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005700822772839039,
      "loss": 2.9109,
      "theoretical_loss": 3.880742571836457,
      "tokens_seen": 542638080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005700657272555324,
      "loss": 2.8382,
      "theoretical_loss": 3.8806928031755454,
      "tokens_seen": 542703616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000570049177227161,
      "loss": 3.054,
      "theoretical_loss": 3.880643042206816,
      "tokens_seen": 542769152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005700326271987895,
      "loss": 2.9292,
      "theoretical_loss": 3.880593288928152,
      "tokens_seen": 542834688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000570016077170418,
      "loss": 2.813,
      "theoretical_loss": 3.8805435433374353,
      "tokens_seen": 542900224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005699995271420466,
      "loss": 2.9243,
      "theoretical_loss": 3.880493805432552,
      "tokens_seen": 542965760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000569982977113675,
      "loss": 2.6172,
      "theoretical_loss": 3.8804440752113862,
      "tokens_seen": 543031296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005699664270853036,
      "loss": 2.7991,
      "theoretical_loss": 3.880394352671824,
      "tokens_seen": 543096832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005699498770569321,
      "loss": 2.994,
      "theoretical_loss": 3.880344637811752,
      "tokens_seen": 543162368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005699333270285606,
      "loss": 3.0728,
      "theoretical_loss": 3.880294930629058,
      "tokens_seen": 543227904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005699167770001892,
      "loss": 3.0684,
      "theoretical_loss": 3.8802452311216302,
      "tokens_seen": 543293440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005699002269718177,
      "loss": 2.8854,
      "theoretical_loss": 3.880195539287358,
      "tokens_seen": 543358976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005698836769434463,
      "loss": 2.8852,
      "theoretical_loss": 3.8801458551241312,
      "tokens_seen": 543424512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005698671269150747,
      "loss": 2.7806,
      "theoretical_loss": 3.8800961786298407,
      "tokens_seen": 543490048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005698505768867032,
      "loss": 3.0099,
      "theoretical_loss": 3.880046509802378,
      "tokens_seen": 543555584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005698340268583318,
      "loss": 2.725,
      "theoretical_loss": 3.879996848639637,
      "tokens_seen": 543621120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005698174768299603,
      "loss": 2.7668,
      "theoretical_loss": 3.8799471951395095,
      "tokens_seen": 543686656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005698009268015888,
      "loss": 2.8192,
      "theoretical_loss": 3.8798975492998906,
      "tokens_seen": 543752192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005697843767732174,
      "loss": 2.7378,
      "theoretical_loss": 3.8798479111186754,
      "tokens_seen": 543817728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005697678267448459,
      "loss": 2.6948,
      "theoretical_loss": 3.879798280593759,
      "tokens_seen": 543883264
    },
    {
      "epoch": 0.19,
      "objective/train/avg_token_score": 0.02929091453552246,
      "objective/train/avg_weight": 0.9655252695083618,
      "objective/train/docs_used": 316612,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2029671669006348,
      "objective/train/std_weight": 0.10290524363517761,
      "objective/train/theoretical_loss": 3.879748657723039,
      "objective/train/tokens_used": 564408800,
      "theoretical_loss": 3.879748657723039,
      "tokens_seen": 543948800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005697512767164745,
      "loss": 2.9395,
      "theoretical_loss": 3.879748657723039,
      "tokens_seen": 543948800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005697347266881029,
      "loss": 2.9647,
      "theoretical_loss": 3.8796990425044124,
      "tokens_seen": 544014336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005697181766597314,
      "loss": 2.9336,
      "theoretical_loss": 3.879649434935778,
      "tokens_seen": 544079872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005697016266313599,
      "loss": 3.0359,
      "theoretical_loss": 3.8795998350150347,
      "tokens_seen": 544145408
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005696850766029884,
      "loss": 2.7354,
      "theoretical_loss": 3.879550242740083,
      "tokens_seen": 544210944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000569668526574617,
      "loss": 2.6773,
      "theoretical_loss": 3.8795006581088236,
      "tokens_seen": 544276480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005696519765462455,
      "loss": 2.7084,
      "theoretical_loss": 3.879451081119158,
      "tokens_seen": 544342016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000569635426517874,
      "loss": 2.8031,
      "theoretical_loss": 3.8794015117689886,
      "tokens_seen": 544407552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005696188764895025,
      "loss": 2.6806,
      "theoretical_loss": 3.8793519500562192,
      "tokens_seen": 544473088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000569602326461131,
      "loss": 2.7595,
      "theoretical_loss": 3.879302395978754,
      "tokens_seen": 544538624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005695857764327596,
      "loss": 2.7307,
      "theoretical_loss": 3.879252849534498,
      "tokens_seen": 544604160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005695692264043881,
      "loss": 2.7738,
      "theoretical_loss": 3.8792033107213566,
      "tokens_seen": 544669696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005695526763760166,
      "loss": 2.825,
      "theoretical_loss": 3.8791537795372366,
      "tokens_seen": 544735232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005695361263476452,
      "loss": 2.7518,
      "theoretical_loss": 3.879104255980046,
      "tokens_seen": 544800768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005695195763192737,
      "loss": 3.0515,
      "theoretical_loss": 3.8790547400476925,
      "tokens_seen": 544866304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005695030262909021,
      "loss": 2.9219,
      "theoretical_loss": 3.879005231738086,
      "tokens_seen": 544931840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005694864762625307,
      "loss": 2.9776,
      "theoretical_loss": 3.878955731049136,
      "tokens_seen": 544997376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005694699262341592,
      "loss": 2.9384,
      "theoretical_loss": 3.878906237978753,
      "tokens_seen": 545062912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005694533762057878,
      "loss": 2.9471,
      "theoretical_loss": 3.8788567525248494,
      "tokens_seen": 545128448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005694368261774163,
      "loss": 2.7994,
      "theoretical_loss": 3.878807274685337,
      "tokens_seen": 545193984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005694202761490448,
      "loss": 2.9106,
      "theoretical_loss": 3.8787578044581292,
      "tokens_seen": 545259520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005694037261206734,
      "loss": 2.8742,
      "theoretical_loss": 3.87870834184114,
      "tokens_seen": 545325056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005693871760923018,
      "loss": 2.8766,
      "theoretical_loss": 3.8786588868322847,
      "tokens_seen": 545390592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005693706260639303,
      "loss": 2.866,
      "theoretical_loss": 3.8786094394294794,
      "tokens_seen": 545456128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005693540760355589,
      "loss": 2.9655,
      "theoretical_loss": 3.8785599996306397,
      "tokens_seen": 545521664
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005693375260071874,
      "loss": 3.0036,
      "theoretical_loss": 3.8785105674336835,
      "tokens_seen": 545587200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000569320975978816,
      "loss": 2.828,
      "theoretical_loss": 3.8784611428365285,
      "tokens_seen": 545652736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005693044259504445,
      "loss": 3.0111,
      "theoretical_loss": 3.8784117258370943,
      "tokens_seen": 545718272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000569287875922073,
      "loss": 2.8901,
      "theoretical_loss": 3.878362316433301,
      "tokens_seen": 545783808
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005692713258937016,
      "loss": 2.7338,
      "theoretical_loss": 3.878312914623068,
      "tokens_seen": 545849344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00056925477586533,
      "loss": 2.8399,
      "theoretical_loss": 3.8782635204043188,
      "tokens_seen": 545914880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0005692382258369585,
      "loss": 3.096,
      "theoretical_loss": 3.8782141337749736,
      "tokens_seen": 545980416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005692216758085871,
      "loss": 2.7218,
      "theoretical_loss": 3.8781647547329565,
      "tokens_seen": 546045952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005692051257802156,
      "loss": 3.0483,
      "theoretical_loss": 3.8781153832761914,
      "tokens_seen": 546111488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005691885757518442,
      "loss": 2.8558,
      "theoretical_loss": 3.8780660194026035,
      "tokens_seen": 546177024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005691720257234727,
      "loss": 2.9109,
      "theoretical_loss": 3.8780166631101176,
      "tokens_seen": 546242560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005691554756951012,
      "loss": 2.8958,
      "theoretical_loss": 3.8779673143966606,
      "tokens_seen": 546308096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005691389256667297,
      "loss": 2.7734,
      "theoretical_loss": 3.8779179732601587,
      "tokens_seen": 546373632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005691223756383582,
      "loss": 2.8726,
      "theoretical_loss": 3.8778686396985416,
      "tokens_seen": 546439168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005691058256099867,
      "loss": 2.8268,
      "theoretical_loss": 3.877819313709736,
      "tokens_seen": 546504704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005690892755816153,
      "loss": 2.7565,
      "theoretical_loss": 3.8777699952916738,
      "tokens_seen": 546570240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005690727255532438,
      "loss": 2.7801,
      "theoretical_loss": 3.8777206844422842,
      "tokens_seen": 546635776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005690561755248724,
      "loss": 2.7608,
      "theoretical_loss": 3.877671381159499,
      "tokens_seen": 546701312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005690396254965009,
      "loss": 2.485,
      "theoretical_loss": 3.8776220854412493,
      "tokens_seen": 546766848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005690230754681293,
      "loss": 2.7562,
      "theoretical_loss": 3.877572797285469,
      "tokens_seen": 546832384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005690065254397579,
      "loss": 3.1246,
      "theoretical_loss": 3.8775235166900917,
      "tokens_seen": 546897920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005689899754113864,
      "loss": 2.7631,
      "theoretical_loss": 3.877474243653051,
      "tokens_seen": 546963456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005689734253830149,
      "loss": 2.9907,
      "theoretical_loss": 3.8774249781722836,
      "tokens_seen": 547028992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005689568753546435,
      "loss": 2.9773,
      "theoretical_loss": 3.877375720245725,
      "tokens_seen": 547094528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000568940325326272,
      "loss": 2.8972,
      "theoretical_loss": 3.8773264698713117,
      "tokens_seen": 547160064
    },
    {
      "epoch": 0.2,
      "objective/train/avg_token_score": 0.007785331923514605,
      "objective/train/avg_weight": 0.9908349514007568,
      "objective/train/docs_used": 318614,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7605834007263184,
      "objective/train/std_weight": 0.03869599103927612,
      "objective/train/theoretical_loss": 3.8772772270469824,
      "objective/train/tokens_used": 567685600,
      "theoretical_loss": 3.8772772270469824,
      "tokens_seen": 547225600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005689237752979006,
      "loss": 2.737,
      "theoretical_loss": 3.8772772270469824,
      "tokens_seen": 547225600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005689072252695291,
      "loss": 2.9874,
      "theoretical_loss": 3.877227991770675,
      "tokens_seen": 547291136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005688906752411575,
      "loss": 2.9183,
      "theoretical_loss": 3.877178764040329,
      "tokens_seen": 547356672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005688741252127861,
      "loss": 2.867,
      "theoretical_loss": 3.8771295438538846,
      "tokens_seen": 547422208
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005688575751844146,
      "loss": 2.8519,
      "theoretical_loss": 3.8770803312092834,
      "tokens_seen": 547487744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005688410251560431,
      "loss": 2.7934,
      "theoretical_loss": 3.877031126104466,
      "tokens_seen": 547553280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005688244751276717,
      "loss": 2.6857,
      "theoretical_loss": 3.8769819285373766,
      "tokens_seen": 547618816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005688079250993002,
      "loss": 2.8751,
      "theoretical_loss": 3.876932738505957,
      "tokens_seen": 547684352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005687913750709288,
      "loss": 2.5881,
      "theoretical_loss": 3.8768835560081527,
      "tokens_seen": 547749888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005687748250425572,
      "loss": 2.7441,
      "theoretical_loss": 3.876834381041908,
      "tokens_seen": 547815424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005687582750141857,
      "loss": 2.7587,
      "theoretical_loss": 3.8767852136051686,
      "tokens_seen": 547880960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005687417249858142,
      "loss": 2.8782,
      "theoretical_loss": 3.8767360536958817,
      "tokens_seen": 547946496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005687251749574427,
      "loss": 2.8489,
      "theoretical_loss": 3.876686901311995,
      "tokens_seen": 548012032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005687086249290713,
      "loss": 2.7234,
      "theoretical_loss": 3.8766377564514563,
      "tokens_seen": 548077568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005686920749006998,
      "loss": 2.8685,
      "theoretical_loss": 3.8765886191122148,
      "tokens_seen": 548143104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005686755248723284,
      "loss": 2.9033,
      "theoretical_loss": 3.87653948929222,
      "tokens_seen": 548208640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005686589748439568,
      "loss": 2.9764,
      "theoretical_loss": 3.876490366989423,
      "tokens_seen": 548274176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005686424248155853,
      "loss": 2.7902,
      "theoretical_loss": 3.876441252201775,
      "tokens_seen": 548339712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005686258747872139,
      "loss": 2.9147,
      "theoretical_loss": 3.876392144927229,
      "tokens_seen": 548405248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005686093247588424,
      "loss": 2.7086,
      "theoretical_loss": 3.876343045163737,
      "tokens_seen": 548470784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005685927747304709,
      "loss": 2.8796,
      "theoretical_loss": 3.8762939529092537,
      "tokens_seen": 548536320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005685762247020995,
      "loss": 2.8729,
      "theoretical_loss": 3.8762448681617334,
      "tokens_seen": 548601856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000568559674673728,
      "loss": 2.6767,
      "theoretical_loss": 3.876195790919132,
      "tokens_seen": 548667392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005685431246453566,
      "loss": 2.5743,
      "theoretical_loss": 3.876146721179405,
      "tokens_seen": 548732928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000568526574616985,
      "loss": 2.757,
      "theoretical_loss": 3.8760976589405107,
      "tokens_seen": 548798464
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005685100245886135,
      "loss": 3.0053,
      "theoretical_loss": 3.8760486042004056,
      "tokens_seen": 548864000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005684934745602421,
      "loss": 2.865,
      "theoretical_loss": 3.8759995569570496,
      "tokens_seen": 548929536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005684769245318706,
      "loss": 2.7942,
      "theoretical_loss": 3.8759505172084014,
      "tokens_seen": 548995072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005684603745034991,
      "loss": 2.7525,
      "theoretical_loss": 3.875901484952422,
      "tokens_seen": 549060608
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005684438244751277,
      "loss": 2.9315,
      "theoretical_loss": 3.8758524601870716,
      "tokens_seen": 549126144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005684272744467562,
      "loss": 3.0008,
      "theoretical_loss": 3.8758034429103128,
      "tokens_seen": 549191680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005684107244183847,
      "loss": 3.1478,
      "theoretical_loss": 3.8757544331201084,
      "tokens_seen": 549257216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005683941743900132,
      "loss": 2.7716,
      "theoretical_loss": 3.875705430814421,
      "tokens_seen": 549322752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005683776243616417,
      "loss": 2.5761,
      "theoretical_loss": 3.8756564359912153,
      "tokens_seen": 549388288
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005683610743332703,
      "loss": 2.7882,
      "theoretical_loss": 3.8756074486484566,
      "tokens_seen": 549453824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005683445243048988,
      "loss": 2.9009,
      "theoretical_loss": 3.8755584687841114,
      "tokens_seen": 549519360
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005683279742765273,
      "loss": 2.5743,
      "theoretical_loss": 3.8755094963961447,
      "tokens_seen": 549584896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005683114242481559,
      "loss": 2.9945,
      "theoretical_loss": 3.875460531482526,
      "tokens_seen": 549650432
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005682948742197843,
      "loss": 2.73,
      "theoretical_loss": 3.8754115740412214,
      "tokens_seen": 549715968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005682783241914129,
      "loss": 2.8048,
      "theoretical_loss": 3.8753626240702013,
      "tokens_seen": 549781504
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005682617741630414,
      "loss": 2.5706,
      "theoretical_loss": 3.8753136815674356,
      "tokens_seen": 549847040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005682452241346699,
      "loss": 2.8518,
      "theoretical_loss": 3.875264746530894,
      "tokens_seen": 549912576
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005682286741062985,
      "loss": 2.8375,
      "theoretical_loss": 3.875215818958549,
      "tokens_seen": 549978112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000568212124077927,
      "loss": 2.838,
      "theoretical_loss": 3.875166898848372,
      "tokens_seen": 550043648
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005681955740495555,
      "loss": 2.9189,
      "theoretical_loss": 3.8751179861983367,
      "tokens_seen": 550109184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005681790240211841,
      "loss": 2.7623,
      "theoretical_loss": 3.8750690810064166,
      "tokens_seen": 550174720
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005681624739928125,
      "loss": 2.6071,
      "theoretical_loss": 3.8750201832705864,
      "tokens_seen": 550240256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005681459239644411,
      "loss": 2.7933,
      "theoretical_loss": 3.874971292988821,
      "tokens_seen": 550305792
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005681293739360696,
      "loss": 2.8942,
      "theoretical_loss": 3.8749224101590976,
      "tokens_seen": 550371328
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005681128239076981,
      "loss": 2.9913,
      "theoretical_loss": 3.8748735347793923,
      "tokens_seen": 550436864
    },
    {
      "epoch": 0.2,
      "objective/train/avg_token_score": 0.02073753997683525,
      "objective/train/avg_weight": 0.9755849838256836,
      "objective/train/docs_used": 320433,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.310927867889404,
      "objective/train/std_weight": 0.13563109934329987,
      "objective/train/theoretical_loss": 3.8748246668476827,
      "objective/train/tokens_used": 570962400,
      "theoretical_loss": 3.8748246668476827,
      "tokens_seen": 550502400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005680962738793267,
      "loss": 2.7844,
      "theoretical_loss": 3.8748246668476827,
      "tokens_seen": 550502400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005680797238509552,
      "loss": 2.5851,
      "theoretical_loss": 3.874775806361948,
      "tokens_seen": 550567936
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005680631738225837,
      "loss": 2.9542,
      "theoretical_loss": 3.8747269533201676,
      "tokens_seen": 550633472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005680466237942123,
      "loss": 2.6341,
      "theoretical_loss": 3.874678107720321,
      "tokens_seen": 550699008
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005680300737658407,
      "loss": 2.7719,
      "theoretical_loss": 3.8746292695603897,
      "tokens_seen": 550764544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005680135237374693,
      "loss": 2.7634,
      "theoretical_loss": 3.874580438838355,
      "tokens_seen": 550830080
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005679969737090978,
      "loss": 2.8439,
      "theoretical_loss": 3.8745316155521996,
      "tokens_seen": 550895616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005679804236807263,
      "loss": 2.9179,
      "theoretical_loss": 3.8744827996999067,
      "tokens_seen": 550961152
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005679638736523549,
      "loss": 2.9752,
      "theoretical_loss": 3.874433991279461,
      "tokens_seen": 551026688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005679473236239834,
      "loss": 2.8199,
      "theoretical_loss": 3.874385190288846,
      "tokens_seen": 551092224
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005679307735956119,
      "loss": 2.8943,
      "theoretical_loss": 3.874336396726049,
      "tokens_seen": 551157760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005679142235672404,
      "loss": 2.7013,
      "theoretical_loss": 3.8742876105890547,
      "tokens_seen": 551223296
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005678976735388689,
      "loss": 2.8094,
      "theoretical_loss": 3.874238831875852,
      "tokens_seen": 551288832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005678811235104975,
      "loss": 2.9459,
      "theoretical_loss": 3.8741900605844273,
      "tokens_seen": 551354368
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000567864573482126,
      "loss": 3.0224,
      "theoretical_loss": 3.8741412967127706,
      "tokens_seen": 551419904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005678480234537545,
      "loss": 2.5615,
      "theoretical_loss": 3.8740925402588706,
      "tokens_seen": 551485440
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005678314734253831,
      "loss": 2.819,
      "theoretical_loss": 3.8740437912207186,
      "tokens_seen": 551550976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005678149233970116,
      "loss": 2.8439,
      "theoretical_loss": 3.873995049596305,
      "tokens_seen": 551616512
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00056779837336864,
      "loss": 3.0657,
      "theoretical_loss": 3.873946315383622,
      "tokens_seen": 551682048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005677818233402686,
      "loss": 3.0279,
      "theoretical_loss": 3.873897588580662,
      "tokens_seen": 551747584
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000567765273311897,
      "loss": 3.0097,
      "theoretical_loss": 3.8738488691854194,
      "tokens_seen": 551813120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005677487232835256,
      "loss": 2.6301,
      "theoretical_loss": 3.873800157195887,
      "tokens_seen": 551878656
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005677321732551541,
      "loss": 2.9597,
      "theoretical_loss": 3.873751452610061,
      "tokens_seen": 551944192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005677156232267827,
      "loss": 2.7569,
      "theoretical_loss": 3.8737027554259367,
      "tokens_seen": 552009728
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005676990731984112,
      "loss": 3.067,
      "theoretical_loss": 3.873654065641511,
      "tokens_seen": 552075264
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005676825231700396,
      "loss": 3.0781,
      "theoretical_loss": 3.8736053832547808,
      "tokens_seen": 552140800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005676659731416682,
      "loss": 2.7379,
      "theoretical_loss": 3.873556708263745,
      "tokens_seen": 552206336
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005676494231132967,
      "loss": 3.059,
      "theoretical_loss": 3.873508040666402,
      "tokens_seen": 552271872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005676328730849252,
      "loss": 2.6448,
      "theoretical_loss": 3.8734593804607513,
      "tokens_seen": 552337408
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005676163230565538,
      "loss": 2.7281,
      "theoretical_loss": 3.873410727644794,
      "tokens_seen": 552402944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005675997730281823,
      "loss": 2.9178,
      "theoretical_loss": 3.8733620822165316,
      "tokens_seen": 552468480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005675832229998109,
      "loss": 2.7099,
      "theoretical_loss": 3.8733134441739647,
      "tokens_seen": 552534016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005675666729714394,
      "loss": 2.8334,
      "theoretical_loss": 3.8732648135150978,
      "tokens_seen": 552599552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005675501229430678,
      "loss": 2.8738,
      "theoretical_loss": 3.8732161902379336,
      "tokens_seen": 552665088
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005675335729146964,
      "loss": 2.9917,
      "theoretical_loss": 3.873167574340477,
      "tokens_seen": 552730624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005675170228863249,
      "loss": 2.9264,
      "theoretical_loss": 3.873118965820733,
      "tokens_seen": 552796160
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005675004728579535,
      "loss": 2.8226,
      "theoretical_loss": 3.873070364676707,
      "tokens_seen": 552861696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000567483922829582,
      "loss": 2.8726,
      "theoretical_loss": 3.8730217709064063,
      "tokens_seen": 552927232
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005674673728012105,
      "loss": 2.8043,
      "theoretical_loss": 3.8729731845078383,
      "tokens_seen": 552992768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005674508227728391,
      "loss": 2.809,
      "theoretical_loss": 3.872924605479011,
      "tokens_seen": 553058304
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005674342727444675,
      "loss": 2.7301,
      "theoretical_loss": 3.872876033817934,
      "tokens_seen": 553123840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000567417722716096,
      "loss": 2.8619,
      "theoretical_loss": 3.8728274695226164,
      "tokens_seen": 553189376
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005674011726877246,
      "loss": 2.9643,
      "theoretical_loss": 3.872778912591069,
      "tokens_seen": 553254912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005673846226593531,
      "loss": 2.8178,
      "theoretical_loss": 3.872730363021304,
      "tokens_seen": 553320448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005673680726309817,
      "loss": 2.7749,
      "theoretical_loss": 3.8726818208113323,
      "tokens_seen": 553385984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005673515226026102,
      "loss": 3.0643,
      "theoretical_loss": 3.872633285959167,
      "tokens_seen": 553451520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005673349725742387,
      "loss": 2.8976,
      "theoretical_loss": 3.872584758462823,
      "tokens_seen": 553517056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005673184225458673,
      "loss": 2.667,
      "theoretical_loss": 3.872536238320313,
      "tokens_seen": 553582592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005673018725174957,
      "loss": 2.758,
      "theoretical_loss": 3.8724877255296533,
      "tokens_seen": 553648128
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005672853224891242,
      "loss": 2.8804,
      "theoretical_loss": 3.872439220088859,
      "tokens_seen": 553713664
    },
    {
      "epoch": 0.2,
      "objective/train/avg_token_score": 0.03743475675582886,
      "objective/train/avg_weight": 0.9559603929519653,
      "objective/train/docs_used": 322451,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.956345796585083,
      "objective/train/std_weight": 0.15378354489803314,
      "objective/train/theoretical_loss": 3.8723907219959486,
      "objective/train/tokens_used": 574239200,
      "theoretical_loss": 3.8723907219959486,
      "tokens_seen": 553779200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005672687724607528,
      "loss": 2.8079,
      "theoretical_loss": 3.8723907219959486,
      "tokens_seen": 553779200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005672522224323813,
      "loss": 2.7404,
      "theoretical_loss": 3.872342231248938,
      "tokens_seen": 553844736
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005672356724040099,
      "loss": 2.8133,
      "theoretical_loss": 3.872293747845846,
      "tokens_seen": 553910272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005672191223756384,
      "loss": 2.8826,
      "theoretical_loss": 3.872245271784692,
      "tokens_seen": 553975808
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005672025723472669,
      "loss": 2.9475,
      "theoretical_loss": 3.872196803063495,
      "tokens_seen": 554041344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005671860223188954,
      "loss": 2.8545,
      "theoretical_loss": 3.872148341680276,
      "tokens_seen": 554106880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005671694722905239,
      "loss": 2.7733,
      "theoretical_loss": 3.8720998876330572,
      "tokens_seen": 554172416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005671529222621524,
      "loss": 2.9583,
      "theoretical_loss": 3.8720514409198596,
      "tokens_seen": 554237952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000567136372233781,
      "loss": 2.9999,
      "theoretical_loss": 3.8720030015387064,
      "tokens_seen": 554303488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005671198222054095,
      "loss": 2.8813,
      "theoretical_loss": 3.871954569487622,
      "tokens_seen": 554369024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005671032721770381,
      "loss": 2.6389,
      "theoretical_loss": 3.8719061447646297,
      "tokens_seen": 554434560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005670867221486666,
      "loss": 2.7285,
      "theoretical_loss": 3.8718577273677557,
      "tokens_seen": 554500096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000567070172120295,
      "loss": 2.7797,
      "theoretical_loss": 3.8718093172950248,
      "tokens_seen": 554565632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005670536220919236,
      "loss": 2.8605,
      "theoretical_loss": 3.8717609145444647,
      "tokens_seen": 554631168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005670370720635521,
      "loss": 2.7812,
      "theoretical_loss": 3.8717125191141024,
      "tokens_seen": 554696704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005670205220351806,
      "loss": 2.8694,
      "theoretical_loss": 3.871664131001967,
      "tokens_seen": 554762240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005670039720068092,
      "loss": 2.8602,
      "theoretical_loss": 3.8716157502060864,
      "tokens_seen": 554827776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005669874219784377,
      "loss": 2.7254,
      "theoretical_loss": 3.8715673767244914,
      "tokens_seen": 554893312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005669708719500663,
      "loss": 2.8455,
      "theoretical_loss": 3.871519010555211,
      "tokens_seen": 554958848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005669543219216948,
      "loss": 2.9608,
      "theoretical_loss": 3.8714706516962787,
      "tokens_seen": 555024384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005669377718933232,
      "loss": 2.773,
      "theoretical_loss": 3.871422300145725,
      "tokens_seen": 555089920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005669212218649518,
      "loss": 2.9256,
      "theoretical_loss": 3.8713739559015834,
      "tokens_seen": 555155456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005669046718365803,
      "loss": 2.7577,
      "theoretical_loss": 3.871325618961887,
      "tokens_seen": 555220992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005668881218082088,
      "loss": 3.0187,
      "theoretical_loss": 3.871277289324671,
      "tokens_seen": 555286528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005668715717798374,
      "loss": 3.0813,
      "theoretical_loss": 3.871228966987969,
      "tokens_seen": 555352064
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005668550217514659,
      "loss": 2.8408,
      "theoretical_loss": 3.871180651949819,
      "tokens_seen": 555417600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005668384717230945,
      "loss": 2.8417,
      "theoretical_loss": 3.871132344208256,
      "tokens_seen": 555483136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005668219216947229,
      "loss": 2.7932,
      "theoretical_loss": 3.871084043761318,
      "tokens_seen": 555548672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005668053716663514,
      "loss": 2.9394,
      "theoretical_loss": 3.8710357506070427,
      "tokens_seen": 555614208
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00056678882163798,
      "loss": 3.0394,
      "theoretical_loss": 3.87098746474347,
      "tokens_seen": 555679744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005667722716096084,
      "loss": 3.1026,
      "theoretical_loss": 3.870939186168638,
      "tokens_seen": 555745280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000566755721581237,
      "loss": 2.6776,
      "theoretical_loss": 3.8708909148805892,
      "tokens_seen": 555810816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005667391715528655,
      "loss": 2.7833,
      "theoretical_loss": 3.8708426508773632,
      "tokens_seen": 555876352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000566722621524494,
      "loss": 2.7136,
      "theoretical_loss": 3.870794394157002,
      "tokens_seen": 555941888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005667060714961225,
      "loss": 2.9603,
      "theoretical_loss": 3.8707461447175495,
      "tokens_seen": 556007424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000566689521467751,
      "loss": 2.8996,
      "theoretical_loss": 3.8706979025570485,
      "tokens_seen": 556072960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005666729714393796,
      "loss": 2.989,
      "theoretical_loss": 3.870649667673543,
      "tokens_seen": 556138496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005666564214110081,
      "loss": 2.5736,
      "theoretical_loss": 3.870601440065078,
      "tokens_seen": 556204032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005666398713826366,
      "loss": 2.9487,
      "theoretical_loss": 3.8705532197296995,
      "tokens_seen": 556269568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005666233213542652,
      "loss": 2.7837,
      "theoretical_loss": 3.8705050066654536,
      "tokens_seen": 556335104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005666067713258937,
      "loss": 3.0138,
      "theoretical_loss": 3.870456800870388,
      "tokens_seen": 556400640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005665902212975221,
      "loss": 3.0404,
      "theoretical_loss": 3.8704086023425512,
      "tokens_seen": 556466176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005665736712691507,
      "loss": 2.7606,
      "theoretical_loss": 3.8703604110799903,
      "tokens_seen": 556531712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005665571212407792,
      "loss": 2.9988,
      "theoretical_loss": 3.870312227080756,
      "tokens_seen": 556597248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005665405712124078,
      "loss": 2.7746,
      "theoretical_loss": 3.8702640503428984,
      "tokens_seen": 556662784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005665240211840363,
      "loss": 2.9542,
      "theoretical_loss": 3.870215880864469,
      "tokens_seen": 556728320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005665074711556648,
      "loss": 2.8186,
      "theoretical_loss": 3.8701677186435184,
      "tokens_seen": 556793856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005664909211272934,
      "loss": 2.6925,
      "theoretical_loss": 3.8701195636781,
      "tokens_seen": 556859392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005664743710989219,
      "loss": 2.9176,
      "theoretical_loss": 3.8700714159662666,
      "tokens_seen": 556924928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005664578210705503,
      "loss": 2.8157,
      "theoretical_loss": 3.870023275506073,
      "tokens_seen": 556990464
    },
    {
      "epoch": 0.2,
      "objective/train/avg_token_score": 0.0075579481199383736,
      "objective/train/avg_weight": 0.991104245185852,
      "objective/train/docs_used": 324068,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.844827175140381,
      "objective/train/std_weight": 0.051473554223775864,
      "objective/train/theoretical_loss": 3.869975142295573,
      "objective/train/tokens_used": 577516000,
      "theoretical_loss": 3.869975142295573,
      "tokens_seen": 557056000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005664412710421789,
      "loss": 2.9618,
      "theoretical_loss": 3.869975142295573,
      "tokens_seen": 557056000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005664247210138074,
      "loss": 3.0513,
      "theoretical_loss": 3.8699270163328223,
      "tokens_seen": 557121536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000566408170985436,
      "loss": 2.8961,
      "theoretical_loss": 3.8698788976158776,
      "tokens_seen": 557187072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005663916209570645,
      "loss": 2.8161,
      "theoretical_loss": 3.8698307861427956,
      "tokens_seen": 557252608
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000566375070928693,
      "loss": 3.0975,
      "theoretical_loss": 3.8697826819116345,
      "tokens_seen": 557318144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005663585209003216,
      "loss": 2.8599,
      "theoretical_loss": 3.8697345849204523,
      "tokens_seen": 557383680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00056634197087195,
      "loss": 2.9138,
      "theoretical_loss": 3.8696864951673087,
      "tokens_seen": 557449216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005663254208435785,
      "loss": 2.867,
      "theoretical_loss": 3.8696384126502634,
      "tokens_seen": 557514752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005663088708152071,
      "loss": 3.0161,
      "theoretical_loss": 3.8695903373673772,
      "tokens_seen": 557580288
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005662923207868356,
      "loss": 2.8689,
      "theoretical_loss": 3.869542269316711,
      "tokens_seen": 557645824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005662757707584642,
      "loss": 2.7853,
      "theoretical_loss": 3.869494208496329,
      "tokens_seen": 557711360
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005662592207300927,
      "loss": 3.1159,
      "theoretical_loss": 3.8694461549042924,
      "tokens_seen": 557776896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005662426707017212,
      "loss": 2.8575,
      "theoretical_loss": 3.8693981085386655,
      "tokens_seen": 557842432
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005662261206733498,
      "loss": 2.7224,
      "theoretical_loss": 3.8693500693975125,
      "tokens_seen": 557907968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005662095706449782,
      "loss": 2.6867,
      "theoretical_loss": 3.8693020374788993,
      "tokens_seen": 557973504
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005661930206166067,
      "loss": 3.0561,
      "theoretical_loss": 3.8692540127808916,
      "tokens_seen": 558039040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005661764705882353,
      "loss": 2.8713,
      "theoretical_loss": 3.8692059953015563,
      "tokens_seen": 558104576
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005661599205598638,
      "loss": 2.9682,
      "theoretical_loss": 3.86915798503896,
      "tokens_seen": 558170112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005661433705314924,
      "loss": 3.1298,
      "theoretical_loss": 3.8691099819911723,
      "tokens_seen": 558235648
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005661268205031209,
      "loss": 3.0455,
      "theoretical_loss": 3.869061986156261,
      "tokens_seen": 558301184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005661102704747494,
      "loss": 3.0617,
      "theoretical_loss": 3.8690139975322966,
      "tokens_seen": 558366720
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000566093720446378,
      "loss": 2.6864,
      "theoretical_loss": 3.8689660161173487,
      "tokens_seen": 558432256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005660771704180064,
      "loss": 2.972,
      "theoretical_loss": 3.8689180419094895,
      "tokens_seen": 558497792
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005660606203896349,
      "loss": 2.9623,
      "theoretical_loss": 3.8688700749067904,
      "tokens_seen": 558563328
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005660440703612635,
      "loss": 2.8387,
      "theoretical_loss": 3.868822115107324,
      "tokens_seen": 558628864
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000566027520332892,
      "loss": 2.9395,
      "theoretical_loss": 3.8687741625091636,
      "tokens_seen": 558694400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005660109703045206,
      "loss": 2.9254,
      "theoretical_loss": 3.8687262171103836,
      "tokens_seen": 558759936
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005659944202761491,
      "loss": 2.9761,
      "theoretical_loss": 3.8686782789090595,
      "tokens_seen": 558825472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005659778702477775,
      "loss": 2.6934,
      "theoretical_loss": 3.868630347903266,
      "tokens_seen": 558891008
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005659613202194061,
      "loss": 2.6834,
      "theoretical_loss": 3.8685824240910804,
      "tokens_seen": 558956544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005659447701910346,
      "loss": 2.805,
      "theoretical_loss": 3.868534507470579,
      "tokens_seen": 559022080
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005659282201626632,
      "loss": 2.9328,
      "theoretical_loss": 3.8684865980398393,
      "tokens_seen": 559087616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005659116701342917,
      "loss": 2.8177,
      "theoretical_loss": 3.868438695796941,
      "tokens_seen": 559153152
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005658951201059202,
      "loss": 3.1159,
      "theoretical_loss": 3.8683908007399634,
      "tokens_seen": 559218688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005658785700775488,
      "loss": 2.64,
      "theoretical_loss": 3.8683429128669857,
      "tokens_seen": 559284224
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005658620200491773,
      "loss": 2.9902,
      "theoretical_loss": 3.8682950321760896,
      "tokens_seen": 559349760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005658454700208057,
      "loss": 2.8941,
      "theoretical_loss": 3.868247158665356,
      "tokens_seen": 559415296
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005658289199924343,
      "loss": 3.0366,
      "theoretical_loss": 3.8681992923328674,
      "tokens_seen": 559480832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005658123699640628,
      "loss": 2.9009,
      "theoretical_loss": 3.868151433176707,
      "tokens_seen": 559546368
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005657958199356913,
      "loss": 3.0345,
      "theoretical_loss": 3.8681035811949585,
      "tokens_seen": 559611904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005657792699073198,
      "loss": 3.0679,
      "theoretical_loss": 3.868055736385706,
      "tokens_seen": 559677440
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005657627198789484,
      "loss": 3.0989,
      "theoretical_loss": 3.868007898747036,
      "tokens_seen": 559742976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005657461698505769,
      "loss": 2.8007,
      "theoretical_loss": 3.867960068277033,
      "tokens_seen": 559808512
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005657296198222053,
      "loss": 2.8801,
      "theoretical_loss": 3.867912244973784,
      "tokens_seen": 559874048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005657130697938339,
      "loss": 2.9544,
      "theoretical_loss": 3.867864428835377,
      "tokens_seen": 559939584
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005656965197654624,
      "loss": 2.8708,
      "theoretical_loss": 3.8678166198598998,
      "tokens_seen": 560005120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005656799697370909,
      "loss": 2.9144,
      "theoretical_loss": 3.8677688180454415,
      "tokens_seen": 560070656
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005656634197087195,
      "loss": 2.7031,
      "theoretical_loss": 3.867721023390091,
      "tokens_seen": 560136192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000565646869680348,
      "loss": 3.0339,
      "theoretical_loss": 3.8676732358919406,
      "tokens_seen": 560201728
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005656303196519766,
      "loss": 2.9469,
      "theoretical_loss": 3.8676254555490788,
      "tokens_seen": 560267264
    },
    {
      "epoch": 0.2,
      "objective/train/avg_token_score": 0.011408049613237381,
      "objective/train/avg_weight": 0.9865744113922119,
      "objective/train/docs_used": 325860,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.061556577682495,
      "objective/train/std_weight": 0.06621329486370087,
      "objective/train/theoretical_loss": 3.8675776823595998,
      "objective/train/tokens_used": 580792800,
      "theoretical_loss": 3.8675776823595998,
      "tokens_seen": 560332800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000565613769623605,
      "loss": 2.9144,
      "theoretical_loss": 3.8675776823595998,
      "tokens_seen": 560332800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005655972195952335,
      "loss": 2.8093,
      "theoretical_loss": 3.867529916321595,
      "tokens_seen": 560398336
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005655806695668621,
      "loss": 3.0466,
      "theoretical_loss": 3.867482157433157,
      "tokens_seen": 560463872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005655641195384906,
      "loss": 3.0723,
      "theoretical_loss": 3.8674344056923813,
      "tokens_seen": 560529408
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005655475695101191,
      "loss": 2.9286,
      "theoretical_loss": 3.8673866610973624,
      "tokens_seen": 560594944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005655310194817477,
      "loss": 3.1956,
      "theoretical_loss": 3.8673389236461952,
      "tokens_seen": 560660480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005655144694533762,
      "loss": 2.6956,
      "theoretical_loss": 3.867291193336976,
      "tokens_seen": 560726016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005654979194250048,
      "loss": 2.7583,
      "theoretical_loss": 3.8672434701678022,
      "tokens_seen": 560791552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005654813693966332,
      "loss": 2.8186,
      "theoretical_loss": 3.867195754136771,
      "tokens_seen": 560857088
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005654648193682617,
      "loss": 3.1345,
      "theoretical_loss": 3.867148045241981,
      "tokens_seen": 560922624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005654482693398903,
      "loss": 2.7941,
      "theoretical_loss": 3.867100343481532,
      "tokens_seen": 560988160
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005654317193115188,
      "loss": 2.769,
      "theoretical_loss": 3.8670526488535226,
      "tokens_seen": 561053696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005654151692831473,
      "loss": 2.8589,
      "theoretical_loss": 3.867004961356054,
      "tokens_seen": 561119232
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005653986192547759,
      "loss": 2.9227,
      "theoretical_loss": 3.8669572809872284,
      "tokens_seen": 561184768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005653820692264044,
      "loss": 3.0006,
      "theoretical_loss": 3.8669096077451464,
      "tokens_seen": 561250304
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000565365519198033,
      "loss": 2.8466,
      "theoretical_loss": 3.8668619416279117,
      "tokens_seen": 561315840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005653489691696614,
      "loss": 3.0166,
      "theoretical_loss": 3.8668142826336274,
      "tokens_seen": 561381376
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005653324191412899,
      "loss": 2.7393,
      "theoretical_loss": 3.866766630760398,
      "tokens_seen": 561446912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005653158691129185,
      "loss": 2.7044,
      "theoretical_loss": 3.8667189860063282,
      "tokens_seen": 561512448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000565299319084547,
      "loss": 2.8312,
      "theoretical_loss": 3.866671348369524,
      "tokens_seen": 561577984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005652827690561755,
      "loss": 2.8037,
      "theoretical_loss": 3.866623717848092,
      "tokens_seen": 561643520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005652662190278041,
      "loss": 3.1379,
      "theoretical_loss": 3.8665760944401377,
      "tokens_seen": 561709056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005652496689994325,
      "loss": 3.0686,
      "theoretical_loss": 3.8665284781437714,
      "tokens_seen": 561774592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005652331189710611,
      "loss": 2.8652,
      "theoretical_loss": 3.8664808689571,
      "tokens_seen": 561840128
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005652331189710611,
      "loss": 2.8202,
      "theoretical_loss": 3.866433266878234,
      "tokens_seen": 561905664
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005652165689426896,
      "loss": 2.661,
      "theoretical_loss": 3.8663856719052827,
      "tokens_seen": 561971200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005652000189143181,
      "loss": 2.873,
      "theoretical_loss": 3.8663380840363564,
      "tokens_seen": 562036736
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005651834688859467,
      "loss": 2.9652,
      "theoretical_loss": 3.866290503269567,
      "tokens_seen": 562102272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005651669188575752,
      "loss": 2.7007,
      "theoretical_loss": 3.866242929603027,
      "tokens_seen": 562167808
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005651503688292037,
      "loss": 2.7661,
      "theoretical_loss": 3.8661953630348496,
      "tokens_seen": 562233344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005651338188008323,
      "loss": 2.854,
      "theoretical_loss": 3.866147803563148,
      "tokens_seen": 562298880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005651172687724607,
      "loss": 3.0835,
      "theoretical_loss": 3.866100251186036,
      "tokens_seen": 562364416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005651007187440893,
      "loss": 3.2007,
      "theoretical_loss": 3.8660527059016294,
      "tokens_seen": 562429952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005650841687157178,
      "loss": 3.0113,
      "theoretical_loss": 3.8660051677080434,
      "tokens_seen": 562495488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005650676186873463,
      "loss": 2.8316,
      "theoretical_loss": 3.8659576366033956,
      "tokens_seen": 562561024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005650510686589749,
      "loss": 2.9195,
      "theoretical_loss": 3.865910112585802,
      "tokens_seen": 562626560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005650345186306034,
      "loss": 2.9524,
      "theoretical_loss": 3.8658625956533816,
      "tokens_seen": 562692096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005650179686022319,
      "loss": 3.0612,
      "theoretical_loss": 3.8658150858042526,
      "tokens_seen": 562757632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005650014185738604,
      "loss": 2.9651,
      "theoretical_loss": 3.8657675830365346,
      "tokens_seen": 562823168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005649848685454889,
      "loss": 2.9175,
      "theoretical_loss": 3.865720087348347,
      "tokens_seen": 562888704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005649683185171175,
      "loss": 2.6587,
      "theoretical_loss": 3.8656725987378118,
      "tokens_seen": 562954240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000564951768488746,
      "loss": 2.6469,
      "theoretical_loss": 3.8656251172030496,
      "tokens_seen": 563019776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005649352184603745,
      "loss": 3.2012,
      "theoretical_loss": 3.865577642742183,
      "tokens_seen": 563085312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005649186684320031,
      "loss": 2.6681,
      "theoretical_loss": 3.865530175353335,
      "tokens_seen": 563150848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005649021184036316,
      "loss": 2.6889,
      "theoretical_loss": 3.86548271503463,
      "tokens_seen": 563216384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00056488556837526,
      "loss": 3.0861,
      "theoretical_loss": 3.865435261784191,
      "tokens_seen": 563281920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005648690183468886,
      "loss": 2.7233,
      "theoretical_loss": 3.865387815600144,
      "tokens_seen": 563347456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005648524683185171,
      "loss": 2.9409,
      "theoretical_loss": 3.8653403764806145,
      "tokens_seen": 563412992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005648359182901457,
      "loss": 2.8288,
      "theoretical_loss": 3.86529294442373,
      "tokens_seen": 563478528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005648193682617742,
      "loss": 2.7757,
      "theoretical_loss": 3.8652455194276163,
      "tokens_seen": 563544064
    },
    {
      "epoch": 0.2,
      "objective/train/avg_token_score": 0.012796124443411827,
      "objective/train/avg_weight": 0.9849420189857483,
      "objective/train/docs_used": 327873,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.031938076019287,
      "objective/train/std_weight": 0.04440978169441223,
      "objective/train/theoretical_loss": 3.8651981014904027,
      "objective/train/tokens_used": 584069600,
      "theoretical_loss": 3.8651981014904027,
      "tokens_seen": 563609600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005648028182334027,
      "loss": 2.6885,
      "theoretical_loss": 3.8651981014904027,
      "tokens_seen": 563609600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005647862682050312,
      "loss": 2.9545,
      "theoretical_loss": 3.8651506906102173,
      "tokens_seen": 563675136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005647697181766596,
      "loss": 2.9334,
      "theoretical_loss": 3.8651032867851893,
      "tokens_seen": 563740672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005647531681482882,
      "loss": 2.6036,
      "theoretical_loss": 3.8650558900134495,
      "tokens_seen": 563806208
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005647366181199167,
      "loss": 2.7671,
      "theoretical_loss": 3.8650085002931283,
      "tokens_seen": 563871744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005647200680915453,
      "loss": 2.7442,
      "theoretical_loss": 3.8649611176223573,
      "tokens_seen": 563937280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005647035180631738,
      "loss": 2.7054,
      "theoretical_loss": 3.8649137419992687,
      "tokens_seen": 564002816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005646869680348023,
      "loss": 2.7611,
      "theoretical_loss": 3.864866373421996,
      "tokens_seen": 564068352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005646704180064309,
      "loss": 2.794,
      "theoretical_loss": 3.8648190118886725,
      "tokens_seen": 564133888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005646538679780594,
      "loss": 2.7876,
      "theoretical_loss": 3.8647716573974327,
      "tokens_seen": 564199424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005646373179496878,
      "loss": 3.0538,
      "theoretical_loss": 3.8647243099464106,
      "tokens_seen": 564264960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005646207679213164,
      "loss": 2.8538,
      "theoretical_loss": 3.864676969533744,
      "tokens_seen": 564330496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005646042178929449,
      "loss": 2.9916,
      "theoretical_loss": 3.8646296361575687,
      "tokens_seen": 564396032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005645876678645735,
      "loss": 2.9693,
      "theoretical_loss": 3.8645823098160212,
      "tokens_seen": 564461568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000564571117836202,
      "loss": 2.8191,
      "theoretical_loss": 3.8645349905072406,
      "tokens_seen": 564527104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005645545678078305,
      "loss": 2.9593,
      "theoretical_loss": 3.8644876782293647,
      "tokens_seen": 564592640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005645380177794591,
      "loss": 2.7767,
      "theoretical_loss": 3.864440372980533,
      "tokens_seen": 564658176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005645214677510875,
      "loss": 2.9016,
      "theoretical_loss": 3.8643930747588855,
      "tokens_seen": 564723712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000564504917722716,
      "loss": 2.8296,
      "theoretical_loss": 3.8643457835625634,
      "tokens_seen": 564789248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005644883676943446,
      "loss": 2.8446,
      "theoretical_loss": 3.8642984993897085,
      "tokens_seen": 564854784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005644718176659731,
      "loss": 2.8805,
      "theoretical_loss": 3.864251222238462,
      "tokens_seen": 564920320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005644552676376017,
      "loss": 2.6961,
      "theoretical_loss": 3.8642039521069678,
      "tokens_seen": 564985856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005644387176092302,
      "loss": 2.7985,
      "theoretical_loss": 3.864156688993369,
      "tokens_seen": 565051392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005644221675808587,
      "loss": 2.8787,
      "theoretical_loss": 3.86410943289581,
      "tokens_seen": 565116928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005644056175524873,
      "loss": 2.7786,
      "theoretical_loss": 3.864062183812436,
      "tokens_seen": 565182464
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005643890675241157,
      "loss": 2.6139,
      "theoretical_loss": 3.864014941741392,
      "tokens_seen": 565248000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005643725174957442,
      "loss": 2.7934,
      "theoretical_loss": 3.8639677066808256,
      "tokens_seen": 565313536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005643559674673728,
      "loss": 2.8059,
      "theoretical_loss": 3.863920478628883,
      "tokens_seen": 565379072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005643394174390013,
      "loss": 2.8108,
      "theoretical_loss": 3.8638732575837125,
      "tokens_seen": 565444608
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005643228674106299,
      "loss": 2.9202,
      "theoretical_loss": 3.8638260435434626,
      "tokens_seen": 565510144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005643063173822584,
      "loss": 2.9219,
      "theoretical_loss": 3.863778836506283,
      "tokens_seen": 565575680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005642897673538869,
      "loss": 2.9749,
      "theoretical_loss": 3.8637316364703227,
      "tokens_seen": 565641216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005642732173255154,
      "loss": 3.1364,
      "theoretical_loss": 3.8636844434337325,
      "tokens_seen": 565706752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005642566672971439,
      "loss": 2.4786,
      "theoretical_loss": 3.863637257394665,
      "tokens_seen": 565772288
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005642401172687724,
      "loss": 3.0367,
      "theoretical_loss": 3.863590078351271,
      "tokens_seen": 565837824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000564223567240401,
      "loss": 2.919,
      "theoretical_loss": 3.8635429063017037,
      "tokens_seen": 565903360
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005642070172120295,
      "loss": 2.9093,
      "theoretical_loss": 3.8634957412441167,
      "tokens_seen": 565968896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005641904671836581,
      "loss": 3.0348,
      "theoretical_loss": 3.8634485831766634,
      "tokens_seen": 566034432
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005641739171552866,
      "loss": 3.048,
      "theoretical_loss": 3.8634014320975005,
      "tokens_seen": 566099968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000564157367126915,
      "loss": 2.6298,
      "theoretical_loss": 3.8633542880047815,
      "tokens_seen": 566165504
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005641408170985436,
      "loss": 2.7048,
      "theoretical_loss": 3.863307150896664,
      "tokens_seen": 566231040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005641242670701721,
      "loss": 2.8793,
      "theoretical_loss": 3.863260020771304,
      "tokens_seen": 566296576
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005641077170418006,
      "loss": 2.7749,
      "theoretical_loss": 3.8632128976268603,
      "tokens_seen": 566362112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005640911670134292,
      "loss": 2.7756,
      "theoretical_loss": 3.8631657814614906,
      "tokens_seen": 566427648
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005640746169850577,
      "loss": 2.7472,
      "theoretical_loss": 3.8631186722733535,
      "tokens_seen": 566493184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005640580669566863,
      "loss": 2.715,
      "theoretical_loss": 3.86307157006061,
      "tokens_seen": 566558720
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005640415169283148,
      "loss": 3.0186,
      "theoretical_loss": 3.8630244748214198,
      "tokens_seen": 566624256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005640249668999432,
      "loss": 2.989,
      "theoretical_loss": 3.862977386553944,
      "tokens_seen": 566689792
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005640084168715718,
      "loss": 2.8721,
      "theoretical_loss": 3.862930305256345,
      "tokens_seen": 566755328
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005639918668432003,
      "loss": 2.9603,
      "theoretical_loss": 3.8628832309267844,
      "tokens_seen": 566820864
    },
    {
      "epoch": 0.2,
      "objective/train/avg_token_score": 0.011976204812526703,
      "objective/train/avg_weight": 0.9859269261360168,
      "objective/train/docs_used": 329813,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0266733169555664,
      "objective/train/std_weight": 0.07899615168571472,
      "objective/train/theoretical_loss": 3.8628361635634265,
      "objective/train/tokens_used": 587346400,
      "theoretical_loss": 3.8628361635634265,
      "tokens_seen": 566886400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005639753168148288,
      "loss": 2.8328,
      "theoretical_loss": 3.8628361635634265,
      "tokens_seen": 566886400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005639587667864574,
      "loss": 2.8425,
      "theoretical_loss": 3.862789103164435,
      "tokens_seen": 566951936
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005639422167580859,
      "loss": 2.9382,
      "theoretical_loss": 3.8627420497279745,
      "tokens_seen": 567017472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005639256667297145,
      "loss": 2.991,
      "theoretical_loss": 3.8626950032522096,
      "tokens_seen": 567083008
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000563909116701343,
      "loss": 2.9863,
      "theoretical_loss": 3.8626479637353075,
      "tokens_seen": 567148544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005638925666729714,
      "loss": 2.8338,
      "theoretical_loss": 3.862600931175434,
      "tokens_seen": 567214080
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005638760166446,
      "loss": 3.0086,
      "theoretical_loss": 3.8625539055707576,
      "tokens_seen": 567279616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005638594666162285,
      "loss": 2.7774,
      "theoretical_loss": 3.862506886919445,
      "tokens_seen": 567345152
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000563842916587857,
      "loss": 2.9248,
      "theoretical_loss": 3.862459875219667,
      "tokens_seen": 567410688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005638263665594855,
      "loss": 2.9782,
      "theoretical_loss": 3.862412870469591,
      "tokens_seen": 567476224
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000563809816531114,
      "loss": 2.8742,
      "theoretical_loss": 3.8623658726673886,
      "tokens_seen": 567541760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005637932665027426,
      "loss": 2.8579,
      "theoretical_loss": 3.86231888181123,
      "tokens_seen": 567607296
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000563776716474371,
      "loss": 3.1707,
      "theoretical_loss": 3.862271897899287,
      "tokens_seen": 567672832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005637601664459996,
      "loss": 2.6879,
      "theoretical_loss": 3.862224920929732,
      "tokens_seen": 567738368
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005637436164176281,
      "loss": 3.0301,
      "theoretical_loss": 3.8621779509007377,
      "tokens_seen": 567803904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005637270663892566,
      "loss": 2.6509,
      "theoretical_loss": 3.8621309878104784,
      "tokens_seen": 567869440
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005637105163608852,
      "loss": 2.7477,
      "theoretical_loss": 3.862084031657128,
      "tokens_seen": 567934976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005636939663325137,
      "loss": 2.7581,
      "theoretical_loss": 3.862037082438861,
      "tokens_seen": 568000512
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005636774163041422,
      "loss": 2.7523,
      "theoretical_loss": 3.8619901401538543,
      "tokens_seen": 568066048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005636608662757707,
      "loss": 2.985,
      "theoretical_loss": 3.8619432048002835,
      "tokens_seen": 568131584
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005636443162473992,
      "loss": 2.8298,
      "theoretical_loss": 3.861896276376326,
      "tokens_seen": 568197120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005636277662190278,
      "loss": 2.7406,
      "theoretical_loss": 3.8618493548801602,
      "tokens_seen": 568262656
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005636112161906563,
      "loss": 2.7924,
      "theoretical_loss": 3.8618024403099636,
      "tokens_seen": 568328192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005635946661622848,
      "loss": 2.9132,
      "theoretical_loss": 3.8617555326639157,
      "tokens_seen": 568393728
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005635781161339134,
      "loss": 2.9316,
      "theoretical_loss": 3.8617086319401968,
      "tokens_seen": 568459264
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005635615661055419,
      "loss": 2.705,
      "theoretical_loss": 3.861661738136987,
      "tokens_seen": 568524800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005635450160771703,
      "loss": 2.8782,
      "theoretical_loss": 3.8616148512524675,
      "tokens_seen": 568590336
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005635284660487989,
      "loss": 2.9835,
      "theoretical_loss": 3.8615679712848205,
      "tokens_seen": 568655872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005635119160204274,
      "loss": 2.9406,
      "theoretical_loss": 3.8615210982322288,
      "tokens_seen": 568721408
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000563495365992056,
      "loss": 2.8787,
      "theoretical_loss": 3.861474232092876,
      "tokens_seen": 568786944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005634788159636845,
      "loss": 2.9256,
      "theoretical_loss": 3.8614273728649446,
      "tokens_seen": 568852480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000563462265935313,
      "loss": 2.8933,
      "theoretical_loss": 3.8613805205466205,
      "tokens_seen": 568918016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005634457159069416,
      "loss": 3.0713,
      "theoretical_loss": 3.8613336751360894,
      "tokens_seen": 568983552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00056342916587857,
      "loss": 2.7886,
      "theoretical_loss": 3.861286836631537,
      "tokens_seen": 569049088
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005634126158501986,
      "loss": 2.6533,
      "theoretical_loss": 3.8612400050311493,
      "tokens_seen": 569114624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005633960658218271,
      "loss": 2.7013,
      "theoretical_loss": 3.8611931803331148,
      "tokens_seen": 569180160
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005633795157934556,
      "loss": 2.7441,
      "theoretical_loss": 3.861146362535621,
      "tokens_seen": 569245696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005633629657650842,
      "loss": 2.752,
      "theoretical_loss": 3.8610995516368565,
      "tokens_seen": 569311232
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005633464157367127,
      "loss": 3.0778,
      "theoretical_loss": 3.8610527476350116,
      "tokens_seen": 569376768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005633298657083412,
      "loss": 2.9027,
      "theoretical_loss": 3.8610059505282757,
      "tokens_seen": 569442304
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005633133156799698,
      "loss": 2.7764,
      "theoretical_loss": 3.86095916031484,
      "tokens_seen": 569507840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005632967656515982,
      "loss": 2.6483,
      "theoretical_loss": 3.860912376992896,
      "tokens_seen": 569573376
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005632802156232268,
      "loss": 2.9058,
      "theoretical_loss": 3.8608656005606354,
      "tokens_seen": 569638912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005632636655948553,
      "loss": 2.9731,
      "theoretical_loss": 3.8608188310162523,
      "tokens_seen": 569704448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005632471155664838,
      "loss": 2.7353,
      "theoretical_loss": 3.8607720683579396,
      "tokens_seen": 569769984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005632305655381124,
      "loss": 2.7164,
      "theoretical_loss": 3.8607253125838916,
      "tokens_seen": 569835520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005632140155097409,
      "loss": 2.6466,
      "theoretical_loss": 3.8606785636923027,
      "tokens_seen": 569901056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005631974654813694,
      "loss": 3.0509,
      "theoretical_loss": 3.8606318216813693,
      "tokens_seen": 569966592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000563180915452998,
      "loss": 2.8322,
      "theoretical_loss": 3.8605850865492872,
      "tokens_seen": 570032128
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005631643654246264,
      "loss": 2.8821,
      "theoretical_loss": 3.860538358294254,
      "tokens_seen": 570097664
    },
    {
      "epoch": 0.2,
      "objective/train/avg_token_score": 0.005783182568848133,
      "objective/train/avg_weight": 0.9931952953338623,
      "objective/train/docs_used": 331022,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.045592784881592,
      "objective/train/std_weight": 0.027224557474255562,
      "objective/train/theoretical_loss": 3.8604916369144666,
      "objective/train/tokens_used": 590623200,
      "theoretical_loss": 3.8604916369144666,
      "tokens_seen": 570163200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000563147815396255,
      "loss": 3.0832,
      "theoretical_loss": 3.8604916369144666,
      "tokens_seen": 570163200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005631312653678835,
      "loss": 2.9536,
      "theoretical_loss": 3.8604449224081234,
      "tokens_seen": 570228736
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000563114715339512,
      "loss": 2.9878,
      "theoretical_loss": 3.860398214773425,
      "tokens_seen": 570294272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005630981653111406,
      "loss": 2.9834,
      "theoretical_loss": 3.8603515140085687,
      "tokens_seen": 570359808
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005630816152827691,
      "loss": 2.6195,
      "theoretical_loss": 3.8603048201117565,
      "tokens_seen": 570425344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005630650652543976,
      "loss": 2.8512,
      "theoretical_loss": 3.8602581330811887,
      "tokens_seen": 570490880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005630485152260261,
      "loss": 2.7947,
      "theoretical_loss": 3.8602114529150673,
      "tokens_seen": 570556416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005630319651976546,
      "loss": 2.9007,
      "theoretical_loss": 3.860164779611595,
      "tokens_seen": 570621952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005630154151692832,
      "loss": 2.9207,
      "theoretical_loss": 3.8601181131689737,
      "tokens_seen": 570687488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005629988651409117,
      "loss": 2.8368,
      "theoretical_loss": 3.8600714535854093,
      "tokens_seen": 570753024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005629823151125402,
      "loss": 2.8679,
      "theoretical_loss": 3.8600248008591045,
      "tokens_seen": 570818560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005629657650841688,
      "loss": 2.9249,
      "theoretical_loss": 3.859978154988265,
      "tokens_seen": 570884096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005629492150557973,
      "loss": 2.6507,
      "theoretical_loss": 3.8599315159710965,
      "tokens_seen": 570949632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005629326650274257,
      "loss": 2.6457,
      "theoretical_loss": 3.8598848838058055,
      "tokens_seen": 571015168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005629161149990543,
      "loss": 2.9222,
      "theoretical_loss": 3.8598382584905995,
      "tokens_seen": 571080704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005628995649706828,
      "loss": 2.9691,
      "theoretical_loss": 3.859791640023685,
      "tokens_seen": 571146240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005628830149423114,
      "loss": 2.9979,
      "theoretical_loss": 3.859745028403273,
      "tokens_seen": 571211776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005628664649139399,
      "loss": 2.8467,
      "theoretical_loss": 3.85969842362757,
      "tokens_seen": 571277312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005628499148855684,
      "loss": 2.8281,
      "theoretical_loss": 3.859651825694787,
      "tokens_seen": 571342848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005628333648571969,
      "loss": 2.9042,
      "theoretical_loss": 3.859605234603135,
      "tokens_seen": 571408384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005628168148288253,
      "loss": 2.8546,
      "theoretical_loss": 3.8595586503508246,
      "tokens_seen": 571473920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005628002648004539,
      "loss": 2.9345,
      "theoretical_loss": 3.8595120729360675,
      "tokens_seen": 571539456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005627837147720824,
      "loss": 2.6581,
      "theoretical_loss": 3.8594655023570765,
      "tokens_seen": 571604992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005627671647437109,
      "loss": 2.802,
      "theoretical_loss": 3.859418938612065,
      "tokens_seen": 571670528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005627506147153395,
      "loss": 3.0485,
      "theoretical_loss": 3.859372381699247,
      "tokens_seen": 571736064
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000562734064686968,
      "loss": 2.8231,
      "theoretical_loss": 3.859325831616836,
      "tokens_seen": 571801600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005627175146585966,
      "loss": 2.6556,
      "theoretical_loss": 3.859279288363048,
      "tokens_seen": 571867136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000562700964630225,
      "loss": 2.7985,
      "theoretical_loss": 3.8592327519360996,
      "tokens_seen": 571932672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005626844146018535,
      "loss": 2.9292,
      "theoretical_loss": 3.8591862223342064,
      "tokens_seen": 571998208
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005626678645734821,
      "loss": 2.8733,
      "theoretical_loss": 3.8591396995555853,
      "tokens_seen": 572063744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005626513145451106,
      "loss": 2.5699,
      "theoretical_loss": 3.859093183598455,
      "tokens_seen": 572129280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005626347645167391,
      "loss": 2.9466,
      "theoretical_loss": 3.859046674461034,
      "tokens_seen": 572194816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005626182144883677,
      "loss": 3.002,
      "theoretical_loss": 3.859000172141541,
      "tokens_seen": 572260352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005626016644599962,
      "loss": 3.0391,
      "theoretical_loss": 3.8589536766381967,
      "tokens_seen": 572325888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005625851144316248,
      "loss": 2.8962,
      "theoretical_loss": 3.858907187949221,
      "tokens_seen": 572391424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005625685644032532,
      "loss": 2.8907,
      "theoretical_loss": 3.8588607060728357,
      "tokens_seen": 572456960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005625520143748817,
      "loss": 2.7931,
      "theoretical_loss": 3.858814231007262,
      "tokens_seen": 572522496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005625354643465103,
      "loss": 3.0262,
      "theoretical_loss": 3.8587677627507233,
      "tokens_seen": 572588032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005625189143181388,
      "loss": 3.0972,
      "theoretical_loss": 3.858721301301442,
      "tokens_seen": 572653568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005625023642897673,
      "loss": 2.7449,
      "theoretical_loss": 3.8586748466576424,
      "tokens_seen": 572719104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005624858142613959,
      "loss": 2.5847,
      "theoretical_loss": 3.8586283988175496,
      "tokens_seen": 572784640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005624692642330244,
      "loss": 2.861,
      "theoretical_loss": 3.8585819577793883,
      "tokens_seen": 572850176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000562452714204653,
      "loss": 2.7016,
      "theoretical_loss": 3.8585355235413843,
      "tokens_seen": 572915712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005624361641762814,
      "loss": 2.8602,
      "theoretical_loss": 3.858489096101765,
      "tokens_seen": 572981248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005624196141479099,
      "loss": 2.8401,
      "theoretical_loss": 3.858442675458756,
      "tokens_seen": 573046784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005624030641195385,
      "loss": 2.7307,
      "theoretical_loss": 3.8583962616105874,
      "tokens_seen": 573112320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000562386514091167,
      "loss": 2.7366,
      "theoretical_loss": 3.858349854555486,
      "tokens_seen": 573177856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005623699640627955,
      "loss": 2.4276,
      "theoretical_loss": 3.858303454291682,
      "tokens_seen": 573243392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005623534140344241,
      "loss": 2.9193,
      "theoretical_loss": 3.8582570608174045,
      "tokens_seen": 573308928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005623368640060526,
      "loss": 3.0301,
      "theoretical_loss": 3.858210674130885,
      "tokens_seen": 573374464
    },
    {
      "epoch": 0.2,
      "objective/train/avg_token_score": 0.013314598239958286,
      "objective/train/avg_weight": 0.9843315482139587,
      "objective/train/docs_used": 332875,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1526834964752197,
      "objective/train/std_weight": 0.06310330331325531,
      "objective/train/theoretical_loss": 3.858164294230354,
      "objective/train/tokens_used": 593900000,
      "theoretical_loss": 3.858164294230354,
      "tokens_seen": 573440000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005623203139776811,
      "loss": 2.9944,
      "theoretical_loss": 3.858164294230354,
      "tokens_seen": 573440000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005623037639493096,
      "loss": 2.7972,
      "theoretical_loss": 3.8581179211140437,
      "tokens_seen": 573505536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005622872139209381,
      "loss": 2.8665,
      "theoretical_loss": 3.8580715547801865,
      "tokens_seen": 573571072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005622706638925667,
      "loss": 2.6774,
      "theoretical_loss": 3.8580251952270164,
      "tokens_seen": 573636608
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005622541138641952,
      "loss": 2.4763,
      "theoretical_loss": 3.8579788424527655,
      "tokens_seen": 573702144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005622375638358237,
      "loss": 2.7658,
      "theoretical_loss": 3.8579324964556703,
      "tokens_seen": 573767680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005622210138074523,
      "loss": 2.9785,
      "theoretical_loss": 3.857886157233965,
      "tokens_seen": 573833216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005622044637790807,
      "loss": 2.7166,
      "theoretical_loss": 3.8578398247858856,
      "tokens_seen": 573898752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0005621879137507093,
      "loss": 2.9572,
      "theoretical_loss": 3.857793499109669,
      "tokens_seen": 573964288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005621713637223378,
      "loss": 2.8587,
      "theoretical_loss": 3.8577471802035515,
      "tokens_seen": 574029824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005621548136939663,
      "loss": 2.9465,
      "theoretical_loss": 3.8577008680657716,
      "tokens_seen": 574095360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005621382636655949,
      "loss": 2.9844,
      "theoretical_loss": 3.8576545626945684,
      "tokens_seen": 574160896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005621217136372234,
      "loss": 2.7764,
      "theoretical_loss": 3.8576082640881797,
      "tokens_seen": 574226432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005621051636088519,
      "loss": 2.9028,
      "theoretical_loss": 3.857561972244846,
      "tokens_seen": 574291968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005620886135804805,
      "loss": 2.9249,
      "theoretical_loss": 3.8575156871628082,
      "tokens_seen": 574357504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005620720635521089,
      "loss": 2.6754,
      "theoretical_loss": 3.8574694088403065,
      "tokens_seen": 574423040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005620555135237375,
      "loss": 2.6179,
      "theoretical_loss": 3.8574231372755836,
      "tokens_seen": 574488576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000562038963495366,
      "loss": 2.7653,
      "theoretical_loss": 3.8573768724668813,
      "tokens_seen": 574554112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005620224134669945,
      "loss": 2.8654,
      "theoretical_loss": 3.8573306144124437,
      "tokens_seen": 574619648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005620058634386231,
      "loss": 2.8421,
      "theoretical_loss": 3.857284363110513,
      "tokens_seen": 574685184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005619893134102516,
      "loss": 2.9861,
      "theoretical_loss": 3.8572381185593354,
      "tokens_seen": 574750720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005619727633818802,
      "loss": 2.9431,
      "theoretical_loss": 3.8571918807571546,
      "tokens_seen": 574816256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005619562133535086,
      "loss": 2.7786,
      "theoretical_loss": 3.8571456497022165,
      "tokens_seen": 574881792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005619396633251371,
      "loss": 2.7614,
      "theoretical_loss": 3.857099425392768,
      "tokens_seen": 574947328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005619231132967657,
      "loss": 2.5884,
      "theoretical_loss": 3.8570532078270565,
      "tokens_seen": 575012864
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005619065632683942,
      "loss": 2.5378,
      "theoretical_loss": 3.857006997003329,
      "tokens_seen": 575078400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005618900132400227,
      "loss": 2.7715,
      "theoretical_loss": 3.8569607929198337,
      "tokens_seen": 575143936
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005618734632116513,
      "loss": 2.9482,
      "theoretical_loss": 3.85691459557482,
      "tokens_seen": 575209472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005618569131832797,
      "loss": 2.66,
      "theoretical_loss": 3.856868404966538,
      "tokens_seen": 575275008
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005618403631549082,
      "loss": 2.9099,
      "theoretical_loss": 3.8568222210932372,
      "tokens_seen": 575340544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005618238131265367,
      "loss": 2.8378,
      "theoretical_loss": 3.856776043953169,
      "tokens_seen": 575406080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005618072630981653,
      "loss": 2.9035,
      "theoretical_loss": 3.8567298735445847,
      "tokens_seen": 575471616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005617907130697938,
      "loss": 2.5495,
      "theoretical_loss": 3.856683709865737,
      "tokens_seen": 575537152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005617741630414223,
      "loss": 2.8367,
      "theoretical_loss": 3.8566375529148793,
      "tokens_seen": 575602688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005617576130130509,
      "loss": 2.925,
      "theoretical_loss": 3.856591402690264,
      "tokens_seen": 575668224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005617410629846794,
      "loss": 3.0251,
      "theoretical_loss": 3.8565452591901463,
      "tokens_seen": 575733760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005617245129563078,
      "loss": 2.862,
      "theoretical_loss": 3.856499122412781,
      "tokens_seen": 575799296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005617079629279364,
      "loss": 2.821,
      "theoretical_loss": 3.8564529923564232,
      "tokens_seen": 575864832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005616914128995649,
      "loss": 2.7504,
      "theoretical_loss": 3.856406869019329,
      "tokens_seen": 575930368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005616748628711935,
      "loss": 2.746,
      "theoretical_loss": 3.8563607523997563,
      "tokens_seen": 575995904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000561658312842822,
      "loss": 2.9949,
      "theoretical_loss": 3.856314642495961,
      "tokens_seen": 576061440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005616417628144505,
      "loss": 2.6014,
      "theoretical_loss": 3.8562685393062024,
      "tokens_seen": 576126976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005616252127860791,
      "loss": 2.8872,
      "theoretical_loss": 3.8562224428287393,
      "tokens_seen": 576192512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005616086627577076,
      "loss": 2.8302,
      "theoretical_loss": 3.85617635306183,
      "tokens_seen": 576258048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000561592112729336,
      "loss": 2.8508,
      "theoretical_loss": 3.856130270003736,
      "tokens_seen": 576323584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005615755627009646,
      "loss": 2.9062,
      "theoretical_loss": 3.856084193652718,
      "tokens_seen": 576389120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005615590126725931,
      "loss": 2.8956,
      "theoretical_loss": 3.8560381240070365,
      "tokens_seen": 576454656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005615424626442217,
      "loss": 2.7209,
      "theoretical_loss": 3.8559920610649536,
      "tokens_seen": 576520192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005615259126158502,
      "loss": 2.8581,
      "theoretical_loss": 3.855946004824733,
      "tokens_seen": 576585728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005615093625874787,
      "loss": 3.0412,
      "theoretical_loss": 3.8558999552846367,
      "tokens_seen": 576651264
    },
    {
      "epoch": 0.21,
      "objective/train/avg_token_score": 0.011753704398870468,
      "objective/train/avg_weight": 0.9862054586410522,
      "objective/train/docs_used": 334850,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0417556762695312,
      "objective/train/std_weight": 0.0582837276160717,
      "objective/train/theoretical_loss": 3.85585391244293,
      "objective/train/tokens_used": 597176800,
      "theoretical_loss": 3.85585391244293,
      "tokens_seen": 576716800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005614928125591073,
      "loss": 2.8783,
      "theoretical_loss": 3.85585391244293,
      "tokens_seen": 576716800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005614762625307357,
      "loss": 2.9808,
      "theoretical_loss": 3.855807876297876,
      "tokens_seen": 576782336
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005614597125023642,
      "loss": 2.7706,
      "theoretical_loss": 3.8557618468477415,
      "tokens_seen": 576847872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005614431624739928,
      "loss": 3.0343,
      "theoretical_loss": 3.8557158240907916,
      "tokens_seen": 576913408
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005614266124456213,
      "loss": 2.765,
      "theoretical_loss": 3.8556698080252936,
      "tokens_seen": 576978944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005614100624172499,
      "loss": 2.809,
      "theoretical_loss": 3.8556237986495137,
      "tokens_seen": 577044480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005613935123888784,
      "loss": 3.0393,
      "theoretical_loss": 3.8555777959617203,
      "tokens_seen": 577110016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005613769623605069,
      "loss": 2.7617,
      "theoretical_loss": 3.8555317999601817,
      "tokens_seen": 577175552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005613604123321355,
      "loss": 2.4767,
      "theoretical_loss": 3.8554858106431675,
      "tokens_seen": 577241088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005613438623037639,
      "loss": 2.6372,
      "theoretical_loss": 3.8554398280089472,
      "tokens_seen": 577306624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005613273122753924,
      "loss": 2.6682,
      "theoretical_loss": 3.8553938520557915,
      "tokens_seen": 577372160
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000561310762247021,
      "loss": 2.7496,
      "theoretical_loss": 3.8553478827819707,
      "tokens_seen": 577437696
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005612942122186495,
      "loss": 2.8942,
      "theoretical_loss": 3.8553019201857577,
      "tokens_seen": 577503232
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005612776621902781,
      "loss": 2.845,
      "theoretical_loss": 3.8552559642654236,
      "tokens_seen": 577568768
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005612611121619066,
      "loss": 2.6617,
      "theoretical_loss": 3.8552100150192423,
      "tokens_seen": 577634304
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005612445621335351,
      "loss": 2.8075,
      "theoretical_loss": 3.855164072445487,
      "tokens_seen": 577699840
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005612280121051636,
      "loss": 2.9233,
      "theoretical_loss": 3.855118136542433,
      "tokens_seen": 577765376
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005612114620767921,
      "loss": 2.8859,
      "theoretical_loss": 3.8550722073083543,
      "tokens_seen": 577830912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005611949120484206,
      "loss": 2.8105,
      "theoretical_loss": 3.855026284741526,
      "tokens_seen": 577896448
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005611783620200492,
      "loss": 2.9436,
      "theoretical_loss": 3.8549803688402253,
      "tokens_seen": 577961984
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005611618119916777,
      "loss": 2.7227,
      "theoretical_loss": 3.8549344596027293,
      "tokens_seen": 578027520
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005611452619633063,
      "loss": 3.0251,
      "theoretical_loss": 3.854888557027315,
      "tokens_seen": 578093056
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005611287119349348,
      "loss": 2.7977,
      "theoretical_loss": 3.8548426611122597,
      "tokens_seen": 578158592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005611121619065632,
      "loss": 2.8378,
      "theoretical_loss": 3.8547967718558436,
      "tokens_seen": 578224128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005610956118781918,
      "loss": 2.8864,
      "theoretical_loss": 3.854750889256345,
      "tokens_seen": 578289664
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005610790618498203,
      "loss": 2.7951,
      "theoretical_loss": 3.854705013312045,
      "tokens_seen": 578355200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005610625118214488,
      "loss": 2.8804,
      "theoretical_loss": 3.8546591440212237,
      "tokens_seen": 578420736
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005610459617930774,
      "loss": 2.6346,
      "theoretical_loss": 3.8546132813821625,
      "tokens_seen": 578486272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005610294117647059,
      "loss": 2.7951,
      "theoretical_loss": 3.8545674253931432,
      "tokens_seen": 578551808
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005610128617363345,
      "loss": 2.9314,
      "theoretical_loss": 3.8545215760524494,
      "tokens_seen": 578617344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000560996311707963,
      "loss": 2.8989,
      "theoretical_loss": 3.8544757333583632,
      "tokens_seen": 578682880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005609797616795914,
      "loss": 2.9576,
      "theoretical_loss": 3.854429897309169,
      "tokens_seen": 578748416
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00056096321165122,
      "loss": 2.6653,
      "theoretical_loss": 3.854384067903151,
      "tokens_seen": 578813952
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005609466616228485,
      "loss": 2.829,
      "theoretical_loss": 3.854338245138595,
      "tokens_seen": 578879488
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000560930111594477,
      "loss": 2.6672,
      "theoretical_loss": 3.8542924290137863,
      "tokens_seen": 578945024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005609135615661056,
      "loss": 2.6981,
      "theoretical_loss": 3.8542466195270118,
      "tokens_seen": 579010560
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005608970115377341,
      "loss": 2.8333,
      "theoretical_loss": 3.8542008166765576,
      "tokens_seen": 579076096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005608804615093626,
      "loss": 2.8917,
      "theoretical_loss": 3.8541550204607127,
      "tokens_seen": 579141632
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000560863911480991,
      "loss": 2.5131,
      "theoretical_loss": 3.8541092308777647,
      "tokens_seen": 579207168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005608473614526196,
      "loss": 2.8219,
      "theoretical_loss": 3.854063447926003,
      "tokens_seen": 579272704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005608308114242481,
      "loss": 2.7384,
      "theoretical_loss": 3.8540176716037164,
      "tokens_seen": 579338240
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005608142613958766,
      "loss": 3.0055,
      "theoretical_loss": 3.853971901909196,
      "tokens_seen": 579403776
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005607977113675052,
      "loss": 2.8932,
      "theoretical_loss": 3.8539261388407327,
      "tokens_seen": 579469312
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005607811613391337,
      "loss": 2.847,
      "theoretical_loss": 3.8538803823966172,
      "tokens_seen": 579534848
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005607646113107623,
      "loss": 2.8381,
      "theoretical_loss": 3.853834632575143,
      "tokens_seen": 579600384
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005607480612823907,
      "loss": 3.1134,
      "theoretical_loss": 3.8537888893746013,
      "tokens_seen": 579665920
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005607315112540192,
      "loss": 3.1545,
      "theoretical_loss": 3.853743152793287,
      "tokens_seen": 579731456
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005607149612256478,
      "loss": 2.7234,
      "theoretical_loss": 3.8536974228294927,
      "tokens_seen": 579796992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005606984111972763,
      "loss": 2.8968,
      "theoretical_loss": 3.8536516994815146,
      "tokens_seen": 579862528
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005606818611689048,
      "loss": 3.0153,
      "theoretical_loss": 3.8536059827476477,
      "tokens_seen": 579928064
    },
    {
      "epoch": 0.21,
      "objective/train/avg_token_score": 0.020048731938004494,
      "objective/train/avg_weight": 0.9764091372489929,
      "objective/train/docs_used": 336483,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2673957347869873,
      "objective/train/std_weight": 0.09058598428964615,
      "objective/train/theoretical_loss": 3.8535602726261864,
      "objective/train/tokens_used": 600453600,
      "theoretical_loss": 3.8535602726261864,
      "tokens_seen": 579993600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005606653111405334,
      "loss": 3.0718,
      "theoretical_loss": 3.8535602726261864,
      "tokens_seen": 579993600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005606487611121619,
      "loss": 2.791,
      "theoretical_loss": 3.853514569115429,
      "tokens_seen": 580059136
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005606322110837905,
      "loss": 2.91,
      "theoretical_loss": 3.8534688722136723,
      "tokens_seen": 580124672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005606156610554189,
      "loss": 2.6435,
      "theoretical_loss": 3.853423181919214,
      "tokens_seen": 580190208
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005605991110270474,
      "loss": 2.8375,
      "theoretical_loss": 3.8533774982303526,
      "tokens_seen": 580255744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000560582560998676,
      "loss": 2.7499,
      "theoretical_loss": 3.853331821145387,
      "tokens_seen": 580321280
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005605660109703045,
      "loss": 2.864,
      "theoretical_loss": 3.8532861506626173,
      "tokens_seen": 580386816
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000560549460941933,
      "loss": 2.996,
      "theoretical_loss": 3.853240486780344,
      "tokens_seen": 580452352
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005605329109135616,
      "loss": 2.7071,
      "theoretical_loss": 3.8531948294968674,
      "tokens_seen": 580517888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005605163608851901,
      "loss": 2.6064,
      "theoretical_loss": 3.85314917881049,
      "tokens_seen": 580583424
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005604998108568186,
      "loss": 2.8208,
      "theoretical_loss": 3.853103534719513,
      "tokens_seen": 580648960
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005604832608284471,
      "loss": 2.8508,
      "theoretical_loss": 3.8530578972222402,
      "tokens_seen": 580714496
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005604667108000756,
      "loss": 2.7942,
      "theoretical_loss": 3.8530122663169752,
      "tokens_seen": 580780032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005604501607717042,
      "loss": 2.9123,
      "theoretical_loss": 3.8529666420020217,
      "tokens_seen": 580845568
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005604336107433327,
      "loss": 2.9904,
      "theoretical_loss": 3.852921024275684,
      "tokens_seen": 580911104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005604170607149612,
      "loss": 2.7639,
      "theoretical_loss": 3.8528754131362684,
      "tokens_seen": 580976640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005604005106865898,
      "loss": 2.7601,
      "theoretical_loss": 3.8528298085820802,
      "tokens_seen": 581042176
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005603839606582182,
      "loss": 2.694,
      "theoretical_loss": 3.8527842106114267,
      "tokens_seen": 581107712
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005603674106298468,
      "loss": 2.7677,
      "theoretical_loss": 3.852738619222615,
      "tokens_seen": 581173248
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005603508606014753,
      "loss": 2.8845,
      "theoretical_loss": 3.8526930344139525,
      "tokens_seen": 581238784
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005603343105731038,
      "loss": 2.8602,
      "theoretical_loss": 3.8526474561837483,
      "tokens_seen": 581304320
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005603177605447324,
      "loss": 2.767,
      "theoretical_loss": 3.852601884530311,
      "tokens_seen": 581369856
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005603012105163609,
      "loss": 3.0202,
      "theoretical_loss": 3.852556319451952,
      "tokens_seen": 581435392
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005602846604879894,
      "loss": 2.8247,
      "theoretical_loss": 3.8525107609469793,
      "tokens_seen": 581500928
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000560268110459618,
      "loss": 2.6399,
      "theoretical_loss": 3.852465209013705,
      "tokens_seen": 581566464
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005602515604312464,
      "loss": 2.6678,
      "theoretical_loss": 3.8524196636504415,
      "tokens_seen": 581632000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000560235010402875,
      "loss": 2.9351,
      "theoretical_loss": 3.8523741248555,
      "tokens_seen": 581697536
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005602184603745035,
      "loss": 2.7831,
      "theoretical_loss": 3.8523285926271935,
      "tokens_seen": 581763072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000560201910346132,
      "loss": 2.8615,
      "theoretical_loss": 3.852283066963836,
      "tokens_seen": 581828608
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005601853603177606,
      "loss": 2.6785,
      "theoretical_loss": 3.8522375478637425,
      "tokens_seen": 581894144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005601688102893891,
      "loss": 2.9864,
      "theoretical_loss": 3.852192035325226,
      "tokens_seen": 581959680
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005601522602610176,
      "loss": 2.907,
      "theoretical_loss": 3.8521465293466024,
      "tokens_seen": 582025216
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005601357102326461,
      "loss": 2.5756,
      "theoretical_loss": 3.852101029926188,
      "tokens_seen": 582090752
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005601191602042746,
      "loss": 2.8509,
      "theoretical_loss": 3.8520555370622995,
      "tokens_seen": 582156288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005601026101759032,
      "loss": 2.7352,
      "theoretical_loss": 3.852010050753254,
      "tokens_seen": 582221824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005600860601475317,
      "loss": 2.9369,
      "theoretical_loss": 3.85196457099737,
      "tokens_seen": 582287360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005600695101191602,
      "loss": 2.7501,
      "theoretical_loss": 3.8519190977929645,
      "tokens_seen": 582352896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005600529600907888,
      "loss": 2.8163,
      "theoretical_loss": 3.851873631138358,
      "tokens_seen": 582418432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005600364100624173,
      "loss": 2.9334,
      "theoretical_loss": 3.85182817103187,
      "tokens_seen": 582483968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005600198600340457,
      "loss": 2.6834,
      "theoretical_loss": 3.8517827174718198,
      "tokens_seen": 582549504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005600033100056743,
      "loss": 2.8445,
      "theoretical_loss": 3.85173727045653,
      "tokens_seen": 582615040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005599867599773028,
      "loss": 2.6751,
      "theoretical_loss": 3.851691829984321,
      "tokens_seen": 582680576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005599702099489314,
      "loss": 2.7778,
      "theoretical_loss": 3.851646396053516,
      "tokens_seen": 582746112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005599536599205599,
      "loss": 2.8335,
      "theoretical_loss": 3.851600968662437,
      "tokens_seen": 582811648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005599371098921884,
      "loss": 2.6557,
      "theoretical_loss": 3.851555547809408,
      "tokens_seen": 582877184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000559920559863817,
      "loss": 2.7112,
      "theoretical_loss": 3.851510133492752,
      "tokens_seen": 582942720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005599040098354455,
      "loss": 2.8693,
      "theoretical_loss": 3.8514647257107955,
      "tokens_seen": 583008256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005598874598070739,
      "loss": 3.0002,
      "theoretical_loss": 3.8514193244618617,
      "tokens_seen": 583073792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005598709097787024,
      "loss": 2.9705,
      "theoretical_loss": 3.8513739297442786,
      "tokens_seen": 583139328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005598543597503309,
      "loss": 3.1362,
      "theoretical_loss": 3.851328541556372,
      "tokens_seen": 583204864
    },
    {
      "epoch": 0.21,
      "objective/train/avg_token_score": 0.013881132006645203,
      "objective/train/avg_weight": 0.9836641550064087,
      "objective/train/docs_used": 338617,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.729675769805908,
      "objective/train/std_weight": 0.06745589524507523,
      "objective/train/theoretical_loss": 3.851283159896468,
      "objective/train/tokens_used": 603730400,
      "theoretical_loss": 3.851283159896468,
      "tokens_seen": 583270400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005598378097219595,
      "loss": 2.7581,
      "theoretical_loss": 3.851283159896468,
      "tokens_seen": 583270400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000559821259693588,
      "loss": 2.8677,
      "theoretical_loss": 3.851237784762896,
      "tokens_seen": 583335936
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005598047096652166,
      "loss": 2.8061,
      "theoretical_loss": 3.8511924161539826,
      "tokens_seen": 583401472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005597881596368451,
      "loss": 2.7386,
      "theoretical_loss": 3.8511470540680586,
      "tokens_seen": 583467008
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005597716096084735,
      "loss": 2.7768,
      "theoretical_loss": 3.8511016985034523,
      "tokens_seen": 583532544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005597550595801021,
      "loss": 2.881,
      "theoretical_loss": 3.851056349458495,
      "tokens_seen": 583598080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005597385095517306,
      "loss": 2.7833,
      "theoretical_loss": 3.8510110069315173,
      "tokens_seen": 583663616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005597219595233591,
      "loss": 2.6168,
      "theoretical_loss": 3.85096567092085,
      "tokens_seen": 583729152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005597054094949877,
      "loss": 2.9716,
      "theoretical_loss": 3.8509203414248256,
      "tokens_seen": 583794688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005596888594666162,
      "loss": 2.8855,
      "theoretical_loss": 3.8508750184417764,
      "tokens_seen": 583860224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005596723094382448,
      "loss": 2.9196,
      "theoretical_loss": 3.850829701970037,
      "tokens_seen": 583925760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005596557594098732,
      "loss": 2.6618,
      "theoretical_loss": 3.85078439200794,
      "tokens_seen": 583991296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005596392093815017,
      "loss": 2.5953,
      "theoretical_loss": 3.85073908855382,
      "tokens_seen": 584056832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005596226593531303,
      "loss": 2.7354,
      "theoretical_loss": 3.8506937916060133,
      "tokens_seen": 584122368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005596061093247588,
      "loss": 2.7897,
      "theoretical_loss": 3.8506485011628544,
      "tokens_seen": 584187904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005595895592963873,
      "loss": 2.6831,
      "theoretical_loss": 3.8506032172226803,
      "tokens_seen": 584253440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005595730092680159,
      "loss": 2.7419,
      "theoretical_loss": 3.850557939783828,
      "tokens_seen": 584318976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005595564592396444,
      "loss": 2.8188,
      "theoretical_loss": 3.850512668844635,
      "tokens_seen": 584384512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000559539909211273,
      "loss": 2.958,
      "theoretical_loss": 3.8504674044034393,
      "tokens_seen": 584450048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005595233591829014,
      "loss": 2.7196,
      "theoretical_loss": 3.8504221464585795,
      "tokens_seen": 584515584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005595068091545299,
      "loss": 2.6964,
      "theoretical_loss": 3.850376895008396,
      "tokens_seen": 584581120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005594902591261585,
      "loss": 2.9342,
      "theoretical_loss": 3.8503316500512286,
      "tokens_seen": 584646656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000559473709097787,
      "loss": 2.928,
      "theoretical_loss": 3.850286411585417,
      "tokens_seen": 584712192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005594571590694156,
      "loss": 2.8877,
      "theoretical_loss": 3.8502411796093035,
      "tokens_seen": 584777728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005594406090410441,
      "loss": 2.7883,
      "theoretical_loss": 3.8501959541212294,
      "tokens_seen": 584843264
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005594240590126726,
      "loss": 3.0148,
      "theoretical_loss": 3.850150735119537,
      "tokens_seen": 584908800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005594075089843012,
      "loss": 2.7753,
      "theoretical_loss": 3.8501055226025707,
      "tokens_seen": 584974336
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005593909589559296,
      "loss": 2.7119,
      "theoretical_loss": 3.8500603165686726,
      "tokens_seen": 585039872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005593744089275581,
      "loss": 2.9821,
      "theoretical_loss": 3.850015117016188,
      "tokens_seen": 585105408
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005593578588991867,
      "loss": 2.7988,
      "theoretical_loss": 3.849969923943461,
      "tokens_seen": 585170944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005593413088708152,
      "loss": 2.9459,
      "theoretical_loss": 3.8499247373488377,
      "tokens_seen": 585236480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005593247588424438,
      "loss": 2.6284,
      "theoretical_loss": 3.8498795572306643,
      "tokens_seen": 585302016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005593082088140723,
      "loss": 2.9226,
      "theoretical_loss": 3.8498343835872872,
      "tokens_seen": 585367552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005592916587857008,
      "loss": 2.8993,
      "theoretical_loss": 3.8497892164170544,
      "tokens_seen": 585433088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005592751087573293,
      "loss": 2.7306,
      "theoretical_loss": 3.8497440557183134,
      "tokens_seen": 585498624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005592585587289578,
      "loss": 2.8944,
      "theoretical_loss": 3.849698901489412,
      "tokens_seen": 585564160
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005592420087005863,
      "loss": 2.8597,
      "theoretical_loss": 3.849653753728701,
      "tokens_seen": 585629696
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005592254586722149,
      "loss": 2.8805,
      "theoretical_loss": 3.849608612434529,
      "tokens_seen": 585695232
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005592089086438434,
      "loss": 3.107,
      "theoretical_loss": 3.8495634776052463,
      "tokens_seen": 585760768
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000559192358615472,
      "loss": 2.5933,
      "theoretical_loss": 3.849518349239205,
      "tokens_seen": 585826304
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005591758085871005,
      "loss": 2.6767,
      "theoretical_loss": 3.849473227334755,
      "tokens_seen": 585891840
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005591592585587289,
      "loss": 2.8031,
      "theoretical_loss": 3.8494281118902505,
      "tokens_seen": 585957376
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005591427085303575,
      "loss": 2.7257,
      "theoretical_loss": 3.849383002904043,
      "tokens_seen": 586022912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000559126158501986,
      "loss": 2.9053,
      "theoretical_loss": 3.8493379003744854,
      "tokens_seen": 586088448
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005591096084736145,
      "loss": 2.7892,
      "theoretical_loss": 3.8492928042999335,
      "tokens_seen": 586153984
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005590930584452431,
      "loss": 2.4786,
      "theoretical_loss": 3.8492477146787403,
      "tokens_seen": 586219520
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005590765084168716,
      "loss": 2.8012,
      "theoretical_loss": 3.849202631509262,
      "tokens_seen": 586285056
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005590599583885002,
      "loss": 2.7392,
      "theoretical_loss": 3.849157554789853,
      "tokens_seen": 586350592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005590434083601287,
      "loss": 2.6023,
      "theoretical_loss": 3.8491124845188716,
      "tokens_seen": 586416128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005590268583317571,
      "loss": 2.966,
      "theoretical_loss": 3.8490674206946736,
      "tokens_seen": 586481664
    },
    {
      "epoch": 0.21,
      "objective/train/avg_token_score": 0.008477779105305672,
      "objective/train/avg_weight": 0.9900242686271667,
      "objective/train/docs_used": 340897,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8460288047790527,
      "objective/train/std_weight": 0.045620422810316086,
      "objective/train/theoretical_loss": 3.8490223633156173,
      "objective/train/tokens_used": 607007200,
      "theoretical_loss": 3.8490223633156173,
      "tokens_seen": 586547200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005590103083033857,
      "loss": 2.7787,
      "theoretical_loss": 3.8490223633156173,
      "tokens_seen": 586547200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005589937582750142,
      "loss": 2.6795,
      "theoretical_loss": 3.84897731238006,
      "tokens_seen": 586612736
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005589772082466427,
      "loss": 2.804,
      "theoretical_loss": 3.8489322678863624,
      "tokens_seen": 586678272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005589606582182713,
      "loss": 2.7278,
      "theoretical_loss": 3.8488872298328816,
      "tokens_seen": 586743808
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005589441081898998,
      "loss": 2.8484,
      "theoretical_loss": 3.848842198217979,
      "tokens_seen": 586809344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005589275581615284,
      "loss": 2.8844,
      "theoretical_loss": 3.8487971730400146,
      "tokens_seen": 586874880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005589110081331567,
      "loss": 2.8266,
      "theoretical_loss": 3.84875215429735,
      "tokens_seen": 586940416
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005588944581047853,
      "loss": 2.6082,
      "theoretical_loss": 3.8487071419883474,
      "tokens_seen": 587005952
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005588779080764138,
      "loss": 2.5478,
      "theoretical_loss": 3.8486621361113684,
      "tokens_seen": 587071488
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005588613580480423,
      "loss": 2.8397,
      "theoretical_loss": 3.848617136664777,
      "tokens_seen": 587137024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005588448080196709,
      "loss": 2.8688,
      "theoretical_loss": 3.848572143646936,
      "tokens_seen": 587202560
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005588282579912994,
      "loss": 2.7463,
      "theoretical_loss": 3.8485271570562096,
      "tokens_seen": 587268096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005588117079629279,
      "loss": 2.8934,
      "theoretical_loss": 3.8484821768909634,
      "tokens_seen": 587333632
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005587951579345564,
      "loss": 2.8032,
      "theoretical_loss": 3.848437203149562,
      "tokens_seen": 587399168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005587786079061849,
      "loss": 2.8511,
      "theoretical_loss": 3.8483922358303717,
      "tokens_seen": 587464704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005587620578778135,
      "loss": 2.9531,
      "theoretical_loss": 3.8483472749317595,
      "tokens_seen": 587530240
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000558745507849442,
      "loss": 2.8354,
      "theoretical_loss": 3.8483023204520928,
      "tokens_seen": 587595776
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005587289578210705,
      "loss": 3.0521,
      "theoretical_loss": 3.848257372389738,
      "tokens_seen": 587661312
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005587124077926991,
      "loss": 2.722,
      "theoretical_loss": 3.848212430743065,
      "tokens_seen": 587726848
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005586958577643276,
      "loss": 2.7925,
      "theoretical_loss": 3.8481674955104426,
      "tokens_seen": 587792384
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000558679307735956,
      "loss": 2.509,
      "theoretical_loss": 3.848122566690239,
      "tokens_seen": 587857920
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005586627577075846,
      "loss": 2.8866,
      "theoretical_loss": 3.8480776442808264,
      "tokens_seen": 587923456
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005586462076792131,
      "loss": 2.779,
      "theoretical_loss": 3.8480327282805744,
      "tokens_seen": 587988992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005586296576508417,
      "loss": 2.6325,
      "theoretical_loss": 3.8479878186878547,
      "tokens_seen": 588054528
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005586131076224702,
      "loss": 2.9559,
      "theoretical_loss": 3.847942915501039,
      "tokens_seen": 588120064
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005585965575940987,
      "loss": 2.8131,
      "theoretical_loss": 3.8478980187185003,
      "tokens_seen": 588185600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005585800075657273,
      "loss": 2.723,
      "theoretical_loss": 3.8478531283386124,
      "tokens_seen": 588251136
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005585634575373558,
      "loss": 2.7709,
      "theoretical_loss": 3.847808244359747,
      "tokens_seen": 588316672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005585469075089842,
      "loss": 2.8111,
      "theoretical_loss": 3.847763366780281,
      "tokens_seen": 588382208
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005585303574806128,
      "loss": 2.9247,
      "theoretical_loss": 3.8477184955985875,
      "tokens_seen": 588447744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005585138074522413,
      "loss": 2.8026,
      "theoretical_loss": 3.8476736308130426,
      "tokens_seen": 588513280
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005584972574238699,
      "loss": 2.6254,
      "theoretical_loss": 3.8476287724220226,
      "tokens_seen": 588578816
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005584807073954984,
      "loss": 2.786,
      "theoretical_loss": 3.8475839204239044,
      "tokens_seen": 588644352
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005584641573671269,
      "loss": 2.7033,
      "theoretical_loss": 3.847539074817065,
      "tokens_seen": 588709888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005584476073387555,
      "loss": 2.7093,
      "theoretical_loss": 3.8474942355998825,
      "tokens_seen": 588775424
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005584310573103839,
      "loss": 2.5549,
      "theoretical_loss": 3.847449402770735,
      "tokens_seen": 588840960
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005584145072820124,
      "loss": 2.7945,
      "theoretical_loss": 3.8474045763280023,
      "tokens_seen": 588906496
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000558397957253641,
      "loss": 2.6375,
      "theoretical_loss": 3.847359756270064,
      "tokens_seen": 588972032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005583814072252695,
      "loss": 2.7379,
      "theoretical_loss": 3.8473149425953,
      "tokens_seen": 589037568
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005583648571968981,
      "loss": 2.8346,
      "theoretical_loss": 3.8472701353020913,
      "tokens_seen": 589103104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005583483071685266,
      "loss": 2.8761,
      "theoretical_loss": 3.8472253343888196,
      "tokens_seen": 589168640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005583317571401551,
      "loss": 3.0091,
      "theoretical_loss": 3.8471805398538663,
      "tokens_seen": 589234176
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005583152071117837,
      "loss": 2.8886,
      "theoretical_loss": 3.8471357516956153,
      "tokens_seen": 589299712
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005582986570834121,
      "loss": 2.7672,
      "theoretical_loss": 3.8470909699124487,
      "tokens_seen": 589365248
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005582821070550406,
      "loss": 2.8616,
      "theoretical_loss": 3.8470461945027505,
      "tokens_seen": 589430784
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005582655570266692,
      "loss": 2.8447,
      "theoretical_loss": 3.8470014254649056,
      "tokens_seen": 589496320
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005582490069982977,
      "loss": 2.5128,
      "theoretical_loss": 3.8469566627972993,
      "tokens_seen": 589561856
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005582324569699263,
      "loss": 2.5206,
      "theoretical_loss": 3.846911906498316,
      "tokens_seen": 589627392
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005582159069415548,
      "loss": 2.8025,
      "theoretical_loss": 3.8468671565663426,
      "tokens_seen": 589692928
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005581993569131833,
      "loss": 2.9218,
      "theoretical_loss": 3.8468224129997663,
      "tokens_seen": 589758464
    },
    {
      "epoch": 0.21,
      "objective/train/avg_token_score": 0.017146486788988113,
      "objective/train/avg_weight": 0.9798255562782288,
      "objective/train/docs_used": 342928,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.770986557006836,
      "objective/train/std_weight": 0.08421693742275238,
      "objective/train/theoretical_loss": 3.846777675796974,
      "objective/train/tokens_used": 610284000,
      "theoretical_loss": 3.846777675796974,
      "tokens_seen": 589824000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005581828068848118,
      "loss": 2.7289,
      "theoretical_loss": 3.846777675796974,
      "tokens_seen": 589824000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005581662568564403,
      "loss": 2.9284,
      "theoretical_loss": 3.8467329449563534,
      "tokens_seen": 589889536
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005581497068280688,
      "loss": 2.8152,
      "theoretical_loss": 3.8466882204762936,
      "tokens_seen": 589955072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005581331567996974,
      "loss": 2.851,
      "theoretical_loss": 3.846643502355183,
      "tokens_seen": 590020608
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005581166067713259,
      "loss": 3.12,
      "theoretical_loss": 3.846598790591412,
      "tokens_seen": 590086144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005581000567429545,
      "loss": 2.8669,
      "theoretical_loss": 3.846554085183371,
      "tokens_seen": 590151680
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000558083506714583,
      "loss": 2.8315,
      "theoretical_loss": 3.8465093861294504,
      "tokens_seen": 590217216
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005580669566862114,
      "loss": 2.7166,
      "theoretical_loss": 3.8464646934280413,
      "tokens_seen": 590282752
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00055805040665784,
      "loss": 3.0601,
      "theoretical_loss": 3.846420007077537,
      "tokens_seen": 590348288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005580338566294685,
      "loss": 2.732,
      "theoretical_loss": 3.8463753270763297,
      "tokens_seen": 590413824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000558017306601097,
      "loss": 2.7089,
      "theoretical_loss": 3.846330653422812,
      "tokens_seen": 590479360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005580007565727256,
      "loss": 2.9005,
      "theoretical_loss": 3.8462859861153778,
      "tokens_seen": 590544896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005579842065443541,
      "loss": 2.7558,
      "theoretical_loss": 3.8462413251524215,
      "tokens_seen": 590610432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005579676565159827,
      "loss": 2.9166,
      "theoretical_loss": 3.846196670532339,
      "tokens_seen": 590675968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005579511064876112,
      "loss": 2.9656,
      "theoretical_loss": 3.8461520222535253,
      "tokens_seen": 590741504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005579345564592396,
      "loss": 2.8561,
      "theoretical_loss": 3.8461073803143764,
      "tokens_seen": 590807040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005579180064308681,
      "loss": 2.6286,
      "theoretical_loss": 3.846062744713289,
      "tokens_seen": 590872576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005579014564024966,
      "loss": 2.6892,
      "theoretical_loss": 3.8460181154486612,
      "tokens_seen": 590938112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005578849063741252,
      "loss": 2.7054,
      "theoretical_loss": 3.8459734925188895,
      "tokens_seen": 591003648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005578683563457537,
      "loss": 2.7569,
      "theoretical_loss": 3.8459288759223735,
      "tokens_seen": 591069184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005578518063173823,
      "loss": 2.7861,
      "theoretical_loss": 3.8458842656575123,
      "tokens_seen": 591134720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005578352562890108,
      "loss": 2.8445,
      "theoretical_loss": 3.845839661722705,
      "tokens_seen": 591200256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005578187062606392,
      "loss": 2.7611,
      "theoretical_loss": 3.8457950641163516,
      "tokens_seen": 591265792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005578021562322678,
      "loss": 2.6926,
      "theoretical_loss": 3.845750472836854,
      "tokens_seen": 591331328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005577856062038963,
      "loss": 2.608,
      "theoretical_loss": 3.845705887882612,
      "tokens_seen": 591396864
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005577690561755248,
      "loss": 2.7104,
      "theoretical_loss": 3.845661309252029,
      "tokens_seen": 591462400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005577525061471534,
      "loss": 3.0526,
      "theoretical_loss": 3.8456167369435073,
      "tokens_seen": 591527936
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005577359561187819,
      "loss": 2.838,
      "theoretical_loss": 3.8455721709554496,
      "tokens_seen": 591593472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005577194060904105,
      "loss": 2.6804,
      "theoretical_loss": 3.84552761128626,
      "tokens_seen": 591659008
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005577028560620389,
      "loss": 2.7029,
      "theoretical_loss": 3.8454830579343424,
      "tokens_seen": 591724544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005576863060336674,
      "loss": 2.5974,
      "theoretical_loss": 3.845438510898102,
      "tokens_seen": 591790080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000557669756005296,
      "loss": 2.747,
      "theoretical_loss": 3.8453939701759436,
      "tokens_seen": 591855616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005576532059769245,
      "loss": 2.8448,
      "theoretical_loss": 3.8453494357662747,
      "tokens_seen": 591921152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000557636655948553,
      "loss": 2.8162,
      "theoretical_loss": 3.8453049076675008,
      "tokens_seen": 591986688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005576201059201816,
      "loss": 2.811,
      "theoretical_loss": 3.8452603858780288,
      "tokens_seen": 592052224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005576035558918101,
      "loss": 2.5474,
      "theoretical_loss": 3.8452158703962676,
      "tokens_seen": 592117760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005575870058634387,
      "loss": 2.8857,
      "theoretical_loss": 3.8451713612206246,
      "tokens_seen": 592183296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005575704558350671,
      "loss": 2.8646,
      "theoretical_loss": 3.8451268583495084,
      "tokens_seen": 592248832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005575539058066956,
      "loss": 2.6795,
      "theoretical_loss": 3.84508236178133,
      "tokens_seen": 592314368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005575373557783242,
      "loss": 2.8815,
      "theoretical_loss": 3.8450378715144984,
      "tokens_seen": 592379904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005575208057499527,
      "loss": 2.8949,
      "theoretical_loss": 3.844993387547425,
      "tokens_seen": 592445440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005575042557215812,
      "loss": 2.685,
      "theoretical_loss": 3.8449489098785197,
      "tokens_seen": 592510976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005574877056932098,
      "loss": 2.5638,
      "theoretical_loss": 3.8449044385061955,
      "tokens_seen": 592576512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005574711556648383,
      "loss": 3.0101,
      "theoretical_loss": 3.8448599734288647,
      "tokens_seen": 592642048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005574546056364668,
      "loss": 2.771,
      "theoretical_loss": 3.8448155146449396,
      "tokens_seen": 592707584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005574380556080953,
      "loss": 2.6127,
      "theoretical_loss": 3.844771062152834,
      "tokens_seen": 592773120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005574215055797238,
      "loss": 2.7132,
      "theoretical_loss": 3.8447266159509628,
      "tokens_seen": 592838656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005574049555513524,
      "loss": 2.6697,
      "theoretical_loss": 3.8446821760377397,
      "tokens_seen": 592904192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005573884055229809,
      "loss": 2.7495,
      "theoretical_loss": 3.84463774241158,
      "tokens_seen": 592969728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005573718554946094,
      "loss": 2.7653,
      "theoretical_loss": 3.8445933150708997,
      "tokens_seen": 593035264
    },
    {
      "epoch": 0.21,
      "objective/train/avg_token_score": 0.006855309940874577,
      "objective/train/avg_weight": 0.9919321537017822,
      "objective/train/docs_used": 344591,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.901031970977783,
      "objective/train/std_weight": 0.043890174478292465,
      "objective/train/theoretical_loss": 3.844548894014116,
      "objective/train/tokens_used": 613560800,
      "theoretical_loss": 3.844548894014116,
      "tokens_seen": 593100800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000557355305466238,
      "loss": 2.8728,
      "theoretical_loss": 3.844548894014116,
      "tokens_seen": 593100800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005573387554378664,
      "loss": 2.9004,
      "theoretical_loss": 3.844504479239645,
      "tokens_seen": 593166336
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000557322205409495,
      "loss": 2.7624,
      "theoretical_loss": 3.844460070745905,
      "tokens_seen": 593231872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005573056553811235,
      "loss": 2.883,
      "theoretical_loss": 3.8444156685313127,
      "tokens_seen": 593297408
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000557289105352752,
      "loss": 2.9267,
      "theoretical_loss": 3.8443712725942882,
      "tokens_seen": 593362944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005572725553243806,
      "loss": 2.5422,
      "theoretical_loss": 3.8443268829332498,
      "tokens_seen": 593428480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005572560052960091,
      "loss": 2.6816,
      "theoretical_loss": 3.844282499546619,
      "tokens_seen": 593494016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005572394552676376,
      "loss": 2.8418,
      "theoretical_loss": 3.844238122432814,
      "tokens_seen": 593559552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005572229052392662,
      "loss": 2.7643,
      "theoretical_loss": 3.8441937515902573,
      "tokens_seen": 593625088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005572063552108946,
      "loss": 2.988,
      "theoretical_loss": 3.84414938701737,
      "tokens_seen": 593690624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005571898051825232,
      "loss": 2.8449,
      "theoretical_loss": 3.8441050287125744,
      "tokens_seen": 593756160
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005571732551541517,
      "loss": 2.8298,
      "theoretical_loss": 3.8440606766742933,
      "tokens_seen": 593821696
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005571567051257802,
      "loss": 2.7041,
      "theoretical_loss": 3.844016330900949,
      "tokens_seen": 593887232
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005571401550974088,
      "loss": 2.7318,
      "theoretical_loss": 3.8439719913909673,
      "tokens_seen": 593952768
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005571236050690373,
      "loss": 2.8606,
      "theoretical_loss": 3.843927658142771,
      "tokens_seen": 594018304
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005571070550406658,
      "loss": 2.6385,
      "theoretical_loss": 3.8438833311547853,
      "tokens_seen": 594083840
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005570905050122943,
      "loss": 2.6719,
      "theoretical_loss": 3.843839010425436,
      "tokens_seen": 594149376
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005570739549839228,
      "loss": 2.8427,
      "theoretical_loss": 3.8437946959531497,
      "tokens_seen": 594214912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005570574049555514,
      "loss": 2.7721,
      "theoretical_loss": 3.8437503877363524,
      "tokens_seen": 594280448
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005570408549271799,
      "loss": 2.822,
      "theoretical_loss": 3.8437060857734715,
      "tokens_seen": 594345984
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005570243048988084,
      "loss": 2.7189,
      "theoretical_loss": 3.843661790062935,
      "tokens_seen": 594411520
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000557007754870437,
      "loss": 2.7274,
      "theoretical_loss": 3.8436175006031714,
      "tokens_seen": 594477056
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005569912048420655,
      "loss": 2.8335,
      "theoretical_loss": 3.84357321739261,
      "tokens_seen": 594542592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000556974654813694,
      "loss": 2.9384,
      "theoretical_loss": 3.843528940429679,
      "tokens_seen": 594608128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005569581047853225,
      "loss": 2.5303,
      "theoretical_loss": 3.8434846697128098,
      "tokens_seen": 594673664
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000556941554756951,
      "loss": 2.8576,
      "theoretical_loss": 3.843440405240433,
      "tokens_seen": 594739200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005569250047285795,
      "loss": 2.7096,
      "theoretical_loss": 3.843396147010979,
      "tokens_seen": 594804736
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000556908454700208,
      "loss": 2.9143,
      "theoretical_loss": 3.843351895022881,
      "tokens_seen": 594870272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005568919046718366,
      "loss": 2.659,
      "theoretical_loss": 3.8433076492745695,
      "tokens_seen": 594935808
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005568753546434651,
      "loss": 2.4306,
      "theoretical_loss": 3.8432634097644787,
      "tokens_seen": 595001344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005568588046150935,
      "loss": 2.8383,
      "theoretical_loss": 3.8432191764910426,
      "tokens_seen": 595066880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005568422545867221,
      "loss": 2.7063,
      "theoretical_loss": 3.8431749494526937,
      "tokens_seen": 595132416
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005568257045583506,
      "loss": 2.6771,
      "theoretical_loss": 3.8431307286478678,
      "tokens_seen": 595197952
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005568091545299792,
      "loss": 2.9104,
      "theoretical_loss": 3.843086514075,
      "tokens_seen": 595263488
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005567926045016077,
      "loss": 2.7634,
      "theoretical_loss": 3.843042305732526,
      "tokens_seen": 595329024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005567760544732362,
      "loss": 2.9047,
      "theoretical_loss": 3.8429981036188816,
      "tokens_seen": 595394560
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005567595044448648,
      "loss": 3.0312,
      "theoretical_loss": 3.8429539077325043,
      "tokens_seen": 595460096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005567429544164933,
      "loss": 2.8526,
      "theoretical_loss": 3.8429097180718315,
      "tokens_seen": 595525632
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005567264043881217,
      "loss": 2.6233,
      "theoretical_loss": 3.842865534635301,
      "tokens_seen": 595591168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005567098543597503,
      "loss": 2.9063,
      "theoretical_loss": 3.842821357421352,
      "tokens_seen": 595656704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005566933043313788,
      "loss": 2.8286,
      "theoretical_loss": 3.842777186428423,
      "tokens_seen": 595722240
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005566767543030074,
      "loss": 2.9183,
      "theoretical_loss": 3.8427330216549533,
      "tokens_seen": 595787776
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005566602042746359,
      "loss": 2.7256,
      "theoretical_loss": 3.8426888630993847,
      "tokens_seen": 595853312
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005566436542462644,
      "loss": 2.8105,
      "theoretical_loss": 3.8426447107601565,
      "tokens_seen": 595918848
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000556627104217893,
      "loss": 2.6329,
      "theoretical_loss": 3.8426005646357106,
      "tokens_seen": 595984384
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005566105541895214,
      "loss": 2.616,
      "theoretical_loss": 3.8425564247244894,
      "tokens_seen": 596049920
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005565940041611499,
      "loss": 2.6733,
      "theoretical_loss": 3.842512291024935,
      "tokens_seen": 596115456
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005565774541327785,
      "loss": 2.8056,
      "theoretical_loss": 3.842468163535491,
      "tokens_seen": 596180992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000556560904104407,
      "loss": 2.9966,
      "theoretical_loss": 3.842424042254601,
      "tokens_seen": 596246528
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005565443540760356,
      "loss": 3.0784,
      "theoretical_loss": 3.8423799271807084,
      "tokens_seen": 596312064
    },
    {
      "epoch": 0.21,
      "objective/train/avg_token_score": 0.01719238981604576,
      "objective/train/avg_weight": 0.9797897934913635,
      "objective/train/docs_used": 346610,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9481544494628906,
      "objective/train/std_weight": 0.07309123128652573,
      "objective/train/theoretical_loss": 3.8423358183122582,
      "objective/train/tokens_used": 616837600,
      "theoretical_loss": 3.8423358183122582,
      "tokens_seen": 596377600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005565278040476641,
      "loss": 2.8963,
      "theoretical_loss": 3.8423358183122582,
      "tokens_seen": 596377600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005565112540192926,
      "loss": 2.8683,
      "theoretical_loss": 3.8422917156476966,
      "tokens_seen": 596443136
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005564947039909212,
      "loss": 3.2326,
      "theoretical_loss": 3.8422476191854686,
      "tokens_seen": 596508672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005564781539625496,
      "loss": 2.5869,
      "theoretical_loss": 3.8422035289240215,
      "tokens_seen": 596574208
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005564616039341781,
      "loss": 2.7058,
      "theoretical_loss": 3.8421594448618013,
      "tokens_seen": 596639744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005564450539058067,
      "loss": 2.6871,
      "theoretical_loss": 3.8421153669972563,
      "tokens_seen": 596705280
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005564285038774352,
      "loss": 2.8235,
      "theoretical_loss": 3.842071295328835,
      "tokens_seen": 596770816
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005564119538490638,
      "loss": 2.7833,
      "theoretical_loss": 3.8420272298549847,
      "tokens_seen": 596836352
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005563954038206923,
      "loss": 2.8531,
      "theoretical_loss": 3.8419831705741556,
      "tokens_seen": 596901888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005563788537923208,
      "loss": 2.6725,
      "theoretical_loss": 3.841939117484798,
      "tokens_seen": 596967424
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005563623037639493,
      "loss": 2.9081,
      "theoretical_loss": 3.841895070585361,
      "tokens_seen": 597032960
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005563457537355778,
      "loss": 2.7585,
      "theoretical_loss": 3.8418510298742965,
      "tokens_seen": 597098496
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005563292037072063,
      "loss": 2.8523,
      "theoretical_loss": 3.8418069953500558,
      "tokens_seen": 597164032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005563126536788349,
      "loss": 2.7182,
      "theoretical_loss": 3.8417629670110904,
      "tokens_seen": 597229568
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005562961036504634,
      "loss": 2.6929,
      "theoretical_loss": 3.841718944855854,
      "tokens_seen": 597295104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000556279553622092,
      "loss": 2.7848,
      "theoretical_loss": 3.8416749288827985,
      "tokens_seen": 597360640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005562630035937205,
      "loss": 2.899,
      "theoretical_loss": 3.8416309190903783,
      "tokens_seen": 597426176
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000556246453565349,
      "loss": 3.0392,
      "theoretical_loss": 3.841586915477048,
      "tokens_seen": 597491712
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005562299035369775,
      "loss": 2.6869,
      "theoretical_loss": 3.8415429180412617,
      "tokens_seen": 597557248
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000556213353508606,
      "loss": 2.7701,
      "theoretical_loss": 3.841498926781475,
      "tokens_seen": 597622784
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005561968034802345,
      "loss": 2.8687,
      "theoretical_loss": 3.8414549416961443,
      "tokens_seen": 597688320
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005561802534518631,
      "loss": 2.8255,
      "theoretical_loss": 3.8414109627837254,
      "tokens_seen": 597753856
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005561637034234916,
      "loss": 2.6235,
      "theoretical_loss": 3.841366990042676,
      "tokens_seen": 597819392
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005561471533951202,
      "loss": 2.7026,
      "theoretical_loss": 3.8413230234714533,
      "tokens_seen": 597884928
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005561306033667487,
      "loss": 2.6106,
      "theoretical_loss": 3.8412790630685154,
      "tokens_seen": 597950464
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005561140533383771,
      "loss": 2.6073,
      "theoretical_loss": 3.841235108832321,
      "tokens_seen": 598016000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005560975033100057,
      "loss": 2.8893,
      "theoretical_loss": 3.84119116076133,
      "tokens_seen": 598081536
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005560809532816342,
      "loss": 2.8993,
      "theoretical_loss": 3.841147218854002,
      "tokens_seen": 598147072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005560644032532627,
      "loss": 2.3401,
      "theoretical_loss": 3.8411032831087963,
      "tokens_seen": 598212608
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005560478532248913,
      "loss": 2.7391,
      "theoretical_loss": 3.841059353524175,
      "tokens_seen": 598278144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005560313031965198,
      "loss": 2.6138,
      "theoretical_loss": 3.841015430098599,
      "tokens_seen": 598343680
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005560147531681484,
      "loss": 2.7376,
      "theoretical_loss": 3.8409715128305306,
      "tokens_seen": 598409216
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005559982031397768,
      "loss": 2.8048,
      "theoretical_loss": 3.8409276017184326,
      "tokens_seen": 598474752
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005559816531114053,
      "loss": 2.4946,
      "theoretical_loss": 3.840883696760768,
      "tokens_seen": 598540288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005559651030830338,
      "loss": 2.7194,
      "theoretical_loss": 3.840839797956,
      "tokens_seen": 598605824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005559485530546623,
      "loss": 2.5659,
      "theoretical_loss": 3.8407959053025937,
      "tokens_seen": 598671360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005559320030262909,
      "loss": 2.6938,
      "theoretical_loss": 3.8407520187990127,
      "tokens_seen": 598736896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005559154529979194,
      "loss": 2.7242,
      "theoretical_loss": 3.8407081384437234,
      "tokens_seen": 598802432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005558989029695479,
      "loss": 2.7826,
      "theoretical_loss": 3.8406642642351914,
      "tokens_seen": 598867968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005558823529411764,
      "loss": 2.831,
      "theoretical_loss": 3.840620396171883,
      "tokens_seen": 598933504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005558658029128049,
      "loss": 2.7439,
      "theoretical_loss": 3.840576534252265,
      "tokens_seen": 598999040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005558492528844335,
      "loss": 2.6896,
      "theoretical_loss": 3.8405326784748057,
      "tokens_seen": 599064576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000555832702856062,
      "loss": 2.8986,
      "theoretical_loss": 3.840488828837973,
      "tokens_seen": 599130112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005558161528276905,
      "loss": 2.7827,
      "theoretical_loss": 3.840444985340235,
      "tokens_seen": 599195648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005557996027993191,
      "loss": 2.6891,
      "theoretical_loss": 3.840401147980061,
      "tokens_seen": 599261184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005557830527709476,
      "loss": 3.0573,
      "theoretical_loss": 3.840357316755921,
      "tokens_seen": 599326720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000555766502742576,
      "loss": 2.7575,
      "theoretical_loss": 3.8403134916662847,
      "tokens_seen": 599392256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005557499527142046,
      "loss": 2.6871,
      "theoretical_loss": 3.840269672709624,
      "tokens_seen": 599457792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005557334026858331,
      "loss": 2.9132,
      "theoretical_loss": 3.8402258598844097,
      "tokens_seen": 599523328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005557168526574617,
      "loss": 2.9562,
      "theoretical_loss": 3.8401820531891135,
      "tokens_seen": 599588864
    },
    {
      "epoch": 0.21,
      "objective/train/avg_token_score": 0.014810123480856419,
      "objective/train/avg_weight": 0.982585608959198,
      "objective/train/docs_used": 348552,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7422080039978027,
      "objective/train/std_weight": 0.07253576070070267,
      "objective/train/theoretical_loss": 3.840138252622208,
      "objective/train/tokens_used": 620114400,
      "theoretical_loss": 3.840138252622208,
      "tokens_seen": 599654400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005557003026290902,
      "loss": 2.5568,
      "theoretical_loss": 3.840138252622208,
      "tokens_seen": 599654400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005556837526007187,
      "loss": 2.799,
      "theoretical_loss": 3.8400944581821674,
      "tokens_seen": 599719936
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005556672025723473,
      "loss": 2.7208,
      "theoretical_loss": 3.8400506698674635,
      "tokens_seen": 599785472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005556506525439758,
      "loss": 2.8638,
      "theoretical_loss": 3.840006887676571,
      "tokens_seen": 599851008
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005556341025156042,
      "loss": 2.8271,
      "theoretical_loss": 3.839963111607965,
      "tokens_seen": 599916544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005556175524872328,
      "loss": 2.927,
      "theoretical_loss": 3.8399193416601207,
      "tokens_seen": 599982080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005556010024588613,
      "loss": 2.7739,
      "theoretical_loss": 3.8398755778315135,
      "tokens_seen": 600047616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005555844524304899,
      "loss": 2.638,
      "theoretical_loss": 3.8398318201206196,
      "tokens_seen": 600113152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005555679024021184,
      "loss": 2.5817,
      "theoretical_loss": 3.8397880685259165,
      "tokens_seen": 600178688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005555513523737469,
      "loss": 2.7935,
      "theoretical_loss": 3.839744323045881,
      "tokens_seen": 600244224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005555348023453755,
      "loss": 2.6573,
      "theoretical_loss": 3.839700583678991,
      "tokens_seen": 600309760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000555518252317004,
      "loss": 2.9314,
      "theoretical_loss": 3.839656850423726,
      "tokens_seen": 600375296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005555017022886324,
      "loss": 2.8445,
      "theoretical_loss": 3.8396131232785637,
      "tokens_seen": 600440832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000555485152260261,
      "loss": 2.9701,
      "theoretical_loss": 3.8395694022419846,
      "tokens_seen": 600506368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005554686022318895,
      "loss": 2.8442,
      "theoretical_loss": 3.8395256873124683,
      "tokens_seen": 600571904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005554520522035181,
      "loss": 2.6267,
      "theoretical_loss": 3.8394819784884957,
      "tokens_seen": 600637440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005554355021751466,
      "loss": 2.6895,
      "theoretical_loss": 3.8394382757685483,
      "tokens_seen": 600702976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005554189521467751,
      "loss": 2.8675,
      "theoretical_loss": 3.839394579151107,
      "tokens_seen": 600768512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005554024021184037,
      "loss": 2.9595,
      "theoretical_loss": 3.8393508886346552,
      "tokens_seen": 600834048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005553858520900321,
      "loss": 2.851,
      "theoretical_loss": 3.8393072042176755,
      "tokens_seen": 600899584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005553693020616607,
      "loss": 2.7664,
      "theoretical_loss": 3.83926352589865,
      "tokens_seen": 600965120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005553527520332892,
      "loss": 2.4812,
      "theoretical_loss": 3.839219853676064,
      "tokens_seen": 601030656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005553362020049177,
      "loss": 2.6792,
      "theoretical_loss": 3.839176187548402,
      "tokens_seen": 601096192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005553196519765463,
      "loss": 2.9374,
      "theoretical_loss": 3.839132527514148,
      "tokens_seen": 601161728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005553031019481748,
      "loss": 2.9537,
      "theoretical_loss": 3.8390888735717885,
      "tokens_seen": 601227264
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005552865519198033,
      "loss": 2.6735,
      "theoretical_loss": 3.8390452257198096,
      "tokens_seen": 601292800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005552700018914319,
      "loss": 2.7596,
      "theoretical_loss": 3.839001583956697,
      "tokens_seen": 601358336
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005552534518630603,
      "loss": 2.8139,
      "theoretical_loss": 3.838957948280939,
      "tokens_seen": 601423872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005552369018346889,
      "loss": 2.8114,
      "theoretical_loss": 3.8389143186910224,
      "tokens_seen": 601489408
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005552203518063174,
      "loss": 2.8812,
      "theoretical_loss": 3.8388706951854354,
      "tokens_seen": 601554944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005552038017779459,
      "loss": 2.7479,
      "theoretical_loss": 3.838827077762668,
      "tokens_seen": 601620480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005551872517495745,
      "loss": 2.7389,
      "theoretical_loss": 3.8387834664212086,
      "tokens_seen": 601686016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000555170701721203,
      "loss": 2.8343,
      "theoretical_loss": 3.8387398611595476,
      "tokens_seen": 601751552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005551541516928315,
      "loss": 2.7765,
      "theoretical_loss": 3.8386962619761746,
      "tokens_seen": 601817088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00055513760166446,
      "loss": 2.6832,
      "theoretical_loss": 3.838652668869581,
      "tokens_seen": 601882624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0005551210516360885,
      "loss": 2.7756,
      "theoretical_loss": 3.8386090818382588,
      "tokens_seen": 601948160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005551045016077171,
      "loss": 2.782,
      "theoretical_loss": 3.838565500880699,
      "tokens_seen": 602013696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005550879515793456,
      "loss": 2.7587,
      "theoretical_loss": 3.838521925995395,
      "tokens_seen": 602079232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005550714015509741,
      "loss": 2.9158,
      "theoretical_loss": 3.83847835718084,
      "tokens_seen": 602144768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005550548515226027,
      "loss": 2.7185,
      "theoretical_loss": 3.838434794435527,
      "tokens_seen": 602210304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005550383014942312,
      "loss": 2.8122,
      "theoretical_loss": 3.8383912377579508,
      "tokens_seen": 602275840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005550217514658596,
      "loss": 2.8382,
      "theoretical_loss": 3.838347687146605,
      "tokens_seen": 602341376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005550052014374882,
      "loss": 2.7868,
      "theoretical_loss": 3.8383041425999864,
      "tokens_seen": 602406912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005549886514091167,
      "loss": 2.9014,
      "theoretical_loss": 3.83826060411659,
      "tokens_seen": 602472448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005549721013807452,
      "loss": 2.4961,
      "theoretical_loss": 3.8382170716949124,
      "tokens_seen": 602537984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005549555513523737,
      "loss": 3.1637,
      "theoretical_loss": 3.83817354533345,
      "tokens_seen": 602603520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005549390013240023,
      "loss": 2.8968,
      "theoretical_loss": 3.8381300250307007,
      "tokens_seen": 602669056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005549224512956308,
      "loss": 2.6398,
      "theoretical_loss": 3.8380865107851623,
      "tokens_seen": 602734592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005549059012672592,
      "loss": 2.6599,
      "theoretical_loss": 3.8380430025953327,
      "tokens_seen": 602800128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005548893512388878,
      "loss": 2.7006,
      "theoretical_loss": 3.837999500459712,
      "tokens_seen": 602865664
    },
    {
      "epoch": 0.22,
      "objective/train/avg_token_score": 0.015218021348118782,
      "objective/train/avg_weight": 0.9820865392684937,
      "objective/train/docs_used": 349996,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0044305324554443,
      "objective/train/std_weight": 0.09596984833478928,
      "objective/train/theoretical_loss": 3.837956004376799,
      "objective/train/tokens_used": 623391200,
      "theoretical_loss": 3.837956004376799,
      "tokens_seen": 602931200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005548728012105163,
      "loss": 2.9553,
      "theoretical_loss": 3.837956004376799,
      "tokens_seen": 602931200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005548562511821448,
      "loss": 2.7759,
      "theoretical_loss": 3.837912514345094,
      "tokens_seen": 602996736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005548397011537734,
      "loss": 2.6235,
      "theoretical_loss": 3.8378690303630973,
      "tokens_seen": 603062272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005548231511254019,
      "loss": 2.6563,
      "theoretical_loss": 3.837825552429311,
      "tokens_seen": 603127808
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005548066010970305,
      "loss": 2.877,
      "theoretical_loss": 3.837782080542236,
      "tokens_seen": 603193344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000554790051068659,
      "loss": 2.8825,
      "theoretical_loss": 3.8377386147003745,
      "tokens_seen": 603258880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005547735010402874,
      "loss": 2.7838,
      "theoretical_loss": 3.837695154902229,
      "tokens_seen": 603324416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000554756951011916,
      "loss": 2.7294,
      "theoretical_loss": 3.837651701146304,
      "tokens_seen": 603389952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005547404009835445,
      "loss": 2.5365,
      "theoretical_loss": 3.8376082534311022,
      "tokens_seen": 603455488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000554723850955173,
      "loss": 2.6076,
      "theoretical_loss": 3.8375648117551284,
      "tokens_seen": 603521024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005547073009268016,
      "loss": 2.9674,
      "theoretical_loss": 3.8375213761168867,
      "tokens_seen": 603586560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005546907508984301,
      "loss": 2.7221,
      "theoretical_loss": 3.837477946514884,
      "tokens_seen": 603652096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005546742008700587,
      "loss": 3.0376,
      "theoretical_loss": 3.8374345229476248,
      "tokens_seen": 603717632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005546576508416871,
      "loss": 2.5987,
      "theoretical_loss": 3.8373911054136167,
      "tokens_seen": 603783168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005546411008133156,
      "loss": 2.6627,
      "theoretical_loss": 3.837347693911366,
      "tokens_seen": 603848704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005546245507849442,
      "loss": 2.8166,
      "theoretical_loss": 3.8373042884393813,
      "tokens_seen": 603914240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005546080007565727,
      "loss": 2.6853,
      "theoretical_loss": 3.837260888996169,
      "tokens_seen": 603979776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005545914507282012,
      "loss": 2.8091,
      "theoretical_loss": 3.837217495580239,
      "tokens_seen": 604045312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005545749006998298,
      "loss": 2.5963,
      "theoretical_loss": 3.8371741081900996,
      "tokens_seen": 604110848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005545583506714583,
      "loss": 2.7952,
      "theoretical_loss": 3.8371307268242614,
      "tokens_seen": 604176384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005545418006430869,
      "loss": 2.853,
      "theoretical_loss": 3.837087351481234,
      "tokens_seen": 604241920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005545252506147153,
      "loss": 2.9243,
      "theoretical_loss": 3.8370439821595284,
      "tokens_seen": 604307456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005545087005863438,
      "loss": 2.9608,
      "theoretical_loss": 3.8370006188576555,
      "tokens_seen": 604372992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005544921505579724,
      "loss": 2.5734,
      "theoretical_loss": 3.8369572615741268,
      "tokens_seen": 604438528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005544756005296009,
      "loss": 2.907,
      "theoretical_loss": 3.836913910307456,
      "tokens_seen": 604504064
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005544590505012294,
      "loss": 2.8282,
      "theoretical_loss": 3.836870565056154,
      "tokens_seen": 604569600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000554442500472858,
      "loss": 2.8601,
      "theoretical_loss": 3.8368272258187366,
      "tokens_seen": 604635136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005544259504444865,
      "loss": 2.7174,
      "theoretical_loss": 3.836783892593715,
      "tokens_seen": 604700672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000554409400416115,
      "loss": 2.7384,
      "theoretical_loss": 3.836740565379606,
      "tokens_seen": 604766208
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005543928503877435,
      "loss": 2.8978,
      "theoretical_loss": 3.836697244174923,
      "tokens_seen": 604831744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000554376300359372,
      "loss": 3.136,
      "theoretical_loss": 3.836653928978182,
      "tokens_seen": 604897280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005543597503310006,
      "loss": 2.6911,
      "theoretical_loss": 3.8366106197878995,
      "tokens_seen": 604962816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005543432003026291,
      "loss": 2.7214,
      "theoretical_loss": 3.8365673166025913,
      "tokens_seen": 605028352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005543266502742576,
      "loss": 2.567,
      "theoretical_loss": 3.8365240194207746,
      "tokens_seen": 605093888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005543101002458862,
      "loss": 2.9411,
      "theoretical_loss": 3.8364807282409674,
      "tokens_seen": 605159424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005542935502175146,
      "loss": 2.6248,
      "theoretical_loss": 3.836437443061688,
      "tokens_seen": 605224960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005542770001891432,
      "loss": 2.7475,
      "theoretical_loss": 3.8363941638814536,
      "tokens_seen": 605290496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005542604501607717,
      "loss": 2.585,
      "theoretical_loss": 3.8363508906987853,
      "tokens_seen": 605356032
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005542439001324002,
      "loss": 2.8468,
      "theoretical_loss": 3.8363076235122016,
      "tokens_seen": 605421568
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005542273501040288,
      "loss": 2.9241,
      "theoretical_loss": 3.8362643623202226,
      "tokens_seen": 605487104
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005542108000756573,
      "loss": 2.8013,
      "theoretical_loss": 3.83622110712137,
      "tokens_seen": 605552640
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005541942500472858,
      "loss": 2.7533,
      "theoretical_loss": 3.8361778579141648,
      "tokens_seen": 605618176
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005541777000189144,
      "loss": 2.7431,
      "theoretical_loss": 3.8361346146971282,
      "tokens_seen": 605683712
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005541611499905428,
      "loss": 2.5588,
      "theoretical_loss": 3.836091377468783,
      "tokens_seen": 605749248
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005541445999621714,
      "loss": 2.8449,
      "theoretical_loss": 3.836048146227652,
      "tokens_seen": 605814784
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005541280499337999,
      "loss": 2.7009,
      "theoretical_loss": 3.8360049209722584,
      "tokens_seen": 605880320
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005541114999054284,
      "loss": 2.6901,
      "theoretical_loss": 3.8359617017011267,
      "tokens_seen": 605945856
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000554094949877057,
      "loss": 2.5222,
      "theoretical_loss": 3.835918488412781,
      "tokens_seen": 606011392
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005540783998486855,
      "loss": 2.6663,
      "theoretical_loss": 3.8358752811057455,
      "tokens_seen": 606076928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000554061849820314,
      "loss": 2.7137,
      "theoretical_loss": 3.8358320797785463,
      "tokens_seen": 606142464
    },
    {
      "epoch": 0.22,
      "objective/train/avg_token_score": 0.021587271243333817,
      "objective/train/avg_weight": 0.9746063947677612,
      "objective/train/docs_used": 351917,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.248901844024658,
      "objective/train/std_weight": 0.12303345650434494,
      "objective/train/theoretical_loss": 3.8357888844297094,
      "objective/train/tokens_used": 626668000,
      "theoretical_loss": 3.8357888844297094,
      "tokens_seen": 606208000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005540452997919425,
      "loss": 2.8586,
      "theoretical_loss": 3.8357888844297094,
      "tokens_seen": 606208000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000554028749763571,
      "loss": 2.8941,
      "theoretical_loss": 3.8357456950577618,
      "tokens_seen": 606273536
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005540121997351996,
      "loss": 2.9295,
      "theoretical_loss": 3.8357025116612293,
      "tokens_seen": 606339072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000553995649706828,
      "loss": 2.5768,
      "theoretical_loss": 3.8356593342386405,
      "tokens_seen": 606404608
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005539790996784566,
      "loss": 2.8179,
      "theoretical_loss": 3.8356161627885235,
      "tokens_seen": 606470144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005539625496500851,
      "loss": 2.8157,
      "theoretical_loss": 3.835572997309406,
      "tokens_seen": 606535680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005539459996217136,
      "loss": 3.0543,
      "theoretical_loss": 3.835529837799818,
      "tokens_seen": 606601216
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005539294495933421,
      "loss": 2.8304,
      "theoretical_loss": 3.8354866842582895,
      "tokens_seen": 606666752
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005539128995649706,
      "loss": 2.7756,
      "theoretical_loss": 3.835443536683349,
      "tokens_seen": 606732288
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005538963495365992,
      "loss": 2.7689,
      "theoretical_loss": 3.835400395073529,
      "tokens_seen": 606797824
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005538797995082277,
      "loss": 2.6975,
      "theoretical_loss": 3.8353572594273597,
      "tokens_seen": 606863360
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005538632494798562,
      "loss": 2.7339,
      "theoretical_loss": 3.835314129743373,
      "tokens_seen": 606928896
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005538466994514848,
      "loss": 2.8162,
      "theoretical_loss": 3.835271006020101,
      "tokens_seen": 606994432
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005538301494231133,
      "loss": 2.9743,
      "theoretical_loss": 3.8352278882560764,
      "tokens_seen": 607059968
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005538135993947417,
      "loss": 2.6925,
      "theoretical_loss": 3.8351847764498332,
      "tokens_seen": 607125504
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005537970493663703,
      "loss": 2.582,
      "theoretical_loss": 3.8351416705999055,
      "tokens_seen": 607191040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005537804993379988,
      "loss": 2.8618,
      "theoretical_loss": 3.835098570704826,
      "tokens_seen": 607256576
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005537639493096274,
      "loss": 2.7568,
      "theoretical_loss": 3.8350554767631304,
      "tokens_seen": 607322112
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005537473992812559,
      "loss": 2.8525,
      "theoretical_loss": 3.8350123887733543,
      "tokens_seen": 607387648
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005537308492528844,
      "loss": 2.793,
      "theoretical_loss": 3.8349693067340334,
      "tokens_seen": 607453184
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000553714299224513,
      "loss": 2.6402,
      "theoretical_loss": 3.8349262306437044,
      "tokens_seen": 607518720
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005536977491961415,
      "loss": 2.8525,
      "theoretical_loss": 3.8348831605009033,
      "tokens_seen": 607584256
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005536811991677699,
      "loss": 3.0432,
      "theoretical_loss": 3.8348400963041684,
      "tokens_seen": 607649792
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005536646491393985,
      "loss": 2.7895,
      "theoretical_loss": 3.8347970380520375,
      "tokens_seen": 607715328
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000553648099111027,
      "loss": 3.022,
      "theoretical_loss": 3.834753985743049,
      "tokens_seen": 607780864
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005536315490826556,
      "loss": 2.9027,
      "theoretical_loss": 3.834710939375742,
      "tokens_seen": 607846400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005536149990542841,
      "loss": 2.8116,
      "theoretical_loss": 3.8346678989486556,
      "tokens_seen": 607911936
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005535984490259126,
      "loss": 2.7626,
      "theoretical_loss": 3.83462486446033,
      "tokens_seen": 607977472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005535818989975412,
      "loss": 2.9724,
      "theoretical_loss": 3.8345818359093062,
      "tokens_seen": 608043008
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005535653489691696,
      "loss": 2.7259,
      "theoretical_loss": 3.8345388132941243,
      "tokens_seen": 608108544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005535487989407981,
      "loss": 2.8489,
      "theoretical_loss": 3.834495796613327,
      "tokens_seen": 608174080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005535322489124267,
      "loss": 2.5288,
      "theoretical_loss": 3.834452785865455,
      "tokens_seen": 608239616
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005535156988840552,
      "loss": 2.5991,
      "theoretical_loss": 3.834409781049053,
      "tokens_seen": 608305152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005534991488556838,
      "loss": 2.7509,
      "theoretical_loss": 3.834366782162662,
      "tokens_seen": 608370688
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005534825988273123,
      "loss": 2.6969,
      "theoretical_loss": 3.8343237892048263,
      "tokens_seen": 608436224
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005534660487989408,
      "loss": 2.9225,
      "theoretical_loss": 3.834280802174091,
      "tokens_seen": 608501760
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005534494987705694,
      "loss": 2.5633,
      "theoretical_loss": 3.8342378210689994,
      "tokens_seen": 608567296
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005534329487421978,
      "loss": 2.8661,
      "theoretical_loss": 3.834194845888097,
      "tokens_seen": 608632832
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005534163987138263,
      "loss": 2.7554,
      "theoretical_loss": 3.83415187662993,
      "tokens_seen": 608698368
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005533998486854549,
      "loss": 2.7814,
      "theoretical_loss": 3.8341089132930444,
      "tokens_seen": 608763904
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005533832986570834,
      "loss": 2.9136,
      "theoretical_loss": 3.834065955875987,
      "tokens_seen": 608829440
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000553366748628712,
      "loss": 2.856,
      "theoretical_loss": 3.8340230043773045,
      "tokens_seen": 608894976
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005533501986003405,
      "loss": 2.5982,
      "theoretical_loss": 3.833980058795545,
      "tokens_seen": 608960512
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000553333648571969,
      "loss": 3.0854,
      "theoretical_loss": 3.8339371191292573,
      "tokens_seen": 609026048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005533170985435975,
      "loss": 2.5764,
      "theoretical_loss": 3.8338941853769892,
      "tokens_seen": 609091584
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000553300548515226,
      "loss": 2.6893,
      "theoretical_loss": 3.8338512575372903,
      "tokens_seen": 609157120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005532839984868545,
      "loss": 2.6281,
      "theoretical_loss": 3.8338083356087105,
      "tokens_seen": 609222656
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005532674484584831,
      "loss": 2.6576,
      "theoretical_loss": 3.8337654195898,
      "tokens_seen": 609288192
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005532508984301116,
      "loss": 2.4961,
      "theoretical_loss": 3.8337225094791103,
      "tokens_seen": 609353728
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005532343484017402,
      "loss": 2.5048,
      "theoretical_loss": 3.8336796052751914,
      "tokens_seen": 609419264
    },
    {
      "epoch": 0.22,
      "objective/train/avg_token_score": 0.022869940847158432,
      "objective/train/avg_weight": 0.9731348752975464,
      "objective/train/docs_used": 353838,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.901431083679199,
      "objective/train/std_weight": 0.13667726516723633,
      "objective/train/theoretical_loss": 3.8336367069765958,
      "objective/train/tokens_used": 629944800,
      "theoretical_loss": 3.8336367069765958,
      "tokens_seen": 609484800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005532177983733687,
      "loss": 2.8563,
      "theoretical_loss": 3.8336367069765958,
      "tokens_seen": 609484800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005532012483449971,
      "loss": 2.688,
      "theoretical_loss": 3.833593814581876,
      "tokens_seen": 609550336
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005531846983166257,
      "loss": 2.6503,
      "theoretical_loss": 3.8335509280895854,
      "tokens_seen": 609615872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005531681482882542,
      "loss": 2.8379,
      "theoretical_loss": 3.833508047498276,
      "tokens_seen": 609681408
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005531515982598827,
      "loss": 2.8059,
      "theoretical_loss": 3.833465172806503,
      "tokens_seen": 609746944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005531350482315113,
      "loss": 2.7772,
      "theoretical_loss": 3.833422304012819,
      "tokens_seen": 609812480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005531184982031398,
      "loss": 2.8524,
      "theoretical_loss": 3.833379441115781,
      "tokens_seen": 609878016
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005531019481747684,
      "loss": 2.6487,
      "theoretical_loss": 3.833336584113943,
      "tokens_seen": 609943552
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005530853981463969,
      "loss": 2.5394,
      "theoretical_loss": 3.8332937330058616,
      "tokens_seen": 610009088
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005530688481180253,
      "loss": 2.7852,
      "theoretical_loss": 3.8332508877900935,
      "tokens_seen": 610074624
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005530522980896539,
      "loss": 2.6383,
      "theoretical_loss": 3.8332080484651945,
      "tokens_seen": 610140160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005530357480612824,
      "loss": 3.009,
      "theoretical_loss": 3.8331652150297226,
      "tokens_seen": 610205696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005530191980329109,
      "loss": 2.5685,
      "theoretical_loss": 3.833122387482236,
      "tokens_seen": 610271232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005530026480045394,
      "loss": 2.8516,
      "theoretical_loss": 3.8330795658212926,
      "tokens_seen": 610336768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000552986097976168,
      "loss": 2.6802,
      "theoretical_loss": 3.833036750045452,
      "tokens_seen": 610402304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005529695479477965,
      "loss": 2.833,
      "theoretical_loss": 3.832993940153273,
      "tokens_seen": 610467840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005529529979194249,
      "loss": 2.8483,
      "theoretical_loss": 3.8329511361433166,
      "tokens_seen": 610533376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005529364478910535,
      "loss": 2.8464,
      "theoretical_loss": 3.832908338014142,
      "tokens_seen": 610598912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000552919897862682,
      "loss": 2.7661,
      "theoretical_loss": 3.832865545764311,
      "tokens_seen": 610664448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005529033478343105,
      "loss": 2.5547,
      "theoretical_loss": 3.8328227593923856,
      "tokens_seen": 610729984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005528867978059391,
      "loss": 2.736,
      "theoretical_loss": 3.832779978896926,
      "tokens_seen": 610795520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005528702477775676,
      "loss": 2.5708,
      "theoretical_loss": 3.832737204276497,
      "tokens_seen": 610861056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005528536977491962,
      "loss": 2.6222,
      "theoretical_loss": 3.8326944355296595,
      "tokens_seen": 610926592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005528371477208246,
      "loss": 2.7767,
      "theoretical_loss": 3.832651672654978,
      "tokens_seen": 610992128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005528205976924531,
      "loss": 2.7954,
      "theoretical_loss": 3.832608915651017,
      "tokens_seen": 611057664
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005528040476640817,
      "loss": 2.7652,
      "theoretical_loss": 3.83256616451634,
      "tokens_seen": 611123200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005527874976357102,
      "loss": 2.5223,
      "theoretical_loss": 3.8325234192495126,
      "tokens_seen": 611188736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005527709476073387,
      "loss": 2.743,
      "theoretical_loss": 3.832480679849101,
      "tokens_seen": 611254272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005527543975789673,
      "loss": 2.8336,
      "theoretical_loss": 3.8324379463136697,
      "tokens_seen": 611319808
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005527378475505958,
      "loss": 2.6189,
      "theoretical_loss": 3.8323952186417864,
      "tokens_seen": 611385344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005527212975222244,
      "loss": 3.0364,
      "theoretical_loss": 3.832352496832018,
      "tokens_seen": 611450880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005527047474938528,
      "loss": 2.7447,
      "theoretical_loss": 3.832309780882932,
      "tokens_seen": 611516416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005526881974654813,
      "loss": 2.8087,
      "theoretical_loss": 3.8322670707930957,
      "tokens_seen": 611581952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005526716474371099,
      "loss": 2.7383,
      "theoretical_loss": 3.832224366561079,
      "tokens_seen": 611647488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005526550974087384,
      "loss": 2.6749,
      "theoretical_loss": 3.83218166818545,
      "tokens_seen": 611713024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005526385473803669,
      "loss": 2.7226,
      "theoretical_loss": 3.8321389756647783,
      "tokens_seen": 611778560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005526219973519955,
      "loss": 2.6428,
      "theoretical_loss": 3.8320962889976347,
      "tokens_seen": 611844096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000552605447323624,
      "loss": 2.7405,
      "theoretical_loss": 3.832053608182589,
      "tokens_seen": 611909632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005525888972952525,
      "loss": 2.6386,
      "theoretical_loss": 3.8320109332182124,
      "tokens_seen": 611975168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000552572347266881,
      "loss": 2.8915,
      "theoretical_loss": 3.831968264103077,
      "tokens_seen": 612040704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005525557972385095,
      "loss": 2.5815,
      "theoretical_loss": 3.831925600835754,
      "tokens_seen": 612106240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005525392472101381,
      "loss": 2.7733,
      "theoretical_loss": 3.831882943414817,
      "tokens_seen": 612171776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005525226971817666,
      "loss": 2.4782,
      "theoretical_loss": 3.8318402918388386,
      "tokens_seen": 612237312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005525061471533951,
      "loss": 2.7952,
      "theoretical_loss": 3.8317976461063923,
      "tokens_seen": 612302848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005524895971250237,
      "loss": 2.9015,
      "theoretical_loss": 3.831755006216052,
      "tokens_seen": 612368384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005524730470966521,
      "loss": 2.7031,
      "theoretical_loss": 3.8317123721663924,
      "tokens_seen": 612433920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005524564970682807,
      "loss": 2.7762,
      "theoretical_loss": 3.831669743955989,
      "tokens_seen": 612499456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005524399470399092,
      "loss": 2.8404,
      "theoretical_loss": 3.8316271215834172,
      "tokens_seen": 612564992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005524233970115377,
      "loss": 2.6886,
      "theoretical_loss": 3.8315845050472532,
      "tokens_seen": 612630528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005524068469831663,
      "loss": 2.7631,
      "theoretical_loss": 3.831541894346073,
      "tokens_seen": 612696064
    },
    {
      "epoch": 0.22,
      "objective/train/avg_token_score": 0.015432137995958328,
      "objective/train/avg_weight": 0.9818339347839355,
      "objective/train/docs_used": 355739,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.938610553741455,
      "objective/train/std_weight": 0.09448862075805664,
      "objective/train/theoretical_loss": 3.8314992894784536,
      "objective/train/tokens_used": 633221600,
      "theoretical_loss": 3.8314992894784536,
      "tokens_seen": 612761600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005523902969547948,
      "loss": 2.5142,
      "theoretical_loss": 3.8314992894784536,
      "tokens_seen": 612761600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005523737469264233,
      "loss": 2.534,
      "theoretical_loss": 3.831456690442974,
      "tokens_seen": 612827136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005523571968980519,
      "loss": 2.7686,
      "theoretical_loss": 3.8314140972382105,
      "tokens_seen": 612892672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005523406468696803,
      "loss": 2.7448,
      "theoretical_loss": 3.831371509862743,
      "tokens_seen": 612958208
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005523240968413089,
      "loss": 2.8054,
      "theoretical_loss": 3.8313289283151497,
      "tokens_seen": 613023744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005523075468129374,
      "loss": 2.869,
      "theoretical_loss": 3.831286352594011,
      "tokens_seen": 613089280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005522909967845659,
      "loss": 2.6621,
      "theoretical_loss": 3.831243782697906,
      "tokens_seen": 613154816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005522744467561945,
      "loss": 2.6784,
      "theoretical_loss": 3.8312012186254156,
      "tokens_seen": 613220352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000552257896727823,
      "loss": 2.5009,
      "theoretical_loss": 3.831158660375122,
      "tokens_seen": 613285888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005522413466994515,
      "loss": 2.7328,
      "theoretical_loss": 3.831116107945605,
      "tokens_seen": 613351424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00055222479667108,
      "loss": 2.796,
      "theoretical_loss": 3.8310735613354483,
      "tokens_seen": 613416960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005522082466427085,
      "loss": 2.765,
      "theoretical_loss": 3.831031020543233,
      "tokens_seen": 613482496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005521916966143371,
      "loss": 2.7283,
      "theoretical_loss": 3.830988485567543,
      "tokens_seen": 613548032
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005521751465859656,
      "loss": 2.6789,
      "theoretical_loss": 3.8309459564069615,
      "tokens_seen": 613613568
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005521585965575941,
      "loss": 2.9603,
      "theoretical_loss": 3.830903433060073,
      "tokens_seen": 613679104
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005521420465292227,
      "loss": 2.8661,
      "theoretical_loss": 3.8308609155254616,
      "tokens_seen": 613744640
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005521254965008512,
      "loss": 2.6761,
      "theoretical_loss": 3.830818403801713,
      "tokens_seen": 613810176
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005521089464724796,
      "loss": 2.8524,
      "theoretical_loss": 3.8307758978874116,
      "tokens_seen": 613875712
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005520923964441082,
      "loss": 2.9292,
      "theoretical_loss": 3.830733397781144,
      "tokens_seen": 613941248
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005520758464157367,
      "loss": 2.6459,
      "theoretical_loss": 3.830690903481497,
      "tokens_seen": 614006784
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005520592963873653,
      "loss": 2.8394,
      "theoretical_loss": 3.830648414987057,
      "tokens_seen": 614072320
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005520427463589938,
      "loss": 2.8105,
      "theoretical_loss": 3.8306059322964128,
      "tokens_seen": 614137856
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005520261963306223,
      "loss": 2.853,
      "theoretical_loss": 3.830563455408151,
      "tokens_seen": 614203392
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005520096463022508,
      "loss": 2.7692,
      "theoretical_loss": 3.8305209843208603,
      "tokens_seen": 614268928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005519930962738792,
      "loss": 3.0735,
      "theoretical_loss": 3.8304785190331305,
      "tokens_seen": 614334464
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005519765462455078,
      "loss": 2.8352,
      "theoretical_loss": 3.830436059543551,
      "tokens_seen": 614400000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005519599962171363,
      "loss": 2.8634,
      "theoretical_loss": 3.830393605850711,
      "tokens_seen": 614465536
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005519434461887648,
      "loss": 2.952,
      "theoretical_loss": 3.8303511579532015,
      "tokens_seen": 614531072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005519268961603934,
      "loss": 2.9798,
      "theoretical_loss": 3.8303087158496134,
      "tokens_seen": 614596608
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005519103461320219,
      "loss": 2.5317,
      "theoretical_loss": 3.8302662795385385,
      "tokens_seen": 614662144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005518937961036505,
      "loss": 2.6744,
      "theoretical_loss": 3.8302238490185676,
      "tokens_seen": 614727680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000551877246075279,
      "loss": 2.8475,
      "theoretical_loss": 3.830181424288295,
      "tokens_seen": 614793216
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005518606960469074,
      "loss": 3.1146,
      "theoretical_loss": 3.830139005346312,
      "tokens_seen": 614858752
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000551844146018536,
      "loss": 2.1989,
      "theoretical_loss": 3.830096592191212,
      "tokens_seen": 614924288
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005518275959901645,
      "loss": 2.8207,
      "theoretical_loss": 3.8300541848215905,
      "tokens_seen": 614989824
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000551811045961793,
      "loss": 2.5993,
      "theoretical_loss": 3.8300117832360403,
      "tokens_seen": 615055360
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005517944959334216,
      "loss": 2.9254,
      "theoretical_loss": 3.8299693874331573,
      "tokens_seen": 615120896
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005517779459050501,
      "loss": 2.7048,
      "theoretical_loss": 3.829926997411537,
      "tokens_seen": 615186432
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005517613958766787,
      "loss": 2.8155,
      "theoretical_loss": 3.8298846131697744,
      "tokens_seen": 615251968
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005517448458483071,
      "loss": 2.6412,
      "theoretical_loss": 3.829842234706466,
      "tokens_seen": 615317504
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005517282958199356,
      "loss": 2.7897,
      "theoretical_loss": 3.8297998620202094,
      "tokens_seen": 615383040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005517117457915642,
      "loss": 2.9796,
      "theoretical_loss": 3.829757495109602,
      "tokens_seen": 615448576
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005516951957631927,
      "loss": 3.0018,
      "theoretical_loss": 3.8297151339732407,
      "tokens_seen": 615514112
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005516786457348212,
      "loss": 2.6186,
      "theoretical_loss": 3.8296727786097247,
      "tokens_seen": 615579648
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005516620957064498,
      "loss": 2.8108,
      "theoretical_loss": 3.8296304290176524,
      "tokens_seen": 615645184
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005516455456780783,
      "loss": 2.7162,
      "theoretical_loss": 3.829588085195623,
      "tokens_seen": 615710720
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005516289956497069,
      "loss": 2.7962,
      "theoretical_loss": 3.8295457471422365,
      "tokens_seen": 615776256
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005516124456213353,
      "loss": 2.689,
      "theoretical_loss": 3.8295034148560934,
      "tokens_seen": 615841792
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005515958955929638,
      "loss": 2.6017,
      "theoretical_loss": 3.8294610883357945,
      "tokens_seen": 615907328
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005515793455645924,
      "loss": 2.9819,
      "theoretical_loss": 3.8294187675799405,
      "tokens_seen": 615972864
    },
    {
      "epoch": 0.22,
      "objective/train/avg_token_score": 0.01557485107332468,
      "objective/train/avg_weight": 0.9816826581954956,
      "objective/train/docs_used": 357525,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.029212236404419,
      "objective/train/std_weight": 0.10043580830097198,
      "objective/train/theoretical_loss": 3.829376452587134,
      "objective/train/tokens_used": 636498400,
      "theoretical_loss": 3.829376452587134,
      "tokens_seen": 616038400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005515627955362209,
      "loss": 2.9861,
      "theoretical_loss": 3.829376452587134,
      "tokens_seen": 616038400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005515462455078494,
      "loss": 2.8538,
      "theoretical_loss": 3.8293341433559767,
      "tokens_seen": 616103936
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000551529695479478,
      "loss": 2.845,
      "theoretical_loss": 3.8292918398850713,
      "tokens_seen": 616169472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005515131454511065,
      "loss": 2.7462,
      "theoretical_loss": 3.829249542173021,
      "tokens_seen": 616235008
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000551496595422735,
      "loss": 2.8999,
      "theoretical_loss": 3.82920725021843,
      "tokens_seen": 616300544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005514800453943635,
      "loss": 2.5318,
      "theoretical_loss": 3.8291649640199017,
      "tokens_seen": 616366080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000551463495365992,
      "loss": 2.9142,
      "theoretical_loss": 3.829122683576042,
      "tokens_seen": 616431616
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005514469453376206,
      "loss": 2.6934,
      "theoretical_loss": 3.8290804088854546,
      "tokens_seen": 616497152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005514303953092491,
      "loss": 2.5577,
      "theoretical_loss": 3.8290381399467464,
      "tokens_seen": 616562688
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005514138452808777,
      "loss": 2.8149,
      "theoretical_loss": 3.8289958767585226,
      "tokens_seen": 616628224
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005513972952525062,
      "loss": 2.7724,
      "theoretical_loss": 3.8289536193193907,
      "tokens_seen": 616693760
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005513807452241346,
      "loss": 2.8107,
      "theoretical_loss": 3.828911367627957,
      "tokens_seen": 616759296
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005513641951957632,
      "loss": 2.9141,
      "theoretical_loss": 3.8288691216828297,
      "tokens_seen": 616824832
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005513476451673917,
      "loss": 2.8229,
      "theoretical_loss": 3.828826881482617,
      "tokens_seen": 616890368
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005513310951390202,
      "loss": 3.1217,
      "theoretical_loss": 3.8287846470259264,
      "tokens_seen": 616955904
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005513145451106488,
      "loss": 2.6304,
      "theoretical_loss": 3.828742418311368,
      "tokens_seen": 617021440
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005512979950822773,
      "loss": 2.8306,
      "theoretical_loss": 3.828700195337551,
      "tokens_seen": 617086976
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005512814450539059,
      "loss": 2.8375,
      "theoretical_loss": 3.8286579781030863,
      "tokens_seen": 617152512
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005512648950255344,
      "loss": 2.7825,
      "theoretical_loss": 3.8286157666065828,
      "tokens_seen": 617218048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005512483449971628,
      "loss": 2.8058,
      "theoretical_loss": 3.828573560846652,
      "tokens_seen": 617283584
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005512317949687914,
      "loss": 2.9036,
      "theoretical_loss": 3.8285313608219056,
      "tokens_seen": 617349120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005512152449404199,
      "loss": 2.7884,
      "theoretical_loss": 3.8284891665309564,
      "tokens_seen": 617414656
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005511986949120484,
      "loss": 2.8103,
      "theoretical_loss": 3.8284469779724155,
      "tokens_seen": 617480192
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000551182144883677,
      "loss": 2.814,
      "theoretical_loss": 3.8284047951448965,
      "tokens_seen": 617545728
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005511655948553055,
      "loss": 2.5332,
      "theoretical_loss": 3.828362618047013,
      "tokens_seen": 617611264
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005511490448269341,
      "loss": 2.8505,
      "theoretical_loss": 3.828320446677378,
      "tokens_seen": 617676800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005511324947985625,
      "loss": 2.8126,
      "theoretical_loss": 3.8282782810346063,
      "tokens_seen": 617742336
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000551115944770191,
      "loss": 2.8822,
      "theoretical_loss": 3.8282361211173135,
      "tokens_seen": 617807872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005510993947418196,
      "loss": 2.6393,
      "theoretical_loss": 3.8281939669241143,
      "tokens_seen": 617873408
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005510828447134481,
      "loss": 2.9992,
      "theoretical_loss": 3.8281518184536245,
      "tokens_seen": 617938944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005510662946850766,
      "loss": 2.9072,
      "theoretical_loss": 3.82810967570446,
      "tokens_seen": 618004480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005510497446567051,
      "loss": 3.0101,
      "theoretical_loss": 3.8280675386752385,
      "tokens_seen": 618070016
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005510331946283336,
      "loss": 2.8756,
      "theoretical_loss": 3.828025407364577,
      "tokens_seen": 618135552
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005510166445999622,
      "loss": 2.9397,
      "theoretical_loss": 3.827983281771093,
      "tokens_seen": 618201088
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005510000945715906,
      "loss": 2.7027,
      "theoretical_loss": 3.8279411618934045,
      "tokens_seen": 618266624
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005509835445432192,
      "loss": 2.7321,
      "theoretical_loss": 3.8278990477301305,
      "tokens_seen": 618332160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005509669945148477,
      "loss": 2.8495,
      "theoretical_loss": 3.82785693927989,
      "tokens_seen": 618397696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005509504444864762,
      "loss": 2.6353,
      "theoretical_loss": 3.827814836541303,
      "tokens_seen": 618463232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005509338944581048,
      "loss": 2.6469,
      "theoretical_loss": 3.82777273951299,
      "tokens_seen": 618528768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005509173444297333,
      "loss": 3.042,
      "theoretical_loss": 3.8277306481935707,
      "tokens_seen": 618594304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005509007944013618,
      "loss": 2.9204,
      "theoretical_loss": 3.8276885625816663,
      "tokens_seen": 618659840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005508842443729903,
      "loss": 2.5471,
      "theoretical_loss": 3.827646482675899,
      "tokens_seen": 618725376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005508676943446188,
      "loss": 2.6105,
      "theoretical_loss": 3.8276044084748904,
      "tokens_seen": 618790912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005508511443162474,
      "loss": 2.7464,
      "theoretical_loss": 3.8275623399772636,
      "tokens_seen": 618856448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005508345942878759,
      "loss": 2.7766,
      "theoretical_loss": 3.827520277181641,
      "tokens_seen": 618921984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005508180442595044,
      "loss": 2.85,
      "theoretical_loss": 3.8274782200866464,
      "tokens_seen": 618987520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000550801494231133,
      "loss": 2.6458,
      "theoretical_loss": 3.8274361686909035,
      "tokens_seen": 619053056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005507849442027615,
      "loss": 2.944,
      "theoretical_loss": 3.827394122993037,
      "tokens_seen": 619118592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005507683941743899,
      "loss": 2.7725,
      "theoretical_loss": 3.827352082991672,
      "tokens_seen": 619184128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005507518441460185,
      "loss": 2.526,
      "theoretical_loss": 3.8273100486854332,
      "tokens_seen": 619249664
    },
    {
      "epoch": 0.22,
      "objective/train/avg_token_score": 0.018784837797284126,
      "objective/train/avg_weight": 0.9779036045074463,
      "objective/train/docs_used": 359292,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.217559814453125,
      "objective/train/std_weight": 0.10310956835746765,
      "objective/train/theoretical_loss": 3.827268020072948,
      "objective/train/tokens_used": 639775200,
      "theoretical_loss": 3.827268020072948,
      "tokens_seen": 619315200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000550735294117647,
      "loss": 2.8672,
      "theoretical_loss": 3.827268020072948,
      "tokens_seen": 619315200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005507187440892756,
      "loss": 2.9378,
      "theoretical_loss": 3.8272259971528406,
      "tokens_seen": 619380736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005507021940609041,
      "loss": 2.8408,
      "theoretical_loss": 3.8271839799237397,
      "tokens_seen": 619446272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005506856440325326,
      "loss": 2.8569,
      "theoretical_loss": 3.8271419683842716,
      "tokens_seen": 619511808
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005506690940041612,
      "loss": 2.6837,
      "theoretical_loss": 3.8270999625330644,
      "tokens_seen": 619577344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005506525439757897,
      "loss": 2.7017,
      "theoretical_loss": 3.8270579623687464,
      "tokens_seen": 619642880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005506359939474181,
      "loss": 3.1771,
      "theoretical_loss": 3.8270159678899462,
      "tokens_seen": 619708416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005506194439190467,
      "loss": 2.6138,
      "theoretical_loss": 3.826973979095293,
      "tokens_seen": 619773952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005506028938906752,
      "loss": 2.7655,
      "theoretical_loss": 3.8269319959834167,
      "tokens_seen": 619839488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005505863438623038,
      "loss": 2.7182,
      "theoretical_loss": 3.8268900185529473,
      "tokens_seen": 619905024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005505697938339323,
      "loss": 2.8179,
      "theoretical_loss": 3.8268480468025157,
      "tokens_seen": 619970560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005505532438055608,
      "loss": 2.839,
      "theoretical_loss": 3.826806080730753,
      "tokens_seen": 620036096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005505366937771894,
      "loss": 3.0407,
      "theoretical_loss": 3.82676412033629,
      "tokens_seen": 620101632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005505201437488178,
      "loss": 2.919,
      "theoretical_loss": 3.8267221656177606,
      "tokens_seen": 620167168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005505035937204463,
      "loss": 2.6959,
      "theoretical_loss": 3.826680216573795,
      "tokens_seen": 620232704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005504870436920749,
      "loss": 2.9363,
      "theoretical_loss": 3.8266382732030277,
      "tokens_seen": 620298240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005504704936637034,
      "loss": 2.9147,
      "theoretical_loss": 3.8265963355040915,
      "tokens_seen": 620363776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000550453943635332,
      "loss": 2.8623,
      "theoretical_loss": 3.8265544034756216,
      "tokens_seen": 620429312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005504373936069605,
      "loss": 2.6776,
      "theoretical_loss": 3.8265124771162506,
      "tokens_seen": 620494848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000550420843578589,
      "loss": 2.7931,
      "theoretical_loss": 3.8264705564246153,
      "tokens_seen": 620560384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005504042935502176,
      "loss": 2.918,
      "theoretical_loss": 3.8264286413993496,
      "tokens_seen": 620625920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000550387743521846,
      "loss": 2.777,
      "theoretical_loss": 3.82638673203909,
      "tokens_seen": 620691456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005503711934934745,
      "loss": 3.0222,
      "theoretical_loss": 3.8263448283424726,
      "tokens_seen": 620756992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005503546434651031,
      "loss": 2.8611,
      "theoretical_loss": 3.8263029303081346,
      "tokens_seen": 620822528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005503380934367316,
      "loss": 2.7765,
      "theoretical_loss": 3.826261037934713,
      "tokens_seen": 620888064
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005503215434083602,
      "loss": 2.7787,
      "theoretical_loss": 3.8262191512208448,
      "tokens_seen": 620953600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005503049933799887,
      "loss": 2.9327,
      "theoretical_loss": 3.8261772701651697,
      "tokens_seen": 621019136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005502884433516172,
      "loss": 2.9375,
      "theoretical_loss": 3.8261353947663252,
      "tokens_seen": 621084672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005502718933232457,
      "loss": 2.5981,
      "theoretical_loss": 3.8260935250229515,
      "tokens_seen": 621150208
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005502553432948742,
      "loss": 2.8176,
      "theoretical_loss": 3.826051660933687,
      "tokens_seen": 621215744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005502387932665027,
      "loss": 2.7183,
      "theoretical_loss": 3.8260098024971727,
      "tokens_seen": 621281280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005502222432381313,
      "loss": 2.694,
      "theoretical_loss": 3.825967949712049,
      "tokens_seen": 621346816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005502056932097598,
      "loss": 2.9172,
      "theoretical_loss": 3.8259261025769566,
      "tokens_seen": 621412352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005501891431813884,
      "loss": 2.9837,
      "theoretical_loss": 3.8258842610905375,
      "tokens_seen": 621477888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005501725931530169,
      "loss": 3.015,
      "theoretical_loss": 3.825842425251434,
      "tokens_seen": 621543424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005501560431246453,
      "loss": 2.8151,
      "theoretical_loss": 3.8258005950582867,
      "tokens_seen": 621608960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005501394930962739,
      "loss": 2.8831,
      "theoretical_loss": 3.8257587705097404,
      "tokens_seen": 621674496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005501229430679024,
      "loss": 2.76,
      "theoretical_loss": 3.825716951604438,
      "tokens_seen": 621740032
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005501063930395309,
      "loss": 2.7031,
      "theoretical_loss": 3.8256751383410235,
      "tokens_seen": 621805568
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005500898430111595,
      "loss": 3.1092,
      "theoretical_loss": 3.82563333071814,
      "tokens_seen": 621871104
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000550073292982788,
      "loss": 2.7811,
      "theoretical_loss": 3.825591528734434,
      "tokens_seen": 621936640
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005500567429544165,
      "loss": 2.7317,
      "theoretical_loss": 3.82554973238855,
      "tokens_seen": 622002176
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005500401929260449,
      "loss": 2.8276,
      "theoretical_loss": 3.8255079416791338,
      "tokens_seen": 622067712
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005500236428976735,
      "loss": 2.7403,
      "theoretical_loss": 3.825466156604831,
      "tokens_seen": 622133248
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000550007092869302,
      "loss": 2.9943,
      "theoretical_loss": 3.8254243771642895,
      "tokens_seen": 622198784
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005499905428409305,
      "loss": 2.6715,
      "theoretical_loss": 3.8253826033561555,
      "tokens_seen": 622264320
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005499739928125591,
      "loss": 2.979,
      "theoretical_loss": 3.8253408351790767,
      "tokens_seen": 622329856
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005499574427841876,
      "loss": 2.6314,
      "theoretical_loss": 3.8252990726317018,
      "tokens_seen": 622395392
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005499408927558162,
      "loss": 2.6569,
      "theoretical_loss": 3.825257315712678,
      "tokens_seen": 622460928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005499243427274447,
      "loss": 2.6917,
      "theoretical_loss": 3.825215564420656,
      "tokens_seen": 622526464
    },
    {
      "epoch": 0.22,
      "objective/train/avg_token_score": 0.01028356608003378,
      "objective/train/avg_weight": 0.987892746925354,
      "objective/train/docs_used": 360813,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7156033515930176,
      "objective/train/std_weight": 0.06568296253681183,
      "objective/train/theoretical_loss": 3.8251738187542843,
      "objective/train/tokens_used": 643052000,
      "theoretical_loss": 3.8251738187542843,
      "tokens_seen": 622592000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005499077926990731,
      "loss": 2.8087,
      "theoretical_loss": 3.8251738187542843,
      "tokens_seen": 622592000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005498912426707017,
      "loss": 2.7445,
      "theoretical_loss": 3.8251320787122127,
      "tokens_seen": 622657536
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005498746926423302,
      "loss": 2.7345,
      "theoretical_loss": 3.8250903442930917,
      "tokens_seen": 622723072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005498581426139587,
      "loss": 2.9707,
      "theoretical_loss": 3.8250486154955725,
      "tokens_seen": 622788608
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005498415925855873,
      "loss": 2.8692,
      "theoretical_loss": 3.8250068923183065,
      "tokens_seen": 622854144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005498250425572158,
      "loss": 3.0001,
      "theoretical_loss": 3.824965174759945,
      "tokens_seen": 622919680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005498084925288444,
      "loss": 2.7789,
      "theoretical_loss": 3.8249234628191404,
      "tokens_seen": 622985216
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005497919425004728,
      "loss": 2.7225,
      "theoretical_loss": 3.8248817564945465,
      "tokens_seen": 623050752
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005497753924721013,
      "loss": 2.7187,
      "theoretical_loss": 3.8248400557848146,
      "tokens_seen": 623116288
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005497588424437299,
      "loss": 2.8818,
      "theoretical_loss": 3.8247983606885994,
      "tokens_seen": 623181824
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005497422924153584,
      "loss": 3.0003,
      "theoretical_loss": 3.824756671204555,
      "tokens_seen": 623247360
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005497257423869869,
      "loss": 2.7678,
      "theoretical_loss": 3.824714987331336,
      "tokens_seen": 623312896
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005497091923586155,
      "loss": 2.6971,
      "theoretical_loss": 3.824673309067598,
      "tokens_seen": 623378432
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000549692642330244,
      "loss": 3.0095,
      "theoretical_loss": 3.8246316364119948,
      "tokens_seen": 623443968
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005496760923018726,
      "loss": 2.6834,
      "theoretical_loss": 3.824589969363184,
      "tokens_seen": 623509504
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000549659542273501,
      "loss": 2.8394,
      "theoretical_loss": 3.8245483079198213,
      "tokens_seen": 623575040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005496429922451295,
      "loss": 2.8321,
      "theoretical_loss": 3.824506652080564,
      "tokens_seen": 623640576
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005496264422167581,
      "loss": 2.8711,
      "theoretical_loss": 3.824465001844069,
      "tokens_seen": 623706112
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005496098921883866,
      "loss": 2.8042,
      "theoretical_loss": 3.8244233572089943,
      "tokens_seen": 623771648
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005495933421600151,
      "loss": 2.8338,
      "theoretical_loss": 3.824381718173999,
      "tokens_seen": 623837184
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005495767921316437,
      "loss": 2.8745,
      "theoretical_loss": 3.8243400847377407,
      "tokens_seen": 623902720
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005495602421032722,
      "loss": 3.0536,
      "theoretical_loss": 3.824298456898879,
      "tokens_seen": 623968256
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005495436920749007,
      "loss": 2.75,
      "theoretical_loss": 3.8242568346560737,
      "tokens_seen": 624033792
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005495271420465292,
      "loss": 2.6532,
      "theoretical_loss": 3.8242152180079847,
      "tokens_seen": 624099328
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005495105920181577,
      "loss": 2.6831,
      "theoretical_loss": 3.8241736069532726,
      "tokens_seen": 624164864
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005494940419897863,
      "loss": 2.7622,
      "theoretical_loss": 3.8241320014905993,
      "tokens_seen": 624230400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005494774919614148,
      "loss": 2.5661,
      "theoretical_loss": 3.824090401618625,
      "tokens_seen": 624295936
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005494609419330433,
      "loss": 2.8299,
      "theoretical_loss": 3.8240488073360126,
      "tokens_seen": 624361472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005494443919046719,
      "loss": 2.887,
      "theoretical_loss": 3.8240072186414245,
      "tokens_seen": 624427008
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005494278418763003,
      "loss": 2.7256,
      "theoretical_loss": 3.8239656355335234,
      "tokens_seen": 624492544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005494112918479289,
      "loss": 2.7172,
      "theoretical_loss": 3.8239240580109723,
      "tokens_seen": 624558080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005493947418195574,
      "loss": 2.5991,
      "theoretical_loss": 3.8238824860724354,
      "tokens_seen": 624623616
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005493781917911859,
      "loss": 2.8707,
      "theoretical_loss": 3.8238409197165772,
      "tokens_seen": 624689152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005493616417628145,
      "loss": 2.8258,
      "theoretical_loss": 3.823799358942062,
      "tokens_seen": 624754688
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000549345091734443,
      "loss": 2.9461,
      "theoretical_loss": 3.8237578037475553,
      "tokens_seen": 624820224
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005493285417060715,
      "loss": 2.7769,
      "theoretical_loss": 3.823716254131723,
      "tokens_seen": 624885760
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005493119916777,
      "loss": 2.8482,
      "theoretical_loss": 3.823674710093231,
      "tokens_seen": 624951296
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005492954416493285,
      "loss": 2.9364,
      "theoretical_loss": 3.8236331716307452,
      "tokens_seen": 625016832
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005492788916209571,
      "loss": 2.6762,
      "theoretical_loss": 3.8235916387429336,
      "tokens_seen": 625082368
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005492623415925856,
      "loss": 2.6052,
      "theoretical_loss": 3.823550111428463,
      "tokens_seen": 625147904
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005492457915642141,
      "loss": 2.8282,
      "theoretical_loss": 3.8235085896860017,
      "tokens_seen": 625213440
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005492292415358427,
      "loss": 2.8334,
      "theoretical_loss": 3.8234670735142187,
      "tokens_seen": 625278976
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005492126915074712,
      "loss": 2.7186,
      "theoretical_loss": 3.8234255629117815,
      "tokens_seen": 625344512
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005491961414790997,
      "loss": 2.771,
      "theoretical_loss": 3.8233840578773606,
      "tokens_seen": 625410048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005491795914507282,
      "loss": 2.7442,
      "theoretical_loss": 3.8233425584096254,
      "tokens_seen": 625475584
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005491630414223567,
      "loss": 2.6514,
      "theoretical_loss": 3.8233010645072456,
      "tokens_seen": 625541120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005491464913939853,
      "loss": 2.9956,
      "theoretical_loss": 3.8232595761688932,
      "tokens_seen": 625606656
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005491299413656138,
      "loss": 2.7881,
      "theoretical_loss": 3.8232180933932383,
      "tokens_seen": 625672192
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005491133913372423,
      "loss": 2.9712,
      "theoretical_loss": 3.8231766161789524,
      "tokens_seen": 625737728
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005490968413088709,
      "loss": 2.8201,
      "theoretical_loss": 3.8231351445247084,
      "tokens_seen": 625803264
    },
    {
      "epoch": 0.22,
      "objective/train/avg_token_score": 0.024863842874765396,
      "objective/train/avg_weight": 0.9707705974578857,
      "objective/train/docs_used": 362724,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6833903789520264,
      "objective/train/std_weight": 0.10043535381555557,
      "objective/train/theoretical_loss": 3.8230936784291787,
      "objective/train/tokens_used": 646328800,
      "theoretical_loss": 3.8230936784291787,
      "tokens_seen": 625868800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005490802912804993,
      "loss": 2.6582,
      "theoretical_loss": 3.8230936784291787,
      "tokens_seen": 625868800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005490637412521278,
      "loss": 2.9885,
      "theoretical_loss": 3.8230522178910356,
      "tokens_seen": 625934336
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005490471912237563,
      "loss": 3.0684,
      "theoretical_loss": 3.823010762908953,
      "tokens_seen": 625999872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005490306411953848,
      "loss": 2.7349,
      "theoretical_loss": 3.8229693134816043,
      "tokens_seen": 626065408
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005490140911670134,
      "loss": 2.8379,
      "theoretical_loss": 3.822927869607664,
      "tokens_seen": 626130944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005489975411386419,
      "loss": 2.9224,
      "theoretical_loss": 3.8228864312858084,
      "tokens_seen": 626196480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005489809911102705,
      "loss": 2.9194,
      "theoretical_loss": 3.822844998514711,
      "tokens_seen": 626262016
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000548964441081899,
      "loss": 2.8918,
      "theoretical_loss": 3.822803571293048,
      "tokens_seen": 626327552
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005489478910535274,
      "loss": 2.7512,
      "theoretical_loss": 3.8227621496194946,
      "tokens_seen": 626393088
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000548931341025156,
      "loss": 2.9212,
      "theoretical_loss": 3.8227207334927296,
      "tokens_seen": 626458624
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005489147909967845,
      "loss": 2.8851,
      "theoretical_loss": 3.822679322911429,
      "tokens_seen": 626524160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005488982409684131,
      "loss": 2.8576,
      "theoretical_loss": 3.8226379178742693,
      "tokens_seen": 626589696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005488816909400416,
      "loss": 2.455,
      "theoretical_loss": 3.82259651837993,
      "tokens_seen": 626655232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005488651409116701,
      "loss": 3.0267,
      "theoretical_loss": 3.822555124427089,
      "tokens_seen": 626720768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005488485908832987,
      "loss": 2.8726,
      "theoretical_loss": 3.8225137360144252,
      "tokens_seen": 626786304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005488320408549272,
      "loss": 2.6929,
      "theoretical_loss": 3.8224723531406175,
      "tokens_seen": 626851840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005488154908265556,
      "loss": 2.8835,
      "theoretical_loss": 3.8224309758043464,
      "tokens_seen": 626917376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005487989407981842,
      "loss": 2.7802,
      "theoretical_loss": 3.8223896040042913,
      "tokens_seen": 626982912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005487823907698127,
      "loss": 2.7241,
      "theoretical_loss": 3.822348237739134,
      "tokens_seen": 627048448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005487658407414413,
      "loss": 2.6739,
      "theoretical_loss": 3.8223068770075552,
      "tokens_seen": 627113984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005487492907130698,
      "loss": 2.8969,
      "theoretical_loss": 3.822265521808236,
      "tokens_seen": 627179520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005487327406846983,
      "loss": 2.7389,
      "theoretical_loss": 3.822224172139859,
      "tokens_seen": 627245056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005487161906563269,
      "loss": 2.8335,
      "theoretical_loss": 3.8221828280011065,
      "tokens_seen": 627310592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005486996406279553,
      "loss": 2.7718,
      "theoretical_loss": 3.8221414893906616,
      "tokens_seen": 627376128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005486830905995838,
      "loss": 3.0263,
      "theoretical_loss": 3.822100156307208,
      "tokens_seen": 627441664
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005486665405712124,
      "loss": 2.8159,
      "theoretical_loss": 3.8220588287494284,
      "tokens_seen": 627507200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005486499905428409,
      "loss": 2.6676,
      "theoretical_loss": 3.8220175067160085,
      "tokens_seen": 627572736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005486334405144695,
      "loss": 2.8889,
      "theoretical_loss": 3.821976190205632,
      "tokens_seen": 627638272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000548616890486098,
      "loss": 2.919,
      "theoretical_loss": 3.821934879216985,
      "tokens_seen": 627703808
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005486003404577265,
      "loss": 2.8036,
      "theoretical_loss": 3.8218935737487527,
      "tokens_seen": 627769344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005485837904293551,
      "loss": 2.84,
      "theoretical_loss": 3.821852273799621,
      "tokens_seen": 627834880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005485672404009835,
      "loss": 2.9261,
      "theoretical_loss": 3.8218109793682773,
      "tokens_seen": 627900416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000548550690372612,
      "loss": 2.8106,
      "theoretical_loss": 3.821769690453407,
      "tokens_seen": 627965952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005485341403442406,
      "loss": 2.8296,
      "theoretical_loss": 3.8217284070536994,
      "tokens_seen": 628031488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005485175903158691,
      "loss": 2.9196,
      "theoretical_loss": 3.8216871291678416,
      "tokens_seen": 628097024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005485010402874977,
      "loss": 2.8287,
      "theoretical_loss": 3.8216458567945217,
      "tokens_seen": 628162560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005484844902591262,
      "loss": 2.9696,
      "theoretical_loss": 3.821604589932429,
      "tokens_seen": 628228096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005484679402307547,
      "loss": 3.0048,
      "theoretical_loss": 3.821563328580253,
      "tokens_seen": 628293632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005484513902023832,
      "loss": 2.6799,
      "theoretical_loss": 3.8215220727366823,
      "tokens_seen": 628359168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005484348401740117,
      "loss": 2.8395,
      "theoretical_loss": 3.8214808224004075,
      "tokens_seen": 628424704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005484182901456402,
      "loss": 3.1072,
      "theoretical_loss": 3.82143957757012,
      "tokens_seen": 628490240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005484017401172688,
      "loss": 3.0179,
      "theoretical_loss": 3.82139833824451,
      "tokens_seen": 628555776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005483851900888973,
      "loss": 2.7736,
      "theoretical_loss": 3.8213571044222694,
      "tokens_seen": 628621312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005483686400605259,
      "loss": 2.7796,
      "theoretical_loss": 3.82131587610209,
      "tokens_seen": 628686848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005483520900321544,
      "loss": 2.6322,
      "theoretical_loss": 3.8212746532826642,
      "tokens_seen": 628752384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005483355400037828,
      "loss": 2.7482,
      "theoretical_loss": 3.821233435962685,
      "tokens_seen": 628817920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005483189899754114,
      "loss": 2.8055,
      "theoretical_loss": 3.8211922241408454,
      "tokens_seen": 628883456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005483024399470399,
      "loss": 2.8437,
      "theoretical_loss": 3.821151017815839,
      "tokens_seen": 628948992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005482858899186684,
      "loss": 2.726,
      "theoretical_loss": 3.8211098169863607,
      "tokens_seen": 629014528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000548269339890297,
      "loss": 2.7375,
      "theoretical_loss": 3.8210686216511043,
      "tokens_seen": 629080064
    },
    {
      "epoch": 0.22,
      "objective/train/avg_token_score": 0.020775306969881058,
      "objective/train/avg_weight": 0.9755895137786865,
      "objective/train/docs_used": 364484,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8620529174804688,
      "objective/train/std_weight": 0.09585858136415482,
      "objective/train/theoretical_loss": 3.8210274318087656,
      "objective/train/tokens_used": 649605600,
      "theoretical_loss": 3.8210274318087656,
      "tokens_seen": 629145600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005482527898619255,
      "loss": 2.9188,
      "theoretical_loss": 3.8210274318087656,
      "tokens_seen": 629145600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005482362398335541,
      "loss": 3.0608,
      "theoretical_loss": 3.8209862474580394,
      "tokens_seen": 629211136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005482196898051826,
      "loss": 2.6601,
      "theoretical_loss": 3.8209450685976227,
      "tokens_seen": 629276672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000548203139776811,
      "loss": 2.8414,
      "theoretical_loss": 3.820903895226211,
      "tokens_seen": 629342208
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005481865897484396,
      "loss": 2.6454,
      "theoretical_loss": 3.820862727342501,
      "tokens_seen": 629407744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005481700397200681,
      "loss": 2.7467,
      "theoretical_loss": 3.8208215649451907,
      "tokens_seen": 629473280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005481534896916966,
      "loss": 2.8142,
      "theoretical_loss": 3.8207804080329772,
      "tokens_seen": 629538816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005481369396633252,
      "loss": 2.7353,
      "theoretical_loss": 3.8207392566045595,
      "tokens_seen": 629604352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005481203896349537,
      "loss": 2.8037,
      "theoretical_loss": 3.8206981106586357,
      "tokens_seen": 629669888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005481038396065822,
      "loss": 2.8602,
      "theoretical_loss": 3.820656970193905,
      "tokens_seen": 629735424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005480872895782106,
      "loss": 2.8293,
      "theoretical_loss": 3.8206158352090664,
      "tokens_seen": 629800960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005480707395498392,
      "loss": 2.7624,
      "theoretical_loss": 3.820574705702821,
      "tokens_seen": 629866496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0005480541895214677,
      "loss": 2.6444,
      "theoretical_loss": 3.8205335816738684,
      "tokens_seen": 629932032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005480376394930962,
      "loss": 2.8708,
      "theoretical_loss": 3.8204924631209094,
      "tokens_seen": 629997568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005480210894647248,
      "loss": 2.728,
      "theoretical_loss": 3.820451350042646,
      "tokens_seen": 630063104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005480045394363533,
      "loss": 2.6703,
      "theoretical_loss": 3.820410242437779,
      "tokens_seen": 630128640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005479879894079818,
      "loss": 2.7552,
      "theoretical_loss": 3.8203691403050106,
      "tokens_seen": 630194176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005479714393796103,
      "loss": 2.9266,
      "theoretical_loss": 3.8203280436430447,
      "tokens_seen": 630259712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005479548893512388,
      "loss": 2.6293,
      "theoretical_loss": 3.8202869524505827,
      "tokens_seen": 630325248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005479383393228674,
      "loss": 2.6227,
      "theoretical_loss": 3.8202458667263297,
      "tokens_seen": 630390784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005479217892944959,
      "loss": 3.03,
      "theoretical_loss": 3.8202047864689885,
      "tokens_seen": 630456320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005479052392661244,
      "loss": 3.0109,
      "theoretical_loss": 3.8201637116772638,
      "tokens_seen": 630521856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000547888689237753,
      "loss": 2.8562,
      "theoretical_loss": 3.820122642349861,
      "tokens_seen": 630587392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005478721392093815,
      "loss": 3.0337,
      "theoretical_loss": 3.8200815784854845,
      "tokens_seen": 630652928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00054785558918101,
      "loss": 2.6894,
      "theoretical_loss": 3.8200405200828405,
      "tokens_seen": 630718464
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005478390391526385,
      "loss": 2.9198,
      "theoretical_loss": 3.8199994671406348,
      "tokens_seen": 630784000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000547822489124267,
      "loss": 2.6007,
      "theoretical_loss": 3.819958419657574,
      "tokens_seen": 630849536
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005478059390958956,
      "loss": 2.7508,
      "theoretical_loss": 3.8199173776323665,
      "tokens_seen": 630915072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005477893890675241,
      "loss": 2.7652,
      "theoretical_loss": 3.819876341063718,
      "tokens_seen": 630980608
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005477728390391526,
      "loss": 2.8543,
      "theoretical_loss": 3.8198353099503373,
      "tokens_seen": 631046144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005477562890107812,
      "loss": 2.6671,
      "theoretical_loss": 3.8197942842909325,
      "tokens_seen": 631111680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005477397389824097,
      "loss": 2.83,
      "theoretical_loss": 3.8197532640842127,
      "tokens_seen": 631177216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005477231889540381,
      "loss": 2.8943,
      "theoretical_loss": 3.8197122493288864,
      "tokens_seen": 631242752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005477066389256667,
      "loss": 2.5596,
      "theoretical_loss": 3.819671240023664,
      "tokens_seen": 631308288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005476900888972952,
      "loss": 2.7909,
      "theoretical_loss": 3.8196302361672556,
      "tokens_seen": 631373824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005476735388689238,
      "loss": 2.8846,
      "theoretical_loss": 3.8195892377583713,
      "tokens_seen": 631439360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005476569888405523,
      "loss": 2.9311,
      "theoretical_loss": 3.8195482447957225,
      "tokens_seen": 631504896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005476404388121808,
      "loss": 2.8518,
      "theoretical_loss": 3.819507257278021,
      "tokens_seen": 631570432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005476238887838094,
      "loss": 2.5773,
      "theoretical_loss": 3.8194662752039776,
      "tokens_seen": 631635968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005476073387554378,
      "loss": 2.8256,
      "theoretical_loss": 3.8194252985723054,
      "tokens_seen": 631701504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005475907887270663,
      "loss": 2.7174,
      "theoretical_loss": 3.8193843273817167,
      "tokens_seen": 631767040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005475742386986949,
      "loss": 2.6152,
      "theoretical_loss": 3.819343361630925,
      "tokens_seen": 631832576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005475576886703234,
      "loss": 2.673,
      "theoretical_loss": 3.8193024013186445,
      "tokens_seen": 631898112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000547541138641952,
      "loss": 2.7377,
      "theoretical_loss": 3.819261446443588,
      "tokens_seen": 631963648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005475245886135805,
      "loss": 2.8531,
      "theoretical_loss": 3.819220497004471,
      "tokens_seen": 632029184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000547508038585209,
      "loss": 2.7856,
      "theoretical_loss": 3.8191795530000077,
      "tokens_seen": 632094720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005474914885568376,
      "loss": 2.7088,
      "theoretical_loss": 3.8191386144289137,
      "tokens_seen": 632160256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000547474938528466,
      "loss": 2.8835,
      "theoretical_loss": 3.8190976812899056,
      "tokens_seen": 632225792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005474583885000946,
      "loss": 2.9468,
      "theoretical_loss": 3.819056753581699,
      "tokens_seen": 632291328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005474418384717231,
      "loss": 2.8038,
      "theoretical_loss": 3.8190158313030103,
      "tokens_seen": 632356864
    },
    {
      "epoch": 0.23,
      "objective/train/avg_token_score": 0.00837642326951027,
      "objective/train/avg_weight": 0.9901455640792847,
      "objective/train/docs_used": 365542,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7984986305236816,
      "objective/train/std_weight": 0.05576125532388687,
      "objective/train/theoretical_loss": 3.818974914452557,
      "objective/train/tokens_used": 652882400,
      "theoretical_loss": 3.818974914452557,
      "tokens_seen": 632422400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005474252884433516,
      "loss": 2.6993,
      "theoretical_loss": 3.818974914452557,
      "tokens_seen": 632422400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005474087384149802,
      "loss": 2.7281,
      "theoretical_loss": 3.8189340030290566,
      "tokens_seen": 632487936
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005473921883866087,
      "loss": 2.8661,
      "theoretical_loss": 3.8188930970312267,
      "tokens_seen": 632553472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005473756383582372,
      "loss": 2.7295,
      "theoretical_loss": 3.8188521964577866,
      "tokens_seen": 632619008
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005473590883298657,
      "loss": 2.9786,
      "theoretical_loss": 3.8188113013074547,
      "tokens_seen": 632684544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005473425383014942,
      "loss": 2.8186,
      "theoretical_loss": 3.8187704115789503,
      "tokens_seen": 632750080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005473259882731228,
      "loss": 2.5844,
      "theoretical_loss": 3.8187295272709934,
      "tokens_seen": 632815616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005473094382447513,
      "loss": 2.7689,
      "theoretical_loss": 3.8186886483823033,
      "tokens_seen": 632881152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005472928882163798,
      "loss": 2.6956,
      "theoretical_loss": 3.818647774911601,
      "tokens_seen": 632946688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005472763381880084,
      "loss": 2.9571,
      "theoretical_loss": 3.8186069068576085,
      "tokens_seen": 633012224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005472597881596369,
      "loss": 2.6618,
      "theoretical_loss": 3.818566044219046,
      "tokens_seen": 633077760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005472432381312653,
      "loss": 2.6999,
      "theoretical_loss": 3.818525186994636,
      "tokens_seen": 633143296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005472266881028939,
      "loss": 2.8,
      "theoretical_loss": 3.8184843351831006,
      "tokens_seen": 633208832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005472101380745224,
      "loss": 2.869,
      "theoretical_loss": 3.8184434887831626,
      "tokens_seen": 633274368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000547193588046151,
      "loss": 2.6804,
      "theoretical_loss": 3.8184026477935458,
      "tokens_seen": 633339904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005471770380177795,
      "loss": 2.7892,
      "theoretical_loss": 3.8183618122129737,
      "tokens_seen": 633405440
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000547160487989408,
      "loss": 2.8997,
      "theoretical_loss": 3.818320982040169,
      "tokens_seen": 633470976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005471439379610366,
      "loss": 2.8494,
      "theoretical_loss": 3.8182801572738585,
      "tokens_seen": 633536512
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005471273879326651,
      "loss": 2.6708,
      "theoretical_loss": 3.8182393379127655,
      "tokens_seen": 633602048
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005471108379042935,
      "loss": 2.7862,
      "theoretical_loss": 3.8181985239556155,
      "tokens_seen": 633667584
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000547094287875922,
      "loss": 2.8044,
      "theoretical_loss": 3.8181577154011346,
      "tokens_seen": 633733120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005470777378475505,
      "loss": 2.6788,
      "theoretical_loss": 3.8181169122480494,
      "tokens_seen": 633798656
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005470611878191791,
      "loss": 2.8396,
      "theoretical_loss": 3.818076114495086,
      "tokens_seen": 633864192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005470446377908076,
      "loss": 2.7485,
      "theoretical_loss": 3.8180353221409717,
      "tokens_seen": 633929728
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005470280877624362,
      "loss": 2.8641,
      "theoretical_loss": 3.817994535184434,
      "tokens_seen": 633995264
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005470115377340647,
      "loss": 2.7616,
      "theoretical_loss": 3.8179537536242005,
      "tokens_seen": 634060800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005469949877056931,
      "loss": 2.7399,
      "theoretical_loss": 3.8179129774590006,
      "tokens_seen": 634126336
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005469784376773217,
      "loss": 2.7641,
      "theoretical_loss": 3.8178722066875626,
      "tokens_seen": 634191872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005469618876489502,
      "loss": 2.9359,
      "theoretical_loss": 3.8178314413086154,
      "tokens_seen": 634257408
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005469453376205787,
      "loss": 2.8415,
      "theoretical_loss": 3.817790681320889,
      "tokens_seen": 634322944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005469287875922073,
      "loss": 2.8691,
      "theoretical_loss": 3.817749926723114,
      "tokens_seen": 634388480
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005469122375638358,
      "loss": 2.6469,
      "theoretical_loss": 3.8177091775140197,
      "tokens_seen": 634454016
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005468956875354644,
      "loss": 2.7271,
      "theoretical_loss": 3.8176684336923383,
      "tokens_seen": 634519552
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005468791375070928,
      "loss": 2.8392,
      "theoretical_loss": 3.8176276952568005,
      "tokens_seen": 634585088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005468625874787213,
      "loss": 2.9058,
      "theoretical_loss": 3.8175869622061382,
      "tokens_seen": 634650624
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005468460374503499,
      "loss": 2.5591,
      "theoretical_loss": 3.8175462345390843,
      "tokens_seen": 634716160
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005468294874219784,
      "loss": 2.8405,
      "theoretical_loss": 3.817505512254371,
      "tokens_seen": 634781696
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005468129373936069,
      "loss": 2.8843,
      "theoretical_loss": 3.8174647953507312,
      "tokens_seen": 634847232
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005467963873652355,
      "loss": 3.0449,
      "theoretical_loss": 3.817424083826899,
      "tokens_seen": 634912768
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000546779837336864,
      "loss": 2.7659,
      "theoretical_loss": 3.817383377681608,
      "tokens_seen": 634978304
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005467632873084926,
      "loss": 2.7599,
      "theoretical_loss": 3.8173426769135927,
      "tokens_seen": 635043840
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000546746737280121,
      "loss": 2.7598,
      "theoretical_loss": 3.8173019815215885,
      "tokens_seen": 635109376
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005467301872517495,
      "loss": 2.5152,
      "theoretical_loss": 3.8172612915043294,
      "tokens_seen": 635174912
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005467136372233781,
      "loss": 2.7972,
      "theoretical_loss": 3.8172206068605523,
      "tokens_seen": 635240448
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005466970871950066,
      "loss": 3.014,
      "theoretical_loss": 3.8171799275889926,
      "tokens_seen": 635305984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005466805371666351,
      "loss": 2.7542,
      "theoretical_loss": 3.8171392536883877,
      "tokens_seen": 635371520
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005466639871382637,
      "loss": 2.8213,
      "theoretical_loss": 3.8170985851574732,
      "tokens_seen": 635437056
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005466474371098922,
      "loss": 2.9156,
      "theoretical_loss": 3.8170579219949876,
      "tokens_seen": 635502592
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005466308870815208,
      "loss": 2.8262,
      "theoretical_loss": 3.8170172641996682,
      "tokens_seen": 635568128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005466143370531492,
      "loss": 2.7046,
      "theoretical_loss": 3.816976611770254,
      "tokens_seen": 635633664
    },
    {
      "epoch": 0.23,
      "objective/train/avg_token_score": 0.015703171491622925,
      "objective/train/avg_weight": 0.9815256595611572,
      "objective/train/docs_used": 367606,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7072391510009766,
      "objective/train/std_weight": 0.08237674087285995,
      "objective/train/theoretical_loss": 3.8169359647054835,
      "objective/train/tokens_used": 656159200,
      "theoretical_loss": 3.8169359647054835,
      "tokens_seen": 635699200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005465977870247777,
      "loss": 2.5069,
      "theoretical_loss": 3.8169359647054835,
      "tokens_seen": 635699200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005465812369964063,
      "loss": 3.0127,
      "theoretical_loss": 3.8168953230040947,
      "tokens_seen": 635764736
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005465646869680348,
      "loss": 2.5758,
      "theoretical_loss": 3.8168546866648283,
      "tokens_seen": 635830272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005465481369396633,
      "loss": 3.0292,
      "theoretical_loss": 3.816814055686424,
      "tokens_seen": 635895808
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005465315869112919,
      "loss": 2.8268,
      "theoretical_loss": 3.816773430067622,
      "tokens_seen": 635961344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005465150368829204,
      "loss": 2.8015,
      "theoretical_loss": 3.816732809807163,
      "tokens_seen": 636026880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005464984868545489,
      "loss": 2.7672,
      "theoretical_loss": 3.816692194903789,
      "tokens_seen": 636092416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005464819368261774,
      "loss": 3.0015,
      "theoretical_loss": 3.816651585356241,
      "tokens_seen": 636157952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005464653867978059,
      "loss": 2.764,
      "theoretical_loss": 3.816610981163261,
      "tokens_seen": 636223488
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005464488367694345,
      "loss": 2.5049,
      "theoretical_loss": 3.816570382323592,
      "tokens_seen": 636289024
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000546432286741063,
      "loss": 2.8959,
      "theoretical_loss": 3.8165297888359766,
      "tokens_seen": 636354560
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005464157367126915,
      "loss": 2.7754,
      "theoretical_loss": 3.8164892006991584,
      "tokens_seen": 636420096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005463991866843201,
      "loss": 2.8162,
      "theoretical_loss": 3.816448617911881,
      "tokens_seen": 636485632
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005463826366559485,
      "loss": 2.7772,
      "theoretical_loss": 3.816408040472889,
      "tokens_seen": 636551168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005463660866275771,
      "loss": 2.6625,
      "theoretical_loss": 3.816367468380926,
      "tokens_seen": 636616704
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005463495365992056,
      "loss": 2.6229,
      "theoretical_loss": 3.8163269016347385,
      "tokens_seen": 636682240
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005463329865708341,
      "loss": 2.8408,
      "theoretical_loss": 3.8162863402330713,
      "tokens_seen": 636747776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005463164365424627,
      "loss": 2.8314,
      "theoretical_loss": 3.81624578417467,
      "tokens_seen": 636813312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005462998865140912,
      "loss": 2.7099,
      "theoretical_loss": 3.816205233458281,
      "tokens_seen": 636878848
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005462833364857197,
      "loss": 2.8523,
      "theoretical_loss": 3.816164688082652,
      "tokens_seen": 636944384
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005462667864573483,
      "loss": 2.6201,
      "theoretical_loss": 3.816124148046529,
      "tokens_seen": 637009920
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005462502364289767,
      "loss": 2.9864,
      "theoretical_loss": 3.81608361334866,
      "tokens_seen": 637075456
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005462336864006053,
      "loss": 2.7098,
      "theoretical_loss": 3.8160430839877937,
      "tokens_seen": 637140992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005462171363722338,
      "loss": 2.6279,
      "theoretical_loss": 3.8160025599626777,
      "tokens_seen": 637206528
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005462005863438623,
      "loss": 2.7608,
      "theoretical_loss": 3.815962041272061,
      "tokens_seen": 637272064
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005461840363154909,
      "loss": 2.822,
      "theoretical_loss": 3.815921527914693,
      "tokens_seen": 637337600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005461674862871194,
      "loss": 2.7283,
      "theoretical_loss": 3.815881019889324,
      "tokens_seen": 637403136
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005461509362587479,
      "loss": 2.6933,
      "theoretical_loss": 3.8158405171947027,
      "tokens_seen": 637468672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005461343862303763,
      "loss": 2.4883,
      "theoretical_loss": 3.8158000198295814,
      "tokens_seen": 637534208
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005461178362020049,
      "loss": 2.7344,
      "theoretical_loss": 3.81575952779271,
      "tokens_seen": 637599744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005461012861736334,
      "loss": 2.6575,
      "theoretical_loss": 3.81571904108284,
      "tokens_seen": 637665280
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005460847361452619,
      "loss": 2.5427,
      "theoretical_loss": 3.815678559698723,
      "tokens_seen": 637730816
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005460681861168905,
      "loss": 2.9668,
      "theoretical_loss": 3.8156380836391124,
      "tokens_seen": 637796352
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000546051636088519,
      "loss": 2.7188,
      "theoretical_loss": 3.8155976129027596,
      "tokens_seen": 637861888
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005460350860601475,
      "loss": 2.9298,
      "theoretical_loss": 3.815557147488418,
      "tokens_seen": 637927424
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000546018536031776,
      "loss": 2.8227,
      "theoretical_loss": 3.8155166873948416,
      "tokens_seen": 637992960
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005460019860034045,
      "loss": 2.5871,
      "theoretical_loss": 3.815476232620784,
      "tokens_seen": 638058496
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005459854359750331,
      "loss": 2.5719,
      "theoretical_loss": 3.8154357831649994,
      "tokens_seen": 638124032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005459688859466616,
      "loss": 2.7535,
      "theoretical_loss": 3.8153953390262427,
      "tokens_seen": 638189568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005459523359182901,
      "loss": 2.6264,
      "theoretical_loss": 3.815354900203269,
      "tokens_seen": 638255104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005459357858899187,
      "loss": 2.8842,
      "theoretical_loss": 3.8153144666948338,
      "tokens_seen": 638320640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005459192358615472,
      "loss": 2.6866,
      "theoretical_loss": 3.8152740384996937,
      "tokens_seen": 638386176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005459026858331756,
      "loss": 2.7366,
      "theoretical_loss": 3.815233615616605,
      "tokens_seen": 638451712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005458861358048042,
      "loss": 2.9373,
      "theoretical_loss": 3.815193198044324,
      "tokens_seen": 638517248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005458695857764327,
      "loss": 2.7376,
      "theoretical_loss": 3.8151527857816077,
      "tokens_seen": 638582784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005458530357480613,
      "loss": 2.634,
      "theoretical_loss": 3.815112378827215,
      "tokens_seen": 638648320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005458364857196898,
      "loss": 2.8289,
      "theoretical_loss": 3.815071977179903,
      "tokens_seen": 638713856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005458199356913183,
      "loss": 3.1475,
      "theoretical_loss": 3.815031580838431,
      "tokens_seen": 638779392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005458033856629469,
      "loss": 2.5627,
      "theoretical_loss": 3.814991189801557,
      "tokens_seen": 638844928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005457868356345754,
      "loss": 2.7123,
      "theoretical_loss": 3.8149508040680415,
      "tokens_seen": 638910464
    },
    {
      "epoch": 0.23,
      "objective/train/avg_token_score": 0.019503163173794746,
      "objective/train/avg_weight": 0.9770482778549194,
      "objective/train/docs_used": 369550,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.521787166595459,
      "objective/train/std_weight": 0.11780939251184464,
      "objective/train/theoretical_loss": 3.8149104236366433,
      "objective/train/tokens_used": 659436000,
      "theoretical_loss": 3.8149104236366433,
      "tokens_seen": 638976000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005457702856062038,
      "loss": 2.9767,
      "theoretical_loss": 3.8149104236366433,
      "tokens_seen": 638976000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005457537355778324,
      "loss": 2.7907,
      "theoretical_loss": 3.8148700485061235,
      "tokens_seen": 639041536
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005457371855494609,
      "loss": 2.9667,
      "theoretical_loss": 3.814829678675242,
      "tokens_seen": 639107072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005457206355210895,
      "loss": 2.7248,
      "theoretical_loss": 3.81478931414276,
      "tokens_seen": 639172608
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000545704085492718,
      "loss": 2.7563,
      "theoretical_loss": 3.814748954907439,
      "tokens_seen": 639238144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005456875354643465,
      "loss": 2.636,
      "theoretical_loss": 3.814708600968041,
      "tokens_seen": 639303680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005456709854359751,
      "loss": 2.8304,
      "theoretical_loss": 3.8146682523233286,
      "tokens_seen": 639369216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005456544354076035,
      "loss": 2.7965,
      "theoretical_loss": 3.8146279089720636,
      "tokens_seen": 639434752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000545637885379232,
      "loss": 2.5981,
      "theoretical_loss": 3.8145875709130093,
      "tokens_seen": 639500288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005456213353508606,
      "loss": 2.4976,
      "theoretical_loss": 3.8145472381449297,
      "tokens_seen": 639565824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005456047853224891,
      "loss": 2.6828,
      "theoretical_loss": 3.814506910666589,
      "tokens_seen": 639631360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005455882352941177,
      "loss": 2.6516,
      "theoretical_loss": 3.814466588476751,
      "tokens_seen": 639696896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005455716852657462,
      "loss": 2.6694,
      "theoretical_loss": 3.81442627157418,
      "tokens_seen": 639762432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005455551352373747,
      "loss": 2.7504,
      "theoretical_loss": 3.814385959957643,
      "tokens_seen": 639827968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005455385852090033,
      "loss": 2.9451,
      "theoretical_loss": 3.8143456536259035,
      "tokens_seen": 639893504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005455220351806317,
      "loss": 2.6565,
      "theoretical_loss": 3.814305352577729,
      "tokens_seen": 639959040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005455054851522602,
      "loss": 2.6989,
      "theoretical_loss": 3.8142650568118848,
      "tokens_seen": 640024576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005454889351238888,
      "loss": 2.7301,
      "theoretical_loss": 3.8142247663271385,
      "tokens_seen": 640090112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005454723850955173,
      "loss": 2.7445,
      "theoretical_loss": 3.814184481122257,
      "tokens_seen": 640155648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005454558350671459,
      "loss": 2.7433,
      "theoretical_loss": 3.8141442011960085,
      "tokens_seen": 640221184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005454392850387744,
      "loss": 2.6948,
      "theoretical_loss": 3.814103926547161,
      "tokens_seen": 640286720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005454227350104029,
      "loss": 2.7699,
      "theoretical_loss": 3.8140636571744824,
      "tokens_seen": 640352256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005454061849820314,
      "loss": 2.686,
      "theoretical_loss": 3.8140233930767424,
      "tokens_seen": 640417792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005453896349536599,
      "loss": 3.0703,
      "theoretical_loss": 3.8139831342527093,
      "tokens_seen": 640483328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005453730849252884,
      "loss": 2.8353,
      "theoretical_loss": 3.8139428807011537,
      "tokens_seen": 640548864
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000545356534896917,
      "loss": 3.0013,
      "theoretical_loss": 3.8139026324208456,
      "tokens_seen": 640614400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005453399848685455,
      "loss": 2.6711,
      "theoretical_loss": 3.8138623894105557,
      "tokens_seen": 640679936
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005453234348401741,
      "loss": 2.5474,
      "theoretical_loss": 3.8138221516690547,
      "tokens_seen": 640745472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005453068848118026,
      "loss": 2.715,
      "theoretical_loss": 3.813781919195114,
      "tokens_seen": 640811008
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000545290334783431,
      "loss": 2.7585,
      "theoretical_loss": 3.813741691987506,
      "tokens_seen": 640876544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005452737847550596,
      "loss": 2.9789,
      "theoretical_loss": 3.813701470045001,
      "tokens_seen": 640942080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005452572347266881,
      "loss": 3.0178,
      "theoretical_loss": 3.8136612533663747,
      "tokens_seen": 641007616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005452406846983166,
      "loss": 2.5221,
      "theoretical_loss": 3.8136210419503977,
      "tokens_seen": 641073152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005452241346699452,
      "loss": 2.8288,
      "theoretical_loss": 3.813580835795844,
      "tokens_seen": 641138688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005452075846415737,
      "loss": 2.846,
      "theoretical_loss": 3.8135406349014884,
      "tokens_seen": 641204224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005451910346132023,
      "loss": 2.6018,
      "theoretical_loss": 3.8135004392661047,
      "tokens_seen": 641269760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005451744845848308,
      "loss": 2.5409,
      "theoretical_loss": 3.8134602488884664,
      "tokens_seen": 641335296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005451579345564592,
      "loss": 2.9039,
      "theoretical_loss": 3.8134200637673503,
      "tokens_seen": 641400832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005451413845280877,
      "loss": 2.6364,
      "theoretical_loss": 3.813379883901531,
      "tokens_seen": 641466368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005451248344997162,
      "loss": 2.7131,
      "theoretical_loss": 3.8133397092897843,
      "tokens_seen": 641531904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005451082844713448,
      "loss": 2.4684,
      "theoretical_loss": 3.8132995399308873,
      "tokens_seen": 641597440
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005450917344429733,
      "loss": 2.6766,
      "theoretical_loss": 3.813259375823616,
      "tokens_seen": 641662976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005450751844146018,
      "loss": 2.8621,
      "theoretical_loss": 3.8132192169667483,
      "tokens_seen": 641728512
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005450586343862304,
      "loss": 2.5399,
      "theoretical_loss": 3.813179063359061,
      "tokens_seen": 641794048
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005450420843578588,
      "loss": 2.7252,
      "theoretical_loss": 3.813138914999332,
      "tokens_seen": 641859584
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005450255343294874,
      "loss": 2.6509,
      "theoretical_loss": 3.8130987718863407,
      "tokens_seen": 641925120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005450089843011159,
      "loss": 2.744,
      "theoretical_loss": 3.813058634018865,
      "tokens_seen": 641990656
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005449924342727444,
      "loss": 2.767,
      "theoretical_loss": 3.8130185013956837,
      "tokens_seen": 642056192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000544975884244373,
      "loss": 2.9713,
      "theoretical_loss": 3.812978374015578,
      "tokens_seen": 642121728
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005449593342160015,
      "loss": 2.9693,
      "theoretical_loss": 3.8129382518773265,
      "tokens_seen": 642187264
    },
    {
      "epoch": 0.23,
      "objective/train/avg_token_score": 0.02458992600440979,
      "objective/train/avg_weight": 0.9710794687271118,
      "objective/train/docs_used": 371294,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2486748695373535,
      "objective/train/std_weight": 0.1181398406624794,
      "objective/train/theoretical_loss": 3.8128981349797098,
      "objective/train/tokens_used": 662712800,
      "theoretical_loss": 3.8128981349797098,
      "tokens_seen": 642252800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005449427841876301,
      "loss": 2.9765,
      "theoretical_loss": 3.8128981349797098,
      "tokens_seen": 642252800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005449262341592585,
      "loss": 2.5073,
      "theoretical_loss": 3.8128580233215095,
      "tokens_seen": 642318336
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000544909684130887,
      "loss": 2.7938,
      "theoretical_loss": 3.8128179169015057,
      "tokens_seen": 642383872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005448931341025156,
      "loss": 2.8989,
      "theoretical_loss": 3.8127778157184817,
      "tokens_seen": 642449408
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005448765840741441,
      "loss": 2.7579,
      "theoretical_loss": 3.8127377197712176,
      "tokens_seen": 642514944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005448600340457726,
      "loss": 2.7574,
      "theoretical_loss": 3.8126976290584973,
      "tokens_seen": 642580480
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005448434840174012,
      "loss": 3.0255,
      "theoretical_loss": 3.812657543579103,
      "tokens_seen": 642646016
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005448269339890297,
      "loss": 2.6011,
      "theoretical_loss": 3.812617463331818,
      "tokens_seen": 642711552
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005448103839606583,
      "loss": 2.6218,
      "theoretical_loss": 3.812577388315426,
      "tokens_seen": 642777088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005447938339322867,
      "loss": 2.7034,
      "theoretical_loss": 3.8125373185287117,
      "tokens_seen": 642842624
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005447772839039152,
      "loss": 2.7197,
      "theoretical_loss": 3.812497253970459,
      "tokens_seen": 642908160
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005447607338755438,
      "loss": 2.7701,
      "theoretical_loss": 3.8124571946394528,
      "tokens_seen": 642973696
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005447441838471723,
      "loss": 2.8626,
      "theoretical_loss": 3.812417140534478,
      "tokens_seen": 643039232
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005447276338188008,
      "loss": 2.6893,
      "theoretical_loss": 3.8123770916543216,
      "tokens_seen": 643104768
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005447110837904294,
      "loss": 2.7786,
      "theoretical_loss": 3.8123370479977687,
      "tokens_seen": 643170304
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005446945337620579,
      "loss": 2.6737,
      "theoretical_loss": 3.812297009563606,
      "tokens_seen": 643235840
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005446779837336864,
      "loss": 2.7154,
      "theoretical_loss": 3.8122569763506204,
      "tokens_seen": 643301376
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005446614337053149,
      "loss": 2.6349,
      "theoretical_loss": 3.8122169483576,
      "tokens_seen": 643366912
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005446448836769434,
      "loss": 2.7928,
      "theoretical_loss": 3.812176925583331,
      "tokens_seen": 643432448
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000544628333648572,
      "loss": 3.0584,
      "theoretical_loss": 3.812136908026603,
      "tokens_seen": 643497984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005446117836202005,
      "loss": 2.6714,
      "theoretical_loss": 3.8120968956862034,
      "tokens_seen": 643563520
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000544595233591829,
      "loss": 2.7725,
      "theoretical_loss": 3.8120568885609223,
      "tokens_seen": 643629056
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005445786835634576,
      "loss": 2.9769,
      "theoretical_loss": 3.8120168866495483,
      "tokens_seen": 643694592
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000544562133535086,
      "loss": 2.8079,
      "theoretical_loss": 3.8119768899508713,
      "tokens_seen": 643760128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005445455835067146,
      "loss": 2.7573,
      "theoretical_loss": 3.8119368984636814,
      "tokens_seen": 643825664
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005445290334783431,
      "loss": 2.5584,
      "theoretical_loss": 3.8118969121867696,
      "tokens_seen": 643891200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005445124834499716,
      "loss": 2.9653,
      "theoretical_loss": 3.8118569311189265,
      "tokens_seen": 643956736
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005444959334216002,
      "loss": 2.7473,
      "theoretical_loss": 3.8118169552589434,
      "tokens_seen": 644022272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005444793833932287,
      "loss": 2.6546,
      "theoretical_loss": 3.811776984605612,
      "tokens_seen": 644087808
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005444628333648572,
      "loss": 2.7267,
      "theoretical_loss": 3.811737019157725,
      "tokens_seen": 644153344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005444462833364858,
      "loss": 2.6627,
      "theoretical_loss": 3.811697058914075,
      "tokens_seen": 644218880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005444297333081142,
      "loss": 2.7697,
      "theoretical_loss": 3.8116571038734546,
      "tokens_seen": 644284416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005444131832797428,
      "loss": 2.8038,
      "theoretical_loss": 3.811617154034657,
      "tokens_seen": 644349952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005443966332513713,
      "loss": 2.866,
      "theoretical_loss": 3.811577209396477,
      "tokens_seen": 644415488
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005443800832229998,
      "loss": 2.64,
      "theoretical_loss": 3.8115372699577073,
      "tokens_seen": 644481024
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005443635331946284,
      "loss": 2.7249,
      "theoretical_loss": 3.8114973357171436,
      "tokens_seen": 644546560
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005443469831662569,
      "loss": 2.8293,
      "theoretical_loss": 3.8114574066735805,
      "tokens_seen": 644612096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005443304331378854,
      "loss": 2.7347,
      "theoretical_loss": 3.8114174828258136,
      "tokens_seen": 644677632
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000544313883109514,
      "loss": 2.8634,
      "theoretical_loss": 3.8113775641726386,
      "tokens_seen": 644743168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005442973330811424,
      "loss": 2.9161,
      "theoretical_loss": 3.8113376507128516,
      "tokens_seen": 644808704
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000544280783052771,
      "loss": 2.4977,
      "theoretical_loss": 3.81129774244525,
      "tokens_seen": 644874240
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005442642330243995,
      "loss": 2.9013,
      "theoretical_loss": 3.811257839368629,
      "tokens_seen": 644939776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000544247682996028,
      "loss": 2.7844,
      "theoretical_loss": 3.8112179414817877,
      "tokens_seen": 645005312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005442311329676566,
      "loss": 2.6275,
      "theoretical_loss": 3.811178048783524,
      "tokens_seen": 645070848
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005442145829392851,
      "loss": 2.8378,
      "theoretical_loss": 3.8111381612726345,
      "tokens_seen": 645136384
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005441980329109135,
      "loss": 2.8051,
      "theoretical_loss": 3.8110982789479193,
      "tokens_seen": 645201920
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005441814828825421,
      "loss": 2.9023,
      "theoretical_loss": 3.811058401808177,
      "tokens_seen": 645267456
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005441649328541705,
      "loss": 3.1403,
      "theoretical_loss": 3.8110185298522063,
      "tokens_seen": 645332992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005441483828257991,
      "loss": 2.8405,
      "theoretical_loss": 3.8109786630788083,
      "tokens_seen": 645398528
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005441318327974276,
      "loss": 2.7625,
      "theoretical_loss": 3.8109388014867824,
      "tokens_seen": 645464064
    },
    {
      "epoch": 0.23,
      "objective/train/avg_token_score": 0.01941216178238392,
      "objective/train/avg_weight": 0.9771456122398376,
      "objective/train/docs_used": 373458,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.894101619720459,
      "objective/train/std_weight": 0.09457415342330933,
      "objective/train/theoretical_loss": 3.8108989450749293,
      "objective/train/tokens_used": 665989600,
      "theoretical_loss": 3.8108989450749293,
      "tokens_seen": 645529600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005441152827690562,
      "loss": 2.768,
      "theoretical_loss": 3.8108989450749293,
      "tokens_seen": 645529600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005440987327406847,
      "loss": 2.7578,
      "theoretical_loss": 3.81085909384205,
      "tokens_seen": 645595136
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005440821827123131,
      "loss": 2.5208,
      "theoretical_loss": 3.810819247786946,
      "tokens_seen": 645660672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005440656326839417,
      "loss": 2.9622,
      "theoretical_loss": 3.810779406908419,
      "tokens_seen": 645726208
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005440490826555702,
      "loss": 2.4986,
      "theoretical_loss": 3.8107395712052714,
      "tokens_seen": 645791744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005440325326271987,
      "loss": 2.7047,
      "theoretical_loss": 3.810699740676306,
      "tokens_seen": 645857280
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005440159825988273,
      "loss": 2.8429,
      "theoretical_loss": 3.8106599153203256,
      "tokens_seen": 645922816
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005439994325704558,
      "loss": 2.7825,
      "theoretical_loss": 3.810620095136133,
      "tokens_seen": 645988352
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005439828825420844,
      "loss": 2.6429,
      "theoretical_loss": 3.810580280122533,
      "tokens_seen": 646053888
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005439663325137129,
      "loss": 2.6309,
      "theoretical_loss": 3.8105404702783288,
      "tokens_seen": 646119424
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005439497824853413,
      "loss": 2.9773,
      "theoretical_loss": 3.8105006656023264,
      "tokens_seen": 646184960
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005439332324569699,
      "loss": 2.7992,
      "theoretical_loss": 3.810460866093329,
      "tokens_seen": 646250496
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005439166824285984,
      "loss": 2.8051,
      "theoretical_loss": 3.8104210717501434,
      "tokens_seen": 646316032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005439001324002269,
      "loss": 2.6175,
      "theoretical_loss": 3.8103812825715746,
      "tokens_seen": 646381568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005438835823718555,
      "loss": 2.9185,
      "theoretical_loss": 3.8103414985564292,
      "tokens_seen": 646447104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000543867032343484,
      "loss": 2.8517,
      "theoretical_loss": 3.8103017197035145,
      "tokens_seen": 646512640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005438504823151126,
      "loss": 2.7658,
      "theoretical_loss": 3.810261946011636,
      "tokens_seen": 646578176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000543833932286741,
      "loss": 3.1514,
      "theoretical_loss": 3.8102221774796012,
      "tokens_seen": 646643712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005438173822583695,
      "loss": 3.0328,
      "theoretical_loss": 3.810182414106219,
      "tokens_seen": 646709248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005438008322299981,
      "loss": 2.8182,
      "theoretical_loss": 3.810142655890297,
      "tokens_seen": 646774784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005437842822016266,
      "loss": 2.7647,
      "theoretical_loss": 3.8101029028306437,
      "tokens_seen": 646840320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005437677321732551,
      "loss": 2.7089,
      "theoretical_loss": 3.810063154926068,
      "tokens_seen": 646905856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005437511821448837,
      "loss": 2.9007,
      "theoretical_loss": 3.8100234121753798,
      "tokens_seen": 646971392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005437346321165122,
      "loss": 2.8165,
      "theoretical_loss": 3.8099836745773885,
      "tokens_seen": 647036928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005437180820881408,
      "loss": 2.9302,
      "theoretical_loss": 3.8099439421309036,
      "tokens_seen": 647102464
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005437015320597692,
      "loss": 2.8223,
      "theoretical_loss": 3.8099042148347366,
      "tokens_seen": 647168000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005436849820313977,
      "loss": 2.94,
      "theoretical_loss": 3.809864492687698,
      "tokens_seen": 647233536
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005436684320030263,
      "loss": 2.828,
      "theoretical_loss": 3.8098247756885995,
      "tokens_seen": 647299072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005436518819746548,
      "loss": 2.6045,
      "theoretical_loss": 3.8097850638362516,
      "tokens_seen": 647364608
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005436353319462833,
      "loss": 2.6386,
      "theoretical_loss": 3.8097453571294686,
      "tokens_seen": 647430144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005436187819179119,
      "loss": 2.5797,
      "theoretical_loss": 3.8097056555670608,
      "tokens_seen": 647495680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005436022318895404,
      "loss": 2.7726,
      "theoretical_loss": 3.809665959147843,
      "tokens_seen": 647561216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000543585681861169,
      "loss": 2.6819,
      "theoretical_loss": 3.8096262678706267,
      "tokens_seen": 647626752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005435691318327974,
      "loss": 2.7559,
      "theoretical_loss": 3.8095865817342274,
      "tokens_seen": 647692288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005435525818044259,
      "loss": 2.7133,
      "theoretical_loss": 3.809546900737457,
      "tokens_seen": 647757824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005435360317760545,
      "loss": 2.6507,
      "theoretical_loss": 3.809507224879132,
      "tokens_seen": 647823360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000543519481747683,
      "loss": 2.7925,
      "theoretical_loss": 3.809467554158067,
      "tokens_seen": 647888896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005435029317193115,
      "loss": 2.7591,
      "theoretical_loss": 3.809427888573077,
      "tokens_seen": 647954432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005434863816909401,
      "loss": 2.7381,
      "theoretical_loss": 3.8093882281229767,
      "tokens_seen": 648019968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005434698316625685,
      "loss": 2.7993,
      "theoretical_loss": 3.8093485728065835,
      "tokens_seen": 648085504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005434532816341971,
      "loss": 2.6067,
      "theoretical_loss": 3.809308922622713,
      "tokens_seen": 648151040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005434367316058256,
      "loss": 2.6866,
      "theoretical_loss": 3.8092692775701824,
      "tokens_seen": 648216576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005434201815774541,
      "loss": 2.9207,
      "theoretical_loss": 3.8092296376478094,
      "tokens_seen": 648282112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005434036315490827,
      "loss": 2.8044,
      "theoretical_loss": 3.8091900028544106,
      "tokens_seen": 648347648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005433870815207112,
      "loss": 2.5287,
      "theoretical_loss": 3.809150373188805,
      "tokens_seen": 648413184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005433705314923398,
      "loss": 2.6965,
      "theoretical_loss": 3.80911074864981,
      "tokens_seen": 648478720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005433539814639683,
      "loss": 2.8266,
      "theoretical_loss": 3.8090711292362456,
      "tokens_seen": 648544256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005433374314355967,
      "loss": 2.6729,
      "theoretical_loss": 3.80903151494693,
      "tokens_seen": 648609792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005433208814072253,
      "loss": 2.772,
      "theoretical_loss": 3.808991905780683,
      "tokens_seen": 648675328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005433043313788538,
      "loss": 2.7255,
      "theoretical_loss": 3.8089523017363254,
      "tokens_seen": 648740864
    },
    {
      "epoch": 0.23,
      "objective/train/avg_token_score": 0.01021789014339447,
      "objective/train/avg_weight": 0.9879812002182007,
      "objective/train/docs_used": 375609,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.884678363800049,
      "objective/train/std_weight": 0.0601215623319149,
      "objective/train/theoretical_loss": 3.8089127028126764,
      "objective/train/tokens_used": 669266400,
      "theoretical_loss": 3.8089127028126764,
      "tokens_seen": 648806400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005432877813504823,
      "loss": 2.6743,
      "theoretical_loss": 3.8089127028126764,
      "tokens_seen": 648806400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005432712313221109,
      "loss": 2.8252,
      "theoretical_loss": 3.8088731090085575,
      "tokens_seen": 648871936
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005432546812937394,
      "loss": 2.9354,
      "theoretical_loss": 3.8088335203227897,
      "tokens_seen": 648937472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000543238131265368,
      "loss": 3.0076,
      "theoretical_loss": 3.8087939367541943,
      "tokens_seen": 649003008
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005432215812369964,
      "loss": 2.4577,
      "theoretical_loss": 3.8087543583015933,
      "tokens_seen": 649068544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005432050312086249,
      "loss": 2.5124,
      "theoretical_loss": 3.8087147849638097,
      "tokens_seen": 649134080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005431884811802535,
      "loss": 2.9279,
      "theoretical_loss": 3.808675216739665,
      "tokens_seen": 649199616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005431719311518819,
      "loss": 2.7643,
      "theoretical_loss": 3.808635653627983,
      "tokens_seen": 649265152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005431553811235105,
      "loss": 2.7378,
      "theoretical_loss": 3.8085960956275873,
      "tokens_seen": 649330688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000543138831095139,
      "loss": 2.8977,
      "theoretical_loss": 3.8085565427373016,
      "tokens_seen": 649396224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005431222810667675,
      "loss": 2.7962,
      "theoretical_loss": 3.8085169949559505,
      "tokens_seen": 649461760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000543105731038396,
      "loss": 2.7807,
      "theoretical_loss": 3.808477452282358,
      "tokens_seen": 649527296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005430891810100245,
      "loss": 2.7308,
      "theoretical_loss": 3.8084379147153493,
      "tokens_seen": 649592832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005430726309816531,
      "loss": 2.912,
      "theoretical_loss": 3.808398382253751,
      "tokens_seen": 649658368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005430560809532816,
      "loss": 2.9705,
      "theoretical_loss": 3.8083588548963867,
      "tokens_seen": 649723904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005430395309249101,
      "loss": 2.8702,
      "theoretical_loss": 3.8083193326420846,
      "tokens_seen": 649789440
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005430229808965387,
      "loss": 2.6575,
      "theoretical_loss": 3.808279815489671,
      "tokens_seen": 649854976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005430064308681672,
      "loss": 2.8305,
      "theoretical_loss": 3.8082403034379713,
      "tokens_seen": 649920512
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005429898808397956,
      "loss": 2.8117,
      "theoretical_loss": 3.808200796485815,
      "tokens_seen": 649986048
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005429733308114242,
      "loss": 2.8312,
      "theoretical_loss": 3.8081612946320287,
      "tokens_seen": 650051584
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005429567807830527,
      "loss": 2.8563,
      "theoretical_loss": 3.8081217978754407,
      "tokens_seen": 650117120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005429402307546813,
      "loss": 3.0813,
      "theoretical_loss": 3.80808230621488,
      "tokens_seen": 650182656
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005429236807263098,
      "loss": 2.994,
      "theoretical_loss": 3.8080428196491747,
      "tokens_seen": 650248192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005429071306979383,
      "loss": 2.9133,
      "theoretical_loss": 3.808003338177154,
      "tokens_seen": 650313728
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005428905806695669,
      "loss": 2.7318,
      "theoretical_loss": 3.8079638617976492,
      "tokens_seen": 650379264
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005428740306411954,
      "loss": 2.9539,
      "theoretical_loss": 3.8079243905094886,
      "tokens_seen": 650444800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005428574806128238,
      "loss": 2.6463,
      "theoretical_loss": 3.8078849243115043,
      "tokens_seen": 650510336
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005428409305844524,
      "loss": 2.7129,
      "theoretical_loss": 3.8078454632025256,
      "tokens_seen": 650575872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005428243805560809,
      "loss": 2.6267,
      "theoretical_loss": 3.8078060071813846,
      "tokens_seen": 650641408
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005428078305277095,
      "loss": 2.8272,
      "theoretical_loss": 3.8077665562469134,
      "tokens_seen": 650706944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000542791280499338,
      "loss": 2.8976,
      "theoretical_loss": 3.807727110397942,
      "tokens_seen": 650772480
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005427747304709665,
      "loss": 2.6915,
      "theoretical_loss": 3.8076876696333057,
      "tokens_seen": 650838016
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005427581804425951,
      "loss": 2.7953,
      "theoretical_loss": 3.807648233951835,
      "tokens_seen": 650903552
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005427416304142235,
      "loss": 2.7207,
      "theoretical_loss": 3.8076088033523643,
      "tokens_seen": 650969088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000542725080385852,
      "loss": 2.6685,
      "theoretical_loss": 3.807569377833726,
      "tokens_seen": 651034624
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005427085303574806,
      "loss": 2.8697,
      "theoretical_loss": 3.8075299573947556,
      "tokens_seen": 651100160
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005426919803291091,
      "loss": 2.6592,
      "theoretical_loss": 3.8074905420342864,
      "tokens_seen": 651165696
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005426754303007377,
      "loss": 2.7535,
      "theoretical_loss": 3.8074511317511535,
      "tokens_seen": 651231232
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005426588802723662,
      "loss": 2.924,
      "theoretical_loss": 3.807411726544191,
      "tokens_seen": 651296768
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005426423302439947,
      "loss": 2.8818,
      "theoretical_loss": 3.8073723264122363,
      "tokens_seen": 651362304
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005426257802156233,
      "loss": 2.7247,
      "theoretical_loss": 3.8073329313541233,
      "tokens_seen": 651427840
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005426092301872517,
      "loss": 2.8685,
      "theoretical_loss": 3.80729354136869,
      "tokens_seen": 651493376
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005425926801588802,
      "loss": 2.8861,
      "theoretical_loss": 3.807254156454772,
      "tokens_seen": 651558912
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005425761301305088,
      "loss": 2.7025,
      "theoretical_loss": 3.807214776611206,
      "tokens_seen": 651624448
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005425595801021373,
      "loss": 2.7804,
      "theoretical_loss": 3.8071754018368305,
      "tokens_seen": 651689984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005425430300737659,
      "loss": 2.784,
      "theoretical_loss": 3.807136032130483,
      "tokens_seen": 651755520
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005425264800453944,
      "loss": 2.7724,
      "theoretical_loss": 3.8070966674910007,
      "tokens_seen": 651821056
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005425099300170229,
      "loss": 2.7623,
      "theoretical_loss": 3.8070573079172227,
      "tokens_seen": 651886592
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005424933799886515,
      "loss": 2.4807,
      "theoretical_loss": 3.807017953407989,
      "tokens_seen": 651952128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005424768299602799,
      "loss": 2.8067,
      "theoretical_loss": 3.806978603962137,
      "tokens_seen": 652017664
    },
    {
      "epoch": 0.23,
      "objective/train/avg_token_score": 0.01072576642036438,
      "objective/train/avg_weight": 0.9873833060264587,
      "objective/train/docs_used": 377261,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6714324951171875,
      "objective/train/std_weight": 0.04792843386530876,
      "objective/train/theoretical_loss": 3.8069392595785083,
      "objective/train/tokens_used": 672543200,
      "theoretical_loss": 3.8069392595785083,
      "tokens_seen": 652083200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005424602799319084,
      "loss": 2.7725,
      "theoretical_loss": 3.8069392595785083,
      "tokens_seen": 652083200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000542443729903537,
      "loss": 2.5257,
      "theoretical_loss": 3.806899920255942,
      "tokens_seen": 652148736
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005424271798751655,
      "loss": 2.922,
      "theoretical_loss": 3.8068605859932783,
      "tokens_seen": 652214272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005424106298467941,
      "loss": 2.6694,
      "theoretical_loss": 3.806821256789359,
      "tokens_seen": 652279808
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005423940798184226,
      "loss": 2.9193,
      "theoretical_loss": 3.806781932643024,
      "tokens_seen": 652345344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000542377529790051,
      "loss": 2.7405,
      "theoretical_loss": 3.8067426135531157,
      "tokens_seen": 652410880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005423609797616796,
      "loss": 2.6422,
      "theoretical_loss": 3.8067032995184764,
      "tokens_seen": 652476416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005423444297333081,
      "loss": 2.6704,
      "theoretical_loss": 3.8066639905379485,
      "tokens_seen": 652541952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005423278797049366,
      "loss": 2.6635,
      "theoretical_loss": 3.806624686610374,
      "tokens_seen": 652607488
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005423113296765652,
      "loss": 2.6793,
      "theoretical_loss": 3.8065853877345965,
      "tokens_seen": 652673024
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005422947796481937,
      "loss": 2.8426,
      "theoretical_loss": 3.8065460939094597,
      "tokens_seen": 652738560
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005422782296198223,
      "loss": 2.7376,
      "theoretical_loss": 3.8065068051338073,
      "tokens_seen": 652804096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005422616795914508,
      "loss": 2.8831,
      "theoretical_loss": 3.806467521406484,
      "tokens_seen": 652869632
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005422451295630792,
      "loss": 2.7103,
      "theoretical_loss": 3.806428242726333,
      "tokens_seen": 652935168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005422285795347078,
      "loss": 2.6565,
      "theoretical_loss": 3.8063889690922013,
      "tokens_seen": 653000704
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005422120295063363,
      "loss": 2.6878,
      "theoretical_loss": 3.806349700502933,
      "tokens_seen": 653066240
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005421954794779648,
      "loss": 2.7807,
      "theoretical_loss": 3.806310436957374,
      "tokens_seen": 653131776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005421789294495933,
      "loss": 2.6634,
      "theoretical_loss": 3.8062711784543715,
      "tokens_seen": 653197312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005421623794212219,
      "loss": 2.6618,
      "theoretical_loss": 3.8062319249927707,
      "tokens_seen": 653262848
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005421458293928504,
      "loss": 2.788,
      "theoretical_loss": 3.8061926765714196,
      "tokens_seen": 653328384
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005421292793644788,
      "loss": 2.8007,
      "theoretical_loss": 3.806153433189165,
      "tokens_seen": 653393920
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005421127293361074,
      "loss": 2.6436,
      "theoretical_loss": 3.806114194844855,
      "tokens_seen": 653459456
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005420961793077359,
      "loss": 2.9448,
      "theoretical_loss": 3.8060749615373375,
      "tokens_seen": 653524992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005420796292793644,
      "loss": 2.7991,
      "theoretical_loss": 3.8060357332654604,
      "tokens_seen": 653590528
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000542063079250993,
      "loss": 2.839,
      "theoretical_loss": 3.8059965100280735,
      "tokens_seen": 653656064
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005420465292226215,
      "loss": 2.8296,
      "theoretical_loss": 3.8059572918240248,
      "tokens_seen": 653721600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005420299791942501,
      "loss": 2.9932,
      "theoretical_loss": 3.8059180786521654,
      "tokens_seen": 653787136
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005420134291658786,
      "loss": 2.862,
      "theoretical_loss": 3.805878870511344,
      "tokens_seen": 653852672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000541996879137507,
      "loss": 3.0156,
      "theoretical_loss": 3.805839667400412,
      "tokens_seen": 653918208
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005419803291091356,
      "loss": 2.6512,
      "theoretical_loss": 3.8058004693182195,
      "tokens_seen": 653983744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005419637790807641,
      "loss": 2.9502,
      "theoretical_loss": 3.8057612762636173,
      "tokens_seen": 654049280
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005419472290523926,
      "loss": 2.8137,
      "theoretical_loss": 3.8057220882354574,
      "tokens_seen": 654114816
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005419306790240212,
      "loss": 2.7125,
      "theoretical_loss": 3.8056829052325916,
      "tokens_seen": 654180352
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005419141289956497,
      "loss": 2.7024,
      "theoretical_loss": 3.805643727253872,
      "tokens_seen": 654245888
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005418975789672783,
      "loss": 2.8959,
      "theoretical_loss": 3.8056045542981516,
      "tokens_seen": 654311424
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005418810289389067,
      "loss": 2.6568,
      "theoretical_loss": 3.805565386364283,
      "tokens_seen": 654376960
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005418644789105352,
      "loss": 2.785,
      "theoretical_loss": 3.8055262234511193,
      "tokens_seen": 654442496
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005418479288821638,
      "loss": 2.8358,
      "theoretical_loss": 3.8054870655575153,
      "tokens_seen": 654508032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005418313788537923,
      "loss": 2.7665,
      "theoretical_loss": 3.805447912682324,
      "tokens_seen": 654573568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005418148288254208,
      "loss": 2.7663,
      "theoretical_loss": 3.8054087648244,
      "tokens_seen": 654639104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005417982787970494,
      "loss": 2.9376,
      "theoretical_loss": 3.8053696219825994,
      "tokens_seen": 654704640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005417817287686779,
      "loss": 2.6015,
      "theoretical_loss": 3.805330484155776,
      "tokens_seen": 654770176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005417651787403065,
      "loss": 2.9768,
      "theoretical_loss": 3.8052913513427864,
      "tokens_seen": 654835712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005417486287119349,
      "loss": 2.9027,
      "theoretical_loss": 3.8052522235424853,
      "tokens_seen": 654901248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005417320786835634,
      "loss": 2.8607,
      "theoretical_loss": 3.805213100753731,
      "tokens_seen": 654966784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000541715528655192,
      "loss": 2.8624,
      "theoretical_loss": 3.8051739829753792,
      "tokens_seen": 655032320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005416989786268205,
      "loss": 2.9333,
      "theoretical_loss": 3.8051348702062864,
      "tokens_seen": 655097856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000541682428598449,
      "loss": 3.0224,
      "theoretical_loss": 3.805095762445311,
      "tokens_seen": 655163392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005416658785700776,
      "loss": 2.6176,
      "theoretical_loss": 3.805056659691311,
      "tokens_seen": 655228928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000541649328541706,
      "loss": 2.8605,
      "theoretical_loss": 3.805017561943144,
      "tokens_seen": 655294464
    },
    {
      "debugging/Self-BLEU-5": 0.5265375629586004,
      "debugging/distinct-1-grams": 0.7435820408094715,
      "debugging/distinct-2-grams": 0.9558103821233092,
      "debugging/entropy-1-grams": 5.931434510687563,
      "debugging/entropy-2-grams": 6.886416755326388,
      "debugging/length": 521.9230769230769,
      "debugging/num_segments": 13,
      "epoch": 0.23,
      "objective/train/avg_token_score": 0.022742915898561478,
      "objective/train/avg_weight": 0.9732370972633362,
      "objective/train/docs_used": 379091,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0681204795837402,
      "objective/train/std_weight": 0.09227747470140457,
      "objective/train/theoretical_loss": 3.804978469199669,
      "objective/train/tokens_used": 675820000,
      "theoretical_loss": 3.804978469199669,
      "tokens_seen": 655360000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005416327785133346,
      "loss": 2.8463,
      "theoretical_loss": 3.804978469199669,
      "tokens_seen": 655360000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005416162284849631,
      "loss": 2.9552,
      "theoretical_loss": 3.804939381459745,
      "tokens_seen": 655425536
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005415996784565916,
      "loss": 2.7021,
      "theoretical_loss": 3.8049002987222313,
      "tokens_seen": 655491072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005415831284282202,
      "loss": 2.9116,
      "theoretical_loss": 3.8048612209859876,
      "tokens_seen": 655556608
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005415665783998487,
      "loss": 2.7677,
      "theoretical_loss": 3.8048221482498743,
      "tokens_seen": 655622144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005415500283714772,
      "loss": 2.7489,
      "theoretical_loss": 3.8047830805127516,
      "tokens_seen": 655687680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005415334783431058,
      "loss": 2.843,
      "theoretical_loss": 3.8047440177734804,
      "tokens_seen": 655753216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005415169283147342,
      "loss": 2.8612,
      "theoretical_loss": 3.8047049600309224,
      "tokens_seen": 655818752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005415003782863628,
      "loss": 2.9335,
      "theoretical_loss": 3.804665907283938,
      "tokens_seen": 655884288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005414838282579913,
      "loss": 2.614,
      "theoretical_loss": 3.804626859531391,
      "tokens_seen": 655949824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005414672782296198,
      "loss": 2.8914,
      "theoretical_loss": 3.804587816772142,
      "tokens_seen": 656015360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005414507282012484,
      "loss": 2.8236,
      "theoretical_loss": 3.8045487790050547,
      "tokens_seen": 656080896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005414341781728769,
      "loss": 2.7509,
      "theoretical_loss": 3.8045097462289923,
      "tokens_seen": 656146432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005414176281445054,
      "loss": 2.862,
      "theoretical_loss": 3.804470718442818,
      "tokens_seen": 656211968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000541401078116134,
      "loss": 2.9009,
      "theoretical_loss": 3.8044316956453956,
      "tokens_seen": 656277504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005413845280877624,
      "loss": 2.6969,
      "theoretical_loss": 3.8043926778355894,
      "tokens_seen": 656343040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000541367978059391,
      "loss": 2.6906,
      "theoretical_loss": 3.804353665012264,
      "tokens_seen": 656408576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005413514280310195,
      "loss": 2.614,
      "theoretical_loss": 3.8043146571742845,
      "tokens_seen": 656474112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000541334878002648,
      "loss": 2.9376,
      "theoretical_loss": 3.8042756543205156,
      "tokens_seen": 656539648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005413183279742766,
      "loss": 2.9649,
      "theoretical_loss": 3.8042366564498242,
      "tokens_seen": 656605184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005413017779459051,
      "loss": 2.8239,
      "theoretical_loss": 3.8041976635610752,
      "tokens_seen": 656670720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005412852279175336,
      "loss": 2.8329,
      "theoretical_loss": 3.8041586756531363,
      "tokens_seen": 656736256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005412686778891621,
      "loss": 2.7458,
      "theoretical_loss": 3.804119692724873,
      "tokens_seen": 656801792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005412521278607906,
      "loss": 2.7847,
      "theoretical_loss": 3.804080714775153,
      "tokens_seen": 656867328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005412355778324192,
      "loss": 2.9878,
      "theoretical_loss": 3.8040417418028443,
      "tokens_seen": 656932864
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005412190278040476,
      "loss": 2.9409,
      "theoretical_loss": 3.8040027738068143,
      "tokens_seen": 656998400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005412024777756762,
      "loss": 2.9558,
      "theoretical_loss": 3.8039638107859317,
      "tokens_seen": 657063936
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005411859277473047,
      "loss": 2.7655,
      "theoretical_loss": 3.803924852739065,
      "tokens_seen": 657129472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005411693777189332,
      "loss": 3.124,
      "theoretical_loss": 3.803885899665083,
      "tokens_seen": 657195008
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005411528276905617,
      "loss": 2.719,
      "theoretical_loss": 3.8038469515628557,
      "tokens_seen": 657260544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005411362776621902,
      "loss": 2.497,
      "theoretical_loss": 3.803808008431252,
      "tokens_seen": 657326080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005411197276338187,
      "loss": 2.9088,
      "theoretical_loss": 3.8037690702691433,
      "tokens_seen": 657391616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005411031776054473,
      "loss": 3.0252,
      "theoretical_loss": 3.8037301370753993,
      "tokens_seen": 657457152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005410866275770758,
      "loss": 2.833,
      "theoretical_loss": 3.803691208848891,
      "tokens_seen": 657522688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005410700775487044,
      "loss": 2.7353,
      "theoretical_loss": 3.8036522855884893,
      "tokens_seen": 657588224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005410535275203329,
      "loss": 2.676,
      "theoretical_loss": 3.803613367293067,
      "tokens_seen": 657653760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005410369774919613,
      "loss": 2.719,
      "theoretical_loss": 3.803574453961495,
      "tokens_seen": 657719296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005410204274635899,
      "loss": 3.0238,
      "theoretical_loss": 3.803535545592646,
      "tokens_seen": 657784832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005410038774352184,
      "loss": 2.7242,
      "theoretical_loss": 3.803496642185393,
      "tokens_seen": 657850368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000540987327406847,
      "loss": 2.6039,
      "theoretical_loss": 3.8034577437386083,
      "tokens_seen": 657915904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0005409707773784755,
      "loss": 2.6574,
      "theoretical_loss": 3.8034188502511665,
      "tokens_seen": 657981440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000540954227350104,
      "loss": 2.7541,
      "theoretical_loss": 3.803379961721941,
      "tokens_seen": 658046976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005409376773217326,
      "loss": 2.3951,
      "theoretical_loss": 3.8033410781498054,
      "tokens_seen": 658112512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000540921127293361,
      "loss": 2.7954,
      "theoretical_loss": 3.803302199533635,
      "tokens_seen": 658178048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005409045772649895,
      "loss": 3.0942,
      "theoretical_loss": 3.8032633258723045,
      "tokens_seen": 658243584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005408880272366181,
      "loss": 2.666,
      "theoretical_loss": 3.8032244571646894,
      "tokens_seen": 658309120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005408714772082466,
      "loss": 2.9545,
      "theoretical_loss": 3.8031855934096654,
      "tokens_seen": 658374656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005408549271798752,
      "loss": 2.7046,
      "theoretical_loss": 3.8031467346061083,
      "tokens_seen": 658440192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005408383771515037,
      "loss": 2.6014,
      "theoretical_loss": 3.8031078807528944,
      "tokens_seen": 658505728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005408218271231322,
      "loss": 2.8416,
      "theoretical_loss": 3.803069031848901,
      "tokens_seen": 658571264
    },
    {
      "epoch": 0.24,
      "objective/train/avg_token_score": 0.009473400190472603,
      "objective/train/avg_weight": 0.9888514876365662,
      "objective/train/docs_used": 380902,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.068744659423828,
      "objective/train/std_weight": 0.03942205011844635,
      "objective/train/theoretical_loss": 3.803030187893005,
      "objective/train/tokens_used": 679096800,
      "theoretical_loss": 3.803030187893005,
      "tokens_seen": 658636800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005408052770947608,
      "loss": 2.9966,
      "theoretical_loss": 3.803030187893005,
      "tokens_seen": 658636800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005407887270663892,
      "loss": 2.8402,
      "theoretical_loss": 3.802991348884084,
      "tokens_seen": 658702336
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005407721770380177,
      "loss": 2.8672,
      "theoretical_loss": 3.8029525148210146,
      "tokens_seen": 658767872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005407556270096463,
      "loss": 2.6418,
      "theoretical_loss": 3.802913685702677,
      "tokens_seen": 658833408
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005407390769812748,
      "loss": 2.7537,
      "theoretical_loss": 3.8028748615279495,
      "tokens_seen": 658898944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005407225269529034,
      "loss": 2.883,
      "theoretical_loss": 3.80283604229571,
      "tokens_seen": 658964480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005407059769245319,
      "loss": 2.6513,
      "theoretical_loss": 3.802797228004839,
      "tokens_seen": 659030016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005406894268961604,
      "loss": 2.9581,
      "theoretical_loss": 3.802758418654215,
      "tokens_seen": 659095552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000540672876867789,
      "loss": 2.7134,
      "theoretical_loss": 3.8027196142427195,
      "tokens_seen": 659161088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005406563268394174,
      "loss": 2.9524,
      "theoretical_loss": 3.802680814769232,
      "tokens_seen": 659226624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005406397768110459,
      "loss": 2.6952,
      "theoretical_loss": 3.802642020232633,
      "tokens_seen": 659292160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005406232267826745,
      "loss": 2.7202,
      "theoretical_loss": 3.8026032306318047,
      "tokens_seen": 659357696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000540606676754303,
      "loss": 2.5644,
      "theoretical_loss": 3.8025644459656283,
      "tokens_seen": 659423232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005405901267259316,
      "loss": 2.7991,
      "theoretical_loss": 3.802525666232985,
      "tokens_seen": 659488768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005405735766975601,
      "loss": 3.0441,
      "theoretical_loss": 3.802486891432758,
      "tokens_seen": 659554304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005405570266691886,
      "loss": 2.878,
      "theoretical_loss": 3.802448121563829,
      "tokens_seen": 659619840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005405404766408171,
      "loss": 2.7639,
      "theoretical_loss": 3.802409356625082,
      "tokens_seen": 659685376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005405239266124456,
      "loss": 2.9136,
      "theoretical_loss": 3.8023705966153996,
      "tokens_seen": 659750912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005405073765840741,
      "loss": 2.6529,
      "theoretical_loss": 3.802331841533666,
      "tokens_seen": 659816448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005404908265557027,
      "loss": 2.9819,
      "theoretical_loss": 3.8022930913787647,
      "tokens_seen": 659881984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005404742765273312,
      "loss": 2.6667,
      "theoretical_loss": 3.8022543461495806,
      "tokens_seen": 659947520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005404577264989598,
      "loss": 2.7007,
      "theoretical_loss": 3.8022156058449985,
      "tokens_seen": 660013056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005404411764705883,
      "loss": 2.7718,
      "theoretical_loss": 3.802176870463904,
      "tokens_seen": 660078592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005404246264422167,
      "loss": 2.9799,
      "theoretical_loss": 3.802138140005182,
      "tokens_seen": 660144128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005404080764138453,
      "loss": 2.911,
      "theoretical_loss": 3.8020994144677185,
      "tokens_seen": 660209664
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005403915263854738,
      "loss": 2.5955,
      "theoretical_loss": 3.8020606938504,
      "tokens_seen": 660275200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005403749763571023,
      "loss": 2.8457,
      "theoretical_loss": 3.8020219781521125,
      "tokens_seen": 660340736
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005403584263287309,
      "loss": 2.6151,
      "theoretical_loss": 3.8019832673717437,
      "tokens_seen": 660406272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005403418763003594,
      "loss": 2.8533,
      "theoretical_loss": 3.801944561508181,
      "tokens_seen": 660471808
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000540325326271988,
      "loss": 2.9833,
      "theoretical_loss": 3.801905860560311,
      "tokens_seen": 660537344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005403087762436165,
      "loss": 2.9575,
      "theoretical_loss": 3.8018671645270237,
      "tokens_seen": 660602880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005402922262152449,
      "loss": 3.0066,
      "theoretical_loss": 3.8018284734072054,
      "tokens_seen": 660668416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005402756761868735,
      "loss": 2.5626,
      "theoretical_loss": 3.8017897871997466,
      "tokens_seen": 660733952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000540259126158502,
      "loss": 2.8285,
      "theoretical_loss": 3.801751105903536,
      "tokens_seen": 660799488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005402425761301305,
      "loss": 2.9569,
      "theoretical_loss": 3.8017124295174622,
      "tokens_seen": 660865024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000540226026101759,
      "loss": 2.8096,
      "theoretical_loss": 3.8016737580404163,
      "tokens_seen": 660930560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005402094760733875,
      "loss": 2.7449,
      "theoretical_loss": 3.8016350914712884,
      "tokens_seen": 660996096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005401929260450161,
      "loss": 2.6127,
      "theoretical_loss": 3.801596429808968,
      "tokens_seen": 661061632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005401763760166445,
      "loss": 2.8155,
      "theoretical_loss": 3.8015577730523473,
      "tokens_seen": 661127168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005401598259882731,
      "loss": 2.7388,
      "theoretical_loss": 3.801519121200317,
      "tokens_seen": 661192704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005401432759599016,
      "loss": 2.9864,
      "theoretical_loss": 3.801480474251769,
      "tokens_seen": 661258240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005401267259315301,
      "loss": 2.7829,
      "theoretical_loss": 3.801441832205595,
      "tokens_seen": 661323776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005401101759031587,
      "loss": 2.6816,
      "theoretical_loss": 3.801403195060688,
      "tokens_seen": 661389312
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005400936258747872,
      "loss": 2.6878,
      "theoretical_loss": 3.801364562815941,
      "tokens_seen": 661454848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005400770758464157,
      "loss": 2.8156,
      "theoretical_loss": 3.8013259354702456,
      "tokens_seen": 661520384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005400605258180442,
      "loss": 3.092,
      "theoretical_loss": 3.801287313022497,
      "tokens_seen": 661585920
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005400439757896727,
      "loss": 2.752,
      "theoretical_loss": 3.801248695471588,
      "tokens_seen": 661651456
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005400274257613013,
      "loss": 2.7645,
      "theoretical_loss": 3.8012100828164135,
      "tokens_seen": 661716992
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005400108757329298,
      "loss": 2.9586,
      "theoretical_loss": 3.801171475055867,
      "tokens_seen": 661782528
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005399943257045583,
      "loss": 2.8256,
      "theoretical_loss": 3.8011328721888447,
      "tokens_seen": 661848064
    },
    {
      "epoch": 0.24,
      "objective/train/avg_token_score": 0.01829039305448532,
      "objective/train/avg_weight": 0.9784629940986633,
      "objective/train/docs_used": 382599,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.082620859146118,
      "objective/train/std_weight": 0.12429077923297882,
      "objective/train/theoretical_loss": 3.8010942742142415,
      "objective/train/tokens_used": 682373600,
      "theoretical_loss": 3.8010942742142415,
      "tokens_seen": 661913600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005399777756761869,
      "loss": 2.5677,
      "theoretical_loss": 3.8010942742142415,
      "tokens_seen": 661913600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005399612256478154,
      "loss": 2.6361,
      "theoretical_loss": 3.8010556811309524,
      "tokens_seen": 661979136
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005399446756194438,
      "loss": 2.8321,
      "theoretical_loss": 3.801017092937874,
      "tokens_seen": 662044672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005399281255910724,
      "loss": 2.889,
      "theoretical_loss": 3.800978509633903,
      "tokens_seen": 662110208
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005399115755627009,
      "loss": 2.637,
      "theoretical_loss": 3.8009399312179353,
      "tokens_seen": 662175744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005398950255343295,
      "loss": 2.9923,
      "theoretical_loss": 3.800901357688868,
      "tokens_seen": 662241280
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000539878475505958,
      "loss": 2.7272,
      "theoretical_loss": 3.800862789045599,
      "tokens_seen": 662306816
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005398619254775865,
      "loss": 2.8888,
      "theoretical_loss": 3.800824225287026,
      "tokens_seen": 662372352
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005398453754492151,
      "loss": 2.8274,
      "theoretical_loss": 3.8007856664120476,
      "tokens_seen": 662437888
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005398288254208436,
      "loss": 2.8271,
      "theoretical_loss": 3.800747112419561,
      "tokens_seen": 662503424
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000539812275392472,
      "loss": 2.8779,
      "theoretical_loss": 3.8007085633084663,
      "tokens_seen": 662568960
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005397957253641006,
      "loss": 2.8728,
      "theoretical_loss": 3.8006700190776623,
      "tokens_seen": 662634496
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005397791753357291,
      "loss": 2.8612,
      "theoretical_loss": 3.8006314797260488,
      "tokens_seen": 662700032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005397626253073577,
      "loss": 2.9296,
      "theoretical_loss": 3.800592945252525,
      "tokens_seen": 662765568
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005397460752789862,
      "loss": 2.9597,
      "theoretical_loss": 3.8005544156559923,
      "tokens_seen": 662831104
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005397295252506147,
      "loss": 2.923,
      "theoretical_loss": 3.80051589093535,
      "tokens_seen": 662896640
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005397129752222433,
      "loss": 3.0012,
      "theoretical_loss": 3.8004773710895003,
      "tokens_seen": 662962176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005396964251938717,
      "loss": 2.7607,
      "theoretical_loss": 3.800438856117344,
      "tokens_seen": 663027712
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005396798751655002,
      "loss": 2.8044,
      "theoretical_loss": 3.800400346017783,
      "tokens_seen": 663093248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005396633251371288,
      "loss": 2.9211,
      "theoretical_loss": 3.8003618407897193,
      "tokens_seen": 663158784
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005396467751087573,
      "loss": 2.667,
      "theoretical_loss": 3.800323340432055,
      "tokens_seen": 663224320
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005396302250803859,
      "loss": 2.8604,
      "theoretical_loss": 3.8002848449436932,
      "tokens_seen": 663289856
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005396136750520144,
      "loss": 3.1033,
      "theoretical_loss": 3.800246354323538,
      "tokens_seen": 663355392
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005395971250236429,
      "loss": 2.8487,
      "theoretical_loss": 3.8002078685704914,
      "tokens_seen": 663420928
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005395805749952715,
      "loss": 2.7798,
      "theoretical_loss": 3.8001693876834572,
      "tokens_seen": 663486464
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005395640249668999,
      "loss": 2.7962,
      "theoretical_loss": 3.8001309116613413,
      "tokens_seen": 663552000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005395474749385284,
      "loss": 2.8891,
      "theoretical_loss": 3.8000924405030467,
      "tokens_seen": 663617536
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000539530924910157,
      "loss": 3.1064,
      "theoretical_loss": 3.8000539742074784,
      "tokens_seen": 663683072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005395143748817855,
      "loss": 2.7653,
      "theoretical_loss": 3.8000155127735424,
      "tokens_seen": 663748608
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005394978248534141,
      "loss": 2.8408,
      "theoretical_loss": 3.7999770562001443,
      "tokens_seen": 663814144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005394812748250426,
      "loss": 2.9662,
      "theoretical_loss": 3.799938604486189,
      "tokens_seen": 663879680
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005394647247966711,
      "loss": 2.7391,
      "theoretical_loss": 3.799900157630584,
      "tokens_seen": 663945216
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005394481747682996,
      "loss": 2.6199,
      "theoretical_loss": 3.799861715632236,
      "tokens_seen": 664010752
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005394316247399281,
      "loss": 2.7667,
      "theoretical_loss": 3.7998232784900514,
      "tokens_seen": 664076288
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005394150747115567,
      "loss": 2.8235,
      "theoretical_loss": 3.7997848462029378,
      "tokens_seen": 664141824
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005393985246831852,
      "loss": 2.8945,
      "theoretical_loss": 3.7997464187698027,
      "tokens_seen": 664207360
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005393819746548137,
      "loss": 2.6138,
      "theoretical_loss": 3.799707996189555,
      "tokens_seen": 664272896
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005393654246264423,
      "loss": 2.7281,
      "theoretical_loss": 3.7996695784611023,
      "tokens_seen": 664338432
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005393488745980708,
      "loss": 2.7783,
      "theoretical_loss": 3.799631165583354,
      "tokens_seen": 664403968
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005393323245696992,
      "loss": 2.962,
      "theoretical_loss": 3.7995927575552186,
      "tokens_seen": 664469504
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005393157745413278,
      "loss": 2.7519,
      "theoretical_loss": 3.7995543543756063,
      "tokens_seen": 664535040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005392992245129563,
      "loss": 2.9978,
      "theoretical_loss": 3.799515956043427,
      "tokens_seen": 664600576
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005392826744845849,
      "loss": 3.0845,
      "theoretical_loss": 3.7994775625575903,
      "tokens_seen": 664666112
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005392661244562134,
      "loss": 2.9069,
      "theoretical_loss": 3.799439173917007,
      "tokens_seen": 664731648
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005392495744278419,
      "loss": 2.9805,
      "theoretical_loss": 3.799400790120588,
      "tokens_seen": 664797184
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005392330243994704,
      "loss": 2.8041,
      "theoretical_loss": 3.799362411167245,
      "tokens_seen": 664862720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005392164743710988,
      "loss": 2.6764,
      "theoretical_loss": 3.7993240370558894,
      "tokens_seen": 664928256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005391999243427274,
      "loss": 2.8586,
      "theoretical_loss": 3.7992856677854334,
      "tokens_seen": 664993792
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005391833743143559,
      "loss": 2.873,
      "theoretical_loss": 3.7992473033547887,
      "tokens_seen": 665059328
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005391668242859844,
      "loss": 2.7663,
      "theoretical_loss": 3.799208943762868,
      "tokens_seen": 665124864
    },
    {
      "epoch": 0.24,
      "objective/train/avg_token_score": 0.009505949914455414,
      "objective/train/avg_weight": 0.9888461232185364,
      "objective/train/docs_used": 384593,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2951087951660156,
      "objective/train/std_weight": 0.045469481498003006,
      "objective/train/theoretical_loss": 3.799170589008585,
      "objective/train/tokens_used": 685650400,
      "theoretical_loss": 3.799170589008585,
      "tokens_seen": 665190400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000539150274257613,
      "loss": 3.1404,
      "theoretical_loss": 3.799170589008585,
      "tokens_seen": 665190400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005391337242292415,
      "loss": 3.003,
      "theoretical_loss": 3.799132239090852,
      "tokens_seen": 665255936
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005391171742008701,
      "loss": 2.9287,
      "theoretical_loss": 3.7990938940085845,
      "tokens_seen": 665321472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005391006241724986,
      "loss": 2.7673,
      "theoretical_loss": 3.799055553760695,
      "tokens_seen": 665387008
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000539084074144127,
      "loss": 3.0764,
      "theoretical_loss": 3.7990172183460995,
      "tokens_seen": 665452544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005390675241157556,
      "loss": 3.0675,
      "theoretical_loss": 3.7989788877637105,
      "tokens_seen": 665518080
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005390509740873841,
      "loss": 2.8823,
      "theoretical_loss": 3.7989405620124455,
      "tokens_seen": 665583616
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005390344240590126,
      "loss": 2.5992,
      "theoretical_loss": 3.7989022410912185,
      "tokens_seen": 665649152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005390178740306412,
      "loss": 2.7293,
      "theoretical_loss": 3.7988639249989458,
      "tokens_seen": 665714688
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005390013240022697,
      "loss": 2.9498,
      "theoretical_loss": 3.7988256137345435,
      "tokens_seen": 665780224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005389847739738983,
      "loss": 2.789,
      "theoretical_loss": 3.7987873072969283,
      "tokens_seen": 665845760
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005389682239455267,
      "loss": 2.774,
      "theoretical_loss": 3.798749005685017,
      "tokens_seen": 665911296
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005389516739171552,
      "loss": 2.6435,
      "theoretical_loss": 3.798710708897727,
      "tokens_seen": 665976832
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005389351238887838,
      "loss": 2.7231,
      "theoretical_loss": 3.7986724169339756,
      "tokens_seen": 666042368
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005389185738604123,
      "loss": 2.9166,
      "theoretical_loss": 3.798634129792681,
      "tokens_seen": 666107904
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005389020238320408,
      "loss": 2.8133,
      "theoretical_loss": 3.798595847472761,
      "tokens_seen": 666173440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005388854738036694,
      "loss": 2.7266,
      "theoretical_loss": 3.798557569973135,
      "tokens_seen": 666238976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005388689237752979,
      "loss": 2.7755,
      "theoretical_loss": 3.7985192972927218,
      "tokens_seen": 666304512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005388523737469265,
      "loss": 2.7831,
      "theoretical_loss": 3.79848102943044,
      "tokens_seen": 666370048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005388358237185549,
      "loss": 3.1047,
      "theoretical_loss": 3.79844276638521,
      "tokens_seen": 666435584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005388192736901834,
      "loss": 3.0283,
      "theoretical_loss": 3.7984045081559517,
      "tokens_seen": 666501120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000538802723661812,
      "loss": 2.925,
      "theoretical_loss": 3.7983662547415853,
      "tokens_seen": 666566656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005387861736334405,
      "loss": 2.9954,
      "theoretical_loss": 3.798328006141032,
      "tokens_seen": 666632192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000538769623605069,
      "loss": 2.8836,
      "theoretical_loss": 3.798289762353212,
      "tokens_seen": 666697728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005387530735766976,
      "loss": 2.8854,
      "theoretical_loss": 3.798251523377048,
      "tokens_seen": 666763264
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005387365235483261,
      "loss": 2.6524,
      "theoretical_loss": 3.79821328921146,
      "tokens_seen": 666828800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005387199735199546,
      "loss": 2.9798,
      "theoretical_loss": 3.798175059855372,
      "tokens_seen": 666894336
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005387034234915831,
      "loss": 2.7963,
      "theoretical_loss": 3.7981368353077056,
      "tokens_seen": 666959872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005386868734632116,
      "loss": 2.9031,
      "theoretical_loss": 3.7980986155673833,
      "tokens_seen": 667025408
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005386703234348402,
      "loss": 2.6495,
      "theoretical_loss": 3.7980604006333287,
      "tokens_seen": 667090944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005386537734064687,
      "loss": 2.6295,
      "theoretical_loss": 3.7980221905044655,
      "tokens_seen": 667156480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005386372233780972,
      "loss": 2.7624,
      "theoretical_loss": 3.797983985179717,
      "tokens_seen": 667222016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005386206733497258,
      "loss": 2.7502,
      "theoretical_loss": 3.797945784658008,
      "tokens_seen": 667287552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005386041233213542,
      "loss": 2.8098,
      "theoretical_loss": 3.797907588938262,
      "tokens_seen": 667353088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005385875732929828,
      "loss": 2.6701,
      "theoretical_loss": 3.7978693980194054,
      "tokens_seen": 667418624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005385710232646113,
      "loss": 2.6734,
      "theoretical_loss": 3.797831211900362,
      "tokens_seen": 667484160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005385544732362398,
      "loss": 2.8103,
      "theoretical_loss": 3.7977930305800585,
      "tokens_seen": 667549696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005385379232078684,
      "loss": 2.8537,
      "theoretical_loss": 3.7977548540574206,
      "tokens_seen": 667615232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005385213731794969,
      "loss": 2.8258,
      "theoretical_loss": 3.7977166823313735,
      "tokens_seen": 667680768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005385048231511254,
      "loss": 2.6255,
      "theoretical_loss": 3.7976785154008454,
      "tokens_seen": 667746304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000538488273122754,
      "loss": 2.85,
      "theoretical_loss": 3.797640353264762,
      "tokens_seen": 667811840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005384717230943824,
      "loss": 2.9453,
      "theoretical_loss": 3.797602195922052,
      "tokens_seen": 667877376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000538455173066011,
      "loss": 2.933,
      "theoretical_loss": 3.7975640433716413,
      "tokens_seen": 667942912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005384386230376395,
      "loss": 2.8059,
      "theoretical_loss": 3.797525895612459,
      "tokens_seen": 668008448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000538422073009268,
      "loss": 2.8651,
      "theoretical_loss": 3.797487752643434,
      "tokens_seen": 668073984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005384055229808966,
      "loss": 2.647,
      "theoretical_loss": 3.7974496144634937,
      "tokens_seen": 668139520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005383889729525251,
      "loss": 2.6555,
      "theoretical_loss": 3.797411481071568,
      "tokens_seen": 668205056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005383724229241536,
      "loss": 2.9656,
      "theoretical_loss": 3.797373352466586,
      "tokens_seen": 668270592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005383558728957821,
      "loss": 2.772,
      "theoretical_loss": 3.7973352286474773,
      "tokens_seen": 668336128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005383393228674106,
      "loss": 2.9123,
      "theoretical_loss": 3.7972971096131722,
      "tokens_seen": 668401664
    },
    {
      "epoch": 0.24,
      "objective/train/avg_token_score": 0.021344289183616638,
      "objective/train/avg_weight": 0.9749338626861572,
      "objective/train/docs_used": 386337,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.121288776397705,
      "objective/train/std_weight": 0.09282240271568298,
      "objective/train/theoretical_loss": 3.7972589953626006,
      "objective/train/tokens_used": 688927200,
      "theoretical_loss": 3.7972589953626006,
      "tokens_seen": 668467200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005383227728390392,
      "loss": 2.874,
      "theoretical_loss": 3.7972589953626006,
      "tokens_seen": 668467200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005383062228106677,
      "loss": 2.7545,
      "theoretical_loss": 3.797220885894694,
      "tokens_seen": 668532736
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005382896727822962,
      "loss": 2.8048,
      "theoretical_loss": 3.7971827812083836,
      "tokens_seen": 668598272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005382731227539248,
      "loss": 2.8056,
      "theoretical_loss": 3.7971446813026,
      "tokens_seen": 668663808
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005382565727255532,
      "loss": 2.8489,
      "theoretical_loss": 3.7971065861762754,
      "tokens_seen": 668729344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005382400226971818,
      "loss": 2.7996,
      "theoretical_loss": 3.7970684958283423,
      "tokens_seen": 668794880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005382234726688102,
      "loss": 2.9347,
      "theoretical_loss": 3.797030410257733,
      "tokens_seen": 668860416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005382069226404388,
      "loss": 3.0435,
      "theoretical_loss": 3.7969923294633796,
      "tokens_seen": 668925952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005381903726120673,
      "loss": 2.7989,
      "theoretical_loss": 3.796954253444216,
      "tokens_seen": 668991488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005381738225836958,
      "loss": 2.8121,
      "theoretical_loss": 3.7969161821991757,
      "tokens_seen": 669057024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005381572725553244,
      "loss": 2.747,
      "theoretical_loss": 3.7968781157271922,
      "tokens_seen": 669122560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005381407225269529,
      "loss": 2.7783,
      "theoretical_loss": 3.7968400540272,
      "tokens_seen": 669188096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005381241724985814,
      "loss": 2.7132,
      "theoretical_loss": 3.796801997098134,
      "tokens_seen": 669253632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005381076224702099,
      "loss": 2.9627,
      "theoretical_loss": 3.7967639449389283,
      "tokens_seen": 669319168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005380910724418384,
      "loss": 2.9489,
      "theoretical_loss": 3.7967258975485185,
      "tokens_seen": 669384704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000538074522413467,
      "loss": 2.7988,
      "theoretical_loss": 3.79668785492584,
      "tokens_seen": 669450240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005380579723850955,
      "loss": 2.8585,
      "theoretical_loss": 3.796649817069828,
      "tokens_seen": 669515776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000538041422356724,
      "loss": 2.8172,
      "theoretical_loss": 3.7966117839794205,
      "tokens_seen": 669581312
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005380248723283526,
      "loss": 2.8031,
      "theoretical_loss": 3.796573755653553,
      "tokens_seen": 669646848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005380083222999811,
      "loss": 2.7294,
      "theoretical_loss": 3.7965357320911624,
      "tokens_seen": 669712384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005379917722716095,
      "loss": 2.7726,
      "theoretical_loss": 3.7964977132911866,
      "tokens_seen": 669777920
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005379752222432381,
      "loss": 2.7513,
      "theoretical_loss": 3.796459699252562,
      "tokens_seen": 669843456
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005379586722148666,
      "loss": 2.8119,
      "theoretical_loss": 3.7964216899742276,
      "tokens_seen": 669908992
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005379421221864952,
      "loss": 2.8478,
      "theoretical_loss": 3.7963836854551216,
      "tokens_seen": 669974528
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005379255721581237,
      "loss": 2.8346,
      "theoretical_loss": 3.796345685694182,
      "tokens_seen": 670040064
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005379090221297522,
      "loss": 2.997,
      "theoretical_loss": 3.796307690690348,
      "tokens_seen": 670105600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005378924721013808,
      "loss": 2.9705,
      "theoretical_loss": 3.7962697004425596,
      "tokens_seen": 670171136
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005378759220730093,
      "loss": 2.8905,
      "theoretical_loss": 3.796231714949756,
      "tokens_seen": 670236672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005378593720446377,
      "loss": 2.7473,
      "theoretical_loss": 3.7961937342108767,
      "tokens_seen": 670302208
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005378428220162663,
      "loss": 3.0141,
      "theoretical_loss": 3.796155758224863,
      "tokens_seen": 670367744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005378262719878948,
      "loss": 2.5732,
      "theoretical_loss": 3.796117786990655,
      "tokens_seen": 670433280
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005378097219595234,
      "loss": 2.7236,
      "theoretical_loss": 3.7960798205071935,
      "tokens_seen": 670498816
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005377931719311519,
      "loss": 2.9714,
      "theoretical_loss": 3.79604185877342,
      "tokens_seen": 670564352
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005377766219027804,
      "loss": 2.8385,
      "theoretical_loss": 3.7960039017882767,
      "tokens_seen": 670629888
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000537760071874409,
      "loss": 2.9616,
      "theoretical_loss": 3.7959659495507045,
      "tokens_seen": 670695424
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005377435218460374,
      "loss": 2.9196,
      "theoretical_loss": 3.7959280020596475,
      "tokens_seen": 670760960
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005377269718176659,
      "loss": 2.8133,
      "theoretical_loss": 3.795890059314047,
      "tokens_seen": 670826496
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005377104217892945,
      "loss": 2.9088,
      "theoretical_loss": 3.795852121312846,
      "tokens_seen": 670892032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000537693871760923,
      "loss": 2.7617,
      "theoretical_loss": 3.7958141880549885,
      "tokens_seen": 670957568
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005376773217325516,
      "loss": 2.6889,
      "theoretical_loss": 3.7957762595394184,
      "tokens_seen": 671023104
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005376607717041801,
      "loss": 2.7676,
      "theoretical_loss": 3.79573833576508,
      "tokens_seen": 671088640
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005376442216758086,
      "loss": 3.0965,
      "theoretical_loss": 3.795700416730916,
      "tokens_seen": 671154176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005376276716474372,
      "loss": 2.7826,
      "theoretical_loss": 3.7956625024358726,
      "tokens_seen": 671219712
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005376111216190656,
      "loss": 2.7515,
      "theoretical_loss": 3.795624592878895,
      "tokens_seen": 671285248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005375945715906941,
      "loss": 2.9291,
      "theoretical_loss": 3.795586688058928,
      "tokens_seen": 671350784
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005375780215623227,
      "loss": 2.732,
      "theoretical_loss": 3.7955487879749175,
      "tokens_seen": 671416320
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005375614715339512,
      "loss": 3.0241,
      "theoretical_loss": 3.795510892625809,
      "tokens_seen": 671481856
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005375449215055798,
      "loss": 2.6088,
      "theoretical_loss": 3.7954730020105503,
      "tokens_seen": 671547392
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005375283714772083,
      "loss": 2.8048,
      "theoretical_loss": 3.795435116128087,
      "tokens_seen": 671612928
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005375118214488368,
      "loss": 2.7496,
      "theoretical_loss": 3.795397234977367,
      "tokens_seen": 671678464
    },
    {
      "epoch": 0.24,
      "objective/train/avg_token_score": 0.010157588869333267,
      "objective/train/avg_weight": 0.9880473613739014,
      "objective/train/docs_used": 388157,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.804244041442871,
      "objective/train/std_weight": 0.05250517651438713,
      "objective/train/theoretical_loss": 3.795359358557337,
      "objective/train/tokens_used": 692204000,
      "theoretical_loss": 3.795359358557337,
      "tokens_seen": 671744000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005374952714204653,
      "loss": 2.8876,
      "theoretical_loss": 3.795359358557337,
      "tokens_seen": 671744000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005374787213920938,
      "loss": 2.7725,
      "theoretical_loss": 3.7953214868669454,
      "tokens_seen": 671809536
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005374621713637223,
      "loss": 2.9025,
      "theoretical_loss": 3.79528361990514,
      "tokens_seen": 671875072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005374456213353509,
      "loss": 2.5769,
      "theoretical_loss": 3.7952457576708687,
      "tokens_seen": 671940608
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005374290713069794,
      "loss": 2.8694,
      "theoretical_loss": 3.7952079001630814,
      "tokens_seen": 672006144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000537412521278608,
      "loss": 2.9245,
      "theoretical_loss": 3.7951700473807266,
      "tokens_seen": 672071680
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005373959712502365,
      "loss": 2.6239,
      "theoretical_loss": 3.795132199322754,
      "tokens_seen": 672137216
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005373794212218649,
      "loss": 2.7958,
      "theoretical_loss": 3.795094355988113,
      "tokens_seen": 672202752
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005373628711934935,
      "loss": 2.766,
      "theoretical_loss": 3.795056517375754,
      "tokens_seen": 672268288
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000537346321165122,
      "loss": 2.8292,
      "theoretical_loss": 3.795018683484628,
      "tokens_seen": 672333824
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005373297711367505,
      "loss": 3.0172,
      "theoretical_loss": 3.7949808543136845,
      "tokens_seen": 672399360
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005373132211083791,
      "loss": 2.6569,
      "theoretical_loss": 3.7949430298618756,
      "tokens_seen": 672464896
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005372966710800076,
      "loss": 2.7867,
      "theoretical_loss": 3.794905210128152,
      "tokens_seen": 672530432
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005372801210516361,
      "loss": 2.5708,
      "theoretical_loss": 3.794867395111467,
      "tokens_seen": 672595968
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005372635710232645,
      "loss": 2.8448,
      "theoretical_loss": 3.7948295848107714,
      "tokens_seen": 672661504
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005372470209948931,
      "loss": 2.7434,
      "theoretical_loss": 3.794791779225018,
      "tokens_seen": 672727040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005372304709665216,
      "loss": 2.7727,
      "theoretical_loss": 3.7947539783531594,
      "tokens_seen": 672792576
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005372139209381501,
      "loss": 2.6522,
      "theoretical_loss": 3.7947161821941497,
      "tokens_seen": 672858112
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005371973709097787,
      "loss": 2.7348,
      "theoretical_loss": 3.794678390746941,
      "tokens_seen": 672923648
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005371808208814072,
      "loss": 2.9346,
      "theoretical_loss": 3.794640604010488,
      "tokens_seen": 672989184
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005371642708530357,
      "loss": 2.735,
      "theoretical_loss": 3.7946028219837453,
      "tokens_seen": 673054720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005371477208246643,
      "loss": 2.6726,
      "theoretical_loss": 3.7945650446656662,
      "tokens_seen": 673120256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005371311707962927,
      "loss": 2.7868,
      "theoretical_loss": 3.794527272055206,
      "tokens_seen": 673185792
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005371146207679213,
      "loss": 2.8585,
      "theoretical_loss": 3.79448950415132,
      "tokens_seen": 673251328
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005370980707395498,
      "loss": 2.7357,
      "theoretical_loss": 3.7944517409529634,
      "tokens_seen": 673316864
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005370815207111783,
      "loss": 2.8331,
      "theoretical_loss": 3.7944139824590923,
      "tokens_seen": 673382400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005370649706828069,
      "loss": 2.8886,
      "theoretical_loss": 3.794376228668663,
      "tokens_seen": 673447936
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005370484206544354,
      "loss": 2.6843,
      "theoretical_loss": 3.794338479580632,
      "tokens_seen": 673513472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005370318706260639,
      "loss": 2.8352,
      "theoretical_loss": 3.7943007351939553,
      "tokens_seen": 673579008
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005370153205976924,
      "loss": 2.7318,
      "theoretical_loss": 3.794262995507591,
      "tokens_seen": 673644544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005369987705693209,
      "loss": 3.0159,
      "theoretical_loss": 3.794225260520496,
      "tokens_seen": 673710080
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005369822205409495,
      "loss": 2.8028,
      "theoretical_loss": 3.794187530231628,
      "tokens_seen": 673775616
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000536965670512578,
      "loss": 2.57,
      "theoretical_loss": 3.7941498046399458,
      "tokens_seen": 673841152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005369491204842065,
      "loss": 2.6526,
      "theoretical_loss": 3.794112083744407,
      "tokens_seen": 673906688
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005369325704558351,
      "loss": 2.6277,
      "theoretical_loss": 3.7940743675439714,
      "tokens_seen": 673972224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005369160204274636,
      "loss": 2.669,
      "theoretical_loss": 3.7940366560375978,
      "tokens_seen": 674037760
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005368994703990922,
      "loss": 2.963,
      "theoretical_loss": 3.793998949224245,
      "tokens_seen": 674103296
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005368829203707206,
      "loss": 2.9313,
      "theoretical_loss": 3.793961247102874,
      "tokens_seen": 674168832
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005368663703423491,
      "loss": 2.6918,
      "theoretical_loss": 3.793923549672444,
      "tokens_seen": 674234368
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005368498203139777,
      "loss": 2.799,
      "theoretical_loss": 3.793885856931916,
      "tokens_seen": 674299904
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005368332702856062,
      "loss": 2.7593,
      "theoretical_loss": 3.7938481688802503,
      "tokens_seen": 674365440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005368167202572347,
      "loss": 2.9122,
      "theoretical_loss": 3.793810485516409,
      "tokens_seen": 674430976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005368001702288633,
      "loss": 2.7556,
      "theoretical_loss": 3.793772806839352,
      "tokens_seen": 674496512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005367836202004918,
      "loss": 2.6828,
      "theoretical_loss": 3.793735132848042,
      "tokens_seen": 674562048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005367670701721203,
      "loss": 2.6711,
      "theoretical_loss": 3.7936974635414416,
      "tokens_seen": 674627584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005367505201437488,
      "loss": 2.6736,
      "theoretical_loss": 3.7936597989185126,
      "tokens_seen": 674693120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005367339701153773,
      "loss": 2.4137,
      "theoretical_loss": 3.793622138978218,
      "tokens_seen": 674758656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005367174200870059,
      "loss": 2.755,
      "theoretical_loss": 3.7935844837195205,
      "tokens_seen": 674824192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005367008700586344,
      "loss": 2.8734,
      "theoretical_loss": 3.7935468331413844,
      "tokens_seen": 674889728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005366843200302629,
      "loss": 2.9308,
      "theoretical_loss": 3.7935091872427726,
      "tokens_seen": 674955264
    },
    {
      "epoch": 0.24,
      "objective/train/avg_token_score": 0.017744513228535652,
      "objective/train/avg_weight": 0.9791145324707031,
      "objective/train/docs_used": 390150,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7876572608947754,
      "objective/train/std_weight": 0.0829954445362091,
      "objective/train/theoretical_loss": 3.79347154602265,
      "objective/train/tokens_used": 695480800,
      "theoretical_loss": 3.79347154602265,
      "tokens_seen": 675020800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005366677700018915,
      "loss": 2.8342,
      "theoretical_loss": 3.79347154602265,
      "tokens_seen": 675020800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005366512199735199,
      "loss": 2.8355,
      "theoretical_loss": 3.7934339094799805,
      "tokens_seen": 675086336
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005366346699451485,
      "loss": 2.875,
      "theoretical_loss": 3.7933962776137284,
      "tokens_seen": 675151872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000536618119916777,
      "loss": 2.9378,
      "theoretical_loss": 3.79335865042286,
      "tokens_seen": 675217408
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005366015698884055,
      "loss": 2.9423,
      "theoretical_loss": 3.79332102790634,
      "tokens_seen": 675282944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005365850198600341,
      "loss": 2.7875,
      "theoretical_loss": 3.7932834100631343,
      "tokens_seen": 675348480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005365684698316626,
      "loss": 2.8378,
      "theoretical_loss": 3.793245796892209,
      "tokens_seen": 675414016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005365519198032911,
      "loss": 2.757,
      "theoretical_loss": 3.7932081883925304,
      "tokens_seen": 675479552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005365353697749197,
      "loss": 2.878,
      "theoretical_loss": 3.7931705845630654,
      "tokens_seen": 675545088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005365188197465481,
      "loss": 2.8398,
      "theoretical_loss": 3.7931329854027807,
      "tokens_seen": 675610624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005365022697181767,
      "loss": 3.1535,
      "theoretical_loss": 3.7930953909106444,
      "tokens_seen": 675676160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005364857196898052,
      "loss": 2.7605,
      "theoretical_loss": 3.7930578010856233,
      "tokens_seen": 675741696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005364691696614337,
      "loss": 2.8743,
      "theoretical_loss": 3.793020215926686,
      "tokens_seen": 675807232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005364526196330623,
      "loss": 2.9007,
      "theoretical_loss": 3.7929826354328013,
      "tokens_seen": 675872768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005364360696046908,
      "loss": 2.8556,
      "theoretical_loss": 3.7929450596029373,
      "tokens_seen": 675938304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005364195195763193,
      "loss": 2.882,
      "theoretical_loss": 3.792907488436063,
      "tokens_seen": 676003840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005364029695479478,
      "loss": 2.7882,
      "theoretical_loss": 3.7928699219311484,
      "tokens_seen": 676069376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005363864195195763,
      "loss": 2.7893,
      "theoretical_loss": 3.7928323600871625,
      "tokens_seen": 676134912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005363698694912049,
      "loss": 2.7434,
      "theoretical_loss": 3.7927948029030754,
      "tokens_seen": 676200448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005363533194628334,
      "loss": 2.9193,
      "theoretical_loss": 3.792757250377858,
      "tokens_seen": 676265984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005363367694344619,
      "loss": 2.7089,
      "theoretical_loss": 3.792719702510481,
      "tokens_seen": 676331520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005363202194060905,
      "loss": 2.8046,
      "theoretical_loss": 3.7926821592999147,
      "tokens_seen": 676397056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005363036693777189,
      "loss": 2.7737,
      "theoretical_loss": 3.79264462074513,
      "tokens_seen": 676462592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005362871193493474,
      "loss": 2.9256,
      "theoretical_loss": 3.7926070868451003,
      "tokens_seen": 676528128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005362705693209759,
      "loss": 2.7067,
      "theoretical_loss": 3.7925695575987963,
      "tokens_seen": 676593664
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005362540192926044,
      "loss": 2.8355,
      "theoretical_loss": 3.792532033005191,
      "tokens_seen": 676659200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000536237469264233,
      "loss": 2.8771,
      "theoretical_loss": 3.7924945130632564,
      "tokens_seen": 676724736
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005362209192358615,
      "loss": 2.74,
      "theoretical_loss": 3.7924569977719655,
      "tokens_seen": 676790272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005362043692074901,
      "loss": 2.7871,
      "theoretical_loss": 3.792419487130292,
      "tokens_seen": 676855808
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005361878191791186,
      "loss": 2.8241,
      "theoretical_loss": 3.792381981137209,
      "tokens_seen": 676921344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000536171269150747,
      "loss": 2.6271,
      "theoretical_loss": 3.7923444797916916,
      "tokens_seen": 676986880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005361547191223756,
      "loss": 2.6909,
      "theoretical_loss": 3.7923069830927125,
      "tokens_seen": 677052416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005361381690940041,
      "loss": 2.7618,
      "theoretical_loss": 3.7922694910392476,
      "tokens_seen": 677117952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005361216190656326,
      "loss": 2.9375,
      "theoretical_loss": 3.792232003630271,
      "tokens_seen": 677183488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005361050690372612,
      "loss": 2.6963,
      "theoretical_loss": 3.7921945208647583,
      "tokens_seen": 677249024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005360885190088897,
      "loss": 2.6259,
      "theoretical_loss": 3.792157042741685,
      "tokens_seen": 677314560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005360719689805183,
      "loss": 2.713,
      "theoretical_loss": 3.792119569260027,
      "tokens_seen": 677380096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005360554189521468,
      "loss": 2.7598,
      "theoretical_loss": 3.79208210041876,
      "tokens_seen": 677445632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005360388689237752,
      "loss": 2.7655,
      "theoretical_loss": 3.792044636216862,
      "tokens_seen": 677511168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005360223188954038,
      "loss": 2.7945,
      "theoretical_loss": 3.7920071766533088,
      "tokens_seen": 677576704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005360057688670323,
      "loss": 3.0165,
      "theoretical_loss": 3.791969721727077,
      "tokens_seen": 677642240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005359892188386608,
      "loss": 2.6904,
      "theoretical_loss": 3.7919322714371457,
      "tokens_seen": 677707776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005359726688102894,
      "loss": 2.8179,
      "theoretical_loss": 3.7918948257824914,
      "tokens_seen": 677773312
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005359561187819179,
      "loss": 2.8047,
      "theoretical_loss": 3.7918573847620927,
      "tokens_seen": 677838848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005359395687535465,
      "loss": 2.7948,
      "theoretical_loss": 3.791819948374928,
      "tokens_seen": 677904384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000535923018725175,
      "loss": 2.8252,
      "theoretical_loss": 3.791782516619977,
      "tokens_seen": 677969920
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005359064686968034,
      "loss": 2.7724,
      "theoretical_loss": 3.7917450894962177,
      "tokens_seen": 678035456
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000535889918668432,
      "loss": 2.8732,
      "theoretical_loss": 3.7917076670026297,
      "tokens_seen": 678100992
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005358733686400605,
      "loss": 2.7299,
      "theoretical_loss": 3.7916702491381935,
      "tokens_seen": 678166528
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000535856818611689,
      "loss": 2.9044,
      "theoretical_loss": 3.791632835901889,
      "tokens_seen": 678232064
    },
    {
      "epoch": 0.24,
      "objective/train/avg_token_score": 0.03200947865843773,
      "objective/train/avg_weight": 0.9623220562934875,
      "objective/train/docs_used": 392039,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7261881828308105,
      "objective/train/std_weight": 0.1342126727104187,
      "objective/train/theoretical_loss": 3.7915954272926955,
      "objective/train/tokens_used": 698757600,
      "theoretical_loss": 3.7915954272926955,
      "tokens_seen": 678297600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005358402685833176,
      "loss": 2.7925,
      "theoretical_loss": 3.7915954272926955,
      "tokens_seen": 678297600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005358237185549461,
      "loss": 2.8453,
      "theoretical_loss": 3.7915580233095953,
      "tokens_seen": 678363136
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005358071685265747,
      "loss": 2.8018,
      "theoretical_loss": 3.791520623951569,
      "tokens_seen": 678428672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005357906184982031,
      "loss": 2.9118,
      "theoretical_loss": 3.7914832292175973,
      "tokens_seen": 678494208
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005357740684698316,
      "loss": 2.8032,
      "theoretical_loss": 3.7914458391066628,
      "tokens_seen": 678559744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005357575184414602,
      "loss": 2.5951,
      "theoretical_loss": 3.7914084536177475,
      "tokens_seen": 678625280
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005357409684130887,
      "loss": 3.1263,
      "theoretical_loss": 3.7913710727498335,
      "tokens_seen": 678690816
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005357244183847172,
      "loss": 2.8368,
      "theoretical_loss": 3.791333696501903,
      "tokens_seen": 678756352
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005357078683563458,
      "loss": 2.7244,
      "theoretical_loss": 3.7912963248729397,
      "tokens_seen": 678821888
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005356913183279743,
      "loss": 2.7438,
      "theoretical_loss": 3.7912589578619267,
      "tokens_seen": 678887424
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005356747682996028,
      "loss": 2.7542,
      "theoretical_loss": 3.7912215954678476,
      "tokens_seen": 678952960
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005356582182712313,
      "loss": 2.4245,
      "theoretical_loss": 3.791184237689687,
      "tokens_seen": 679018496
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005356416682428598,
      "loss": 2.8404,
      "theoretical_loss": 3.791146884526428,
      "tokens_seen": 679084032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005356251182144884,
      "loss": 2.7148,
      "theoretical_loss": 3.791109535977056,
      "tokens_seen": 679149568
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005356085681861169,
      "loss": 2.6532,
      "theoretical_loss": 3.7910721920405566,
      "tokens_seen": 679215104
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005355920181577454,
      "loss": 2.8581,
      "theoretical_loss": 3.7910348527159137,
      "tokens_seen": 679280640
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000535575468129374,
      "loss": 2.8283,
      "theoretical_loss": 3.790997518002113,
      "tokens_seen": 679346176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005355589181010024,
      "loss": 2.742,
      "theoretical_loss": 3.790960187898142,
      "tokens_seen": 679411712
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000535542368072631,
      "loss": 2.8765,
      "theoretical_loss": 3.7909228624029856,
      "tokens_seen": 679477248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005355258180442595,
      "loss": 3.0194,
      "theoretical_loss": 3.79088554151563,
      "tokens_seen": 679542784
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000535509268015888,
      "loss": 2.6594,
      "theoretical_loss": 3.790848225235063,
      "tokens_seen": 679608320
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005354927179875166,
      "loss": 2.705,
      "theoretical_loss": 3.7908109135602714,
      "tokens_seen": 679673856
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005354761679591451,
      "loss": 2.8105,
      "theoretical_loss": 3.790773606490243,
      "tokens_seen": 679739392
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005354596179307737,
      "loss": 3.0158,
      "theoretical_loss": 3.7907363040239654,
      "tokens_seen": 679804928
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005354430679024022,
      "loss": 2.8304,
      "theoretical_loss": 3.790699006160427,
      "tokens_seen": 679870464
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005354265178740306,
      "loss": 2.634,
      "theoretical_loss": 3.7906617128986153,
      "tokens_seen": 679936000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005354099678456592,
      "loss": 2.6121,
      "theoretical_loss": 3.79062442423752,
      "tokens_seen": 680001536
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005353934178172877,
      "loss": 2.4107,
      "theoretical_loss": 3.7905871401761306,
      "tokens_seen": 680067072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005353768677889162,
      "loss": 2.6814,
      "theoretical_loss": 3.790549860713435,
      "tokens_seen": 680132608
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005353603177605448,
      "loss": 2.7147,
      "theoretical_loss": 3.7905125858484245,
      "tokens_seen": 680198144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005353437677321733,
      "loss": 2.7476,
      "theoretical_loss": 3.7904753155800885,
      "tokens_seen": 680263680
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005353272177038019,
      "loss": 2.8544,
      "theoretical_loss": 3.7904380499074177,
      "tokens_seen": 680329216
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005353106676754302,
      "loss": 2.527,
      "theoretical_loss": 3.7904007888294027,
      "tokens_seen": 680394752
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005352941176470588,
      "loss": 2.7657,
      "theoretical_loss": 3.790363532345034,
      "tokens_seen": 680460288
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005352775676186873,
      "loss": 2.7302,
      "theoretical_loss": 3.790326280453303,
      "tokens_seen": 680525824
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005352610175903158,
      "loss": 2.6133,
      "theoretical_loss": 3.790289033153202,
      "tokens_seen": 680591360
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005352444675619444,
      "loss": 2.8383,
      "theoretical_loss": 3.7902517904437225,
      "tokens_seen": 680656896
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005352279175335729,
      "loss": 2.7487,
      "theoretical_loss": 3.7902145523238575,
      "tokens_seen": 680722432
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005352113675052014,
      "loss": 2.9102,
      "theoretical_loss": 3.790177318792599,
      "tokens_seen": 680787968
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00053519481747683,
      "loss": 2.2762,
      "theoretical_loss": 3.79014008984894,
      "tokens_seen": 680853504
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005351782674484584,
      "loss": 2.8537,
      "theoretical_loss": 3.790102865491874,
      "tokens_seen": 680919040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000535161717420087,
      "loss": 2.7988,
      "theoretical_loss": 3.7900656457203934,
      "tokens_seen": 680984576
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005351451673917155,
      "loss": 2.95,
      "theoretical_loss": 3.7900284305334937,
      "tokens_seen": 681050112
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000535128617363344,
      "loss": 2.8673,
      "theoretical_loss": 3.7899912199301684,
      "tokens_seen": 681115648
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005351120673349726,
      "loss": 2.6277,
      "theoretical_loss": 3.789954013909412,
      "tokens_seen": 681181184
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005350955173066011,
      "loss": 2.7589,
      "theoretical_loss": 3.78991681247022,
      "tokens_seen": 681246720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005350789672782295,
      "loss": 2.8374,
      "theoretical_loss": 3.7898796156115866,
      "tokens_seen": 681312256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005350624172498581,
      "loss": 2.9235,
      "theoretical_loss": 3.7898424233325074,
      "tokens_seen": 681377792
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005350458672214866,
      "loss": 2.8597,
      "theoretical_loss": 3.7898052356319787,
      "tokens_seen": 681443328
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005350293171931152,
      "loss": 2.8901,
      "theoretical_loss": 3.7897680525089967,
      "tokens_seen": 681508864
    },
    {
      "epoch": 0.24,
      "objective/train/avg_token_score": 0.010815569199621677,
      "objective/train/avg_weight": 0.9872714877128601,
      "objective/train/docs_used": 393961,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.531698703765869,
      "objective/train/std_weight": 0.0735975131392479,
      "objective/train/theoretical_loss": 3.789730873962557,
      "objective/train/tokens_used": 702034400,
      "theoretical_loss": 3.789730873962557,
      "tokens_seen": 681574400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005350127671647437,
      "loss": 2.9377,
      "theoretical_loss": 3.789730873962557,
      "tokens_seen": 681574400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005349962171363722,
      "loss": 2.686,
      "theoretical_loss": 3.789693699991657,
      "tokens_seen": 681639936
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005349796671080008,
      "loss": 2.6616,
      "theoretical_loss": 3.7896565305952934,
      "tokens_seen": 681705472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005349631170796293,
      "loss": 2.6324,
      "theoretical_loss": 3.789619365772464,
      "tokens_seen": 681771008
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005349465670512577,
      "loss": 2.7563,
      "theoretical_loss": 3.7895822055221657,
      "tokens_seen": 681836544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005349300170228863,
      "loss": 2.462,
      "theoretical_loss": 3.7895450498433982,
      "tokens_seen": 681902080
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005349134669945148,
      "loss": 2.6631,
      "theoretical_loss": 3.789507898735158,
      "tokens_seen": 681967616
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005348969169661434,
      "loss": 2.6819,
      "theoretical_loss": 3.7894707521964444,
      "tokens_seen": 682033152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005348803669377719,
      "loss": 2.5972,
      "theoretical_loss": 3.789433610226256,
      "tokens_seen": 682098688
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005348638169094004,
      "loss": 2.5314,
      "theoretical_loss": 3.789396472823593,
      "tokens_seen": 682164224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000534847266881029,
      "loss": 2.6966,
      "theoretical_loss": 3.7893593399874548,
      "tokens_seen": 682229760
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005348307168526574,
      "loss": 2.6969,
      "theoretical_loss": 3.78932221171684,
      "tokens_seen": 682295296
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005348141668242859,
      "loss": 2.8155,
      "theoretical_loss": 3.7892850880107503,
      "tokens_seen": 682360832
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005347976167959145,
      "loss": 2.756,
      "theoretical_loss": 3.789247968868186,
      "tokens_seen": 682426368
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000534781066767543,
      "loss": 2.8281,
      "theoretical_loss": 3.7892108542881475,
      "tokens_seen": 682491904
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005347645167391716,
      "loss": 2.8795,
      "theoretical_loss": 3.7891737442696356,
      "tokens_seen": 682557440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005347479667108001,
      "loss": 2.8077,
      "theoretical_loss": 3.7891366388116525,
      "tokens_seen": 682622976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005347314166824286,
      "loss": 2.8233,
      "theoretical_loss": 3.7890995379131995,
      "tokens_seen": 682688512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005347148666540572,
      "loss": 2.5065,
      "theoretical_loss": 3.7890624415732796,
      "tokens_seen": 682754048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005346983166256856,
      "loss": 2.8805,
      "theoretical_loss": 3.789025349790894,
      "tokens_seen": 682819584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005346817665973141,
      "loss": 2.8042,
      "theoretical_loss": 3.7889882625650464,
      "tokens_seen": 682885120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005346652165689427,
      "loss": 2.6765,
      "theoretical_loss": 3.788951179894739,
      "tokens_seen": 682950656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005346486665405712,
      "loss": 2.806,
      "theoretical_loss": 3.788914101778976,
      "tokens_seen": 683016192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005346321165121998,
      "loss": 2.8222,
      "theoretical_loss": 3.7888770282167608,
      "tokens_seen": 683081728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005346155664838283,
      "loss": 2.5682,
      "theoretical_loss": 3.788839959207097,
      "tokens_seen": 683147264
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005345990164554568,
      "loss": 2.7436,
      "theoretical_loss": 3.788802894748989,
      "tokens_seen": 683212800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005345824664270853,
      "loss": 2.6953,
      "theoretical_loss": 3.788765834841442,
      "tokens_seen": 683278336
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005345659163987138,
      "loss": 2.5996,
      "theoretical_loss": 3.78872877948346,
      "tokens_seen": 683343872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005345493663703423,
      "loss": 2.7544,
      "theoretical_loss": 3.788691728674049,
      "tokens_seen": 683409408
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005345328163419709,
      "loss": 2.8737,
      "theoretical_loss": 3.788654682412215,
      "tokens_seen": 683474944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005345162663135994,
      "loss": 2.536,
      "theoretical_loss": 3.7886176406969625,
      "tokens_seen": 683540480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000534499716285228,
      "loss": 2.4941,
      "theoretical_loss": 3.788580603527298,
      "tokens_seen": 683606016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005344831662568565,
      "loss": 2.563,
      "theoretical_loss": 3.7885435709022284,
      "tokens_seen": 683671552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000534466616228485,
      "loss": 2.8035,
      "theoretical_loss": 3.788506542820761,
      "tokens_seen": 683737088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005344500662001135,
      "loss": 2.8471,
      "theoretical_loss": 3.788469519281902,
      "tokens_seen": 683802624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000534433516171742,
      "loss": 2.704,
      "theoretical_loss": 3.7884325002846593,
      "tokens_seen": 683868160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005344169661433705,
      "loss": 2.9298,
      "theoretical_loss": 3.7883954858280404,
      "tokens_seen": 683933696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005344004161149991,
      "loss": 2.806,
      "theoretical_loss": 3.788358475911054,
      "tokens_seen": 683999232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005343838660866276,
      "loss": 2.5946,
      "theoretical_loss": 3.788321470532707,
      "tokens_seen": 684064768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005343673160582562,
      "loss": 2.7748,
      "theoretical_loss": 3.7882844696920097,
      "tokens_seen": 684130304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005343507660298847,
      "loss": 2.5034,
      "theoretical_loss": 3.78824747338797,
      "tokens_seen": 684195840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005343342160015131,
      "loss": 2.8212,
      "theoretical_loss": 3.788210481619598,
      "tokens_seen": 684261376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005343176659731416,
      "loss": 2.617,
      "theoretical_loss": 3.7881734943859025,
      "tokens_seen": 684326912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005343011159447701,
      "loss": 3.0091,
      "theoretical_loss": 3.788136511685894,
      "tokens_seen": 684392448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005342845659163987,
      "loss": 2.826,
      "theoretical_loss": 3.788099533518582,
      "tokens_seen": 684457984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005342680158880272,
      "loss": 2.5999,
      "theoretical_loss": 3.7880625598829782,
      "tokens_seen": 684523520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005342514658596558,
      "loss": 2.7523,
      "theoretical_loss": 3.7880255907780924,
      "tokens_seen": 684589056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005342349158312843,
      "loss": 2.8548,
      "theoretical_loss": 3.787988626202937,
      "tokens_seen": 684654592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005342183658029127,
      "loss": 2.6805,
      "theoretical_loss": 3.7879516661565216,
      "tokens_seen": 684720128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005342018157745413,
      "loss": 2.6334,
      "theoretical_loss": 3.7879147106378595,
      "tokens_seen": 684785664
    },
    {
      "epoch": 0.24,
      "objective/train/avg_token_score": 0.026597652584314346,
      "objective/train/avg_weight": 0.968704104423523,
      "objective/train/docs_used": 395905,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6906678676605225,
      "objective/train/std_weight": 0.10204198211431503,
      "objective/train/theoretical_loss": 3.787877759645963,
      "objective/train/tokens_used": 705311200,
      "theoretical_loss": 3.787877759645963,
      "tokens_seen": 684851200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005341852657461698,
      "loss": 2.6093,
      "theoretical_loss": 3.787877759645963,
      "tokens_seen": 684851200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005341687157177983,
      "loss": 2.7828,
      "theoretical_loss": 3.7878408131798427,
      "tokens_seen": 684916736
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005341521656894269,
      "loss": 2.7326,
      "theoretical_loss": 3.787803871238513,
      "tokens_seen": 684982272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005341356156610554,
      "loss": 2.8481,
      "theoretical_loss": 3.7877669338209863,
      "tokens_seen": 685047808
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000534119065632684,
      "loss": 2.7151,
      "theoretical_loss": 3.787730000926276,
      "tokens_seen": 685113344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005341025156043124,
      "loss": 2.8128,
      "theoretical_loss": 3.787693072553396,
      "tokens_seen": 685178880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005340859655759409,
      "loss": 2.731,
      "theoretical_loss": 3.78765614870136,
      "tokens_seen": 685244416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005340694155475695,
      "loss": 2.7673,
      "theoretical_loss": 3.7876192293691817,
      "tokens_seen": 685309952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000534052865519198,
      "loss": 2.4536,
      "theoretical_loss": 3.787582314555877,
      "tokens_seen": 685375488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005340363154908265,
      "loss": 2.8183,
      "theoretical_loss": 3.7875454042604595,
      "tokens_seen": 685441024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005340197654624551,
      "loss": 2.5553,
      "theoretical_loss": 3.787508498481945,
      "tokens_seen": 685506560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005340032154340836,
      "loss": 2.9527,
      "theoretical_loss": 3.7874715972193487,
      "tokens_seen": 685572096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005339866654057122,
      "loss": 2.933,
      "theoretical_loss": 3.7874347004716866,
      "tokens_seen": 685637632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005339701153773406,
      "loss": 2.7259,
      "theoretical_loss": 3.787397808237975,
      "tokens_seen": 685703168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005339535653489691,
      "loss": 2.7707,
      "theoretical_loss": 3.78736092051723,
      "tokens_seen": 685768704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005339370153205977,
      "loss": 2.9135,
      "theoretical_loss": 3.7873240373084687,
      "tokens_seen": 685834240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005339204652922262,
      "loss": 2.5685,
      "theoretical_loss": 3.7872871586107077,
      "tokens_seen": 685899776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0005339039152638547,
      "loss": 2.7033,
      "theoretical_loss": 3.7872502844229645,
      "tokens_seen": 685965312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005338873652354833,
      "loss": 2.9096,
      "theoretical_loss": 3.787213414744257,
      "tokens_seen": 686030848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005338708152071118,
      "loss": 2.7261,
      "theoretical_loss": 3.787176549573603,
      "tokens_seen": 686096384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005338542651787404,
      "loss": 2.8748,
      "theoretical_loss": 3.7871396889100204,
      "tokens_seen": 686161920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005338377151503688,
      "loss": 2.621,
      "theoretical_loss": 3.7871028327525287,
      "tokens_seen": 686227456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005338211651219973,
      "loss": 2.9106,
      "theoretical_loss": 3.7870659811001457,
      "tokens_seen": 686292992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005338046150936259,
      "loss": 2.8767,
      "theoretical_loss": 3.787029133951891,
      "tokens_seen": 686358528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005337880650652544,
      "loss": 2.8239,
      "theoretical_loss": 3.7869922913067846,
      "tokens_seen": 686424064
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005337715150368829,
      "loss": 2.7938,
      "theoretical_loss": 3.786955453163846,
      "tokens_seen": 686489600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005337549650085115,
      "loss": 2.7169,
      "theoretical_loss": 3.7869186195220945,
      "tokens_seen": 686555136
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00053373841498014,
      "loss": 2.4926,
      "theoretical_loss": 3.786881790380552,
      "tokens_seen": 686620672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005337218649517685,
      "loss": 2.6892,
      "theoretical_loss": 3.7868449657382377,
      "tokens_seen": 686686208
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000533705314923397,
      "loss": 2.722,
      "theoretical_loss": 3.786808145594174,
      "tokens_seen": 686751744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005336887648950255,
      "loss": 2.6069,
      "theoretical_loss": 3.7867713299473813,
      "tokens_seen": 686817280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005336722148666541,
      "loss": 2.8018,
      "theoretical_loss": 3.786734518796882,
      "tokens_seen": 686882816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005336556648382826,
      "loss": 2.8514,
      "theoretical_loss": 3.786697712141697,
      "tokens_seen": 686948352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005336391148099111,
      "loss": 2.523,
      "theoretical_loss": 3.7866609099808493,
      "tokens_seen": 687013888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005336225647815397,
      "loss": 2.7794,
      "theoretical_loss": 3.7866241123133615,
      "tokens_seen": 687079424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005336060147531681,
      "loss": 2.8276,
      "theoretical_loss": 3.786587319138256,
      "tokens_seen": 687144960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005335894647247967,
      "loss": 2.5609,
      "theoretical_loss": 3.786550530454557,
      "tokens_seen": 687210496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005335729146964252,
      "loss": 2.6735,
      "theoretical_loss": 3.786513746261287,
      "tokens_seen": 687276032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005335563646680537,
      "loss": 2.8517,
      "theoretical_loss": 3.78647696655747,
      "tokens_seen": 687341568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005335398146396823,
      "loss": 2.7815,
      "theoretical_loss": 3.7864401913421295,
      "tokens_seen": 687407104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005335232646113108,
      "loss": 2.9601,
      "theoretical_loss": 3.7864034206142914,
      "tokens_seen": 687472640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005335067145829393,
      "loss": 2.8223,
      "theoretical_loss": 3.786366654372979,
      "tokens_seen": 687538176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005334901645545679,
      "loss": 2.3908,
      "theoretical_loss": 3.7863298926172186,
      "tokens_seen": 687603712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005334736145261963,
      "loss": 2.6163,
      "theoretical_loss": 3.786293135346034,
      "tokens_seen": 687669248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005334570644978249,
      "loss": 2.8052,
      "theoretical_loss": 3.786256382558452,
      "tokens_seen": 687734784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005334405144694534,
      "loss": 2.7234,
      "theoretical_loss": 3.7862196342534977,
      "tokens_seen": 687800320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005334239644410819,
      "loss": 2.8958,
      "theoretical_loss": 3.786182890430198,
      "tokens_seen": 687865856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005334074144127105,
      "loss": 2.7173,
      "theoretical_loss": 3.7861461510875793,
      "tokens_seen": 687931392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000533390864384339,
      "loss": 2.6722,
      "theoretical_loss": 3.786109416224668,
      "tokens_seen": 687996928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005333743143559675,
      "loss": 2.5914,
      "theoretical_loss": 3.786072685840492,
      "tokens_seen": 688062464
    },
    {
      "epoch": 0.25,
      "objective/train/avg_token_score": 0.033809758722782135,
      "objective/train/avg_weight": 0.9602457284927368,
      "objective/train/docs_used": 397858,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.849191665649414,
      "objective/train/std_weight": 0.1457771211862564,
      "objective/train/theoretical_loss": 3.7860359599340776,
      "objective/train/tokens_used": 708588000,
      "theoretical_loss": 3.7860359599340776,
      "tokens_seen": 688128000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000533357764327596,
      "loss": 2.66,
      "theoretical_loss": 3.7860359599340776,
      "tokens_seen": 688128000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005333412142992244,
      "loss": 3.0678,
      "theoretical_loss": 3.7859992385044534,
      "tokens_seen": 688193536
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000533324664270853,
      "loss": 2.7956,
      "theoretical_loss": 3.7859625215506476,
      "tokens_seen": 688259072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005333081142424815,
      "loss": 3.1515,
      "theoretical_loss": 3.7859258090716876,
      "tokens_seen": 688324608
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005332915642141101,
      "loss": 2.7408,
      "theoretical_loss": 3.785889101066603,
      "tokens_seen": 688390144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005332750141857386,
      "loss": 2.9237,
      "theoretical_loss": 3.785852397534423,
      "tokens_seen": 688455680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000533258464157367,
      "loss": 2.8102,
      "theoretical_loss": 3.7858156984741758,
      "tokens_seen": 688521216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005332419141289956,
      "loss": 2.7627,
      "theoretical_loss": 3.7857790038848913,
      "tokens_seen": 688586752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005332253641006241,
      "loss": 2.748,
      "theoretical_loss": 3.7857423137655997,
      "tokens_seen": 688652288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005332088140722526,
      "loss": 2.6541,
      "theoretical_loss": 3.785705628115331,
      "tokens_seen": 688717824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005331922640438812,
      "loss": 2.552,
      "theoretical_loss": 3.7856689469331153,
      "tokens_seen": 688783360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005331757140155097,
      "loss": 2.7287,
      "theoretical_loss": 3.785632270217984,
      "tokens_seen": 688848896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005331591639871383,
      "loss": 2.7735,
      "theoretical_loss": 3.7855955979689684,
      "tokens_seen": 688914432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005331426139587668,
      "loss": 2.8464,
      "theoretical_loss": 3.7855589301850987,
      "tokens_seen": 688979968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005331260639303952,
      "loss": 2.9733,
      "theoretical_loss": 3.7855222668654074,
      "tokens_seen": 689045504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005331095139020238,
      "loss": 2.6789,
      "theoretical_loss": 3.7854856080089268,
      "tokens_seen": 689111040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005330929638736523,
      "loss": 2.8499,
      "theoretical_loss": 3.7854489536146883,
      "tokens_seen": 689176576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005330764138452808,
      "loss": 2.9189,
      "theoretical_loss": 3.785412303681725,
      "tokens_seen": 689242112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005330598638169094,
      "loss": 2.8201,
      "theoretical_loss": 3.7853756582090705,
      "tokens_seen": 689307648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005330433137885379,
      "loss": 2.5837,
      "theoretical_loss": 3.7853390171957564,
      "tokens_seen": 689373184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005330267637601665,
      "loss": 2.9248,
      "theoretical_loss": 3.7853023806408173,
      "tokens_seen": 689438720
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000533010213731795,
      "loss": 2.7699,
      "theoretical_loss": 3.785265748543287,
      "tokens_seen": 689504256
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005329936637034234,
      "loss": 2.5333,
      "theoretical_loss": 3.7852291209021987,
      "tokens_seen": 689569792
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000532977113675052,
      "loss": 2.6746,
      "theoretical_loss": 3.785192497716588,
      "tokens_seen": 689635328
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005329605636466805,
      "loss": 2.528,
      "theoretical_loss": 3.7851558789854884,
      "tokens_seen": 689700864
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005329440136183091,
      "loss": 2.572,
      "theoretical_loss": 3.785119264707936,
      "tokens_seen": 689766400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005329274635899376,
      "loss": 2.9888,
      "theoretical_loss": 3.7850826548829652,
      "tokens_seen": 689831936
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005329109135615661,
      "loss": 2.5766,
      "theoretical_loss": 3.7850460495096123,
      "tokens_seen": 689897472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005328943635331947,
      "loss": 2.6542,
      "theoretical_loss": 3.785009448586913,
      "tokens_seen": 689963008
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005328778135048231,
      "loss": 2.7533,
      "theoretical_loss": 3.784972852113903,
      "tokens_seen": 690028544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005328612634764516,
      "loss": 2.7812,
      "theoretical_loss": 3.78493626008962,
      "tokens_seen": 690094080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005328447134480802,
      "loss": 2.7124,
      "theoretical_loss": 3.7848996725130992,
      "tokens_seen": 690159616
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005328281634197087,
      "loss": 2.6421,
      "theoretical_loss": 3.7848630893833786,
      "tokens_seen": 690225152
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005328116133913373,
      "loss": 2.6816,
      "theoretical_loss": 3.784826510699496,
      "tokens_seen": 690290688
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005327950633629658,
      "loss": 2.6675,
      "theoretical_loss": 3.7847899364604887,
      "tokens_seen": 690356224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005327785133345943,
      "loss": 2.9107,
      "theoretical_loss": 3.7847533666653943,
      "tokens_seen": 690421760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005327619633062229,
      "loss": 2.8578,
      "theoretical_loss": 3.7847168013132513,
      "tokens_seen": 690487296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005327454132778513,
      "loss": 2.6667,
      "theoretical_loss": 3.7846802404030986,
      "tokens_seen": 690552832
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005327288632494798,
      "loss": 2.7725,
      "theoretical_loss": 3.7846436839339743,
      "tokens_seen": 690618368
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005327123132211084,
      "loss": 2.859,
      "theoretical_loss": 3.7846071319049193,
      "tokens_seen": 690683904
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005326957631927369,
      "loss": 2.825,
      "theoretical_loss": 3.7845705843149715,
      "tokens_seen": 690749440
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005326792131643655,
      "loss": 2.9283,
      "theoretical_loss": 3.7845340411631714,
      "tokens_seen": 690814976
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000532662663135994,
      "loss": 2.6569,
      "theoretical_loss": 3.7844975024485583,
      "tokens_seen": 690880512
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005326461131076225,
      "loss": 2.998,
      "theoretical_loss": 3.7844609681701735,
      "tokens_seen": 690946048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000532629563079251,
      "loss": 2.8847,
      "theoretical_loss": 3.7844244383270578,
      "tokens_seen": 691011584
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005326130130508795,
      "loss": 2.8062,
      "theoretical_loss": 3.7843879129182514,
      "tokens_seen": 691077120
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000532596463022508,
      "loss": 2.9409,
      "theoretical_loss": 3.7843513919427965,
      "tokens_seen": 691142656
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005325799129941366,
      "loss": 2.758,
      "theoretical_loss": 3.7843148753997333,
      "tokens_seen": 691208192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005325633629657651,
      "loss": 2.8815,
      "theoretical_loss": 3.784278363288105,
      "tokens_seen": 691273728
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005325468129373937,
      "loss": 2.9483,
      "theoretical_loss": 3.7842418556069535,
      "tokens_seen": 691339264
    },
    {
      "epoch": 0.25,
      "objective/train/avg_token_score": 0.032341502606868744,
      "objective/train/avg_weight": 0.9619433283805847,
      "objective/train/docs_used": 398758,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.668466567993164,
      "objective/train/std_weight": 0.12431105226278305,
      "objective/train/theoretical_loss": 3.784205352355321,
      "objective/train/tokens_used": 711864800,
      "theoretical_loss": 3.784205352355321,
      "tokens_seen": 691404800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005325302629090222,
      "loss": 2.5914,
      "theoretical_loss": 3.784205352355321,
      "tokens_seen": 691404800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005325137128806506,
      "loss": 2.9112,
      "theoretical_loss": 3.7841688535322504,
      "tokens_seen": 691470336
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005324971628522792,
      "loss": 2.8166,
      "theoretical_loss": 3.7841323591367844,
      "tokens_seen": 691535872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005324806128239077,
      "loss": 2.9544,
      "theoretical_loss": 3.7840958691679667,
      "tokens_seen": 691601408
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005324640627955362,
      "loss": 2.8537,
      "theoretical_loss": 3.7840593836248413,
      "tokens_seen": 691666944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005324475127671648,
      "loss": 3.0442,
      "theoretical_loss": 3.784022902506452,
      "tokens_seen": 691732480
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005324309627387933,
      "loss": 2.8908,
      "theoretical_loss": 3.7839864258118427,
      "tokens_seen": 691798016
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005324144127104219,
      "loss": 2.8261,
      "theoretical_loss": 3.7839499535400583,
      "tokens_seen": 691863552
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005323978626820504,
      "loss": 2.9487,
      "theoretical_loss": 3.7839134856901437,
      "tokens_seen": 691929088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005323813126536788,
      "loss": 2.8817,
      "theoretical_loss": 3.7838770222611435,
      "tokens_seen": 691994624
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005323647626253073,
      "loss": 2.9274,
      "theoretical_loss": 3.7838405632521037,
      "tokens_seen": 692060160
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005323482125969358,
      "loss": 2.8239,
      "theoretical_loss": 3.7838041086620704,
      "tokens_seen": 692125696
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005323316625685644,
      "loss": 2.883,
      "theoretical_loss": 3.7837676584900883,
      "tokens_seen": 692191232
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005323151125401929,
      "loss": 2.7754,
      "theoretical_loss": 3.7837312127352054,
      "tokens_seen": 692256768
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005322985625118214,
      "loss": 3.0694,
      "theoretical_loss": 3.7836947713964673,
      "tokens_seen": 692322304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00053228201248345,
      "loss": 2.6831,
      "theoretical_loss": 3.783658334472921,
      "tokens_seen": 692387840
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005322654624550784,
      "loss": 2.8742,
      "theoretical_loss": 3.783621901963614,
      "tokens_seen": 692453376
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000532248912426707,
      "loss": 2.7261,
      "theoretical_loss": 3.7835854738675936,
      "tokens_seen": 692518912
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005322323623983355,
      "loss": 2.9879,
      "theoretical_loss": 3.7835490501839075,
      "tokens_seen": 692584448
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000532215812369964,
      "loss": 2.87,
      "theoretical_loss": 3.7835126309116047,
      "tokens_seen": 692649984
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005321992623415926,
      "loss": 2.9237,
      "theoretical_loss": 3.7834762160497326,
      "tokens_seen": 692715520
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005321827123132211,
      "loss": 2.4817,
      "theoretical_loss": 3.7834398055973404,
      "tokens_seen": 692781056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005321661622848496,
      "loss": 2.6831,
      "theoretical_loss": 3.7834033995534764,
      "tokens_seen": 692846592
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005321496122564781,
      "loss": 2.7903,
      "theoretical_loss": 3.7833669979171916,
      "tokens_seen": 692912128
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005321330622281066,
      "loss": 2.757,
      "theoretical_loss": 3.7833306006875334,
      "tokens_seen": 692977664
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005321165121997352,
      "loss": 2.7952,
      "theoretical_loss": 3.7832942078635527,
      "tokens_seen": 693043200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005320999621713637,
      "loss": 2.7467,
      "theoretical_loss": 3.7832578194443007,
      "tokens_seen": 693108736
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005320834121429922,
      "loss": 2.8614,
      "theoretical_loss": 3.7832214354288265,
      "tokens_seen": 693174272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005320668621146208,
      "loss": 2.8823,
      "theoretical_loss": 3.783185055816181,
      "tokens_seen": 693239808
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005320503120862493,
      "loss": 2.8439,
      "theoretical_loss": 3.7831486806054158,
      "tokens_seen": 693305344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005320337620578777,
      "loss": 2.6964,
      "theoretical_loss": 3.783112309795582,
      "tokens_seen": 693370880
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005320172120295063,
      "loss": 2.7865,
      "theoretical_loss": 3.783075943385731,
      "tokens_seen": 693436416
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005320006620011348,
      "loss": 2.68,
      "theoretical_loss": 3.783039581374915,
      "tokens_seen": 693501952
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005319841119727634,
      "loss": 2.8678,
      "theoretical_loss": 3.7830032237621865,
      "tokens_seen": 693567488
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005319675619443919,
      "loss": 2.6032,
      "theoretical_loss": 3.7829668705465975,
      "tokens_seen": 693633024
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005319510119160204,
      "loss": 2.6484,
      "theoretical_loss": 3.7829305217272013,
      "tokens_seen": 693698560
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000531934461887649,
      "loss": 2.7979,
      "theoretical_loss": 3.7828941773030507,
      "tokens_seen": 693764096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005319179118592775,
      "loss": 2.7016,
      "theoretical_loss": 3.782857837273199,
      "tokens_seen": 693829632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005319013618309059,
      "loss": 2.7223,
      "theoretical_loss": 3.782821501636701,
      "tokens_seen": 693895168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005318848118025345,
      "loss": 2.964,
      "theoretical_loss": 3.782785170392609,
      "tokens_seen": 693960704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000531868261774163,
      "loss": 2.8863,
      "theoretical_loss": 3.782748843539978,
      "tokens_seen": 694026240
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005318517117457916,
      "loss": 2.7809,
      "theoretical_loss": 3.7827125210778636,
      "tokens_seen": 694091776
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005318351617174201,
      "loss": 2.9251,
      "theoretical_loss": 3.7826762030053196,
      "tokens_seen": 694157312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005318186116890486,
      "loss": 2.8652,
      "theoretical_loss": 3.7826398893214006,
      "tokens_seen": 694222848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005318020616606772,
      "loss": 2.8958,
      "theoretical_loss": 3.782603580025164,
      "tokens_seen": 694288384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005317855116323056,
      "loss": 2.9096,
      "theoretical_loss": 3.782567275115663,
      "tokens_seen": 694353920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005317689616039341,
      "loss": 2.8767,
      "theoretical_loss": 3.782530974591956,
      "tokens_seen": 694419456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005317524115755627,
      "loss": 2.8972,
      "theoretical_loss": 3.782494678453098,
      "tokens_seen": 694484992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005317358615471912,
      "loss": 2.778,
      "theoretical_loss": 3.7824583866981465,
      "tokens_seen": 694550528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005317193115188198,
      "loss": 2.8456,
      "theoretical_loss": 3.7824220993261575,
      "tokens_seen": 694616064
    },
    {
      "epoch": 0.25,
      "objective/train/avg_token_score": 0.010049034841358662,
      "objective/train/avg_weight": 0.9881696701049805,
      "objective/train/docs_used": 400726,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9497294425964355,
      "objective/train/std_weight": 0.07258860766887665,
      "objective/train/theoretical_loss": 3.782385816336189,
      "objective/train/tokens_used": 715141600,
      "theoretical_loss": 3.782385816336189,
      "tokens_seen": 694681600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005317027614904483,
      "loss": 2.8835,
      "theoretical_loss": 3.782385816336189,
      "tokens_seen": 694681600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005316862114620768,
      "loss": 2.7661,
      "theoretical_loss": 3.7823495377272978,
      "tokens_seen": 694747136
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005316696614337054,
      "loss": 2.5204,
      "theoretical_loss": 3.7823132634985424,
      "tokens_seen": 694812672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005316531114053338,
      "loss": 2.9119,
      "theoretical_loss": 3.7822769936489804,
      "tokens_seen": 694878208
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005316365613769623,
      "loss": 2.5819,
      "theoretical_loss": 3.78224072817767,
      "tokens_seen": 694943744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005316200113485909,
      "loss": 2.6278,
      "theoretical_loss": 3.7822044670836705,
      "tokens_seen": 695009280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005316034613202194,
      "loss": 2.9965,
      "theoretical_loss": 3.782168210366041,
      "tokens_seen": 695074816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000531586911291848,
      "loss": 2.7814,
      "theoretical_loss": 3.78213195802384,
      "tokens_seen": 695140352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005315703612634765,
      "loss": 3.0353,
      "theoretical_loss": 3.7820957100561277,
      "tokens_seen": 695205888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000531553811235105,
      "loss": 2.7452,
      "theoretical_loss": 3.782059466461963,
      "tokens_seen": 695271424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005315372612067335,
      "loss": 2.5655,
      "theoretical_loss": 3.782023227240407,
      "tokens_seen": 695336960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000531520711178362,
      "loss": 2.8615,
      "theoretical_loss": 3.78198699239052,
      "tokens_seen": 695402496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005315041611499905,
      "loss": 2.8088,
      "theoretical_loss": 3.781950761911362,
      "tokens_seen": 695468032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005314876111216191,
      "loss": 2.9137,
      "theoretical_loss": 3.7819145358019948,
      "tokens_seen": 695533568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005314710610932476,
      "loss": 2.7433,
      "theoretical_loss": 3.7818783140614793,
      "tokens_seen": 695599104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005314545110648762,
      "loss": 3.1121,
      "theoretical_loss": 3.7818420966888775,
      "tokens_seen": 695664640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005314379610365047,
      "loss": 2.7824,
      "theoretical_loss": 3.7818058836832504,
      "tokens_seen": 695730176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005314214110081331,
      "loss": 2.9783,
      "theoretical_loss": 3.7817696750436607,
      "tokens_seen": 695795712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005314048609797617,
      "loss": 2.8979,
      "theoretical_loss": 3.781733470769171,
      "tokens_seen": 695861248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005313883109513901,
      "loss": 2.9442,
      "theoretical_loss": 3.7816972708588446,
      "tokens_seen": 695926784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005313717609230187,
      "loss": 2.9323,
      "theoretical_loss": 3.7816610753117432,
      "tokens_seen": 695992320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005313552108946472,
      "loss": 2.5385,
      "theoretical_loss": 3.7816248841269307,
      "tokens_seen": 696057856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005313386608662758,
      "loss": 2.8499,
      "theoretical_loss": 3.7815886973034707,
      "tokens_seen": 696123392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005313221108379043,
      "loss": 2.747,
      "theoretical_loss": 3.781552514840427,
      "tokens_seen": 696188928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005313055608095327,
      "loss": 2.8637,
      "theoretical_loss": 3.781516336736864,
      "tokens_seen": 696254464
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005312890107811613,
      "loss": 2.5193,
      "theoretical_loss": 3.781480162991847,
      "tokens_seen": 696320000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005312724607527898,
      "loss": 2.8031,
      "theoretical_loss": 3.7814439936044386,
      "tokens_seen": 696385536
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005312559107244183,
      "loss": 2.9468,
      "theoretical_loss": 3.781407828573706,
      "tokens_seen": 696451072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005312393606960469,
      "loss": 2.7771,
      "theoretical_loss": 3.7813716678987133,
      "tokens_seen": 696516608
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005312228106676754,
      "loss": 2.6136,
      "theoretical_loss": 3.781335511578526,
      "tokens_seen": 696582144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000531206260639304,
      "loss": 2.8737,
      "theoretical_loss": 3.781299359612211,
      "tokens_seen": 696647680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005311897106109325,
      "loss": 2.7258,
      "theoretical_loss": 3.781263211998834,
      "tokens_seen": 696713216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005311731605825609,
      "loss": 2.5537,
      "theoretical_loss": 3.7812270687374614,
      "tokens_seen": 696778752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005311566105541895,
      "loss": 3.0041,
      "theoretical_loss": 3.7811909298271607,
      "tokens_seen": 696844288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000531140060525818,
      "loss": 2.9651,
      "theoretical_loss": 3.7811547952669975,
      "tokens_seen": 696909824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005311235104974465,
      "loss": 2.7159,
      "theoretical_loss": 3.78111866505604,
      "tokens_seen": 696975360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005311069604690751,
      "loss": 2.816,
      "theoretical_loss": 3.7810825391933562,
      "tokens_seen": 697040896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005310904104407036,
      "loss": 2.5321,
      "theoretical_loss": 3.781046417678014,
      "tokens_seen": 697106432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005310738604123322,
      "loss": 2.8725,
      "theoretical_loss": 3.781010300509081,
      "tokens_seen": 697171968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005310573103839606,
      "loss": 2.8155,
      "theoretical_loss": 3.7809741876856258,
      "tokens_seen": 697237504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005310407603555891,
      "loss": 2.8957,
      "theoretical_loss": 3.780938079206718,
      "tokens_seen": 697303040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005310242103272177,
      "loss": 2.597,
      "theoretical_loss": 3.7809019750714254,
      "tokens_seen": 697368576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005310076602988462,
      "loss": 2.9417,
      "theoretical_loss": 3.7808658752788187,
      "tokens_seen": 697434112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005309911102704747,
      "loss": 2.8985,
      "theoretical_loss": 3.7808297798279664,
      "tokens_seen": 697499648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005309745602421033,
      "loss": 2.5488,
      "theoretical_loss": 3.7807936887179396,
      "tokens_seen": 697565184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005309580102137318,
      "loss": 2.6454,
      "theoretical_loss": 3.780757601947807,
      "tokens_seen": 697630720
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005309414601853604,
      "loss": 2.6647,
      "theoretical_loss": 3.780721519516641,
      "tokens_seen": 697696256
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005309249101569888,
      "loss": 2.7757,
      "theoretical_loss": 3.7806854414235103,
      "tokens_seen": 697761792
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005309083601286173,
      "loss": 2.6224,
      "theoretical_loss": 3.780649367667488,
      "tokens_seen": 697827328
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005308918101002459,
      "loss": 2.866,
      "theoretical_loss": 3.780613298247644,
      "tokens_seen": 697892864
    },
    {
      "epoch": 0.25,
      "objective/train/avg_token_score": 0.012157727032899857,
      "objective/train/avg_weight": 0.9856868386268616,
      "objective/train/docs_used": 402482,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7887966632843018,
      "objective/train/std_weight": 0.07723906636238098,
      "objective/train/theoretical_loss": 3.7805772331630516,
      "objective/train/tokens_used": 718418400,
      "theoretical_loss": 3.7805772331630516,
      "tokens_seen": 697958400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005308752600718744,
      "loss": 2.7627,
      "theoretical_loss": 3.7805772331630516,
      "tokens_seen": 697958400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005308587100435029,
      "loss": 3.0054,
      "theoretical_loss": 3.7805411724127804,
      "tokens_seen": 698023936
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005308421600151315,
      "loss": 2.987,
      "theoretical_loss": 3.780505115995904,
      "tokens_seen": 698089472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00053082560998676,
      "loss": 2.6812,
      "theoretical_loss": 3.7804690639114957,
      "tokens_seen": 698155008
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005308090599583885,
      "loss": 2.5896,
      "theoretical_loss": 3.780433016158627,
      "tokens_seen": 698220544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000530792509930017,
      "loss": 2.812,
      "theoretical_loss": 3.780396972736371,
      "tokens_seen": 698286080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005307759599016455,
      "loss": 2.8008,
      "theoretical_loss": 3.780360933643802,
      "tokens_seen": 698351616
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005307594098732741,
      "loss": 2.8172,
      "theoretical_loss": 3.780324898879993,
      "tokens_seen": 698417152
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005307428598449026,
      "loss": 2.8952,
      "theoretical_loss": 3.780288868444018,
      "tokens_seen": 698482688
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005307263098165311,
      "loss": 2.76,
      "theoretical_loss": 3.7802528423349515,
      "tokens_seen": 698548224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005307097597881597,
      "loss": 2.8392,
      "theoretical_loss": 3.7802168205518676,
      "tokens_seen": 698613760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005306932097597881,
      "loss": 2.8186,
      "theoretical_loss": 3.7801808030938417,
      "tokens_seen": 698679296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005306766597314167,
      "loss": 3.0396,
      "theoretical_loss": 3.780144789959948,
      "tokens_seen": 698744832
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005306601097030452,
      "loss": 2.8699,
      "theoretical_loss": 3.780108781149263,
      "tokens_seen": 698810368
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005306435596746737,
      "loss": 2.6834,
      "theoretical_loss": 3.7800727766608615,
      "tokens_seen": 698875904
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005306270096463023,
      "loss": 2.7559,
      "theoretical_loss": 3.7800367764938194,
      "tokens_seen": 698941440
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005306104596179308,
      "loss": 3.0497,
      "theoretical_loss": 3.7800007806472133,
      "tokens_seen": 699006976
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005305939095895593,
      "loss": 2.5479,
      "theoretical_loss": 3.77996478912012,
      "tokens_seen": 699072512
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005305773595611879,
      "loss": 2.8305,
      "theoretical_loss": 3.7799288019116153,
      "tokens_seen": 699138048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005305608095328163,
      "loss": 2.6594,
      "theoretical_loss": 3.7798928190207772,
      "tokens_seen": 699203584
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005305442595044449,
      "loss": 2.7114,
      "theoretical_loss": 3.7798568404466826,
      "tokens_seen": 699269120
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005305277094760734,
      "loss": 2.4913,
      "theoretical_loss": 3.779820866188409,
      "tokens_seen": 699334656
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005305111594477019,
      "loss": 2.9389,
      "theoretical_loss": 3.779784896245035,
      "tokens_seen": 699400192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005304946094193305,
      "loss": 2.6806,
      "theoretical_loss": 3.7797489306156384,
      "tokens_seen": 699465728
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000530478059390959,
      "loss": 2.7676,
      "theoretical_loss": 3.7797129692992977,
      "tokens_seen": 699531264
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005304615093625875,
      "loss": 2.7071,
      "theoretical_loss": 3.779677012295091,
      "tokens_seen": 699596800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000530444959334216,
      "loss": 2.6745,
      "theoretical_loss": 3.7796410596020986,
      "tokens_seen": 699662336
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005304284093058445,
      "loss": 2.7769,
      "theoretical_loss": 3.779605111219399,
      "tokens_seen": 699727872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005304118592774731,
      "loss": 2.8749,
      "theoretical_loss": 3.7795691671460725,
      "tokens_seen": 699793408
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005303953092491015,
      "loss": 2.6371,
      "theoretical_loss": 3.779533227381198,
      "tokens_seen": 699858944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005303787592207301,
      "loss": 2.8516,
      "theoretical_loss": 3.7794972919238563,
      "tokens_seen": 699924480
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005303622091923586,
      "loss": 3.0717,
      "theoretical_loss": 3.779461360773128,
      "tokens_seen": 699990016
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005303456591639871,
      "loss": 2.4592,
      "theoretical_loss": 3.7794254339280937,
      "tokens_seen": 700055552
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005303291091356156,
      "loss": 2.7755,
      "theoretical_loss": 3.7793895113878344,
      "tokens_seen": 700121088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005303125591072441,
      "loss": 2.5891,
      "theoretical_loss": 3.7793535931514315,
      "tokens_seen": 700186624
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005302960090788727,
      "loss": 2.6255,
      "theoretical_loss": 3.779317679217966,
      "tokens_seen": 700252160
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005302794590505012,
      "loss": 2.8024,
      "theoretical_loss": 3.779281769586521,
      "tokens_seen": 700317696
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005302629090221297,
      "loss": 2.5149,
      "theoretical_loss": 3.7792458642561777,
      "tokens_seen": 700383232
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005302463589937583,
      "loss": 2.6276,
      "theoretical_loss": 3.7792099632260188,
      "tokens_seen": 700448768
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005302298089653868,
      "loss": 2.5516,
      "theoretical_loss": 3.779174066495127,
      "tokens_seen": 700514304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005302132589370152,
      "loss": 2.7421,
      "theoretical_loss": 3.7791381740625853,
      "tokens_seen": 700579840
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005301967089086438,
      "loss": 2.6991,
      "theoretical_loss": 3.779102285927477,
      "tokens_seen": 700645376
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005301801588802723,
      "loss": 2.4666,
      "theoretical_loss": 3.779066402088886,
      "tokens_seen": 700710912
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005301636088519009,
      "loss": 2.8306,
      "theoretical_loss": 3.779030522545896,
      "tokens_seen": 700776448
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005301470588235294,
      "loss": 2.6808,
      "theoretical_loss": 3.7789946472975906,
      "tokens_seen": 700841984
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005301305087951579,
      "loss": 2.7987,
      "theoretical_loss": 3.7789587763430545,
      "tokens_seen": 700907520
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005301139587667865,
      "loss": 2.6457,
      "theoretical_loss": 3.7789229096813726,
      "tokens_seen": 700973056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000530097408738415,
      "loss": 2.8705,
      "theoretical_loss": 3.77888704731163,
      "tokens_seen": 701038592
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005300808587100434,
      "loss": 2.867,
      "theoretical_loss": 3.7788511892329115,
      "tokens_seen": 701104128
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000530064308681672,
      "loss": 2.6655,
      "theoretical_loss": 3.7788153354443024,
      "tokens_seen": 701169664
    },
    {
      "epoch": 0.25,
      "objective/train/avg_token_score": 0.009921053424477577,
      "objective/train/avg_weight": 0.988320529460907,
      "objective/train/docs_used": 404817,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8440604209899902,
      "objective/train/std_weight": 0.055882129818201065,
      "objective/train/theoretical_loss": 3.7787794859448898,
      "objective/train/tokens_used": 721695200,
      "theoretical_loss": 3.7787794859448898,
      "tokens_seen": 701235200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005300477586533005,
      "loss": 2.9044,
      "theoretical_loss": 3.7787794859448898,
      "tokens_seen": 701235200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005300312086249291,
      "loss": 2.7416,
      "theoretical_loss": 3.7787436407337585,
      "tokens_seen": 701300736
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005300146585965576,
      "loss": 2.779,
      "theoretical_loss": 3.7787077998099954,
      "tokens_seen": 701366272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005299981085681861,
      "loss": 2.736,
      "theoretical_loss": 3.778671963172687,
      "tokens_seen": 701431808
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005299815585398147,
      "loss": 2.691,
      "theoretical_loss": 3.77863613082092,
      "tokens_seen": 701497344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005299650085114431,
      "loss": 2.7924,
      "theoretical_loss": 3.778600302753783,
      "tokens_seen": 701562880
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005299484584830716,
      "loss": 2.6837,
      "theoretical_loss": 3.7785644789703614,
      "tokens_seen": 701628416
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005299319084547002,
      "loss": 2.6004,
      "theoretical_loss": 3.7785286594697443,
      "tokens_seen": 701693952
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005299153584263287,
      "loss": 2.97,
      "theoretical_loss": 3.7784928442510193,
      "tokens_seen": 701759488
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005298988083979573,
      "loss": 2.7423,
      "theoretical_loss": 3.778457033313275,
      "tokens_seen": 701825024
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005298822583695858,
      "loss": 2.6827,
      "theoretical_loss": 3.7784212266556008,
      "tokens_seen": 701890560
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005298657083412143,
      "loss": 2.6622,
      "theoretical_loss": 3.778385424277084,
      "tokens_seen": 701956096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005298491583128429,
      "loss": 2.8583,
      "theoretical_loss": 3.778349626176814,
      "tokens_seen": 702021632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005298326082844713,
      "loss": 2.9417,
      "theoretical_loss": 3.7783138323538807,
      "tokens_seen": 702087168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005298160582560998,
      "loss": 2.875,
      "theoretical_loss": 3.778278042807375,
      "tokens_seen": 702152704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005297995082277284,
      "loss": 2.677,
      "theoretical_loss": 3.7782422575363848,
      "tokens_seen": 702218240
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005297829581993569,
      "loss": 2.7377,
      "theoretical_loss": 3.778206476540001,
      "tokens_seen": 702283776
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005297664081709855,
      "loss": 2.5879,
      "theoretical_loss": 3.7781706998173155,
      "tokens_seen": 702349312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000529749858142614,
      "loss": 2.5649,
      "theoretical_loss": 3.778134927367417,
      "tokens_seen": 702414848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005297333081142425,
      "loss": 2.5312,
      "theoretical_loss": 3.778099159189399,
      "tokens_seen": 702480384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000529716758085871,
      "loss": 2.744,
      "theoretical_loss": 3.7780633952823512,
      "tokens_seen": 702545920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005297002080574995,
      "loss": 2.6248,
      "theoretical_loss": 3.7780276356453655,
      "tokens_seen": 702611456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000529683658029128,
      "loss": 2.7473,
      "theoretical_loss": 3.7779918802775345,
      "tokens_seen": 702676992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005296671080007566,
      "loss": 2.8534,
      "theoretical_loss": 3.7779561291779498,
      "tokens_seen": 702742528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005296505579723851,
      "loss": 2.7377,
      "theoretical_loss": 3.777920382345704,
      "tokens_seen": 702808064
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005296340079440137,
      "loss": 2.8108,
      "theoretical_loss": 3.7778846397798898,
      "tokens_seen": 702873600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005296174579156422,
      "loss": 2.741,
      "theoretical_loss": 3.777848901479601,
      "tokens_seen": 702939136
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005296009078872707,
      "loss": 2.4115,
      "theoretical_loss": 3.77781316744393,
      "tokens_seen": 703004672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005295843578588992,
      "loss": 2.7768,
      "theoretical_loss": 3.777777437671971,
      "tokens_seen": 703070208
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005295678078305277,
      "loss": 2.592,
      "theoretical_loss": 3.777741712162818,
      "tokens_seen": 703135744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005295512578021562,
      "loss": 2.6955,
      "theoretical_loss": 3.777705990915565,
      "tokens_seen": 703201280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005295347077737848,
      "loss": 2.7328,
      "theoretical_loss": 3.7776702739293055,
      "tokens_seen": 703266816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005295181577454133,
      "loss": 2.8561,
      "theoretical_loss": 3.777634561203136,
      "tokens_seen": 703332352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005295016077170419,
      "loss": 2.5589,
      "theoretical_loss": 3.7775988527361504,
      "tokens_seen": 703397888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005294850576886704,
      "loss": 2.9552,
      "theoretical_loss": 3.7775631485274435,
      "tokens_seen": 703463424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005294685076602988,
      "loss": 2.8036,
      "theoretical_loss": 3.7775274485761123,
      "tokens_seen": 703528960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005294519576319274,
      "loss": 2.705,
      "theoretical_loss": 3.7774917528812515,
      "tokens_seen": 703594496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005294354076035559,
      "loss": 2.5779,
      "theoretical_loss": 3.7774560614419577,
      "tokens_seen": 703660032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005294188575751844,
      "loss": 2.7632,
      "theoretical_loss": 3.777420374257327,
      "tokens_seen": 703725568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005294023075468129,
      "loss": 2.485,
      "theoretical_loss": 3.7773846913264566,
      "tokens_seen": 703791104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005293857575184414,
      "loss": 2.7094,
      "theoretical_loss": 3.777349012648443,
      "tokens_seen": 703856640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00052936920749007,
      "loss": 2.7152,
      "theoretical_loss": 3.777313338222383,
      "tokens_seen": 703922176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005293526574616984,
      "loss": 2.8269,
      "theoretical_loss": 3.7772776680473745,
      "tokens_seen": 703987712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000529336107433327,
      "loss": 2.9214,
      "theoretical_loss": 3.7772420021225157,
      "tokens_seen": 704053248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005293195574049555,
      "loss": 2.5125,
      "theoretical_loss": 3.777206340446904,
      "tokens_seen": 704118784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000529303007376584,
      "loss": 2.881,
      "theoretical_loss": 3.7771706830196385,
      "tokens_seen": 704184320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005292864573482126,
      "loss": 2.9098,
      "theoretical_loss": 3.777135029839817,
      "tokens_seen": 704249856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005292699073198411,
      "loss": 2.7467,
      "theoretical_loss": 3.777099380906538,
      "tokens_seen": 704315392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005292533572914696,
      "loss": 2.758,
      "theoretical_loss": 3.777063736218902,
      "tokens_seen": 704380928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005292368072630982,
      "loss": 2.8125,
      "theoretical_loss": 3.7770280957760076,
      "tokens_seen": 704446464
    },
    {
      "epoch": 0.25,
      "objective/train/avg_token_score": 0.04208316653966904,
      "objective/train/avg_weight": 0.95048987865448,
      "objective/train/docs_used": 406758,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5287270545959473,
      "objective/train/std_weight": 0.18069475889205933,
      "objective/train/theoretical_loss": 3.7769924595769546,
      "objective/train/tokens_used": 724972000,
      "theoretical_loss": 3.7769924595769546,
      "tokens_seen": 704512000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005292202572347266,
      "loss": 2.6015,
      "theoretical_loss": 3.7769924595769546,
      "tokens_seen": 704512000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005292037072063552,
      "loss": 2.9537,
      "theoretical_loss": 3.776956827620843,
      "tokens_seen": 704577536
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005291871571779837,
      "loss": 2.7961,
      "theoretical_loss": 3.7769211999067727,
      "tokens_seen": 704643072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005291706071496122,
      "loss": 2.9086,
      "theoretical_loss": 3.7768855764338447,
      "tokens_seen": 704708608
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005291540571212408,
      "loss": 2.6872,
      "theoretical_loss": 3.776849957201159,
      "tokens_seen": 704774144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005291375070928693,
      "loss": 2.5961,
      "theoretical_loss": 3.776814342207818,
      "tokens_seen": 704839680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005291209570644978,
      "loss": 2.727,
      "theoretical_loss": 3.7767787314529215,
      "tokens_seen": 704905216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005291044070361263,
      "loss": 2.6704,
      "theoretical_loss": 3.7767431249355727,
      "tokens_seen": 704970752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005290878570077548,
      "loss": 2.6965,
      "theoretical_loss": 3.7767075226548723,
      "tokens_seen": 705036288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005290713069793834,
      "loss": 2.7792,
      "theoretical_loss": 3.7766719246099227,
      "tokens_seen": 705101824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005290547569510119,
      "loss": 2.8811,
      "theoretical_loss": 3.7766363307998265,
      "tokens_seen": 705167360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005290382069226404,
      "loss": 2.7257,
      "theoretical_loss": 3.7766007412236866,
      "tokens_seen": 705232896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000529021656894269,
      "loss": 2.8949,
      "theoretical_loss": 3.7765651558806055,
      "tokens_seen": 705298432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005290051068658975,
      "loss": 2.5568,
      "theoretical_loss": 3.776529574769687,
      "tokens_seen": 705363968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000528988556837526,
      "loss": 2.5713,
      "theoretical_loss": 3.776493997890034,
      "tokens_seen": 705429504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005289720068091545,
      "loss": 2.8738,
      "theoretical_loss": 3.776458425240751,
      "tokens_seen": 705495040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000528955456780783,
      "loss": 2.798,
      "theoretical_loss": 3.776422856820941,
      "tokens_seen": 705560576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005289389067524116,
      "loss": 2.6777,
      "theoretical_loss": 3.7763872926297095,
      "tokens_seen": 705626112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005289223567240401,
      "loss": 2.5836,
      "theoretical_loss": 3.7763517326661606,
      "tokens_seen": 705691648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005289058066956686,
      "loss": 3.0898,
      "theoretical_loss": 3.7763161769293996,
      "tokens_seen": 705757184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005288892566672972,
      "loss": 2.7254,
      "theoretical_loss": 3.776280625418531,
      "tokens_seen": 705822720
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005288727066389257,
      "loss": 2.5824,
      "theoretical_loss": 3.77624507813266,
      "tokens_seen": 705888256
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005288561566105542,
      "loss": 2.6344,
      "theoretical_loss": 3.7762095350708935,
      "tokens_seen": 705953792
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005288396065821827,
      "loss": 2.8439,
      "theoretical_loss": 3.7761739962323366,
      "tokens_seen": 706019328
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005288230565538112,
      "loss": 2.6564,
      "theoretical_loss": 3.776138461616096,
      "tokens_seen": 706084864
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005288065065254398,
      "loss": 2.6649,
      "theoretical_loss": 3.7761029312212777,
      "tokens_seen": 706150400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005287899564970683,
      "loss": 2.8125,
      "theoretical_loss": 3.776067405046989,
      "tokens_seen": 706215936
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005287734064686968,
      "loss": 2.7757,
      "theoretical_loss": 3.7760318830923367,
      "tokens_seen": 706281472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005287734064686968,
      "loss": 2.8574,
      "theoretical_loss": 3.775996365356428,
      "tokens_seen": 706347008
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005287568564403254,
      "loss": 2.6794,
      "theoretical_loss": 3.775960851838371,
      "tokens_seen": 706412544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005287403064119538,
      "loss": 2.8917,
      "theoretical_loss": 3.7759253425372736,
      "tokens_seen": 706478080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005287237563835824,
      "loss": 2.7125,
      "theoretical_loss": 3.775889837452243,
      "tokens_seen": 706543616
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005287072063552109,
      "loss": 2.7811,
      "theoretical_loss": 3.7758543365823884,
      "tokens_seen": 706609152
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005286906563268394,
      "loss": 2.5457,
      "theoretical_loss": 3.775818839926819,
      "tokens_seen": 706674688
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000528674106298468,
      "loss": 2.7414,
      "theoretical_loss": 3.7757833474846425,
      "tokens_seen": 706740224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005286575562700965,
      "loss": 2.5327,
      "theoretical_loss": 3.775747859254969,
      "tokens_seen": 706805760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000528641006241725,
      "loss": 2.899,
      "theoretical_loss": 3.7757123752369077,
      "tokens_seen": 706871296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005286244562133536,
      "loss": 2.6091,
      "theoretical_loss": 3.775676895429569,
      "tokens_seen": 706936832
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000528607906184982,
      "loss": 2.7922,
      "theoretical_loss": 3.775641419832062,
      "tokens_seen": 707002368
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005285913561566106,
      "loss": 2.6458,
      "theoretical_loss": 3.775605948443497,
      "tokens_seen": 707067904
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005285748061282391,
      "loss": 2.8845,
      "theoretical_loss": 3.7755704812629856,
      "tokens_seen": 707133440
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005285582560998676,
      "loss": 2.6469,
      "theoretical_loss": 3.775535018289638,
      "tokens_seen": 707198976
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005285417060714962,
      "loss": 2.7554,
      "theoretical_loss": 3.7754995595225656,
      "tokens_seen": 707264512
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005285251560431247,
      "loss": 2.5971,
      "theoretical_loss": 3.7754641049608795,
      "tokens_seen": 707330048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005285086060147532,
      "loss": 2.5365,
      "theoretical_loss": 3.775428654603691,
      "tokens_seen": 707395584
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005284920559863817,
      "loss": 2.6926,
      "theoretical_loss": 3.7753932084501134,
      "tokens_seen": 707461120
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005284755059580102,
      "loss": 2.86,
      "theoretical_loss": 3.7753577664992575,
      "tokens_seen": 707526656
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005284589559296388,
      "loss": 2.5807,
      "theoretical_loss": 3.775322328750237,
      "tokens_seen": 707592192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005284424059012673,
      "loss": 2.6124,
      "theoretical_loss": 3.7752868952021634,
      "tokens_seen": 707657728
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005284258558728958,
      "loss": 2.3314,
      "theoretical_loss": 3.7752514658541503,
      "tokens_seen": 707723264
    },
    {
      "epoch": 0.25,
      "objective/train/avg_token_score": 0.01784474402666092,
      "objective/train/avg_weight": 0.9789999723434448,
      "objective/train/docs_used": 408602,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4648513793945312,
      "objective/train/std_weight": 0.09039363265037537,
      "objective/train/theoretical_loss": 3.7752160407053115,
      "objective/train/tokens_used": 728248800,
      "theoretical_loss": 3.7752160407053115,
      "tokens_seen": 707788800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005284093058445243,
      "loss": 2.5113,
      "theoretical_loss": 3.7752160407053115,
      "tokens_seen": 707788800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005283927558161528,
      "loss": 2.5046,
      "theoretical_loss": 3.7751806197547593,
      "tokens_seen": 707854336
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005283762057877813,
      "loss": 2.8183,
      "theoretical_loss": 3.775145203001609,
      "tokens_seen": 707919872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005283596557594098,
      "loss": 2.8233,
      "theoretical_loss": 3.775109790444974,
      "tokens_seen": 707985408
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005283431057310383,
      "loss": 2.5222,
      "theoretical_loss": 3.775074382083968,
      "tokens_seen": 708050944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005283265557026669,
      "loss": 2.7429,
      "theoretical_loss": 3.775038977917707,
      "tokens_seen": 708116480
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005283100056742954,
      "loss": 2.6641,
      "theoretical_loss": 3.775003577945305,
      "tokens_seen": 708182016
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000528293455645924,
      "loss": 2.7998,
      "theoretical_loss": 3.7749681821658774,
      "tokens_seen": 708247552
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005282769056175525,
      "loss": 2.5925,
      "theoretical_loss": 3.7749327905785393,
      "tokens_seen": 708313088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005282603555891809,
      "loss": 2.7742,
      "theoretical_loss": 3.774897403182407,
      "tokens_seen": 708378624
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005282438055608095,
      "loss": 2.8061,
      "theoretical_loss": 3.774862019976596,
      "tokens_seen": 708444160
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000528227255532438,
      "loss": 2.8835,
      "theoretical_loss": 3.7748266409602236,
      "tokens_seen": 708509696
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005282107055040665,
      "loss": 2.6091,
      "theoretical_loss": 3.7747912661324046,
      "tokens_seen": 708575232
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005281941554756951,
      "loss": 2.9908,
      "theoretical_loss": 3.774755895492257,
      "tokens_seen": 708640768
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005281776054473236,
      "loss": 2.5301,
      "theoretical_loss": 3.7747205290388974,
      "tokens_seen": 708706304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005281610554189522,
      "loss": 2.7098,
      "theoretical_loss": 3.7746851667714436,
      "tokens_seen": 708771840
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005281445053905807,
      "loss": 2.8517,
      "theoretical_loss": 3.774649808689013,
      "tokens_seen": 708837376
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005281279553622091,
      "loss": 2.5118,
      "theoretical_loss": 3.774614454790722,
      "tokens_seen": 708902912
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005281114053338377,
      "loss": 2.8089,
      "theoretical_loss": 3.7745791050756914,
      "tokens_seen": 708968448
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005280948553054662,
      "loss": 2.9114,
      "theoretical_loss": 3.774543759543037,
      "tokens_seen": 709033984
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005280783052770947,
      "loss": 2.5625,
      "theoretical_loss": 3.7745084181918793,
      "tokens_seen": 709099520
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005280617552487233,
      "loss": 2.7871,
      "theoretical_loss": 3.7744730810213367,
      "tokens_seen": 709165056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005280452052203518,
      "loss": 2.5735,
      "theoretical_loss": 3.774437748030528,
      "tokens_seen": 709230592
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005280286551919804,
      "loss": 2.6624,
      "theoretical_loss": 3.774402419218573,
      "tokens_seen": 709296128
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005280121051636088,
      "loss": 2.7941,
      "theoretical_loss": 3.7743670945845915,
      "tokens_seen": 709361664
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005279955551352373,
      "loss": 2.6264,
      "theoretical_loss": 3.774331774127703,
      "tokens_seen": 709427200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005279790051068659,
      "loss": 2.554,
      "theoretical_loss": 3.7742964578470284,
      "tokens_seen": 709492736
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005279624550784944,
      "loss": 2.5662,
      "theoretical_loss": 3.7742611457416877,
      "tokens_seen": 709558272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005279459050501229,
      "loss": 2.4755,
      "theoretical_loss": 3.7742258378108016,
      "tokens_seen": 709623808
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005279293550217515,
      "loss": 2.626,
      "theoretical_loss": 3.774190534053492,
      "tokens_seen": 709689344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00052791280499338,
      "loss": 2.6888,
      "theoretical_loss": 3.7741552344688793,
      "tokens_seen": 709754880
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005278962549650086,
      "loss": 2.8534,
      "theoretical_loss": 3.774119939056086,
      "tokens_seen": 709820416
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000527879704936637,
      "loss": 2.7813,
      "theoretical_loss": 3.7740846478142327,
      "tokens_seen": 709885952
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005278631549082655,
      "loss": 2.7336,
      "theoretical_loss": 3.774049360742443,
      "tokens_seen": 709951488
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005278466048798941,
      "loss": 2.6182,
      "theoretical_loss": 3.774014077839838,
      "tokens_seen": 710017024
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005278300548515226,
      "loss": 2.7544,
      "theoretical_loss": 3.7739787991055413,
      "tokens_seen": 710082560
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005278135048231511,
      "loss": 2.7285,
      "theoretical_loss": 3.773943524538675,
      "tokens_seen": 710148096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005277969547947797,
      "loss": 2.6282,
      "theoretical_loss": 3.773908254138363,
      "tokens_seen": 710213632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005277804047664082,
      "loss": 2.7373,
      "theoretical_loss": 3.7738729879037285,
      "tokens_seen": 710279168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005277638547380367,
      "loss": 2.5634,
      "theoretical_loss": 3.7738377258338947,
      "tokens_seen": 710344704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005277473047096652,
      "loss": 2.8102,
      "theoretical_loss": 3.773802467927986,
      "tokens_seen": 710410240
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005277307546812937,
      "loss": 2.7974,
      "theoretical_loss": 3.7737672141851273,
      "tokens_seen": 710475776
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005277142046529223,
      "loss": 2.3804,
      "theoretical_loss": 3.7737319646044423,
      "tokens_seen": 710541312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005276976546245508,
      "loss": 2.7514,
      "theoretical_loss": 3.7736967191850557,
      "tokens_seen": 710606848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005276811045961793,
      "loss": 2.7128,
      "theoretical_loss": 3.773661477926093,
      "tokens_seen": 710672384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005276645545678079,
      "loss": 2.8743,
      "theoretical_loss": 3.773626240826679,
      "tokens_seen": 710737920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005276480045394363,
      "loss": 2.7983,
      "theoretical_loss": 3.7735910078859396,
      "tokens_seen": 710803456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005276314545110649,
      "loss": 2.7763,
      "theoretical_loss": 3.7735557791030008,
      "tokens_seen": 710868992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005276149044826934,
      "loss": 2.6118,
      "theoretical_loss": 3.773520554476989,
      "tokens_seen": 710934528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005275983544543219,
      "loss": 2.7069,
      "theoretical_loss": 3.7734853340070296,
      "tokens_seen": 711000064
    },
    {
      "epoch": 0.25,
      "objective/train/avg_token_score": 0.010946712456643581,
      "objective/train/avg_weight": 0.9871148467063904,
      "objective/train/docs_used": 410204,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.768474578857422,
      "objective/train/std_weight": 0.06671565026044846,
      "objective/train/theoretical_loss": 3.7734501176922493,
      "objective/train/tokens_used": 731525600,
      "theoretical_loss": 3.7734501176922493,
      "tokens_seen": 711065600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005275818044259505,
      "loss": 2.8084,
      "theoretical_loss": 3.7734501176922493,
      "tokens_seen": 711065600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000527565254397579,
      "loss": 2.6792,
      "theoretical_loss": 3.7734149055317756,
      "tokens_seen": 711131136
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005275487043692075,
      "loss": 2.6099,
      "theoretical_loss": 3.773379697524736,
      "tokens_seen": 711196672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005275321543408361,
      "loss": 2.7246,
      "theoretical_loss": 3.773344493670257,
      "tokens_seen": 711262208
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005275156043124645,
      "loss": 2.6753,
      "theoretical_loss": 3.7733092939674666,
      "tokens_seen": 711327744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005274990542840931,
      "loss": 2.8515,
      "theoretical_loss": 3.7732740984154924,
      "tokens_seen": 711393280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005274825042557216,
      "loss": 2.7018,
      "theoretical_loss": 3.7732389070134635,
      "tokens_seen": 711458816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005274659542273501,
      "loss": 2.7489,
      "theoretical_loss": 3.7732037197605077,
      "tokens_seen": 711524352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005274494041989786,
      "loss": 2.624,
      "theoretical_loss": 3.7731685366557537,
      "tokens_seen": 711589888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005274328541706071,
      "loss": 3.175,
      "theoretical_loss": 3.773133357698331,
      "tokens_seen": 711655424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005274163041422357,
      "loss": 2.481,
      "theoretical_loss": 3.7730981828873684,
      "tokens_seen": 711720960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005273997541138641,
      "loss": 2.6899,
      "theoretical_loss": 3.7730630122219955,
      "tokens_seen": 711786496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005273832040854927,
      "loss": 2.5072,
      "theoretical_loss": 3.773027845701342,
      "tokens_seen": 711852032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005273666540571212,
      "loss": 2.596,
      "theoretical_loss": 3.7729926833245386,
      "tokens_seen": 711917568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005273501040287497,
      "loss": 2.7221,
      "theoretical_loss": 3.772957525090715,
      "tokens_seen": 711983104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005273335540003783,
      "loss": 2.7257,
      "theoretical_loss": 3.7729223709990016,
      "tokens_seen": 712048640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005273170039720068,
      "loss": 2.9171,
      "theoretical_loss": 3.7728872210485296,
      "tokens_seen": 712114176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005273004539436353,
      "loss": 2.714,
      "theoretical_loss": 3.77285207523843,
      "tokens_seen": 712179712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005272839039152638,
      "loss": 2.9214,
      "theoretical_loss": 3.7728169335678343,
      "tokens_seen": 712245248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005272673538868923,
      "loss": 2.6803,
      "theoretical_loss": 3.772781796035874,
      "tokens_seen": 712310784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005272508038585209,
      "loss": 2.8418,
      "theoretical_loss": 3.772746662641681,
      "tokens_seen": 712376320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005272342538301494,
      "loss": 2.7933,
      "theoretical_loss": 3.7727115333843875,
      "tokens_seen": 712441856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005272177038017779,
      "loss": 2.7575,
      "theoretical_loss": 3.772676408263126,
      "tokens_seen": 712507392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005272011537734065,
      "loss": 2.6028,
      "theoretical_loss": 3.772641287277029,
      "tokens_seen": 712572928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000527184603745035,
      "loss": 2.9259,
      "theoretical_loss": 3.772606170425229,
      "tokens_seen": 712638464
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005271680537166634,
      "loss": 2.9292,
      "theoretical_loss": 3.77257105770686,
      "tokens_seen": 712704000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000527151503688292,
      "loss": 2.7162,
      "theoretical_loss": 3.772535949121055,
      "tokens_seen": 712769536
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005271349536599205,
      "loss": 2.6408,
      "theoretical_loss": 3.772500844666948,
      "tokens_seen": 712835072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005271184036315491,
      "loss": 2.922,
      "theoretical_loss": 3.7724657443436724,
      "tokens_seen": 712900608
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005271018536031776,
      "loss": 2.6239,
      "theoretical_loss": 3.7724306481503636,
      "tokens_seen": 712966144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005270853035748061,
      "loss": 2.8627,
      "theoretical_loss": 3.7723955560861544,
      "tokens_seen": 713031680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005270687535464347,
      "loss": 2.501,
      "theoretical_loss": 3.7723604681501812,
      "tokens_seen": 713097216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005270522035180632,
      "loss": 2.8218,
      "theoretical_loss": 3.7723253843415785,
      "tokens_seen": 713162752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005270356534896916,
      "loss": 2.7886,
      "theoretical_loss": 3.7722903046594807,
      "tokens_seen": 713228288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005270191034613202,
      "loss": 2.7284,
      "theoretical_loss": 3.772255229103025,
      "tokens_seen": 713293824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005270025534329487,
      "loss": 2.6985,
      "theoretical_loss": 3.772220157671345,
      "tokens_seen": 713359360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005269860034045773,
      "loss": 2.7936,
      "theoretical_loss": 3.7721850903635787,
      "tokens_seen": 713424896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005269694533762058,
      "loss": 2.4495,
      "theoretical_loss": 3.7721500271788617,
      "tokens_seen": 713490432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005269529033478343,
      "loss": 2.8619,
      "theoretical_loss": 3.772114968116331,
      "tokens_seen": 713555968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005269363533194629,
      "loss": 2.5806,
      "theoretical_loss": 3.772079913175123,
      "tokens_seen": 713621504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005269198032910913,
      "loss": 2.7691,
      "theoretical_loss": 3.7720448623543743,
      "tokens_seen": 713687040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005269032532627198,
      "loss": 2.7587,
      "theoretical_loss": 3.7720098156532234,
      "tokens_seen": 713752576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005268867032343484,
      "loss": 2.7473,
      "theoretical_loss": 3.771974773070807,
      "tokens_seen": 713818112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005268701532059769,
      "loss": 2.9618,
      "theoretical_loss": 3.7719397346062635,
      "tokens_seen": 713883648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0005268536031776055,
      "loss": 2.7191,
      "theoretical_loss": 3.771904700258731,
      "tokens_seen": 713949184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000526837053149234,
      "loss": 2.7747,
      "theoretical_loss": 3.771869670027348,
      "tokens_seen": 714014720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005268205031208625,
      "loss": 2.6389,
      "theoretical_loss": 3.7718346439112524,
      "tokens_seen": 714080256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005268039530924911,
      "loss": 2.7115,
      "theoretical_loss": 3.771799621909584,
      "tokens_seen": 714145792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005267874030641195,
      "loss": 2.719,
      "theoretical_loss": 3.771764604021482,
      "tokens_seen": 714211328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000526770853035748,
      "loss": 2.8538,
      "theoretical_loss": 3.771729590246085,
      "tokens_seen": 714276864
    },
    {
      "epoch": 0.26,
      "objective/train/avg_token_score": 0.022645477205514908,
      "objective/train/avg_weight": 0.9733468294143677,
      "objective/train/docs_used": 412084,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8671722412109375,
      "objective/train/std_weight": 0.1320657581090927,
      "objective/train/theoretical_loss": 3.7716945805825337,
      "objective/train/tokens_used": 734802400,
      "theoretical_loss": 3.7716945805825337,
      "tokens_seen": 714342400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005267543030073766,
      "loss": 2.7956,
      "theoretical_loss": 3.7716945805825337,
      "tokens_seen": 714342400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005267377529790051,
      "loss": 2.8424,
      "theoretical_loss": 3.7716595750299673,
      "tokens_seen": 714407936
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005267212029506337,
      "loss": 2.7549,
      "theoretical_loss": 3.771624573587526,
      "tokens_seen": 714473472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005267046529222622,
      "loss": 3.0485,
      "theoretical_loss": 3.771589576254351,
      "tokens_seen": 714539008
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005266881028938907,
      "loss": 2.799,
      "theoretical_loss": 3.7715545830295825,
      "tokens_seen": 714604544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005266715528655192,
      "loss": 2.9975,
      "theoretical_loss": 3.7715195939123616,
      "tokens_seen": 714670080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005266550028371477,
      "loss": 2.8582,
      "theoretical_loss": 3.7714846089018295,
      "tokens_seen": 714735616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005266384528087762,
      "loss": 2.8029,
      "theoretical_loss": 3.7714496279971277,
      "tokens_seen": 714801152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005266219027804048,
      "loss": 2.713,
      "theoretical_loss": 3.7714146511973983,
      "tokens_seen": 714866688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005266053527520333,
      "loss": 2.7587,
      "theoretical_loss": 3.7713796785017824,
      "tokens_seen": 714932224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005265888027236619,
      "loss": 2.681,
      "theoretical_loss": 3.7713447099094233,
      "tokens_seen": 714997760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005265722526952904,
      "loss": 2.6648,
      "theoretical_loss": 3.7713097454194635,
      "tokens_seen": 715063296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005265557026669188,
      "loss": 2.8985,
      "theoretical_loss": 3.7712747850310446,
      "tokens_seen": 715128832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005265391526385474,
      "loss": 2.8141,
      "theoretical_loss": 3.7712398287433113,
      "tokens_seen": 715194368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005265226026101759,
      "loss": 2.7543,
      "theoretical_loss": 3.771204876555406,
      "tokens_seen": 715259904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005265060525818044,
      "loss": 2.8625,
      "theoretical_loss": 3.771169928466472,
      "tokens_seen": 715325440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000526489502553433,
      "loss": 2.477,
      "theoretical_loss": 3.7711349844756543,
      "tokens_seen": 715390976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005264729525250615,
      "loss": 3.0411,
      "theoretical_loss": 3.771100044582096,
      "tokens_seen": 715456512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00052645640249669,
      "loss": 2.6771,
      "theoretical_loss": 3.771065108784941,
      "tokens_seen": 715522048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005264398524683184,
      "loss": 2.8879,
      "theoretical_loss": 3.771030177083335,
      "tokens_seen": 715587584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000526423302439947,
      "loss": 2.6406,
      "theoretical_loss": 3.7709952494764227,
      "tokens_seen": 715653120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005264067524115755,
      "loss": 2.6527,
      "theoretical_loss": 3.7709603259633484,
      "tokens_seen": 715718656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000526390202383204,
      "loss": 2.6371,
      "theoretical_loss": 3.7709254065432583,
      "tokens_seen": 715784192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005263736523548326,
      "loss": 2.71,
      "theoretical_loss": 3.770890491215298,
      "tokens_seen": 715849728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005263571023264611,
      "loss": 2.6921,
      "theoretical_loss": 3.7708555799786128,
      "tokens_seen": 715915264
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005263405522980897,
      "loss": 2.6324,
      "theoretical_loss": 3.7708206728323495,
      "tokens_seen": 715980800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005263240022697182,
      "loss": 2.6378,
      "theoretical_loss": 3.770785769775654,
      "tokens_seen": 716046336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005263074522413466,
      "loss": 3.0044,
      "theoretical_loss": 3.7707508708076727,
      "tokens_seen": 716111872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005262909022129752,
      "loss": 2.9614,
      "theoretical_loss": 3.770715975927553,
      "tokens_seen": 716177408
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005262743521846037,
      "loss": 2.8048,
      "theoretical_loss": 3.7706810851344423,
      "tokens_seen": 716242944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005262578021562322,
      "loss": 2.9709,
      "theoretical_loss": 3.7706461984274875,
      "tokens_seen": 716308480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005262412521278608,
      "loss": 2.6826,
      "theoretical_loss": 3.7706113158058363,
      "tokens_seen": 716374016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005262247020994893,
      "loss": 2.6925,
      "theoretical_loss": 3.7705764372686374,
      "tokens_seen": 716439552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005262081520711179,
      "loss": 2.7412,
      "theoretical_loss": 3.770541562815038,
      "tokens_seen": 716505088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005261916020427463,
      "loss": 2.8408,
      "theoretical_loss": 3.770506692444187,
      "tokens_seen": 716570624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005261750520143748,
      "loss": 2.7556,
      "theoretical_loss": 3.7704718261552332,
      "tokens_seen": 716636160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005261585019860034,
      "loss": 2.8293,
      "theoretical_loss": 3.770436963947325,
      "tokens_seen": 716701696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005261419519576319,
      "loss": 2.7483,
      "theoretical_loss": 3.7704021058196124,
      "tokens_seen": 716767232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005261254019292604,
      "loss": 2.9194,
      "theoretical_loss": 3.770367251771245,
      "tokens_seen": 716832768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000526108851900889,
      "loss": 2.842,
      "theoretical_loss": 3.7703324018013706,
      "tokens_seen": 716898304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005260923018725175,
      "loss": 2.8941,
      "theoretical_loss": 3.7702975559091416,
      "tokens_seen": 716963840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005260757518441461,
      "loss": 2.6166,
      "theoretical_loss": 3.7702627140937075,
      "tokens_seen": 717029376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005260592018157745,
      "loss": 2.9111,
      "theoretical_loss": 3.7702278763542174,
      "tokens_seen": 717094912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000526042651787403,
      "loss": 2.7658,
      "theoretical_loss": 3.770193042689824,
      "tokens_seen": 717160448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005260261017590316,
      "loss": 2.753,
      "theoretical_loss": 3.770158213099677,
      "tokens_seen": 717225984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005260095517306601,
      "loss": 2.745,
      "theoretical_loss": 3.7701233875829283,
      "tokens_seen": 717291520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005259930017022886,
      "loss": 2.754,
      "theoretical_loss": 3.7700885661387287,
      "tokens_seen": 717357056
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005259764516739172,
      "loss": 2.9819,
      "theoretical_loss": 3.770053748766231,
      "tokens_seen": 717422592
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005259599016455457,
      "loss": 2.8203,
      "theoretical_loss": 3.770018935464586,
      "tokens_seen": 717488128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005259433516171742,
      "loss": 2.8079,
      "theoretical_loss": 3.7699841262329477,
      "tokens_seen": 717553664
    },
    {
      "epoch": 0.26,
      "objective/train/avg_token_score": 0.026916783303022385,
      "objective/train/avg_weight": 0.9683293700218201,
      "objective/train/docs_used": 413912,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5533580780029297,
      "objective/train/std_weight": 0.12172781676054001,
      "objective/train/theoretical_loss": 3.7699493210704667,
      "objective/train/tokens_used": 738079200,
      "theoretical_loss": 3.7699493210704667,
      "tokens_seen": 717619200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005259268015888027,
      "loss": 2.7241,
      "theoretical_loss": 3.7699493210704667,
      "tokens_seen": 717619200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005259102515604312,
      "loss": 2.9564,
      "theoretical_loss": 3.7699145199762967,
      "tokens_seen": 717684736
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005258937015320598,
      "loss": 2.6593,
      "theoretical_loss": 3.7698797229495904,
      "tokens_seen": 717750272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005258771515036883,
      "loss": 2.6912,
      "theoretical_loss": 3.7698449299895014,
      "tokens_seen": 717815808
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005258606014753168,
      "loss": 2.87,
      "theoretical_loss": 3.769810141095183,
      "tokens_seen": 717881344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005258440514469454,
      "loss": 2.7918,
      "theoretical_loss": 3.76977535626579,
      "tokens_seen": 717946880
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005258275014185738,
      "loss": 2.8898,
      "theoretical_loss": 3.769740575500475,
      "tokens_seen": 718012416
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005258109513902024,
      "loss": 2.6156,
      "theoretical_loss": 3.769705798798393,
      "tokens_seen": 718077952
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005257944013618309,
      "loss": 2.8557,
      "theoretical_loss": 3.769671026158698,
      "tokens_seen": 718143488
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005257778513334594,
      "loss": 2.8874,
      "theoretical_loss": 3.769636257580545,
      "tokens_seen": 718209024
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000525761301305088,
      "loss": 2.8934,
      "theoretical_loss": 3.76960149306309,
      "tokens_seen": 718274560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005257447512767165,
      "loss": 2.9392,
      "theoretical_loss": 3.769566732605487,
      "tokens_seen": 718340096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000525728201248345,
      "loss": 2.7271,
      "theoretical_loss": 3.769531976206893,
      "tokens_seen": 718405632
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005257116512199736,
      "loss": 2.8851,
      "theoretical_loss": 3.7694972238664617,
      "tokens_seen": 718471168
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000525695101191602,
      "loss": 2.7017,
      "theoretical_loss": 3.7694624755833512,
      "tokens_seen": 718536704
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005256785511632306,
      "loss": 2.7764,
      "theoretical_loss": 3.769427731356717,
      "tokens_seen": 718602240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005256620011348591,
      "loss": 3.0425,
      "theoretical_loss": 3.7693929911857156,
      "tokens_seen": 718667776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005256454511064876,
      "loss": 2.7548,
      "theoretical_loss": 3.7693582550695037,
      "tokens_seen": 718733312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005256289010781162,
      "loss": 2.5299,
      "theoretical_loss": 3.7693235230072384,
      "tokens_seen": 718798848
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005256123510497447,
      "loss": 2.7564,
      "theoretical_loss": 3.7692887949980776,
      "tokens_seen": 718864384
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005255958010213732,
      "loss": 2.6774,
      "theoretical_loss": 3.7692540710411784,
      "tokens_seen": 718929920
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005255792509930017,
      "loss": 2.8796,
      "theoretical_loss": 3.769219351135699,
      "tokens_seen": 718995456
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005255627009646302,
      "loss": 2.7709,
      "theoretical_loss": 3.769184635280797,
      "tokens_seen": 719060992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005255461509362588,
      "loss": 2.7275,
      "theoretical_loss": 3.7691499234756303,
      "tokens_seen": 719126528
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005255296009078873,
      "loss": 2.7132,
      "theoretical_loss": 3.7691152157193586,
      "tokens_seen": 719192064
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005255130508795158,
      "loss": 2.752,
      "theoretical_loss": 3.76908051201114,
      "tokens_seen": 719257600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005254965008511444,
      "loss": 2.7683,
      "theoretical_loss": 3.769045812350134,
      "tokens_seen": 719323136
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005254799508227728,
      "loss": 2.8218,
      "theoretical_loss": 3.7690111167355,
      "tokens_seen": 719388672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005254634007944014,
      "loss": 2.6031,
      "theoretical_loss": 3.768976425166397,
      "tokens_seen": 719454208
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005254468507660298,
      "loss": 2.7192,
      "theoretical_loss": 3.7689417376419856,
      "tokens_seen": 719519744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005254303007376583,
      "loss": 2.7373,
      "theoretical_loss": 3.768907054161425,
      "tokens_seen": 719585280
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005254137507092869,
      "loss": 2.757,
      "theoretical_loss": 3.768872374723876,
      "tokens_seen": 719650816
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005253972006809154,
      "loss": 2.7258,
      "theoretical_loss": 3.7688376993284995,
      "tokens_seen": 719716352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000525380650652544,
      "loss": 2.8086,
      "theoretical_loss": 3.768803027974456,
      "tokens_seen": 719781888
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005253641006241725,
      "loss": 2.6162,
      "theoretical_loss": 3.768768360660907,
      "tokens_seen": 719847424
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000525347550595801,
      "loss": 2.5971,
      "theoretical_loss": 3.768733697387013,
      "tokens_seen": 719912960
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005253310005674295,
      "loss": 2.636,
      "theoretical_loss": 3.7686990381519365,
      "tokens_seen": 719978496
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000525314450539058,
      "loss": 2.4787,
      "theoretical_loss": 3.7686643829548387,
      "tokens_seen": 720044032
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005252979005106865,
      "loss": 2.5872,
      "theoretical_loss": 3.768629731794882,
      "tokens_seen": 720109568
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005252813504823151,
      "loss": 2.6925,
      "theoretical_loss": 3.7685950846712286,
      "tokens_seen": 720175104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005252648004539436,
      "loss": 2.5832,
      "theoretical_loss": 3.768560441583041,
      "tokens_seen": 720240640
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005252482504255722,
      "loss": 2.8046,
      "theoretical_loss": 3.7685258025294823,
      "tokens_seen": 720306176
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005252317003972007,
      "loss": 2.8552,
      "theoretical_loss": 3.768491167509716,
      "tokens_seen": 720371712
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005252151503688291,
      "loss": 2.7942,
      "theoretical_loss": 3.768456536522905,
      "tokens_seen": 720437248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005251986003404577,
      "loss": 2.7162,
      "theoretical_loss": 3.768421909568212,
      "tokens_seen": 720502784
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005251820503120862,
      "loss": 2.6311,
      "theoretical_loss": 3.768387286644802,
      "tokens_seen": 720568320
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005251655002837147,
      "loss": 3.0294,
      "theoretical_loss": 3.768352667751839,
      "tokens_seen": 720633856
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005251489502553433,
      "loss": 2.7749,
      "theoretical_loss": 3.768318052888487,
      "tokens_seen": 720699392
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005251324002269718,
      "loss": 2.9186,
      "theoretical_loss": 3.768283442053911,
      "tokens_seen": 720764928
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005251158501986004,
      "loss": 2.5942,
      "theoretical_loss": 3.7682488352472756,
      "tokens_seen": 720830464
    },
    {
      "epoch": 0.26,
      "objective/train/avg_token_score": 0.015001632273197174,
      "objective/train/avg_weight": 0.9823410511016846,
      "objective/train/docs_used": 415834,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.848790407180786,
      "objective/train/std_weight": 0.08708380162715912,
      "objective/train/theoretical_loss": 3.7682142324677455,
      "objective/train/tokens_used": 741356000,
      "theoretical_loss": 3.7682142324677455,
      "tokens_seen": 720896000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005250993001702289,
      "loss": 2.6315,
      "theoretical_loss": 3.7682142324677455,
      "tokens_seen": 720896000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005250827501418573,
      "loss": 2.6571,
      "theoretical_loss": 3.7681796337144866,
      "tokens_seen": 720961536
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005250662001134859,
      "loss": 2.7513,
      "theoretical_loss": 3.768145038986664,
      "tokens_seen": 721027072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005250496500851144,
      "loss": 2.9361,
      "theoretical_loss": 3.768110448283444,
      "tokens_seen": 721092608
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005250331000567429,
      "loss": 2.7511,
      "theoretical_loss": 3.768075861603993,
      "tokens_seen": 721158144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005250165500283715,
      "loss": 2.8523,
      "theoretical_loss": 3.7680412789474764,
      "tokens_seen": 721223680
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000525,
      "loss": 2.9038,
      "theoretical_loss": 3.7680067003130615,
      "tokens_seen": 721289216
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005249834499716286,
      "loss": 3.0006,
      "theoretical_loss": 3.767972125699915,
      "tokens_seen": 721354752
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000524966899943257,
      "loss": 2.6178,
      "theoretical_loss": 3.767937555107203,
      "tokens_seen": 721420288
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005249503499148855,
      "loss": 2.8723,
      "theoretical_loss": 3.767902988534094,
      "tokens_seen": 721485824
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005249337998865141,
      "loss": 2.8167,
      "theoretical_loss": 3.7678684259797555,
      "tokens_seen": 721551360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005249172498581426,
      "loss": 2.805,
      "theoretical_loss": 3.767833867443355,
      "tokens_seen": 721616896
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005249006998297712,
      "loss": 2.8112,
      "theoretical_loss": 3.767799312924061,
      "tokens_seen": 721682432
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005248841498013997,
      "loss": 2.6443,
      "theoretical_loss": 3.7677647624210406,
      "tokens_seen": 721747968
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005248675997730282,
      "loss": 2.8072,
      "theoretical_loss": 3.7677302159334642,
      "tokens_seen": 721813504
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005248510497446568,
      "loss": 2.7321,
      "theoretical_loss": 3.7676956734604987,
      "tokens_seen": 721879040
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005248344997162852,
      "loss": 2.9585,
      "theoretical_loss": 3.7676611350013145,
      "tokens_seen": 721944576
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005248179496879137,
      "loss": 2.9869,
      "theoretical_loss": 3.7676266005550803,
      "tokens_seen": 722010112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005248013996595423,
      "loss": 2.7047,
      "theoretical_loss": 3.7675920701209664,
      "tokens_seen": 722075648
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005247848496311708,
      "loss": 2.9715,
      "theoretical_loss": 3.7675575436981417,
      "tokens_seen": 722141184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005247682996027994,
      "loss": 2.7358,
      "theoretical_loss": 3.767523021285777,
      "tokens_seen": 722206720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005247517495744279,
      "loss": 2.6759,
      "theoretical_loss": 3.7674885028830416,
      "tokens_seen": 722272256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005247351995460564,
      "loss": 2.7871,
      "theoretical_loss": 3.767453988489107,
      "tokens_seen": 722337792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005247186495176849,
      "loss": 2.9233,
      "theoretical_loss": 3.767419478103144,
      "tokens_seen": 722403328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005247020994893134,
      "loss": 2.5435,
      "theoretical_loss": 3.7673849717243226,
      "tokens_seen": 722468864
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005246855494609419,
      "loss": 2.74,
      "theoretical_loss": 3.7673504693518156,
      "tokens_seen": 722534400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005246689994325705,
      "loss": 2.8674,
      "theoretical_loss": 3.767315970984793,
      "tokens_seen": 722599936
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000524652449404199,
      "loss": 2.6785,
      "theoretical_loss": 3.7672814766224274,
      "tokens_seen": 722665472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005246358993758276,
      "loss": 2.8324,
      "theoretical_loss": 3.767246986263891,
      "tokens_seen": 722731008
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005246193493474561,
      "loss": 2.7769,
      "theoretical_loss": 3.7672124999083554,
      "tokens_seen": 722796544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005246027993190845,
      "loss": 2.9256,
      "theoretical_loss": 3.767178017554994,
      "tokens_seen": 722862080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005245862492907131,
      "loss": 2.8464,
      "theoretical_loss": 3.7671435392029786,
      "tokens_seen": 722927616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005245696992623416,
      "loss": 2.9349,
      "theoretical_loss": 3.767109064851483,
      "tokens_seen": 722993152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005245531492339701,
      "loss": 2.9057,
      "theoretical_loss": 3.7670745944996797,
      "tokens_seen": 723058688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005245365992055987,
      "loss": 2.8269,
      "theoretical_loss": 3.7670401281467427,
      "tokens_seen": 723124224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005245200491772272,
      "loss": 2.8682,
      "theoretical_loss": 3.767005665791846,
      "tokens_seen": 723189760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005245034991488557,
      "loss": 2.7927,
      "theoretical_loss": 3.766971207434163,
      "tokens_seen": 723255296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005244869491204841,
      "loss": 2.7973,
      "theoretical_loss": 3.766936753072869,
      "tokens_seen": 723320832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005244703990921127,
      "loss": 2.7671,
      "theoretical_loss": 3.7669023027071367,
      "tokens_seen": 723386368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005244538490637412,
      "loss": 2.7824,
      "theoretical_loss": 3.766867856336142,
      "tokens_seen": 723451904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005244372990353697,
      "loss": 2.6486,
      "theoretical_loss": 3.76683341395906,
      "tokens_seen": 723517440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005244207490069983,
      "loss": 2.8234,
      "theoretical_loss": 3.7667989755750657,
      "tokens_seen": 723582976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005244041989786268,
      "loss": 2.7564,
      "theoretical_loss": 3.766764541183334,
      "tokens_seen": 723648512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005243876489502553,
      "loss": 2.8208,
      "theoretical_loss": 3.7667301107830413,
      "tokens_seen": 723714048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005243710989218839,
      "loss": 3.0146,
      "theoretical_loss": 3.766695684373363,
      "tokens_seen": 723779584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005243545488935123,
      "loss": 2.8931,
      "theoretical_loss": 3.766661261953476,
      "tokens_seen": 723845120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005243379988651409,
      "loss": 2.7798,
      "theoretical_loss": 3.7666268435225563,
      "tokens_seen": 723910656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005243214488367694,
      "loss": 2.6306,
      "theoretical_loss": 3.7665924290797808,
      "tokens_seen": 723976192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005243048988083979,
      "loss": 2.7672,
      "theoretical_loss": 3.766558018624326,
      "tokens_seen": 724041728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005242883487800265,
      "loss": 2.7184,
      "theoretical_loss": 3.7665236121553693,
      "tokens_seen": 724107264
    },
    {
      "epoch": 0.26,
      "objective/train/avg_token_score": 0.017583779990673065,
      "objective/train/avg_weight": 0.9793280959129333,
      "objective/train/docs_used": 417692,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1279635429382324,
      "objective/train/std_weight": 0.09767508506774902,
      "objective/train/theoretical_loss": 3.7664892096720886,
      "objective/train/tokens_used": 744632800,
      "theoretical_loss": 3.7664892096720886,
      "tokens_seen": 724172800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000524271798751655,
      "loss": 3.1462,
      "theoretical_loss": 3.7664892096720886,
      "tokens_seen": 724172800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005242552487232835,
      "loss": 2.8037,
      "theoretical_loss": 3.766454811173661,
      "tokens_seen": 724238336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000524238698694912,
      "loss": 2.5846,
      "theoretical_loss": 3.7664204166592645,
      "tokens_seen": 724303872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005242221486665405,
      "loss": 2.8708,
      "theoretical_loss": 3.7663860261280773,
      "tokens_seen": 724369408
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005242055986381691,
      "loss": 2.4323,
      "theoretical_loss": 3.7663516395792778,
      "tokens_seen": 724434944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005241890486097976,
      "loss": 2.6534,
      "theoretical_loss": 3.7663172570120445,
      "tokens_seen": 724500480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005241724985814261,
      "loss": 2.8771,
      "theoretical_loss": 3.7662828784255566,
      "tokens_seen": 724566016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005241559485530547,
      "loss": 3.0063,
      "theoretical_loss": 3.766248503818993,
      "tokens_seen": 724631552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005241393985246832,
      "loss": 2.8967,
      "theoretical_loss": 3.766214133191533,
      "tokens_seen": 724697088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005241228484963116,
      "loss": 2.7987,
      "theoretical_loss": 3.7661797665423564,
      "tokens_seen": 724762624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005241062984679402,
      "loss": 2.8241,
      "theoretical_loss": 3.766145403870643,
      "tokens_seen": 724828160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005240897484395687,
      "loss": 2.6387,
      "theoretical_loss": 3.766111045175573,
      "tokens_seen": 724893696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005240731984111973,
      "loss": 2.7872,
      "theoretical_loss": 3.7660766904563268,
      "tokens_seen": 724959232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005240566483828258,
      "loss": 2.7741,
      "theoretical_loss": 3.7660423397120844,
      "tokens_seen": 725024768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005240400983544543,
      "loss": 2.7933,
      "theoretical_loss": 3.7660079929420274,
      "tokens_seen": 725090304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005240235483260829,
      "loss": 2.695,
      "theoretical_loss": 3.765973650145336,
      "tokens_seen": 725155840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005240069982977114,
      "loss": 2.7986,
      "theoretical_loss": 3.7659393113211923,
      "tokens_seen": 725221376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005239904482693398,
      "loss": 2.7449,
      "theoretical_loss": 3.7659049764687778,
      "tokens_seen": 725286912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005239738982409684,
      "loss": 2.8704,
      "theoretical_loss": 3.765870645587274,
      "tokens_seen": 725352448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005239573482125969,
      "loss": 2.6,
      "theoretical_loss": 3.7658363186758628,
      "tokens_seen": 725417984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005239407981842255,
      "loss": 2.8813,
      "theoretical_loss": 3.765801995733727,
      "tokens_seen": 725483520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000523924248155854,
      "loss": 2.8873,
      "theoretical_loss": 3.7657676767600488,
      "tokens_seen": 725549056
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005239076981274825,
      "loss": 2.8186,
      "theoretical_loss": 3.7657333617540107,
      "tokens_seen": 725614592
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005238911480991111,
      "loss": 2.7561,
      "theoretical_loss": 3.7656990507147965,
      "tokens_seen": 725680128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005238745980707395,
      "loss": 2.7194,
      "theoretical_loss": 3.765664743641589,
      "tokens_seen": 725745664
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000523858048042368,
      "loss": 2.8557,
      "theoretical_loss": 3.765630440533571,
      "tokens_seen": 725811200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005238414980139966,
      "loss": 2.6093,
      "theoretical_loss": 3.7655961413899277,
      "tokens_seen": 725876736
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005238249479856251,
      "loss": 2.7792,
      "theoretical_loss": 3.765561846209842,
      "tokens_seen": 725942272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005238083979572537,
      "loss": 2.8306,
      "theoretical_loss": 3.7655275549924987,
      "tokens_seen": 726007808
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005237918479288822,
      "loss": 2.7264,
      "theoretical_loss": 3.7654932677370816,
      "tokens_seen": 726073344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005237752979005107,
      "loss": 2.532,
      "theoretical_loss": 3.765458984442776,
      "tokens_seen": 726138880
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005237587478721393,
      "loss": 2.8673,
      "theoretical_loss": 3.7654247051087664,
      "tokens_seen": 726204416
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005237421978437677,
      "loss": 2.6066,
      "theoretical_loss": 3.765390429734238,
      "tokens_seen": 726269952
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005237256478153962,
      "loss": 2.7949,
      "theoretical_loss": 3.765356158318377,
      "tokens_seen": 726335488
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005237090977870248,
      "loss": 2.6672,
      "theoretical_loss": 3.765321890860368,
      "tokens_seen": 726401024
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005236925477586533,
      "loss": 2.4266,
      "theoretical_loss": 3.7652876273593976,
      "tokens_seen": 726466560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005236759977302819,
      "loss": 2.522,
      "theoretical_loss": 3.765253367814652,
      "tokens_seen": 726532096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005236594477019104,
      "loss": 2.7682,
      "theoretical_loss": 3.765219112225317,
      "tokens_seen": 726597632
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005236428976735389,
      "loss": 2.6872,
      "theoretical_loss": 3.7651848605905798,
      "tokens_seen": 726663168
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005236263476451674,
      "loss": 2.6072,
      "theoretical_loss": 3.7651506129096273,
      "tokens_seen": 726728704
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005236097976167959,
      "loss": 2.7553,
      "theoretical_loss": 3.765116369181646,
      "tokens_seen": 726794240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005235932475884244,
      "loss": 2.5508,
      "theoretical_loss": 3.7650821294058234,
      "tokens_seen": 726859776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000523576697560053,
      "loss": 2.8278,
      "theoretical_loss": 3.7650478935813476,
      "tokens_seen": 726925312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005235601475316815,
      "loss": 2.7553,
      "theoretical_loss": 3.7650136617074064,
      "tokens_seen": 726990848
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005235435975033101,
      "loss": 2.8471,
      "theoretical_loss": 3.7649794337831874,
      "tokens_seen": 727056384
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005235270474749386,
      "loss": 2.8125,
      "theoretical_loss": 3.7649452098078795,
      "tokens_seen": 727121920
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000523510497446567,
      "loss": 2.7308,
      "theoretical_loss": 3.7649109897806703,
      "tokens_seen": 727187456
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005234939474181955,
      "loss": 3.0895,
      "theoretical_loss": 3.7648767737007494,
      "tokens_seen": 727252992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000523477397389824,
      "loss": 2.8056,
      "theoretical_loss": 3.7648425615673053,
      "tokens_seen": 727318528
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005234608473614526,
      "loss": 2.8365,
      "theoretical_loss": 3.7648083533795282,
      "tokens_seen": 727384064
    },
    {
      "epoch": 0.26,
      "objective/train/avg_token_score": 0.014879835769534111,
      "objective/train/avg_weight": 0.9825495481491089,
      "objective/train/docs_used": 419961,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.4176673889160156,
      "objective/train/std_weight": 0.09299308806657791,
      "objective/train/theoretical_loss": 3.7647741491366067,
      "objective/train/tokens_used": 747909600,
      "theoretical_loss": 3.7647741491366067,
      "tokens_seen": 727449600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005234442973330811,
      "loss": 2.8506,
      "theoretical_loss": 3.7647741491366067,
      "tokens_seen": 727449600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005234277473047097,
      "loss": 2.6843,
      "theoretical_loss": 3.764739948837731,
      "tokens_seen": 727515136
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005234111972763382,
      "loss": 2.9497,
      "theoretical_loss": 3.764705752482091,
      "tokens_seen": 727580672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005233946472479666,
      "loss": 2.6243,
      "theoretical_loss": 3.7646715600688765,
      "tokens_seen": 727646208
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005233780972195952,
      "loss": 2.6817,
      "theoretical_loss": 3.7646373715972787,
      "tokens_seen": 727711744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005233615471912237,
      "loss": 2.7541,
      "theoretical_loss": 3.7646031870664878,
      "tokens_seen": 727777280
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005233449971628522,
      "loss": 2.7496,
      "theoretical_loss": 3.764569006475695,
      "tokens_seen": 727842816
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005233284471344808,
      "loss": 2.8695,
      "theoretical_loss": 3.7645348298240915,
      "tokens_seen": 727908352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005233118971061093,
      "loss": 2.5241,
      "theoretical_loss": 3.7645006571108683,
      "tokens_seen": 727973888
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005232953470777379,
      "loss": 2.7743,
      "theoretical_loss": 3.7644664883352172,
      "tokens_seen": 728039424
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005232787970493664,
      "loss": 2.8512,
      "theoretical_loss": 3.764432323496331,
      "tokens_seen": 728104960
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005232622470209948,
      "loss": 3.0436,
      "theoretical_loss": 3.764398162593401,
      "tokens_seen": 728170496
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005232456969926234,
      "loss": 2.631,
      "theoretical_loss": 3.764364005625619,
      "tokens_seen": 728236032
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005232291469642519,
      "loss": 2.9728,
      "theoretical_loss": 3.764329852592179,
      "tokens_seen": 728301568
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005232125969358804,
      "loss": 2.6639,
      "theoretical_loss": 3.764295703492273,
      "tokens_seen": 728367104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000523196046907509,
      "loss": 2.6079,
      "theoretical_loss": 3.764261558325094,
      "tokens_seen": 728432640
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005231794968791375,
      "loss": 2.6853,
      "theoretical_loss": 3.764227417089836,
      "tokens_seen": 728498176
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005231629468507661,
      "loss": 2.8011,
      "theoretical_loss": 3.764193279785692,
      "tokens_seen": 728563712
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005231463968223945,
      "loss": 2.7831,
      "theoretical_loss": 3.7641591464118562,
      "tokens_seen": 728629248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000523129846794023,
      "loss": 2.9188,
      "theoretical_loss": 3.7641250169675224,
      "tokens_seen": 728694784
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005231132967656516,
      "loss": 2.4778,
      "theoretical_loss": 3.7640908914518847,
      "tokens_seen": 728760320
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005230967467372801,
      "loss": 2.7833,
      "theoretical_loss": 3.764056769864138,
      "tokens_seen": 728825856
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005230801967089086,
      "loss": 2.7283,
      "theoretical_loss": 3.7640226522034768,
      "tokens_seen": 728891392
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005230636466805372,
      "loss": 2.8422,
      "theoretical_loss": 3.7639885384690963,
      "tokens_seen": 728956928
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005230470966521657,
      "loss": 2.7193,
      "theoretical_loss": 3.763954428660192,
      "tokens_seen": 729022464
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005230305466237943,
      "loss": 2.8004,
      "theoretical_loss": 3.7639203227759586,
      "tokens_seen": 729088000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005230139965954227,
      "loss": 2.6396,
      "theoretical_loss": 3.763886220815592,
      "tokens_seen": 729153536
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005229974465670512,
      "loss": 2.7333,
      "theoretical_loss": 3.763852122778289,
      "tokens_seen": 729219072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005229808965386798,
      "loss": 2.7502,
      "theoretical_loss": 3.7638180286632443,
      "tokens_seen": 729284608
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005229643465103083,
      "loss": 2.7642,
      "theoretical_loss": 3.763783938469656,
      "tokens_seen": 729350144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005229477964819368,
      "loss": 2.8442,
      "theoretical_loss": 3.763749852196719,
      "tokens_seen": 729415680
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005229312464535654,
      "loss": 2.56,
      "theoretical_loss": 3.763715769843632,
      "tokens_seen": 729481216
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005229146964251939,
      "loss": 2.9524,
      "theoretical_loss": 3.763681691409591,
      "tokens_seen": 729546752
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005228981463968224,
      "loss": 3.0134,
      "theoretical_loss": 3.7636476168937936,
      "tokens_seen": 729612288
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005228815963684509,
      "loss": 2.7644,
      "theoretical_loss": 3.7636135462954368,
      "tokens_seen": 729677824
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005228650463400794,
      "loss": 2.742,
      "theoretical_loss": 3.76357947961372,
      "tokens_seen": 729743360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000522848496311708,
      "loss": 2.7685,
      "theoretical_loss": 3.76354541684784,
      "tokens_seen": 729808896
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005228319462833365,
      "loss": 2.8807,
      "theoretical_loss": 3.763511357996995,
      "tokens_seen": 729874432
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000522815396254965,
      "loss": 2.6662,
      "theoretical_loss": 3.7634773030603847,
      "tokens_seen": 729939968
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005227988462265936,
      "loss": 2.7323,
      "theoretical_loss": 3.7634432520372068,
      "tokens_seen": 730005504
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000522782296198222,
      "loss": 2.5815,
      "theoretical_loss": 3.7634092049266608,
      "tokens_seen": 730071040
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005227657461698506,
      "loss": 2.7764,
      "theoretical_loss": 3.763375161727946,
      "tokens_seen": 730136576
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005227491961414791,
      "loss": 3.0243,
      "theoretical_loss": 3.7633411224402615,
      "tokens_seen": 730202112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005227326461131076,
      "loss": 2.8238,
      "theoretical_loss": 3.7633070870628074,
      "tokens_seen": 730267648
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005227160960847362,
      "loss": 2.7046,
      "theoretical_loss": 3.7632730555947833,
      "tokens_seen": 730333184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005226995460563647,
      "loss": 2.7037,
      "theoretical_loss": 3.76323902803539,
      "tokens_seen": 730398720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005226829960279932,
      "loss": 2.9339,
      "theoretical_loss": 3.763205004383827,
      "tokens_seen": 730464256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005226664459996218,
      "loss": 2.878,
      "theoretical_loss": 3.7631709846392964,
      "tokens_seen": 730529792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005226498959712502,
      "loss": 2.6746,
      "theoretical_loss": 3.763136968800998,
      "tokens_seen": 730595328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005226333459428788,
      "loss": 2.7449,
      "theoretical_loss": 3.7631029568681327,
      "tokens_seen": 730660864
    },
    {
      "epoch": 0.26,
      "objective/train/avg_token_score": 0.026552706956863403,
      "objective/train/avg_weight": 0.968758225440979,
      "objective/train/docs_used": 421797,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.105583667755127,
      "objective/train/std_weight": 0.13363923132419586,
      "objective/train/theoretical_loss": 3.7630689488399027,
      "objective/train/tokens_used": 751186400,
      "theoretical_loss": 3.7630689488399027,
      "tokens_seen": 730726400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005226167959145073,
      "loss": 2.861,
      "theoretical_loss": 3.7630689488399027,
      "tokens_seen": 730726400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005226002458861358,
      "loss": 2.5792,
      "theoretical_loss": 3.763034944715509,
      "tokens_seen": 730791936
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005225836958577644,
      "loss": 2.7951,
      "theoretical_loss": 3.763000944494154,
      "tokens_seen": 730857472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005225671458293929,
      "loss": 2.7467,
      "theoretical_loss": 3.7629669481750394,
      "tokens_seen": 730923008
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005225505958010214,
      "loss": 2.6896,
      "theoretical_loss": 3.7629329557573676,
      "tokens_seen": 730988544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005225340457726498,
      "loss": 2.4998,
      "theoretical_loss": 3.762898967240341,
      "tokens_seen": 731054080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005225174957442783,
      "loss": 2.8829,
      "theoretical_loss": 3.762864982623163,
      "tokens_seen": 731119616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005225009457159069,
      "loss": 2.7057,
      "theoretical_loss": 3.762831001905036,
      "tokens_seen": 731185152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005224843956875354,
      "loss": 2.6043,
      "theoretical_loss": 3.7627970250851632,
      "tokens_seen": 731250688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000522467845659164,
      "loss": 2.5856,
      "theoretical_loss": 3.7627630521627484,
      "tokens_seen": 731316224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005224512956307925,
      "loss": 2.6191,
      "theoretical_loss": 3.7627290831369953,
      "tokens_seen": 731381760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000522434745602421,
      "loss": 2.7422,
      "theoretical_loss": 3.762695118007107,
      "tokens_seen": 731447296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005224181955740495,
      "loss": 2.465,
      "theoretical_loss": 3.7626611567722894,
      "tokens_seen": 731512832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000522401645545678,
      "loss": 2.7718,
      "theoretical_loss": 3.7626271994317455,
      "tokens_seen": 731578368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005223850955173066,
      "loss": 2.5492,
      "theoretical_loss": 3.7625932459846805,
      "tokens_seen": 731643904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005223685454889351,
      "loss": 2.6849,
      "theoretical_loss": 3.7625592964302994,
      "tokens_seen": 731709440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005223519954605636,
      "loss": 2.8482,
      "theoretical_loss": 3.762525350767807,
      "tokens_seen": 731774976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005223354454321922,
      "loss": 2.6927,
      "theoretical_loss": 3.762491408996409,
      "tokens_seen": 731840512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005223188954038207,
      "loss": 2.581,
      "theoretical_loss": 3.7624574711153107,
      "tokens_seen": 731906048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005223023453754491,
      "loss": 2.8581,
      "theoretical_loss": 3.7624235371237176,
      "tokens_seen": 731971584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005222857953470777,
      "loss": 2.7941,
      "theoretical_loss": 3.7623896070208365,
      "tokens_seen": 732037120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005222692453187062,
      "loss": 2.6712,
      "theoretical_loss": 3.7623556808058733,
      "tokens_seen": 732102656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005222526952903348,
      "loss": 2.739,
      "theoretical_loss": 3.762321758478034,
      "tokens_seen": 732168192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005222361452619633,
      "loss": 2.8825,
      "theoretical_loss": 3.762287840036527,
      "tokens_seen": 732233728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005222195952335918,
      "loss": 2.618,
      "theoretical_loss": 3.7622539254805574,
      "tokens_seen": 732299264
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005222030452052204,
      "loss": 2.8746,
      "theoretical_loss": 3.7622200148093334,
      "tokens_seen": 732364800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005221864951768489,
      "loss": 2.7945,
      "theoretical_loss": 3.7621861080220627,
      "tokens_seen": 732430336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005221699451484773,
      "loss": 2.6478,
      "theoretical_loss": 3.7621522051179523,
      "tokens_seen": 732495872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005221533951201059,
      "loss": 2.6533,
      "theoretical_loss": 3.7621183060962102,
      "tokens_seen": 732561408
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005221368450917344,
      "loss": 2.8677,
      "theoretical_loss": 3.762084410956045,
      "tokens_seen": 732626944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000522120295063363,
      "loss": 2.8737,
      "theoretical_loss": 3.7620505196966647,
      "tokens_seen": 732692480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005221037450349915,
      "loss": 2.7313,
      "theoretical_loss": 3.762016632317278,
      "tokens_seen": 732758016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00052208719500662,
      "loss": 2.7702,
      "theoretical_loss": 3.761982748817094,
      "tokens_seen": 732823552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005220706449782486,
      "loss": 2.7224,
      "theoretical_loss": 3.7619488691953213,
      "tokens_seen": 732889088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000522054094949877,
      "loss": 2.7117,
      "theoretical_loss": 3.7619149934511693,
      "tokens_seen": 732954624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005220375449215055,
      "loss": 2.8024,
      "theoretical_loss": 3.761881121583848,
      "tokens_seen": 733020160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005220209948931341,
      "loss": 2.7212,
      "theoretical_loss": 3.761847253592567,
      "tokens_seen": 733085696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005220044448647626,
      "loss": 2.81,
      "theoretical_loss": 3.7618133894765364,
      "tokens_seen": 733151232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005219878948363912,
      "loss": 2.7191,
      "theoretical_loss": 3.761779529234966,
      "tokens_seen": 733216768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005219713448080197,
      "loss": 2.7861,
      "theoretical_loss": 3.761745672867066,
      "tokens_seen": 733282304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005219547947796482,
      "loss": 2.7423,
      "theoretical_loss": 3.7617118203720485,
      "tokens_seen": 733347840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005219382447512768,
      "loss": 2.8432,
      "theoretical_loss": 3.7616779717491227,
      "tokens_seen": 733413376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005219216947229052,
      "loss": 2.7463,
      "theoretical_loss": 3.761644126997501,
      "tokens_seen": 733478912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005219051446945337,
      "loss": 2.9515,
      "theoretical_loss": 3.7616102861163947,
      "tokens_seen": 733544448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005218885946661623,
      "loss": 2.8057,
      "theoretical_loss": 3.7615764491050143,
      "tokens_seen": 733609984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005218720446377908,
      "loss": 2.7574,
      "theoretical_loss": 3.761542615962573,
      "tokens_seen": 733675520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005218554946094194,
      "loss": 2.6377,
      "theoretical_loss": 3.7615087866882826,
      "tokens_seen": 733741056
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005218389445810479,
      "loss": 2.8476,
      "theoretical_loss": 3.7614749612813547,
      "tokens_seen": 733806592
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005218223945526764,
      "loss": 2.9445,
      "theoretical_loss": 3.761441139741003,
      "tokens_seen": 733872128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000521805844524305,
      "loss": 2.7269,
      "theoretical_loss": 3.7614073220664395,
      "tokens_seen": 733937664
    },
    {
      "epoch": 0.26,
      "objective/train/avg_token_score": 0.027903227135539055,
      "objective/train/avg_weight": 0.9671719074249268,
      "objective/train/docs_used": 423825,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.264462471008301,
      "objective/train/std_weight": 0.13173261284828186,
      "objective/train/theoretical_loss": 3.7613735082568764,
      "objective/train/tokens_used": 754463200,
      "theoretical_loss": 3.7613735082568764,
      "tokens_seen": 734003200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005217892944959334,
      "loss": 2.7871,
      "theoretical_loss": 3.7613735082568764,
      "tokens_seen": 734003200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005217727444675619,
      "loss": 3.039,
      "theoretical_loss": 3.7613396983115286,
      "tokens_seen": 734068736
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005217561944391905,
      "loss": 2.6624,
      "theoretical_loss": 3.761305892229609,
      "tokens_seen": 734134272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000521739644410819,
      "loss": 2.7477,
      "theoretical_loss": 3.7612720900103307,
      "tokens_seen": 734199808
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005217230943824476,
      "loss": 2.7314,
      "theoretical_loss": 3.7612382916529086,
      "tokens_seen": 734265344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005217065443540761,
      "loss": 2.7255,
      "theoretical_loss": 3.7612044971565557,
      "tokens_seen": 734330880
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005216899943257045,
      "loss": 2.875,
      "theoretical_loss": 3.7611707065204874,
      "tokens_seen": 734396416
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005216734442973331,
      "loss": 2.7832,
      "theoretical_loss": 3.761136919743918,
      "tokens_seen": 734461952
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005216568942689616,
      "loss": 2.724,
      "theoretical_loss": 3.761103136826062,
      "tokens_seen": 734527488
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005216403442405901,
      "loss": 2.6233,
      "theoretical_loss": 3.761069357766135,
      "tokens_seen": 734593024
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005216237942122187,
      "loss": 2.7412,
      "theoretical_loss": 3.761035582563352,
      "tokens_seen": 734658560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005216072441838472,
      "loss": 2.7921,
      "theoretical_loss": 3.7610018112169286,
      "tokens_seen": 734724096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005215906941554758,
      "loss": 2.8359,
      "theoretical_loss": 3.760968043726081,
      "tokens_seen": 734789632
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005215741441271043,
      "loss": 2.8945,
      "theoretical_loss": 3.7609342800900247,
      "tokens_seen": 734855168
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005215575940987327,
      "loss": 3.0489,
      "theoretical_loss": 3.7609005203079757,
      "tokens_seen": 734920704
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005215410440703612,
      "loss": 2.9476,
      "theoretical_loss": 3.760866764379151,
      "tokens_seen": 734986240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005215244940419897,
      "loss": 2.7878,
      "theoretical_loss": 3.7608330123027667,
      "tokens_seen": 735051776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005215079440136183,
      "loss": 2.8496,
      "theoretical_loss": 3.7607992640780408,
      "tokens_seen": 735117312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005214913939852468,
      "loss": 2.8054,
      "theoretical_loss": 3.760765519704189,
      "tokens_seen": 735182848
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005214748439568753,
      "loss": 2.6731,
      "theoretical_loss": 3.7607317791804302,
      "tokens_seen": 735248384
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005214582939285039,
      "loss": 2.7844,
      "theoretical_loss": 3.7606980425059806,
      "tokens_seen": 735313920
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005214417439001323,
      "loss": 2.8297,
      "theoretical_loss": 3.7606643096800587,
      "tokens_seen": 735379456
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005214251938717609,
      "loss": 2.8043,
      "theoretical_loss": 3.760630580701883,
      "tokens_seen": 735444992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005214086438433894,
      "loss": 2.8756,
      "theoretical_loss": 3.760596855570671,
      "tokens_seen": 735510528
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005213920938150179,
      "loss": 2.8174,
      "theoretical_loss": 3.7605631342856416,
      "tokens_seen": 735576064
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005213755437866465,
      "loss": 2.9557,
      "theoretical_loss": 3.7605294168460133,
      "tokens_seen": 735641600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000521358993758275,
      "loss": 2.8984,
      "theoretical_loss": 3.7604957032510056,
      "tokens_seen": 735707136
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005213424437299035,
      "loss": 2.6605,
      "theoretical_loss": 3.760461993499837,
      "tokens_seen": 735772672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000521325893701532,
      "loss": 2.6569,
      "theoretical_loss": 3.7604282875917274,
      "tokens_seen": 735838208
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005213093436731605,
      "loss": 2.6799,
      "theoretical_loss": 3.760394585525896,
      "tokens_seen": 735903744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005212927936447891,
      "loss": 2.7218,
      "theoretical_loss": 3.7603608873015633,
      "tokens_seen": 735969280
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005212762436164176,
      "loss": 2.6568,
      "theoretical_loss": 3.7603271929179494,
      "tokens_seen": 736034816
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005212596935880461,
      "loss": 2.6313,
      "theoretical_loss": 3.7602935023742736,
      "tokens_seen": 736100352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005212431435596747,
      "loss": 2.6206,
      "theoretical_loss": 3.760259815669758,
      "tokens_seen": 736165888
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005212265935313032,
      "loss": 2.6899,
      "theoretical_loss": 3.7602261328036226,
      "tokens_seen": 736231424
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005212100435029316,
      "loss": 2.8357,
      "theoretical_loss": 3.760192453775088,
      "tokens_seen": 736296960
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005211934934745602,
      "loss": 2.5985,
      "theoretical_loss": 3.760158778583376,
      "tokens_seen": 736362496
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005211769434461887,
      "loss": 2.7153,
      "theoretical_loss": 3.7601251072277084,
      "tokens_seen": 736428032
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005211603934178173,
      "loss": 2.7148,
      "theoretical_loss": 3.760091439707306,
      "tokens_seen": 736493568
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005211438433894458,
      "loss": 2.8128,
      "theoretical_loss": 3.760057776021392,
      "tokens_seen": 736559104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005211272933610743,
      "loss": 2.8496,
      "theoretical_loss": 3.7600241161691876,
      "tokens_seen": 736624640
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005211107433327029,
      "loss": 2.8359,
      "theoretical_loss": 3.759990460149915,
      "tokens_seen": 736690176
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005210941933043314,
      "loss": 2.898,
      "theoretical_loss": 3.759956807962797,
      "tokens_seen": 736755712
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005210776432759598,
      "loss": 2.5612,
      "theoretical_loss": 3.759923159607057,
      "tokens_seen": 736821248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005210610932475884,
      "loss": 2.5506,
      "theoretical_loss": 3.759889515081918,
      "tokens_seen": 736886784
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005210445432192169,
      "loss": 3.0308,
      "theoretical_loss": 3.759855874386602,
      "tokens_seen": 736952320
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005210279931908455,
      "loss": 2.6898,
      "theoretical_loss": 3.7598222375203347,
      "tokens_seen": 737017856
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000521011443162474,
      "loss": 2.7929,
      "theoretical_loss": 3.7597886044823383,
      "tokens_seen": 737083392
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005209948931341025,
      "loss": 2.8322,
      "theoretical_loss": 3.7597549752718367,
      "tokens_seen": 737148928
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005209783431057311,
      "loss": 2.7062,
      "theoretical_loss": 3.7597213498880553,
      "tokens_seen": 737214464
    },
    {
      "epoch": 0.26,
      "objective/train/avg_token_score": 0.012162305414676666,
      "objective/train/avg_weight": 0.985683262348175,
      "objective/train/docs_used": 425209,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.899125576019287,
      "objective/train/std_weight": 0.0780651792883873,
      "objective/train/theoretical_loss": 3.759687728330217,
      "objective/train/tokens_used": 757740000,
      "theoretical_loss": 3.759687728330217,
      "tokens_seen": 737280000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005209617930773596,
      "loss": 2.7055,
      "theoretical_loss": 3.759687728330217,
      "tokens_seen": 737280000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005209452430489881,
      "loss": 2.9487,
      "theoretical_loss": 3.759654110597548,
      "tokens_seen": 737345536
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005209286930206166,
      "loss": 2.8457,
      "theoretical_loss": 3.759620496689272,
      "tokens_seen": 737411072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005209121429922451,
      "loss": 2.6322,
      "theoretical_loss": 3.7595868866046143,
      "tokens_seen": 737476608
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005208955929638737,
      "loss": 2.6392,
      "theoretical_loss": 3.759553280342801,
      "tokens_seen": 737542144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005208790429355022,
      "loss": 2.505,
      "theoretical_loss": 3.7595196779030564,
      "tokens_seen": 737607680
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005208624929071307,
      "loss": 2.7602,
      "theoretical_loss": 3.759486079284607,
      "tokens_seen": 737673216
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005208459428787593,
      "loss": 2.7084,
      "theoretical_loss": 3.7594524844866797,
      "tokens_seen": 737738752
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005208293928503877,
      "loss": 2.5829,
      "theoretical_loss": 3.7594188935084993,
      "tokens_seen": 737804288
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005208128428220163,
      "loss": 2.7612,
      "theoretical_loss": 3.7593853063492926,
      "tokens_seen": 737869824
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005207962927936448,
      "loss": 2.9355,
      "theoretical_loss": 3.7593517230082867,
      "tokens_seen": 737935360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005207797427652733,
      "loss": 2.6754,
      "theoretical_loss": 3.759318143484708,
      "tokens_seen": 738000896
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005207631927369019,
      "loss": 2.6454,
      "theoretical_loss": 3.759284567777784,
      "tokens_seen": 738066432
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005207466427085304,
      "loss": 2.6719,
      "theoretical_loss": 3.7592509958867426,
      "tokens_seen": 738131968
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005207300926801589,
      "loss": 2.8072,
      "theoretical_loss": 3.75921742781081,
      "tokens_seen": 738197504
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005207135426517875,
      "loss": 2.7298,
      "theoretical_loss": 3.759183863549215,
      "tokens_seen": 738263040
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005206969926234159,
      "loss": 2.657,
      "theoretical_loss": 3.759150303101185,
      "tokens_seen": 738328576
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005206804425950445,
      "loss": 2.854,
      "theoretical_loss": 3.759116746465949,
      "tokens_seen": 738394112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000520663892566673,
      "loss": 2.7727,
      "theoretical_loss": 3.7590831936427356,
      "tokens_seen": 738459648
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005206473425383015,
      "loss": 2.9662,
      "theoretical_loss": 3.759049644630773,
      "tokens_seen": 738525184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005206307925099301,
      "loss": 2.8503,
      "theoretical_loss": 3.7590160994292896,
      "tokens_seen": 738590720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005206142424815586,
      "loss": 2.7923,
      "theoretical_loss": 3.7589825580375154,
      "tokens_seen": 738656256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000520597692453187,
      "loss": 2.7756,
      "theoretical_loss": 3.758949020454679,
      "tokens_seen": 738721792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005205811424248156,
      "loss": 2.8287,
      "theoretical_loss": 3.758915486680011,
      "tokens_seen": 738787328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000520564592396444,
      "loss": 2.7539,
      "theoretical_loss": 3.7588819567127407,
      "tokens_seen": 738852864
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005205480423680726,
      "loss": 2.8836,
      "theoretical_loss": 3.7588484305520984,
      "tokens_seen": 738918400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005205314923397011,
      "loss": 2.7834,
      "theoretical_loss": 3.7588149081973143,
      "tokens_seen": 738983936
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005205149423113297,
      "loss": 2.6719,
      "theoretical_loss": 3.7587813896476185,
      "tokens_seen": 739049472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005204983922829582,
      "loss": 2.6885,
      "theoretical_loss": 3.758747874902242,
      "tokens_seen": 739115008
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005204818422545867,
      "loss": 2.82,
      "theoretical_loss": 3.758714363960416,
      "tokens_seen": 739180544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005204652922262152,
      "loss": 2.8244,
      "theoretical_loss": 3.758680856821371,
      "tokens_seen": 739246080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005204487421978437,
      "loss": 2.829,
      "theoretical_loss": 3.7586473534843394,
      "tokens_seen": 739311616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005204321921694722,
      "loss": 2.9884,
      "theoretical_loss": 3.7586138539485514,
      "tokens_seen": 739377152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005204156421411008,
      "loss": 3.0983,
      "theoretical_loss": 3.7585803582132407,
      "tokens_seen": 739442688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005203990921127293,
      "loss": 2.8232,
      "theoretical_loss": 3.7585468662776376,
      "tokens_seen": 739508224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005203825420843579,
      "loss": 2.722,
      "theoretical_loss": 3.758513378140975,
      "tokens_seen": 739573760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005203659920559864,
      "loss": 2.5825,
      "theoretical_loss": 3.7584798938024866,
      "tokens_seen": 739639296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005203494420276148,
      "loss": 3.0436,
      "theoretical_loss": 3.7584464132614035,
      "tokens_seen": 739704832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005203328919992434,
      "loss": 2.6705,
      "theoretical_loss": 3.758412936516959,
      "tokens_seen": 739770368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005203163419708719,
      "loss": 2.7191,
      "theoretical_loss": 3.758379463568387,
      "tokens_seen": 739835904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005202997919425004,
      "loss": 2.6051,
      "theoretical_loss": 3.7583459944149205,
      "tokens_seen": 739901440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000520283241914129,
      "loss": 2.796,
      "theoretical_loss": 3.7583125290557926,
      "tokens_seen": 739966976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005202666918857575,
      "loss": 2.7509,
      "theoretical_loss": 3.7582790674902378,
      "tokens_seen": 740032512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005202501418573861,
      "loss": 3.1222,
      "theoretical_loss": 3.7582456097174903,
      "tokens_seen": 740098048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005202335918290146,
      "loss": 2.6793,
      "theoretical_loss": 3.758212155736784,
      "tokens_seen": 740163584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000520217041800643,
      "loss": 2.4132,
      "theoretical_loss": 3.758178705547353,
      "tokens_seen": 740229120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005202004917722716,
      "loss": 2.4712,
      "theoretical_loss": 3.7581452591484323,
      "tokens_seen": 740294656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005201839417439001,
      "loss": 2.6646,
      "theoretical_loss": 3.758111816539258,
      "tokens_seen": 740360192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005201673917155286,
      "loss": 2.8221,
      "theoretical_loss": 3.7580783777190643,
      "tokens_seen": 740425728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005201508416871572,
      "loss": 2.8872,
      "theoretical_loss": 3.7580449426870866,
      "tokens_seen": 740491264
    },
    {
      "epoch": 0.26,
      "objective/train/avg_token_score": 0.008828648366034031,
      "objective/train/avg_weight": 0.9896084070205688,
      "objective/train/docs_used": 428014,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6653997898101807,
      "objective/train/std_weight": 0.0541449598968029,
      "objective/train/theoretical_loss": 3.75801151144256,
      "objective/train/tokens_used": 761016800,
      "theoretical_loss": 3.75801151144256,
      "tokens_seen": 740556800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005201342916587857,
      "loss": 2.8538,
      "theoretical_loss": 3.75801151144256,
      "tokens_seen": 740556800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005201177416304143,
      "loss": 2.804,
      "theoretical_loss": 3.757978083984722,
      "tokens_seen": 740622336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005201011916020427,
      "loss": 2.7083,
      "theoretical_loss": 3.7579446603128064,
      "tokens_seen": 740687872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005200846415736712,
      "loss": 2.6515,
      "theoretical_loss": 3.7579112404260515,
      "tokens_seen": 740753408
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005200680915452998,
      "loss": 2.7199,
      "theoretical_loss": 3.7578778243236934,
      "tokens_seen": 740818944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005200515415169283,
      "loss": 2.7613,
      "theoretical_loss": 3.757844412004968,
      "tokens_seen": 740884480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005200349914885568,
      "loss": 2.7951,
      "theoretical_loss": 3.7578110034691123,
      "tokens_seen": 740950016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005200184414601854,
      "loss": 2.9229,
      "theoretical_loss": 3.7577775987153648,
      "tokens_seen": 741015552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005200018914318139,
      "loss": 2.8216,
      "theoretical_loss": 3.7577441977429613,
      "tokens_seen": 741081088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005199853414034425,
      "loss": 2.9189,
      "theoretical_loss": 3.7577108005511404,
      "tokens_seen": 741146624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005199687913750709,
      "loss": 2.8948,
      "theoretical_loss": 3.7576774071391403,
      "tokens_seen": 741212160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005199522413466994,
      "loss": 2.8844,
      "theoretical_loss": 3.7576440175061974,
      "tokens_seen": 741277696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000519935691318328,
      "loss": 2.5612,
      "theoretical_loss": 3.757610631651552,
      "tokens_seen": 741343232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005199191412899565,
      "loss": 2.7105,
      "theoretical_loss": 3.757577249574441,
      "tokens_seen": 741408768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000519902591261585,
      "loss": 2.6941,
      "theoretical_loss": 3.7575438712741036,
      "tokens_seen": 741474304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005198860412332136,
      "loss": 2.9207,
      "theoretical_loss": 3.75751049674978,
      "tokens_seen": 741539840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000519869491204842,
      "loss": 2.8254,
      "theoretical_loss": 3.7574771260007074,
      "tokens_seen": 741605376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005198529411764706,
      "loss": 2.5922,
      "theoretical_loss": 3.7574437590261263,
      "tokens_seen": 741670912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005198363911480991,
      "loss": 2.6559,
      "theoretical_loss": 3.757410395825276,
      "tokens_seen": 741736448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005198198411197276,
      "loss": 2.6145,
      "theoretical_loss": 3.757377036397396,
      "tokens_seen": 741801984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005198032910913562,
      "loss": 2.532,
      "theoretical_loss": 3.7573436807417275,
      "tokens_seen": 741867520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0005197867410629847,
      "loss": 2.9255,
      "theoretical_loss": 3.7573103288575096,
      "tokens_seen": 741933056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005197701910346132,
      "loss": 2.9014,
      "theoretical_loss": 3.757276980743983,
      "tokens_seen": 741998592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005197536410062418,
      "loss": 2.785,
      "theoretical_loss": 3.757243636400389,
      "tokens_seen": 742064128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005197370909778702,
      "loss": 2.7937,
      "theoretical_loss": 3.757210295825968,
      "tokens_seen": 742129664
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005197205409494988,
      "loss": 3.0351,
      "theoretical_loss": 3.7571769590199615,
      "tokens_seen": 742195200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005197039909211273,
      "loss": 2.4556,
      "theoretical_loss": 3.7571436259816107,
      "tokens_seen": 742260736
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005196874408927558,
      "loss": 2.8807,
      "theoretical_loss": 3.757110296710157,
      "tokens_seen": 742326272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005196708908643844,
      "loss": 2.6799,
      "theoretical_loss": 3.7570769712048424,
      "tokens_seen": 742391808
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005196543408360129,
      "loss": 2.8676,
      "theoretical_loss": 3.757043649464909,
      "tokens_seen": 742457344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005196377908076414,
      "loss": 2.5299,
      "theoretical_loss": 3.7570103314895986,
      "tokens_seen": 742522880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00051962124077927,
      "loss": 2.8484,
      "theoretical_loss": 3.756977017278154,
      "tokens_seen": 742588416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005196046907508984,
      "loss": 2.6063,
      "theoretical_loss": 3.7569437068298184,
      "tokens_seen": 742653952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005195881407225269,
      "loss": 2.6767,
      "theoretical_loss": 3.756910400143833,
      "tokens_seen": 742719488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005195715906941554,
      "loss": 2.6071,
      "theoretical_loss": 3.7568770972194434,
      "tokens_seen": 742785024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000519555040665784,
      "loss": 2.8781,
      "theoretical_loss": 3.7568437980558915,
      "tokens_seen": 742850560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005195384906374125,
      "loss": 2.5376,
      "theoretical_loss": 3.756810502652421,
      "tokens_seen": 742916096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000519521940609041,
      "loss": 2.9516,
      "theoretical_loss": 3.7567772110082753,
      "tokens_seen": 742981632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005195053905806696,
      "loss": 2.7474,
      "theoretical_loss": 3.756743923122699,
      "tokens_seen": 743047168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000519488840552298,
      "loss": 2.7495,
      "theoretical_loss": 3.756710638994936,
      "tokens_seen": 743112704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005194722905239266,
      "loss": 2.6871,
      "theoretical_loss": 3.756677358624231,
      "tokens_seen": 743178240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005194557404955551,
      "loss": 2.7685,
      "theoretical_loss": 3.7566440820098284,
      "tokens_seen": 743243776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005194391904671836,
      "loss": 2.6421,
      "theoretical_loss": 3.756610809150973,
      "tokens_seen": 743309312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005194226404388122,
      "loss": 2.8152,
      "theoretical_loss": 3.7565775400469104,
      "tokens_seen": 743374848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005194060904104407,
      "loss": 2.9482,
      "theoretical_loss": 3.7565442746968856,
      "tokens_seen": 743440384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005193895403820692,
      "loss": 2.542,
      "theoretical_loss": 3.7565110131001442,
      "tokens_seen": 743505920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005193729903536977,
      "loss": 2.7155,
      "theoretical_loss": 3.756477755255932,
      "tokens_seen": 743571456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005193564403253262,
      "loss": 2.6085,
      "theoretical_loss": 3.7564445011634944,
      "tokens_seen": 743636992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005193398902969548,
      "loss": 2.7505,
      "theoretical_loss": 3.7564112508220777,
      "tokens_seen": 743702528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005193233402685833,
      "loss": 2.7239,
      "theoretical_loss": 3.756378004230929,
      "tokens_seen": 743768064
    },
    {
      "epoch": 0.27,
      "objective/train/avg_token_score": 0.007486060261726379,
      "objective/train/avg_weight": 0.9911874532699585,
      "objective/train/docs_used": 429315,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7764382362365723,
      "objective/train/std_weight": 0.034586504101753235,
      "objective/train/theoretical_loss": 3.756344761389295,
      "objective/train/tokens_used": 764293600,
      "theoretical_loss": 3.756344761389295,
      "tokens_seen": 743833600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005193067902402118,
      "loss": 2.7918,
      "theoretical_loss": 3.756344761389295,
      "tokens_seen": 743833600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005192902402118404,
      "loss": 2.7635,
      "theoretical_loss": 3.7563115222964214,
      "tokens_seen": 743899136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005192736901834689,
      "loss": 2.7004,
      "theoretical_loss": 3.7562782869515567,
      "tokens_seen": 743964672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005192571401550973,
      "loss": 2.7376,
      "theoretical_loss": 3.756245055353947,
      "tokens_seen": 744030208
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005192405901267259,
      "loss": 2.7647,
      "theoretical_loss": 3.75621182750284,
      "tokens_seen": 744095744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005192240400983544,
      "loss": 2.7939,
      "theoretical_loss": 3.7561786033974833,
      "tokens_seen": 744161280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000519207490069983,
      "loss": 2.5175,
      "theoretical_loss": 3.7561453830371256,
      "tokens_seen": 744226816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005191909400416115,
      "loss": 2.8844,
      "theoretical_loss": 3.7561121664210146,
      "tokens_seen": 744292352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00051917439001324,
      "loss": 2.7343,
      "theoretical_loss": 3.756078953548398,
      "tokens_seen": 744357888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005191578399848686,
      "loss": 2.799,
      "theoretical_loss": 3.756045744418526,
      "tokens_seen": 744423424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000519141289956497,
      "loss": 2.7604,
      "theoretical_loss": 3.7560125390306456,
      "tokens_seen": 744488960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005191247399281255,
      "loss": 2.7985,
      "theoretical_loss": 3.755979337384007,
      "tokens_seen": 744554496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005191081898997541,
      "loss": 2.7465,
      "theoretical_loss": 3.7559461394778584,
      "tokens_seen": 744620032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005190916398713826,
      "loss": 2.9792,
      "theoretical_loss": 3.755912945311451,
      "tokens_seen": 744685568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005190750898430112,
      "loss": 2.5654,
      "theoretical_loss": 3.755879754884032,
      "tokens_seen": 744751104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005190585398146397,
      "loss": 2.9419,
      "theoretical_loss": 3.755846568194853,
      "tokens_seen": 744816640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005190419897862682,
      "loss": 2.7721,
      "theoretical_loss": 3.755813385243164,
      "tokens_seen": 744882176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005190254397578968,
      "loss": 2.6415,
      "theoretical_loss": 3.7557802060282146,
      "tokens_seen": 744947712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005190088897295252,
      "loss": 2.9118,
      "theoretical_loss": 3.755747030549256,
      "tokens_seen": 745013248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005189923397011537,
      "loss": 2.5085,
      "theoretical_loss": 3.7557138588055388,
      "tokens_seen": 745078784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005189757896727823,
      "loss": 2.6128,
      "theoretical_loss": 3.7556806907963134,
      "tokens_seen": 745144320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005189592396444108,
      "loss": 2.6334,
      "theoretical_loss": 3.7556475265208316,
      "tokens_seen": 745209856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005189426896160394,
      "loss": 2.6932,
      "theoretical_loss": 3.7556143659783445,
      "tokens_seen": 745275392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005189261395876679,
      "loss": 2.6856,
      "theoretical_loss": 3.755581209168104,
      "tokens_seen": 745340928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005189095895592964,
      "loss": 2.8248,
      "theoretical_loss": 3.7555480560893617,
      "tokens_seen": 745406464
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000518893039530925,
      "loss": 2.8636,
      "theoretical_loss": 3.7555149067413693,
      "tokens_seen": 745472000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005188764895025534,
      "loss": 2.5972,
      "theoretical_loss": 3.755481761123379,
      "tokens_seen": 745537536
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005188599394741819,
      "loss": 2.6524,
      "theoretical_loss": 3.7554486192346443,
      "tokens_seen": 745603072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005188433894458105,
      "loss": 2.4915,
      "theoretical_loss": 3.7554154810744174,
      "tokens_seen": 745668608
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000518826839417439,
      "loss": 2.8428,
      "theoretical_loss": 3.7553823466419507,
      "tokens_seen": 745734144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005188102893890676,
      "loss": 2.7802,
      "theoretical_loss": 3.755349215936498,
      "tokens_seen": 745799680
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005187937393606961,
      "loss": 2.577,
      "theoretical_loss": 3.755316088957312,
      "tokens_seen": 745865216
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005187771893323246,
      "loss": 2.6168,
      "theoretical_loss": 3.755282965703646,
      "tokens_seen": 745930752
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005187606393039531,
      "loss": 2.4253,
      "theoretical_loss": 3.755249846174755,
      "tokens_seen": 745996288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005187440892755816,
      "loss": 2.8514,
      "theoretical_loss": 3.7552167303698925,
      "tokens_seen": 746061824
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005187275392472101,
      "loss": 2.7457,
      "theoretical_loss": 3.755183618288312,
      "tokens_seen": 746127360
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005187109892188387,
      "loss": 2.7616,
      "theoretical_loss": 3.755150509929268,
      "tokens_seen": 746192896
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005186944391904672,
      "loss": 2.6199,
      "theoretical_loss": 3.7551174052920158,
      "tokens_seen": 746258432
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005186778891620958,
      "loss": 2.7353,
      "theoretical_loss": 3.75508430437581,
      "tokens_seen": 746323968
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005186613391337243,
      "loss": 2.7599,
      "theoretical_loss": 3.755051207179905,
      "tokens_seen": 746389504
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005186447891053527,
      "loss": 2.4247,
      "theoretical_loss": 3.7550181137035574,
      "tokens_seen": 746455040
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005186282390769813,
      "loss": 2.9667,
      "theoretical_loss": 3.7549850239460216,
      "tokens_seen": 746520576
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005186116890486098,
      "loss": 2.4966,
      "theoretical_loss": 3.7549519379065535,
      "tokens_seen": 746586112
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005185951390202383,
      "loss": 2.743,
      "theoretical_loss": 3.7549188555844095,
      "tokens_seen": 746651648
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005185785889918668,
      "loss": 2.6818,
      "theoretical_loss": 3.7548857769788446,
      "tokens_seen": 746717184
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005185620389634953,
      "loss": 2.5031,
      "theoretical_loss": 3.7548527020891163,
      "tokens_seen": 746782720
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005185454889351239,
      "loss": 2.7378,
      "theoretical_loss": 3.754819630914481,
      "tokens_seen": 746848256
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005185289389067523,
      "loss": 2.8225,
      "theoretical_loss": 3.7547865634541946,
      "tokens_seen": 746913792
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005185123888783809,
      "loss": 2.8513,
      "theoretical_loss": 3.7547534997075145,
      "tokens_seen": 746979328
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005184958388500094,
      "loss": 2.7396,
      "theoretical_loss": 3.754720439673698,
      "tokens_seen": 747044864
    },
    {
      "epoch": 0.27,
      "objective/train/avg_token_score": 0.0051964083686470985,
      "objective/train/avg_weight": 0.9938863515853882,
      "objective/train/docs_used": 431156,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.625262498855591,
      "objective/train/std_weight": 0.044133711606264114,
      "objective/train/theoretical_loss": 3.754687383352003,
      "objective/train/tokens_used": 767570400,
      "theoretical_loss": 3.754687383352003,
      "tokens_seen": 747110400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005184792888216379,
      "loss": 2.6966,
      "theoretical_loss": 3.754687383352003,
      "tokens_seen": 747110400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005184627387932665,
      "loss": 2.566,
      "theoretical_loss": 3.754654330741687,
      "tokens_seen": 747175936
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000518446188764895,
      "loss": 2.9993,
      "theoretical_loss": 3.754621281842007,
      "tokens_seen": 747241472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005184296387365236,
      "loss": 2.7286,
      "theoretical_loss": 3.7545882366522214,
      "tokens_seen": 747307008
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005184130887081521,
      "loss": 2.8025,
      "theoretical_loss": 3.7545551951715885,
      "tokens_seen": 747372544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005183965386797805,
      "loss": 2.504,
      "theoretical_loss": 3.754522157399367,
      "tokens_seen": 747438080
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005183799886514091,
      "loss": 2.7554,
      "theoretical_loss": 3.7544891233348157,
      "tokens_seen": 747503616
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005183634386230376,
      "loss": 2.861,
      "theoretical_loss": 3.7544560929771924,
      "tokens_seen": 747569152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005183468885946661,
      "loss": 2.5233,
      "theoretical_loss": 3.7544230663257583,
      "tokens_seen": 747634688
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005183303385662947,
      "loss": 2.8098,
      "theoretical_loss": 3.7543900433797703,
      "tokens_seen": 747700224
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005183137885379232,
      "loss": 2.9054,
      "theoretical_loss": 3.75435702413849,
      "tokens_seen": 747765760
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005182972385095518,
      "loss": 2.7228,
      "theoretical_loss": 3.7543240086011753,
      "tokens_seen": 747831296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005182806884811802,
      "loss": 2.7721,
      "theoretical_loss": 3.7542909967670877,
      "tokens_seen": 747896832
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005182641384528087,
      "loss": 2.7914,
      "theoretical_loss": 3.754257988635487,
      "tokens_seen": 747962368
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005182475884244373,
      "loss": 2.6882,
      "theoretical_loss": 3.754224984205633,
      "tokens_seen": 748027904
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005182310383960658,
      "loss": 2.6408,
      "theoretical_loss": 3.7541919834767867,
      "tokens_seen": 748093440
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005182144883676943,
      "loss": 2.704,
      "theoretical_loss": 3.754158986448209,
      "tokens_seen": 748158976
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005181979383393229,
      "loss": 2.7259,
      "theoretical_loss": 3.7541259931191604,
      "tokens_seen": 748224512
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005181813883109514,
      "loss": 2.8096,
      "theoretical_loss": 3.7540930034889035,
      "tokens_seen": 748290048
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00051816483828258,
      "loss": 2.826,
      "theoretical_loss": 3.7540600175566983,
      "tokens_seen": 748355584
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005181482882542084,
      "loss": 3.0027,
      "theoretical_loss": 3.7540270353218066,
      "tokens_seen": 748421120
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005181317382258369,
      "loss": 2.8251,
      "theoretical_loss": 3.7539940567834913,
      "tokens_seen": 748486656
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005181151881974655,
      "loss": 2.7202,
      "theoretical_loss": 3.753961081941013,
      "tokens_seen": 748552192
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000518098638169094,
      "loss": 2.5971,
      "theoretical_loss": 3.7539281107936358,
      "tokens_seen": 748617728
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005180820881407225,
      "loss": 2.455,
      "theoretical_loss": 3.7538951433406202,
      "tokens_seen": 748683264
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005180655381123511,
      "loss": 2.6678,
      "theoretical_loss": 3.7538621795812306,
      "tokens_seen": 748748800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005180489880839796,
      "loss": 2.6395,
      "theoretical_loss": 3.7538292195147287,
      "tokens_seen": 748814336
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005180324380556081,
      "loss": 2.6261,
      "theoretical_loss": 3.753796263140379,
      "tokens_seen": 748879872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005180158880272366,
      "loss": 2.7098,
      "theoretical_loss": 3.753763310457444,
      "tokens_seen": 748945408
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005179993379988651,
      "loss": 2.7171,
      "theoretical_loss": 3.7537303614651867,
      "tokens_seen": 749010944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005179827879704937,
      "loss": 2.4344,
      "theoretical_loss": 3.753697416162872,
      "tokens_seen": 749076480
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005179662379421222,
      "loss": 2.647,
      "theoretical_loss": 3.753664474549763,
      "tokens_seen": 749142016
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005179496879137507,
      "loss": 2.437,
      "theoretical_loss": 3.753631536625125,
      "tokens_seen": 749207552
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005179331378853793,
      "loss": 2.7003,
      "theoretical_loss": 3.7535986023882213,
      "tokens_seen": 749273088
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005179165878570077,
      "loss": 2.7404,
      "theoretical_loss": 3.753565671838317,
      "tokens_seen": 749338624
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005179000378286363,
      "loss": 2.6215,
      "theoretical_loss": 3.753532744974677,
      "tokens_seen": 749404160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005178834878002648,
      "loss": 2.758,
      "theoretical_loss": 3.753499821796566,
      "tokens_seen": 749469696
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005178669377718933,
      "loss": 2.6231,
      "theoretical_loss": 3.7534669023032494,
      "tokens_seen": 749535232
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005178503877435219,
      "loss": 2.6118,
      "theoretical_loss": 3.7534339864939934,
      "tokens_seen": 749600768
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005178338377151504,
      "loss": 2.9245,
      "theoretical_loss": 3.7534010743680626,
      "tokens_seen": 749666304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005178172876867789,
      "loss": 2.737,
      "theoretical_loss": 3.7533681659247233,
      "tokens_seen": 749731840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005178007376584075,
      "loss": 2.6353,
      "theoretical_loss": 3.7533352611632416,
      "tokens_seen": 749797376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005177841876300359,
      "loss": 2.8469,
      "theoretical_loss": 3.753302360082884,
      "tokens_seen": 749862912
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005177676376016645,
      "loss": 2.666,
      "theoretical_loss": 3.7532694626829164,
      "tokens_seen": 749928448
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000517751087573293,
      "loss": 2.7642,
      "theoretical_loss": 3.753236568962606,
      "tokens_seen": 749993984
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005177345375449215,
      "loss": 2.6223,
      "theoretical_loss": 3.7532036789212206,
      "tokens_seen": 750059520
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005177179875165501,
      "loss": 2.8777,
      "theoretical_loss": 3.7531707925580253,
      "tokens_seen": 750125056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005177014374881786,
      "loss": 2.5626,
      "theoretical_loss": 3.7531379098722897,
      "tokens_seen": 750190592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005176848874598071,
      "loss": 2.9726,
      "theoretical_loss": 3.7531050308632796,
      "tokens_seen": 750256128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005176683374314356,
      "loss": 2.7958,
      "theoretical_loss": 3.7530721555302637,
      "tokens_seen": 750321664
    },
    {
      "epoch": 0.27,
      "objective/train/avg_token_score": 0.010190621018409729,
      "objective/train/avg_weight": 0.9880073666572571,
      "objective/train/docs_used": 433094,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.608158588409424,
      "objective/train/std_weight": 0.04930504038929939,
      "objective/train/theoretical_loss": 3.7530392838725097,
      "objective/train/tokens_used": 770847200,
      "theoretical_loss": 3.7530392838725097,
      "tokens_seen": 750387200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005176517874030641,
      "loss": 2.5669,
      "theoretical_loss": 3.7530392838725097,
      "tokens_seen": 750387200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005176352373746927,
      "loss": 3.1204,
      "theoretical_loss": 3.7530064158892857,
      "tokens_seen": 750452736
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005176186873463211,
      "loss": 2.8824,
      "theoretical_loss": 3.7529735515798603,
      "tokens_seen": 750518272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005176021373179497,
      "loss": 2.496,
      "theoretical_loss": 3.752940690943502,
      "tokens_seen": 750583808
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005175855872895782,
      "loss": 2.7772,
      "theoretical_loss": 3.7529078339794806,
      "tokens_seen": 750649344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005175690372612067,
      "loss": 2.6389,
      "theoretical_loss": 3.7528749806870634,
      "tokens_seen": 750714880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005175524872328352,
      "loss": 2.4833,
      "theoretical_loss": 3.7528421310655204,
      "tokens_seen": 750780416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005175359372044637,
      "loss": 2.8066,
      "theoretical_loss": 3.7528092851141213,
      "tokens_seen": 750845952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005175193871760922,
      "loss": 2.6133,
      "theoretical_loss": 3.7527764428321357,
      "tokens_seen": 750911488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005175028371477208,
      "loss": 2.4889,
      "theoretical_loss": 3.752743604218834,
      "tokens_seen": 750977024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005174862871193493,
      "loss": 2.6279,
      "theoretical_loss": 3.7527107692734845,
      "tokens_seen": 751042560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005174697370909779,
      "loss": 2.5404,
      "theoretical_loss": 3.752677937995359,
      "tokens_seen": 751108096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005174531870626064,
      "loss": 2.5957,
      "theoretical_loss": 3.752645110383728,
      "tokens_seen": 751173632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005174366370342348,
      "loss": 2.6173,
      "theoretical_loss": 3.752612286437862,
      "tokens_seen": 751239168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005174200870058634,
      "loss": 2.9473,
      "theoretical_loss": 3.752579466157031,
      "tokens_seen": 751304704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005174035369774919,
      "loss": 2.6347,
      "theoretical_loss": 3.7525466495405073,
      "tokens_seen": 751370240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005173869869491204,
      "loss": 2.7576,
      "theoretical_loss": 3.7525138365875623,
      "tokens_seen": 751435776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000517370436920749,
      "loss": 2.683,
      "theoretical_loss": 3.752481027297466,
      "tokens_seen": 751501312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005173538868923775,
      "loss": 2.6684,
      "theoretical_loss": 3.7524482216694923,
      "tokens_seen": 751566848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005173373368640061,
      "loss": 2.6923,
      "theoretical_loss": 3.7524154197029116,
      "tokens_seen": 751632384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005173207868356346,
      "loss": 2.7316,
      "theoretical_loss": 3.752382621396997,
      "tokens_seen": 751697920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000517304236807263,
      "loss": 2.6424,
      "theoretical_loss": 3.7523498267510202,
      "tokens_seen": 751763456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005172876867788916,
      "loss": 2.8415,
      "theoretical_loss": 3.7523170357642535,
      "tokens_seen": 751828992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005172711367505201,
      "loss": 2.9641,
      "theoretical_loss": 3.752284248435971,
      "tokens_seen": 751894528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005172545867221486,
      "loss": 2.6615,
      "theoretical_loss": 3.752251464765444,
      "tokens_seen": 751960064
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005172380366937772,
      "loss": 2.8684,
      "theoretical_loss": 3.7522186847519476,
      "tokens_seen": 752025600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005172214866654057,
      "loss": 2.8235,
      "theoretical_loss": 3.752185908394754,
      "tokens_seen": 752091136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005172049366370343,
      "loss": 2.9192,
      "theoretical_loss": 3.7521531356931366,
      "tokens_seen": 752156672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005171883866086627,
      "loss": 2.8128,
      "theoretical_loss": 3.75212036664637,
      "tokens_seen": 752222208
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005171718365802912,
      "loss": 2.5639,
      "theoretical_loss": 3.7520876012537276,
      "tokens_seen": 752287744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005171552865519198,
      "loss": 2.7558,
      "theoretical_loss": 3.7520548395144844,
      "tokens_seen": 752353280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005171387365235483,
      "loss": 2.4672,
      "theoretical_loss": 3.7520220814279144,
      "tokens_seen": 752418816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005171221864951768,
      "loss": 2.7804,
      "theoretical_loss": 3.751989326993292,
      "tokens_seen": 752484352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005171056364668054,
      "loss": 2.6889,
      "theoretical_loss": 3.751956576209892,
      "tokens_seen": 752549888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005170890864384339,
      "loss": 2.8665,
      "theoretical_loss": 3.7519238290769907,
      "tokens_seen": 752615424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005170725364100625,
      "loss": 2.8841,
      "theoretical_loss": 3.751891085593862,
      "tokens_seen": 752680960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005170559863816909,
      "loss": 2.6666,
      "theoretical_loss": 3.751858345759782,
      "tokens_seen": 752746496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005170394363533194,
      "loss": 2.6391,
      "theoretical_loss": 3.751825609574026,
      "tokens_seen": 752812032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000517022886324948,
      "loss": 2.582,
      "theoretical_loss": 3.7517928770358706,
      "tokens_seen": 752877568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005170063362965765,
      "loss": 2.7745,
      "theoretical_loss": 3.7517601481445917,
      "tokens_seen": 752943104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005169897862682051,
      "loss": 2.7795,
      "theoretical_loss": 3.7517274228994646,
      "tokens_seen": 753008640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005169732362398336,
      "loss": 2.7579,
      "theoretical_loss": 3.7516947012997672,
      "tokens_seen": 753074176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005169566862114621,
      "loss": 2.6974,
      "theoretical_loss": 3.7516619833447757,
      "tokens_seen": 753139712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005169401361830907,
      "loss": 2.7391,
      "theoretical_loss": 3.7516292690337667,
      "tokens_seen": 753205248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005169235861547191,
      "loss": 2.8542,
      "theoretical_loss": 3.7515965583660176,
      "tokens_seen": 753270784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005169070361263476,
      "loss": 2.5704,
      "theoretical_loss": 3.751563851340806,
      "tokens_seen": 753336320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005168904860979762,
      "loss": 2.7317,
      "theoretical_loss": 3.7515311479574094,
      "tokens_seen": 753401856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005168739360696047,
      "loss": 2.8586,
      "theoretical_loss": 3.7514984482151053,
      "tokens_seen": 753467392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005168573860412333,
      "loss": 2.7737,
      "theoretical_loss": 3.7514657521131713,
      "tokens_seen": 753532928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005168408360128618,
      "loss": 2.6793,
      "theoretical_loss": 3.7514330596508865,
      "tokens_seen": 753598464
    },
    {
      "epoch": 0.27,
      "objective/train/avg_token_score": 0.012395842000842094,
      "objective/train/avg_weight": 0.985407292842865,
      "objective/train/docs_used": 434283,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2925596237182617,
      "objective/train/std_weight": 0.08026266098022461,
      "objective/train/theoretical_loss": 3.751400370827529,
      "objective/train/tokens_used": 774124000,
      "theoretical_loss": 3.751400370827529,
      "tokens_seen": 753664000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005168242859844903,
      "loss": 2.905,
      "theoretical_loss": 3.751400370827529,
      "tokens_seen": 753664000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005168077359561188,
      "loss": 2.7287,
      "theoretical_loss": 3.7513676856423768,
      "tokens_seen": 753729536
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005167911859277473,
      "loss": 2.8771,
      "theoretical_loss": 3.751335004094709,
      "tokens_seen": 753795072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005167746358993758,
      "loss": 2.5597,
      "theoretical_loss": 3.751302326183805,
      "tokens_seen": 753860608
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005167580858710044,
      "loss": 3.1214,
      "theoretical_loss": 3.7512696519089435,
      "tokens_seen": 753926144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005167415358426329,
      "loss": 2.6886,
      "theoretical_loss": 3.751236981269404,
      "tokens_seen": 753991680
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005167249858142615,
      "loss": 2.7108,
      "theoretical_loss": 3.7512043142644664,
      "tokens_seen": 754057216
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00051670843578589,
      "loss": 2.9639,
      "theoretical_loss": 3.7511716508934105,
      "tokens_seen": 754122752
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005166918857575184,
      "loss": 2.7337,
      "theoretical_loss": 3.7511389911555155,
      "tokens_seen": 754188288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000516675335729147,
      "loss": 2.6361,
      "theoretical_loss": 3.7511063350500624,
      "tokens_seen": 754253824
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005166587857007755,
      "loss": 2.9265,
      "theoretical_loss": 3.7510736825763322,
      "tokens_seen": 754319360
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000516642235672404,
      "loss": 2.8503,
      "theoretical_loss": 3.751041033733604,
      "tokens_seen": 754384896
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005166256856440325,
      "loss": 2.8495,
      "theoretical_loss": 3.75100838852116,
      "tokens_seen": 754450432
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000516609135615661,
      "loss": 2.9731,
      "theoretical_loss": 3.7509757469382805,
      "tokens_seen": 754515968
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005165925855872896,
      "loss": 2.7581,
      "theoretical_loss": 3.750943108984247,
      "tokens_seen": 754581504
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000516576035558918,
      "loss": 2.9198,
      "theoretical_loss": 3.7509104746583413,
      "tokens_seen": 754647040
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005165594855305466,
      "loss": 2.6627,
      "theoretical_loss": 3.7508778439598442,
      "tokens_seen": 754712576
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005165429355021751,
      "loss": 2.8173,
      "theoretical_loss": 3.750845216888038,
      "tokens_seen": 754778112
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005165263854738036,
      "loss": 2.6106,
      "theoretical_loss": 3.7508125934422054,
      "tokens_seen": 754843648
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005165098354454322,
      "loss": 2.699,
      "theoretical_loss": 3.7507799736216274,
      "tokens_seen": 754909184
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005164932854170607,
      "loss": 2.9014,
      "theoretical_loss": 3.7507473574255874,
      "tokens_seen": 754974720
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005164767353886892,
      "loss": 2.6139,
      "theoretical_loss": 3.7507147448533678,
      "tokens_seen": 755040256
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005164601853603178,
      "loss": 2.7227,
      "theoretical_loss": 3.7506821359042517,
      "tokens_seen": 755105792
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005164436353319462,
      "loss": 2.7509,
      "theoretical_loss": 3.7506495305775225,
      "tokens_seen": 755171328
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005164270853035748,
      "loss": 2.9505,
      "theoretical_loss": 3.7506169288724625,
      "tokens_seen": 755236864
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005164105352752033,
      "loss": 2.7235,
      "theoretical_loss": 3.7505843307883557,
      "tokens_seen": 755302400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005163939852468318,
      "loss": 2.7124,
      "theoretical_loss": 3.750551736324486,
      "tokens_seen": 755367936
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005163774352184604,
      "loss": 2.7597,
      "theoretical_loss": 3.750519145480137,
      "tokens_seen": 755433472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005163608851900889,
      "loss": 2.7402,
      "theoretical_loss": 3.750486558254593,
      "tokens_seen": 755499008
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005163443351617174,
      "loss": 2.7833,
      "theoretical_loss": 3.7504539746471384,
      "tokens_seen": 755564544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005163277851333459,
      "loss": 2.997,
      "theoretical_loss": 3.7504213946570575,
      "tokens_seen": 755630080
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005163112351049744,
      "loss": 2.7157,
      "theoretical_loss": 3.7503888182836356,
      "tokens_seen": 755695616
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000516294685076603,
      "loss": 2.7128,
      "theoretical_loss": 3.7503562455261568,
      "tokens_seen": 755761152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005162781350482315,
      "loss": 2.679,
      "theoretical_loss": 3.750323676383906,
      "tokens_seen": 755826688
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00051626158501986,
      "loss": 2.8735,
      "theoretical_loss": 3.7502911108561694,
      "tokens_seen": 755892224
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005162450349914886,
      "loss": 2.533,
      "theoretical_loss": 3.7502585489422327,
      "tokens_seen": 755957760
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005162284849631171,
      "loss": 2.8412,
      "theoretical_loss": 3.750225990641381,
      "tokens_seen": 756023296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005162119349347455,
      "loss": 2.6164,
      "theoretical_loss": 3.7501934359529,
      "tokens_seen": 756088832
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005161953849063741,
      "loss": 2.6984,
      "theoretical_loss": 3.7501608848760766,
      "tokens_seen": 756154368
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005161788348780026,
      "loss": 2.7581,
      "theoretical_loss": 3.750128337410197,
      "tokens_seen": 756219904
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005161622848496312,
      "loss": 2.7659,
      "theoretical_loss": 3.7500957935545474,
      "tokens_seen": 756285440
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005161457348212597,
      "loss": 2.8051,
      "theoretical_loss": 3.750063253308414,
      "tokens_seen": 756350976
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005161291847928882,
      "loss": 2.7421,
      "theoretical_loss": 3.750030716671086,
      "tokens_seen": 756416512
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005161126347645168,
      "loss": 2.7109,
      "theoretical_loss": 3.7499981836418477,
      "tokens_seen": 756482048
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005160960847361453,
      "loss": 2.4597,
      "theoretical_loss": 3.7499656542199884,
      "tokens_seen": 756547584
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005160795347077737,
      "loss": 2.5031,
      "theoretical_loss": 3.7499331284047948,
      "tokens_seen": 756613120
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005160629846794023,
      "loss": 2.8525,
      "theoretical_loss": 3.749900606195555,
      "tokens_seen": 756678656
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005160464346510308,
      "loss": 2.6769,
      "theoretical_loss": 3.749868087591557,
      "tokens_seen": 756744192
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005160298846226594,
      "loss": 2.7483,
      "theoretical_loss": 3.7498355725920884,
      "tokens_seen": 756809728
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005160133345942879,
      "loss": 2.8332,
      "theoretical_loss": 3.7498030611964386,
      "tokens_seen": 756875264
    },
    {
      "epoch": 0.27,
      "objective/train/avg_token_score": 0.013377543538808823,
      "objective/train/avg_weight": 0.9842540621757507,
      "objective/train/docs_used": 435567,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.778474807739258,
      "objective/train/std_weight": 0.0755251944065094,
      "objective/train/theoretical_loss": 3.749770553403895,
      "objective/train/tokens_used": 777400800,
      "theoretical_loss": 3.749770553403895,
      "tokens_seen": 756940800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005159967845659164,
      "loss": 2.6417,
      "theoretical_loss": 3.749770553403895,
      "tokens_seen": 756940800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000515980234537545,
      "loss": 2.8687,
      "theoretical_loss": 3.7497380492137475,
      "tokens_seen": 757006336
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005159636845091734,
      "loss": 2.7988,
      "theoretical_loss": 3.7497055486252844,
      "tokens_seen": 757071872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005159471344808019,
      "loss": 2.558,
      "theoretical_loss": 3.749673051637795,
      "tokens_seen": 757137408
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005159305844524305,
      "loss": 2.6461,
      "theoretical_loss": 3.7496405582505687,
      "tokens_seen": 757202944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000515914034424059,
      "loss": 2.9344,
      "theoretical_loss": 3.7496080684628956,
      "tokens_seen": 757268480
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005158974843956876,
      "loss": 2.6134,
      "theoretical_loss": 3.7495755822740646,
      "tokens_seen": 757334016
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005158809343673161,
      "loss": 2.8808,
      "theoretical_loss": 3.749543099683366,
      "tokens_seen": 757399552
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005158643843389446,
      "loss": 2.8333,
      "theoretical_loss": 3.7495106206900903,
      "tokens_seen": 757465088
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005158478343105732,
      "loss": 2.7757,
      "theoretical_loss": 3.7494781452935273,
      "tokens_seen": 757530624
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005158312842822016,
      "loss": 2.5396,
      "theoretical_loss": 3.749445673492968,
      "tokens_seen": 757596160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005158147342538301,
      "loss": 2.7258,
      "theoretical_loss": 3.7494132052877034,
      "tokens_seen": 757661696
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005157981842254587,
      "loss": 2.6937,
      "theoretical_loss": 3.749380740677024,
      "tokens_seen": 757727232
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005157816341970872,
      "loss": 2.7706,
      "theoretical_loss": 3.7493482796602216,
      "tokens_seen": 757792768
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005157650841687158,
      "loss": 2.7648,
      "theoretical_loss": 3.749315822236587,
      "tokens_seen": 757858304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005157485341403443,
      "loss": 2.7084,
      "theoretical_loss": 3.749283368405412,
      "tokens_seen": 757923840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005157319841119728,
      "loss": 2.8613,
      "theoretical_loss": 3.7492509181659885,
      "tokens_seen": 757989376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005157154340836013,
      "loss": 2.816,
      "theoretical_loss": 3.749218471517608,
      "tokens_seen": 758054912
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005156988840552298,
      "loss": 2.9095,
      "theoretical_loss": 3.749186028459564,
      "tokens_seen": 758120448
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005156823340268583,
      "loss": 2.8811,
      "theoretical_loss": 3.749153588991147,
      "tokens_seen": 758185984
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005156657839984869,
      "loss": 2.7809,
      "theoretical_loss": 3.7491211531116515,
      "tokens_seen": 758251520
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005156492339701154,
      "loss": 2.7135,
      "theoretical_loss": 3.749088720820369,
      "tokens_seen": 758317056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005156326839417439,
      "loss": 2.7206,
      "theoretical_loss": 3.749056292116593,
      "tokens_seen": 758382592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005156161339133724,
      "loss": 2.803,
      "theoretical_loss": 3.7490238669996163,
      "tokens_seen": 758448128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005155995838850009,
      "loss": 2.7432,
      "theoretical_loss": 3.748991445468733,
      "tokens_seen": 758513664
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005155830338566294,
      "loss": 2.7252,
      "theoretical_loss": 3.7489590275232367,
      "tokens_seen": 758579200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005155664838282579,
      "loss": 2.3707,
      "theoretical_loss": 3.7489266131624204,
      "tokens_seen": 758644736
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005155499337998865,
      "loss": 2.5853,
      "theoretical_loss": 3.7488942023855785,
      "tokens_seen": 758710272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000515533383771515,
      "loss": 2.94,
      "theoretical_loss": 3.7488617951920054,
      "tokens_seen": 758775808
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005155168337431436,
      "loss": 2.8239,
      "theoretical_loss": 3.748829391580995,
      "tokens_seen": 758841344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005155002837147721,
      "loss": 2.7938,
      "theoretical_loss": 3.748796991551843,
      "tokens_seen": 758906880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005154837336864005,
      "loss": 2.6531,
      "theoretical_loss": 3.7487645951038426,
      "tokens_seen": 758972416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005154671836580291,
      "loss": 2.7167,
      "theoretical_loss": 3.7487322022362903,
      "tokens_seen": 759037952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005154506336296576,
      "loss": 2.8529,
      "theoretical_loss": 3.7486998129484803,
      "tokens_seen": 759103488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005154340836012861,
      "loss": 2.7177,
      "theoretical_loss": 3.7486674272397082,
      "tokens_seen": 759169024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005154175335729147,
      "loss": 2.7516,
      "theoretical_loss": 3.7486350451092703,
      "tokens_seen": 759234560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005154009835445432,
      "loss": 2.7583,
      "theoretical_loss": 3.7486026665564616,
      "tokens_seen": 759300096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005153844335161718,
      "loss": 2.5545,
      "theoretical_loss": 3.7485702915805783,
      "tokens_seen": 759365632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005153678834878003,
      "loss": 2.4074,
      "theoretical_loss": 3.7485379201809166,
      "tokens_seen": 759431168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005153513334594287,
      "loss": 2.5712,
      "theoretical_loss": 3.7485055523567734,
      "tokens_seen": 759496704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005153347834310573,
      "loss": 2.8122,
      "theoretical_loss": 3.748473188107444,
      "tokens_seen": 759562240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005153182334026858,
      "loss": 2.7014,
      "theoretical_loss": 3.7484408274322267,
      "tokens_seen": 759627776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005153016833743143,
      "loss": 2.7054,
      "theoretical_loss": 3.7484084703304177,
      "tokens_seen": 759693312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005152851333459429,
      "loss": 2.8429,
      "theoretical_loss": 3.7483761168013143,
      "tokens_seen": 759758848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005152685833175714,
      "loss": 2.7332,
      "theoretical_loss": 3.748343766844214,
      "tokens_seen": 759824384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005152520332892,
      "loss": 2.5851,
      "theoretical_loss": 3.7483114204584145,
      "tokens_seen": 759889920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005152354832608284,
      "loss": 2.7018,
      "theoretical_loss": 3.748279077643213,
      "tokens_seen": 759955456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005152189332324569,
      "loss": 2.6753,
      "theoretical_loss": 3.748246738397908,
      "tokens_seen": 760020992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005152023832040855,
      "loss": 2.6575,
      "theoretical_loss": 3.748214402721798,
      "tokens_seen": 760086528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000515185833175714,
      "loss": 2.8513,
      "theoretical_loss": 3.7481820706141806,
      "tokens_seen": 760152064
    },
    {
      "epoch": 0.27,
      "objective/train/avg_token_score": 0.019080501049757004,
      "objective/train/avg_weight": 0.9775392413139343,
      "objective/train/docs_used": 437507,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.877164840698242,
      "objective/train/std_weight": 0.10107860714197159,
      "objective/train/theoretical_loss": 3.748149742074355,
      "objective/train/tokens_used": 780677600,
      "theoretical_loss": 3.748149742074355,
      "tokens_seen": 760217600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005151692831473425,
      "loss": 2.4579,
      "theoretical_loss": 3.748149742074355,
      "tokens_seen": 760217600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005151527331189711,
      "loss": 2.7006,
      "theoretical_loss": 3.7481174171016196,
      "tokens_seen": 760283136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005151361830905996,
      "loss": 2.7767,
      "theoretical_loss": 3.7480850956952736,
      "tokens_seen": 760348672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005151196330622282,
      "loss": 2.5884,
      "theoretical_loss": 3.7480527778546167,
      "tokens_seen": 760414208
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005151030830338566,
      "loss": 2.696,
      "theoretical_loss": 3.7480204635789462,
      "tokens_seen": 760479744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005150865330054851,
      "loss": 2.905,
      "theoretical_loss": 3.7479881528675643,
      "tokens_seen": 760545280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005150699829771137,
      "loss": 3.0069,
      "theoretical_loss": 3.74795584571977,
      "tokens_seen": 760610816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005150534329487422,
      "loss": 2.7185,
      "theoretical_loss": 3.7479235421348616,
      "tokens_seen": 760676352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005150368829203707,
      "loss": 2.7443,
      "theoretical_loss": 3.747891242112142,
      "tokens_seen": 760741888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005150203328919993,
      "loss": 2.7691,
      "theoretical_loss": 3.747858945650909,
      "tokens_seen": 760807424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005150037828636278,
      "loss": 2.8751,
      "theoretical_loss": 3.747826652750465,
      "tokens_seen": 760872960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005149872328352563,
      "loss": 2.6623,
      "theoretical_loss": 3.74779436341011,
      "tokens_seen": 760938496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005149706828068848,
      "loss": 2.6684,
      "theoretical_loss": 3.7477620776291447,
      "tokens_seen": 761004032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005149541327785133,
      "loss": 2.7434,
      "theoretical_loss": 3.7477297954068707,
      "tokens_seen": 761069568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005149375827501419,
      "loss": 2.8587,
      "theoretical_loss": 3.747697516742589,
      "tokens_seen": 761135104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005149210327217704,
      "loss": 2.8236,
      "theoretical_loss": 3.7476652416356018,
      "tokens_seen": 761200640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005149044826933989,
      "loss": 2.798,
      "theoretical_loss": 3.7476329700852107,
      "tokens_seen": 761266176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005148879326650275,
      "loss": 2.9164,
      "theoretical_loss": 3.7476007020907165,
      "tokens_seen": 761331712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005148713826366559,
      "loss": 2.9671,
      "theoretical_loss": 3.747568437651423,
      "tokens_seen": 761397248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005148548326082845,
      "loss": 2.7652,
      "theoretical_loss": 3.7475361767666318,
      "tokens_seen": 761462784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000514838282579913,
      "loss": 2.6487,
      "theoretical_loss": 3.747503919435645,
      "tokens_seen": 761528320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005148217325515415,
      "loss": 2.8544,
      "theoretical_loss": 3.747471665657766,
      "tokens_seen": 761593856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005148051825231701,
      "loss": 2.75,
      "theoretical_loss": 3.7474394154322974,
      "tokens_seen": 761659392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005147886324947986,
      "loss": 2.8846,
      "theoretical_loss": 3.747407168758542,
      "tokens_seen": 761724928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005147720824664271,
      "loss": 2.8728,
      "theoretical_loss": 3.747374925635804,
      "tokens_seen": 761790464
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005147555324380557,
      "loss": 2.752,
      "theoretical_loss": 3.7473426860633863,
      "tokens_seen": 761856000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005147389824096841,
      "loss": 2.8202,
      "theoretical_loss": 3.747310450040593,
      "tokens_seen": 761921536
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005147224323813127,
      "loss": 2.6501,
      "theoretical_loss": 3.747278217566728,
      "tokens_seen": 761987072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005147058823529412,
      "loss": 2.4309,
      "theoretical_loss": 3.7472459886410947,
      "tokens_seen": 762052608
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005146893323245697,
      "loss": 2.6745,
      "theoretical_loss": 3.747213763262998,
      "tokens_seen": 762118144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005146727822961982,
      "loss": 2.7336,
      "theoretical_loss": 3.7471815414317424,
      "tokens_seen": 762183680
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005146562322678267,
      "loss": 2.8647,
      "theoretical_loss": 3.7471493231466324,
      "tokens_seen": 762249216
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005146396822394553,
      "loss": 2.7631,
      "theoretical_loss": 3.747117108406973,
      "tokens_seen": 762314752
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005146231322110837,
      "loss": 2.6611,
      "theoretical_loss": 3.747084897212069,
      "tokens_seen": 762380288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005146065821827122,
      "loss": 2.6435,
      "theoretical_loss": 3.747052689561227,
      "tokens_seen": 762445824
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005145900321543408,
      "loss": 2.8521,
      "theoretical_loss": 3.74702048545375,
      "tokens_seen": 762511360
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005145734821259693,
      "loss": 2.5211,
      "theoretical_loss": 3.746988284888946,
      "tokens_seen": 762576896
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005145569320975979,
      "loss": 2.6445,
      "theoretical_loss": 3.7469560878661197,
      "tokens_seen": 762642432
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005145403820692264,
      "loss": 2.5032,
      "theoretical_loss": 3.746923894384578,
      "tokens_seen": 762707968
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005145238320408549,
      "loss": 2.926,
      "theoretical_loss": 3.7468917044436267,
      "tokens_seen": 762773504
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005145072820124834,
      "loss": 2.8923,
      "theoretical_loss": 3.7468595180425712,
      "tokens_seen": 762839040
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005144907319841119,
      "loss": 2.8729,
      "theoretical_loss": 3.74682733518072,
      "tokens_seen": 762904576
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005144741819557405,
      "loss": 2.8745,
      "theoretical_loss": 3.746795155857379,
      "tokens_seen": 762970112
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000514457631927369,
      "loss": 2.7817,
      "theoretical_loss": 3.746762980071855,
      "tokens_seen": 763035648
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005144410818989975,
      "loss": 2.7375,
      "theoretical_loss": 3.7467308078234565,
      "tokens_seen": 763101184
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005144245318706261,
      "loss": 2.6144,
      "theoretical_loss": 3.7466986391114894,
      "tokens_seen": 763166720
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005144079818422546,
      "loss": 2.8182,
      "theoretical_loss": 3.746666473935262,
      "tokens_seen": 763232256
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000514391431813883,
      "loss": 2.7719,
      "theoretical_loss": 3.7466343122940815,
      "tokens_seen": 763297792
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005143748817855116,
      "loss": 2.7091,
      "theoretical_loss": 3.7466021541872574,
      "tokens_seen": 763363328
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005143583317571401,
      "loss": 2.6909,
      "theoretical_loss": 3.7465699996140964,
      "tokens_seen": 763428864
    },
    {
      "epoch": 0.27,
      "objective/train/avg_token_score": 0.01802533119916916,
      "objective/train/avg_weight": 0.9787824153900146,
      "objective/train/docs_used": 439068,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.731436252593994,
      "objective/train/std_weight": 0.08805865794420242,
      "objective/train/theoretical_loss": 3.746537848573908,
      "objective/train/tokens_used": 783954400,
      "theoretical_loss": 3.746537848573908,
      "tokens_seen": 763494400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005143417817287687,
      "loss": 2.7185,
      "theoretical_loss": 3.746537848573908,
      "tokens_seen": 763494400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005143252317003972,
      "loss": 2.8945,
      "theoretical_loss": 3.746505701066001,
      "tokens_seen": 763559936
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005143086816720257,
      "loss": 2.7973,
      "theoretical_loss": 3.7464735570896823,
      "tokens_seen": 763625472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005142921316436543,
      "loss": 2.7114,
      "theoretical_loss": 3.746441416644263,
      "tokens_seen": 763691008
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005142755816152828,
      "loss": 2.5689,
      "theoretical_loss": 3.7464092797290505,
      "tokens_seen": 763756544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005142590315869112,
      "loss": 2.7299,
      "theoretical_loss": 3.746377146343356,
      "tokens_seen": 763822080
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005142424815585398,
      "loss": 2.871,
      "theoretical_loss": 3.746345016486488,
      "tokens_seen": 763887616
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005142259315301683,
      "loss": 2.817,
      "theoretical_loss": 3.7463128901577565,
      "tokens_seen": 763953152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005142093815017969,
      "loss": 2.7277,
      "theoretical_loss": 3.7462807673564713,
      "tokens_seen": 764018688
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005141928314734254,
      "loss": 2.8318,
      "theoretical_loss": 3.746248648081943,
      "tokens_seen": 764084224
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005141762814450539,
      "loss": 2.7385,
      "theoretical_loss": 3.7462165323334813,
      "tokens_seen": 764149760
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005141597314166825,
      "loss": 2.9877,
      "theoretical_loss": 3.746184420110397,
      "tokens_seen": 764215296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000514143181388311,
      "loss": 2.7858,
      "theoretical_loss": 3.7461523114120014,
      "tokens_seen": 764280832
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005141266313599394,
      "loss": 2.7121,
      "theoretical_loss": 3.746120206237605,
      "tokens_seen": 764346368
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000514110081331568,
      "loss": 2.8594,
      "theoretical_loss": 3.7460881045865184,
      "tokens_seen": 764411904
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005140935313031965,
      "loss": 2.8926,
      "theoretical_loss": 3.746056006458054,
      "tokens_seen": 764477440
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005140769812748251,
      "loss": 2.7436,
      "theoretical_loss": 3.7460239118515224,
      "tokens_seen": 764542976
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005140604312464536,
      "loss": 2.8624,
      "theoretical_loss": 3.745991820766236,
      "tokens_seen": 764608512
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005140438812180821,
      "loss": 2.5747,
      "theoretical_loss": 3.745959733201506,
      "tokens_seen": 764674048
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005140273311897107,
      "loss": 2.825,
      "theoretical_loss": 3.7459276491566453,
      "tokens_seen": 764739584
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005140107811613391,
      "loss": 2.7635,
      "theoretical_loss": 3.7458955686309654,
      "tokens_seen": 764805120
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005139942311329676,
      "loss": 3.0171,
      "theoretical_loss": 3.74586349162378,
      "tokens_seen": 764870656
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005139776811045962,
      "loss": 2.9747,
      "theoretical_loss": 3.7458314181344,
      "tokens_seen": 764936192
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005139611310762247,
      "loss": 2.6598,
      "theoretical_loss": 3.74579934816214,
      "tokens_seen": 765001728
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005139445810478533,
      "loss": 2.8649,
      "theoretical_loss": 3.7457672817063115,
      "tokens_seen": 765067264
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005139280310194818,
      "loss": 2.9303,
      "theoretical_loss": 3.745735218766229,
      "tokens_seen": 765132800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005139114809911103,
      "loss": 2.6577,
      "theoretical_loss": 3.745703159341205,
      "tokens_seen": 765198336
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005138949309627388,
      "loss": 2.8389,
      "theoretical_loss": 3.745671103430554,
      "tokens_seen": 765263872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005138783809343673,
      "loss": 2.6129,
      "theoretical_loss": 3.7456390510335895,
      "tokens_seen": 765329408
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005138618309059958,
      "loss": 2.8511,
      "theoretical_loss": 3.7456070021496255,
      "tokens_seen": 765394944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005138452808776244,
      "loss": 2.5245,
      "theoretical_loss": 3.7455749567779764,
      "tokens_seen": 765460480
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005138287308492529,
      "loss": 2.5365,
      "theoretical_loss": 3.745542914917956,
      "tokens_seen": 765526016
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005138121808208815,
      "loss": 2.8121,
      "theoretical_loss": 3.74551087656888,
      "tokens_seen": 765591552
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00051379563079251,
      "loss": 2.9157,
      "theoretical_loss": 3.745478841730062,
      "tokens_seen": 765657088
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005137790807641384,
      "loss": 2.757,
      "theoretical_loss": 3.7454468104008174,
      "tokens_seen": 765722624
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000513762530735767,
      "loss": 2.6283,
      "theoretical_loss": 3.7454147825804616,
      "tokens_seen": 765788160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005137459807073955,
      "loss": 2.8162,
      "theoretical_loss": 3.74538275826831,
      "tokens_seen": 765853696
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000513729430679024,
      "loss": 2.7827,
      "theoretical_loss": 3.745350737463678,
      "tokens_seen": 765919232
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005137128806506526,
      "loss": 2.8414,
      "theoretical_loss": 3.7453187201658817,
      "tokens_seen": 765984768
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005136963306222811,
      "loss": 2.476,
      "theoretical_loss": 3.7452867063742366,
      "tokens_seen": 766050304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005136797805939096,
      "loss": 2.6813,
      "theoretical_loss": 3.7452546960880593,
      "tokens_seen": 766115840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000513663230565538,
      "loss": 2.9679,
      "theoretical_loss": 3.745222689306665,
      "tokens_seen": 766181376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005136466805371666,
      "loss": 2.8151,
      "theoretical_loss": 3.745190686029372,
      "tokens_seen": 766246912
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005136301305087951,
      "loss": 2.8664,
      "theoretical_loss": 3.745158686255496,
      "tokens_seen": 766312448
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005136135804804236,
      "loss": 2.565,
      "theoretical_loss": 3.7451266899843536,
      "tokens_seen": 766377984
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005135970304520522,
      "loss": 2.7642,
      "theoretical_loss": 3.7450946972152623,
      "tokens_seen": 766443520
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005135804804236807,
      "loss": 2.8155,
      "theoretical_loss": 3.7450627079475396,
      "tokens_seen": 766509056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005135639303953092,
      "loss": 2.8042,
      "theoretical_loss": 3.745030722180503,
      "tokens_seen": 766574592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005135473803669378,
      "loss": 2.7061,
      "theoretical_loss": 3.74499873991347,
      "tokens_seen": 766640128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005135308303385662,
      "loss": 2.6653,
      "theoretical_loss": 3.7449667611457578,
      "tokens_seen": 766705664
    },
    {
      "epoch": 0.27,
      "objective/train/avg_token_score": 0.03261525183916092,
      "objective/train/avg_weight": 0.9616137146949768,
      "objective/train/docs_used": 441051,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.021949291229248,
      "objective/train/std_weight": 0.11532315611839294,
      "objective/train/theoretical_loss": 3.744934785876686,
      "objective/train/tokens_used": 787231200,
      "theoretical_loss": 3.744934785876686,
      "tokens_seen": 766771200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005135142803101948,
      "loss": 2.8177,
      "theoretical_loss": 3.744934785876686,
      "tokens_seen": 766771200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005134977302818233,
      "loss": 2.6154,
      "theoretical_loss": 3.744902814105571,
      "tokens_seen": 766836736
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005134811802534518,
      "loss": 2.4816,
      "theoretical_loss": 3.7448708458317332,
      "tokens_seen": 766902272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005134646302250804,
      "loss": 2.81,
      "theoretical_loss": 3.7448388810544895,
      "tokens_seen": 766967808
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005134480801967089,
      "loss": 2.8517,
      "theoretical_loss": 3.7448069197731595,
      "tokens_seen": 767033344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005134315301683374,
      "loss": 2.739,
      "theoretical_loss": 3.744774961987063,
      "tokens_seen": 767098880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000513414980139966,
      "loss": 2.8316,
      "theoretical_loss": 3.7447430076955177,
      "tokens_seen": 767164416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005133984301115944,
      "loss": 2.8077,
      "theoretical_loss": 3.7447110568978434,
      "tokens_seen": 767229952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000513381880083223,
      "loss": 2.6451,
      "theoretical_loss": 3.7446791095933603,
      "tokens_seen": 767295488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005133653300548515,
      "loss": 2.8253,
      "theoretical_loss": 3.7446471657813882,
      "tokens_seen": 767361024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00051334878002648,
      "loss": 2.8616,
      "theoretical_loss": 3.744615225461246,
      "tokens_seen": 767426560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005133322299981086,
      "loss": 2.7679,
      "theoretical_loss": 3.744583288632255,
      "tokens_seen": 767492096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005133156799697371,
      "loss": 2.9018,
      "theoretical_loss": 3.7445513552937353,
      "tokens_seen": 767557632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005132991299413655,
      "loss": 2.8398,
      "theoretical_loss": 3.7445194254450067,
      "tokens_seen": 767623168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005132825799129941,
      "loss": 2.8283,
      "theoretical_loss": 3.7444874990853907,
      "tokens_seen": 767688704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005132660298846226,
      "loss": 2.9057,
      "theoretical_loss": 3.744455576214208,
      "tokens_seen": 767754240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005132494798562512,
      "loss": 2.7939,
      "theoretical_loss": 3.7444236568307803,
      "tokens_seen": 767819776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005132329298278797,
      "loss": 2.8397,
      "theoretical_loss": 3.7443917409344274,
      "tokens_seen": 767885312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005132163797995082,
      "loss": 2.8015,
      "theoretical_loss": 3.7443598285244724,
      "tokens_seen": 767950848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005131998297711368,
      "loss": 2.7367,
      "theoretical_loss": 3.7443279196002357,
      "tokens_seen": 768016384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005131832797427653,
      "loss": 2.7544,
      "theoretical_loss": 3.74429601416104,
      "tokens_seen": 768081920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005131667297143937,
      "loss": 2.7736,
      "theoretical_loss": 3.7442641122062073,
      "tokens_seen": 768147456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005131501796860223,
      "loss": 2.9282,
      "theoretical_loss": 3.7442322137350597,
      "tokens_seen": 768212992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005131336296576508,
      "loss": 2.6813,
      "theoretical_loss": 3.7442003187469193,
      "tokens_seen": 768278528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005131170796292794,
      "loss": 2.8298,
      "theoretical_loss": 3.7441684272411093,
      "tokens_seen": 768344064
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005131005296009079,
      "loss": 2.748,
      "theoretical_loss": 3.744136539216952,
      "tokens_seen": 768409600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005130839795725364,
      "loss": 2.8142,
      "theoretical_loss": 3.744104654673771,
      "tokens_seen": 768475136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000513067429544165,
      "loss": 2.6993,
      "theoretical_loss": 3.7440727736108887,
      "tokens_seen": 768540672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005130508795157934,
      "loss": 2.5744,
      "theoretical_loss": 3.744040896027629,
      "tokens_seen": 768606208
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005130343294874219,
      "loss": 2.7402,
      "theoretical_loss": 3.7440090219233157,
      "tokens_seen": 768671744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005130177794590505,
      "loss": 2.6624,
      "theoretical_loss": 3.7439771512972726,
      "tokens_seen": 768737280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000513001229430679,
      "loss": 2.8324,
      "theoretical_loss": 3.7439452841488228,
      "tokens_seen": 768802816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005129846794023076,
      "loss": 2.5674,
      "theoretical_loss": 3.7439134204772913,
      "tokens_seen": 768868352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005129681293739361,
      "loss": 2.812,
      "theoretical_loss": 3.7438815602820017,
      "tokens_seen": 768933888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005129515793455646,
      "loss": 2.7113,
      "theoretical_loss": 3.743849703562279,
      "tokens_seen": 768999424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005129350293171932,
      "loss": 2.7961,
      "theoretical_loss": 3.743817850317448,
      "tokens_seen": 769064960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005129184792888216,
      "loss": 2.7944,
      "theoretical_loss": 3.7437860005468337,
      "tokens_seen": 769130496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005129019292604502,
      "loss": 2.9287,
      "theoretical_loss": 3.7437541542497605,
      "tokens_seen": 769196032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005128853792320787,
      "loss": 2.6371,
      "theoretical_loss": 3.7437223114255542,
      "tokens_seen": 769261568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005128688292037072,
      "loss": 2.7325,
      "theoretical_loss": 3.74369047207354,
      "tokens_seen": 769327104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005128522791753358,
      "loss": 2.8674,
      "theoretical_loss": 3.7436586361930444,
      "tokens_seen": 769392640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005128357291469643,
      "loss": 2.8545,
      "theoretical_loss": 3.743626803783392,
      "tokens_seen": 769458176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005128191791185928,
      "loss": 2.7538,
      "theoretical_loss": 3.7435949748439095,
      "tokens_seen": 769523712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005128026290902213,
      "loss": 2.8189,
      "theoretical_loss": 3.743563149373923,
      "tokens_seen": 769589248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005127860790618498,
      "loss": 2.6279,
      "theoretical_loss": 3.743531327372759,
      "tokens_seen": 769654784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005127695290334784,
      "loss": 2.8094,
      "theoretical_loss": 3.743499508839744,
      "tokens_seen": 769720320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005127529790051069,
      "loss": 2.6404,
      "theoretical_loss": 3.7434676937742046,
      "tokens_seen": 769785856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005127364289767354,
      "loss": 2.9312,
      "theoretical_loss": 3.743435882175468,
      "tokens_seen": 769851392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000512719878948364,
      "loss": 2.9495,
      "theoretical_loss": 3.7434040740428616,
      "tokens_seen": 769916928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0005127033289199924,
      "loss": 2.9384,
      "theoretical_loss": 3.743372269375712,
      "tokens_seen": 769982464
    },
    {
      "epoch": 0.27,
      "objective/train/avg_token_score": 0.012731542810797691,
      "objective/train/avg_weight": 0.9850170016288757,
      "objective/train/docs_used": 442749,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9628005027770996,
      "objective/train/std_weight": 0.06771586835384369,
      "objective/train/theoretical_loss": 3.7433404681733475,
      "objective/train/tokens_used": 790508000,
      "theoretical_loss": 3.7433404681733475,
      "tokens_seen": 770048000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000512686778891621,
      "loss": 2.8971,
      "theoretical_loss": 3.7433404681733475,
      "tokens_seen": 770048000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005126702288632494,
      "loss": 2.7117,
      "theoretical_loss": 3.7433086704350957,
      "tokens_seen": 770113536
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005126536788348779,
      "loss": 2.7271,
      "theoretical_loss": 3.7432768761602837,
      "tokens_seen": 770179072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005126371288065065,
      "loss": 2.9068,
      "theoretical_loss": 3.743245085348241,
      "tokens_seen": 770244608
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000512620578778135,
      "loss": 2.5394,
      "theoretical_loss": 3.7432132979982944,
      "tokens_seen": 770310144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005126040287497636,
      "loss": 2.9966,
      "theoretical_loss": 3.7431815141097733,
      "tokens_seen": 770375680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005125874787213921,
      "loss": 2.6805,
      "theoretical_loss": 3.7431497336820065,
      "tokens_seen": 770441216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005125709286930206,
      "loss": 2.8719,
      "theoretical_loss": 3.743117956714322,
      "tokens_seen": 770506752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005125543786646491,
      "loss": 2.8237,
      "theoretical_loss": 3.7430861832060494,
      "tokens_seen": 770572288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005125378286362776,
      "loss": 3.0234,
      "theoretical_loss": 3.7430544131565178,
      "tokens_seen": 770637824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005125212786079061,
      "loss": 2.9471,
      "theoretical_loss": 3.7430226465650573,
      "tokens_seen": 770703360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005125047285795347,
      "loss": 2.9239,
      "theoretical_loss": 3.7429908834309957,
      "tokens_seen": 770768896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005124881785511632,
      "loss": 2.6736,
      "theoretical_loss": 3.742959123753665,
      "tokens_seen": 770834432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005124716285227918,
      "loss": 2.991,
      "theoretical_loss": 3.7429273675323933,
      "tokens_seen": 770899968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005124550784944203,
      "loss": 2.6894,
      "theoretical_loss": 3.742895614766512,
      "tokens_seen": 770965504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005124385284660487,
      "loss": 2.7878,
      "theoretical_loss": 3.7428638654553508,
      "tokens_seen": 771031040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005124219784376773,
      "loss": 2.7381,
      "theoretical_loss": 3.74283211959824,
      "tokens_seen": 771096576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005124054284093058,
      "loss": 2.7709,
      "theoretical_loss": 3.7428003771945115,
      "tokens_seen": 771162112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005123888783809343,
      "loss": 2.8781,
      "theoretical_loss": 3.7427686382434953,
      "tokens_seen": 771227648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005123723283525629,
      "loss": 2.8991,
      "theoretical_loss": 3.742736902744522,
      "tokens_seen": 771293184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005123557783241914,
      "loss": 2.784,
      "theoretical_loss": 3.742705170696924,
      "tokens_seen": 771358720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00051233922829582,
      "loss": 2.6116,
      "theoretical_loss": 3.742673442100032,
      "tokens_seen": 771424256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005123226782674485,
      "loss": 2.9006,
      "theoretical_loss": 3.7426417169531776,
      "tokens_seen": 771489792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005123061282390769,
      "loss": 2.6018,
      "theoretical_loss": 3.7426099952556937,
      "tokens_seen": 771555328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005122895782107055,
      "loss": 2.7669,
      "theoretical_loss": 3.742578277006911,
      "tokens_seen": 771620864
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000512273028182334,
      "loss": 2.8906,
      "theoretical_loss": 3.7425465622061624,
      "tokens_seen": 771686400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005122564781539625,
      "loss": 2.7926,
      "theoretical_loss": 3.74251485085278,
      "tokens_seen": 771751936
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005122399281255911,
      "loss": 2.8223,
      "theoretical_loss": 3.7424831429460967,
      "tokens_seen": 771817472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005122233780972196,
      "loss": 2.7136,
      "theoretical_loss": 3.7424514384854453,
      "tokens_seen": 771883008
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005122068280688482,
      "loss": 2.8423,
      "theoretical_loss": 3.742419737470158,
      "tokens_seen": 771948544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005121902780404766,
      "loss": 2.7688,
      "theoretical_loss": 3.742388039899569,
      "tokens_seen": 772014080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005121737280121051,
      "loss": 2.8691,
      "theoretical_loss": 3.742356345773011,
      "tokens_seen": 772079616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005121571779837337,
      "loss": 2.7236,
      "theoretical_loss": 3.742324655089817,
      "tokens_seen": 772145152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005121406279553622,
      "loss": 2.8646,
      "theoretical_loss": 3.742292967849322,
      "tokens_seen": 772210688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005121240779269907,
      "loss": 3.1358,
      "theoretical_loss": 3.742261284050859,
      "tokens_seen": 772276224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005121075278986193,
      "loss": 2.6907,
      "theoretical_loss": 3.7422296036937617,
      "tokens_seen": 772341760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005120909778702478,
      "loss": 2.7804,
      "theoretical_loss": 3.7421979267773655,
      "tokens_seen": 772407296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005120744278418764,
      "loss": 2.6337,
      "theoretical_loss": 3.742166253301004,
      "tokens_seen": 772472832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005120578778135048,
      "loss": 2.8968,
      "theoretical_loss": 3.7421345832640123,
      "tokens_seen": 772538368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005120413277851333,
      "loss": 2.9285,
      "theoretical_loss": 3.7421029166657243,
      "tokens_seen": 772603904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005120247777567619,
      "loss": 2.9969,
      "theoretical_loss": 3.7420712535054763,
      "tokens_seen": 772669440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005120082277283904,
      "loss": 2.9773,
      "theoretical_loss": 3.742039593782602,
      "tokens_seen": 772734976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005119916777000189,
      "loss": 2.87,
      "theoretical_loss": 3.742007937496438,
      "tokens_seen": 772800512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005119751276716475,
      "loss": 2.599,
      "theoretical_loss": 3.7419762846463196,
      "tokens_seen": 772866048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000511958577643276,
      "loss": 2.7545,
      "theoretical_loss": 3.741944635231582,
      "tokens_seen": 772931584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005119420276149045,
      "loss": 2.7952,
      "theoretical_loss": 3.7419129892515617,
      "tokens_seen": 772997120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000511925477586533,
      "loss": 2.7157,
      "theoretical_loss": 3.7418813467055942,
      "tokens_seen": 773062656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005119089275581615,
      "loss": 2.4355,
      "theoretical_loss": 3.7418497075930164,
      "tokens_seen": 773128192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005118923775297901,
      "loss": 2.8556,
      "theoretical_loss": 3.741818071913164,
      "tokens_seen": 773193728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005118758275014186,
      "loss": 2.5943,
      "theoretical_loss": 3.7417864396653746,
      "tokens_seen": 773259264
    },
    {
      "epoch": 0.28,
      "objective/train/avg_token_score": 0.03155338391661644,
      "objective/train/avg_weight": 0.9628701210021973,
      "objective/train/docs_used": 444718,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.4094901084899902,
      "objective/train/std_weight": 0.15757034718990326,
      "objective/train/theoretical_loss": 3.7417548108489846,
      "objective/train/tokens_used": 793784800,
      "theoretical_loss": 3.7417548108489846,
      "tokens_seen": 773324800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005118592774730471,
      "loss": 2.6706,
      "theoretical_loss": 3.7417548108489846,
      "tokens_seen": 773324800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005118427274446757,
      "loss": 2.5241,
      "theoretical_loss": 3.741723185463331,
      "tokens_seen": 773390336
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005118261774163041,
      "loss": 2.7408,
      "theoretical_loss": 3.7416915635077506,
      "tokens_seen": 773455872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005118096273879327,
      "loss": 2.7345,
      "theoretical_loss": 3.7416599449815813,
      "tokens_seen": 773521408
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005117930773595612,
      "loss": 2.6853,
      "theoretical_loss": 3.7416283298841604,
      "tokens_seen": 773586944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005117765273311897,
      "loss": 2.7635,
      "theoretical_loss": 3.7415967182148266,
      "tokens_seen": 773652480
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005117599773028183,
      "loss": 2.7672,
      "theoretical_loss": 3.7415651099729166,
      "tokens_seen": 773718016
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005117434272744468,
      "loss": 2.9254,
      "theoretical_loss": 3.7415335051577685,
      "tokens_seen": 773783552
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005117268772460753,
      "loss": 2.7481,
      "theoretical_loss": 3.7415019037687216,
      "tokens_seen": 773849088
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005117103272177037,
      "loss": 2.8188,
      "theoretical_loss": 3.7414703058051137,
      "tokens_seen": 773914624
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005116937771893323,
      "loss": 2.7872,
      "theoretical_loss": 3.7414387112662837,
      "tokens_seen": 773980160
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005116772271609608,
      "loss": 2.6236,
      "theoretical_loss": 3.74140712015157,
      "tokens_seen": 774045696
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005116606771325893,
      "loss": 2.9516,
      "theoretical_loss": 3.741375532460313,
      "tokens_seen": 774111232
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005116441271042179,
      "loss": 2.566,
      "theoretical_loss": 3.74134394819185,
      "tokens_seen": 774176768
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005116275770758464,
      "loss": 2.9693,
      "theoretical_loss": 3.741312367345522,
      "tokens_seen": 774242304
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005116110270474749,
      "loss": 2.6825,
      "theoretical_loss": 3.7412807899206677,
      "tokens_seen": 774307840
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005115944770191035,
      "loss": 2.7444,
      "theoretical_loss": 3.7412492159166275,
      "tokens_seen": 774373376
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005115779269907319,
      "loss": 2.8194,
      "theoretical_loss": 3.741217645332741,
      "tokens_seen": 774438912
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005115613769623605,
      "loss": 2.6357,
      "theoretical_loss": 3.7411860781683477,
      "tokens_seen": 774504448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000511544826933989,
      "loss": 2.8084,
      "theoretical_loss": 3.741154514422789,
      "tokens_seen": 774569984
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005115282769056175,
      "loss": 2.6701,
      "theoretical_loss": 3.741122954095405,
      "tokens_seen": 774635520
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005115117268772461,
      "loss": 2.6487,
      "theoretical_loss": 3.7410913971855364,
      "tokens_seen": 774701056
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005114951768488746,
      "loss": 2.7763,
      "theoretical_loss": 3.7410598436925238,
      "tokens_seen": 774766592
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000511478626820503,
      "loss": 2.8933,
      "theoretical_loss": 3.7410282936157087,
      "tokens_seen": 774832128
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005114620767921316,
      "loss": 2.5746,
      "theoretical_loss": 3.7409967469544325,
      "tokens_seen": 774897664
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005114455267637601,
      "loss": 2.9017,
      "theoretical_loss": 3.7409652037080363,
      "tokens_seen": 774963200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005114289767353887,
      "loss": 2.9171,
      "theoretical_loss": 3.740933663875861,
      "tokens_seen": 775028736
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005114124267070172,
      "loss": 2.9252,
      "theoretical_loss": 3.7409021274572494,
      "tokens_seen": 775094272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005113958766786457,
      "loss": 2.6351,
      "theoretical_loss": 3.7408705944515432,
      "tokens_seen": 775159808
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005113793266502743,
      "loss": 2.7507,
      "theoretical_loss": 3.740839064858085,
      "tokens_seen": 775225344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005113627766219028,
      "loss": 2.821,
      "theoretical_loss": 3.740807538676216,
      "tokens_seen": 775290880
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005113462265935312,
      "loss": 2.7262,
      "theoretical_loss": 3.7407760159052796,
      "tokens_seen": 775356416
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005113296765651598,
      "loss": 2.7604,
      "theoretical_loss": 3.740744496544618,
      "tokens_seen": 775421952
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005113131265367883,
      "loss": 2.7948,
      "theoretical_loss": 3.7407129805935746,
      "tokens_seen": 775487488
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005112965765084169,
      "loss": 2.7748,
      "theoretical_loss": 3.740681468051492,
      "tokens_seen": 775553024
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005112800264800454,
      "loss": 2.9476,
      "theoretical_loss": 3.740649958917714,
      "tokens_seen": 775618560
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005112634764516739,
      "loss": 2.8724,
      "theoretical_loss": 3.7406184531915834,
      "tokens_seen": 775684096
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005112469264233025,
      "loss": 2.8686,
      "theoretical_loss": 3.7405869508724443,
      "tokens_seen": 775749632
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000511230376394931,
      "loss": 2.6507,
      "theoretical_loss": 3.74055545195964,
      "tokens_seen": 775815168
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005112138263665594,
      "loss": 2.8928,
      "theoretical_loss": 3.740523956452515,
      "tokens_seen": 775880704
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000511197276338188,
      "loss": 3.0084,
      "theoretical_loss": 3.7404924643504134,
      "tokens_seen": 775946240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005111807263098165,
      "loss": 3.0942,
      "theoretical_loss": 3.740460975652679,
      "tokens_seen": 776011776
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005111641762814451,
      "loss": 2.9357,
      "theoretical_loss": 3.740429490358657,
      "tokens_seen": 776077312
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005111476262530736,
      "loss": 2.7829,
      "theoretical_loss": 3.7403980084676918,
      "tokens_seen": 776142848
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005111310762247021,
      "loss": 2.615,
      "theoretical_loss": 3.740366529979128,
      "tokens_seen": 776208384
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005111145261963307,
      "loss": 2.5479,
      "theoretical_loss": 3.740335054892311,
      "tokens_seen": 776273920
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005110979761679591,
      "loss": 2.6949,
      "theoretical_loss": 3.740303583206586,
      "tokens_seen": 776339456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005110814261395876,
      "loss": 2.6567,
      "theoretical_loss": 3.7402721149212983,
      "tokens_seen": 776404992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005110648761112162,
      "loss": 2.6321,
      "theoretical_loss": 3.7402406500357936,
      "tokens_seen": 776470528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005110483260828447,
      "loss": 2.7229,
      "theoretical_loss": 3.740209188549418,
      "tokens_seen": 776536064
    },
    {
      "epoch": 0.28,
      "objective/train/avg_token_score": 0.01584375835955143,
      "objective/train/avg_weight": 0.9813547730445862,
      "objective/train/docs_used": 446660,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8512210845947266,
      "objective/train/std_weight": 0.0958196371793747,
      "objective/train/theoretical_loss": 3.740177730461517,
      "objective/train/tokens_used": 797061600,
      "theoretical_loss": 3.740177730461517,
      "tokens_seen": 776601600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005110317760544733,
      "loss": 2.7663,
      "theoretical_loss": 3.740177730461517,
      "tokens_seen": 776601600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005110152260261018,
      "loss": 2.6914,
      "theoretical_loss": 3.740146275771437,
      "tokens_seen": 776667136
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005109986759977303,
      "loss": 2.8485,
      "theoretical_loss": 3.740114824478524,
      "tokens_seen": 776732672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005109821259693589,
      "loss": 2.8385,
      "theoretical_loss": 3.740083376582125,
      "tokens_seen": 776798208
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005109655759409873,
      "loss": 2.6977,
      "theoretical_loss": 3.7400519320815864,
      "tokens_seen": 776863744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005109490259126158,
      "loss": 2.921,
      "theoretical_loss": 3.740020490976255,
      "tokens_seen": 776929280
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005109324758842444,
      "loss": 3.2959,
      "theoretical_loss": 3.7399890532654783,
      "tokens_seen": 776994816
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005109159258558729,
      "loss": 2.9279,
      "theoretical_loss": 3.739957618948603,
      "tokens_seen": 777060352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005108993758275015,
      "loss": 2.7745,
      "theoretical_loss": 3.7399261880249766,
      "tokens_seen": 777125888
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00051088282579913,
      "loss": 2.8145,
      "theoretical_loss": 3.7398947604939474,
      "tokens_seen": 777191424
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005108662757707585,
      "loss": 2.8149,
      "theoretical_loss": 3.739863336354862,
      "tokens_seen": 777256960
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000510849725742387,
      "loss": 2.9205,
      "theoretical_loss": 3.7398319156070694,
      "tokens_seen": 777322496
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005108331757140155,
      "loss": 3.0183,
      "theoretical_loss": 3.739800498249917,
      "tokens_seen": 777388032
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000510816625685644,
      "loss": 2.7549,
      "theoretical_loss": 3.7397690842827536,
      "tokens_seen": 777453568
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005108000756572726,
      "loss": 2.6136,
      "theoretical_loss": 3.7397376737049277,
      "tokens_seen": 777519104
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005107835256289011,
      "loss": 2.7794,
      "theoretical_loss": 3.7397062665157876,
      "tokens_seen": 777584640
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005107669756005297,
      "loss": 2.5518,
      "theoretical_loss": 3.7396748627146827,
      "tokens_seen": 777650176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005107504255721582,
      "loss": 2.8449,
      "theoretical_loss": 3.739643462300961,
      "tokens_seen": 777715712
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005107338755437866,
      "loss": 2.832,
      "theoretical_loss": 3.739612065273973,
      "tokens_seen": 777781248
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005107173255154151,
      "loss": 2.7855,
      "theoretical_loss": 3.7395806716330675,
      "tokens_seen": 777846784
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005107007754870436,
      "loss": 2.6929,
      "theoretical_loss": 3.7395492813775943,
      "tokens_seen": 777912320
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005106842254586722,
      "loss": 2.9907,
      "theoretical_loss": 3.7395178945069025,
      "tokens_seen": 777977856
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005106676754303007,
      "loss": 2.9032,
      "theoretical_loss": 3.739486511020343,
      "tokens_seen": 778043392
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005106511254019292,
      "loss": 2.5905,
      "theoretical_loss": 3.739455130917265,
      "tokens_seen": 778108928
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005106345753735578,
      "loss": 2.8018,
      "theoretical_loss": 3.739423754197019,
      "tokens_seen": 778174464
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005106180253451862,
      "loss": 3.0092,
      "theoretical_loss": 3.7393923808589564,
      "tokens_seen": 778240000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005106014753168148,
      "loss": 2.7304,
      "theoretical_loss": 3.7393610109024267,
      "tokens_seen": 778305536
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005105849252884433,
      "loss": 2.9187,
      "theoretical_loss": 3.7393296443267814,
      "tokens_seen": 778371072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005105683752600718,
      "loss": 2.6998,
      "theoretical_loss": 3.7392982811313713,
      "tokens_seen": 778436608
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005105518252317004,
      "loss": 2.733,
      "theoretical_loss": 3.739266921315547,
      "tokens_seen": 778502144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005105352752033289,
      "loss": 2.8684,
      "theoretical_loss": 3.739235564878661,
      "tokens_seen": 778567680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005105187251749574,
      "loss": 2.651,
      "theoretical_loss": 3.7392042118200646,
      "tokens_seen": 778633216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000510502175146586,
      "loss": 2.7892,
      "theoretical_loss": 3.739172862139109,
      "tokens_seen": 778698752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005104856251182144,
      "loss": 2.7008,
      "theoretical_loss": 3.739141515835146,
      "tokens_seen": 778764288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000510469075089843,
      "loss": 2.9366,
      "theoretical_loss": 3.7391101729075285,
      "tokens_seen": 778829824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005104525250614715,
      "loss": 2.6826,
      "theoretical_loss": 3.7390788333556086,
      "tokens_seen": 778895360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005104359750331,
      "loss": 3.0577,
      "theoretical_loss": 3.7390474971787375,
      "tokens_seen": 778960896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005104194250047286,
      "loss": 3.067,
      "theoretical_loss": 3.7390161643762694,
      "tokens_seen": 779026432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005104028749763571,
      "loss": 2.5383,
      "theoretical_loss": 3.7389848349475567,
      "tokens_seen": 779091968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005103863249479857,
      "loss": 2.8246,
      "theoretical_loss": 3.738953508891952,
      "tokens_seen": 779157504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005103697749196141,
      "loss": 2.8604,
      "theoretical_loss": 3.7389221862088085,
      "tokens_seen": 779223040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005103532248912426,
      "loss": 2.7569,
      "theoretical_loss": 3.7388908668974796,
      "tokens_seen": 779288576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005103366748628712,
      "loss": 2.9527,
      "theoretical_loss": 3.738859550957319,
      "tokens_seen": 779354112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005103201248344997,
      "loss": 2.9284,
      "theoretical_loss": 3.73882823838768,
      "tokens_seen": 779419648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005103035748061282,
      "loss": 2.8812,
      "theoretical_loss": 3.7387969291879175,
      "tokens_seen": 779485184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005102870247777568,
      "loss": 2.8204,
      "theoretical_loss": 3.7387656233573843,
      "tokens_seen": 779550720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005102704747493853,
      "loss": 3.06,
      "theoretical_loss": 3.7387343208954356,
      "tokens_seen": 779616256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005102539247210139,
      "loss": 2.8024,
      "theoretical_loss": 3.7387030218014248,
      "tokens_seen": 779681792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005102373746926423,
      "loss": 2.7607,
      "theoretical_loss": 3.738671726074707,
      "tokens_seen": 779747328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005102208246642708,
      "loss": 2.8277,
      "theoretical_loss": 3.7386404337146373,
      "tokens_seen": 779812864
    },
    {
      "epoch": 0.28,
      "objective/train/avg_token_score": 0.023475702852010727,
      "objective/train/avg_weight": 0.9723640084266663,
      "objective/train/docs_used": 448564,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7229108810424805,
      "objective/train/std_weight": 0.10961704701185226,
      "objective/train/theoretical_loss": 3.73860914472057,
      "objective/train/tokens_used": 800338400,
      "theoretical_loss": 3.73860914472057,
      "tokens_seen": 779878400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005102042746358994,
      "loss": 2.7216,
      "theoretical_loss": 3.73860914472057,
      "tokens_seen": 779878400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005101877246075279,
      "loss": 2.8073,
      "theoretical_loss": 3.738577859091861,
      "tokens_seen": 779943936
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005101711745791564,
      "loss": 3.0046,
      "theoretical_loss": 3.738546576827865,
      "tokens_seen": 780009472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000510154624550785,
      "loss": 2.872,
      "theoretical_loss": 3.738515297927938,
      "tokens_seen": 780075008
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005101380745224135,
      "loss": 2.9401,
      "theoretical_loss": 3.7384840223914346,
      "tokens_seen": 780140544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000510121524494042,
      "loss": 2.8143,
      "theoretical_loss": 3.7384527502177116,
      "tokens_seen": 780206080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005101049744656705,
      "loss": 2.7848,
      "theoretical_loss": 3.738421481406125,
      "tokens_seen": 780271616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000510088424437299,
      "loss": 2.9159,
      "theoretical_loss": 3.7383902159560307,
      "tokens_seen": 780337152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005100718744089276,
      "loss": 2.6349,
      "theoretical_loss": 3.7383589538667845,
      "tokens_seen": 780402688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005100553243805561,
      "loss": 2.8121,
      "theoretical_loss": 3.7383276951377438,
      "tokens_seen": 780468224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005100387743521846,
      "loss": 2.9714,
      "theoretical_loss": 3.7382964397682654,
      "tokens_seen": 780533760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005100222243238132,
      "loss": 2.6505,
      "theoretical_loss": 3.7382651877577056,
      "tokens_seen": 780599296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005100056742954416,
      "loss": 2.7436,
      "theoretical_loss": 3.7382339391054216,
      "tokens_seen": 780664832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005099891242670702,
      "loss": 2.8058,
      "theoretical_loss": 3.738202693810771,
      "tokens_seen": 780730368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005099725742386987,
      "loss": 2.6733,
      "theoretical_loss": 3.738171451873111,
      "tokens_seen": 780795904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005099560242103272,
      "loss": 2.8059,
      "theoretical_loss": 3.738140213291799,
      "tokens_seen": 780861440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005099394741819558,
      "loss": 2.69,
      "theoretical_loss": 3.7381089780661925,
      "tokens_seen": 780926976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005099229241535843,
      "loss": 2.9561,
      "theoretical_loss": 3.7380777461956503,
      "tokens_seen": 780992512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005099063741252128,
      "loss": 2.5635,
      "theoretical_loss": 3.73804651767953,
      "tokens_seen": 781058048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005098898240968414,
      "loss": 2.6929,
      "theoretical_loss": 3.738015292517191,
      "tokens_seen": 781123584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005098732740684698,
      "loss": 2.8626,
      "theoretical_loss": 3.7379840707079897,
      "tokens_seen": 781189120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005098567240400984,
      "loss": 2.8674,
      "theoretical_loss": 3.7379528522512864,
      "tokens_seen": 781254656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005098401740117269,
      "loss": 2.7469,
      "theoretical_loss": 3.7379216371464388,
      "tokens_seen": 781320192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005098236239833554,
      "loss": 2.7333,
      "theoretical_loss": 3.7378904253928074,
      "tokens_seen": 781385728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000509807073954984,
      "loss": 2.8054,
      "theoretical_loss": 3.73785921698975,
      "tokens_seen": 781451264
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005097905239266125,
      "loss": 2.7163,
      "theoretical_loss": 3.7378280119366267,
      "tokens_seen": 781516800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000509773973898241,
      "loss": 2.6612,
      "theoretical_loss": 3.7377968102327968,
      "tokens_seen": 781582336
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005097574238698694,
      "loss": 2.9144,
      "theoretical_loss": 3.7377656118776206,
      "tokens_seen": 781647872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005097408738414979,
      "loss": 2.5864,
      "theoretical_loss": 3.737734416870457,
      "tokens_seen": 781713408
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005097243238131265,
      "loss": 3.0578,
      "theoretical_loss": 3.737703225210667,
      "tokens_seen": 781778944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000509707773784755,
      "loss": 2.6621,
      "theoretical_loss": 3.7376720368976093,
      "tokens_seen": 781844480
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005096912237563836,
      "loss": 2.7175,
      "theoretical_loss": 3.7376408519306463,
      "tokens_seen": 781910016
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005096746737280121,
      "loss": 2.5708,
      "theoretical_loss": 3.7376096703091375,
      "tokens_seen": 781975552
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005096581236996406,
      "loss": 2.5183,
      "theoretical_loss": 3.737578492032444,
      "tokens_seen": 782041088
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005096415736712691,
      "loss": 2.9141,
      "theoretical_loss": 3.7375473170999265,
      "tokens_seen": 782106624
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005096250236428976,
      "loss": 2.8637,
      "theoretical_loss": 3.7375161455109467,
      "tokens_seen": 782172160
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005096084736145261,
      "loss": 2.8288,
      "theoretical_loss": 3.737484977264865,
      "tokens_seen": 782237696
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005095919235861547,
      "loss": 2.6682,
      "theoretical_loss": 3.7374538123610432,
      "tokens_seen": 782303232
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005095753735577832,
      "loss": 2.74,
      "theoretical_loss": 3.7374226507988437,
      "tokens_seen": 782368768
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005095588235294118,
      "loss": 2.8397,
      "theoretical_loss": 3.7373914925776273,
      "tokens_seen": 782434304
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005095422735010403,
      "loss": 2.852,
      "theoretical_loss": 3.7373603376967566,
      "tokens_seen": 782499840
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005095257234726687,
      "loss": 2.5606,
      "theoretical_loss": 3.737329186155593,
      "tokens_seen": 782565376
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005095091734442973,
      "loss": 2.8086,
      "theoretical_loss": 3.7372980379535004,
      "tokens_seen": 782630912
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005094926234159258,
      "loss": 3.0298,
      "theoretical_loss": 3.73726689308984,
      "tokens_seen": 782696448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005094760733875543,
      "loss": 2.9463,
      "theoretical_loss": 3.7372357515639747,
      "tokens_seen": 782761984
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005094595233591829,
      "loss": 2.925,
      "theoretical_loss": 3.737204613375268,
      "tokens_seen": 782827520
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005094429733308114,
      "loss": 2.9419,
      "theoretical_loss": 3.7371734785230823,
      "tokens_seen": 782893056
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00050942642330244,
      "loss": 2.6224,
      "theoretical_loss": 3.737142347006781,
      "tokens_seen": 782958592
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005094098732740685,
      "loss": 2.7304,
      "theoretical_loss": 3.737111218825728,
      "tokens_seen": 783024128
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005093933232456969,
      "loss": 2.9582,
      "theoretical_loss": 3.737080093979286,
      "tokens_seen": 783089664
    },
    {
      "epoch": 0.28,
      "objective/train/avg_token_score": 0.016343753784894943,
      "objective/train/avg_weight": 0.980783998966217,
      "objective/train/docs_used": 450552,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5750229358673096,
      "objective/train/std_weight": 0.07510800659656525,
      "objective/train/theoretical_loss": 3.7370489724668197,
      "objective/train/tokens_used": 803615200,
      "theoretical_loss": 3.7370489724668197,
      "tokens_seen": 783155200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005093767732173255,
      "loss": 2.6094,
      "theoretical_loss": 3.7370489724668197,
      "tokens_seen": 783155200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000509360223188954,
      "loss": 2.7667,
      "theoretical_loss": 3.737017854287692,
      "tokens_seen": 783220736
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005093436731605825,
      "loss": 2.6631,
      "theoretical_loss": 3.7369867394412677,
      "tokens_seen": 783286272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005093271231322111,
      "loss": 2.5079,
      "theoretical_loss": 3.7369556279269105,
      "tokens_seen": 783351808
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005093105731038396,
      "loss": 2.7117,
      "theoretical_loss": 3.736924519743986,
      "tokens_seen": 783417344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005092940230754682,
      "loss": 2.7629,
      "theoretical_loss": 3.736893414891858,
      "tokens_seen": 783482880
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005092774730470966,
      "loss": 2.8646,
      "theoretical_loss": 3.7368623133698913,
      "tokens_seen": 783548416
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005092609230187251,
      "loss": 2.8228,
      "theoretical_loss": 3.7368312151774505,
      "tokens_seen": 783613952
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005092443729903537,
      "loss": 2.8748,
      "theoretical_loss": 3.7368001203139016,
      "tokens_seen": 783679488
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005092278229619822,
      "loss": 2.5737,
      "theoretical_loss": 3.7367690287786095,
      "tokens_seen": 783745024
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005092112729336107,
      "loss": 2.632,
      "theoretical_loss": 3.73673794057094,
      "tokens_seen": 783810560
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005091947229052393,
      "loss": 2.9741,
      "theoretical_loss": 3.736706855690258,
      "tokens_seen": 783876096
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005091781728768678,
      "loss": 2.8985,
      "theoretical_loss": 3.7366757741359304,
      "tokens_seen": 783941632
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005091616228484964,
      "loss": 3.0159,
      "theoretical_loss": 3.736644695907322,
      "tokens_seen": 784007168
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005091450728201248,
      "loss": 2.6937,
      "theoretical_loss": 3.7366136210038006,
      "tokens_seen": 784072704
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005091285227917533,
      "loss": 2.7572,
      "theoretical_loss": 3.736582549424731,
      "tokens_seen": 784138240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005091119727633819,
      "loss": 2.906,
      "theoretical_loss": 3.73655148116948,
      "tokens_seen": 784203776
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005090954227350104,
      "loss": 2.6553,
      "theoretical_loss": 3.7365204162374153,
      "tokens_seen": 784269312
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005090788727066389,
      "loss": 2.8658,
      "theoretical_loss": 3.736489354627903,
      "tokens_seen": 784334848
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005090788727066389,
      "loss": 2.7811,
      "theoretical_loss": 3.7364582963403103,
      "tokens_seen": 784400384
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005090623226782675,
      "loss": 2.7975,
      "theoretical_loss": 3.7364272413740047,
      "tokens_seen": 784465920
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000509045772649896,
      "loss": 2.8018,
      "theoretical_loss": 3.736396189728353,
      "tokens_seen": 784531456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005090292226215245,
      "loss": 2.8139,
      "theoretical_loss": 3.7363651414027235,
      "tokens_seen": 784596992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000509012672593153,
      "loss": 2.856,
      "theoretical_loss": 3.7363340963964835,
      "tokens_seen": 784662528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005089961225647815,
      "loss": 2.6503,
      "theoretical_loss": 3.7363030547090013,
      "tokens_seen": 784728064
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005089795725364101,
      "loss": 2.8402,
      "theoretical_loss": 3.736272016339644,
      "tokens_seen": 784793600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005089630225080386,
      "loss": 2.8905,
      "theoretical_loss": 3.7362409812877813,
      "tokens_seen": 784859136
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005089464724796672,
      "loss": 2.8628,
      "theoretical_loss": 3.7362099495527805,
      "tokens_seen": 784924672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005089299224512957,
      "loss": 2.6511,
      "theoretical_loss": 3.736178921134011,
      "tokens_seen": 784990208
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005089133724229241,
      "loss": 2.831,
      "theoretical_loss": 3.7361478960308414,
      "tokens_seen": 785055744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005088968223945527,
      "loss": 2.7786,
      "theoretical_loss": 3.73611687424264,
      "tokens_seen": 785121280
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005088802723661812,
      "loss": 2.6598,
      "theoretical_loss": 3.736085855768777,
      "tokens_seen": 785186816
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005088637223378097,
      "loss": 2.8936,
      "theoretical_loss": 3.736054840608621,
      "tokens_seen": 785252352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005088471723094383,
      "loss": 2.8533,
      "theoretical_loss": 3.7360238287615415,
      "tokens_seen": 785317888
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005088306222810668,
      "loss": 2.5781,
      "theoretical_loss": 3.7359928202269086,
      "tokens_seen": 785383424
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005088140722526954,
      "loss": 3.0528,
      "theoretical_loss": 3.735961815004092,
      "tokens_seen": 785448960
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005087975222243239,
      "loss": 2.6851,
      "theoretical_loss": 3.735930813092461,
      "tokens_seen": 785514496
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005087809721959523,
      "loss": 3.018,
      "theoretical_loss": 3.7358998144913866,
      "tokens_seen": 785580032
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005087644221675808,
      "loss": 2.6588,
      "theoretical_loss": 3.735868819200239,
      "tokens_seen": 785645568
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005087478721392093,
      "loss": 2.9071,
      "theoretical_loss": 3.7358378272183885,
      "tokens_seen": 785711104
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005087313221108379,
      "loss": 2.4797,
      "theoretical_loss": 3.735806838545206,
      "tokens_seen": 785776640
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005087147720824664,
      "loss": 2.5754,
      "theoretical_loss": 3.735775853180062,
      "tokens_seen": 785842176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005086982220540949,
      "loss": 2.9272,
      "theoretical_loss": 3.735744871122328,
      "tokens_seen": 785907712
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005086816720257235,
      "loss": 3.096,
      "theoretical_loss": 3.7357138923713746,
      "tokens_seen": 785973248
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005086651219973519,
      "loss": 2.8639,
      "theoretical_loss": 3.7356829169265744,
      "tokens_seen": 786038784
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005086485719689805,
      "loss": 2.7567,
      "theoretical_loss": 3.735651944787297,
      "tokens_seen": 786104320
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000508632021940609,
      "loss": 2.9008,
      "theoretical_loss": 3.7356209759529158,
      "tokens_seen": 786169856
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005086154719122375,
      "loss": 2.9144,
      "theoretical_loss": 3.7355900104228024,
      "tokens_seen": 786235392
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005085989218838661,
      "loss": 2.795,
      "theoretical_loss": 3.735559048196328,
      "tokens_seen": 786300928
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005085823718554946,
      "loss": 2.7458,
      "theoretical_loss": 3.735528089272866,
      "tokens_seen": 786366464
    },
    {
      "epoch": 0.28,
      "objective/train/avg_token_score": 0.02094729244709015,
      "objective/train/avg_weight": 0.9753710031509399,
      "objective/train/docs_used": 452240,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6858911514282227,
      "objective/train/std_weight": 0.09311465919017792,
      "objective/train/theoretical_loss": 3.735497133651788,
      "objective/train/tokens_used": 806892000,
      "theoretical_loss": 3.735497133651788,
      "tokens_seen": 786432000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005085658218271231,
      "loss": 2.5639,
      "theoretical_loss": 3.735497133651788,
      "tokens_seen": 786432000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005085492717987516,
      "loss": 2.8156,
      "theoretical_loss": 3.735466181332467,
      "tokens_seen": 786497536
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005085327217703801,
      "loss": 2.8498,
      "theoretical_loss": 3.735435232314275,
      "tokens_seen": 786563072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005085161717420087,
      "loss": 2.7533,
      "theoretical_loss": 3.7354042865965864,
      "tokens_seen": 786628608
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005084996217136372,
      "loss": 2.8369,
      "theoretical_loss": 3.7353733441787726,
      "tokens_seen": 786694144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005084830716852657,
      "loss": 2.7534,
      "theoretical_loss": 3.735342405060208,
      "tokens_seen": 786759680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005084665216568943,
      "loss": 2.8989,
      "theoretical_loss": 3.7353114692402656,
      "tokens_seen": 786825216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005084499716285228,
      "loss": 2.7233,
      "theoretical_loss": 3.7352805367183195,
      "tokens_seen": 786890752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005084334216001512,
      "loss": 2.6219,
      "theoretical_loss": 3.7352496074937425,
      "tokens_seen": 786956288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005084168715717798,
      "loss": 2.7922,
      "theoretical_loss": 3.735218681565909,
      "tokens_seen": 787021824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005084003215434083,
      "loss": 2.8086,
      "theoretical_loss": 3.7351877589341944,
      "tokens_seen": 787087360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005083837715150369,
      "loss": 2.6288,
      "theoretical_loss": 3.7351568395979706,
      "tokens_seen": 787152896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005083672214866654,
      "loss": 2.7166,
      "theoretical_loss": 3.7351259235566143,
      "tokens_seen": 787218432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005083506714582939,
      "loss": 2.7649,
      "theoretical_loss": 3.7350950108094985,
      "tokens_seen": 787283968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005083341214299225,
      "loss": 2.784,
      "theoretical_loss": 3.7350641013559986,
      "tokens_seen": 787349504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000508317571401551,
      "loss": 2.8607,
      "theoretical_loss": 3.73503319519549,
      "tokens_seen": 787415040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005083010213731794,
      "loss": 2.9051,
      "theoretical_loss": 3.7350022923273474,
      "tokens_seen": 787480576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000508284471344808,
      "loss": 2.7646,
      "theoretical_loss": 3.7349713927509463,
      "tokens_seen": 787546112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005082679213164365,
      "loss": 2.485,
      "theoretical_loss": 3.7349404964656614,
      "tokens_seen": 787611648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005082513712880651,
      "loss": 2.7225,
      "theoretical_loss": 3.7349096034708698,
      "tokens_seen": 787677184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005082348212596936,
      "loss": 2.7853,
      "theoretical_loss": 3.734878713765946,
      "tokens_seen": 787742720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005082182712313221,
      "loss": 2.7866,
      "theoretical_loss": 3.734847827350267,
      "tokens_seen": 787808256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005082017212029507,
      "loss": 2.6232,
      "theoretical_loss": 3.7348169442232084,
      "tokens_seen": 787873792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005081851711745792,
      "loss": 2.7415,
      "theoretical_loss": 3.7347860643841466,
      "tokens_seen": 787939328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005081686211462076,
      "loss": 2.8726,
      "theoretical_loss": 3.7347551878324587,
      "tokens_seen": 788004864
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005081520711178362,
      "loss": 2.6169,
      "theoretical_loss": 3.73472431456752,
      "tokens_seen": 788070400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005081355210894647,
      "loss": 3.1118,
      "theoretical_loss": 3.7346934445887086,
      "tokens_seen": 788135936
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005081189710610933,
      "loss": 2.8472,
      "theoretical_loss": 3.734662577895401,
      "tokens_seen": 788201472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005081024210327218,
      "loss": 2.7496,
      "theoretical_loss": 3.7346317144869747,
      "tokens_seen": 788267008
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005080858710043503,
      "loss": 2.6043,
      "theoretical_loss": 3.7346008543628066,
      "tokens_seen": 788332544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005080693209759789,
      "loss": 2.9031,
      "theoretical_loss": 3.734569997522275,
      "tokens_seen": 788398080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005080527709476073,
      "loss": 2.7736,
      "theoretical_loss": 3.734539143964757,
      "tokens_seen": 788463616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005080362209192358,
      "loss": 2.5802,
      "theoretical_loss": 3.73450829368963,
      "tokens_seen": 788529152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005080196708908644,
      "loss": 2.6006,
      "theoretical_loss": 3.734477446696273,
      "tokens_seen": 788594688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005080031208624929,
      "loss": 2.9312,
      "theoretical_loss": 3.7344466029840637,
      "tokens_seen": 788660224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005079865708341215,
      "loss": 2.7759,
      "theoretical_loss": 3.7344157625523806,
      "tokens_seen": 788725760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00050797002080575,
      "loss": 2.6871,
      "theoretical_loss": 3.7343849254006027,
      "tokens_seen": 788791296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005079534707773785,
      "loss": 2.619,
      "theoretical_loss": 3.7343540915281075,
      "tokens_seen": 788856832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000507936920749007,
      "loss": 2.7051,
      "theoretical_loss": 3.734323260934275,
      "tokens_seen": 788922368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005079203707206355,
      "loss": 2.8966,
      "theoretical_loss": 3.734292433618484,
      "tokens_seen": 788987904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000507903820692264,
      "loss": 3.0059,
      "theoretical_loss": 3.7342616095801136,
      "tokens_seen": 789053440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005078872706638926,
      "loss": 2.7692,
      "theoretical_loss": 3.734230788818543,
      "tokens_seen": 789118976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005078707206355211,
      "loss": 2.7194,
      "theoretical_loss": 3.734199971333152,
      "tokens_seen": 789184512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005078541706071497,
      "loss": 2.9164,
      "theoretical_loss": 3.73416915712332,
      "tokens_seen": 789250048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005078376205787782,
      "loss": 2.8762,
      "theoretical_loss": 3.734138346188428,
      "tokens_seen": 789315584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005078210705504067,
      "loss": 2.9138,
      "theoretical_loss": 3.734107538527854,
      "tokens_seen": 789381120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005078045205220352,
      "loss": 2.8538,
      "theoretical_loss": 3.734076734140981,
      "tokens_seen": 789446656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005077879704936636,
      "loss": 2.9074,
      "theoretical_loss": 3.7340459330271867,
      "tokens_seen": 789512192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005077714204652922,
      "loss": 2.719,
      "theoretical_loss": 3.7340151351858535,
      "tokens_seen": 789577728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005077548704369207,
      "loss": 2.7959,
      "theoretical_loss": 3.7339843406163613,
      "tokens_seen": 789643264
    },
    {
      "epoch": 0.28,
      "objective/train/avg_token_score": 0.010169348679482937,
      "objective/train/avg_weight": 0.9880306124687195,
      "objective/train/docs_used": 454142,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7576723098754883,
      "objective/train/std_weight": 0.049474943429231644,
      "objective/train/theoretical_loss": 3.733953549318091,
      "objective/train/tokens_used": 810168800,
      "theoretical_loss": 3.733953549318091,
      "tokens_seen": 789708800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005077383204085493,
      "loss": 2.5722,
      "theoretical_loss": 3.733953549318091,
      "tokens_seen": 789708800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005077217703801778,
      "loss": 2.7234,
      "theoretical_loss": 3.733922761290424,
      "tokens_seen": 789774336
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005077052203518063,
      "loss": 2.6306,
      "theoretical_loss": 3.733891976532742,
      "tokens_seen": 789839872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005076886703234348,
      "loss": 2.7926,
      "theoretical_loss": 3.733861195044425,
      "tokens_seen": 789905408
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005076721202950633,
      "loss": 2.839,
      "theoretical_loss": 3.7338304168248566,
      "tokens_seen": 789970944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005076555702666918,
      "loss": 2.6456,
      "theoretical_loss": 3.733799641873417,
      "tokens_seen": 790036480
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005076390202383204,
      "loss": 2.5785,
      "theoretical_loss": 3.733768870189488,
      "tokens_seen": 790102016
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005076224702099489,
      "loss": 3.0392,
      "theoretical_loss": 3.7337381017724525,
      "tokens_seen": 790167552
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005076059201815775,
      "loss": 2.709,
      "theoretical_loss": 3.733707336621692,
      "tokens_seen": 790233088
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000507589370153206,
      "loss": 2.8512,
      "theoretical_loss": 3.7336765747365903,
      "tokens_seen": 790298624
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005075728201248344,
      "loss": 2.717,
      "theoretical_loss": 3.7336458161165287,
      "tokens_seen": 790364160
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000507556270096463,
      "loss": 2.9743,
      "theoretical_loss": 3.73361506076089,
      "tokens_seen": 790429696
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005075397200680915,
      "loss": 2.7413,
      "theoretical_loss": 3.7335843086690583,
      "tokens_seen": 790495232
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00050752317003972,
      "loss": 2.8302,
      "theoretical_loss": 3.7335535598404155,
      "tokens_seen": 790560768
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005075066200113486,
      "loss": 2.8572,
      "theoretical_loss": 3.733522814274345,
      "tokens_seen": 790626304
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005074900699829771,
      "loss": 2.8295,
      "theoretical_loss": 3.7334920719702307,
      "tokens_seen": 790691840
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005074735199546057,
      "loss": 2.8481,
      "theoretical_loss": 3.733461332927456,
      "tokens_seen": 790757376
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005074569699262342,
      "loss": 2.9271,
      "theoretical_loss": 3.7334305971454045,
      "tokens_seen": 790822912
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005074404198978626,
      "loss": 2.897,
      "theoretical_loss": 3.7333998646234603,
      "tokens_seen": 790888448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005074238698694912,
      "loss": 2.827,
      "theoretical_loss": 3.7333691353610075,
      "tokens_seen": 790953984
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005074073198411197,
      "loss": 2.8209,
      "theoretical_loss": 3.7333384093574304,
      "tokens_seen": 791019520
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005073907698127482,
      "loss": 2.6626,
      "theoretical_loss": 3.7333076866121138,
      "tokens_seen": 791085056
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005073742197843768,
      "loss": 2.6877,
      "theoretical_loss": 3.733276967124441,
      "tokens_seen": 791150592
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005073576697560053,
      "loss": 2.6521,
      "theoretical_loss": 3.7332462508937985,
      "tokens_seen": 791216128
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005073411197276339,
      "loss": 2.7224,
      "theoretical_loss": 3.7332155379195697,
      "tokens_seen": 791281664
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005073245696992623,
      "loss": 2.7951,
      "theoretical_loss": 3.733184828201141,
      "tokens_seen": 791347200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005073080196708908,
      "loss": 3.0435,
      "theoretical_loss": 3.733154121737897,
      "tokens_seen": 791412736
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005072914696425194,
      "loss": 2.8497,
      "theoretical_loss": 3.733123418529223,
      "tokens_seen": 791478272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005072749196141479,
      "loss": 2.8359,
      "theoretical_loss": 3.733092718574505,
      "tokens_seen": 791543808
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005072583695857764,
      "loss": 2.5656,
      "theoretical_loss": 3.7330620218731285,
      "tokens_seen": 791609344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000507241819557405,
      "loss": 2.7202,
      "theoretical_loss": 3.7330313284244796,
      "tokens_seen": 791674880
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005072252695290335,
      "loss": 2.7846,
      "theoretical_loss": 3.7330006382279444,
      "tokens_seen": 791740416
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000507208719500662,
      "loss": 2.821,
      "theoretical_loss": 3.732969951282909,
      "tokens_seen": 791805952
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005071921694722905,
      "loss": 2.8415,
      "theoretical_loss": 3.73293926758876,
      "tokens_seen": 791871488
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000507175619443919,
      "loss": 2.6507,
      "theoretical_loss": 3.732908587144884,
      "tokens_seen": 791937024
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005071590694155476,
      "loss": 2.9444,
      "theoretical_loss": 3.732877909950668,
      "tokens_seen": 792002560
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005071425193871761,
      "loss": 2.7707,
      "theoretical_loss": 3.7328472360054983,
      "tokens_seen": 792068096
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005071259693588046,
      "loss": 2.8225,
      "theoretical_loss": 3.7328165653087626,
      "tokens_seen": 792133632
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005071094193304332,
      "loss": 2.7168,
      "theoretical_loss": 3.7327858978598476,
      "tokens_seen": 792199168
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005070928693020617,
      "loss": 2.7339,
      "theoretical_loss": 3.7327552336581418,
      "tokens_seen": 792264704
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005070763192736902,
      "loss": 2.6631,
      "theoretical_loss": 3.732724572703032,
      "tokens_seen": 792330240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005070597692453187,
      "loss": 2.6627,
      "theoretical_loss": 3.732693914993906,
      "tokens_seen": 792395776
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005070432192169472,
      "loss": 2.7381,
      "theoretical_loss": 3.732663260530151,
      "tokens_seen": 792461312
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005070266691885758,
      "loss": 2.9183,
      "theoretical_loss": 3.732632609311157,
      "tokens_seen": 792526848
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005070101191602043,
      "loss": 2.9024,
      "theoretical_loss": 3.7326019613363104,
      "tokens_seen": 792592384
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005069935691318328,
      "loss": 2.9213,
      "theoretical_loss": 3.732571316605001,
      "tokens_seen": 792657920
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005069770191034614,
      "loss": 3.0689,
      "theoretical_loss": 3.7325406751166166,
      "tokens_seen": 792723456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005069604690750898,
      "loss": 2.9329,
      "theoretical_loss": 3.732510036870546,
      "tokens_seen": 792788992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005069439190467184,
      "loss": 2.7308,
      "theoretical_loss": 3.7324794018661787,
      "tokens_seen": 792854528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005069273690183469,
      "loss": 3.035,
      "theoretical_loss": 3.732448770102903,
      "tokens_seen": 792920064
    },
    {
      "epoch": 0.28,
      "objective/train/avg_token_score": 0.023099083453416824,
      "objective/train/avg_weight": 0.9728018045425415,
      "objective/train/docs_used": 455856,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5004849433898926,
      "objective/train/std_weight": 0.11675141751766205,
      "objective/train/theoretical_loss": 3.7324181415801094,
      "objective/train/tokens_used": 813445600,
      "theoretical_loss": 3.7324181415801094,
      "tokens_seen": 792985600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005069108189899754,
      "loss": 2.535,
      "theoretical_loss": 3.7324181415801094,
      "tokens_seen": 792985600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000506894268961604,
      "loss": 2.7402,
      "theoretical_loss": 3.7323875162971856,
      "tokens_seen": 793051136
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005068777189332325,
      "loss": 2.8247,
      "theoretical_loss": 3.7323568942535226,
      "tokens_seen": 793116672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000506861168904861,
      "loss": 2.7273,
      "theoretical_loss": 3.73232627544851,
      "tokens_seen": 793182208
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005068446188764896,
      "loss": 2.4947,
      "theoretical_loss": 3.7322956598815367,
      "tokens_seen": 793247744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000506828068848118,
      "loss": 2.8238,
      "theoretical_loss": 3.7322650475519934,
      "tokens_seen": 793313280
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005068115188197465,
      "loss": 2.6992,
      "theoretical_loss": 3.7322344384592707,
      "tokens_seen": 793378816
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000506794968791375,
      "loss": 2.7795,
      "theoretical_loss": 3.7322038326027593,
      "tokens_seen": 793444352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005067784187630036,
      "loss": 2.8302,
      "theoretical_loss": 3.7321732299818486,
      "tokens_seen": 793509888
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005067618687346321,
      "loss": 2.7825,
      "theoretical_loss": 3.73214263059593,
      "tokens_seen": 793575424
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005067453187062606,
      "loss": 2.8674,
      "theoretical_loss": 3.7321120344443948,
      "tokens_seen": 793640960
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005067287686778892,
      "loss": 2.4278,
      "theoretical_loss": 3.732081441526634,
      "tokens_seen": 793706496
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005067122186495176,
      "loss": 2.8525,
      "theoretical_loss": 3.732050851842038,
      "tokens_seen": 793772032
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005066956686211461,
      "loss": 2.8209,
      "theoretical_loss": 3.7320202653899983,
      "tokens_seen": 793837568
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005066791185927747,
      "loss": 2.8319,
      "theoretical_loss": 3.7319896821699077,
      "tokens_seen": 793903104
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005066625685644032,
      "loss": 2.9176,
      "theoretical_loss": 3.731959102181157,
      "tokens_seen": 793968640
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005066460185360318,
      "loss": 2.8143,
      "theoretical_loss": 3.7319285254231382,
      "tokens_seen": 794034176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005066294685076603,
      "loss": 2.6491,
      "theoretical_loss": 3.7318979518952435,
      "tokens_seen": 794099712
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005066129184792888,
      "loss": 3.1138,
      "theoretical_loss": 3.7318673815968655,
      "tokens_seen": 794165248
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005065963684509173,
      "loss": 2.677,
      "theoretical_loss": 3.7318368145273952,
      "tokens_seen": 794230784
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005065798184225458,
      "loss": 3.0012,
      "theoretical_loss": 3.731806250686227,
      "tokens_seen": 794296320
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005065632683941743,
      "loss": 2.845,
      "theoretical_loss": 3.7317756900727526,
      "tokens_seen": 794361856
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005065467183658029,
      "loss": 2.7642,
      "theoretical_loss": 3.731745132686365,
      "tokens_seen": 794427392
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005065301683374314,
      "loss": 2.6622,
      "theoretical_loss": 3.731714578526457,
      "tokens_seen": 794492928
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00050651361830906,
      "loss": 2.795,
      "theoretical_loss": 3.7316840275924226,
      "tokens_seen": 794558464
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005064970682806885,
      "loss": 2.7976,
      "theoretical_loss": 3.7316534798836543,
      "tokens_seen": 794624000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005064805182523169,
      "loss": 2.6755,
      "theoretical_loss": 3.731622935399547,
      "tokens_seen": 794689536
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005064639682239455,
      "loss": 2.8605,
      "theoretical_loss": 3.7315923941394926,
      "tokens_seen": 794755072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000506447418195574,
      "loss": 2.7702,
      "theoretical_loss": 3.7315618561028865,
      "tokens_seen": 794820608
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005064308681672026,
      "loss": 2.7442,
      "theoretical_loss": 3.731531321289121,
      "tokens_seen": 794886144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005064143181388311,
      "loss": 2.9046,
      "theoretical_loss": 3.7315007896975927,
      "tokens_seen": 794951680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005063977681104596,
      "loss": 2.6212,
      "theoretical_loss": 3.731470261327694,
      "tokens_seen": 795017216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005063812180820882,
      "loss": 2.7951,
      "theoretical_loss": 3.73143973617882,
      "tokens_seen": 795082752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005063646680537167,
      "loss": 2.9512,
      "theoretical_loss": 3.731409214250366,
      "tokens_seen": 795148288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005063481180253451,
      "loss": 2.6436,
      "theoretical_loss": 3.731378695541726,
      "tokens_seen": 795213824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005063315679969737,
      "loss": 2.9036,
      "theoretical_loss": 3.7313481800522954,
      "tokens_seen": 795279360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005063150179686022,
      "loss": 2.7672,
      "theoretical_loss": 3.73131766778147,
      "tokens_seen": 795344896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005062984679402308,
      "loss": 2.8412,
      "theoretical_loss": 3.7312871587286436,
      "tokens_seen": 795410432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005062819179118593,
      "loss": 3.1466,
      "theoretical_loss": 3.7312566528932134,
      "tokens_seen": 795475968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005062653678834878,
      "loss": 2.664,
      "theoretical_loss": 3.7312261502745736,
      "tokens_seen": 795541504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005062488178551164,
      "loss": 2.9513,
      "theoretical_loss": 3.731195650872121,
      "tokens_seen": 795607040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005062322678267448,
      "loss": 2.8183,
      "theoretical_loss": 3.731165154685252,
      "tokens_seen": 795672576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005062157177983733,
      "loss": 2.6479,
      "theoretical_loss": 3.7311346617133614,
      "tokens_seen": 795738112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005061991677700019,
      "loss": 2.7594,
      "theoretical_loss": 3.731104171955847,
      "tokens_seen": 795803648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005061826177416304,
      "loss": 2.7375,
      "theoretical_loss": 3.731073685412104,
      "tokens_seen": 795869184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000506166067713259,
      "loss": 2.853,
      "theoretical_loss": 3.73104320208153,
      "tokens_seen": 795934720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005061495176848875,
      "loss": 2.8991,
      "theoretical_loss": 3.7310127219635207,
      "tokens_seen": 796000256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000506132967656516,
      "loss": 2.9063,
      "theoretical_loss": 3.730982245057474,
      "tokens_seen": 796065792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005061164176281446,
      "loss": 2.9106,
      "theoretical_loss": 3.730951771362787,
      "tokens_seen": 796131328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000506099867599773,
      "loss": 2.6994,
      "theoretical_loss": 3.7309213008788573,
      "tokens_seen": 796196864
    },
    {
      "epoch": 0.28,
      "objective/train/avg_token_score": 0.01525387354195118,
      "objective/train/avg_weight": 0.9820623397827148,
      "objective/train/docs_used": 457654,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.707221031188965,
      "objective/train/std_weight": 0.0881563350558281,
      "objective/train/theoretical_loss": 3.7308908336050814,
      "objective/train/tokens_used": 816722400,
      "theoretical_loss": 3.7308908336050814,
      "tokens_seen": 796262400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005060833175714015,
      "loss": 2.7381,
      "theoretical_loss": 3.7308908336050814,
      "tokens_seen": 796262400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005060667675430301,
      "loss": 2.7566,
      "theoretical_loss": 3.730860369540858,
      "tokens_seen": 796327936
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005060502175146586,
      "loss": 2.7453,
      "theoretical_loss": 3.7308299086855836,
      "tokens_seen": 796393472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005060336674862872,
      "loss": 2.9676,
      "theoretical_loss": 3.7307994510386577,
      "tokens_seen": 796459008
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005060171174579157,
      "loss": 2.8261,
      "theoretical_loss": 3.730768996599477,
      "tokens_seen": 796524544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005060005674295442,
      "loss": 2.9249,
      "theoretical_loss": 3.7307385453674407,
      "tokens_seen": 796590080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005059840174011727,
      "loss": 2.626,
      "theoretical_loss": 3.730708097341947,
      "tokens_seen": 796655616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005059674673728012,
      "loss": 2.5825,
      "theoretical_loss": 3.7306776525223944,
      "tokens_seen": 796721152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005059509173444297,
      "loss": 2.9566,
      "theoretical_loss": 3.730647210908182,
      "tokens_seen": 796786688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005059343673160583,
      "loss": 2.9022,
      "theoretical_loss": 3.730616772498708,
      "tokens_seen": 796852224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005059178172876868,
      "loss": 2.6881,
      "theoretical_loss": 3.7305863372933725,
      "tokens_seen": 796917760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005059012672593154,
      "loss": 2.6634,
      "theoretical_loss": 3.730555905291574,
      "tokens_seen": 796983296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005058847172309439,
      "loss": 2.5637,
      "theoretical_loss": 3.730525476492712,
      "tokens_seen": 797048832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005058681672025723,
      "loss": 2.8836,
      "theoretical_loss": 3.7304950508961863,
      "tokens_seen": 797114368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005058516171742009,
      "loss": 3.0302,
      "theoretical_loss": 3.7304646285013963,
      "tokens_seen": 797179904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005058350671458294,
      "loss": 3.0508,
      "theoretical_loss": 3.7304342093077425,
      "tokens_seen": 797245440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005058185171174579,
      "loss": 2.6775,
      "theoretical_loss": 3.7304037933146246,
      "tokens_seen": 797310976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005058019670890864,
      "loss": 2.8155,
      "theoretical_loss": 3.7303733805214425,
      "tokens_seen": 797376512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005057854170607149,
      "loss": 2.667,
      "theoretical_loss": 3.7303429709275973,
      "tokens_seen": 797442048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005057688670323435,
      "loss": 2.9966,
      "theoretical_loss": 3.730312564532489,
      "tokens_seen": 797507584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000505752317003972,
      "loss": 2.8228,
      "theoretical_loss": 3.7302821613355186,
      "tokens_seen": 797573120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005057357669756005,
      "loss": 2.8158,
      "theoretical_loss": 3.7302517613360866,
      "tokens_seen": 797638656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000505719216947229,
      "loss": 2.7997,
      "theoretical_loss": 3.7302213645335947,
      "tokens_seen": 797704192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005057026669188575,
      "loss": 2.925,
      "theoretical_loss": 3.7301909709274437,
      "tokens_seen": 797769728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005056861168904861,
      "loss": 2.8615,
      "theoretical_loss": 3.730160580517035,
      "tokens_seen": 797835264
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005056695668621146,
      "loss": 2.5203,
      "theoretical_loss": 3.73013019330177,
      "tokens_seen": 797900800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0005056530168337431,
      "loss": 2.6011,
      "theoretical_loss": 3.7300998092810502,
      "tokens_seen": 797966336
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005056364668053717,
      "loss": 2.7612,
      "theoretical_loss": 3.7300694284542786,
      "tokens_seen": 798031872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005056199167770001,
      "loss": 2.5061,
      "theoretical_loss": 3.730039050820855,
      "tokens_seen": 798097408
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005056033667486287,
      "loss": 2.7476,
      "theoretical_loss": 3.7300086763801836,
      "tokens_seen": 798162944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005055868167202572,
      "loss": 2.8393,
      "theoretical_loss": 3.729978305131666,
      "tokens_seen": 798228480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005055702666918857,
      "loss": 3.0329,
      "theoretical_loss": 3.729947937074705,
      "tokens_seen": 798294016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005055537166635143,
      "loss": 2.9494,
      "theoretical_loss": 3.7299175722087026,
      "tokens_seen": 798359552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005055371666351428,
      "loss": 2.8395,
      "theoretical_loss": 3.729887210533062,
      "tokens_seen": 798425088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005055206166067713,
      "loss": 2.8435,
      "theoretical_loss": 3.7298568520471855,
      "tokens_seen": 798490624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005055040665783998,
      "loss": 2.9479,
      "theoretical_loss": 3.7298264967504773,
      "tokens_seen": 798556160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005054875165500283,
      "loss": 2.6912,
      "theoretical_loss": 3.72979614464234,
      "tokens_seen": 798621696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005054709665216569,
      "loss": 2.7865,
      "theoretical_loss": 3.7297657957221784,
      "tokens_seen": 798687232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005054544164932854,
      "loss": 2.8575,
      "theoretical_loss": 3.729735449989394,
      "tokens_seen": 798752768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005054378664649139,
      "loss": 2.7816,
      "theoretical_loss": 3.729705107443391,
      "tokens_seen": 798818304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005054213164365425,
      "loss": 2.815,
      "theoretical_loss": 3.7296747680835747,
      "tokens_seen": 798883840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000505404766408171,
      "loss": 2.6552,
      "theoretical_loss": 3.7296444319093487,
      "tokens_seen": 798949376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005053882163797994,
      "loss": 2.7222,
      "theoretical_loss": 3.7296140989201163,
      "tokens_seen": 799014912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000505371666351428,
      "loss": 2.8027,
      "theoretical_loss": 3.7295837691152824,
      "tokens_seen": 799080448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005053551163230565,
      "loss": 3.0255,
      "theoretical_loss": 3.729553442494252,
      "tokens_seen": 799145984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005053385662946851,
      "loss": 2.6326,
      "theoretical_loss": 3.72952311905643,
      "tokens_seen": 799211520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005053220162663136,
      "loss": 2.731,
      "theoretical_loss": 3.7294927988012203,
      "tokens_seen": 799277056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005053054662379421,
      "loss": 2.7023,
      "theoretical_loss": 3.729462481728029,
      "tokens_seen": 799342592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005052889162095707,
      "loss": 2.8262,
      "theoretical_loss": 3.7294321678362596,
      "tokens_seen": 799408128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005052723661811992,
      "loss": 2.7355,
      "theoretical_loss": 3.72940185712532,
      "tokens_seen": 799473664
    },
    {
      "epoch": 0.29,
      "objective/train/avg_token_score": 0.01235914696007967,
      "objective/train/avg_weight": 0.9854694604873657,
      "objective/train/docs_used": 459747,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.907264232635498,
      "objective/train/std_weight": 0.08462017774581909,
      "objective/train/theoretical_loss": 3.729371549594614,
      "objective/train/tokens_used": 819999200,
      "theoretical_loss": 3.729371549594614,
      "tokens_seen": 799539200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005052558161528276,
      "loss": 2.7546,
      "theoretical_loss": 3.729371549594614,
      "tokens_seen": 799539200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005052392661244562,
      "loss": 2.9199,
      "theoretical_loss": 3.729341245243547,
      "tokens_seen": 799604736
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005052227160960847,
      "loss": 2.8583,
      "theoretical_loss": 3.7293109440715266,
      "tokens_seen": 799670272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005052061660677133,
      "loss": 2.7619,
      "theoretical_loss": 3.729280646077957,
      "tokens_seen": 799735808
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005051896160393418,
      "loss": 2.6187,
      "theoretical_loss": 3.7292503512622455,
      "tokens_seen": 799801344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005051730660109703,
      "loss": 2.7586,
      "theoretical_loss": 3.729220059623798,
      "tokens_seen": 799866880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005051565159825989,
      "loss": 2.5366,
      "theoretical_loss": 3.729189771162021,
      "tokens_seen": 799932416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005051399659542273,
      "loss": 2.7736,
      "theoretical_loss": 3.7291594858763215,
      "tokens_seen": 799997952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005051234159258558,
      "loss": 2.8778,
      "theoretical_loss": 3.7291292037661057,
      "tokens_seen": 800063488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005051068658974844,
      "loss": 2.8607,
      "theoretical_loss": 3.7290989248307813,
      "tokens_seen": 800129024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005050903158691129,
      "loss": 2.7237,
      "theoretical_loss": 3.729068649069754,
      "tokens_seen": 800194560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005050737658407415,
      "loss": 2.7445,
      "theoretical_loss": 3.729038376482433,
      "tokens_seen": 800260096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00050505721581237,
      "loss": 2.9267,
      "theoretical_loss": 3.729008107068225,
      "tokens_seen": 800325632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005050406657839985,
      "loss": 2.6681,
      "theoretical_loss": 3.728977840826537,
      "tokens_seen": 800391168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005050241157556271,
      "loss": 2.7893,
      "theoretical_loss": 3.7289475777567773,
      "tokens_seen": 800456704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005050075657272555,
      "loss": 2.8328,
      "theoretical_loss": 3.728917317858354,
      "tokens_seen": 800522240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000504991015698884,
      "loss": 2.8793,
      "theoretical_loss": 3.7288870611306746,
      "tokens_seen": 800587776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005049744656705126,
      "loss": 2.8232,
      "theoretical_loss": 3.7288568075731474,
      "tokens_seen": 800653312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005049579156421411,
      "loss": 2.8708,
      "theoretical_loss": 3.7288265571851817,
      "tokens_seen": 800718848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005049413656137697,
      "loss": 2.7097,
      "theoretical_loss": 3.728796309966185,
      "tokens_seen": 800784384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005049248155853982,
      "loss": 3.0123,
      "theoretical_loss": 3.7287660659155666,
      "tokens_seen": 800849920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005049082655570267,
      "loss": 2.8209,
      "theoretical_loss": 3.7287358250327354,
      "tokens_seen": 800915456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005048917155286552,
      "loss": 3.1771,
      "theoretical_loss": 3.7287055873171,
      "tokens_seen": 800980992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005048751655002837,
      "loss": 2.7932,
      "theoretical_loss": 3.7286753527680703,
      "tokens_seen": 801046528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005048586154719123,
      "loss": 2.7392,
      "theoretical_loss": 3.7286451213850547,
      "tokens_seen": 801112064
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005048420654435407,
      "loss": 2.868,
      "theoretical_loss": 3.7286148931674634,
      "tokens_seen": 801177600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005048255154151693,
      "loss": 2.8806,
      "theoretical_loss": 3.728584668114706,
      "tokens_seen": 801243136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005048089653867978,
      "loss": 2.8785,
      "theoretical_loss": 3.728554446226192,
      "tokens_seen": 801308672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005047924153584263,
      "loss": 2.8341,
      "theoretical_loss": 3.7285242275013317,
      "tokens_seen": 801374208
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005047758653300548,
      "loss": 2.6237,
      "theoretical_loss": 3.728494011939535,
      "tokens_seen": 801439744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005047593153016833,
      "loss": 2.865,
      "theoretical_loss": 3.7284637995402123,
      "tokens_seen": 801505280
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005047427652733118,
      "loss": 2.8683,
      "theoretical_loss": 3.7284335903027745,
      "tokens_seen": 801570816
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005047262152449404,
      "loss": 2.7766,
      "theoretical_loss": 3.728403384226632,
      "tokens_seen": 801636352
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005047096652165689,
      "loss": 2.5524,
      "theoretical_loss": 3.728373181311195,
      "tokens_seen": 801701888
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005046931151881975,
      "loss": 2.9268,
      "theoretical_loss": 3.728342981555875,
      "tokens_seen": 801767424
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000504676565159826,
      "loss": 2.7841,
      "theoretical_loss": 3.7283127849600826,
      "tokens_seen": 801832960
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005046600151314544,
      "loss": 2.7202,
      "theoretical_loss": 3.7282825915232296,
      "tokens_seen": 801898496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000504643465103083,
      "loss": 2.7236,
      "theoretical_loss": 3.728252401244727,
      "tokens_seen": 801964032
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005046269150747115,
      "loss": 2.6483,
      "theoretical_loss": 3.7282222141239867,
      "tokens_seen": 802029568
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00050461036504634,
      "loss": 2.7578,
      "theoretical_loss": 3.7281920301604208,
      "tokens_seen": 802095104
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005045938150179686,
      "loss": 2.6248,
      "theoretical_loss": 3.72816184935344,
      "tokens_seen": 802160640
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005045772649895971,
      "loss": 2.6896,
      "theoretical_loss": 3.728131671702457,
      "tokens_seen": 802226176
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005045607149612257,
      "loss": 3.0626,
      "theoretical_loss": 3.728101497206884,
      "tokens_seen": 802291712
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005045441649328542,
      "loss": 2.8339,
      "theoretical_loss": 3.728071325866134,
      "tokens_seen": 802357248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005045276149044826,
      "loss": 2.8902,
      "theoretical_loss": 3.7280411576796184,
      "tokens_seen": 802422784
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005045110648761112,
      "loss": 2.8989,
      "theoretical_loss": 3.7280109926467504,
      "tokens_seen": 802488320
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005044945148477397,
      "loss": 2.9666,
      "theoretical_loss": 3.727980830766943,
      "tokens_seen": 802553856
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005044779648193682,
      "loss": 2.7959,
      "theoretical_loss": 3.7279506720396083,
      "tokens_seen": 802619392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005044614147909968,
      "loss": 2.9966,
      "theoretical_loss": 3.7279205164641604,
      "tokens_seen": 802684928
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005044448647626253,
      "loss": 2.4908,
      "theoretical_loss": 3.7278903640400127,
      "tokens_seen": 802750464
    },
    {
      "epoch": 0.29,
      "objective/train/avg_token_score": 0.023677125573158264,
      "objective/train/avg_weight": 0.9721720218658447,
      "objective/train/docs_used": 461200,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8991169929504395,
      "objective/train/std_weight": 0.14121496677398682,
      "objective/train/theoretical_loss": 3.7278602147665776,
      "objective/train/tokens_used": 823276000,
      "theoretical_loss": 3.7278602147665776,
      "tokens_seen": 802816000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005044283147342539,
      "loss": 2.6292,
      "theoretical_loss": 3.7278602147665776,
      "tokens_seen": 802816000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005044117647058823,
      "loss": 2.9095,
      "theoretical_loss": 3.7278300686432697,
      "tokens_seen": 802881536
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005043952146775108,
      "loss": 2.8088,
      "theoretical_loss": 3.7277999256695025,
      "tokens_seen": 802947072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005043786646491394,
      "loss": 2.6142,
      "theoretical_loss": 3.7277697858446897,
      "tokens_seen": 803012608
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005043621146207679,
      "loss": 2.8434,
      "theoretical_loss": 3.727739649168245,
      "tokens_seen": 803078144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005043455645923964,
      "loss": 2.6737,
      "theoretical_loss": 3.7277095156395843,
      "tokens_seen": 803143680
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000504329014564025,
      "loss": 2.7493,
      "theoretical_loss": 3.7276793852581203,
      "tokens_seen": 803209216
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005043124645356535,
      "loss": 2.6787,
      "theoretical_loss": 3.727649258023268,
      "tokens_seen": 803274752
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005042959145072821,
      "loss": 2.9092,
      "theoretical_loss": 3.7276191339344424,
      "tokens_seen": 803340288
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005042793644789105,
      "loss": 2.8039,
      "theoretical_loss": 3.727589012991058,
      "tokens_seen": 803405824
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000504262814450539,
      "loss": 2.8743,
      "theoretical_loss": 3.72755889519253,
      "tokens_seen": 803471360
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005042462644221676,
      "loss": 2.7979,
      "theoretical_loss": 3.7275287805382744,
      "tokens_seen": 803536896
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005042297143937961,
      "loss": 2.9327,
      "theoretical_loss": 3.7274986690277045,
      "tokens_seen": 803602432
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005042131643654246,
      "loss": 2.6102,
      "theoretical_loss": 3.727468560660238,
      "tokens_seen": 803667968
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005041966143370532,
      "loss": 2.8552,
      "theoretical_loss": 3.727438455435289,
      "tokens_seen": 803733504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005041800643086817,
      "loss": 2.7718,
      "theoretical_loss": 3.727408353352274,
      "tokens_seen": 803799040
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005041635142803102,
      "loss": 2.9735,
      "theoretical_loss": 3.727378254410609,
      "tokens_seen": 803864576
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005041469642519387,
      "loss": 2.68,
      "theoretical_loss": 3.72734815860971,
      "tokens_seen": 803930112
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005041304142235672,
      "loss": 2.6071,
      "theoretical_loss": 3.7273180659489933,
      "tokens_seen": 803995648
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005041138641951958,
      "loss": 2.9,
      "theoretical_loss": 3.727287976427875,
      "tokens_seen": 804061184
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005040973141668243,
      "loss": 2.8798,
      "theoretical_loss": 3.727257890045772,
      "tokens_seen": 804126720
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005040807641384528,
      "loss": 2.783,
      "theoretical_loss": 3.7272278068021008,
      "tokens_seen": 804192256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005040642141100814,
      "loss": 2.8952,
      "theoretical_loss": 3.7271977266962786,
      "tokens_seen": 804257792
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005040476640817099,
      "loss": 2.884,
      "theoretical_loss": 3.727167649727723,
      "tokens_seen": 804323328
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005040311140533384,
      "loss": 2.817,
      "theoretical_loss": 3.7271375758958496,
      "tokens_seen": 804388864
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005040145640249669,
      "loss": 2.6926,
      "theoretical_loss": 3.7271075052000775,
      "tokens_seen": 804454400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005039980139965954,
      "loss": 2.8324,
      "theoretical_loss": 3.7270774376398226,
      "tokens_seen": 804519936
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000503981463968224,
      "loss": 2.6611,
      "theoretical_loss": 3.727047373214504,
      "tokens_seen": 804585472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005039649139398525,
      "loss": 2.5994,
      "theoretical_loss": 3.727017311923539,
      "tokens_seen": 804651008
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000503948363911481,
      "loss": 2.7701,
      "theoretical_loss": 3.726987253766345,
      "tokens_seen": 804716544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005039318138831096,
      "loss": 2.7944,
      "theoretical_loss": 3.7269571987423413,
      "tokens_seen": 804782080
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000503915263854738,
      "loss": 2.6196,
      "theoretical_loss": 3.726927146850945,
      "tokens_seen": 804847616
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005038987138263666,
      "loss": 2.6843,
      "theoretical_loss": 3.726897098091576,
      "tokens_seen": 804913152
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005038821637979951,
      "loss": 2.8225,
      "theoretical_loss": 3.7268670524636507,
      "tokens_seen": 804978688
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005038656137696236,
      "loss": 2.6743,
      "theoretical_loss": 3.7268370099665904,
      "tokens_seen": 805044224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005038490637412521,
      "loss": 2.7943,
      "theoretical_loss": 3.7268069705998124,
      "tokens_seen": 805109760
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005038325137128806,
      "loss": 2.8908,
      "theoretical_loss": 3.7267769343627357,
      "tokens_seen": 805175296
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005038159636845092,
      "loss": 2.6376,
      "theoretical_loss": 3.726746901254781,
      "tokens_seen": 805240832
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005037994136561376,
      "loss": 2.5821,
      "theoretical_loss": 3.726716871275366,
      "tokens_seen": 805306368
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005037828636277662,
      "loss": 2.5677,
      "theoretical_loss": 3.726686844423911,
      "tokens_seen": 805371904
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005037663135993947,
      "loss": 3.047,
      "theoretical_loss": 3.726656820699836,
      "tokens_seen": 805437440
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005037497635710232,
      "loss": 2.8419,
      "theoretical_loss": 3.72662680010256,
      "tokens_seen": 805502976
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005037332135426518,
      "loss": 2.8466,
      "theoretical_loss": 3.726596782631504,
      "tokens_seen": 805568512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005037166635142803,
      "loss": 2.7336,
      "theoretical_loss": 3.7265667682860872,
      "tokens_seen": 805634048
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005037001134859088,
      "loss": 2.9473,
      "theoretical_loss": 3.7265367570657304,
      "tokens_seen": 805699584
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005036835634575374,
      "loss": 2.901,
      "theoretical_loss": 3.7265067489698547,
      "tokens_seen": 805765120
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005036670134291658,
      "loss": 2.793,
      "theoretical_loss": 3.726476743997879,
      "tokens_seen": 805830656
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005036504634007944,
      "loss": 3.0148,
      "theoretical_loss": 3.7264467421492258,
      "tokens_seen": 805896192
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005036339133724229,
      "loss": 2.6354,
      "theoretical_loss": 3.7264167434233153,
      "tokens_seen": 805961728
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005036173633440514,
      "loss": 2.8105,
      "theoretical_loss": 3.7263867478195687,
      "tokens_seen": 806027264
    },
    {
      "epoch": 0.29,
      "objective/train/avg_token_score": 0.011998748406767845,
      "objective/train/avg_weight": 0.985874354839325,
      "objective/train/docs_used": 463166,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0515246391296387,
      "objective/train/std_weight": 0.08968957513570786,
      "objective/train/theoretical_loss": 3.726356755337407,
      "objective/train/tokens_used": 826552800,
      "theoretical_loss": 3.726356755337407,
      "tokens_seen": 806092800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00050360081331568,
      "loss": 2.8215,
      "theoretical_loss": 3.726356755337407,
      "tokens_seen": 806092800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005035842632873085,
      "loss": 2.8418,
      "theoretical_loss": 3.7263267659762516,
      "tokens_seen": 806158336
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000503567713258937,
      "loss": 2.9005,
      "theoretical_loss": 3.7262967797355246,
      "tokens_seen": 806223872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005035511632305655,
      "loss": 2.8042,
      "theoretical_loss": 3.7262667966146474,
      "tokens_seen": 806289408
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000503534613202194,
      "loss": 2.7488,
      "theoretical_loss": 3.7262368166130413,
      "tokens_seen": 806354944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005035180631738226,
      "loss": 2.8255,
      "theoretical_loss": 3.7262068397301293,
      "tokens_seen": 806420480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005035015131454511,
      "loss": 2.7554,
      "theoretical_loss": 3.726176865965333,
      "tokens_seen": 806486016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005034849631170796,
      "loss": 2.6496,
      "theoretical_loss": 3.7261468953180747,
      "tokens_seen": 806551552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005034684130887082,
      "loss": 2.5453,
      "theoretical_loss": 3.7261169277877766,
      "tokens_seen": 806617088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005034518630603367,
      "loss": 2.8086,
      "theoretical_loss": 3.7260869633738625,
      "tokens_seen": 806682624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005034353130319651,
      "loss": 2.9443,
      "theoretical_loss": 3.7260570020757537,
      "tokens_seen": 806748160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005034187630035937,
      "loss": 2.7853,
      "theoretical_loss": 3.7260270438928744,
      "tokens_seen": 806813696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005034022129752222,
      "loss": 2.8057,
      "theoretical_loss": 3.7259970888246468,
      "tokens_seen": 806879232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005033856629468508,
      "loss": 2.6871,
      "theoretical_loss": 3.7259671368704943,
      "tokens_seen": 806944768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005033691129184793,
      "loss": 2.7695,
      "theoretical_loss": 3.7259371880298406,
      "tokens_seen": 807010304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005033525628901078,
      "loss": 2.7995,
      "theoretical_loss": 3.7259072423021085,
      "tokens_seen": 807075840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005033360128617364,
      "loss": 2.6757,
      "theoretical_loss": 3.725877299686723,
      "tokens_seen": 807141376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005033194628333649,
      "loss": 2.6565,
      "theoretical_loss": 3.7258473601831072,
      "tokens_seen": 807206912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005033029128049933,
      "loss": 2.7488,
      "theoretical_loss": 3.725817423790685,
      "tokens_seen": 807272448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005032863627766219,
      "loss": 2.5614,
      "theoretical_loss": 3.72578749050888,
      "tokens_seen": 807337984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005032698127482504,
      "loss": 2.7434,
      "theoretical_loss": 3.7257575603371187,
      "tokens_seen": 807403520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000503253262719879,
      "loss": 2.8471,
      "theoretical_loss": 3.7257276332748233,
      "tokens_seen": 807469056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005032367126915075,
      "loss": 2.6423,
      "theoretical_loss": 3.725697709321419,
      "tokens_seen": 807534592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000503220162663136,
      "loss": 2.5137,
      "theoretical_loss": 3.725667788476331,
      "tokens_seen": 807600128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005032036126347646,
      "loss": 2.7,
      "theoretical_loss": 3.7256378707389843,
      "tokens_seen": 807665664
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000503187062606393,
      "loss": 2.7587,
      "theoretical_loss": 3.725607956108804,
      "tokens_seen": 807731200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005031705125780215,
      "loss": 3.0028,
      "theoretical_loss": 3.7255780445852142,
      "tokens_seen": 807796736
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005031539625496501,
      "loss": 3.1155,
      "theoretical_loss": 3.7255481361676415,
      "tokens_seen": 807862272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005031374125212786,
      "loss": 2.6625,
      "theoretical_loss": 3.725518230855511,
      "tokens_seen": 807927808
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005031208624929072,
      "loss": 3.0203,
      "theoretical_loss": 3.7254883286482485,
      "tokens_seen": 807993344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005031043124645357,
      "loss": 2.8087,
      "theoretical_loss": 3.72545842954528,
      "tokens_seen": 808058880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005030877624361642,
      "loss": 2.6544,
      "theoretical_loss": 3.7254285335460313,
      "tokens_seen": 808124416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005030712124077928,
      "loss": 3.0387,
      "theoretical_loss": 3.725398640649929,
      "tokens_seen": 808189952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005030546623794212,
      "loss": 2.8295,
      "theoretical_loss": 3.725368750856398,
      "tokens_seen": 808255488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005030381123510497,
      "loss": 2.7819,
      "theoretical_loss": 3.7253388641648666,
      "tokens_seen": 808321024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005030215623226783,
      "loss": 2.7191,
      "theoretical_loss": 3.72530898057476,
      "tokens_seen": 808386560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005030050122943068,
      "loss": 2.5783,
      "theoretical_loss": 3.7252791000855057,
      "tokens_seen": 808452096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005029884622659354,
      "loss": 3.0029,
      "theoretical_loss": 3.7252492226965304,
      "tokens_seen": 808517632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005029719122375639,
      "loss": 2.7945,
      "theoretical_loss": 3.7252193484072613,
      "tokens_seen": 808583168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005029553622091924,
      "loss": 2.8277,
      "theoretical_loss": 3.7251894772171257,
      "tokens_seen": 808648704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005029388121808209,
      "loss": 2.6975,
      "theoretical_loss": 3.72515960912555,
      "tokens_seen": 808714240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005029222621524494,
      "loss": 3.0393,
      "theoretical_loss": 3.725129744131963,
      "tokens_seen": 808779776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005029057121240779,
      "loss": 3.0276,
      "theoretical_loss": 3.725099882235792,
      "tokens_seen": 808845312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005028891620957065,
      "loss": 2.6702,
      "theoretical_loss": 3.725070023436465,
      "tokens_seen": 808910848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005028726120673349,
      "loss": 2.5697,
      "theoretical_loss": 3.7250401677334093,
      "tokens_seen": 808976384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005028560620389635,
      "loss": 2.7202,
      "theoretical_loss": 3.7250103151260534,
      "tokens_seen": 809041920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000502839512010592,
      "loss": 2.833,
      "theoretical_loss": 3.7249804656138257,
      "tokens_seen": 809107456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005028229619822205,
      "loss": 2.8876,
      "theoretical_loss": 3.7249506191961546,
      "tokens_seen": 809172992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000502806411953849,
      "loss": 2.7054,
      "theoretical_loss": 3.7249207758724685,
      "tokens_seen": 809238528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005027898619254775,
      "loss": 2.8415,
      "theoretical_loss": 3.7248909356421964,
      "tokens_seen": 809304064
    },
    {
      "epoch": 0.29,
      "objective/train/avg_token_score": 0.02637120708823204,
      "objective/train/avg_weight": 0.9689534306526184,
      "objective/train/docs_used": 465044,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9010791778564453,
      "objective/train/std_weight": 0.13047672808170319,
      "objective/train/theoretical_loss": 3.724861098504767,
      "objective/train/tokens_used": 829829600,
      "theoretical_loss": 3.724861098504767,
      "tokens_seen": 809369600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005027733118971061,
      "loss": 2.7774,
      "theoretical_loss": 3.724861098504767,
      "tokens_seen": 809369600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005027567618687346,
      "loss": 2.6019,
      "theoretical_loss": 3.7248312644596098,
      "tokens_seen": 809435136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005027402118403631,
      "loss": 2.9349,
      "theoretical_loss": 3.724801433506153,
      "tokens_seen": 809500672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005027236618119917,
      "loss": 2.7292,
      "theoretical_loss": 3.724771605643827,
      "tokens_seen": 809566208
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005027071117836201,
      "loss": 2.5502,
      "theoretical_loss": 3.7247417808720606,
      "tokens_seen": 809631744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005026905617552487,
      "loss": 2.7641,
      "theoretical_loss": 3.7247119591902837,
      "tokens_seen": 809697280
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005026740117268772,
      "loss": 2.8114,
      "theoretical_loss": 3.7246821405979267,
      "tokens_seen": 809762816
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005026574616985057,
      "loss": 2.7314,
      "theoretical_loss": 3.724652325094418,
      "tokens_seen": 809828352
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005026409116701343,
      "loss": 2.7662,
      "theoretical_loss": 3.724622512679189,
      "tokens_seen": 809893888
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005026243616417628,
      "loss": 2.6885,
      "theoretical_loss": 3.7245927033516697,
      "tokens_seen": 809959424
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005026078116133913,
      "loss": 2.7992,
      "theoretical_loss": 3.7245628971112907,
      "tokens_seen": 810024960
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005025912615850199,
      "loss": 2.8752,
      "theoretical_loss": 3.724533093957482,
      "tokens_seen": 810090496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005025747115566483,
      "loss": 2.8764,
      "theoretical_loss": 3.7245032938896743,
      "tokens_seen": 810156032
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005025581615282769,
      "loss": 2.7625,
      "theoretical_loss": 3.724473496907299,
      "tokens_seen": 810221568
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005025416114999054,
      "loss": 2.628,
      "theoretical_loss": 3.724443703009787,
      "tokens_seen": 810287104
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005025250614715339,
      "loss": 2.9049,
      "theoretical_loss": 3.724413912196569,
      "tokens_seen": 810352640
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005025085114431625,
      "loss": 2.6888,
      "theoretical_loss": 3.7243841244670772,
      "tokens_seen": 810418176
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000502491961414791,
      "loss": 2.744,
      "theoretical_loss": 3.724354339820742,
      "tokens_seen": 810483712
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005024754113864196,
      "loss": 2.6676,
      "theoretical_loss": 3.7243245582569955,
      "tokens_seen": 810549248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000502458861358048,
      "loss": 2.8227,
      "theoretical_loss": 3.7242947797752697,
      "tokens_seen": 810614784
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005024423113296765,
      "loss": 2.8124,
      "theoretical_loss": 3.7242650043749963,
      "tokens_seen": 810680320
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005024257613013051,
      "loss": 2.9276,
      "theoretical_loss": 3.724235232055607,
      "tokens_seen": 810745856
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005024092112729336,
      "loss": 2.6638,
      "theoretical_loss": 3.7242054628165344,
      "tokens_seen": 810811392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005023926612445621,
      "loss": 2.6134,
      "theoretical_loss": 3.724175696657211,
      "tokens_seen": 810876928
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005023761112161907,
      "loss": 2.7433,
      "theoretical_loss": 3.724145933577069,
      "tokens_seen": 810942464
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005023595611878192,
      "loss": 2.8042,
      "theoretical_loss": 3.724116173575542,
      "tokens_seen": 811008000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005023430111594478,
      "loss": 2.7605,
      "theoretical_loss": 3.724086416652061,
      "tokens_seen": 811073536
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005023264611310762,
      "loss": 2.592,
      "theoretical_loss": 3.7240566628060607,
      "tokens_seen": 811139072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005023099111027047,
      "loss": 2.7906,
      "theoretical_loss": 3.7240269120369733,
      "tokens_seen": 811204608
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005022933610743333,
      "loss": 2.7826,
      "theoretical_loss": 3.7239971643442322,
      "tokens_seen": 811270144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005022768110459618,
      "loss": 2.6091,
      "theoretical_loss": 3.7239674197272707,
      "tokens_seen": 811335680
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005022602610175903,
      "loss": 2.7421,
      "theoretical_loss": 3.7239376781855222,
      "tokens_seen": 811401216
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005022437109892189,
      "loss": 2.7268,
      "theoretical_loss": 3.723907939718422,
      "tokens_seen": 811466752
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005022271609608474,
      "loss": 2.5993,
      "theoretical_loss": 3.723878204325402,
      "tokens_seen": 811532288
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005022106109324759,
      "loss": 2.735,
      "theoretical_loss": 3.7238484720058973,
      "tokens_seen": 811597824
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005021940609041044,
      "loss": 2.6859,
      "theoretical_loss": 3.7238187427593417,
      "tokens_seen": 811663360
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005021775108757329,
      "loss": 2.8203,
      "theoretical_loss": 3.7237890165851693,
      "tokens_seen": 811728896
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005021609608473615,
      "loss": 2.5966,
      "theoretical_loss": 3.723759293482815,
      "tokens_seen": 811794432
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00050214441081899,
      "loss": 2.5757,
      "theoretical_loss": 3.7237295734517133,
      "tokens_seen": 811859968
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005021278607906185,
      "loss": 2.6781,
      "theoretical_loss": 3.7236998564912986,
      "tokens_seen": 811925504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005021113107622471,
      "loss": 2.6244,
      "theoretical_loss": 3.723670142601007,
      "tokens_seen": 811991040
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005020947607338755,
      "loss": 2.7865,
      "theoretical_loss": 3.7236404317802716,
      "tokens_seen": 812056576
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005020782107055041,
      "loss": 2.7821,
      "theoretical_loss": 3.7236107240285294,
      "tokens_seen": 812122112
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005020616606771326,
      "loss": 2.9512,
      "theoretical_loss": 3.723581019345215,
      "tokens_seen": 812187648
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005020451106487611,
      "loss": 2.823,
      "theoretical_loss": 3.723551317729764,
      "tokens_seen": 812253184
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005020285606203897,
      "loss": 2.7453,
      "theoretical_loss": 3.723521619181612,
      "tokens_seen": 812318720
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005020120105920182,
      "loss": 2.8642,
      "theoretical_loss": 3.7234919237001947,
      "tokens_seen": 812384256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005019954605636467,
      "loss": 2.7933,
      "theoretical_loss": 3.723462231284949,
      "tokens_seen": 812449792
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005019789105352753,
      "loss": 2.6372,
      "theoretical_loss": 3.72343254193531,
      "tokens_seen": 812515328
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005019623605069037,
      "loss": 2.7088,
      "theoretical_loss": 3.7234028556507135,
      "tokens_seen": 812580864
    },
    {
      "epoch": 0.29,
      "objective/train/avg_token_score": 0.014298880472779274,
      "objective/train/avg_weight": 0.983168363571167,
      "objective/train/docs_used": 466971,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0626397132873535,
      "objective/train/std_weight": 0.08236566931009293,
      "objective/train/theoretical_loss": 3.7233731724305974,
      "objective/train/tokens_used": 833106400,
      "theoretical_loss": 3.7233731724305974,
      "tokens_seen": 812646400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005019458104785323,
      "loss": 3.0138,
      "theoretical_loss": 3.7233731724305974,
      "tokens_seen": 812646400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005019292604501608,
      "loss": 2.8115,
      "theoretical_loss": 3.723343492274397,
      "tokens_seen": 812711936
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005019127104217893,
      "loss": 2.6592,
      "theoretical_loss": 3.72331381518155,
      "tokens_seen": 812777472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005018961603934179,
      "loss": 2.706,
      "theoretical_loss": 3.7232841411514923,
      "tokens_seen": 812843008
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005018796103650463,
      "loss": 2.7887,
      "theoretical_loss": 3.7232544701836616,
      "tokens_seen": 812908544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005018630603366749,
      "loss": 2.7391,
      "theoretical_loss": 3.7232248022774943,
      "tokens_seen": 812974080
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005018465103083033,
      "loss": 2.737,
      "theoretical_loss": 3.723195137432429,
      "tokens_seen": 813039616
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005018299602799318,
      "loss": 2.7323,
      "theoretical_loss": 3.7231654756479022,
      "tokens_seen": 813105152
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005018134102515604,
      "loss": 2.8955,
      "theoretical_loss": 3.723135816923352,
      "tokens_seen": 813170688
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005017968602231889,
      "loss": 2.8285,
      "theoretical_loss": 3.7231061612582153,
      "tokens_seen": 813236224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005017803101948175,
      "loss": 2.6755,
      "theoretical_loss": 3.7230765086519306,
      "tokens_seen": 813301760
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000501763760166446,
      "loss": 2.6091,
      "theoretical_loss": 3.723046859103936,
      "tokens_seen": 813367296
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005017472101380745,
      "loss": 2.8359,
      "theoretical_loss": 3.7230172126136694,
      "tokens_seen": 813432832
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000501730660109703,
      "loss": 2.8239,
      "theoretical_loss": 3.7229875691805696,
      "tokens_seen": 813498368
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005017141100813315,
      "loss": 2.9605,
      "theoretical_loss": 3.7229579288040746,
      "tokens_seen": 813563904
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00050169756005296,
      "loss": 2.7295,
      "theoretical_loss": 3.7229282914836235,
      "tokens_seen": 813629440
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005016810100245886,
      "loss": 2.9581,
      "theoretical_loss": 3.7228986572186553,
      "tokens_seen": 813694976
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005016644599962171,
      "loss": 2.507,
      "theoretical_loss": 3.722869026008608,
      "tokens_seen": 813760512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005016479099678457,
      "loss": 2.6935,
      "theoretical_loss": 3.722839397852921,
      "tokens_seen": 813826048
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005016313599394742,
      "loss": 2.7892,
      "theoretical_loss": 3.7228097727510336,
      "tokens_seen": 813891584
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005016148099111026,
      "loss": 2.8935,
      "theoretical_loss": 3.722780150702386,
      "tokens_seen": 813957120
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005015982598827312,
      "loss": 2.819,
      "theoretical_loss": 3.7227505317064162,
      "tokens_seen": 814022656
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005015817098543597,
      "loss": 2.9307,
      "theoretical_loss": 3.722720915762565,
      "tokens_seen": 814088192
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005015651598259882,
      "loss": 2.7287,
      "theoretical_loss": 3.722691302870272,
      "tokens_seen": 814153728
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005015486097976168,
      "loss": 3.0164,
      "theoretical_loss": 3.7226616930289773,
      "tokens_seen": 814219264
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005015320597692453,
      "loss": 2.9789,
      "theoretical_loss": 3.7226320862381206,
      "tokens_seen": 814284800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005015155097408739,
      "loss": 2.8205,
      "theoretical_loss": 3.722602482497143,
      "tokens_seen": 814350336
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005014989597125024,
      "loss": 2.7289,
      "theoretical_loss": 3.7225728818054833,
      "tokens_seen": 814415872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005014824096841308,
      "loss": 2.5602,
      "theoretical_loss": 3.7225432841625836,
      "tokens_seen": 814481408
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005014658596557594,
      "loss": 2.6755,
      "theoretical_loss": 3.722513689567884,
      "tokens_seen": 814546944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005014493096273879,
      "loss": 2.8566,
      "theoretical_loss": 3.722484098020826,
      "tokens_seen": 814612480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005014327595990164,
      "loss": 2.8347,
      "theoretical_loss": 3.7224545095208494,
      "tokens_seen": 814678016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000501416209570645,
      "loss": 2.8007,
      "theoretical_loss": 3.722424924067396,
      "tokens_seen": 814743552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005013996595422735,
      "loss": 2.6684,
      "theoretical_loss": 3.722395341659907,
      "tokens_seen": 814809088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005013831095139021,
      "loss": 2.8462,
      "theoretical_loss": 3.722365762297825,
      "tokens_seen": 814874624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005013665594855305,
      "loss": 2.666,
      "theoretical_loss": 3.72233618598059,
      "tokens_seen": 814940160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000501350009457159,
      "loss": 2.5982,
      "theoretical_loss": 3.7223066127076443,
      "tokens_seen": 815005696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005013334594287876,
      "loss": 2.7844,
      "theoretical_loss": 3.7222770424784297,
      "tokens_seen": 815071232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005013169094004161,
      "loss": 3.0278,
      "theoretical_loss": 3.722247475292388,
      "tokens_seen": 815136768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005013003593720446,
      "loss": 2.7746,
      "theoretical_loss": 3.7222179111489626,
      "tokens_seen": 815202304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005012838093436732,
      "loss": 2.7389,
      "theoretical_loss": 3.7221883500475945,
      "tokens_seen": 815267840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005012672593153017,
      "loss": 2.7717,
      "theoretical_loss": 3.7221587919877264,
      "tokens_seen": 815333376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005012507092869303,
      "loss": 2.9006,
      "theoretical_loss": 3.7221292369688017,
      "tokens_seen": 815398912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005012341592585587,
      "loss": 2.4543,
      "theoretical_loss": 3.722099684990263,
      "tokens_seen": 815464448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005012176092301872,
      "loss": 2.844,
      "theoretical_loss": 3.722070136051552,
      "tokens_seen": 815529984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005012010592018158,
      "loss": 2.6989,
      "theoretical_loss": 3.7220405901521127,
      "tokens_seen": 815595520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005011845091734443,
      "loss": 2.7222,
      "theoretical_loss": 3.7220110472913888,
      "tokens_seen": 815661056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005011679591450728,
      "loss": 3.0739,
      "theoretical_loss": 3.721981507468823,
      "tokens_seen": 815726592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005011514091167014,
      "loss": 2.8227,
      "theoretical_loss": 3.7219519706838584,
      "tokens_seen": 815792128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005011348590883299,
      "loss": 2.7346,
      "theoretical_loss": 3.72192243693594,
      "tokens_seen": 815857664
    },
    {
      "epoch": 0.29,
      "objective/train/avg_token_score": 0.024799689650535583,
      "objective/train/avg_weight": 0.9708377718925476,
      "objective/train/docs_used": 469033,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.301227569580078,
      "objective/train/std_weight": 0.13943512737751007,
      "objective/train/theoretical_loss": 3.7218929062245105,
      "objective/train/tokens_used": 836383200,
      "theoretical_loss": 3.7218929062245105,
      "tokens_seen": 815923200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005011183090599584,
      "loss": 2.73,
      "theoretical_loss": 3.7218929062245105,
      "tokens_seen": 815923200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005011017590315869,
      "loss": 2.7857,
      "theoretical_loss": 3.7218633785490143,
      "tokens_seen": 815988736
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005010852090032154,
      "loss": 2.97,
      "theoretical_loss": 3.721833853908895,
      "tokens_seen": 816054272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000501068658974844,
      "loss": 2.5755,
      "theoretical_loss": 3.721804332303597,
      "tokens_seen": 816119808
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005010521089464725,
      "loss": 2.8708,
      "theoretical_loss": 3.7217748137325652,
      "tokens_seen": 816185344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000501035558918101,
      "loss": 2.5758,
      "theoretical_loss": 3.721745298195244,
      "tokens_seen": 816250880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005010190088897296,
      "loss": 2.6131,
      "theoretical_loss": 3.7217157856910776,
      "tokens_seen": 816316416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000501002458861358,
      "loss": 2.4382,
      "theoretical_loss": 3.7216862762195113,
      "tokens_seen": 816381952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005009859088329866,
      "loss": 2.7817,
      "theoretical_loss": 3.72165676977999,
      "tokens_seen": 816447488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005009693588046151,
      "loss": 2.9101,
      "theoretical_loss": 3.721627266371958,
      "tokens_seen": 816513024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005009528087762436,
      "loss": 2.4695,
      "theoretical_loss": 3.721597765994862,
      "tokens_seen": 816578560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005009362587478722,
      "loss": 2.6175,
      "theoretical_loss": 3.721568268648146,
      "tokens_seen": 816644096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005009197087195007,
      "loss": 2.5884,
      "theoretical_loss": 3.7215387743312567,
      "tokens_seen": 816709632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005009031586911292,
      "loss": 2.8764,
      "theoretical_loss": 3.7215092830436394,
      "tokens_seen": 816775168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005008866086627576,
      "loss": 2.3906,
      "theoretical_loss": 3.72147979478474,
      "tokens_seen": 816840704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005008700586343862,
      "loss": 2.7951,
      "theoretical_loss": 3.7214503095540037,
      "tokens_seen": 816906240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005008535086060147,
      "loss": 2.9747,
      "theoretical_loss": 3.721420827350878,
      "tokens_seen": 816971776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005008369585776432,
      "loss": 2.6423,
      "theoretical_loss": 3.7213913481748087,
      "tokens_seen": 817037312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005008204085492718,
      "loss": 2.7634,
      "theoretical_loss": 3.7213618720252413,
      "tokens_seen": 817102848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005008038585209003,
      "loss": 2.8661,
      "theoretical_loss": 3.7213323989016236,
      "tokens_seen": 817168384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005007873084925288,
      "loss": 2.7778,
      "theoretical_loss": 3.721302928803402,
      "tokens_seen": 817233920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005007707584641574,
      "loss": 2.3922,
      "theoretical_loss": 3.7212734617300223,
      "tokens_seen": 817299456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005007542084357858,
      "loss": 2.8295,
      "theoretical_loss": 3.7212439976809337,
      "tokens_seen": 817364992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005007376584074144,
      "loss": 2.6121,
      "theoretical_loss": 3.721214536655581,
      "tokens_seen": 817430528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005007211083790429,
      "loss": 2.657,
      "theoretical_loss": 3.721185078653413,
      "tokens_seen": 817496064
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005007045583506714,
      "loss": 2.6737,
      "theoretical_loss": 3.7211556236738765,
      "tokens_seen": 817561600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005006880083223,
      "loss": 2.7339,
      "theoretical_loss": 3.7211261717164197,
      "tokens_seen": 817627136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005006714582939285,
      "loss": 2.7372,
      "theoretical_loss": 3.72109672278049,
      "tokens_seen": 817692672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000500654908265557,
      "loss": 2.6699,
      "theoretical_loss": 3.7210672768655355,
      "tokens_seen": 817758208
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005006383582371855,
      "loss": 2.9091,
      "theoretical_loss": 3.7210378339710033,
      "tokens_seen": 817823744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000500621808208814,
      "loss": 2.8258,
      "theoretical_loss": 3.7210083940963425,
      "tokens_seen": 817889280
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005006052581804426,
      "loss": 2.7848,
      "theoretical_loss": 3.7209789572410013,
      "tokens_seen": 817954816
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005005887081520711,
      "loss": 2.6208,
      "theoretical_loss": 3.7209495234044283,
      "tokens_seen": 818020352
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005005721581236996,
      "loss": 2.9302,
      "theoretical_loss": 3.7209200925860717,
      "tokens_seen": 818085888
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005005556080953282,
      "loss": 2.591,
      "theoretical_loss": 3.7208906647853803,
      "tokens_seen": 818151424
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005005390580669567,
      "loss": 2.6968,
      "theoretical_loss": 3.720861240001803,
      "tokens_seen": 818216960
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005005225080385851,
      "loss": 2.7407,
      "theoretical_loss": 3.720831818234789,
      "tokens_seen": 818282496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005005059580102137,
      "loss": 2.7247,
      "theoretical_loss": 3.7208023994837878,
      "tokens_seen": 818348032
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005004894079818422,
      "loss": 2.5134,
      "theoretical_loss": 3.7207729837482484,
      "tokens_seen": 818413568
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005004728579534708,
      "loss": 2.839,
      "theoretical_loss": 3.7207435710276195,
      "tokens_seen": 818479104
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005004563079250993,
      "loss": 2.8638,
      "theoretical_loss": 3.720714161321352,
      "tokens_seen": 818544640
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005004397578967278,
      "loss": 2.8007,
      "theoretical_loss": 3.7206847546288957,
      "tokens_seen": 818610176
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005004232078683564,
      "loss": 2.7147,
      "theoretical_loss": 3.7206553509496993,
      "tokens_seen": 818675712
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005004066578399849,
      "loss": 3.0624,
      "theoretical_loss": 3.720625950283214,
      "tokens_seen": 818741248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005003901078116133,
      "loss": 2.7607,
      "theoretical_loss": 3.720596552628889,
      "tokens_seen": 818806784
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005003735577832419,
      "loss": 2.8093,
      "theoretical_loss": 3.720567157986175,
      "tokens_seen": 818872320
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005003570077548704,
      "loss": 2.734,
      "theoretical_loss": 3.7205377663545236,
      "tokens_seen": 818937856
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000500340457726499,
      "loss": 2.7463,
      "theoretical_loss": 3.720508377733384,
      "tokens_seen": 819003392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005003239076981275,
      "loss": 2.7308,
      "theoretical_loss": 3.7204789921222075,
      "tokens_seen": 819068928
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000500307357669756,
      "loss": 2.8116,
      "theoretical_loss": 3.720449609520445,
      "tokens_seen": 819134464
    },
    {
      "epoch": 0.29,
      "objective/train/avg_token_score": 0.009068925864994526,
      "objective/train/avg_weight": 0.9893243908882141,
      "objective/train/docs_used": 471128,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7218191623687744,
      "objective/train/std_weight": 0.06846398115158081,
      "objective/train/theoretical_loss": 3.7204202299275475,
      "objective/train/tokens_used": 839660000,
      "theoretical_loss": 3.7204202299275475,
      "tokens_seen": 819200000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005002908076413846,
      "loss": 2.6078,
      "theoretical_loss": 3.7204202299275475,
      "tokens_seen": 819200000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000500274257613013,
      "loss": 2.76,
      "theoretical_loss": 3.7203908533429666,
      "tokens_seen": 819265536
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005002577075846415,
      "loss": 2.7137,
      "theoretical_loss": 3.7203614797661535,
      "tokens_seen": 819331072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005002411575562701,
      "loss": 2.6485,
      "theoretical_loss": 3.720332109196559,
      "tokens_seen": 819396608
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005002246075278986,
      "loss": 2.6888,
      "theoretical_loss": 3.7203027416336356,
      "tokens_seen": 819462144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005002080574995272,
      "loss": 2.7157,
      "theoretical_loss": 3.7202733770768344,
      "tokens_seen": 819527680
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005001915074711557,
      "loss": 2.7132,
      "theoretical_loss": 3.7202440155256076,
      "tokens_seen": 819593216
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005001749574427842,
      "loss": 2.6824,
      "theoretical_loss": 3.720214656979408,
      "tokens_seen": 819658752
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005001584074144128,
      "loss": 2.6737,
      "theoretical_loss": 3.720185301437687,
      "tokens_seen": 819724288
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005001418573860412,
      "loss": 2.6982,
      "theoretical_loss": 3.720155948899897,
      "tokens_seen": 819789824
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005001253073576697,
      "loss": 2.5902,
      "theoretical_loss": 3.720126599365491,
      "tokens_seen": 819855360
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005001087573292983,
      "loss": 3.0654,
      "theoretical_loss": 3.720097252833921,
      "tokens_seen": 819920896
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005000922073009268,
      "loss": 2.542,
      "theoretical_loss": 3.72006790930464,
      "tokens_seen": 819986432
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005000756572725554,
      "loss": 2.9867,
      "theoretical_loss": 3.720038568777101,
      "tokens_seen": 820051968
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005000591072441839,
      "loss": 2.438,
      "theoretical_loss": 3.7200092312507573,
      "tokens_seen": 820117504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005000425572158124,
      "loss": 2.6711,
      "theoretical_loss": 3.719979896725062,
      "tokens_seen": 820183040
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000500026007187441,
      "loss": 2.7644,
      "theoretical_loss": 3.7199505651994684,
      "tokens_seen": 820248576
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0005000094571590694,
      "loss": 2.8421,
      "theoretical_loss": 3.71992123667343,
      "tokens_seen": 820314112
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004999929071306979,
      "loss": 2.891,
      "theoretical_loss": 3.7198919111464006,
      "tokens_seen": 820379648
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004999763571023265,
      "loss": 2.7144,
      "theoretical_loss": 3.7198625886178336,
      "tokens_seen": 820445184
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000499959807073955,
      "loss": 2.7298,
      "theoretical_loss": 3.719833269087183,
      "tokens_seen": 820510720
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004999432570455836,
      "loss": 2.7762,
      "theoretical_loss": 3.719803952553904,
      "tokens_seen": 820576256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000499926707017212,
      "loss": 2.8241,
      "theoretical_loss": 3.719774639017449,
      "tokens_seen": 820641792
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004999101569888405,
      "loss": 2.5851,
      "theoretical_loss": 3.7197453284772735,
      "tokens_seen": 820707328
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000499893606960469,
      "loss": 2.8182,
      "theoretical_loss": 3.719716020932832,
      "tokens_seen": 820772864
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004998770569320975,
      "loss": 2.8654,
      "theoretical_loss": 3.719686716383579,
      "tokens_seen": 820838400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004998605069037261,
      "loss": 2.7124,
      "theoretical_loss": 3.719657414828969,
      "tokens_seen": 820903936
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004998439568753546,
      "loss": 2.9345,
      "theoretical_loss": 3.7196281162684572,
      "tokens_seen": 820969472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004998274068469832,
      "loss": 2.8386,
      "theoretical_loss": 3.7195988207014983,
      "tokens_seen": 821035008
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004998108568186117,
      "loss": 2.7689,
      "theoretical_loss": 3.7195695281275487,
      "tokens_seen": 821100544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004997943067902402,
      "loss": 2.6345,
      "theoretical_loss": 3.7195402385460623,
      "tokens_seen": 821166080
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004997777567618687,
      "loss": 2.8826,
      "theoretical_loss": 3.7195109519564955,
      "tokens_seen": 821231616
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004997612067334972,
      "loss": 2.8188,
      "theoretical_loss": 3.719481668358304,
      "tokens_seen": 821297152
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004997446567051257,
      "loss": 2.6142,
      "theoretical_loss": 3.7194523877509424,
      "tokens_seen": 821362688
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004997281066767543,
      "loss": 2.8785,
      "theoretical_loss": 3.719423110133868,
      "tokens_seen": 821428224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004997115566483828,
      "loss": 2.8418,
      "theoretical_loss": 3.719393835506537,
      "tokens_seen": 821493760
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004996950066200114,
      "loss": 2.8032,
      "theoretical_loss": 3.719364563868404,
      "tokens_seen": 821559296
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004996784565916399,
      "loss": 2.671,
      "theoretical_loss": 3.719335295218927,
      "tokens_seen": 821624832
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004996619065632683,
      "loss": 2.8924,
      "theoretical_loss": 3.719306029557562,
      "tokens_seen": 821690368
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004996453565348969,
      "loss": 2.7488,
      "theoretical_loss": 3.719276766883765,
      "tokens_seen": 821755904
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004996288065065254,
      "loss": 2.7035,
      "theoretical_loss": 3.719247507196994,
      "tokens_seen": 821821440
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004996122564781539,
      "loss": 2.8949,
      "theoretical_loss": 3.719218250496705,
      "tokens_seen": 821886976
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004995957064497825,
      "loss": 2.7746,
      "theoretical_loss": 3.7191889967823557,
      "tokens_seen": 821952512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000499579156421411,
      "loss": 2.8153,
      "theoretical_loss": 3.7191597460534025,
      "tokens_seen": 822018048
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004995626063930396,
      "loss": 2.8242,
      "theoretical_loss": 3.7191304983093034,
      "tokens_seen": 822083584
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000499546056364668,
      "loss": 2.59,
      "theoretical_loss": 3.719101253549516,
      "tokens_seen": 822149120
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004995295063362965,
      "loss": 2.7023,
      "theoretical_loss": 3.719072011773497,
      "tokens_seen": 822214656
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004995129563079251,
      "loss": 2.6663,
      "theoretical_loss": 3.719042772980705,
      "tokens_seen": 822280192
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004994964062795536,
      "loss": 2.777,
      "theoretical_loss": 3.719013537170598,
      "tokens_seen": 822345728
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004994798562511821,
      "loss": 2.8479,
      "theoretical_loss": 3.718984304342634,
      "tokens_seen": 822411264
    },
    {
      "epoch": 0.29,
      "objective/train/avg_token_score": 0.016127336770296097,
      "objective/train/avg_weight": 0.9810256361961365,
      "objective/train/docs_used": 472950,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.844233989715576,
      "objective/train/std_weight": 0.08944837003946304,
      "objective/train/theoretical_loss": 3.7189550744962707,
      "objective/train/tokens_used": 842936800,
      "theoretical_loss": 3.7189550744962707,
      "tokens_seen": 822476800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004994633062228107,
      "loss": 2.796,
      "theoretical_loss": 3.7189550744962707,
      "tokens_seen": 822476800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004994467561944392,
      "loss": 2.7306,
      "theoretical_loss": 3.718925847630967,
      "tokens_seen": 822542336
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004994302061660678,
      "loss": 2.7341,
      "theoretical_loss": 3.7188966237461814,
      "tokens_seen": 822607872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004994136561376962,
      "loss": 2.6721,
      "theoretical_loss": 3.718867402841372,
      "tokens_seen": 822673408
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004993971061093247,
      "loss": 2.8884,
      "theoretical_loss": 3.718838184915998,
      "tokens_seen": 822738944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004993805560809533,
      "loss": 2.6216,
      "theoretical_loss": 3.718808969969518,
      "tokens_seen": 822804480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004993640060525818,
      "loss": 2.7118,
      "theoretical_loss": 3.718779758001392,
      "tokens_seen": 822870016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004993474560242103,
      "loss": 2.5898,
      "theoretical_loss": 3.7187505490110775,
      "tokens_seen": 822935552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004993309059958389,
      "loss": 2.8139,
      "theoretical_loss": 3.7187213429980357,
      "tokens_seen": 823001088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004993143559674674,
      "loss": 2.4398,
      "theoretical_loss": 3.7186921399617248,
      "tokens_seen": 823066624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000499297805939096,
      "loss": 2.7645,
      "theoretical_loss": 3.7186629399016047,
      "tokens_seen": 823132160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004992812559107244,
      "loss": 2.7722,
      "theoretical_loss": 3.7186337428171354,
      "tokens_seen": 823197696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004992647058823529,
      "loss": 2.8278,
      "theoretical_loss": 3.7186045487077766,
      "tokens_seen": 823263232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004992481558539815,
      "loss": 2.6495,
      "theoretical_loss": 3.7185753575729885,
      "tokens_seen": 823328768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00049923160582561,
      "loss": 2.6477,
      "theoretical_loss": 3.7185461694122313,
      "tokens_seen": 823394304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004992150557972385,
      "loss": 2.7561,
      "theoretical_loss": 3.7185169842249652,
      "tokens_seen": 823459840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004991985057688671,
      "loss": 2.8178,
      "theoretical_loss": 3.7184878020106504,
      "tokens_seen": 823525376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004991819557404956,
      "loss": 2.6455,
      "theoretical_loss": 3.718458622768748,
      "tokens_seen": 823590912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004991654057121241,
      "loss": 2.7599,
      "theoretical_loss": 3.718429446498719,
      "tokens_seen": 823656448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004991488556837526,
      "loss": 2.576,
      "theoretical_loss": 3.7184002732000234,
      "tokens_seen": 823721984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004991323056553811,
      "loss": 2.894,
      "theoretical_loss": 3.7183711028721227,
      "tokens_seen": 823787520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004991157556270097,
      "loss": 2.7429,
      "theoretical_loss": 3.7183419355144784,
      "tokens_seen": 823853056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004990992055986382,
      "loss": 2.7362,
      "theoretical_loss": 3.718312771126551,
      "tokens_seen": 823918592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004990826555702667,
      "loss": 2.9343,
      "theoretical_loss": 3.718283609707803,
      "tokens_seen": 823984128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004990661055418953,
      "loss": 2.6825,
      "theoretical_loss": 3.718254451257695,
      "tokens_seen": 824049664
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004990495555135237,
      "loss": 2.7213,
      "theoretical_loss": 3.7182252957756896,
      "tokens_seen": 824115200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004990330054851523,
      "loss": 2.7551,
      "theoretical_loss": 3.718196143261248,
      "tokens_seen": 824180736
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004990164554567808,
      "loss": 2.6357,
      "theoretical_loss": 3.718166993713832,
      "tokens_seen": 824246272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004989999054284093,
      "loss": 2.6722,
      "theoretical_loss": 3.7181378471329047,
      "tokens_seen": 824311808
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004989833554000379,
      "loss": 2.7003,
      "theoretical_loss": 3.7181087035179283,
      "tokens_seen": 824377344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004989668053716664,
      "loss": 2.7672,
      "theoretical_loss": 3.718079562868364,
      "tokens_seen": 824442880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004989502553432949,
      "loss": 2.6913,
      "theoretical_loss": 3.7180504251836757,
      "tokens_seen": 824508416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004989337053149233,
      "loss": 2.5302,
      "theoretical_loss": 3.7180212904633256,
      "tokens_seen": 824573952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004989171552865518,
      "loss": 2.6524,
      "theoretical_loss": 3.7179921587067764,
      "tokens_seen": 824639488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004989006052581804,
      "loss": 2.8137,
      "theoretical_loss": 3.7179630299134914,
      "tokens_seen": 824705024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004988840552298089,
      "loss": 2.5009,
      "theoretical_loss": 3.7179339040829333,
      "tokens_seen": 824770560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004988675052014375,
      "loss": 2.7811,
      "theoretical_loss": 3.7179047812145662,
      "tokens_seen": 824836096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000498850955173066,
      "loss": 2.8478,
      "theoretical_loss": 3.7178756613078527,
      "tokens_seen": 824901632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004988344051446945,
      "loss": 2.6858,
      "theoretical_loss": 3.717846544362257,
      "tokens_seen": 824967168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000498817855116323,
      "loss": 2.784,
      "theoretical_loss": 3.717817430377242,
      "tokens_seen": 825032704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004988013050879515,
      "loss": 2.4973,
      "theoretical_loss": 3.7177883193522723,
      "tokens_seen": 825098240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00049878475505958,
      "loss": 2.8127,
      "theoretical_loss": 3.717759211286811,
      "tokens_seen": 825163776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004987682050312086,
      "loss": 2.4457,
      "theoretical_loss": 3.7177301061803236,
      "tokens_seen": 825229312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004987516550028371,
      "loss": 2.6802,
      "theoretical_loss": 3.717701004032273,
      "tokens_seen": 825294848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004987351049744657,
      "loss": 2.8132,
      "theoretical_loss": 3.7176719048421245,
      "tokens_seen": 825360384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004987185549460942,
      "loss": 2.8245,
      "theoretical_loss": 3.717642808609342,
      "tokens_seen": 825425920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004987020049177227,
      "loss": 2.7102,
      "theoretical_loss": 3.717613715333391,
      "tokens_seen": 825491456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004986854548893512,
      "loss": 2.9295,
      "theoretical_loss": 3.7175846250137345,
      "tokens_seen": 825556992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004986689048609797,
      "loss": 2.5846,
      "theoretical_loss": 3.7175555376498397,
      "tokens_seen": 825622528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004986523548326082,
      "loss": 3.0079,
      "theoretical_loss": 3.7175264532411703,
      "tokens_seen": 825688064
    },
    {
      "epoch": 0.29,
      "objective/train/avg_token_score": 0.014618747867643833,
      "objective/train/avg_weight": 0.9827982187271118,
      "objective/train/docs_used": 475156,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0948524475097656,
      "objective/train/std_weight": 0.08955591917037964,
      "objective/train/theoretical_loss": 3.717497371787192,
      "objective/train/tokens_used": 846213600,
      "theoretical_loss": 3.717497371787192,
      "tokens_seen": 825753600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004986358048042368,
      "loss": 2.8094,
      "theoretical_loss": 3.717497371787192,
      "tokens_seen": 825753600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004986192547758653,
      "loss": 2.6502,
      "theoretical_loss": 3.7174682932873706,
      "tokens_seen": 825819136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004986027047474939,
      "loss": 3.035,
      "theoretical_loss": 3.7174392177411706,
      "tokens_seen": 825884672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0004985861547191224,
      "loss": 2.9444,
      "theoretical_loss": 3.717410145148058,
      "tokens_seen": 825950208
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004985696046907508,
      "loss": 2.6832,
      "theoretical_loss": 3.717381075507499,
      "tokens_seen": 826015744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004985530546623794,
      "loss": 2.7272,
      "theoretical_loss": 3.717352008818959,
      "tokens_seen": 826081280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004985365046340079,
      "loss": 2.9587,
      "theoretical_loss": 3.7173229450819045,
      "tokens_seen": 826146816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004985199546056364,
      "loss": 2.7452,
      "theoretical_loss": 3.7172938842958017,
      "tokens_seen": 826212352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000498503404577265,
      "loss": 2.8571,
      "theoretical_loss": 3.7172648264601165,
      "tokens_seen": 826277888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004984868545488935,
      "loss": 2.6337,
      "theoretical_loss": 3.7172357715743156,
      "tokens_seen": 826343424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004984703045205221,
      "loss": 2.7886,
      "theoretical_loss": 3.7172067196378658,
      "tokens_seen": 826408960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004984537544921506,
      "loss": 2.7223,
      "theoretical_loss": 3.7171776706502335,
      "tokens_seen": 826474496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000498437204463779,
      "loss": 2.7601,
      "theoretical_loss": 3.717148624610886,
      "tokens_seen": 826540032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004984206544354076,
      "loss": 2.851,
      "theoretical_loss": 3.71711958151929,
      "tokens_seen": 826605568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004984041044070361,
      "loss": 2.8014,
      "theoretical_loss": 3.717090541374913,
      "tokens_seen": 826671104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004983875543786647,
      "loss": 2.7785,
      "theoretical_loss": 3.7170615041772224,
      "tokens_seen": 826736640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004983710043502932,
      "loss": 2.6226,
      "theoretical_loss": 3.7170324699256847,
      "tokens_seen": 826802176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004983544543219217,
      "loss": 2.7037,
      "theoretical_loss": 3.717003438619769,
      "tokens_seen": 826867712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004983379042935503,
      "loss": 2.8802,
      "theoretical_loss": 3.7169744102589415,
      "tokens_seen": 826933248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004983213542651787,
      "loss": 2.598,
      "theoretical_loss": 3.7169453848426715,
      "tokens_seen": 826998784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004983048042368072,
      "loss": 2.762,
      "theoretical_loss": 3.7169163623704256,
      "tokens_seen": 827064320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004982882542084358,
      "loss": 2.8145,
      "theoretical_loss": 3.716887342841673,
      "tokens_seen": 827129856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004982717041800643,
      "loss": 2.7182,
      "theoretical_loss": 3.716858326255881,
      "tokens_seen": 827195392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004982551541516929,
      "loss": 2.7139,
      "theoretical_loss": 3.7168293126125196,
      "tokens_seen": 827260928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004982386041233214,
      "loss": 2.528,
      "theoretical_loss": 3.7168003019110554,
      "tokens_seen": 827326464
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004982220540949499,
      "loss": 2.917,
      "theoretical_loss": 3.7167712941509587,
      "tokens_seen": 827392000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004982055040665785,
      "loss": 2.3883,
      "theoretical_loss": 3.7167422893316977,
      "tokens_seen": 827457536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004981889540382069,
      "loss": 2.5172,
      "theoretical_loss": 3.716713287452741,
      "tokens_seen": 827523072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004981724040098354,
      "loss": 2.5698,
      "theoretical_loss": 3.7166842885135583,
      "tokens_seen": 827588608
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000498155853981464,
      "loss": 2.8406,
      "theoretical_loss": 3.716655292513619,
      "tokens_seen": 827654144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004981393039530925,
      "loss": 2.8137,
      "theoretical_loss": 3.7166262994523915,
      "tokens_seen": 827719680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004981227539247211,
      "loss": 2.5545,
      "theoretical_loss": 3.7165973093293463,
      "tokens_seen": 827785216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004981062038963496,
      "loss": 2.4661,
      "theoretical_loss": 3.7165683221439525,
      "tokens_seen": 827850752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000498089653867978,
      "loss": 2.7133,
      "theoretical_loss": 3.7165393378956804,
      "tokens_seen": 827916288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004980731038396066,
      "loss": 2.847,
      "theoretical_loss": 3.716510356583999,
      "tokens_seen": 827981824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004980565538112351,
      "loss": 2.6591,
      "theoretical_loss": 3.71648137820838,
      "tokens_seen": 828047360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004980400037828636,
      "loss": 2.7663,
      "theoretical_loss": 3.716452402768292,
      "tokens_seen": 828112896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004980234537544922,
      "loss": 2.9592,
      "theoretical_loss": 3.716423430263206,
      "tokens_seen": 828178432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004980069037261207,
      "loss": 2.8239,
      "theoretical_loss": 3.7163944606925927,
      "tokens_seen": 828243968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004979903536977493,
      "loss": 2.8898,
      "theoretical_loss": 3.716365494055922,
      "tokens_seen": 828309504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004979738036693778,
      "loss": 2.7142,
      "theoretical_loss": 3.716336530352666,
      "tokens_seen": 828375040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004979572536410062,
      "loss": 2.946,
      "theoretical_loss": 3.716307569582294,
      "tokens_seen": 828440576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004979407036126347,
      "loss": 2.7213,
      "theoretical_loss": 3.716278611744278,
      "tokens_seen": 828506112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004979241535842632,
      "loss": 2.4248,
      "theoretical_loss": 3.7162496568380887,
      "tokens_seen": 828571648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004979076035558918,
      "loss": 2.67,
      "theoretical_loss": 3.716220704863198,
      "tokens_seen": 828637184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004978910535275203,
      "loss": 2.657,
      "theoretical_loss": 3.716191755819077,
      "tokens_seen": 828702720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004978745034991488,
      "loss": 3.0256,
      "theoretical_loss": 3.716162809705197,
      "tokens_seen": 828768256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004978579534707774,
      "loss": 2.8011,
      "theoretical_loss": 3.71613386652103,
      "tokens_seen": 828833792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004978414034424058,
      "loss": 2.6317,
      "theoretical_loss": 3.7161049262660484,
      "tokens_seen": 828899328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004978248534140344,
      "loss": 2.5848,
      "theoretical_loss": 3.7160759889397235,
      "tokens_seen": 828964864
    },
    {
      "epoch": 0.3,
      "objective/train/avg_token_score": 0.01202075183391571,
      "objective/train/avg_weight": 0.9858548641204834,
      "objective/train/docs_used": 477026,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8157782554626465,
      "objective/train/std_weight": 0.06933365762233734,
      "objective/train/theoretical_loss": 3.7160470545415274,
      "objective/train/tokens_used": 849490400,
      "theoretical_loss": 3.7160470545415274,
      "tokens_seen": 829030400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004978083033856629,
      "loss": 2.7552,
      "theoretical_loss": 3.7160470545415274,
      "tokens_seen": 829030400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004977917533572914,
      "loss": 2.8195,
      "theoretical_loss": 3.7160181230709326,
      "tokens_seen": 829095936
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00049777520332892,
      "loss": 2.8131,
      "theoretical_loss": 3.7159891945274115,
      "tokens_seen": 829161472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004977586533005485,
      "loss": 2.7812,
      "theoretical_loss": 3.715960268910437,
      "tokens_seen": 829227008
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000497742103272177,
      "loss": 2.7528,
      "theoretical_loss": 3.715931346219481,
      "tokens_seen": 829292544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004977255532438056,
      "loss": 2.6261,
      "theoretical_loss": 3.715902426454017,
      "tokens_seen": 829358080
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000497709003215434,
      "loss": 2.8595,
      "theoretical_loss": 3.715873509613517,
      "tokens_seen": 829423616
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004976924531870626,
      "loss": 2.6274,
      "theoretical_loss": 3.7158445956974555,
      "tokens_seen": 829489152
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004976759031586911,
      "loss": 2.5103,
      "theoretical_loss": 3.715815684705304,
      "tokens_seen": 829554688
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004976593531303196,
      "loss": 2.6525,
      "theoretical_loss": 3.715786776636538,
      "tokens_seen": 829620224
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004976428031019482,
      "loss": 2.9031,
      "theoretical_loss": 3.715757871490629,
      "tokens_seen": 829685760
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004976262530735767,
      "loss": 2.626,
      "theoretical_loss": 3.715728969267052,
      "tokens_seen": 829751296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004976097030452052,
      "loss": 2.7266,
      "theoretical_loss": 3.71570006996528,
      "tokens_seen": 829816832
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004975931530168337,
      "loss": 2.525,
      "theoretical_loss": 3.715671173584787,
      "tokens_seen": 829882368
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004975766029884622,
      "loss": 2.6882,
      "theoretical_loss": 3.7156422801250475,
      "tokens_seen": 829947904
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004975600529600908,
      "loss": 2.3594,
      "theoretical_loss": 3.7156133895855357,
      "tokens_seen": 830013440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004975435029317193,
      "loss": 2.5986,
      "theoretical_loss": 3.7155845019657248,
      "tokens_seen": 830078976
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004975269529033478,
      "loss": 2.8496,
      "theoretical_loss": 3.7155556172650903,
      "tokens_seen": 830144512
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004975104028749764,
      "loss": 2.4201,
      "theoretical_loss": 3.715526735483107,
      "tokens_seen": 830210048
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004974938528466049,
      "loss": 2.9032,
      "theoretical_loss": 3.7154978566192485,
      "tokens_seen": 830275584
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004974773028182333,
      "loss": 2.7859,
      "theoretical_loss": 3.715468980672991,
      "tokens_seen": 830341120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004974607527898619,
      "loss": 2.7001,
      "theoretical_loss": 3.715440107643808,
      "tokens_seen": 830406656
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004974442027614904,
      "loss": 2.62,
      "theoretical_loss": 3.7154112375311765,
      "tokens_seen": 830472192
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000497427652733119,
      "loss": 2.7123,
      "theoretical_loss": 3.71538237033457,
      "tokens_seen": 830537728
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004974111027047475,
      "loss": 2.7094,
      "theoretical_loss": 3.7153535060534644,
      "tokens_seen": 830603264
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000497394552676376,
      "loss": 2.7213,
      "theoretical_loss": 3.7153246446873363,
      "tokens_seen": 830668800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004973780026480046,
      "loss": 2.9984,
      "theoretical_loss": 3.71529578623566,
      "tokens_seen": 830734336
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004973614526196331,
      "loss": 2.6067,
      "theoretical_loss": 3.715266930697912,
      "tokens_seen": 830799872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004973449025912615,
      "loss": 2.4507,
      "theoretical_loss": 3.715238078073568,
      "tokens_seen": 830865408
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004973283525628901,
      "loss": 2.878,
      "theoretical_loss": 3.715209228362104,
      "tokens_seen": 830930944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004973118025345186,
      "loss": 2.6489,
      "theoretical_loss": 3.715180381562997,
      "tokens_seen": 830996480
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004972952525061472,
      "loss": 2.6785,
      "theoretical_loss": 3.7151515376757223,
      "tokens_seen": 831062016
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004972787024777757,
      "loss": 2.5143,
      "theoretical_loss": 3.7151226966997566,
      "tokens_seen": 831127552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004972621524494042,
      "loss": 2.8371,
      "theoretical_loss": 3.715093858634577,
      "tokens_seen": 831193088
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004972456024210328,
      "loss": 2.6273,
      "theoretical_loss": 3.71506502347966,
      "tokens_seen": 831258624
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004972290523926612,
      "loss": 2.5932,
      "theoretical_loss": 3.715036191234483,
      "tokens_seen": 831324160
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004972125023642897,
      "loss": 2.8775,
      "theoretical_loss": 3.715007361898522,
      "tokens_seen": 831389696
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004971959523359183,
      "loss": 2.794,
      "theoretical_loss": 3.7149785354712552,
      "tokens_seen": 831455232
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004971794023075468,
      "loss": 2.5204,
      "theoretical_loss": 3.7149497119521593,
      "tokens_seen": 831520768
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004971628522791754,
      "loss": 2.8429,
      "theoretical_loss": 3.7149208913407117,
      "tokens_seen": 831586304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004971463022508039,
      "loss": 2.6857,
      "theoretical_loss": 3.71489207363639,
      "tokens_seen": 831651840
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004971297522224324,
      "loss": 2.6322,
      "theoretical_loss": 3.714863258838672,
      "tokens_seen": 831717376
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000497113202194061,
      "loss": 2.8863,
      "theoretical_loss": 3.7148344469470356,
      "tokens_seen": 831782912
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004970966521656894,
      "loss": 2.8463,
      "theoretical_loss": 3.714805637960959,
      "tokens_seen": 831848448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004970801021373179,
      "loss": 2.861,
      "theoretical_loss": 3.71477683187992,
      "tokens_seen": 831913984
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004970635521089465,
      "loss": 2.7588,
      "theoretical_loss": 3.7147480287033967,
      "tokens_seen": 831979520
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000497047002080575,
      "loss": 2.7306,
      "theoretical_loss": 3.7147192284308677,
      "tokens_seen": 832045056
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004970304520522036,
      "loss": 2.6141,
      "theoretical_loss": 3.7146904310618116,
      "tokens_seen": 832110592
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004970139020238321,
      "loss": 2.6183,
      "theoretical_loss": 3.7146616365957073,
      "tokens_seen": 832176128
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004969973519954606,
      "loss": 2.9332,
      "theoretical_loss": 3.7146328450320327,
      "tokens_seen": 832241664
    },
    {
      "epoch": 0.3,
      "objective/train/avg_token_score": 0.01314801536500454,
      "objective/train/avg_weight": 0.9845210909843445,
      "objective/train/docs_used": 478220,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2295846939086914,
      "objective/train/std_weight": 0.09336137026548386,
      "objective/train/theoretical_loss": 3.714604056370267,
      "objective/train/tokens_used": 852767200,
      "theoretical_loss": 3.714604056370267,
      "tokens_seen": 832307200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004969808019670891,
      "loss": 2.8176,
      "theoretical_loss": 3.714604056370267,
      "tokens_seen": 832307200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004969642519387175,
      "loss": 2.8116,
      "theoretical_loss": 3.71457527060989,
      "tokens_seen": 832372736
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004969477019103461,
      "loss": 2.7066,
      "theoretical_loss": 3.7145464877503804,
      "tokens_seen": 832438272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004969311518819746,
      "loss": 2.6629,
      "theoretical_loss": 3.7145177077912175,
      "tokens_seen": 832503808
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004969146018536032,
      "loss": 2.8531,
      "theoretical_loss": 3.714488930731881,
      "tokens_seen": 832569344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004968980518252317,
      "loss": 2.8455,
      "theoretical_loss": 3.7144601565718496,
      "tokens_seen": 832634880
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004968815017968602,
      "loss": 2.4549,
      "theoretical_loss": 3.714431385310604,
      "tokens_seen": 832700416
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004968649517684887,
      "loss": 2.6079,
      "theoretical_loss": 3.7144026169476243,
      "tokens_seen": 832765952
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004968484017401172,
      "loss": 2.4199,
      "theoretical_loss": 3.714373851482389,
      "tokens_seen": 832831488
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004968318517117457,
      "loss": 2.8183,
      "theoretical_loss": 3.7143450889143796,
      "tokens_seen": 832897024
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004968153016833743,
      "loss": 2.6873,
      "theoretical_loss": 3.7143163292430765,
      "tokens_seen": 832962560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004967987516550028,
      "loss": 2.7469,
      "theoretical_loss": 3.714287572467959,
      "tokens_seen": 833028096
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004967822016266314,
      "loss": 2.8106,
      "theoretical_loss": 3.714258818588508,
      "tokens_seen": 833093632
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004967656515982599,
      "loss": 3.0093,
      "theoretical_loss": 3.7142300676042046,
      "tokens_seen": 833159168
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004967491015698883,
      "loss": 2.5191,
      "theoretical_loss": 3.7142013195145287,
      "tokens_seen": 833224704
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004967325515415169,
      "loss": 2.7605,
      "theoretical_loss": 3.7141725743189626,
      "tokens_seen": 833290240
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004967160015131454,
      "loss": 2.6768,
      "theoretical_loss": 3.714143832016987,
      "tokens_seen": 833355776
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004966994514847739,
      "loss": 2.583,
      "theoretical_loss": 3.7141150926080817,
      "tokens_seen": 833421312
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004966829014564025,
      "loss": 2.5958,
      "theoretical_loss": 3.7140863560917294,
      "tokens_seen": 833486848
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000496666351428031,
      "loss": 2.7978,
      "theoretical_loss": 3.7140576224674113,
      "tokens_seen": 833552384
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004966498013996596,
      "loss": 2.5358,
      "theoretical_loss": 3.714028891734609,
      "tokens_seen": 833617920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004966332513712881,
      "loss": 2.8181,
      "theoretical_loss": 3.714000163892804,
      "tokens_seen": 833683456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004966167013429165,
      "loss": 2.7148,
      "theoretical_loss": 3.7139714389414786,
      "tokens_seen": 833748992
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004966001513145451,
      "loss": 2.8622,
      "theoretical_loss": 3.7139427168801147,
      "tokens_seen": 833814528
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004965836012861736,
      "loss": 2.8178,
      "theoretical_loss": 3.713913997708194,
      "tokens_seen": 833880064
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004965670512578021,
      "loss": 2.6569,
      "theoretical_loss": 3.713885281425199,
      "tokens_seen": 833945600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004965505012294307,
      "loss": 2.8169,
      "theoretical_loss": 3.7138565680306117,
      "tokens_seen": 834011136
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004965339512010592,
      "loss": 2.6412,
      "theoretical_loss": 3.7138278575239156,
      "tokens_seen": 834076672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004965174011726878,
      "loss": 2.7128,
      "theoretical_loss": 3.7137991499045926,
      "tokens_seen": 834142208
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004965008511443162,
      "loss": 2.7224,
      "theoretical_loss": 3.713770445172126,
      "tokens_seen": 834207744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004964843011159447,
      "loss": 2.5008,
      "theoretical_loss": 3.7137417433259987,
      "tokens_seen": 834273280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004964677510875733,
      "loss": 2.6306,
      "theoretical_loss": 3.7137130443656936,
      "tokens_seen": 834338816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004964512010592018,
      "loss": 2.5919,
      "theoretical_loss": 3.713684348290693,
      "tokens_seen": 834404352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004964346510308303,
      "loss": 2.6704,
      "theoretical_loss": 3.713655655100482,
      "tokens_seen": 834469888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004964181010024589,
      "loss": 2.7836,
      "theoretical_loss": 3.713626964794543,
      "tokens_seen": 834535424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004964015509740874,
      "loss": 2.8147,
      "theoretical_loss": 3.7135982773723595,
      "tokens_seen": 834600960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000496385000945716,
      "loss": 2.6733,
      "theoretical_loss": 3.713569592833416,
      "tokens_seen": 834666496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004963684509173444,
      "loss": 2.8315,
      "theoretical_loss": 3.713540911177195,
      "tokens_seen": 834732032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004963519008889729,
      "loss": 2.8383,
      "theoretical_loss": 3.713512232403182,
      "tokens_seen": 834797568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004963353508606015,
      "loss": 2.8917,
      "theoretical_loss": 3.7134835565108606,
      "tokens_seen": 834863104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00049631880083223,
      "loss": 2.8079,
      "theoretical_loss": 3.7134548834997148,
      "tokens_seen": 834928640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004963022508038585,
      "loss": 2.5967,
      "theoretical_loss": 3.713426213369229,
      "tokens_seen": 834994176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004962857007754871,
      "loss": 2.669,
      "theoretical_loss": 3.7133975461188884,
      "tokens_seen": 835059712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004962691507471156,
      "loss": 2.6732,
      "theoretical_loss": 3.7133688817481767,
      "tokens_seen": 835125248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004962526007187441,
      "loss": 2.6987,
      "theoretical_loss": 3.7133402202565793,
      "tokens_seen": 835190784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004962360506903726,
      "loss": 2.9966,
      "theoretical_loss": 3.713311561643581,
      "tokens_seen": 835256320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004962195006620011,
      "loss": 2.5525,
      "theoretical_loss": 3.7132829059086667,
      "tokens_seen": 835321856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004962029506336297,
      "loss": 2.8148,
      "theoretical_loss": 3.713254253051322,
      "tokens_seen": 835387392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004961864006052582,
      "loss": 2.5145,
      "theoretical_loss": 3.7132256030710318,
      "tokens_seen": 835452928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004961698505768867,
      "loss": 2.7422,
      "theoretical_loss": 3.713196955967282,
      "tokens_seen": 835518464
    },
    {
      "epoch": 0.3,
      "objective/train/avg_token_score": 0.01864147186279297,
      "objective/train/avg_weight": 0.9780945181846619,
      "objective/train/docs_used": 480064,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0167973041534424,
      "objective/train/std_weight": 0.12047401070594788,
      "objective/train/theoretical_loss": 3.713168311739558,
      "objective/train/tokens_used": 856044000,
      "theoretical_loss": 3.713168311739558,
      "tokens_seen": 835584000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004961533005485153,
      "loss": 2.7568,
      "theoretical_loss": 3.713168311739558,
      "tokens_seen": 835584000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004961367505201437,
      "loss": 2.6809,
      "theoretical_loss": 3.7131396703873456,
      "tokens_seen": 835649536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004961202004917723,
      "loss": 2.5674,
      "theoretical_loss": 3.71311103191013,
      "tokens_seen": 835715072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004961036504634008,
      "loss": 2.7147,
      "theoretical_loss": 3.713082396307398,
      "tokens_seen": 835780608
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004960871004350293,
      "loss": 2.8649,
      "theoretical_loss": 3.7130537635786354,
      "tokens_seen": 835846144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004960705504066579,
      "loss": 2.8425,
      "theoretical_loss": 3.7130251337233284,
      "tokens_seen": 835911680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004960540003782864,
      "loss": 2.6699,
      "theoretical_loss": 3.7129965067409634,
      "tokens_seen": 835977216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004960374503499149,
      "loss": 2.7602,
      "theoretical_loss": 3.712967882631027,
      "tokens_seen": 836042752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004960209003215435,
      "loss": 2.6279,
      "theoretical_loss": 3.7129392613930063,
      "tokens_seen": 836108288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004960043502931719,
      "loss": 2.5982,
      "theoretical_loss": 3.712910643026387,
      "tokens_seen": 836173824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004959878002648004,
      "loss": 2.7996,
      "theoretical_loss": 3.712882027530657,
      "tokens_seen": 836239360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004959712502364289,
      "loss": 2.5703,
      "theoretical_loss": 3.712853414905303,
      "tokens_seen": 836304896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004959547002080575,
      "loss": 2.6449,
      "theoretical_loss": 3.712824805149812,
      "tokens_seen": 836370432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000495938150179686,
      "loss": 2.526,
      "theoretical_loss": 3.7127961982636712,
      "tokens_seen": 836435968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004959216001513145,
      "loss": 2.7126,
      "theoretical_loss": 3.712767594246369,
      "tokens_seen": 836501504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004959050501229431,
      "loss": 2.6637,
      "theoretical_loss": 3.7127389930973917,
      "tokens_seen": 836567040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004958885000945715,
      "loss": 2.7335,
      "theoretical_loss": 3.712710394816228,
      "tokens_seen": 836632576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004958719500662001,
      "loss": 2.7072,
      "theoretical_loss": 3.7126817994023655,
      "tokens_seen": 836698112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004958554000378286,
      "loss": 2.4949,
      "theoretical_loss": 3.7126532068552915,
      "tokens_seen": 836763648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004958388500094571,
      "loss": 2.6165,
      "theoretical_loss": 3.7126246171744945,
      "tokens_seen": 836829184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004958222999810857,
      "loss": 2.5598,
      "theoretical_loss": 3.712596030359463,
      "tokens_seen": 836894720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004958057499527142,
      "loss": 2.8644,
      "theoretical_loss": 3.712567446409685,
      "tokens_seen": 836960256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004957891999243427,
      "loss": 2.8367,
      "theoretical_loss": 3.7125388653246496,
      "tokens_seen": 837025792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004957726498959712,
      "loss": 2.5738,
      "theoretical_loss": 3.7125102871038447,
      "tokens_seen": 837091328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004957560998675997,
      "loss": 2.6938,
      "theoretical_loss": 3.7124817117467597,
      "tokens_seen": 837156864
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004957395498392283,
      "loss": 2.6345,
      "theoretical_loss": 3.712453139252883,
      "tokens_seen": 837222400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004957229998108568,
      "loss": 2.8664,
      "theoretical_loss": 3.712424569621703,
      "tokens_seen": 837287936
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004957064497824853,
      "loss": 2.7944,
      "theoretical_loss": 3.712396002852711,
      "tokens_seen": 837353472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004956898997541139,
      "loss": 2.8555,
      "theoretical_loss": 3.712367438945394,
      "tokens_seen": 837419008
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004956733497257424,
      "loss": 2.82,
      "theoretical_loss": 3.712338877899242,
      "tokens_seen": 837484544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004956567996973708,
      "loss": 2.9007,
      "theoretical_loss": 3.7123103197137457,
      "tokens_seen": 837550080
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004956402496689994,
      "loss": 2.4497,
      "theoretical_loss": 3.7122817643883934,
      "tokens_seen": 837615616
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004956236996406279,
      "loss": 2.8682,
      "theoretical_loss": 3.7122532119226754,
      "tokens_seen": 837681152
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004956071496122565,
      "loss": 2.9203,
      "theoretical_loss": 3.712224662316082,
      "tokens_seen": 837746688
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000495590599583885,
      "loss": 2.6054,
      "theoretical_loss": 3.7121961155681027,
      "tokens_seen": 837812224
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004955740495555135,
      "loss": 2.549,
      "theoretical_loss": 3.712167571678228,
      "tokens_seen": 837877760
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004955574995271421,
      "loss": 2.6231,
      "theoretical_loss": 3.712139030645948,
      "tokens_seen": 837943296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004955409494987706,
      "loss": 2.6096,
      "theoretical_loss": 3.7121104924707535,
      "tokens_seen": 838008832
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000495524399470399,
      "loss": 2.8574,
      "theoretical_loss": 3.712081957152135,
      "tokens_seen": 838074368
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004955078494420276,
      "loss": 2.7688,
      "theoretical_loss": 3.712053424689583,
      "tokens_seen": 838139904
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004954912994136561,
      "loss": 2.7757,
      "theoretical_loss": 3.712024895082588,
      "tokens_seen": 838205440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004954747493852847,
      "loss": 2.498,
      "theoretical_loss": 3.711996368330642,
      "tokens_seen": 838270976
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004954581993569132,
      "loss": 2.678,
      "theoretical_loss": 3.7119678444332354,
      "tokens_seen": 838336512
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004954416493285417,
      "loss": 2.7437,
      "theoretical_loss": 3.71193932338986,
      "tokens_seen": 838402048
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004954250993001703,
      "loss": 2.8547,
      "theoretical_loss": 3.711910805200006,
      "tokens_seen": 838467584
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004954085492717988,
      "loss": 2.5818,
      "theoretical_loss": 3.7118822898631665,
      "tokens_seen": 838533120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004953919992434272,
      "loss": 2.6333,
      "theoretical_loss": 3.711853777378832,
      "tokens_seen": 838598656
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004953754492150558,
      "loss": 2.5494,
      "theoretical_loss": 3.7118252677464945,
      "tokens_seen": 838664192
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004953588991866843,
      "loss": 2.7783,
      "theoretical_loss": 3.7117967609656457,
      "tokens_seen": 838729728
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004953423491583129,
      "loss": 2.8569,
      "theoretical_loss": 3.7117682570357786,
      "tokens_seen": 838795264
    },
    {
      "epoch": 0.3,
      "objective/train/avg_token_score": 0.02161218225955963,
      "objective/train/avg_weight": 0.974566638469696,
      "objective/train/docs_used": 482017,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.256676197052002,
      "objective/train/std_weight": 0.09085341542959213,
      "objective/train/theoretical_loss": 3.7117397559563843,
      "objective/train/tokens_used": 859320800,
      "theoretical_loss": 3.7117397559563843,
      "tokens_seen": 838860800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004953257991299414,
      "loss": 2.7116,
      "theoretical_loss": 3.7117397559563843,
      "tokens_seen": 838860800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004953092491015699,
      "loss": 2.7719,
      "theoretical_loss": 3.7117112577269555,
      "tokens_seen": 838926336
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004952926990731985,
      "loss": 2.6558,
      "theoretical_loss": 3.7116827623469844,
      "tokens_seen": 838991872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004952761490448269,
      "loss": 2.88,
      "theoretical_loss": 3.7116542698159636,
      "tokens_seen": 839057408
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004952595990164554,
      "loss": 2.7178,
      "theoretical_loss": 3.711625780133386,
      "tokens_seen": 839122944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000495243048988084,
      "loss": 2.6582,
      "theoretical_loss": 3.711597293298744,
      "tokens_seen": 839188480
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004952264989597125,
      "loss": 2.7889,
      "theoretical_loss": 3.7115688093115313,
      "tokens_seen": 839254016
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004952099489313411,
      "loss": 2.5621,
      "theoretical_loss": 3.7115403281712407,
      "tokens_seen": 839319552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004951933989029696,
      "loss": 2.6407,
      "theoretical_loss": 3.7115118498773647,
      "tokens_seen": 839385088
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004951768488745981,
      "loss": 2.6951,
      "theoretical_loss": 3.711483374429397,
      "tokens_seen": 839450624
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004951602988462267,
      "loss": 2.6006,
      "theoretical_loss": 3.7114549018268317,
      "tokens_seen": 839516160
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004951437488178551,
      "loss": 2.5837,
      "theoretical_loss": 3.7114264320691612,
      "tokens_seen": 839581696
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004951271987894836,
      "loss": 2.6021,
      "theoretical_loss": 3.7113979651558804,
      "tokens_seen": 839647232
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004951106487611122,
      "loss": 2.6086,
      "theoretical_loss": 3.7113695010864824,
      "tokens_seen": 839712768
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004950940987327407,
      "loss": 2.5418,
      "theoretical_loss": 3.7113410398604616,
      "tokens_seen": 839778304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004950775487043693,
      "loss": 2.4193,
      "theoretical_loss": 3.711312581477311,
      "tokens_seen": 839843840
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004950609986759978,
      "loss": 2.7974,
      "theoretical_loss": 3.7112841259365266,
      "tokens_seen": 839909376
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004950444486476263,
      "loss": 2.6223,
      "theoretical_loss": 3.7112556732376016,
      "tokens_seen": 839974912
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004950278986192548,
      "loss": 2.7446,
      "theoretical_loss": 3.711227223380031,
      "tokens_seen": 840040448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004950113485908832,
      "loss": 2.6254,
      "theoretical_loss": 3.7111987763633083,
      "tokens_seen": 840105984
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004949947985625118,
      "loss": 2.5077,
      "theoretical_loss": 3.71117033218693,
      "tokens_seen": 840171520
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004949782485341403,
      "loss": 2.6237,
      "theoretical_loss": 3.71114189085039,
      "tokens_seen": 840237056
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004949616985057688,
      "loss": 2.5176,
      "theoretical_loss": 3.7111134523531835,
      "tokens_seen": 840302592
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004949451484773974,
      "loss": 2.4988,
      "theoretical_loss": 3.711085016694805,
      "tokens_seen": 840368128
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004949285984490259,
      "loss": 2.7028,
      "theoretical_loss": 3.7110565838747505,
      "tokens_seen": 840433664
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004949120484206544,
      "loss": 2.742,
      "theoretical_loss": 3.7110281538925154,
      "tokens_seen": 840499200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004948954983922829,
      "loss": 2.7266,
      "theoretical_loss": 3.7109997267475947,
      "tokens_seen": 840564736
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004948789483639114,
      "loss": 2.6348,
      "theoretical_loss": 3.710971302439485,
      "tokens_seen": 840630272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00049486239833554,
      "loss": 2.7129,
      "theoretical_loss": 3.710942880967681,
      "tokens_seen": 840695808
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004948458483071685,
      "loss": 2.7384,
      "theoretical_loss": 3.710914462331679,
      "tokens_seen": 840761344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000494829298278797,
      "loss": 2.7249,
      "theoretical_loss": 3.710886046530975,
      "tokens_seen": 840826880
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004948127482504256,
      "loss": 2.8159,
      "theoretical_loss": 3.710857633565066,
      "tokens_seen": 840892416
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000494796198222054,
      "loss": 2.87,
      "theoretical_loss": 3.7108292234334472,
      "tokens_seen": 840957952
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004947796481936826,
      "loss": 2.749,
      "theoretical_loss": 3.710800816135615,
      "tokens_seen": 841023488
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004947630981653111,
      "loss": 2.7665,
      "theoretical_loss": 3.7107724116710665,
      "tokens_seen": 841089024
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004947465481369396,
      "loss": 2.6749,
      "theoretical_loss": 3.7107440100392988,
      "tokens_seen": 841154560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004947299981085682,
      "loss": 2.5371,
      "theoretical_loss": 3.7107156112398076,
      "tokens_seen": 841220096
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004947134480801967,
      "loss": 2.9169,
      "theoretical_loss": 3.710687215272091,
      "tokens_seen": 841285632
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004946968980518252,
      "loss": 2.6511,
      "theoretical_loss": 3.7106588221356445,
      "tokens_seen": 841351168
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004946803480234538,
      "loss": 2.6888,
      "theoretical_loss": 3.7106304318299674,
      "tokens_seen": 841416704
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004946637979950822,
      "loss": 2.7102,
      "theoretical_loss": 3.7106020443545553,
      "tokens_seen": 841482240
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004946472479667108,
      "loss": 2.4678,
      "theoretical_loss": 3.7105736597089063,
      "tokens_seen": 841547776
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004946306979383393,
      "loss": 2.7064,
      "theoretical_loss": 3.7105452778925176,
      "tokens_seen": 841613312
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004946141479099678,
      "loss": 2.7001,
      "theoretical_loss": 3.7105168989048876,
      "tokens_seen": 841678848
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004945975978815964,
      "loss": 2.6508,
      "theoretical_loss": 3.7104885227455138,
      "tokens_seen": 841744384
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004945810478532249,
      "loss": 2.7396,
      "theoretical_loss": 3.710460149413894,
      "tokens_seen": 841809920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004945644978248534,
      "loss": 2.6041,
      "theoretical_loss": 3.7104317789095265,
      "tokens_seen": 841875456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004945479477964819,
      "loss": 2.716,
      "theoretical_loss": 3.7104034112319093,
      "tokens_seen": 841940992
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004945313977681104,
      "loss": 2.5348,
      "theoretical_loss": 3.710375046380541,
      "tokens_seen": 842006528
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000494514847739739,
      "loss": 2.7082,
      "theoretical_loss": 3.7103466843549207,
      "tokens_seen": 842072064
    },
    {
      "epoch": 0.3,
      "objective/train/avg_token_score": 0.008207079023122787,
      "objective/train/avg_weight": 0.9903393983840942,
      "objective/train/docs_used": 484166,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0806450843811035,
      "objective/train/std_weight": 0.060667045414447784,
      "objective/train/theoretical_loss": 3.710318325154545,
      "objective/train/tokens_used": 862597600,
      "theoretical_loss": 3.710318325154545,
      "tokens_seen": 842137600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004944982977113675,
      "loss": 2.7269,
      "theoretical_loss": 3.710318325154545,
      "tokens_seen": 842137600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000494481747682996,
      "loss": 2.4663,
      "theoretical_loss": 3.7102899687789153,
      "tokens_seen": 842203136
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004944651976546246,
      "loss": 2.7808,
      "theoretical_loss": 3.710261615227528,
      "tokens_seen": 842268672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004944486476262531,
      "loss": 2.7812,
      "theoretical_loss": 3.710233264499884,
      "tokens_seen": 842334208
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004944320975978817,
      "loss": 2.7048,
      "theoretical_loss": 3.7102049165954814,
      "tokens_seen": 842399744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004944155475695101,
      "loss": 2.7771,
      "theoretical_loss": 3.710176571513819,
      "tokens_seen": 842465280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004943989975411386,
      "loss": 2.4336,
      "theoretical_loss": 3.710148229254398,
      "tokens_seen": 842530816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004943824475127672,
      "loss": 2.7738,
      "theoretical_loss": 3.710119889816716,
      "tokens_seen": 842596352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004943658974843957,
      "loss": 2.8701,
      "theoretical_loss": 3.7100915532002734,
      "tokens_seen": 842661888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004943493474560242,
      "loss": 2.9432,
      "theoretical_loss": 3.71006321940457,
      "tokens_seen": 842727424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004943327974276528,
      "loss": 2.5758,
      "theoretical_loss": 3.7100348884291057,
      "tokens_seen": 842792960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004943162473992813,
      "loss": 2.7466,
      "theoretical_loss": 3.7100065602733805,
      "tokens_seen": 842858496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004942996973709098,
      "loss": 2.6334,
      "theoretical_loss": 3.7099782349368944,
      "tokens_seen": 842924032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004942831473425383,
      "loss": 2.6482,
      "theoretical_loss": 3.7099499124191473,
      "tokens_seen": 842989568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004942665973141668,
      "loss": 2.9393,
      "theoretical_loss": 3.709921592719641,
      "tokens_seen": 843055104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004942500472857954,
      "loss": 2.5342,
      "theoretical_loss": 3.7098932758378744,
      "tokens_seen": 843120640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004942334972574239,
      "loss": 2.6796,
      "theoretical_loss": 3.7098649617733486,
      "tokens_seen": 843186176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004942169472290524,
      "loss": 2.8417,
      "theoretical_loss": 3.709836650525565,
      "tokens_seen": 843251712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000494200397200681,
      "loss": 2.8465,
      "theoretical_loss": 3.709808342094024,
      "tokens_seen": 843317248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004941838471723094,
      "loss": 2.7718,
      "theoretical_loss": 3.709780036478227,
      "tokens_seen": 843382784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000494167297143938,
      "loss": 2.4828,
      "theoretical_loss": 3.7097517336776744,
      "tokens_seen": 843448320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004941507471155665,
      "loss": 2.8197,
      "theoretical_loss": 3.7097234336918685,
      "tokens_seen": 843513856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000494134197087195,
      "loss": 2.8882,
      "theoretical_loss": 3.7096951365203097,
      "tokens_seen": 843579392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004941176470588236,
      "loss": 2.758,
      "theoretical_loss": 3.7096668421625,
      "tokens_seen": 843644928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004941010970304521,
      "loss": 2.764,
      "theoretical_loss": 3.7096385506179415,
      "tokens_seen": 843710464
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004940845470020806,
      "loss": 2.8034,
      "theoretical_loss": 3.7096102618861355,
      "tokens_seen": 843776000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004940679969737092,
      "loss": 2.8576,
      "theoretical_loss": 3.709581975966584,
      "tokens_seen": 843841536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004940514469453376,
      "loss": 2.6034,
      "theoretical_loss": 3.709553692858789,
      "tokens_seen": 843907072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004940348969169662,
      "loss": 2.8352,
      "theoretical_loss": 3.709525412562253,
      "tokens_seen": 843972608
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004940183468885946,
      "loss": 2.4003,
      "theoretical_loss": 3.709497135076478,
      "tokens_seen": 844038144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004940017968602232,
      "loss": 2.6532,
      "theoretical_loss": 3.709468860400966,
      "tokens_seen": 844103680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004939852468318517,
      "loss": 2.8821,
      "theoretical_loss": 3.7094405885352204,
      "tokens_seen": 844169216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004939686968034802,
      "loss": 2.8183,
      "theoretical_loss": 3.7094123194787434,
      "tokens_seen": 844234752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004939521467751088,
      "loss": 2.5829,
      "theoretical_loss": 3.709384053231038,
      "tokens_seen": 844300288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004939355967467372,
      "loss": 2.3738,
      "theoretical_loss": 3.7093557897916067,
      "tokens_seen": 844365824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004939190467183657,
      "loss": 2.646,
      "theoretical_loss": 3.7093275291599532,
      "tokens_seen": 844431360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004939024966899943,
      "loss": 2.6171,
      "theoretical_loss": 3.7092992713355804,
      "tokens_seen": 844496896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004938859466616228,
      "loss": 2.7839,
      "theoretical_loss": 3.7092710163179916,
      "tokens_seen": 844562432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004938693966332514,
      "loss": 2.7262,
      "theoretical_loss": 3.7092427641066905,
      "tokens_seen": 844627968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004938528466048799,
      "loss": 2.5523,
      "theoretical_loss": 3.7092145147011806,
      "tokens_seen": 844693504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004938362965765084,
      "loss": 2.773,
      "theoretical_loss": 3.709186268100965,
      "tokens_seen": 844759040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004938197465481369,
      "loss": 2.7108,
      "theoretical_loss": 3.7091580243055478,
      "tokens_seen": 844824576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004938031965197654,
      "loss": 2.8156,
      "theoretical_loss": 3.709129783314433,
      "tokens_seen": 844890112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004937866464913939,
      "loss": 2.8925,
      "theoretical_loss": 3.709101545127125,
      "tokens_seen": 844955648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004937700964630225,
      "loss": 2.6056,
      "theoretical_loss": 3.709073309743128,
      "tokens_seen": 845021184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000493753546434651,
      "loss": 3.0209,
      "theoretical_loss": 3.709045077161946,
      "tokens_seen": 845086720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004937369964062796,
      "loss": 2.705,
      "theoretical_loss": 3.7090168473830833,
      "tokens_seen": 845152256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004937204463779081,
      "loss": 2.7569,
      "theoretical_loss": 3.708988620406045,
      "tokens_seen": 845217792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004937038963495365,
      "loss": 2.8281,
      "theoretical_loss": 3.708960396230335,
      "tokens_seen": 845283328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004936873463211651,
      "loss": 2.7417,
      "theoretical_loss": 3.7089321748554593,
      "tokens_seen": 845348864
    },
    {
      "epoch": 0.3,
      "objective/train/avg_token_score": 0.024202242493629456,
      "objective/train/avg_weight": 0.9715213179588318,
      "objective/train/docs_used": 485899,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6180062294006348,
      "objective/train/std_weight": 0.10681289434432983,
      "objective/train/theoretical_loss": 3.7089039562809223,
      "objective/train/tokens_used": 865874400,
      "theoretical_loss": 3.7089039562809223,
      "tokens_seen": 845414400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004936873463211651,
      "loss": 2.4516,
      "theoretical_loss": 3.7089039562809223,
      "tokens_seen": 845414400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004936707962927936,
      "loss": 2.099,
      "theoretical_loss": 3.7088757405062283,
      "tokens_seen": 845479936
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004936542462644221,
      "loss": 2.6118,
      "theoretical_loss": 3.7088475275308834,
      "tokens_seen": 845545472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004936376962360507,
      "loss": 2.9072,
      "theoretical_loss": 3.708819317354393,
      "tokens_seen": 845611008
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004936211462076792,
      "loss": 2.6568,
      "theoretical_loss": 3.708791109976262,
      "tokens_seen": 845676544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004936045961793078,
      "loss": 2.5284,
      "theoretical_loss": 3.7087629053959965,
      "tokens_seen": 845742080
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004935880461509363,
      "loss": 2.7172,
      "theoretical_loss": 3.7087347036131018,
      "tokens_seen": 845807616
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004935714961225647,
      "loss": 2.5166,
      "theoretical_loss": 3.7087065046270844,
      "tokens_seen": 845873152
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004935549460941933,
      "loss": 2.7155,
      "theoretical_loss": 3.708678308437449,
      "tokens_seen": 845938688
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004935383960658218,
      "loss": 2.5962,
      "theoretical_loss": 3.708650115043703,
      "tokens_seen": 846004224
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004935218460374503,
      "loss": 2.786,
      "theoretical_loss": 3.7086219244453518,
      "tokens_seen": 846069760
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004935052960090789,
      "loss": 2.8501,
      "theoretical_loss": 3.7085937366419017,
      "tokens_seen": 846135296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004934887459807074,
      "loss": 2.539,
      "theoretical_loss": 3.7085655516328604,
      "tokens_seen": 846200832
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000493472195952336,
      "loss": 2.7302,
      "theoretical_loss": 3.7085373694177326,
      "tokens_seen": 846266368
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004934556459239644,
      "loss": 2.8423,
      "theoretical_loss": 3.7085091899960263,
      "tokens_seen": 846331904
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004934390958955929,
      "loss": 2.7588,
      "theoretical_loss": 3.7084810133672486,
      "tokens_seen": 846397440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004934225458672215,
      "loss": 2.8499,
      "theoretical_loss": 3.708452839530905,
      "tokens_seen": 846462976
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00049340599583885,
      "loss": 2.5203,
      "theoretical_loss": 3.7084246684865043,
      "tokens_seen": 846528512
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004933894458104785,
      "loss": 2.8991,
      "theoretical_loss": 3.708396500233552,
      "tokens_seen": 846594048
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004933728957821071,
      "loss": 2.8889,
      "theoretical_loss": 3.708368334771557,
      "tokens_seen": 846659584
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004933563457537356,
      "loss": 2.7365,
      "theoretical_loss": 3.7083401721000255,
      "tokens_seen": 846725120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004933397957253642,
      "loss": 2.7139,
      "theoretical_loss": 3.708312012218466,
      "tokens_seen": 846790656
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004933232456969926,
      "loss": 2.8348,
      "theoretical_loss": 3.708283855126386,
      "tokens_seen": 846856192
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004933066956686211,
      "loss": 3.005,
      "theoretical_loss": 3.708255700823293,
      "tokens_seen": 846921728
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004932901456402497,
      "loss": 2.8568,
      "theoretical_loss": 3.7082275493086954,
      "tokens_seen": 846987264
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004932735956118782,
      "loss": 2.3864,
      "theoretical_loss": 3.708199400582101,
      "tokens_seen": 847052800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004932570455835067,
      "loss": 2.667,
      "theoretical_loss": 3.708171254643018,
      "tokens_seen": 847118336
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004932404955551353,
      "loss": 2.5913,
      "theoretical_loss": 3.708143111490955,
      "tokens_seen": 847183872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004932239455267638,
      "loss": 2.6193,
      "theoretical_loss": 3.70811497112542,
      "tokens_seen": 847249408
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004932073954983923,
      "loss": 2.7286,
      "theoretical_loss": 3.708086833545922,
      "tokens_seen": 847314944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004931908454700208,
      "loss": 2.6385,
      "theoretical_loss": 3.70805869875197,
      "tokens_seen": 847380480
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004931742954416493,
      "loss": 2.7979,
      "theoretical_loss": 3.708030566743072,
      "tokens_seen": 847446016
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004931577454132779,
      "loss": 2.7513,
      "theoretical_loss": 3.7080024375187373,
      "tokens_seen": 847511552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004931411953849064,
      "loss": 2.7455,
      "theoretical_loss": 3.7079743110784755,
      "tokens_seen": 847577088
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004931246453565349,
      "loss": 2.7567,
      "theoretical_loss": 3.707946187421795,
      "tokens_seen": 847642624
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004931080953281635,
      "loss": 2.5778,
      "theoretical_loss": 3.707918066548206,
      "tokens_seen": 847708160
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000493091545299792,
      "loss": 2.855,
      "theoretical_loss": 3.707889948457217,
      "tokens_seen": 847773696
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004930749952714205,
      "loss": 2.7118,
      "theoretical_loss": 3.7078618331483386,
      "tokens_seen": 847839232
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000493058445243049,
      "loss": 2.564,
      "theoretical_loss": 3.7078337206210796,
      "tokens_seen": 847904768
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004930418952146775,
      "loss": 2.7341,
      "theoretical_loss": 3.70780561087495,
      "tokens_seen": 847970304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000493025345186306,
      "loss": 2.9213,
      "theoretical_loss": 3.7077775039094605,
      "tokens_seen": 848035840
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004930087951579345,
      "loss": 2.6272,
      "theoretical_loss": 3.70774939972412,
      "tokens_seen": 848101376
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004929922451295631,
      "loss": 2.6819,
      "theoretical_loss": 3.7077212983184404,
      "tokens_seen": 848166912
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004929756951011915,
      "loss": 2.2241,
      "theoretical_loss": 3.70769319969193,
      "tokens_seen": 848232448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004929591450728201,
      "loss": 2.7179,
      "theoretical_loss": 3.707665103844101,
      "tokens_seen": 848297984
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004929425950444486,
      "loss": 2.7999,
      "theoretical_loss": 3.7076370107744627,
      "tokens_seen": 848363520
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004929260450160771,
      "loss": 2.5943,
      "theoretical_loss": 3.7076089204825267,
      "tokens_seen": 848429056
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004929094949877057,
      "loss": 2.8065,
      "theoretical_loss": 3.7075808329678033,
      "tokens_seen": 848494592
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004928929449593342,
      "loss": 2.5905,
      "theoretical_loss": 3.707552748229803,
      "tokens_seen": 848560128
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004928763949309627,
      "loss": 2.2379,
      "theoretical_loss": 3.7075246662680383,
      "tokens_seen": 848625664
    },
    {
      "epoch": 0.3,
      "objective/train/avg_token_score": 0.03921729698777199,
      "objective/train/avg_weight": 0.9538445472717285,
      "objective/train/docs_used": 487462,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1473522186279297,
      "objective/train/std_weight": 0.1519465148448944,
      "objective/train/theoretical_loss": 3.7074965870820193,
      "objective/train/tokens_used": 869151200,
      "theoretical_loss": 3.7074965870820193,
      "tokens_seen": 848691200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004928598449025913,
      "loss": 2.7244,
      "theoretical_loss": 3.7074965870820193,
      "tokens_seen": 848691200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004928432948742197,
      "loss": 2.5644,
      "theoretical_loss": 3.7074685106712577,
      "tokens_seen": 848756736
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004928267448458483,
      "loss": 2.7595,
      "theoretical_loss": 3.7074404370352645,
      "tokens_seen": 848822272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004928101948174768,
      "loss": 2.7505,
      "theoretical_loss": 3.707412366173552,
      "tokens_seen": 848887808
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004927936447891053,
      "loss": 2.8057,
      "theoretical_loss": 3.707384298085631,
      "tokens_seen": 848953344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004927770947607339,
      "loss": 2.5877,
      "theoretical_loss": 3.7073562327710143,
      "tokens_seen": 849018880
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004927605447323624,
      "loss": 2.5553,
      "theoretical_loss": 3.7073281702292133,
      "tokens_seen": 849084416
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004927439947039909,
      "loss": 2.5447,
      "theoretical_loss": 3.70730011045974,
      "tokens_seen": 849149952
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004927274446756194,
      "loss": 2.8678,
      "theoretical_loss": 3.707272053462107,
      "tokens_seen": 849215488
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004927108946472479,
      "loss": 2.8575,
      "theoretical_loss": 3.7072439992358257,
      "tokens_seen": 849281024
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004926943446188765,
      "loss": 2.6313,
      "theoretical_loss": 3.7072159477804094,
      "tokens_seen": 849346560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000492677794590505,
      "loss": 2.6665,
      "theoretical_loss": 3.707187899095371,
      "tokens_seen": 849412096
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004926612445621335,
      "loss": 2.9646,
      "theoretical_loss": 3.707159853180222,
      "tokens_seen": 849477632
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004926446945337621,
      "loss": 2.8392,
      "theoretical_loss": 3.7071318100344754,
      "tokens_seen": 849543168
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004926281445053906,
      "loss": 2.846,
      "theoretical_loss": 3.707103769657645,
      "tokens_seen": 849608704
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000492611594477019,
      "loss": 2.7704,
      "theoretical_loss": 3.707075732049244,
      "tokens_seen": 849674240
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004925950444486476,
      "loss": 2.5915,
      "theoretical_loss": 3.7070476972087842,
      "tokens_seen": 849739776
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004925784944202761,
      "loss": 2.6268,
      "theoretical_loss": 3.7070196651357796,
      "tokens_seen": 849805312
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004925619443919047,
      "loss": 2.3574,
      "theoretical_loss": 3.706991635829744,
      "tokens_seen": 849870848
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004925453943635332,
      "loss": 2.4634,
      "theoretical_loss": 3.70696360929019,
      "tokens_seen": 849936384
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004925288443351617,
      "loss": 2.4958,
      "theoretical_loss": 3.706935585516632,
      "tokens_seen": 850001920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004925122943067903,
      "loss": 2.7725,
      "theoretical_loss": 3.7069075645085845,
      "tokens_seen": 850067456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004924957442784188,
      "loss": 2.6687,
      "theoretical_loss": 3.7068795462655597,
      "tokens_seen": 850132992
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004924791942500472,
      "loss": 2.7816,
      "theoretical_loss": 3.7068515307870733,
      "tokens_seen": 850198528
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004924626442216758,
      "loss": 2.6693,
      "theoretical_loss": 3.706823518072638,
      "tokens_seen": 850264064
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004924460941933043,
      "loss": 2.5309,
      "theoretical_loss": 3.706795508121769,
      "tokens_seen": 850329600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004924295441649329,
      "loss": 2.8646,
      "theoretical_loss": 3.7067675009339807,
      "tokens_seen": 850395136
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004924129941365614,
      "loss": 2.8801,
      "theoretical_loss": 3.7067394965087868,
      "tokens_seen": 850460672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004923964441081899,
      "loss": 2.7284,
      "theoretical_loss": 3.7067114948457025,
      "tokens_seen": 850526208
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004923798940798185,
      "loss": 2.7946,
      "theoretical_loss": 3.706683495944243,
      "tokens_seen": 850591744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000492363344051447,
      "loss": 2.6639,
      "theoretical_loss": 3.7066554998039227,
      "tokens_seen": 850657280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004923467940230754,
      "loss": 2.7585,
      "theoretical_loss": 3.7066275064242564,
      "tokens_seen": 850722816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000492330243994704,
      "loss": 2.7688,
      "theoretical_loss": 3.70659951580476,
      "tokens_seen": 850788352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004923136939663325,
      "loss": 2.6549,
      "theoretical_loss": 3.706571527944948,
      "tokens_seen": 850853888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004922971439379611,
      "loss": 2.8093,
      "theoretical_loss": 3.7065435428443365,
      "tokens_seen": 850919424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004922805939095896,
      "loss": 2.7413,
      "theoretical_loss": 3.70651556050244,
      "tokens_seen": 850984960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004922640438812181,
      "loss": 2.5968,
      "theoretical_loss": 3.706487580918775,
      "tokens_seen": 851050496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004922474938528467,
      "loss": 2.6751,
      "theoretical_loss": 3.7064596040928572,
      "tokens_seen": 851116032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004922309438244751,
      "loss": 2.7339,
      "theoretical_loss": 3.706431630024202,
      "tokens_seen": 851181568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004922143937961036,
      "loss": 2.7456,
      "theoretical_loss": 3.7064036587123255,
      "tokens_seen": 851247104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004921978437677322,
      "loss": 2.7347,
      "theoretical_loss": 3.7063756901567437,
      "tokens_seen": 851312640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004921812937393607,
      "loss": 2.7405,
      "theoretical_loss": 3.7063477243569736,
      "tokens_seen": 851378176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004921647437109893,
      "loss": 2.6353,
      "theoretical_loss": 3.706319761312531,
      "tokens_seen": 851443712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004921481936826178,
      "loss": 2.7894,
      "theoretical_loss": 3.7062918010229327,
      "tokens_seen": 851509248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004921316436542463,
      "loss": 2.8139,
      "theoretical_loss": 3.7062638434876947,
      "tokens_seen": 851574784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004921150936258748,
      "loss": 2.6576,
      "theoretical_loss": 3.706235888706334,
      "tokens_seen": 851640320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004920985435975033,
      "loss": 2.8111,
      "theoretical_loss": 3.706207936678368,
      "tokens_seen": 851705856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004920819935691318,
      "loss": 2.8419,
      "theoretical_loss": 3.7061799874033126,
      "tokens_seen": 851771392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004920654435407604,
      "loss": 2.7311,
      "theoretical_loss": 3.706152040880686,
      "tokens_seen": 851836928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004920488935123888,
      "loss": 2.6535,
      "theoretical_loss": 3.7061240971100045,
      "tokens_seen": 851902464
    },
    {
      "epoch": 0.3,
      "objective/train/avg_token_score": 0.021799368783831596,
      "objective/train/avg_weight": 0.9743759632110596,
      "objective/train/docs_used": 489283,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5381689071655273,
      "objective/train/std_weight": 0.09412091225385666,
      "objective/train/theoretical_loss": 3.7060961560907857,
      "objective/train/tokens_used": 872428000,
      "theoretical_loss": 3.7060961560907857,
      "tokens_seen": 851968000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004920323434840174,
      "loss": 2.6319,
      "theoretical_loss": 3.7060961560907857,
      "tokens_seen": 851968000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004920157934556459,
      "loss": 2.6808,
      "theoretical_loss": 3.7060682178225477,
      "tokens_seen": 852033536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004919992434272744,
      "loss": 2.9656,
      "theoretical_loss": 3.7060402823048073,
      "tokens_seen": 852099072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004919826933989029,
      "loss": 2.8662,
      "theoretical_loss": 3.7060123495370827,
      "tokens_seen": 852164608
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004919661433705314,
      "loss": 2.83,
      "theoretical_loss": 3.705984419518891,
      "tokens_seen": 852230144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00049194959334216,
      "loss": 2.661,
      "theoretical_loss": 3.7059564922497517,
      "tokens_seen": 852295680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004919330433137885,
      "loss": 3.0099,
      "theoretical_loss": 3.7059285677291807,
      "tokens_seen": 852361216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004919164932854171,
      "loss": 2.6567,
      "theoretical_loss": 3.705900645956698,
      "tokens_seen": 852426752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004918999432570456,
      "loss": 2.6544,
      "theoretical_loss": 3.705872726931821,
      "tokens_seen": 852492288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000491883393228674,
      "loss": 2.6236,
      "theoretical_loss": 3.7058448106540682,
      "tokens_seen": 852557824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004918668432003026,
      "loss": 3.0012,
      "theoretical_loss": 3.7058168971229586,
      "tokens_seen": 852623360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004918502931719311,
      "loss": 2.7534,
      "theoretical_loss": 3.7057889863380105,
      "tokens_seen": 852688896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004918337431435596,
      "loss": 2.7611,
      "theoretical_loss": 3.7057610782987433,
      "tokens_seen": 852754432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004918171931151882,
      "loss": 2.72,
      "theoretical_loss": 3.7057331730046745,
      "tokens_seen": 852819968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004918006430868167,
      "loss": 2.8072,
      "theoretical_loss": 3.705705270455325,
      "tokens_seen": 852885504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004917840930584453,
      "loss": 2.869,
      "theoretical_loss": 3.7056773706502124,
      "tokens_seen": 852951040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004917675430300738,
      "loss": 2.6294,
      "theoretical_loss": 3.7056494735888563,
      "tokens_seen": 853016576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004917509930017022,
      "loss": 2.6841,
      "theoretical_loss": 3.705621579270777,
      "tokens_seen": 853082112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004917344429733308,
      "loss": 2.7365,
      "theoretical_loss": 3.705593687695493,
      "tokens_seen": 853147648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004917178929449593,
      "loss": 2.7468,
      "theoretical_loss": 3.7055657988625246,
      "tokens_seen": 853213184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004917013429165878,
      "loss": 2.811,
      "theoretical_loss": 3.705537912771391,
      "tokens_seen": 853278720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004916847928882164,
      "loss": 2.6935,
      "theoretical_loss": 3.7055100294216126,
      "tokens_seen": 853344256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004916682428598449,
      "loss": 2.5893,
      "theoretical_loss": 3.7054821488127097,
      "tokens_seen": 853409792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004916516928314735,
      "loss": 2.751,
      "theoretical_loss": 3.705454270944201,
      "tokens_seen": 853475328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000491635142803102,
      "loss": 2.7137,
      "theoretical_loss": 3.7054263958156084,
      "tokens_seen": 853540864
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004916185927747304,
      "loss": 2.8951,
      "theoretical_loss": 3.7053985234264513,
      "tokens_seen": 853606400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000491602042746359,
      "loss": 2.8064,
      "theoretical_loss": 3.7053706537762503,
      "tokens_seen": 853671936
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004915854927179875,
      "loss": 2.7656,
      "theoretical_loss": 3.7053427868645263,
      "tokens_seen": 853737472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000491568942689616,
      "loss": 2.6514,
      "theoretical_loss": 3.7053149226908,
      "tokens_seen": 853803008
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004915523926612446,
      "loss": 2.6189,
      "theoretical_loss": 3.7052870612545927,
      "tokens_seen": 853868544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0004915358426328731,
      "loss": 2.807,
      "theoretical_loss": 3.705259202555424,
      "tokens_seen": 853934080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004915192926045017,
      "loss": 2.6763,
      "theoretical_loss": 3.705231346592816,
      "tokens_seen": 853999616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004915027425761301,
      "loss": 2.636,
      "theoretical_loss": 3.70520349336629,
      "tokens_seen": 854065152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004914861925477586,
      "loss": 2.873,
      "theoretical_loss": 3.705175642875367,
      "tokens_seen": 854130688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004914696425193872,
      "loss": 2.8499,
      "theoretical_loss": 3.7051477951195677,
      "tokens_seen": 854196224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004914530924910157,
      "loss": 2.7499,
      "theoretical_loss": 3.7051199500984158,
      "tokens_seen": 854261760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004914365424626442,
      "loss": 2.9736,
      "theoretical_loss": 3.705092107811431,
      "tokens_seen": 854327296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004914199924342728,
      "loss": 2.6794,
      "theoretical_loss": 3.705064268258136,
      "tokens_seen": 854392832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004914034424059013,
      "loss": 2.7392,
      "theoretical_loss": 3.705036431438052,
      "tokens_seen": 854458368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004913868923775298,
      "loss": 2.7233,
      "theoretical_loss": 3.7050085973507025,
      "tokens_seen": 854523904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004913703423491583,
      "loss": 3.0411,
      "theoretical_loss": 3.704980765995608,
      "tokens_seen": 854589440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004913537923207868,
      "loss": 2.8957,
      "theoretical_loss": 3.7049529373722923,
      "tokens_seen": 854654976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004913372422924154,
      "loss": 2.5661,
      "theoretical_loss": 3.704925111480277,
      "tokens_seen": 854720512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004913206922640439,
      "loss": 2.7054,
      "theoretical_loss": 3.704897288319084,
      "tokens_seen": 854786048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004913041422356724,
      "loss": 2.762,
      "theoretical_loss": 3.704869467888237,
      "tokens_seen": 854851584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000491287592207301,
      "loss": 2.9285,
      "theoretical_loss": 3.7048416501872588,
      "tokens_seen": 854917120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004912710421789295,
      "loss": 2.6871,
      "theoretical_loss": 3.7048138352156723,
      "tokens_seen": 854982656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000491254492150558,
      "loss": 2.6738,
      "theoretical_loss": 3.7047860229729994,
      "tokens_seen": 855048192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004912379421221865,
      "loss": 2.7855,
      "theoretical_loss": 3.7047582134587644,
      "tokens_seen": 855113728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000491221392093815,
      "loss": 2.6895,
      "theoretical_loss": 3.70473040667249,
      "tokens_seen": 855179264
    },
    {
      "epoch": 0.31,
      "objective/train/avg_token_score": 0.020525820553302765,
      "objective/train/avg_weight": 0.9758756756782532,
      "objective/train/docs_used": 491311,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.380526065826416,
      "objective/train/std_weight": 0.11475884169340134,
      "objective/train/theoretical_loss": 3.7047026026137,
      "objective/train/tokens_used": 875704800,
      "theoretical_loss": 3.7047026026137,
      "tokens_seen": 855244800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004912048420654436,
      "loss": 2.7845,
      "theoretical_loss": 3.7047026026137,
      "tokens_seen": 855244800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004911882920370721,
      "loss": 2.7715,
      "theoretical_loss": 3.7046748012819175,
      "tokens_seen": 855310336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004911717420087006,
      "loss": 2.8311,
      "theoretical_loss": 3.7046470026766665,
      "tokens_seen": 855375872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004911551919803292,
      "loss": 2.77,
      "theoretical_loss": 3.70461920679747,
      "tokens_seen": 855441408
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004911386419519576,
      "loss": 2.6855,
      "theoretical_loss": 3.704591413643853,
      "tokens_seen": 855506944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004911220919235862,
      "loss": 2.7255,
      "theoretical_loss": 3.7045636232153383,
      "tokens_seen": 855572480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004911055418952147,
      "loss": 2.9263,
      "theoretical_loss": 3.704535835511451,
      "tokens_seen": 855638016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004910889918668432,
      "loss": 2.7918,
      "theoretical_loss": 3.7045080505317145,
      "tokens_seen": 855703552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004910724418384717,
      "loss": 2.636,
      "theoretical_loss": 3.7044802682756535,
      "tokens_seen": 855769088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004910558918101002,
      "loss": 2.8199,
      "theoretical_loss": 3.7044524887427928,
      "tokens_seen": 855834624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004910393417817288,
      "loss": 2.8803,
      "theoretical_loss": 3.7044247119326568,
      "tokens_seen": 855900160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004910227917533572,
      "loss": 2.7685,
      "theoretical_loss": 3.7043969378447694,
      "tokens_seen": 855965696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004910062417249857,
      "loss": 2.7999,
      "theoretical_loss": 3.7043691664786564,
      "tokens_seen": 856031232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004909896916966143,
      "loss": 2.6086,
      "theoretical_loss": 3.7043413978338426,
      "tokens_seen": 856096768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004909731416682428,
      "loss": 2.6521,
      "theoretical_loss": 3.7043136319098524,
      "tokens_seen": 856162304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004909565916398714,
      "loss": 2.7773,
      "theoretical_loss": 3.7042858687062115,
      "tokens_seen": 856227840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004909400416114999,
      "loss": 2.6334,
      "theoretical_loss": 3.7042581082224446,
      "tokens_seen": 856293376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004909234915831284,
      "loss": 2.9079,
      "theoretical_loss": 3.7042303504580785,
      "tokens_seen": 856358912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000490906941554757,
      "loss": 2.8248,
      "theoretical_loss": 3.704202595412637,
      "tokens_seen": 856424448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004908903915263854,
      "loss": 2.9021,
      "theoretical_loss": 3.7041748430856467,
      "tokens_seen": 856489984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004908738414980139,
      "loss": 2.6022,
      "theoretical_loss": 3.7041470934766334,
      "tokens_seen": 856555520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004908572914696425,
      "loss": 2.554,
      "theoretical_loss": 3.7041193465851228,
      "tokens_seen": 856621056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000490840741441271,
      "loss": 2.6476,
      "theoretical_loss": 3.704091602410641,
      "tokens_seen": 856686592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004908241914128996,
      "loss": 2.5942,
      "theoretical_loss": 3.704063860952714,
      "tokens_seen": 856752128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004908076413845281,
      "loss": 2.979,
      "theoretical_loss": 3.704036122210868,
      "tokens_seen": 856817664
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004907910913561566,
      "loss": 2.6449,
      "theoretical_loss": 3.704008386184629,
      "tokens_seen": 856883200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004907745413277851,
      "loss": 2.8892,
      "theoretical_loss": 3.7039806528735246,
      "tokens_seen": 856948736
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004907579912994136,
      "loss": 2.8889,
      "theoretical_loss": 3.7039529222770806,
      "tokens_seen": 857014272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004907414412710421,
      "loss": 2.559,
      "theoretical_loss": 3.7039251943948237,
      "tokens_seen": 857079808
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004907248912426707,
      "loss": 2.7763,
      "theoretical_loss": 3.703897469226281,
      "tokens_seen": 857145344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004907083412142992,
      "loss": 2.7244,
      "theoretical_loss": 3.7038697467709794,
      "tokens_seen": 857210880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004906917911859278,
      "loss": 2.7193,
      "theoretical_loss": 3.703842027028445,
      "tokens_seen": 857276416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004906752411575563,
      "loss": 2.9708,
      "theoretical_loss": 3.703814309998207,
      "tokens_seen": 857341952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004906586911291847,
      "loss": 2.7542,
      "theoretical_loss": 3.703786595679791,
      "tokens_seen": 857407488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004906421411008133,
      "loss": 2.7109,
      "theoretical_loss": 3.703758884072725,
      "tokens_seen": 857473024
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004906255910724418,
      "loss": 2.6029,
      "theoretical_loss": 3.7037311751765367,
      "tokens_seen": 857538560
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004906090410440703,
      "loss": 2.692,
      "theoretical_loss": 3.7037034689907538,
      "tokens_seen": 857604096
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004905924910156989,
      "loss": 2.6991,
      "theoretical_loss": 3.7036757655149044,
      "tokens_seen": 857669632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004905759409873274,
      "loss": 2.7758,
      "theoretical_loss": 3.7036480647485153,
      "tokens_seen": 857735168
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000490559390958956,
      "loss": 2.8635,
      "theoretical_loss": 3.7036203666911147,
      "tokens_seen": 857800704
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004905428409305845,
      "loss": 2.6659,
      "theoretical_loss": 3.703592671342232,
      "tokens_seen": 857866240
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004905262909022129,
      "loss": 2.4549,
      "theoretical_loss": 3.703564978701394,
      "tokens_seen": 857931776
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004905097408738415,
      "loss": 2.8273,
      "theoretical_loss": 3.70353728876813,
      "tokens_seen": 857997312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00049049319084547,
      "loss": 2.4757,
      "theoretical_loss": 3.703509601541968,
      "tokens_seen": 858062848
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004904766408170986,
      "loss": 2.4946,
      "theoretical_loss": 3.703481917022437,
      "tokens_seen": 858128384
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004904600907887271,
      "loss": 2.515,
      "theoretical_loss": 3.703454235209066,
      "tokens_seen": 858193920
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004904435407603556,
      "loss": 2.7441,
      "theoretical_loss": 3.7034265561013826,
      "tokens_seen": 858259456
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004904269907319842,
      "loss": 2.6679,
      "theoretical_loss": 3.7033988796989172,
      "tokens_seen": 858324992
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004904104407036126,
      "loss": 2.8357,
      "theoretical_loss": 3.703371206001198,
      "tokens_seen": 858390528
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004903938906752411,
      "loss": 2.628,
      "theoretical_loss": 3.703343535007754,
      "tokens_seen": 858456064
    },
    {
      "epoch": 0.31,
      "objective/train/avg_token_score": 0.02647998556494713,
      "objective/train/avg_weight": 0.9688270092010498,
      "objective/train/docs_used": 493471,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0034778118133545,
      "objective/train/std_weight": 0.11831440031528473,
      "objective/train/theoretical_loss": 3.7033158667181154,
      "objective/train/tokens_used": 878981600,
      "theoretical_loss": 3.7033158667181154,
      "tokens_seen": 858521600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004903773406468697,
      "loss": 2.9928,
      "theoretical_loss": 3.7033158667181154,
      "tokens_seen": 858521600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004903607906184982,
      "loss": 2.9003,
      "theoretical_loss": 3.7032882011318113,
      "tokens_seen": 858587136
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004903442405901268,
      "loss": 2.7984,
      "theoretical_loss": 3.703260538248371,
      "tokens_seen": 858652672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004903276905617553,
      "loss": 2.7086,
      "theoretical_loss": 3.7032328780673245,
      "tokens_seen": 858718208
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004903111405333838,
      "loss": 2.5351,
      "theoretical_loss": 3.7032052205882007,
      "tokens_seen": 858783744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004902945905050124,
      "loss": 2.7171,
      "theoretical_loss": 3.703177565810531,
      "tokens_seen": 858849280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004902780404766408,
      "loss": 3.0313,
      "theoretical_loss": 3.7031499137338444,
      "tokens_seen": 858914816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004902614904482693,
      "loss": 2.8136,
      "theoretical_loss": 3.7031222643576713,
      "tokens_seen": 858980352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004902449404198979,
      "loss": 2.844,
      "theoretical_loss": 3.7030946176815425,
      "tokens_seen": 859045888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004902283903915264,
      "loss": 2.5432,
      "theoretical_loss": 3.703066973704987,
      "tokens_seen": 859111424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000490211840363155,
      "loss": 3.0972,
      "theoretical_loss": 3.7030393324275366,
      "tokens_seen": 859176960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004901952903347835,
      "loss": 2.9112,
      "theoretical_loss": 3.7030116938487208,
      "tokens_seen": 859242496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000490178740306412,
      "loss": 2.8608,
      "theoretical_loss": 3.702984057968072,
      "tokens_seen": 859308032
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004901621902780405,
      "loss": 2.8047,
      "theoretical_loss": 3.702956424785119,
      "tokens_seen": 859373568
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000490145640249669,
      "loss": 2.8492,
      "theoretical_loss": 3.7029287942993943,
      "tokens_seen": 859439104
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004901290902212975,
      "loss": 2.6581,
      "theoretical_loss": 3.7029011665104283,
      "tokens_seen": 859504640
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004901125401929261,
      "loss": 2.5557,
      "theoretical_loss": 3.7028735414177527,
      "tokens_seen": 859570176
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004900959901645545,
      "loss": 2.8574,
      "theoretical_loss": 3.7028459190208984,
      "tokens_seen": 859635712
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004900794401361831,
      "loss": 2.9141,
      "theoretical_loss": 3.7028182993193965,
      "tokens_seen": 859701248
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004900628901078116,
      "loss": 2.7094,
      "theoretical_loss": 3.70279068231278,
      "tokens_seen": 859766784
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004900463400794401,
      "loss": 2.7037,
      "theoretical_loss": 3.7027630680005785,
      "tokens_seen": 859832320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004900297900510686,
      "loss": 2.8009,
      "theoretical_loss": 3.702735456382325,
      "tokens_seen": 859897856
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004900132400226971,
      "loss": 2.7395,
      "theoretical_loss": 3.7027078474575514,
      "tokens_seen": 859963392
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004899966899943257,
      "loss": 2.8492,
      "theoretical_loss": 3.7026802412257895,
      "tokens_seen": 860028928
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004899801399659542,
      "loss": 2.8218,
      "theoretical_loss": 3.7026526376865716,
      "tokens_seen": 860094464
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004899635899375827,
      "loss": 2.7228,
      "theoretical_loss": 3.7026250368394296,
      "tokens_seen": 860160000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004899470399092113,
      "loss": 2.5098,
      "theoretical_loss": 3.7025974386838962,
      "tokens_seen": 860225536
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004899304898808397,
      "loss": 2.9443,
      "theoretical_loss": 3.7025698432195044,
      "tokens_seen": 860291072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004899139398524683,
      "loss": 2.9958,
      "theoretical_loss": 3.7025422504457857,
      "tokens_seen": 860356608
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004898973898240968,
      "loss": 2.8402,
      "theoretical_loss": 3.702514660362273,
      "tokens_seen": 860422144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004898808397957253,
      "loss": 2.402,
      "theoretical_loss": 3.7024870729684998,
      "tokens_seen": 860487680
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004898642897673539,
      "loss": 2.8217,
      "theoretical_loss": 3.7024594882639987,
      "tokens_seen": 860553216
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004898477397389824,
      "loss": 2.528,
      "theoretical_loss": 3.702431906248303,
      "tokens_seen": 860618752
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004898311897106109,
      "loss": 3.0404,
      "theoretical_loss": 3.702404326920945,
      "tokens_seen": 860684288
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004898146396822395,
      "loss": 2.9598,
      "theoretical_loss": 3.7023767502814593,
      "tokens_seen": 860749824
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004897980896538679,
      "loss": 2.8515,
      "theoretical_loss": 3.7023491763293785,
      "tokens_seen": 860815360
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004897815396254965,
      "loss": 2.7587,
      "theoretical_loss": 3.7023216050642365,
      "tokens_seen": 860880896
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000489764989597125,
      "loss": 2.8292,
      "theoretical_loss": 3.7022940364855668,
      "tokens_seen": 860946432
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004897484395687535,
      "loss": 2.9599,
      "theoretical_loss": 3.702266470592903,
      "tokens_seen": 861011968
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004897318895403821,
      "loss": 2.806,
      "theoretical_loss": 3.7022389073857793,
      "tokens_seen": 861077504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004897153395120106,
      "loss": 2.6446,
      "theoretical_loss": 3.70221134686373,
      "tokens_seen": 861143040
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000489698789483639,
      "loss": 2.6766,
      "theoretical_loss": 3.7021837890262876,
      "tokens_seen": 861208576
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004896822394552676,
      "loss": 2.6665,
      "theoretical_loss": 3.7021562338729885,
      "tokens_seen": 861274112
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004896656894268961,
      "loss": 2.7678,
      "theoretical_loss": 3.702128681403366,
      "tokens_seen": 861339648
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004896491393985247,
      "loss": 2.9519,
      "theoretical_loss": 3.702101131616954,
      "tokens_seen": 861405184
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004896325893701532,
      "loss": 2.8097,
      "theoretical_loss": 3.702073584513288,
      "tokens_seen": 861470720
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004896160393417817,
      "loss": 2.7398,
      "theoretical_loss": 3.7020460400919024,
      "tokens_seen": 861536256
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004895994893134103,
      "loss": 2.7986,
      "theoretical_loss": 3.702018498352332,
      "tokens_seen": 861601792
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004895829392850388,
      "loss": 2.9578,
      "theoretical_loss": 3.701990959294112,
      "tokens_seen": 861667328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004895663892566672,
      "loss": 2.631,
      "theoretical_loss": 3.7019634229167773,
      "tokens_seen": 861732864
    },
    {
      "epoch": 0.31,
      "objective/train/avg_token_score": 0.013004355132579803,
      "objective/train/avg_weight": 0.984716534614563,
      "objective/train/docs_used": 495231,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9183006286621094,
      "objective/train/std_weight": 0.09929699450731277,
      "objective/train/theoretical_loss": 3.701935889219863,
      "objective/train/tokens_used": 882258400,
      "theoretical_loss": 3.701935889219863,
      "tokens_seen": 861798400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004895498392282958,
      "loss": 2.8066,
      "theoretical_loss": 3.701935889219863,
      "tokens_seen": 861798400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004895332891999243,
      "loss": 2.8801,
      "theoretical_loss": 3.7019083582029038,
      "tokens_seen": 861863936
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004895167391715529,
      "loss": 2.7393,
      "theoretical_loss": 3.701880829865436,
      "tokens_seen": 861929472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004895001891431814,
      "loss": 2.7791,
      "theoretical_loss": 3.7018533042069945,
      "tokens_seen": 861995008
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004894836391148099,
      "loss": 2.7638,
      "theoretical_loss": 3.7018257812271154,
      "tokens_seen": 862060544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004894670890864385,
      "loss": 2.9223,
      "theoretical_loss": 3.701798260925335,
      "tokens_seen": 862126080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000489450539058067,
      "loss": 2.9573,
      "theoretical_loss": 3.7017707433011875,
      "tokens_seen": 862191616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004894339890296954,
      "loss": 2.5867,
      "theoretical_loss": 3.7017432283542098,
      "tokens_seen": 862257152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000489417439001324,
      "loss": 2.7885,
      "theoretical_loss": 3.701715716083938,
      "tokens_seen": 862322688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004894008889729525,
      "loss": 2.6392,
      "theoretical_loss": 3.7016882064899086,
      "tokens_seen": 862388224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004893843389445811,
      "loss": 2.977,
      "theoretical_loss": 3.701660699571657,
      "tokens_seen": 862453760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004893677889162096,
      "loss": 2.6847,
      "theoretical_loss": 3.701633195328721,
      "tokens_seen": 862519296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004893512388878381,
      "loss": 2.7706,
      "theoretical_loss": 3.701605693760636,
      "tokens_seen": 862584832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004893346888594667,
      "loss": 2.5958,
      "theoretical_loss": 3.701578194866939,
      "tokens_seen": 862650368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004893181388310951,
      "loss": 2.6782,
      "theoretical_loss": 3.701550698647167,
      "tokens_seen": 862715904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004893015888027236,
      "loss": 2.6947,
      "theoretical_loss": 3.701523205100856,
      "tokens_seen": 862781440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004892850387743522,
      "loss": 2.765,
      "theoretical_loss": 3.7014957142275446,
      "tokens_seen": 862846976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004892684887459807,
      "loss": 2.7838,
      "theoretical_loss": 3.701468226026769,
      "tokens_seen": 862912512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004892519387176093,
      "loss": 2.7695,
      "theoretical_loss": 3.7014407404980663,
      "tokens_seen": 862978048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004892353886892378,
      "loss": 2.808,
      "theoretical_loss": 3.7014132576409744,
      "tokens_seen": 863043584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004892188386608663,
      "loss": 2.865,
      "theoretical_loss": 3.7013857774550303,
      "tokens_seen": 863109120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004892022886324949,
      "loss": 2.7274,
      "theoretical_loss": 3.7013582999397716,
      "tokens_seen": 863174656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004891857386041233,
      "loss": 2.8007,
      "theoretical_loss": 3.701330825094736,
      "tokens_seen": 863240192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004891691885757518,
      "loss": 2.908,
      "theoretical_loss": 3.701303352919462,
      "tokens_seen": 863305728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004891526385473804,
      "loss": 2.8452,
      "theoretical_loss": 3.701275883413487,
      "tokens_seen": 863371264
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004891360885190089,
      "loss": 2.9882,
      "theoretical_loss": 3.7012484165763486,
      "tokens_seen": 863436800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004891195384906375,
      "loss": 2.6693,
      "theoretical_loss": 3.701220952407586,
      "tokens_seen": 863502336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004891029884622659,
      "loss": 2.9218,
      "theoretical_loss": 3.7011934909067365,
      "tokens_seen": 863567872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004890864384338945,
      "loss": 2.7948,
      "theoretical_loss": 3.7011660320733393,
      "tokens_seen": 863633408
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004890698884055229,
      "loss": 2.6497,
      "theoretical_loss": 3.701138575906932,
      "tokens_seen": 863698944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004890533383771514,
      "loss": 2.6809,
      "theoretical_loss": 3.7011111224070543,
      "tokens_seen": 863764480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048903678834878,
      "loss": 2.6266,
      "theoretical_loss": 3.701083671573244,
      "tokens_seen": 863830016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004890202383204085,
      "loss": 2.7788,
      "theoretical_loss": 3.7010562234050406,
      "tokens_seen": 863895552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004890036882920371,
      "loss": 2.6324,
      "theoretical_loss": 3.701028777901983,
      "tokens_seen": 863961088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004889871382636656,
      "loss": 2.6538,
      "theoretical_loss": 3.7010013350636095,
      "tokens_seen": 864026624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004889705882352941,
      "loss": 2.847,
      "theoretical_loss": 3.7009738948894597,
      "tokens_seen": 864092160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004889540382069226,
      "loss": 2.7601,
      "theoretical_loss": 3.7009464573790742,
      "tokens_seen": 864157696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004889374881785511,
      "loss": 2.7575,
      "theoretical_loss": 3.7009190225319903,
      "tokens_seen": 864223232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004889209381501796,
      "loss": 2.8254,
      "theoretical_loss": 3.7008915903477497,
      "tokens_seen": 864288768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004889043881218082,
      "loss": 2.6699,
      "theoretical_loss": 3.70086416082589,
      "tokens_seen": 864354304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004888878380934367,
      "loss": 2.6959,
      "theoretical_loss": 3.7008367339659523,
      "tokens_seen": 864419840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004888712880650653,
      "loss": 2.7475,
      "theoretical_loss": 3.700809309767476,
      "tokens_seen": 864485376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004888547380366938,
      "loss": 2.8485,
      "theoretical_loss": 3.7007818882300016,
      "tokens_seen": 864550912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004888381880083222,
      "loss": 2.5393,
      "theoretical_loss": 3.7007544693530683,
      "tokens_seen": 864616448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004888216379799508,
      "loss": 2.7558,
      "theoretical_loss": 3.7007270531362173,
      "tokens_seen": 864681984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004888050879515793,
      "loss": 2.5742,
      "theoretical_loss": 3.700699639578988,
      "tokens_seen": 864747520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004887885379232078,
      "loss": 2.7102,
      "theoretical_loss": 3.700672228680922,
      "tokens_seen": 864813056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004887719878948364,
      "loss": 2.7405,
      "theoretical_loss": 3.700644820441559,
      "tokens_seen": 864878592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004887554378664649,
      "loss": 2.7208,
      "theoretical_loss": 3.70061741486044,
      "tokens_seen": 864944128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004887388878380935,
      "loss": 2.6465,
      "theoretical_loss": 3.7005900119371056,
      "tokens_seen": 865009664
    },
    {
      "epoch": 0.31,
      "objective/train/avg_token_score": 0.031225278973579407,
      "objective/train/avg_weight": 0.9632714986801147,
      "objective/train/docs_used": 497313,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.393155574798584,
      "objective/train/std_weight": 0.1505284607410431,
      "objective/train/theoretical_loss": 3.7005626116710966,
      "objective/train/tokens_used": 885535200,
      "theoretical_loss": 3.7005626116710966,
      "tokens_seen": 865075200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000488722337809722,
      "loss": 2.5888,
      "theoretical_loss": 3.7005626116710966,
      "tokens_seen": 865075200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004887057877813504,
      "loss": 2.7796,
      "theoretical_loss": 3.700535214061955,
      "tokens_seen": 865140736
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000488689237752979,
      "loss": 2.8417,
      "theoretical_loss": 3.7005078191092204,
      "tokens_seen": 865206272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004886726877246075,
      "loss": 2.8062,
      "theoretical_loss": 3.7004804268124354,
      "tokens_seen": 865271808
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000488656137696236,
      "loss": 2.7582,
      "theoretical_loss": 3.700453037171141,
      "tokens_seen": 865337344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004886395876678646,
      "loss": 2.7286,
      "theoretical_loss": 3.700425650184878,
      "tokens_seen": 865402880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004886230376394931,
      "loss": 3.0561,
      "theoretical_loss": 3.7003982658531887,
      "tokens_seen": 865468416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004886064876111217,
      "loss": 2.9019,
      "theoretical_loss": 3.700370884175615,
      "tokens_seen": 865533952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004885899375827501,
      "loss": 2.8415,
      "theoretical_loss": 3.7003435051516975,
      "tokens_seen": 865599488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004885733875543786,
      "loss": 2.5464,
      "theoretical_loss": 3.7003161287809796,
      "tokens_seen": 865665024
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004885568375260072,
      "loss": 2.7972,
      "theoretical_loss": 3.700288755063003,
      "tokens_seen": 865730560
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004885402874976357,
      "loss": 2.686,
      "theoretical_loss": 3.7002613839973097,
      "tokens_seen": 865796096
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004885237374692642,
      "loss": 2.8898,
      "theoretical_loss": 3.7002340155834412,
      "tokens_seen": 865861632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004885071874408928,
      "loss": 2.8422,
      "theoretical_loss": 3.7002066498209416,
      "tokens_seen": 865927168
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004884906374125213,
      "loss": 2.8041,
      "theoretical_loss": 3.7001792867093517,
      "tokens_seen": 865992704
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004884740873841499,
      "loss": 2.6818,
      "theoretical_loss": 3.700151926248215,
      "tokens_seen": 866058240
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004884575373557783,
      "loss": 2.5525,
      "theoretical_loss": 3.700124568437074,
      "tokens_seen": 866123776
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004884409873274068,
      "loss": 2.437,
      "theoretical_loss": 3.7000972132754715,
      "tokens_seen": 866189312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004884244372990354,
      "loss": 2.7448,
      "theoretical_loss": 3.7000698607629507,
      "tokens_seen": 866254848
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004884078872706639,
      "loss": 2.6629,
      "theoretical_loss": 3.7000425108990553,
      "tokens_seen": 866320384
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004883913372422924,
      "loss": 2.6699,
      "theoretical_loss": 3.700015163683327,
      "tokens_seen": 866385920
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000488374787213921,
      "loss": 2.888,
      "theoretical_loss": 3.69998781911531,
      "tokens_seen": 866451456
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004883582371855495,
      "loss": 2.972,
      "theoretical_loss": 3.6999604771945473,
      "tokens_seen": 866516992
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000488341687157178,
      "loss": 2.6551,
      "theoretical_loss": 3.699933137920583,
      "tokens_seen": 866582528
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004883251371288065,
      "loss": 2.4513,
      "theoretical_loss": 3.6999058012929606,
      "tokens_seen": 866648064
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000488308587100435,
      "loss": 2.8272,
      "theoretical_loss": 3.6998784673112235,
      "tokens_seen": 866713600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004882920370720636,
      "loss": 2.7136,
      "theoretical_loss": 3.6998511359749164,
      "tokens_seen": 866779136
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004882754870436921,
      "loss": 2.6931,
      "theoretical_loss": 3.6998238072835816,
      "tokens_seen": 866844672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004882589370153206,
      "loss": 2.905,
      "theoretical_loss": 3.699796481236765,
      "tokens_seen": 866910208
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048824238698694916,
      "loss": 2.7645,
      "theoretical_loss": 3.6997691578340097,
      "tokens_seen": 866975744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048822583695857765,
      "loss": 2.9577,
      "theoretical_loss": 3.6997418370748605,
      "tokens_seen": 867041280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004882092869302062,
      "loss": 2.6583,
      "theoretical_loss": 3.699714518958862,
      "tokens_seen": 867106816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004881927369018347,
      "loss": 2.5775,
      "theoretical_loss": 3.699687203485558,
      "tokens_seen": 867172352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048817618687346326,
      "loss": 2.8316,
      "theoretical_loss": 3.699659890654494,
      "tokens_seen": 867237888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048815963684509174,
      "loss": 2.8299,
      "theoretical_loss": 3.699632580465215,
      "tokens_seen": 867303424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004881430868167203,
      "loss": 2.9303,
      "theoretical_loss": 3.699605272917264,
      "tokens_seen": 867368960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004881265367883487,
      "loss": 2.811,
      "theoretical_loss": 3.699577968010188,
      "tokens_seen": 867434496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048810998675997725,
      "loss": 2.8544,
      "theoretical_loss": 3.6995506657435313,
      "tokens_seen": 867500032
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004880934367316058,
      "loss": 2.2972,
      "theoretical_loss": 3.699523366116839,
      "tokens_seen": 867565568
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004880768867032343,
      "loss": 2.7413,
      "theoretical_loss": 3.6994960691296574,
      "tokens_seen": 867631104
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004880603366748628,
      "loss": 2.8607,
      "theoretical_loss": 3.6994687747815305,
      "tokens_seen": 867696640
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048804378664649134,
      "loss": 2.961,
      "theoretical_loss": 3.699441483072005,
      "tokens_seen": 867762176
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004880272366181199,
      "loss": 2.8279,
      "theoretical_loss": 3.6994141940006253,
      "tokens_seen": 867827712
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004880106865897484,
      "loss": 3.0012,
      "theoretical_loss": 3.6993869075669386,
      "tokens_seen": 867893248
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048799413656137696,
      "loss": 2.9785,
      "theoretical_loss": 3.6993596237704907,
      "tokens_seen": 867958784
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048797758653300544,
      "loss": 2.3966,
      "theoretical_loss": 3.699332342610827,
      "tokens_seen": 868024320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000487961036504634,
      "loss": 2.658,
      "theoretical_loss": 3.6993050640874934,
      "tokens_seen": 868089856
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004879444864762625,
      "loss": 2.9497,
      "theoretical_loss": 3.6992777882000363,
      "tokens_seen": 868155392
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048792793644789105,
      "loss": 2.7123,
      "theoretical_loss": 3.6992505149480026,
      "tokens_seen": 868220928
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048791138641951954,
      "loss": 2.6653,
      "theoretical_loss": 3.6992232443309385,
      "tokens_seen": 868286464
    },
    {
      "epoch": 0.31,
      "objective/train/avg_token_score": 0.014921732246875763,
      "objective/train/avg_weight": 0.982434868812561,
      "objective/train/docs_used": 499341,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6704978942871094,
      "objective/train/std_weight": 0.09517405927181244,
      "objective/train/theoretical_loss": 3.69919597634839,
      "objective/train/tokens_used": 888812000,
      "theoretical_loss": 3.69919597634839,
      "tokens_seen": 868352000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004878948363911481,
      "loss": 2.8714,
      "theoretical_loss": 3.69919597634839,
      "tokens_seen": 868352000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004878782863627766,
      "loss": 2.7111,
      "theoretical_loss": 3.699168710999905,
      "tokens_seen": 868417536
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048786173633440515,
      "loss": 2.8015,
      "theoretical_loss": 3.6991414482850296,
      "tokens_seen": 868483072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048784518630603363,
      "loss": 2.6969,
      "theoretical_loss": 3.69911418820331,
      "tokens_seen": 868548608
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048782863627766217,
      "loss": 2.7754,
      "theoretical_loss": 3.6990869307542944,
      "tokens_seen": 868614144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004878120862492907,
      "loss": 2.7515,
      "theoretical_loss": 3.699059675937529,
      "tokens_seen": 868679680
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048779553622091924,
      "loss": 2.4078,
      "theoretical_loss": 3.6990324237525622,
      "tokens_seen": 868745216
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048777898619254773,
      "loss": 2.7039,
      "theoretical_loss": 3.6990051741989403,
      "tokens_seen": 868810752
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048776243616417627,
      "loss": 2.7762,
      "theoretical_loss": 3.698977927276211,
      "tokens_seen": 868876288
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004877458861358048,
      "loss": 2.7068,
      "theoretical_loss": 3.6989506829839227,
      "tokens_seen": 868941824
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048772933610743334,
      "loss": 2.7081,
      "theoretical_loss": 3.6989234413216217,
      "tokens_seen": 869007360
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004877127860790618,
      "loss": 2.8082,
      "theoretical_loss": 3.6988962022888563,
      "tokens_seen": 869072896
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048769623605069036,
      "loss": 2.624,
      "theoretical_loss": 3.698868965885175,
      "tokens_seen": 869138432
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004876796860223189,
      "loss": 2.6411,
      "theoretical_loss": 3.6988417321101252,
      "tokens_seen": 869203968
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048766313599394744,
      "loss": 2.5375,
      "theoretical_loss": 3.6988145009632563,
      "tokens_seen": 869269504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004876465859655759,
      "loss": 2.5159,
      "theoretical_loss": 3.6987872724441146,
      "tokens_seen": 869335040
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048763003593720446,
      "loss": 2.71,
      "theoretical_loss": 3.698760046552249,
      "tokens_seen": 869400576
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000487613485908833,
      "loss": 2.7732,
      "theoretical_loss": 3.6987328232872096,
      "tokens_seen": 869466112
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048759693588046153,
      "loss": 3.018,
      "theoretical_loss": 3.698705602648543,
      "tokens_seen": 869531648
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048758038585209,
      "loss": 2.6672,
      "theoretical_loss": 3.698678384635799,
      "tokens_seen": 869597184
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048756383582371855,
      "loss": 2.7,
      "theoretical_loss": 3.698651169248526,
      "tokens_seen": 869662720
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004875472857953471,
      "loss": 2.5643,
      "theoretical_loss": 3.698623956486273,
      "tokens_seen": 869728256
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048753073576697563,
      "loss": 2.641,
      "theoretical_loss": 3.6985967463485894,
      "tokens_seen": 869793792
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004875141857386041,
      "loss": 2.7476,
      "theoretical_loss": 3.6985695388350233,
      "tokens_seen": 869859328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048749763571023265,
      "loss": 2.7916,
      "theoretical_loss": 3.698542333945125,
      "tokens_seen": 869924864
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004874810856818612,
      "loss": 2.4193,
      "theoretical_loss": 3.6985151316784437,
      "tokens_seen": 869990400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004874645356534897,
      "loss": 2.64,
      "theoretical_loss": 3.6984879320345287,
      "tokens_seen": 870055936
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004874479856251182,
      "loss": 2.8772,
      "theoretical_loss": 3.6984607350129295,
      "tokens_seen": 870121472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048743143559674675,
      "loss": 2.7647,
      "theoretical_loss": 3.6984335406131956,
      "tokens_seen": 870187008
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048743143559674675,
      "loss": 4.3025,
      "theoretical_loss": 3.6984063488348777,
      "tokens_seen": 870252544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004874148855683753,
      "loss": 3.0256,
      "theoretical_loss": 3.698379159677525,
      "tokens_seen": 870318080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004873983355400038,
      "loss": 2.6322,
      "theoretical_loss": 3.6983519731406878,
      "tokens_seen": 870383616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004873817855116323,
      "loss": 2.6202,
      "theoretical_loss": 3.6983247892239155,
      "tokens_seen": 870449152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048736523548326084,
      "loss": 2.8699,
      "theoretical_loss": 3.6982976079267598,
      "tokens_seen": 870514688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004873486854548894,
      "loss": 2.7488,
      "theoretical_loss": 3.69827042924877,
      "tokens_seen": 870580224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004873321354265179,
      "loss": 2.7565,
      "theoretical_loss": 3.6982432531894966,
      "tokens_seen": 870645760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004873155853981464,
      "loss": 2.7434,
      "theoretical_loss": 3.69821607974849,
      "tokens_seen": 870711296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048729903536977494,
      "loss": 2.6767,
      "theoretical_loss": 3.698188908925302,
      "tokens_seen": 870776832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004872824853414035,
      "loss": 2.9729,
      "theoretical_loss": 3.698161740719483,
      "tokens_seen": 870842368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000487265935313032,
      "loss": 2.5398,
      "theoretical_loss": 3.6981345751305827,
      "tokens_seen": 870907904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004872493852846605,
      "loss": 2.7119,
      "theoretical_loss": 3.698107412158154,
      "tokens_seen": 870973440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048723283525628903,
      "loss": 2.7278,
      "theoretical_loss": 3.6980802518017466,
      "tokens_seen": 871038976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048721628522791757,
      "loss": 2.7468,
      "theoretical_loss": 3.6980530940609126,
      "tokens_seen": 871104512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004871997351995461,
      "loss": 2.7822,
      "theoretical_loss": 3.6980259389352033,
      "tokens_seen": 871170048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004871831851711746,
      "loss": 2.7782,
      "theoretical_loss": 3.6979987864241695,
      "tokens_seen": 871235584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048716663514280313,
      "loss": 3.2244,
      "theoretical_loss": 3.697971636527364,
      "tokens_seen": 871301120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048715008511443167,
      "loss": 2.6083,
      "theoretical_loss": 3.697944489244337,
      "tokens_seen": 871366656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004871335350860601,
      "loss": 2.8983,
      "theoretical_loss": 3.697917344574641,
      "tokens_seen": 871432192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048711698505768863,
      "loss": 2.7645,
      "theoretical_loss": 3.6978902025178284,
      "tokens_seen": 871497728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048710043502931717,
      "loss": 3.0709,
      "theoretical_loss": 3.697863063073451,
      "tokens_seen": 871563264
    },
    {
      "epoch": 0.31,
      "objective/train/avg_token_score": 0.010347530245780945,
      "objective/train/avg_weight": 0.9878193736076355,
      "objective/train/docs_used": 501132,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0692243576049805,
      "objective/train/std_weight": 0.04914121702313423,
      "objective/train/theoretical_loss": 3.6978359262410603,
      "objective/train/tokens_used": 892088800,
      "theoretical_loss": 3.6978359262410603,
      "tokens_seen": 871628800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048708388500094565,
      "loss": 2.8042,
      "theoretical_loss": 3.6978359262410603,
      "tokens_seen": 871628800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004870673349725742,
      "loss": 2.8413,
      "theoretical_loss": 3.6978087920202087,
      "tokens_seen": 871694336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048705078494420273,
      "loss": 2.9681,
      "theoretical_loss": 3.6977816604104494,
      "tokens_seen": 871759872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048703423491583127,
      "loss": 2.7102,
      "theoretical_loss": 3.6977545314113343,
      "tokens_seen": 871825408
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048701768488745975,
      "loss": 2.8144,
      "theoretical_loss": 3.697727405022416,
      "tokens_seen": 871890944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004870011348590883,
      "loss": 2.9153,
      "theoretical_loss": 3.6977002812432476,
      "tokens_seen": 871956480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004869845848307168,
      "loss": 2.7258,
      "theoretical_loss": 3.6976731600733808,
      "tokens_seen": 872022016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048696803480234536,
      "loss": 2.7422,
      "theoretical_loss": 3.6976460415123693,
      "tokens_seen": 872087552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048695148477397385,
      "loss": 2.7853,
      "theoretical_loss": 3.6976189255597665,
      "tokens_seen": 872153088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004869349347456024,
      "loss": 2.6802,
      "theoretical_loss": 3.697591812215125,
      "tokens_seen": 872218624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004869183847172309,
      "loss": 2.7113,
      "theoretical_loss": 3.697564701477998,
      "tokens_seen": 872284160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048690183468885946,
      "loss": 2.973,
      "theoretical_loss": 3.6975375933479393,
      "tokens_seen": 872349696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048688528466048794,
      "loss": 2.8125,
      "theoretical_loss": 3.6975104878245015,
      "tokens_seen": 872415232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004868687346321165,
      "loss": 2.8097,
      "theoretical_loss": 3.6974833849072395,
      "tokens_seen": 872480768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000486852184603745,
      "loss": 2.8603,
      "theoretical_loss": 3.6974562845957055,
      "tokens_seen": 872546304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048683563457537355,
      "loss": 2.8049,
      "theoretical_loss": 3.697429186889454,
      "tokens_seen": 872611840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048681908454700204,
      "loss": 3.0393,
      "theoretical_loss": 3.6974020917880397,
      "tokens_seen": 872677376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004868025345186306,
      "loss": 2.8471,
      "theoretical_loss": 3.697374999291015,
      "tokens_seen": 872742912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004867859844902591,
      "loss": 2.5484,
      "theoretical_loss": 3.6973479093979353,
      "tokens_seen": 872808448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048676943446188765,
      "loss": 2.6116,
      "theoretical_loss": 3.697320822108354,
      "tokens_seen": 872873984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048675288443351613,
      "loss": 2.9336,
      "theoretical_loss": 3.697293737421826,
      "tokens_seen": 872939520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048673633440514467,
      "loss": 2.3106,
      "theoretical_loss": 3.697266655337906,
      "tokens_seen": 873005056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004867197843767732,
      "loss": 2.7571,
      "theoretical_loss": 3.6972395758561483,
      "tokens_seen": 873070592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048670323434840175,
      "loss": 2.6607,
      "theoretical_loss": 3.6972124989761066,
      "tokens_seen": 873136128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048668668432003023,
      "loss": 2.6938,
      "theoretical_loss": 3.697185424697337,
      "tokens_seen": 873201664
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048667013429165877,
      "loss": 2.6612,
      "theoretical_loss": 3.6971583530193937,
      "tokens_seen": 873267200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004866535842632873,
      "loss": 2.7366,
      "theoretical_loss": 3.697131283941832,
      "tokens_seen": 873332736
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048663703423491584,
      "loss": 2.8717,
      "theoretical_loss": 3.697104217464207,
      "tokens_seen": 873398272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004866204842065443,
      "loss": 2.6558,
      "theoretical_loss": 3.697077153586074,
      "tokens_seen": 873463808
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048660393417817286,
      "loss": 2.6166,
      "theoretical_loss": 3.697050092306988,
      "tokens_seen": 873529344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004865873841498014,
      "loss": 2.6654,
      "theoretical_loss": 3.697023033626504,
      "tokens_seen": 873594880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048657083412142994,
      "loss": 2.5243,
      "theoretical_loss": 3.696995977544179,
      "tokens_seen": 873660416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004865542840930584,
      "loss": 2.7239,
      "theoretical_loss": 3.6969689240595676,
      "tokens_seen": 873725952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048653773406468696,
      "loss": 2.7731,
      "theoretical_loss": 3.6969418731722254,
      "tokens_seen": 873791488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004865211840363155,
      "loss": 2.6803,
      "theoretical_loss": 3.696914824881709,
      "tokens_seen": 873857024
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048650463400794403,
      "loss": 2.61,
      "theoretical_loss": 3.6968877791875743,
      "tokens_seen": 873922560
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048648808397957257,
      "loss": 2.6686,
      "theoretical_loss": 3.6968607360893766,
      "tokens_seen": 873988096
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048647153395120106,
      "loss": 2.7905,
      "theoretical_loss": 3.6968336955866725,
      "tokens_seen": 874053632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004864549839228296,
      "loss": 3.1713,
      "theoretical_loss": 3.6968066576790193,
      "tokens_seen": 874119168
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048643843389445813,
      "loss": 2.7661,
      "theoretical_loss": 3.6967796223659724,
      "tokens_seen": 874184704
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048642188386608667,
      "loss": 2.3528,
      "theoretical_loss": 3.6967525896470876,
      "tokens_seen": 874250240
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048640533383771515,
      "loss": 2.8917,
      "theoretical_loss": 3.6967255595219233,
      "tokens_seen": 874315776
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004863887838093437,
      "loss": 2.5933,
      "theoretical_loss": 3.6966985319900356,
      "tokens_seen": 874381312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004863722337809722,
      "loss": 2.8976,
      "theoretical_loss": 3.6966715070509806,
      "tokens_seen": 874446848
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048635568375260076,
      "loss": 2.5647,
      "theoretical_loss": 3.696644484704316,
      "tokens_seen": 874512384
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048633913372422925,
      "loss": 2.7753,
      "theoretical_loss": 3.6966174649495986,
      "tokens_seen": 874577920
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004863225836958578,
      "loss": 2.4786,
      "theoretical_loss": 3.696590447786386,
      "tokens_seen": 874643456
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004863060336674863,
      "loss": 2.8333,
      "theoretical_loss": 3.6965634332142345,
      "tokens_seen": 874708992
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048628948363911486,
      "loss": 2.7207,
      "theoretical_loss": 3.696536421232703,
      "tokens_seen": 874774528
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048627293361074334,
      "loss": 3.0541,
      "theoretical_loss": 3.696509411841348,
      "tokens_seen": 874840064
    },
    {
      "epoch": 0.31,
      "objective/train/avg_token_score": 0.01772496849298477,
      "objective/train/avg_weight": 0.979145884513855,
      "objective/train/docs_used": 502873,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1205387115478516,
      "objective/train/std_weight": 0.08095329254865646,
      "objective/train/theoretical_loss": 3.6964824050397276,
      "objective/train/tokens_used": 895365600,
      "theoretical_loss": 3.6964824050397276,
      "tokens_seen": 874905600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004862563835823719,
      "loss": 2.8256,
      "theoretical_loss": 3.6964824050397276,
      "tokens_seen": 874905600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004862398335540004,
      "loss": 2.8786,
      "theoretical_loss": 3.696455400827398,
      "tokens_seen": 874971136
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048622328352562896,
      "loss": 2.8355,
      "theoretical_loss": 3.6964283992039197,
      "tokens_seen": 875036672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048620673349725744,
      "loss": 2.6729,
      "theoretical_loss": 3.6964014001688486,
      "tokens_seen": 875102208
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000486190183468886,
      "loss": 2.72,
      "theoretical_loss": 3.6963744037217436,
      "tokens_seen": 875167744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004861736334405145,
      "loss": 2.7877,
      "theoretical_loss": 3.6963474098621623,
      "tokens_seen": 875233280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048615708341214294,
      "loss": 2.5067,
      "theoretical_loss": 3.6963204185896643,
      "tokens_seen": 875298816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004861405333837715,
      "loss": 2.617,
      "theoretical_loss": 3.6962934299038057,
      "tokens_seen": 875364352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004861239833554,
      "loss": 3.0218,
      "theoretical_loss": 3.6962664438041473,
      "tokens_seen": 875429888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004861074333270285,
      "loss": 2.6476,
      "theoretical_loss": 3.696239460290246,
      "tokens_seen": 875495424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048609088329865704,
      "loss": 2.5488,
      "theoretical_loss": 3.6962124793616615,
      "tokens_seen": 875560960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004860743332702856,
      "loss": 2.637,
      "theoretical_loss": 3.6961855010179527,
      "tokens_seen": 875626496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004860577832419141,
      "loss": 2.5344,
      "theoretical_loss": 3.696158525258678,
      "tokens_seen": 875692032
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004860412332135426,
      "loss": 2.7626,
      "theoretical_loss": 3.696131552083396,
      "tokens_seen": 875757568
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048602468318517114,
      "loss": 2.6654,
      "theoretical_loss": 3.696104581491667,
      "tokens_seen": 875823104
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048600813315679967,
      "loss": 2.5445,
      "theoretical_loss": 3.696077613483049,
      "tokens_seen": 875888640
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004859915831284282,
      "loss": 2.6359,
      "theoretical_loss": 3.6960506480571027,
      "tokens_seen": 875954176
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004859750331000567,
      "loss": 2.9001,
      "theoretical_loss": 3.6960236852133868,
      "tokens_seen": 876019712
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048595848307168523,
      "loss": 2.6224,
      "theoretical_loss": 3.6959967249514607,
      "tokens_seen": 876085248
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048594193304331377,
      "loss": 2.5736,
      "theoretical_loss": 3.6959697672708844,
      "tokens_seen": 876150784
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004859253830149423,
      "loss": 2.6966,
      "theoretical_loss": 3.695942812171218,
      "tokens_seen": 876216320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004859088329865708,
      "loss": 2.7976,
      "theoretical_loss": 3.6959158596520205,
      "tokens_seen": 876281856
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048589228295819933,
      "loss": 2.6426,
      "theoretical_loss": 3.6958889097128527,
      "tokens_seen": 876347392
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048587573292982786,
      "loss": 2.6958,
      "theoretical_loss": 3.6958619623532742,
      "tokens_seen": 876412928
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004858591829014564,
      "loss": 2.4132,
      "theoretical_loss": 3.6958350175728456,
      "tokens_seen": 876478464
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004858426328730849,
      "loss": 2.5263,
      "theoretical_loss": 3.695808075371127,
      "tokens_seen": 876544000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004858260828447134,
      "loss": 2.7514,
      "theoretical_loss": 3.695781135747679,
      "tokens_seen": 876609536
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048580953281634196,
      "loss": 2.8371,
      "theoretical_loss": 3.695754198702062,
      "tokens_seen": 876675072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004857929827879705,
      "loss": 2.6709,
      "theoretical_loss": 3.6957272642338372,
      "tokens_seen": 876740608
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000485776432759599,
      "loss": 2.8364,
      "theoretical_loss": 3.695700332342564,
      "tokens_seen": 876806144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004857598827312275,
      "loss": 2.6866,
      "theoretical_loss": 3.6956734030278047,
      "tokens_seen": 876871680
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048574333270285606,
      "loss": 2.7567,
      "theoretical_loss": 3.6956464762891192,
      "tokens_seen": 876937216
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004857267826744846,
      "loss": 2.8594,
      "theoretical_loss": 3.69561955212607,
      "tokens_seen": 877002752
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004857102326461131,
      "loss": 2.6624,
      "theoretical_loss": 3.6955926305382167,
      "tokens_seen": 877068288
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004856936826177416,
      "loss": 2.7634,
      "theoretical_loss": 3.6955657115251213,
      "tokens_seen": 877133824
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048567713258937015,
      "loss": 2.5674,
      "theoretical_loss": 3.695538795086345,
      "tokens_seen": 877199360
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004856605825609987,
      "loss": 2.7314,
      "theoretical_loss": 3.69551188122145,
      "tokens_seen": 877264896
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004856440325326272,
      "loss": 2.652,
      "theoretical_loss": 3.695484969929997,
      "tokens_seen": 877330432
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004856274825042557,
      "loss": 2.593,
      "theoretical_loss": 3.695458061211548,
      "tokens_seen": 877395968
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048561093247588425,
      "loss": 2.5798,
      "theoretical_loss": 3.6954311550656658,
      "tokens_seen": 877461504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004855943824475128,
      "loss": 2.604,
      "theoretical_loss": 3.6954042514919108,
      "tokens_seen": 877527040
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048557783241914127,
      "loss": 2.5095,
      "theoretical_loss": 3.6953773504898457,
      "tokens_seen": 877592576
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004855612823907698,
      "loss": 2.4952,
      "theoretical_loss": 3.6953504520590332,
      "tokens_seen": 877658112
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048554473236239834,
      "loss": 2.8848,
      "theoretical_loss": 3.695323556199035,
      "tokens_seen": 877723648
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004855281823340269,
      "loss": 2.7311,
      "theoretical_loss": 3.6952966629094135,
      "tokens_seen": 877789184
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048551163230565537,
      "loss": 2.6775,
      "theoretical_loss": 3.695269772189731,
      "tokens_seen": 877854720
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004854950822772839,
      "loss": 2.6721,
      "theoretical_loss": 3.6952428840395504,
      "tokens_seen": 877920256
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048547853224891244,
      "loss": 2.789,
      "theoretical_loss": 3.6952159984584343,
      "tokens_seen": 877985792
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000485461982220541,
      "loss": 2.7019,
      "theoretical_loss": 3.6951891154459453,
      "tokens_seen": 878051328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048544543219216946,
      "loss": 2.4714,
      "theoretical_loss": 3.6951622350016473,
      "tokens_seen": 878116864
    },
    {
      "epoch": 0.31,
      "objective/train/avg_token_score": 0.01299184001982212,
      "objective/train/avg_weight": 0.984708845615387,
      "objective/train/docs_used": 504313,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.827441453933716,
      "objective/train/std_weight": 0.06926679611206055,
      "objective/train/theoretical_loss": 3.6951353571251015,
      "objective/train/tokens_used": 898642400,
      "theoretical_loss": 3.6951353571251015,
      "tokens_seen": 878182400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000485428882163798,
      "loss": 2.5564,
      "theoretical_loss": 3.6951353571251015,
      "tokens_seen": 878182400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048541233213542654,
      "loss": 2.6089,
      "theoretical_loss": 3.695108481815873,
      "tokens_seen": 878247936
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004853957821070551,
      "loss": 2.6096,
      "theoretical_loss": 3.6950816090735232,
      "tokens_seen": 878313472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048537923207868356,
      "loss": 2.7295,
      "theoretical_loss": 3.6950547388976167,
      "tokens_seen": 878379008
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004853626820503121,
      "loss": 2.631,
      "theoretical_loss": 3.6950278712877163,
      "tokens_seen": 878444544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048534613202194063,
      "loss": 2.338,
      "theoretical_loss": 3.6950010062433862,
      "tokens_seen": 878510080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048532958199356917,
      "loss": 2.6564,
      "theoretical_loss": 3.694974143764189,
      "tokens_seen": 878575616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048531303196519765,
      "loss": 2.7819,
      "theoretical_loss": 3.6949472838496895,
      "tokens_seen": 878641152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004852964819368262,
      "loss": 2.7341,
      "theoretical_loss": 3.6949204264994515,
      "tokens_seen": 878706688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048527993190845473,
      "loss": 2.694,
      "theoretical_loss": 3.6948935717130382,
      "tokens_seen": 878772224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048526338188008327,
      "loss": 2.7396,
      "theoretical_loss": 3.6948667194900144,
      "tokens_seen": 878837760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048524683185171175,
      "loss": 2.9174,
      "theoretical_loss": 3.6948398698299436,
      "tokens_seen": 878903296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004852302818233403,
      "loss": 2.5191,
      "theoretical_loss": 3.694813022732391,
      "tokens_seen": 878968832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004852137317949688,
      "loss": 2.8276,
      "theoretical_loss": 3.69478617819692,
      "tokens_seen": 879034368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048519718176659736,
      "loss": 2.6628,
      "theoretical_loss": 3.694759336223096,
      "tokens_seen": 879099904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004851806317382258,
      "loss": 2.8651,
      "theoretical_loss": 3.694732496810483,
      "tokens_seen": 879165440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048516408170985433,
      "loss": 2.7258,
      "theoretical_loss": 3.694705659958646,
      "tokens_seen": 879230976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048514753168148287,
      "loss": 2.6919,
      "theoretical_loss": 3.6946788256671494,
      "tokens_seen": 879296512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048513098165311135,
      "loss": 2.8053,
      "theoretical_loss": 3.694651993935559,
      "tokens_seen": 879362048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004851144316247399,
      "loss": 2.6185,
      "theoretical_loss": 3.6946251647634396,
      "tokens_seen": 879427584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004850978815963684,
      "loss": 2.769,
      "theoretical_loss": 3.6945983381503558,
      "tokens_seen": 879493120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048508133156799696,
      "loss": 2.8712,
      "theoretical_loss": 3.6945715140958733,
      "tokens_seen": 879558656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048506478153962545,
      "loss": 2.6104,
      "theoretical_loss": 3.694544692599557,
      "tokens_seen": 879624192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000485048231511254,
      "loss": 2.5511,
      "theoretical_loss": 3.6945178736609723,
      "tokens_seen": 879689728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004850316814828825,
      "loss": 2.57,
      "theoretical_loss": 3.6944910572796856,
      "tokens_seen": 879755264
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048501513145451106,
      "loss": 2.8647,
      "theoretical_loss": 3.694464243455262,
      "tokens_seen": 879820800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048499858142613954,
      "loss": 2.8279,
      "theoretical_loss": 3.694437432187268,
      "tokens_seen": 879886336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004849820313977681,
      "loss": 2.7993,
      "theoretical_loss": 3.694410623475268,
      "tokens_seen": 879951872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004849654813693966,
      "loss": 2.6828,
      "theoretical_loss": 3.694383817318829,
      "tokens_seen": 880017408
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048494893134102515,
      "loss": 2.767,
      "theoretical_loss": 3.694357013717517,
      "tokens_seen": 880082944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048493238131265364,
      "loss": 2.6119,
      "theoretical_loss": 3.6943302126708986,
      "tokens_seen": 880148480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004849158312842822,
      "loss": 2.6626,
      "theoretical_loss": 3.6943034141785396,
      "tokens_seen": 880214016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004848992812559107,
      "loss": 2.4834,
      "theoretical_loss": 3.694276618240006,
      "tokens_seen": 880279552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048488273122753925,
      "loss": 2.7815,
      "theoretical_loss": 3.6942498248548654,
      "tokens_seen": 880345088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048486618119916773,
      "loss": 2.5881,
      "theoretical_loss": 3.6942230340226834,
      "tokens_seen": 880410624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048484963117079627,
      "loss": 2.6124,
      "theoretical_loss": 3.6941962457430275,
      "tokens_seen": 880476160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004848330811424248,
      "loss": 2.7078,
      "theoretical_loss": 3.6941694600154644,
      "tokens_seen": 880541696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048481653111405335,
      "loss": 2.6304,
      "theoretical_loss": 3.6941426768395607,
      "tokens_seen": 880607232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048479998108568183,
      "loss": 2.6055,
      "theoretical_loss": 3.6941158962148837,
      "tokens_seen": 880672768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048478343105731037,
      "loss": 2.8439,
      "theoretical_loss": 3.694089118141,
      "tokens_seen": 880738304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004847668810289389,
      "loss": 2.7657,
      "theoretical_loss": 3.694062342617478,
      "tokens_seen": 880803840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048475033100056744,
      "loss": 2.6022,
      "theoretical_loss": 3.6940355696438845,
      "tokens_seen": 880869376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004847337809721959,
      "loss": 2.7449,
      "theoretical_loss": 3.6940087992197865,
      "tokens_seen": 880934912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048471723094382446,
      "loss": 2.8905,
      "theoretical_loss": 3.693982031344752,
      "tokens_seen": 881000448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000484700680915453,
      "loss": 2.6577,
      "theoretical_loss": 3.6939552660183486,
      "tokens_seen": 881065984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048468413088708154,
      "loss": 2.818,
      "theoretical_loss": 3.6939285032401443,
      "tokens_seen": 881131520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048466758085871,
      "loss": 2.5362,
      "theoretical_loss": 3.6939017430097074,
      "tokens_seen": 881197056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048465103083033856,
      "loss": 2.6019,
      "theoretical_loss": 3.6938749853266044,
      "tokens_seen": 881262592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004846344808019671,
      "loss": 2.5425,
      "theoretical_loss": 3.6938482301904054,
      "tokens_seen": 881328128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048461793077359563,
      "loss": 2.6478,
      "theoretical_loss": 3.693821477600677,
      "tokens_seen": 881393664
    },
    {
      "epoch": 0.31,
      "objective/train/avg_token_score": 0.014565665274858475,
      "objective/train/avg_weight": 0.9828512072563171,
      "objective/train/docs_used": 506500,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.051870584487915,
      "objective/train/std_weight": 0.07804255187511444,
      "objective/train/theoretical_loss": 3.693794727556988,
      "objective/train/tokens_used": 901919200,
      "theoretical_loss": 3.693794727556988,
      "tokens_seen": 881459200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004846013807452241,
      "loss": 2.7557,
      "theoretical_loss": 3.693794727556988,
      "tokens_seen": 881459200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048458483071685265,
      "loss": 2.4218,
      "theoretical_loss": 3.693767980058907,
      "tokens_seen": 881524736
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004845682806884812,
      "loss": 2.2577,
      "theoretical_loss": 3.693741235106003,
      "tokens_seen": 881590272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048455173066010973,
      "loss": 2.6882,
      "theoretical_loss": 3.6937144926978434,
      "tokens_seen": 881655808
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004845351806317382,
      "loss": 2.6873,
      "theoretical_loss": 3.693687752833998,
      "tokens_seen": 881721344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00048451863060336675,
      "loss": 2.9891,
      "theoretical_loss": 3.693661015514035,
      "tokens_seen": 881786880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004845020805749953,
      "loss": 2.591,
      "theoretical_loss": 3.693634280737524,
      "tokens_seen": 881852416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004844855305466238,
      "loss": 2.8864,
      "theoretical_loss": 3.6936075485040334,
      "tokens_seen": 881917952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0004844689805182523,
      "loss": 2.7219,
      "theoretical_loss": 3.6935808188131327,
      "tokens_seen": 881983488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048445243048988085,
      "loss": 2.565,
      "theoretical_loss": 3.6935540916643914,
      "tokens_seen": 882049024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004844358804615094,
      "loss": 2.6295,
      "theoretical_loss": 3.6935273670573787,
      "tokens_seen": 882114560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004844193304331379,
      "loss": 2.8067,
      "theoretical_loss": 3.693500644991664,
      "tokens_seen": 882180096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004844027804047664,
      "loss": 2.3784,
      "theoretical_loss": 3.693473925466816,
      "tokens_seen": 882245632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048438623037639494,
      "loss": 2.6811,
      "theoretical_loss": 3.693447208482406,
      "tokens_seen": 882311168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004843696803480235,
      "loss": 2.6554,
      "theoretical_loss": 3.693420494038003,
      "tokens_seen": 882376704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000484353130319652,
      "loss": 2.5583,
      "theoretical_loss": 3.6933937821331764,
      "tokens_seen": 882442240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004843365802912805,
      "loss": 2.7433,
      "theoretical_loss": 3.693367072767497,
      "tokens_seen": 882507776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048432003026290904,
      "loss": 2.8619,
      "theoretical_loss": 3.6933403659405344,
      "tokens_seen": 882573312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004843034802345376,
      "loss": 2.7924,
      "theoretical_loss": 3.6933136616518594,
      "tokens_seen": 882638848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004842869302061661,
      "loss": 2.6076,
      "theoretical_loss": 3.693286959901042,
      "tokens_seen": 882704384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004842703801777946,
      "loss": 2.7801,
      "theoretical_loss": 3.6932602606876523,
      "tokens_seen": 882769920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048425383014942313,
      "loss": 2.9724,
      "theoretical_loss": 3.693233564011261,
      "tokens_seen": 882835456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048423728012105167,
      "loss": 2.7787,
      "theoretical_loss": 3.693206869871439,
      "tokens_seen": 882900992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004842207300926802,
      "loss": 2.8001,
      "theoretical_loss": 3.6931801782677565,
      "tokens_seen": 882966528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004842041800643087,
      "loss": 2.6033,
      "theoretical_loss": 3.693153489199785,
      "tokens_seen": 883032064
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004841876300359372,
      "loss": 2.709,
      "theoretical_loss": 3.693126802667095,
      "tokens_seen": 883097600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048417108000756566,
      "loss": 2.522,
      "theoretical_loss": 3.693100118669258,
      "tokens_seen": 883163136
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004841545299791942,
      "loss": 2.6159,
      "theoretical_loss": 3.693073437205844,
      "tokens_seen": 883228672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048413797995082273,
      "loss": 2.7148,
      "theoretical_loss": 3.6930467582764255,
      "tokens_seen": 883294208
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048412142992245127,
      "loss": 2.7733,
      "theoretical_loss": 3.6930200818805736,
      "tokens_seen": 883359744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048410487989407976,
      "loss": 2.6233,
      "theoretical_loss": 3.6929934080178595,
      "tokens_seen": 883425280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004840883298657083,
      "loss": 2.6906,
      "theoretical_loss": 3.692966736687855,
      "tokens_seen": 883490816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048407177983733683,
      "loss": 2.9689,
      "theoretical_loss": 3.692940067890131,
      "tokens_seen": 883556352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048405522980896537,
      "loss": 2.6891,
      "theoretical_loss": 3.6929134016242604,
      "tokens_seen": 883621888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048403867978059385,
      "loss": 2.8683,
      "theoretical_loss": 3.6928867378898147,
      "tokens_seen": 883687424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004840221297522224,
      "loss": 2.7848,
      "theoretical_loss": 3.6928600766863653,
      "tokens_seen": 883752960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004840055797238509,
      "loss": 2.9683,
      "theoretical_loss": 3.6928334180134854,
      "tokens_seen": 883818496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048398902969547946,
      "loss": 2.6698,
      "theoretical_loss": 3.692806761870746,
      "tokens_seen": 883884032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000483972479667108,
      "loss": 2.6134,
      "theoretical_loss": 3.69278010825772,
      "tokens_seen": 883949568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004839559296387365,
      "loss": 2.7737,
      "theoretical_loss": 3.6927534571739797,
      "tokens_seen": 884015104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000483939379610365,
      "loss": 2.8624,
      "theoretical_loss": 3.692726808619098,
      "tokens_seen": 884080640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048392282958199356,
      "loss": 2.6168,
      "theoretical_loss": 3.6927001625926463,
      "tokens_seen": 884146176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004839062795536221,
      "loss": 2.5895,
      "theoretical_loss": 3.692673519094199,
      "tokens_seen": 884211712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004838897295252506,
      "loss": 2.8587,
      "theoretical_loss": 3.6926468781233277,
      "tokens_seen": 884277248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004838731794968791,
      "loss": 2.4494,
      "theoretical_loss": 3.6926202396796057,
      "tokens_seen": 884342784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048385662946850766,
      "loss": 2.8451,
      "theoretical_loss": 3.692593603762606,
      "tokens_seen": 884408320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004838400794401362,
      "loss": 2.5597,
      "theoretical_loss": 3.692566970371902,
      "tokens_seen": 884473856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004838235294117647,
      "loss": 2.8183,
      "theoretical_loss": 3.692540339507066,
      "tokens_seen": 884539392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004838069793833932,
      "loss": 2.7235,
      "theoretical_loss": 3.692513711167673,
      "tokens_seen": 884604928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048379042935502175,
      "loss": 2.9855,
      "theoretical_loss": 3.692487085353295,
      "tokens_seen": 884670464
    },
    {
      "epoch": 0.32,
      "objective/train/avg_token_score": 0.0073553090915083885,
      "objective/train/avg_weight": 0.9913473129272461,
      "objective/train/docs_used": 508372,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6066365242004395,
      "objective/train/std_weight": 0.03385532274842262,
      "objective/train/theoretical_loss": 3.692460462063506,
      "objective/train/tokens_used": 905196000,
      "theoretical_loss": 3.692460462063506,
      "tokens_seen": 884736000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004837738793266503,
      "loss": 2.5817,
      "theoretical_loss": 3.692460462063506,
      "tokens_seen": 884736000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004837573292982788,
      "loss": 2.852,
      "theoretical_loss": 3.6924338412978797,
      "tokens_seen": 884801536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004837407792699073,
      "loss": 2.7711,
      "theoretical_loss": 3.6924072230559895,
      "tokens_seen": 884867072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048372422924153585,
      "loss": 2.7346,
      "theoretical_loss": 3.6923806073374097,
      "tokens_seen": 884932608
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004837076792131644,
      "loss": 2.5655,
      "theoretical_loss": 3.692353994141714,
      "tokens_seen": 884998144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048369112918479287,
      "loss": 2.8375,
      "theoretical_loss": 3.6923273834684767,
      "tokens_seen": 885063680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004836745791564214,
      "loss": 2.6984,
      "theoretical_loss": 3.6923007753172716,
      "tokens_seen": 885129216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048365802912804994,
      "loss": 2.6734,
      "theoretical_loss": 3.692274169687673,
      "tokens_seen": 885194752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004836414790996785,
      "loss": 2.5202,
      "theoretical_loss": 3.692247566579256,
      "tokens_seen": 885260288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048362492907130696,
      "loss": 2.7217,
      "theoretical_loss": 3.692220965991594,
      "tokens_seen": 885325824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004836083790429355,
      "loss": 2.9022,
      "theoretical_loss": 3.6921943679242624,
      "tokens_seen": 885391360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048359182901456404,
      "loss": 2.8692,
      "theoretical_loss": 3.692167772376836,
      "tokens_seen": 885456896
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004835752789861926,
      "loss": 2.7459,
      "theoretical_loss": 3.692141179348888,
      "tokens_seen": 885522432
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048355872895782106,
      "loss": 2.7744,
      "theoretical_loss": 3.6921145888399955,
      "tokens_seen": 885587968
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004835421789294496,
      "loss": 2.7821,
      "theoretical_loss": 3.692088000849732,
      "tokens_seen": 885653504
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048352562890107814,
      "loss": 2.6223,
      "theoretical_loss": 3.692061415377673,
      "tokens_seen": 885719040
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004835090788727067,
      "loss": 2.7575,
      "theoretical_loss": 3.692034832423394,
      "tokens_seen": 885784576
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048349252884433516,
      "loss": 2.7736,
      "theoretical_loss": 3.6920082519864694,
      "tokens_seen": 885850112
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004834759788159637,
      "loss": 2.5591,
      "theoretical_loss": 3.6919816740664757,
      "tokens_seen": 885915648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048345942878759223,
      "loss": 2.8273,
      "theoretical_loss": 3.691955098662988,
      "tokens_seen": 885981184
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048344287875922077,
      "loss": 2.9457,
      "theoretical_loss": 3.6919285257755816,
      "tokens_seen": 886046720
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048342632873084925,
      "loss": 2.7047,
      "theoretical_loss": 3.6919019554038326,
      "tokens_seen": 886112256
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004834097787024778,
      "loss": 2.6615,
      "theoretical_loss": 3.6918753875473165,
      "tokens_seen": 886177792
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048339322867410633,
      "loss": 2.5747,
      "theoretical_loss": 3.6918488222056096,
      "tokens_seen": 886243328
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048337667864573487,
      "loss": 2.9203,
      "theoretical_loss": 3.691822259378287,
      "tokens_seen": 886308864
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048336012861736335,
      "loss": 2.7408,
      "theoretical_loss": 3.6917956990649263,
      "tokens_seen": 886374400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004833435785889919,
      "loss": 2.5791,
      "theoretical_loss": 3.6917691412651026,
      "tokens_seen": 886439936
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004833270285606204,
      "loss": 2.8353,
      "theoretical_loss": 3.6917425859783926,
      "tokens_seen": 886505472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048331047853224896,
      "loss": 2.6332,
      "theoretical_loss": 3.691716033204373,
      "tokens_seen": 886571008
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048329392850387744,
      "loss": 2.6024,
      "theoretical_loss": 3.6916894829426194,
      "tokens_seen": 886636544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000483277378475506,
      "loss": 2.723,
      "theoretical_loss": 3.6916629351927095,
      "tokens_seen": 886702080
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004832608284471345,
      "loss": 2.847,
      "theoretical_loss": 3.69163638995422,
      "tokens_seen": 886767616
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048324427841876306,
      "loss": 2.524,
      "theoretical_loss": 3.6916098472267267,
      "tokens_seen": 886833152
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048322772839039154,
      "loss": 2.7648,
      "theoretical_loss": 3.691583307009807,
      "tokens_seen": 886898688
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048321117836202,
      "loss": 2.6159,
      "theoretical_loss": 3.6915567693030384,
      "tokens_seen": 886964224
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004831946283336485,
      "loss": 2.7323,
      "theoretical_loss": 3.691530234105998,
      "tokens_seen": 887029760
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048317807830527704,
      "loss": 2.7764,
      "theoretical_loss": 3.6915037014182626,
      "tokens_seen": 887095296
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004831615282769056,
      "loss": 2.4479,
      "theoretical_loss": 3.69147717123941,
      "tokens_seen": 887160832
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004831449782485341,
      "loss": 2.7373,
      "theoretical_loss": 3.6914506435690173,
      "tokens_seen": 887226368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004831284282201626,
      "loss": 2.9152,
      "theoretical_loss": 3.691424118406662,
      "tokens_seen": 887291904
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048311187819179114,
      "loss": 3.1419,
      "theoretical_loss": 3.691397595751922,
      "tokens_seen": 887357440
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004830953281634197,
      "loss": 3.0141,
      "theoretical_loss": 3.6913710756043754,
      "tokens_seen": 887422976
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004830787781350482,
      "loss": 2.74,
      "theoretical_loss": 3.6913445579635997,
      "tokens_seen": 887488512
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004830622281066767,
      "loss": 2.9578,
      "theoretical_loss": 3.691318042829172,
      "tokens_seen": 887554048
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048304567807830524,
      "loss": 2.6003,
      "theoretical_loss": 3.691291530200672,
      "tokens_seen": 887619584
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004830291280499338,
      "loss": 2.8209,
      "theoretical_loss": 3.691265020077677,
      "tokens_seen": 887685120
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004830125780215623,
      "loss": 2.8679,
      "theoretical_loss": 3.6912385124597655,
      "tokens_seen": 887750656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004829960279931908,
      "loss": 2.8538,
      "theoretical_loss": 3.6912120073465156,
      "tokens_seen": 887816192
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048297947796481933,
      "loss": 2.7326,
      "theoretical_loss": 3.6911855047375055,
      "tokens_seen": 887881728
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048296292793644787,
      "loss": 2.6227,
      "theoretical_loss": 3.691159004632315,
      "tokens_seen": 887947264
    },
    {
      "epoch": 0.32,
      "objective/train/avg_token_score": 0.024096213281154633,
      "objective/train/avg_weight": 0.9716524481773376,
      "objective/train/docs_used": 510118,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.8386683464050293,
      "objective/train/std_weight": 0.14509572088718414,
      "objective/train/theoretical_loss": 3.691132507030521,
      "objective/train/tokens_used": 908472800,
      "theoretical_loss": 3.691132507030521,
      "tokens_seen": 888012800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004829463779080764,
      "loss": 2.8071,
      "theoretical_loss": 3.691132507030521,
      "tokens_seen": 888012800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004829298278797049,
      "loss": 2.6467,
      "theoretical_loss": 3.691106011931704,
      "tokens_seen": 888078336
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048291327785133343,
      "loss": 2.6394,
      "theoretical_loss": 3.6910795193354415,
      "tokens_seen": 888143872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048289672782296197,
      "loss": 2.7202,
      "theoretical_loss": 3.6910530292413135,
      "tokens_seen": 888209408
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004828801777945905,
      "loss": 2.5903,
      "theoretical_loss": 3.691026541648899,
      "tokens_seen": 888274944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000482863627766219,
      "loss": 2.6429,
      "theoretical_loss": 3.691000056557777,
      "tokens_seen": 888340480
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004828470777378475,
      "loss": 2.7014,
      "theoretical_loss": 3.6909735739675265,
      "tokens_seen": 888406016
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048283052770947606,
      "loss": 2.6873,
      "theoretical_loss": 3.690947093877727,
      "tokens_seen": 888471552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004828139776811046,
      "loss": 2.7275,
      "theoretical_loss": 3.690920616287958,
      "tokens_seen": 888537088
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004827974276527331,
      "loss": 2.9009,
      "theoretical_loss": 3.6908941411977993,
      "tokens_seen": 888602624
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004827808776243616,
      "loss": 2.7115,
      "theoretical_loss": 3.6908676686068307,
      "tokens_seen": 888668160
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048276432759599016,
      "loss": 2.7473,
      "theoretical_loss": 3.690841198514632,
      "tokens_seen": 888733696
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004827477775676187,
      "loss": 2.8492,
      "theoretical_loss": 3.6908147309207826,
      "tokens_seen": 888799232
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004827312275392472,
      "loss": 2.702,
      "theoretical_loss": 3.690788265824863,
      "tokens_seen": 888864768
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004827146775108757,
      "loss": 2.9792,
      "theoretical_loss": 3.6907618032264526,
      "tokens_seen": 888930304
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048269812748250425,
      "loss": 2.6082,
      "theoretical_loss": 3.6907353431251324,
      "tokens_seen": 888995840
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004826815774541328,
      "loss": 2.7589,
      "theoretical_loss": 3.6907088855204826,
      "tokens_seen": 889061376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004826650274257613,
      "loss": 2.8178,
      "theoretical_loss": 3.6906824304120835,
      "tokens_seen": 889126912
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004826484773973898,
      "loss": 2.6583,
      "theoretical_loss": 3.6906559777995156,
      "tokens_seen": 889192448
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048263192736901835,
      "loss": 2.9121,
      "theoretical_loss": 3.690629527682359,
      "tokens_seen": 889257984
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004826153773406469,
      "loss": 2.7964,
      "theoretical_loss": 3.690603080060195,
      "tokens_seen": 889323520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048259882731227537,
      "loss": 2.8525,
      "theoretical_loss": 3.690576634932605,
      "tokens_seen": 889389056
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004825822772839039,
      "loss": 2.5893,
      "theoretical_loss": 3.6905501922991686,
      "tokens_seen": 889454592
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048256572725553245,
      "loss": 2.8914,
      "theoretical_loss": 3.690523752159468,
      "tokens_seen": 889520128
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000482549177227161,
      "loss": 2.6549,
      "theoretical_loss": 3.690497314513083,
      "tokens_seen": 889585664
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048253262719878947,
      "loss": 2.7195,
      "theoretical_loss": 3.690470879359596,
      "tokens_seen": 889651200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000482516077170418,
      "loss": 2.8524,
      "theoretical_loss": 3.6904444466985877,
      "tokens_seen": 889716736
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048249952714204654,
      "loss": 2.7224,
      "theoretical_loss": 3.69041801652964,
      "tokens_seen": 889782272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004824829771136751,
      "loss": 2.659,
      "theoretical_loss": 3.6903915888523344,
      "tokens_seen": 889847808
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004824664270853036,
      "loss": 2.8204,
      "theoretical_loss": 3.690365163666252,
      "tokens_seen": 889913344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004824498770569321,
      "loss": 2.7468,
      "theoretical_loss": 3.690338740970975,
      "tokens_seen": 889978880
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048243332702856064,
      "loss": 2.7744,
      "theoretical_loss": 3.6903123207660844,
      "tokens_seen": 890044416
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004824167770001892,
      "loss": 2.8335,
      "theoretical_loss": 3.6902859030511634,
      "tokens_seen": 890109952
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004824002269718177,
      "loss": 3.0101,
      "theoretical_loss": 3.6902594878257933,
      "tokens_seen": 890175488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004823836769434462,
      "loss": 2.5684,
      "theoretical_loss": 3.6902330750895564,
      "tokens_seen": 890241024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048236712691507473,
      "loss": 2.8703,
      "theoretical_loss": 3.6902066648420346,
      "tokens_seen": 890306560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048235057688670327,
      "loss": 2.7084,
      "theoretical_loss": 3.690180257082811,
      "tokens_seen": 890372096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004823340268583318,
      "loss": 3.0025,
      "theoretical_loss": 3.690153851811467,
      "tokens_seen": 890437632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004823174768299603,
      "loss": 2.5757,
      "theoretical_loss": 3.690127449027586,
      "tokens_seen": 890503168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048230092680158883,
      "loss": 2.7077,
      "theoretical_loss": 3.6901010487307504,
      "tokens_seen": 890568704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048228437677321737,
      "loss": 2.9704,
      "theoretical_loss": 3.6900746509205424,
      "tokens_seen": 890634240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004822678267448459,
      "loss": 2.786,
      "theoretical_loss": 3.6900482555965457,
      "tokens_seen": 890699776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004822512767164744,
      "loss": 2.6393,
      "theoretical_loss": 3.6900218627583428,
      "tokens_seen": 890765312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004822347266881029,
      "loss": 2.7882,
      "theoretical_loss": 3.689995472405516,
      "tokens_seen": 890830848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048221817665973136,
      "loss": 2.7613,
      "theoretical_loss": 3.6899690845376503,
      "tokens_seen": 890896384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004822016266313599,
      "loss": 2.9101,
      "theoretical_loss": 3.689942699154327,
      "tokens_seen": 890961920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048218507660298843,
      "loss": 2.8195,
      "theoretical_loss": 3.6899163162551307,
      "tokens_seen": 891027456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048216852657461697,
      "loss": 2.9025,
      "theoretical_loss": 3.689889935839645,
      "tokens_seen": 891092992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048215197654624545,
      "loss": 2.8241,
      "theoretical_loss": 3.6898635579074517,
      "tokens_seen": 891158528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000482135426517874,
      "loss": 2.7555,
      "theoretical_loss": 3.6898371824581364,
      "tokens_seen": 891224064
    },
    {
      "epoch": 0.32,
      "objective/train/avg_token_score": 0.010467197746038437,
      "objective/train/avg_weight": 0.987676739692688,
      "objective/train/docs_used": 511934,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3958497047424316,
      "objective/train/std_weight": 0.06989966332912445,
      "objective/train/theoretical_loss": 3.6898108094912816,
      "objective/train/tokens_used": 911749600,
      "theoretical_loss": 3.6898108094912816,
      "tokens_seen": 891289600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004821188764895025,
      "loss": 2.9067,
      "theoretical_loss": 3.6898108094912816,
      "tokens_seen": 891289600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048210232646113106,
      "loss": 2.9261,
      "theoretical_loss": 3.6897844390064716,
      "tokens_seen": 891355136
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048208577643275955,
      "loss": 2.785,
      "theoretical_loss": 3.6897580710032907,
      "tokens_seen": 891420672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004820692264043881,
      "loss": 2.7531,
      "theoretical_loss": 3.6897317054813223,
      "tokens_seen": 891486208
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004820526763760166,
      "loss": 2.9893,
      "theoretical_loss": 3.689705342440151,
      "tokens_seen": 891551744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048203612634764516,
      "loss": 2.6014,
      "theoretical_loss": 3.6896789818793607,
      "tokens_seen": 891617280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048201957631927364,
      "loss": 2.7652,
      "theoretical_loss": 3.689652623798536,
      "tokens_seen": 891682816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004820030262909022,
      "loss": 2.9747,
      "theoretical_loss": 3.689626268197261,
      "tokens_seen": 891748352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004819864762625307,
      "loss": 2.5403,
      "theoretical_loss": 3.689599915075121,
      "tokens_seen": 891813888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048196992623415926,
      "loss": 2.6274,
      "theoretical_loss": 3.6895735644317,
      "tokens_seen": 891879424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048195337620578774,
      "loss": 2.7086,
      "theoretical_loss": 3.6895472162665826,
      "tokens_seen": 891944960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004819368261774163,
      "loss": 2.9116,
      "theoretical_loss": 3.6895208705793534,
      "tokens_seen": 892010496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004819202761490448,
      "loss": 2.727,
      "theoretical_loss": 3.6894945273695985,
      "tokens_seen": 892076032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048190372612067335,
      "loss": 2.6741,
      "theoretical_loss": 3.6894681866369026,
      "tokens_seen": 892141568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048188717609230183,
      "loss": 2.5902,
      "theoretical_loss": 3.6894418483808504,
      "tokens_seen": 892207104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048187062606393037,
      "loss": 2.6728,
      "theoretical_loss": 3.689415512601027,
      "tokens_seen": 892272640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004818540760355589,
      "loss": 2.4574,
      "theoretical_loss": 3.6893891792970184,
      "tokens_seen": 892338176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048183752600718745,
      "loss": 2.7834,
      "theoretical_loss": 3.6893628484684093,
      "tokens_seen": 892403712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048182097597881593,
      "loss": 2.865,
      "theoretical_loss": 3.689336520114786,
      "tokens_seen": 892469248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048180442595044447,
      "loss": 2.9001,
      "theoretical_loss": 3.689310194235733,
      "tokens_seen": 892534784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000481787875922073,
      "loss": 2.7448,
      "theoretical_loss": 3.6892838708308378,
      "tokens_seen": 892600320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048177132589370154,
      "loss": 2.9281,
      "theoretical_loss": 3.6892575498996845,
      "tokens_seen": 892665856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048175477586533,
      "loss": 3.0222,
      "theoretical_loss": 3.6892312314418607,
      "tokens_seen": 892731392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048173822583695856,
      "loss": 2.6402,
      "theoretical_loss": 3.689204915456951,
      "tokens_seen": 892796928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004817216758085871,
      "loss": 2.862,
      "theoretical_loss": 3.689178601944542,
      "tokens_seen": 892862464
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048170512578021564,
      "loss": 2.9655,
      "theoretical_loss": 3.68915229090422,
      "tokens_seen": 892928000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004816885757518441,
      "loss": 2.76,
      "theoretical_loss": 3.6891259823355718,
      "tokens_seen": 892993536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048167202572347266,
      "loss": 2.8827,
      "theoretical_loss": 3.689099676238183,
      "tokens_seen": 893059072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004816554756951012,
      "loss": 2.9415,
      "theoretical_loss": 3.6890733726116407,
      "tokens_seen": 893124608
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048163892566672974,
      "loss": 2.5598,
      "theoretical_loss": 3.6890470714555317,
      "tokens_seen": 893190144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004816223756383582,
      "loss": 2.4545,
      "theoretical_loss": 3.6890207727694415,
      "tokens_seen": 893255680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048160582560998676,
      "loss": 2.9899,
      "theoretical_loss": 3.688994476552959,
      "tokens_seen": 893321216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004815892755816153,
      "loss": 2.4572,
      "theoretical_loss": 3.6889681828056693,
      "tokens_seen": 893386752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048157272555324383,
      "loss": 2.7098,
      "theoretical_loss": 3.6889418915271603,
      "tokens_seen": 893452288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004815561755248723,
      "loss": 2.8535,
      "theoretical_loss": 3.688915602717019,
      "tokens_seen": 893517824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048153962549650085,
      "loss": 2.9687,
      "theoretical_loss": 3.688889316374833,
      "tokens_seen": 893583360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004815230754681294,
      "loss": 2.854,
      "theoretical_loss": 3.688863032500189,
      "tokens_seen": 893648896
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048150652543975793,
      "loss": 2.7407,
      "theoretical_loss": 3.6888367510926745,
      "tokens_seen": 893714432
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004814899754113864,
      "loss": 2.6118,
      "theoretical_loss": 3.6888104721518777,
      "tokens_seen": 893779968
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048147342538301495,
      "loss": 2.4742,
      "theoretical_loss": 3.6887841956773855,
      "tokens_seen": 893845504
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004814568753546435,
      "loss": 2.6643,
      "theoretical_loss": 3.688757921668786,
      "tokens_seen": 893911040
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000481440325326272,
      "loss": 2.7105,
      "theoretical_loss": 3.688731650125667,
      "tokens_seen": 893976576
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004814237752979005,
      "loss": 2.5489,
      "theoretical_loss": 3.6887053810476162,
      "tokens_seen": 894042112
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048140722526952904,
      "loss": 3.0621,
      "theoretical_loss": 3.6886791144342217,
      "tokens_seen": 894107648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004813906752411576,
      "loss": 2.5843,
      "theoretical_loss": 3.688652850285072,
      "tokens_seen": 894173184
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004813741252127861,
      "loss": 2.7156,
      "theoretical_loss": 3.688626588599755,
      "tokens_seen": 894238720
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004813575751844146,
      "loss": 2.7644,
      "theoretical_loss": 3.6886003293778584,
      "tokens_seen": 894304256
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048134102515604314,
      "loss": 2.6619,
      "theoretical_loss": 3.688574072618972,
      "tokens_seen": 894369792
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004813244751276717,
      "loss": 2.7297,
      "theoretical_loss": 3.6885478183226836,
      "tokens_seen": 894435328
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004813079250993002,
      "loss": 2.7689,
      "theoretical_loss": 3.688521566488582,
      "tokens_seen": 894500864
    },
    {
      "epoch": 0.32,
      "objective/train/avg_token_score": 0.023056253790855408,
      "objective/train/avg_weight": 0.9728865623474121,
      "objective/train/docs_used": 513625,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.849482536315918,
      "objective/train/std_weight": 0.11544700711965561,
      "objective/train/theoretical_loss": 3.6884953171162556,
      "objective/train/tokens_used": 915026400,
      "theoretical_loss": 3.6884953171162556,
      "tokens_seen": 894566400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004812913750709287,
      "loss": 2.8776,
      "theoretical_loss": 3.6884953171162556,
      "tokens_seen": 894566400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048127482504255724,
      "loss": 2.9664,
      "theoretical_loss": 3.688469070205293,
      "tokens_seen": 894631936
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004812582750141858,
      "loss": 2.7871,
      "theoretical_loss": 3.688442825755284,
      "tokens_seen": 894697472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004812417249858142,
      "loss": 2.8519,
      "theoretical_loss": 3.688416583765817,
      "tokens_seen": 894763008
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048122517495744274,
      "loss": 2.6209,
      "theoretical_loss": 3.688390344236481,
      "tokens_seen": 894828544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004812086249290713,
      "loss": 2.8767,
      "theoretical_loss": 3.688364107166866,
      "tokens_seen": 894894080
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004811920749006998,
      "loss": 2.6072,
      "theoretical_loss": 3.6883378725565605,
      "tokens_seen": 894959616
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004811755248723283,
      "loss": 2.7136,
      "theoretical_loss": 3.6883116404051544,
      "tokens_seen": 895025152
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048115897484395684,
      "loss": 2.7458,
      "theoretical_loss": 3.6882854107122363,
      "tokens_seen": 895090688
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004811424248155854,
      "loss": 2.7257,
      "theoretical_loss": 3.6882591834773972,
      "tokens_seen": 895156224
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004811258747872139,
      "loss": 2.6344,
      "theoretical_loss": 3.688232958700226,
      "tokens_seen": 895221760
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004811093247588424,
      "loss": 2.6541,
      "theoretical_loss": 3.688206736380313,
      "tokens_seen": 895287296
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048109277473047093,
      "loss": 2.847,
      "theoretical_loss": 3.688180516517247,
      "tokens_seen": 895352832
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048107622470209947,
      "loss": 2.5829,
      "theoretical_loss": 3.6881542991106193,
      "tokens_seen": 895418368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000481059674673728,
      "loss": 2.853,
      "theoretical_loss": 3.6881280841600192,
      "tokens_seen": 895483904
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004810431246453565,
      "loss": 2.9713,
      "theoretical_loss": 3.6881018716650376,
      "tokens_seen": 895549440
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048102657461698503,
      "loss": 2.685,
      "theoretical_loss": 3.688075661625264,
      "tokens_seen": 895614976
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048101002458861357,
      "loss": 2.7516,
      "theoretical_loss": 3.6880494540402893,
      "tokens_seen": 895680512
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004809934745602421,
      "loss": 2.814,
      "theoretical_loss": 3.688023248909704,
      "tokens_seen": 895746048
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004809769245318706,
      "loss": 2.8163,
      "theoretical_loss": 3.6879970462330975,
      "tokens_seen": 895811584
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004809603745034991,
      "loss": 2.5786,
      "theoretical_loss": 3.687970846010063,
      "tokens_seen": 895877120
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048094382447512766,
      "loss": 2.7082,
      "theoretical_loss": 3.6879446482401885,
      "tokens_seen": 895942656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004809272744467562,
      "loss": 2.7949,
      "theoretical_loss": 3.6879184529230673,
      "tokens_seen": 896008192
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004809107244183847,
      "loss": 2.6158,
      "theoretical_loss": 3.687892260058289,
      "tokens_seen": 896073728
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004808941743900132,
      "loss": 2.5123,
      "theoretical_loss": 3.6878660696454446,
      "tokens_seen": 896139264
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048087762436164176,
      "loss": 2.6193,
      "theoretical_loss": 3.6878398816841265,
      "tokens_seen": 896204800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004808610743332703,
      "loss": 2.5377,
      "theoretical_loss": 3.6878136961739245,
      "tokens_seen": 896270336
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004808445243048988,
      "loss": 2.8099,
      "theoretical_loss": 3.687787513114431,
      "tokens_seen": 896335872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004808279742765273,
      "loss": 2.7343,
      "theoretical_loss": 3.6877613325052367,
      "tokens_seen": 896401408
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048081142424815585,
      "loss": 2.8561,
      "theoretical_loss": 3.687735154345934,
      "tokens_seen": 896466944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004807948742197844,
      "loss": 2.7733,
      "theoretical_loss": 3.687708978636114,
      "tokens_seen": 896532480
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004807783241914129,
      "loss": 2.5723,
      "theoretical_loss": 3.6876828053753687,
      "tokens_seen": 896598016
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004807617741630414,
      "loss": 2.6905,
      "theoretical_loss": 3.68765663456329,
      "tokens_seen": 896663552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048074522413466995,
      "loss": 2.7707,
      "theoretical_loss": 3.6876304661994697,
      "tokens_seen": 896729088
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004807286741062985,
      "loss": 2.7705,
      "theoretical_loss": 3.6876043002834997,
      "tokens_seen": 896794624
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048071212407792697,
      "loss": 2.9555,
      "theoretical_loss": 3.687578136814973,
      "tokens_seen": 896860160
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004806955740495555,
      "loss": 2.6996,
      "theoretical_loss": 3.6875519757934807,
      "tokens_seen": 896925696
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048067902402118405,
      "loss": 2.7213,
      "theoretical_loss": 3.6875258172186163,
      "tokens_seen": 896991232
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004806624739928126,
      "loss": 3.0353,
      "theoretical_loss": 3.687499661089971,
      "tokens_seen": 897056768
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048064592396444107,
      "loss": 2.6013,
      "theoretical_loss": 3.6874735074071383,
      "tokens_seen": 897122304
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004806293739360696,
      "loss": 2.5667,
      "theoretical_loss": 3.687447356169711,
      "tokens_seen": 897187840
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048061282390769814,
      "loss": 2.7204,
      "theoretical_loss": 3.687421207377281,
      "tokens_seen": 897253376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004805962738793267,
      "loss": 2.904,
      "theoretical_loss": 3.6873950610294415,
      "tokens_seen": 897318912
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048057972385095516,
      "loss": 2.658,
      "theoretical_loss": 3.687368917125786,
      "tokens_seen": 897384448
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004805631738225837,
      "loss": 2.8589,
      "theoretical_loss": 3.6873427756659067,
      "tokens_seen": 897449984
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048054662379421224,
      "loss": 2.7455,
      "theoretical_loss": 3.6873166366493972,
      "tokens_seen": 897515520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004805300737658408,
      "loss": 2.5058,
      "theoretical_loss": 3.687290500075851,
      "tokens_seen": 897581056
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048051352373746926,
      "loss": 2.7604,
      "theoretical_loss": 3.687264365944861,
      "tokens_seen": 897646592
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004804969737090978,
      "loss": 2.7146,
      "theoretical_loss": 3.6872382342560206,
      "tokens_seen": 897712128
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048048042368072633,
      "loss": 2.7701,
      "theoretical_loss": 3.687212105008923,
      "tokens_seen": 897777664
    },
    {
      "epoch": 0.32,
      "objective/train/avg_token_score": 0.016011035069823265,
      "objective/train/avg_weight": 0.9811726212501526,
      "objective/train/docs_used": 515235,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.773860454559326,
      "objective/train/std_weight": 0.07490324974060059,
      "objective/train/theoretical_loss": 3.6871859782031624,
      "objective/train/tokens_used": 918303200,
      "theoretical_loss": 3.6871859782031624,
      "tokens_seen": 897843200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048046387365235487,
      "loss": 2.8225,
      "theoretical_loss": 3.6871859782031624,
      "tokens_seen": 897843200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048044732362398335,
      "loss": 2.8788,
      "theoretical_loss": 3.687159853838333,
      "tokens_seen": 897908736
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004804307735956119,
      "loss": 2.8328,
      "theoretical_loss": 3.6871337319140274,
      "tokens_seen": 897974272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048041422356724043,
      "loss": 2.9266,
      "theoretical_loss": 3.6871076124298403,
      "tokens_seen": 898039808
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048039767353886897,
      "loss": 2.7443,
      "theoretical_loss": 3.6870814953853657,
      "tokens_seen": 898105344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048038112351049745,
      "loss": 2.8561,
      "theoretical_loss": 3.6870553807801976,
      "tokens_seen": 898170880
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000480364573482126,
      "loss": 2.7218,
      "theoretical_loss": 3.68702926861393,
      "tokens_seen": 898236416
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004803480234537545,
      "loss": 2.6828,
      "theoretical_loss": 3.687003158886158,
      "tokens_seen": 898301952
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048033147342538306,
      "loss": 2.6878,
      "theoretical_loss": 3.6869770515964744,
      "tokens_seen": 898367488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048031492339701155,
      "loss": 2.8187,
      "theoretical_loss": 3.6869509467444757,
      "tokens_seen": 898433024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004802983733686401,
      "loss": 2.6338,
      "theoretical_loss": 3.686924844329755,
      "tokens_seen": 898498560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004802818233402686,
      "loss": 2.4053,
      "theoretical_loss": 3.686898744351907,
      "tokens_seen": 898564096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048026527331189705,
      "loss": 2.7323,
      "theoretical_loss": 3.6868726468105275,
      "tokens_seen": 898629632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004802487232835256,
      "loss": 2.6578,
      "theoretical_loss": 3.6868465517052105,
      "tokens_seen": 898695168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004802321732551541,
      "loss": 2.8277,
      "theoretical_loss": 3.686820459035552,
      "tokens_seen": 898760704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004802156232267826,
      "loss": 2.7819,
      "theoretical_loss": 3.686794368801146,
      "tokens_seen": 898826240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048019907319841115,
      "loss": 2.9763,
      "theoretical_loss": 3.686768281001588,
      "tokens_seen": 898891776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004801825231700397,
      "loss": 2.6194,
      "theoretical_loss": 3.6867421956364734,
      "tokens_seen": 898957312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004801659731416682,
      "loss": 2.7019,
      "theoretical_loss": 3.6867161127053976,
      "tokens_seen": 899022848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004801494231132967,
      "loss": 2.7916,
      "theoretical_loss": 3.686690032207956,
      "tokens_seen": 899088384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048013287308492524,
      "loss": 2.5209,
      "theoretical_loss": 3.6866639541437447,
      "tokens_seen": 899153920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004801163230565538,
      "loss": 2.8152,
      "theoretical_loss": 3.686637878512358,
      "tokens_seen": 899219456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004800997730281823,
      "loss": 2.8028,
      "theoretical_loss": 3.686611805313393,
      "tokens_seen": 899284992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004800832229998108,
      "loss": 2.7902,
      "theoretical_loss": 3.686585734546445,
      "tokens_seen": 899350528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048006667297143934,
      "loss": 2.5816,
      "theoretical_loss": 3.68655966621111,
      "tokens_seen": 899416064
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004800501229430679,
      "loss": 2.62,
      "theoretical_loss": 3.6865336003069844,
      "tokens_seen": 899481600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004800335729146964,
      "loss": 2.6882,
      "theoretical_loss": 3.6865075368336635,
      "tokens_seen": 899547136
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004800170228863249,
      "loss": 2.6908,
      "theoretical_loss": 3.6864814757907443,
      "tokens_seen": 899612672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00048000047285795343,
      "loss": 2.6879,
      "theoretical_loss": 3.686455417177823,
      "tokens_seen": 899678208
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047998392282958197,
      "loss": 2.894,
      "theoretical_loss": 3.6864293609944956,
      "tokens_seen": 899743744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004799673728012105,
      "loss": 2.702,
      "theoretical_loss": 3.6864033072403593,
      "tokens_seen": 899809280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047995082277283905,
      "loss": 2.8352,
      "theoretical_loss": 3.6863772559150103,
      "tokens_seen": 899874816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047993427274446753,
      "loss": 2.5477,
      "theoretical_loss": 3.6863512070180455,
      "tokens_seen": 899940352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047991772271609607,
      "loss": 2.7897,
      "theoretical_loss": 3.6863251605490612,
      "tokens_seen": 900005888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004799011726877246,
      "loss": 2.8511,
      "theoretical_loss": 3.686299116507655,
      "tokens_seen": 900071424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047988462265935314,
      "loss": 2.7972,
      "theoretical_loss": 3.686273074893424,
      "tokens_seen": 900136960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004798680726309816,
      "loss": 2.6015,
      "theoretical_loss": 3.6862470357059642,
      "tokens_seen": 900202496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047985152260261016,
      "loss": 2.7992,
      "theoretical_loss": 3.6862209989448744,
      "tokens_seen": 900268032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004798349725742387,
      "loss": 2.8537,
      "theoretical_loss": 3.6861949646097507,
      "tokens_seen": 900333568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047981842254586724,
      "loss": 2.8217,
      "theoretical_loss": 3.686168932700191,
      "tokens_seen": 900399104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004798018725174957,
      "loss": 2.8359,
      "theoretical_loss": 3.6861429032157926,
      "tokens_seen": 900464640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047978532248912426,
      "loss": 2.6036,
      "theoretical_loss": 3.686116876156153,
      "tokens_seen": 900530176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004797687724607528,
      "loss": 2.744,
      "theoretical_loss": 3.6860908515208703,
      "tokens_seen": 900595712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047975222243238134,
      "loss": 2.8244,
      "theoretical_loss": 3.686064829309542,
      "tokens_seen": 900661248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004797356724040098,
      "loss": 2.6484,
      "theoretical_loss": 3.686038809521766,
      "tokens_seen": 900726784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047971912237563836,
      "loss": 2.8345,
      "theoretical_loss": 3.6860127921571406,
      "tokens_seen": 900792320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004797025723472669,
      "loss": 2.745,
      "theoretical_loss": 3.685986777215263,
      "tokens_seen": 900857856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047968602231889543,
      "loss": 2.6879,
      "theoretical_loss": 3.685960764695732,
      "tokens_seen": 900923392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004796694722905239,
      "loss": 2.7295,
      "theoretical_loss": 3.6859347545981462,
      "tokens_seen": 900988928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047965292226215245,
      "loss": 2.5069,
      "theoretical_loss": 3.6859087469221032,
      "tokens_seen": 901054464
    },
    {
      "epoch": 0.32,
      "objective/train/avg_token_score": 0.01596631482243538,
      "objective/train/avg_weight": 0.9812052845954895,
      "objective/train/docs_used": 516962,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9867377281188965,
      "objective/train/std_weight": 0.08526885509490967,
      "objective/train/theoretical_loss": 3.685882741667202,
      "objective/train/tokens_used": 921580000,
      "theoretical_loss": 3.685882741667202,
      "tokens_seen": 901120000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000479636372233781,
      "loss": 2.856,
      "theoretical_loss": 3.685882741667202,
      "tokens_seen": 901120000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004796198222054095,
      "loss": 2.6459,
      "theoretical_loss": 3.6858567388330408,
      "tokens_seen": 901185536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000479603272177038,
      "loss": 2.5389,
      "theoretical_loss": 3.6858307384192184,
      "tokens_seen": 901251072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047958672214866655,
      "loss": 2.5933,
      "theoretical_loss": 3.6858047404253336,
      "tokens_seen": 901316608
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004795701721202951,
      "loss": 2.7896,
      "theoretical_loss": 3.6857787448509853,
      "tokens_seen": 901382144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004795536220919236,
      "loss": 2.4952,
      "theoretical_loss": 3.685752751695772,
      "tokens_seen": 901447680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004795370720635521,
      "loss": 2.7286,
      "theoretical_loss": 3.685726760959293,
      "tokens_seen": 901513216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047952052203518064,
      "loss": 2.7994,
      "theoretical_loss": 3.6857007726411477,
      "tokens_seen": 901578752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004795039720068092,
      "loss": 2.9393,
      "theoretical_loss": 3.685674786740935,
      "tokens_seen": 901644288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004794874219784377,
      "loss": 2.9454,
      "theoretical_loss": 3.6856488032582546,
      "tokens_seen": 901709824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004794708719500662,
      "loss": 2.7088,
      "theoretical_loss": 3.6856228221927054,
      "tokens_seen": 901775360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047945432192169474,
      "loss": 2.6468,
      "theoretical_loss": 3.6855968435438866,
      "tokens_seen": 901840896
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004794377718933233,
      "loss": 2.601,
      "theoretical_loss": 3.685570867311399,
      "tokens_seen": 901906432
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004794212218649518,
      "loss": 2.9741,
      "theoretical_loss": 3.685544893494841,
      "tokens_seen": 901971968
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004794046718365803,
      "loss": 2.74,
      "theoretical_loss": 3.6855189220938134,
      "tokens_seen": 902037504
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047938812180820884,
      "loss": 2.6276,
      "theoretical_loss": 3.6854929531079157,
      "tokens_seen": 902103040
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004793715717798374,
      "loss": 2.8122,
      "theoretical_loss": 3.6854669865367473,
      "tokens_seen": 902168576
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004793550217514659,
      "loss": 2.607,
      "theoretical_loss": 3.6854410223799094,
      "tokens_seen": 902234112
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004793384717230944,
      "loss": 2.8905,
      "theoretical_loss": 3.685415060637002,
      "tokens_seen": 902299648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047932192169472293,
      "loss": 2.8465,
      "theoretical_loss": 3.6853891013076243,
      "tokens_seen": 902365184
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047930537166635147,
      "loss": 2.9985,
      "theoretical_loss": 3.6853631443913772,
      "tokens_seen": 902430720
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047928882163798,
      "loss": 2.7668,
      "theoretical_loss": 3.6853371898878615,
      "tokens_seen": 902496256
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047927227160960844,
      "loss": 2.6868,
      "theoretical_loss": 3.685311237796678,
      "tokens_seen": 902561792
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000479255721581237,
      "loss": 2.6229,
      "theoretical_loss": 3.6852852881174263,
      "tokens_seen": 902627328
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047923917155286546,
      "loss": 2.652,
      "theoretical_loss": 3.685259340849708,
      "tokens_seen": 902692864
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000479222621524494,
      "loss": 2.8747,
      "theoretical_loss": 3.685233395993124,
      "tokens_seen": 902758400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047920607149612253,
      "loss": 2.638,
      "theoretical_loss": 3.6852074535472745,
      "tokens_seen": 902823936
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047918952146775107,
      "loss": 2.4473,
      "theoretical_loss": 3.685181513511761,
      "tokens_seen": 902889472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047917297143937955,
      "loss": 2.685,
      "theoretical_loss": 3.685155575886185,
      "tokens_seen": 902955008
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004791564214110081,
      "loss": 2.7249,
      "theoretical_loss": 3.6851296406701466,
      "tokens_seen": 903020544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047913987138263663,
      "loss": 2.8706,
      "theoretical_loss": 3.6851037078632483,
      "tokens_seen": 903086080
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047912332135426517,
      "loss": 2.7493,
      "theoretical_loss": 3.685077777465091,
      "tokens_seen": 903151616
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047910677132589365,
      "loss": 2.8969,
      "theoretical_loss": 3.685051849475276,
      "tokens_seen": 903217152
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004790902212975222,
      "loss": 2.635,
      "theoretical_loss": 3.6850259238934053,
      "tokens_seen": 903282688
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004790736712691507,
      "loss": 2.5858,
      "theoretical_loss": 3.6850000007190804,
      "tokens_seen": 903348224
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047905712124077926,
      "loss": 2.7602,
      "theoretical_loss": 3.684974079951903,
      "tokens_seen": 903413760
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047904057121240774,
      "loss": 2.6911,
      "theoretical_loss": 3.684948161591475,
      "tokens_seen": 903479296
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004790240211840363,
      "loss": 2.4928,
      "theoretical_loss": 3.6849222456373987,
      "tokens_seen": 903544832
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004790074711556648,
      "loss": 2.9184,
      "theoretical_loss": 3.684896332089276,
      "tokens_seen": 903610368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047899092112729336,
      "loss": 2.6383,
      "theoretical_loss": 3.6848704209467087,
      "tokens_seen": 903675904
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047897437109892184,
      "loss": 2.8263,
      "theoretical_loss": 3.6848445122093,
      "tokens_seen": 903741440
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004789578210705504,
      "loss": 2.757,
      "theoretical_loss": 3.684818605876651,
      "tokens_seen": 903806976
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004789412710421789,
      "loss": 2.5882,
      "theoretical_loss": 3.6847927019483646,
      "tokens_seen": 903872512
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047892472101380745,
      "loss": 2.6544,
      "theoretical_loss": 3.6847668004240437,
      "tokens_seen": 903938048
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047890817098543594,
      "loss": 2.6653,
      "theoretical_loss": 3.6847409013032912,
      "tokens_seen": 904003584
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004788916209570645,
      "loss": 2.8643,
      "theoretical_loss": 3.684715004585709,
      "tokens_seen": 904069120
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000478875070928693,
      "loss": 2.7584,
      "theoretical_loss": 3.6846891102709005,
      "tokens_seen": 904134656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047885852090032155,
      "loss": 2.7651,
      "theoretical_loss": 3.684663218358468,
      "tokens_seen": 904200192
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047884197087195003,
      "loss": 2.6981,
      "theoretical_loss": 3.684637328848016,
      "tokens_seen": 904265728
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047882542084357857,
      "loss": 2.7804,
      "theoretical_loss": 3.684611441739145,
      "tokens_seen": 904331264
    },
    {
      "epoch": 0.32,
      "objective/train/avg_token_score": 0.008878730237483978,
      "objective/train/avg_weight": 0.9895593523979187,
      "objective/train/docs_used": 518997,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.745586633682251,
      "objective/train/std_weight": 0.059738609939813614,
      "objective/train/theoretical_loss": 3.684585557031461,
      "objective/train/tokens_used": 924856800,
      "theoretical_loss": 3.684585557031461,
      "tokens_seen": 904396800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004788088708152071,
      "loss": 2.536,
      "theoretical_loss": 3.684585557031461,
      "tokens_seen": 904396800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047879232078683565,
      "loss": 2.7722,
      "theoretical_loss": 3.684559674724566,
      "tokens_seen": 904462336
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047877577075846413,
      "loss": 2.9113,
      "theoretical_loss": 3.684533794818063,
      "tokens_seen": 904527872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047875922073009267,
      "loss": 2.4287,
      "theoretical_loss": 3.6845079173115565,
      "tokens_seen": 904593408
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004787426707017212,
      "loss": 2.6826,
      "theoretical_loss": 3.684482042204649,
      "tokens_seen": 904658944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047872612067334974,
      "loss": 2.8718,
      "theoretical_loss": 3.6844561694969453,
      "tokens_seen": 904724480
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004787095706449782,
      "loss": 2.9651,
      "theoretical_loss": 3.6844302991880484,
      "tokens_seen": 904790016
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047869302061660676,
      "loss": 2.7142,
      "theoretical_loss": 3.684404431277562,
      "tokens_seen": 904855552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004786764705882353,
      "loss": 2.7501,
      "theoretical_loss": 3.6843785657650905,
      "tokens_seen": 904921088
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047865992055986384,
      "loss": 2.801,
      "theoretical_loss": 3.684352702650238,
      "tokens_seen": 904986624
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004786433705314923,
      "loss": 2.6907,
      "theoretical_loss": 3.684326841932609,
      "tokens_seen": 905052160
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047862682050312086,
      "loss": 2.7047,
      "theoretical_loss": 3.684300983611807,
      "tokens_seen": 905117696
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004786102704747494,
      "loss": 2.7737,
      "theoretical_loss": 3.684275127687436,
      "tokens_seen": 905183232
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047859372044637793,
      "loss": 2.3329,
      "theoretical_loss": 3.684249274159102,
      "tokens_seen": 905248768
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004785771704180064,
      "loss": 2.5775,
      "theoretical_loss": 3.684223423026408,
      "tokens_seen": 905314304
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047856062038963495,
      "loss": 2.5857,
      "theoretical_loss": 3.6841975742889588,
      "tokens_seen": 905379840
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004785440703612635,
      "loss": 2.8257,
      "theoretical_loss": 3.68417172794636,
      "tokens_seen": 905445376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047852752033289203,
      "loss": 2.8408,
      "theoretical_loss": 3.684145883998216,
      "tokens_seen": 905510912
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004785109703045205,
      "loss": 2.7793,
      "theoretical_loss": 3.684120042444132,
      "tokens_seen": 905576448
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047849442027614905,
      "loss": 2.9225,
      "theoretical_loss": 3.6840942032837116,
      "tokens_seen": 905641984
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004784778702477776,
      "loss": 2.7993,
      "theoretical_loss": 3.6840683665165614,
      "tokens_seen": 905707520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004784613202194061,
      "loss": 2.7558,
      "theoretical_loss": 3.6840425321422856,
      "tokens_seen": 905773056
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047844477019103466,
      "loss": 2.6949,
      "theoretical_loss": 3.6840167001604907,
      "tokens_seen": 905838592
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047842822016266315,
      "loss": 2.7485,
      "theoretical_loss": 3.683990870570781,
      "tokens_seen": 905904128
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004784116701342917,
      "loss": 2.8201,
      "theoretical_loss": 3.683965043372762,
      "tokens_seen": 905969664
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004783951201059202,
      "loss": 2.6539,
      "theoretical_loss": 3.6839392185660396,
      "tokens_seen": 906035200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047837857007754876,
      "loss": 2.1147,
      "theoretical_loss": 3.6839133961502197,
      "tokens_seen": 906100736
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047836202004917724,
      "loss": 2.9974,
      "theoretical_loss": 3.6838875761249072,
      "tokens_seen": 906166272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004783454700208058,
      "loss": 2.7451,
      "theoretical_loss": 3.6838617584897086,
      "tokens_seen": 906231808
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004783289199924343,
      "loss": 2.8398,
      "theoretical_loss": 3.68383594324423,
      "tokens_seen": 906297344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047831236996406285,
      "loss": 2.6689,
      "theoretical_loss": 3.6838101303880766,
      "tokens_seen": 906362880
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004782958199356913,
      "loss": 2.9122,
      "theoretical_loss": 3.6837843199208558,
      "tokens_seen": 906428416
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004782792699073198,
      "loss": 2.7587,
      "theoretical_loss": 3.6837585118421723,
      "tokens_seen": 906493952
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004782627198789483,
      "loss": 2.8357,
      "theoretical_loss": 3.683732706151633,
      "tokens_seen": 906559488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047824616985057684,
      "loss": 2.9677,
      "theoretical_loss": 3.6837069028488454,
      "tokens_seen": 906625024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004782296198222054,
      "loss": 2.6694,
      "theoretical_loss": 3.683681101933414,
      "tokens_seen": 906690560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004782130697938339,
      "loss": 2.6515,
      "theoretical_loss": 3.683655303404947,
      "tokens_seen": 906756096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004781965197654624,
      "loss": 2.7483,
      "theoretical_loss": 3.68362950726305,
      "tokens_seen": 906821632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047817996973709094,
      "loss": 2.6781,
      "theoretical_loss": 3.6836037135073307,
      "tokens_seen": 906887168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004781634197087195,
      "loss": 2.9093,
      "theoretical_loss": 3.6835779221373954,
      "tokens_seen": 906952704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000478146869680348,
      "loss": 2.78,
      "theoretical_loss": 3.683552133152851,
      "tokens_seen": 907018240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004781303196519765,
      "loss": 2.5715,
      "theoretical_loss": 3.683526346553305,
      "tokens_seen": 907083776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047811376962360503,
      "loss": 2.7705,
      "theoretical_loss": 3.683500562338364,
      "tokens_seen": 907149312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047809721959523357,
      "loss": 2.7754,
      "theoretical_loss": 3.6834747805076353,
      "tokens_seen": 907214848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004780806695668621,
      "loss": 2.6959,
      "theoretical_loss": 3.6834490010607266,
      "tokens_seen": 907280384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004780641195384906,
      "loss": 2.9118,
      "theoretical_loss": 3.683423223997245,
      "tokens_seen": 907345920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047804756951011913,
      "loss": 2.7376,
      "theoretical_loss": 3.6833974493167982,
      "tokens_seen": 907411456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047803101948174767,
      "loss": 2.7222,
      "theoretical_loss": 3.683371677018994,
      "tokens_seen": 907476992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004780144694533762,
      "loss": 2.7162,
      "theoretical_loss": 3.6833459071034396,
      "tokens_seen": 907542528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004779979194250047,
      "loss": 2.8864,
      "theoretical_loss": 3.6833201395697435,
      "tokens_seen": 907608064
    },
    {
      "epoch": 0.32,
      "objective/train/avg_token_score": 0.01694336161017418,
      "objective/train/avg_weight": 0.9800611138343811,
      "objective/train/docs_used": 520740,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.664320945739746,
      "objective/train/std_weight": 0.07448679208755493,
      "objective/train/theoretical_loss": 3.6832943744175126,
      "objective/train/tokens_used": 928133600,
      "theoretical_loss": 3.6832943744175126,
      "tokens_seen": 907673600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004779813693966332,
      "loss": 2.463,
      "theoretical_loss": 3.6832943744175126,
      "tokens_seen": 907673600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047796481936826176,
      "loss": 2.7095,
      "theoretical_loss": 3.6832686116463558,
      "tokens_seen": 907739136
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004779482693398903,
      "loss": 2.5222,
      "theoretical_loss": 3.683242851255881,
      "tokens_seen": 907804672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004779317193115188,
      "loss": 2.7643,
      "theoretical_loss": 3.6832170932456956,
      "tokens_seen": 907870208
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004779151692831473,
      "loss": 2.5586,
      "theoretical_loss": 3.6831913376154084,
      "tokens_seen": 907935744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047789861925477586,
      "loss": 2.9325,
      "theoretical_loss": 3.683165584364628,
      "tokens_seen": 908001280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004778820692264044,
      "loss": 2.7819,
      "theoretical_loss": 3.6831398334929624,
      "tokens_seen": 908066816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004778655191980329,
      "loss": 2.6556,
      "theoretical_loss": 3.6831140850000206,
      "tokens_seen": 908132352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004778489691696614,
      "loss": 2.6961,
      "theoretical_loss": 3.683088338885411,
      "tokens_seen": 908197888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047783241914128996,
      "loss": 2.662,
      "theoretical_loss": 3.6830625951487423,
      "tokens_seen": 908263424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004778158691129185,
      "loss": 2.6581,
      "theoretical_loss": 3.6830368537896225,
      "tokens_seen": 908328960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000477799319084547,
      "loss": 2.8576,
      "theoretical_loss": 3.6830111148076625,
      "tokens_seen": 908394496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004777827690561755,
      "loss": 2.8525,
      "theoretical_loss": 3.6829853782024697,
      "tokens_seen": 908460032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047776621902780405,
      "loss": 2.706,
      "theoretical_loss": 3.6829596439736534,
      "tokens_seen": 908525568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004777496689994326,
      "loss": 2.7053,
      "theoretical_loss": 3.6829339121208227,
      "tokens_seen": 908591104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047773311897106107,
      "loss": 2.5044,
      "theoretical_loss": 3.6829081826435877,
      "tokens_seen": 908656640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004777165689426896,
      "loss": 2.8394,
      "theoretical_loss": 3.682882455541557,
      "tokens_seen": 908722176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047770001891431815,
      "loss": 2.6934,
      "theoretical_loss": 3.6828567308143403,
      "tokens_seen": 908787712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004776834688859467,
      "loss": 2.8624,
      "theoretical_loss": 3.682831008461547,
      "tokens_seen": 908853248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047766691885757517,
      "loss": 2.7941,
      "theoretical_loss": 3.682805288482787,
      "tokens_seen": 908918784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004776503688292037,
      "loss": 2.5627,
      "theoretical_loss": 3.6827795708776705,
      "tokens_seen": 908984320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047763381880083224,
      "loss": 2.644,
      "theoretical_loss": 3.6827538556458057,
      "tokens_seen": 909049856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004776172687724608,
      "loss": 2.7707,
      "theoretical_loss": 3.6827281427868046,
      "tokens_seen": 909115392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047760071874408926,
      "loss": 2.6549,
      "theoretical_loss": 3.6827024323002755,
      "tokens_seen": 909180928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004775841687157178,
      "loss": 2.6044,
      "theoretical_loss": 3.6826767241858294,
      "tokens_seen": 909246464
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047756761868734634,
      "loss": 2.7816,
      "theoretical_loss": 3.6826510184430763,
      "tokens_seen": 909312000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004775510686589749,
      "loss": 2.5706,
      "theoretical_loss": 3.6826253150716264,
      "tokens_seen": 909377536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047753451863060336,
      "loss": 2.7869,
      "theoretical_loss": 3.6825996140710906,
      "tokens_seen": 909443072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0004775179686022319,
      "loss": 2.7779,
      "theoretical_loss": 3.682573915441078,
      "tokens_seen": 909508608
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047750141857386044,
      "loss": 2.7165,
      "theoretical_loss": 3.682548219181201,
      "tokens_seen": 909574144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047748486854548897,
      "loss": 2.8943,
      "theoretical_loss": 3.682522525291069,
      "tokens_seen": 909639680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047746831851711746,
      "loss": 2.6378,
      "theoretical_loss": 3.6824968337702932,
      "tokens_seen": 909705216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000477451768488746,
      "loss": 2.7873,
      "theoretical_loss": 3.6824711446184843,
      "tokens_seen": 909770752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047743521846037453,
      "loss": 2.9621,
      "theoretical_loss": 3.6824454578352537,
      "tokens_seen": 909836288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047741866843200307,
      "loss": 2.6602,
      "theoretical_loss": 3.6824197734202118,
      "tokens_seen": 909901824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00047740211840363155,
      "loss": 2.7286,
      "theoretical_loss": 3.6823940913729696,
      "tokens_seen": 909967360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004773855683752601,
      "loss": 2.7855,
      "theoretical_loss": 3.6823684116931386,
      "tokens_seen": 910032896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047736901834688863,
      "loss": 2.9147,
      "theoretical_loss": 3.682342734380331,
      "tokens_seen": 910098432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047735246831851716,
      "loss": 2.7442,
      "theoretical_loss": 3.6823170594341565,
      "tokens_seen": 910163968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047733591829014565,
      "loss": 2.7434,
      "theoretical_loss": 3.682291386854228,
      "tokens_seen": 910229504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004773193682617742,
      "loss": 2.5346,
      "theoretical_loss": 3.6822657166401562,
      "tokens_seen": 910295040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047730281823340267,
      "loss": 2.8006,
      "theoretical_loss": 3.682240048791553,
      "tokens_seen": 910360576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047728626820503115,
      "loss": 2.5558,
      "theoretical_loss": 3.6822143833080307,
      "tokens_seen": 910426112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004772697181766597,
      "loss": 2.6502,
      "theoretical_loss": 3.6821887201892007,
      "tokens_seen": 910491648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047725316814828823,
      "loss": 2.8632,
      "theoretical_loss": 3.682163059434674,
      "tokens_seen": 910557184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047723661811991676,
      "loss": 2.3929,
      "theoretical_loss": 3.6821374010440646,
      "tokens_seen": 910622720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047722006809154525,
      "loss": 2.998,
      "theoretical_loss": 3.6821117450169836,
      "tokens_seen": 910688256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004772035180631738,
      "loss": 2.7928,
      "theoretical_loss": 3.682086091353043,
      "tokens_seen": 910753792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004771869680348023,
      "loss": 2.7413,
      "theoretical_loss": 3.6820604400518553,
      "tokens_seen": 910819328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047717041800643086,
      "loss": 2.5708,
      "theoretical_loss": 3.6820347911130327,
      "tokens_seen": 910884864
    },
    {
      "epoch": 0.33,
      "objective/train/avg_token_score": 0.011827781796455383,
      "objective/train/avg_weight": 0.9860789775848389,
      "objective/train/docs_used": 522796,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.45194673538208,
      "objective/train/std_weight": 0.0444519966840744,
      "objective/train/theoretical_loss": 3.682009144536188,
      "objective/train/tokens_used": 931410400,
      "theoretical_loss": 3.682009144536188,
      "tokens_seen": 910950400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047715386797805934,
      "loss": 2.5558,
      "theoretical_loss": 3.682009144536188,
      "tokens_seen": 910950400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004771373179496879,
      "loss": 2.8766,
      "theoretical_loss": 3.6819835003209342,
      "tokens_seen": 911015936
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004771207679213164,
      "loss": 2.7899,
      "theoretical_loss": 3.681957858466883,
      "tokens_seen": 911081472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047710421789294496,
      "loss": 2.7551,
      "theoretical_loss": 3.681932218973648,
      "tokens_seen": 911147008
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047708766786457344,
      "loss": 2.8486,
      "theoretical_loss": 3.681906581840842,
      "tokens_seen": 911212544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000477071117836202,
      "loss": 2.6568,
      "theoretical_loss": 3.681880947068077,
      "tokens_seen": 911278080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004770545678078305,
      "loss": 2.7395,
      "theoretical_loss": 3.681855314654967,
      "tokens_seen": 911343616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047703801777945905,
      "loss": 2.7784,
      "theoretical_loss": 3.681829684601125,
      "tokens_seen": 911409152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047702146775108754,
      "loss": 2.7488,
      "theoretical_loss": 3.681804056906164,
      "tokens_seen": 911474688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004770049177227161,
      "loss": 2.5481,
      "theoretical_loss": 3.6817784315696978,
      "tokens_seen": 911540224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004769883676943446,
      "loss": 2.6955,
      "theoretical_loss": 3.6817528085913382,
      "tokens_seen": 911605760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047697181766597315,
      "loss": 2.8005,
      "theoretical_loss": 3.681727187970701,
      "tokens_seen": 911671296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047695526763760163,
      "loss": 2.6579,
      "theoretical_loss": 3.6817015697073985,
      "tokens_seen": 911736832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047693871760923017,
      "loss": 2.8333,
      "theoretical_loss": 3.6816759538010446,
      "tokens_seen": 911802368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004769221675808587,
      "loss": 2.9337,
      "theoretical_loss": 3.6816503402512533,
      "tokens_seen": 911867904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047690561755248724,
      "loss": 2.781,
      "theoretical_loss": 3.6816247290576376,
      "tokens_seen": 911933440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047688906752411573,
      "loss": 2.6901,
      "theoretical_loss": 3.681599120219812,
      "tokens_seen": 911998976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047687251749574427,
      "loss": 2.9763,
      "theoretical_loss": 3.681573513737391,
      "tokens_seen": 912064512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004768559674673728,
      "loss": 2.8607,
      "theoretical_loss": 3.6815479096099883,
      "tokens_seen": 912130048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047683941743900134,
      "loss": 2.5449,
      "theoretical_loss": 3.681522307837218,
      "tokens_seen": 912195584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004768228674106298,
      "loss": 2.6881,
      "theoretical_loss": 3.6814967084186945,
      "tokens_seen": 912261120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047680631738225836,
      "loss": 2.8464,
      "theoretical_loss": 3.681471111354033,
      "tokens_seen": 912326656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004767897673538869,
      "loss": 2.4921,
      "theoretical_loss": 3.681445516642847,
      "tokens_seen": 912392192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047677321732551544,
      "loss": 2.7657,
      "theoretical_loss": 3.681419924284751,
      "tokens_seen": 912457728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004767566672971439,
      "loss": 2.4759,
      "theoretical_loss": 3.68139433427936,
      "tokens_seen": 912523264
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047674011726877246,
      "loss": 2.5482,
      "theoretical_loss": 3.681368746626289,
      "tokens_seen": 912588800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000476723567240401,
      "loss": 2.7734,
      "theoretical_loss": 3.6813431613251533,
      "tokens_seen": 912654336
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047670701721202953,
      "loss": 2.7492,
      "theoretical_loss": 3.6813175783755665,
      "tokens_seen": 912719872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000476690467183658,
      "loss": 2.8122,
      "theoretical_loss": 3.6812919977771448,
      "tokens_seen": 912785408
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047667391715528655,
      "loss": 2.7218,
      "theoretical_loss": 3.6812664195295026,
      "tokens_seen": 912850944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004766573671269151,
      "loss": 2.7956,
      "theoretical_loss": 3.6812408436322555,
      "tokens_seen": 912916480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047664081709854363,
      "loss": 2.6491,
      "theoretical_loss": 3.681215270085019,
      "tokens_seen": 912982016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004766242670701721,
      "loss": 2.6733,
      "theoretical_loss": 3.681189698887408,
      "tokens_seen": 913047552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047660771704180065,
      "loss": 2.9049,
      "theoretical_loss": 3.681164130039039,
      "tokens_seen": 913113088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004765911670134292,
      "loss": 2.7716,
      "theoretical_loss": 3.681138563539526,
      "tokens_seen": 913178624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004765746169850577,
      "loss": 2.6377,
      "theoretical_loss": 3.6811129993884864,
      "tokens_seen": 913244160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004765580669566862,
      "loss": 2.6836,
      "theoretical_loss": 3.6810874375855342,
      "tokens_seen": 913309696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047654151692831475,
      "loss": 2.76,
      "theoretical_loss": 3.6810618781302864,
      "tokens_seen": 913375232
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004765249668999433,
      "loss": 2.7303,
      "theoretical_loss": 3.681036321022359,
      "tokens_seen": 913440768
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004765084168715718,
      "loss": 2.7292,
      "theoretical_loss": 3.681010766261368,
      "tokens_seen": 913506304
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004764918668432003,
      "loss": 2.7884,
      "theoretical_loss": 3.680985213846929,
      "tokens_seen": 913571840
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047647531681482884,
      "loss": 2.9485,
      "theoretical_loss": 3.6809596637786584,
      "tokens_seen": 913637376
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004764587667864574,
      "loss": 2.7758,
      "theoretical_loss": 3.6809341160561724,
      "tokens_seen": 913702912
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004764422167580859,
      "loss": 2.508,
      "theoretical_loss": 3.6809085706790876,
      "tokens_seen": 913768448
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004764256667297144,
      "loss": 3.0182,
      "theoretical_loss": 3.6808830276470212,
      "tokens_seen": 913833984
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047640911670134294,
      "loss": 2.9329,
      "theoretical_loss": 3.6808574869595887,
      "tokens_seen": 913899520
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004763925666729715,
      "loss": 2.8724,
      "theoretical_loss": 3.6808319486164067,
      "tokens_seen": 913965056
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004763760166446,
      "loss": 2.6904,
      "theoretical_loss": 3.680806412617093,
      "tokens_seen": 914030592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004763594666162285,
      "loss": 2.6997,
      "theoretical_loss": 3.6807808789612633,
      "tokens_seen": 914096128
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047634291658785703,
      "loss": 2.687,
      "theoretical_loss": 3.680755347648536,
      "tokens_seen": 914161664
    },
    {
      "epoch": 0.33,
      "objective/train/avg_token_score": 0.019573185592889786,
      "objective/train/avg_weight": 0.9769989848136902,
      "objective/train/docs_used": 524442,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.20746111869812,
      "objective/train/std_weight": 0.08950918912887573,
      "objective/train/theoretical_loss": 3.680729818678526,
      "objective/train/tokens_used": 934687200,
      "theoretical_loss": 3.680729818678526,
      "tokens_seen": 914227200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047632636655948546,
      "loss": 2.8527,
      "theoretical_loss": 3.680729818678526,
      "tokens_seen": 914227200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000476309816531114,
      "loss": 2.7145,
      "theoretical_loss": 3.6807042920508533,
      "tokens_seen": 914292736
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047629326650274254,
      "loss": 2.437,
      "theoretical_loss": 3.6806787677651323,
      "tokens_seen": 914358272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004762767164743711,
      "loss": 2.8909,
      "theoretical_loss": 3.6806532458209817,
      "tokens_seen": 914423808
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047626016644599956,
      "loss": 2.6178,
      "theoretical_loss": 3.680627726218019,
      "tokens_seen": 914489344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004762436164176281,
      "loss": 2.5618,
      "theoretical_loss": 3.680602208955861,
      "tokens_seen": 914554880
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047622706638925663,
      "loss": 2.7706,
      "theoretical_loss": 3.6805766940341256,
      "tokens_seen": 914620416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047621051636088517,
      "loss": 2.7936,
      "theoretical_loss": 3.6805511814524308,
      "tokens_seen": 914685952
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047619396633251365,
      "loss": 2.5683,
      "theoretical_loss": 3.6805256712103938,
      "tokens_seen": 914751488
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004761774163041422,
      "loss": 2.5243,
      "theoretical_loss": 3.6805001633076326,
      "tokens_seen": 914817024
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047616086627577073,
      "loss": 2.8265,
      "theoretical_loss": 3.6804746577437655,
      "tokens_seen": 914882560
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047614431624739927,
      "loss": 2.5892,
      "theoretical_loss": 3.68044915451841,
      "tokens_seen": 914948096
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047612776621902775,
      "loss": 2.7871,
      "theoretical_loss": 3.680423653631184,
      "tokens_seen": 915013632
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004761112161906563,
      "loss": 2.8909,
      "theoretical_loss": 3.680398155081707,
      "tokens_seen": 915079168
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004760946661622848,
      "loss": 2.7674,
      "theoretical_loss": 3.680372658869596,
      "tokens_seen": 915144704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047607811613391336,
      "loss": 2.641,
      "theoretical_loss": 3.68034716499447,
      "tokens_seen": 915210240
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047606156610554185,
      "loss": 2.8044,
      "theoretical_loss": 3.680321673455947,
      "tokens_seen": 915275776
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004760450160771704,
      "loss": 2.7478,
      "theoretical_loss": 3.6802961842536464,
      "tokens_seen": 915341312
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004760284660487989,
      "loss": 2.7514,
      "theoretical_loss": 3.6802706973871855,
      "tokens_seen": 915406848
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047601191602042746,
      "loss": 2.8062,
      "theoretical_loss": 3.6802452128561844,
      "tokens_seen": 915472384
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000475995365992056,
      "loss": 2.7211,
      "theoretical_loss": 3.680219730660261,
      "tokens_seen": 915537920
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004759788159636845,
      "loss": 2.7426,
      "theoretical_loss": 3.6801942507990346,
      "tokens_seen": 915603456
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000475962265935313,
      "loss": 2.8169,
      "theoretical_loss": 3.680168773272124,
      "tokens_seen": 915668992
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047594571590694155,
      "loss": 3.0684,
      "theoretical_loss": 3.680143298079149,
      "tokens_seen": 915734528
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004759291658785701,
      "loss": 2.5344,
      "theoretical_loss": 3.680117825219728,
      "tokens_seen": 915800064
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004759126158501986,
      "loss": 2.9828,
      "theoretical_loss": 3.6800923546934805,
      "tokens_seen": 915865600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004758960658218271,
      "loss": 2.9511,
      "theoretical_loss": 3.680066886500026,
      "tokens_seen": 915931136
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047587951579345565,
      "loss": 2.6884,
      "theoretical_loss": 3.6800414206389833,
      "tokens_seen": 915996672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004758629657650842,
      "loss": 2.6917,
      "theoretical_loss": 3.6800159571099735,
      "tokens_seen": 916062208
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047584641573671267,
      "loss": 2.9443,
      "theoretical_loss": 3.679990495912614,
      "tokens_seen": 916127744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004758298657083412,
      "loss": 2.7799,
      "theoretical_loss": 3.679965037046527,
      "tokens_seen": 916193280
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047581331567996975,
      "loss": 2.8559,
      "theoretical_loss": 3.67993958051133,
      "tokens_seen": 916258816
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004757967656515983,
      "loss": 2.9227,
      "theoretical_loss": 3.6799141263066444,
      "tokens_seen": 916324352
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047578021562322677,
      "loss": 2.838,
      "theoretical_loss": 3.67988867443209,
      "tokens_seen": 916389888
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004757636655948553,
      "loss": 2.7131,
      "theoretical_loss": 3.6798632248872862,
      "tokens_seen": 916455424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047574711556648384,
      "loss": 2.926,
      "theoretical_loss": 3.679837777671854,
      "tokens_seen": 916520960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004757305655381124,
      "loss": 2.5229,
      "theoretical_loss": 3.6798123327854126,
      "tokens_seen": 916586496
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047571401550974086,
      "loss": 2.5755,
      "theoretical_loss": 3.679786890227583,
      "tokens_seen": 916652032
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004756974654813694,
      "loss": 2.6937,
      "theoretical_loss": 3.6797614499979865,
      "tokens_seen": 916717568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047568091545299794,
      "loss": 2.5584,
      "theoretical_loss": 3.6797360120962423,
      "tokens_seen": 916783104
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004756643654246265,
      "loss": 2.6483,
      "theoretical_loss": 3.679710576521971,
      "tokens_seen": 916848640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047564781539625496,
      "loss": 2.6359,
      "theoretical_loss": 3.679685143274794,
      "tokens_seen": 916914176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004756312653678835,
      "loss": 2.688,
      "theoretical_loss": 3.679659712354332,
      "tokens_seen": 916979712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047561471533951203,
      "loss": 2.3389,
      "theoretical_loss": 3.6796342837602056,
      "tokens_seen": 917045248
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047559816531114057,
      "loss": 2.721,
      "theoretical_loss": 3.6796088574920356,
      "tokens_seen": 917110784
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047558161528276906,
      "loss": 2.7975,
      "theoretical_loss": 3.6795834335494444,
      "tokens_seen": 917176320
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004755650652543976,
      "loss": 2.8695,
      "theoretical_loss": 3.679558011932051,
      "tokens_seen": 917241856
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047554851522602613,
      "loss": 2.8573,
      "theoretical_loss": 3.679532592639478,
      "tokens_seen": 917307392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047553196519765467,
      "loss": 2.672,
      "theoretical_loss": 3.6795071756713464,
      "tokens_seen": 917372928
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047551541516928315,
      "loss": 2.5547,
      "theoretical_loss": 3.6794817610272776,
      "tokens_seen": 917438464
    },
    {
      "epoch": 0.33,
      "objective/train/avg_token_score": 0.004547202959656715,
      "objective/train/avg_weight": 0.9946557283401489,
      "objective/train/docs_used": 526356,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.621753215789795,
      "objective/train/std_weight": 0.032646842300891876,
      "objective/train/theoretical_loss": 3.6794563487068936,
      "objective/train/tokens_used": 937964000,
      "theoretical_loss": 3.6794563487068936,
      "tokens_seen": 917504000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004754988651409117,
      "loss": 2.6535,
      "theoretical_loss": 3.6794563487068936,
      "tokens_seen": 917504000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004754823151125402,
      "loss": 2.7974,
      "theoretical_loss": 3.679430938709815,
      "tokens_seen": 917569536
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047546576508416876,
      "loss": 2.8889,
      "theoretical_loss": 3.6794055310356644,
      "tokens_seen": 917635072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047544921505579725,
      "loss": 2.7067,
      "theoretical_loss": 3.6793801256840633,
      "tokens_seen": 917700608
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004754326650274258,
      "loss": 2.7505,
      "theoretical_loss": 3.679354722654633,
      "tokens_seen": 917766144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004754161149990543,
      "loss": 2.962,
      "theoretical_loss": 3.6793293219469962,
      "tokens_seen": 917831680
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047539956497068286,
      "loss": 2.8471,
      "theoretical_loss": 3.6793039235607745,
      "tokens_seen": 917897216
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047538301494231134,
      "loss": 2.6496,
      "theoretical_loss": 3.6792785274955904,
      "tokens_seen": 917962752
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004753664649139399,
      "loss": 3.0325,
      "theoretical_loss": 3.6792531337510654,
      "tokens_seen": 918028288
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004753499148855683,
      "loss": 2.9021,
      "theoretical_loss": 3.6792277423268223,
      "tokens_seen": 918093824
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047533336485719685,
      "loss": 2.8407,
      "theoretical_loss": 3.6792023532224833,
      "tokens_seen": 918159360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004753168148288254,
      "loss": 2.5799,
      "theoretical_loss": 3.6791769664376712,
      "tokens_seen": 918224896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004753002648004539,
      "loss": 3.0053,
      "theoretical_loss": 3.6791515819720084,
      "tokens_seen": 918290432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004752837147720824,
      "loss": 2.7692,
      "theoretical_loss": 3.679126199825118,
      "tokens_seen": 918355968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047526716474371094,
      "loss": 2.7692,
      "theoretical_loss": 3.6791008199966218,
      "tokens_seen": 918421504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004752506147153395,
      "loss": 2.7904,
      "theoretical_loss": 3.679075442486143,
      "tokens_seen": 918487040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000475234064686968,
      "loss": 2.8012,
      "theoretical_loss": 3.6790500672933044,
      "tokens_seen": 918552576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004752175146585965,
      "loss": 2.6841,
      "theoretical_loss": 3.6790246944177296,
      "tokens_seen": 918618112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047520096463022504,
      "loss": 2.8463,
      "theoretical_loss": 3.678999323859041,
      "tokens_seen": 918683648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004751844146018536,
      "loss": 2.7562,
      "theoretical_loss": 3.678973955616862,
      "tokens_seen": 918749184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004751678645734821,
      "loss": 2.5814,
      "theoretical_loss": 3.6789485896908163,
      "tokens_seen": 918814720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004751513145451106,
      "loss": 2.7408,
      "theoretical_loss": 3.678923226080527,
      "tokens_seen": 918880256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047513476451673914,
      "loss": 2.7195,
      "theoretical_loss": 3.6788978647856174,
      "tokens_seen": 918945792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004751182144883677,
      "loss": 2.7338,
      "theoretical_loss": 3.678872505805711,
      "tokens_seen": 919011328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004751016644599962,
      "loss": 2.7446,
      "theoretical_loss": 3.678847149140431,
      "tokens_seen": 919076864
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004750851144316247,
      "loss": 2.7861,
      "theoretical_loss": 3.678821794789402,
      "tokens_seen": 919142400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047506856440325323,
      "loss": 2.7155,
      "theoretical_loss": 3.6787964427522475,
      "tokens_seen": 919207936
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047505201437488177,
      "loss": 2.9236,
      "theoretical_loss": 3.678771093028591,
      "tokens_seen": 919273472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004750354643465103,
      "loss": 2.6561,
      "theoretical_loss": 3.678745745618057,
      "tokens_seen": 919339008
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004750189143181388,
      "loss": 2.6448,
      "theoretical_loss": 3.6787204005202687,
      "tokens_seen": 919404544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047500236428976733,
      "loss": 2.4705,
      "theoretical_loss": 3.678695057734852,
      "tokens_seen": 919470080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047498581426139587,
      "loss": 2.8269,
      "theoretical_loss": 3.678669717261429,
      "tokens_seen": 919535616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004749692642330244,
      "loss": 2.8077,
      "theoretical_loss": 3.6786443790996257,
      "tokens_seen": 919601152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004749527142046529,
      "loss": 2.4029,
      "theoretical_loss": 3.6786190432490655,
      "tokens_seen": 919666688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004749361641762814,
      "loss": 2.6857,
      "theoretical_loss": 3.6785937097093733,
      "tokens_seen": 919732224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047491961414790996,
      "loss": 2.7688,
      "theoretical_loss": 3.6785683784801737,
      "tokens_seen": 919797760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004749030641195385,
      "loss": 2.7488,
      "theoretical_loss": 3.678543049561091,
      "tokens_seen": 919863296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000474886514091167,
      "loss": 2.8694,
      "theoretical_loss": 3.6785177229517507,
      "tokens_seen": 919928832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004748699640627955,
      "loss": 2.8599,
      "theoretical_loss": 3.678492398651777,
      "tokens_seen": 919994368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047485341403442406,
      "loss": 2.7639,
      "theoretical_loss": 3.678467076660795,
      "tokens_seen": 920059904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004748368640060526,
      "loss": 2.8455,
      "theoretical_loss": 3.6784417569784296,
      "tokens_seen": 920125440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004748203139776811,
      "loss": 2.8454,
      "theoretical_loss": 3.6784164396043066,
      "tokens_seen": 920190976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004748037639493096,
      "loss": 2.7389,
      "theoretical_loss": 3.6783911245380505,
      "tokens_seen": 920256512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047478721392093815,
      "loss": 2.8401,
      "theoretical_loss": 3.678365811779287,
      "tokens_seen": 920322048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004747706638925667,
      "loss": 2.7892,
      "theoretical_loss": 3.6783405013276402,
      "tokens_seen": 920387584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004747541138641952,
      "loss": 2.7419,
      "theoretical_loss": 3.678315193182738,
      "tokens_seen": 920453120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004747375638358237,
      "loss": 2.6858,
      "theoretical_loss": 3.678289887344204,
      "tokens_seen": 920518656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047472101380745225,
      "loss": 2.8483,
      "theoretical_loss": 3.6782645838116643,
      "tokens_seen": 920584192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004747044637790808,
      "loss": 2.7893,
      "theoretical_loss": 3.6782392825847445,
      "tokens_seen": 920649728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047468791375070927,
      "loss": 2.7648,
      "theoretical_loss": 3.6782139836630714,
      "tokens_seen": 920715264
    },
    {
      "epoch": 0.33,
      "objective/train/avg_token_score": 0.01529812440276146,
      "objective/train/avg_weight": 0.9820030331611633,
      "objective/train/docs_used": 528450,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6673381328582764,
      "objective/train/std_weight": 0.08610226213932037,
      "objective/train/theoretical_loss": 3.6781886870462692,
      "objective/train/tokens_used": 941240800,
      "theoretical_loss": 3.6781886870462692,
      "tokens_seen": 920780800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004746713637223378,
      "loss": 2.572,
      "theoretical_loss": 3.6781886870462692,
      "tokens_seen": 920780800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047465481369396634,
      "loss": 2.7037,
      "theoretical_loss": 3.678163392733966,
      "tokens_seen": 920846336
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004746382636655949,
      "loss": 2.8479,
      "theoretical_loss": 3.6781381007257856,
      "tokens_seen": 920911872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047462171363722337,
      "loss": 2.7648,
      "theoretical_loss": 3.6781128110213555,
      "tokens_seen": 920977408
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004746051636088519,
      "loss": 2.8611,
      "theoretical_loss": 3.6780875236203023,
      "tokens_seen": 921042944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047458861358048044,
      "loss": 2.7294,
      "theoretical_loss": 3.6780622385222514,
      "tokens_seen": 921108480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000474572063552109,
      "loss": 2.9463,
      "theoretical_loss": 3.6780369557268298,
      "tokens_seen": 921174016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047455551352373746,
      "loss": 3.0509,
      "theoretical_loss": 3.6780116752336633,
      "tokens_seen": 921239552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000474538963495366,
      "loss": 2.7915,
      "theoretical_loss": 3.6779863970423796,
      "tokens_seen": 921305088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047452241346699454,
      "loss": 2.6227,
      "theoretical_loss": 3.677961121152604,
      "tokens_seen": 921370624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004745058634386231,
      "loss": 2.647,
      "theoretical_loss": 3.677935847563965,
      "tokens_seen": 921436160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047448931341025156,
      "loss": 2.6581,
      "theoretical_loss": 3.677910576276088,
      "tokens_seen": 921501696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004744727633818801,
      "loss": 2.6131,
      "theoretical_loss": 3.6778853072886006,
      "tokens_seen": 921567232
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047445621335350863,
      "loss": 2.9029,
      "theoretical_loss": 3.6778600406011295,
      "tokens_seen": 921632768
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047443966332513717,
      "loss": 2.7057,
      "theoretical_loss": 3.6778347762133023,
      "tokens_seen": 921698304
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004744231132967657,
      "loss": 2.7488,
      "theoretical_loss": 3.677809514124746,
      "tokens_seen": 921763840
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004744065632683942,
      "loss": 2.8659,
      "theoretical_loss": 3.677784254335087,
      "tokens_seen": 921829376
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047439001324002273,
      "loss": 2.7618,
      "theoretical_loss": 3.677758996843955,
      "tokens_seen": 921894912
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047437346321165127,
      "loss": 2.6686,
      "theoretical_loss": 3.6777337416509743,
      "tokens_seen": 921960448
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004743569131832797,
      "loss": 3.0249,
      "theoretical_loss": 3.6777084887557754,
      "tokens_seen": 922025984
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047434036315490823,
      "loss": 2.6577,
      "theoretical_loss": 3.677683238157984,
      "tokens_seen": 922091520
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047432381312653677,
      "loss": 2.7658,
      "theoretical_loss": 3.677657989857229,
      "tokens_seen": 922157056
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047430726309816525,
      "loss": 2.8654,
      "theoretical_loss": 3.6776327438531373,
      "tokens_seen": 922222592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004742907130697938,
      "loss": 2.6702,
      "theoretical_loss": 3.677607500145337,
      "tokens_seen": 922288128
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047427416304142233,
      "loss": 2.8764,
      "theoretical_loss": 3.6775822587334566,
      "tokens_seen": 922353664
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047425761301305087,
      "loss": 2.8045,
      "theoretical_loss": 3.677557019617124,
      "tokens_seen": 922419200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047424106298467935,
      "loss": 2.7452,
      "theoretical_loss": 3.6775317827959673,
      "tokens_seen": 922484736
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004742245129563079,
      "loss": 2.7039,
      "theoretical_loss": 3.6775065482696148,
      "tokens_seen": 922550272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004742079629279364,
      "loss": 2.7597,
      "theoretical_loss": 3.677481316037694,
      "tokens_seen": 922615808
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047419141289956496,
      "loss": 2.6369,
      "theoretical_loss": 3.6774560860998347,
      "tokens_seen": 922681344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047417486287119345,
      "loss": 2.9492,
      "theoretical_loss": 3.6774308584556645,
      "tokens_seen": 922746880
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000474158312842822,
      "loss": 2.7721,
      "theoretical_loss": 3.6774056331048124,
      "tokens_seen": 922812416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004741417628144505,
      "loss": 2.5478,
      "theoretical_loss": 3.677380410046907,
      "tokens_seen": 922877952
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047412521278607906,
      "loss": 2.8051,
      "theoretical_loss": 3.6773551892815766,
      "tokens_seen": 922943488
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047410866275770754,
      "loss": 2.6737,
      "theoretical_loss": 3.677329970808451,
      "tokens_seen": 923009024
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004740921127293361,
      "loss": 2.6986,
      "theoretical_loss": 3.6773047546271584,
      "tokens_seen": 923074560
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004740755627009646,
      "loss": 2.6959,
      "theoretical_loss": 3.677279540737328,
      "tokens_seen": 923140096
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047405901267259315,
      "loss": 2.8933,
      "theoretical_loss": 3.677254329138589,
      "tokens_seen": 923205632
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047404246264422164,
      "loss": 2.7301,
      "theoretical_loss": 3.67722911983057,
      "tokens_seen": 923271168
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004740259126158502,
      "loss": 2.7659,
      "theoretical_loss": 3.6772039128129013,
      "tokens_seen": 923336704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004740093625874787,
      "loss": 2.8769,
      "theoretical_loss": 3.6771787080852123,
      "tokens_seen": 923402240
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047399281255910725,
      "loss": 2.8746,
      "theoretical_loss": 3.677153505647132,
      "tokens_seen": 923467776
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047397626253073573,
      "loss": 2.6207,
      "theoretical_loss": 3.6771283054982886,
      "tokens_seen": 923533312
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047395971250236427,
      "loss": 2.6169,
      "theoretical_loss": 3.677103107638314,
      "tokens_seen": 923598848
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004739431624739928,
      "loss": 2.9233,
      "theoretical_loss": 3.677077912066837,
      "tokens_seen": 923664384
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047392661244562135,
      "loss": 2.5482,
      "theoretical_loss": 3.677052718783487,
      "tokens_seen": 923729920
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047391006241724983,
      "loss": 2.7314,
      "theoretical_loss": 3.6770275277878945,
      "tokens_seen": 923795456
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047389351238887837,
      "loss": 2.8247,
      "theoretical_loss": 3.677002339079689,
      "tokens_seen": 923860992
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004738769623605069,
      "loss": 2.3519,
      "theoretical_loss": 3.676977152658501,
      "tokens_seen": 923926528
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047386041233213544,
      "loss": 2.9036,
      "theoretical_loss": 3.6769519685239604,
      "tokens_seen": 923992064
    },
    {
      "epoch": 0.33,
      "objective/train/avg_token_score": 0.014201238751411438,
      "objective/train/avg_weight": 0.9832841157913208,
      "objective/train/docs_used": 529792,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0170297622680664,
      "objective/train/std_weight": 0.08190172910690308,
      "objective/train/theoretical_loss": 3.676926786675698,
      "objective/train/tokens_used": 944517600,
      "theoretical_loss": 3.676926786675698,
      "tokens_seen": 924057600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004738438623037639,
      "loss": 2.7,
      "theoretical_loss": 3.676926786675698,
      "tokens_seen": 924057600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047382731227539246,
      "loss": 2.5238,
      "theoretical_loss": 3.6769016071133427,
      "tokens_seen": 924123136
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000473810762247021,
      "loss": 2.7141,
      "theoretical_loss": 3.6768764298365264,
      "tokens_seen": 924188672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047379421221864954,
      "loss": 2.8693,
      "theoretical_loss": 3.6768512548448786,
      "tokens_seen": 924254208
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000473777662190278,
      "loss": 2.5888,
      "theoretical_loss": 3.6768260821380307,
      "tokens_seen": 924319744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047376111216190656,
      "loss": 2.7469,
      "theoretical_loss": 3.6768009117156124,
      "tokens_seen": 924385280
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004737445621335351,
      "loss": 2.7295,
      "theoretical_loss": 3.6767757435772555,
      "tokens_seen": 924450816
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047372801210516363,
      "loss": 2.7251,
      "theoretical_loss": 3.67675057772259,
      "tokens_seen": 924516352
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004737114620767921,
      "loss": 2.7475,
      "theoretical_loss": 3.6767254141512478,
      "tokens_seen": 924581888
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047369491204842066,
      "loss": 2.6064,
      "theoretical_loss": 3.6767002528628585,
      "tokens_seen": 924647424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004736783620200492,
      "loss": 2.7251,
      "theoretical_loss": 3.676675093857054,
      "tokens_seen": 924712960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047366181199167773,
      "loss": 2.6874,
      "theoretical_loss": 3.676649937133466,
      "tokens_seen": 924778496
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004736452619633062,
      "loss": 2.6213,
      "theoretical_loss": 3.6766247826917247,
      "tokens_seen": 924844032
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047362871193493475,
      "loss": 2.6277,
      "theoretical_loss": 3.676599630531462,
      "tokens_seen": 924909568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004736121619065633,
      "loss": 2.7881,
      "theoretical_loss": 3.6765744806523095,
      "tokens_seen": 924975104
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004735956118781918,
      "loss": 2.9624,
      "theoretical_loss": 3.6765493330538987,
      "tokens_seen": 925040640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004735790618498203,
      "loss": 2.6703,
      "theoretical_loss": 3.6765241877358603,
      "tokens_seen": 925106176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047356251182144885,
      "loss": 2.453,
      "theoretical_loss": 3.6764990446978274,
      "tokens_seen": 925171712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004735459617930774,
      "loss": 2.7668,
      "theoretical_loss": 3.6764739039394305,
      "tokens_seen": 925237248
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004735294117647059,
      "loss": 2.6204,
      "theoretical_loss": 3.676448765460303,
      "tokens_seen": 925302784
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004735128617363344,
      "loss": 2.6473,
      "theoretical_loss": 3.676423629260075,
      "tokens_seen": 925368320
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047349631170796294,
      "loss": 2.8352,
      "theoretical_loss": 3.67639849533838,
      "tokens_seen": 925433856
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004734797616795915,
      "loss": 2.5404,
      "theoretical_loss": 3.676373363694849,
      "tokens_seen": 925499392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047346321165122,
      "loss": 2.9009,
      "theoretical_loss": 3.676348234329115,
      "tokens_seen": 925564928
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004734466616228485,
      "loss": 2.7993,
      "theoretical_loss": 3.67632310724081,
      "tokens_seen": 925630464
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047343011159447704,
      "loss": 2.7109,
      "theoretical_loss": 3.6762979824295665,
      "tokens_seen": 925696000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004734135615661056,
      "loss": 2.8497,
      "theoretical_loss": 3.6762728598950174,
      "tokens_seen": 925761536
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004733970115377341,
      "loss": 2.6967,
      "theoretical_loss": 3.676247739636794,
      "tokens_seen": 925827072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047338046150936254,
      "loss": 2.6323,
      "theoretical_loss": 3.6762226216545297,
      "tokens_seen": 925892608
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004733639114809911,
      "loss": 2.5731,
      "theoretical_loss": 3.676197505947858,
      "tokens_seen": 925958144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004733473614526196,
      "loss": 2.8345,
      "theoretical_loss": 3.67617239251641,
      "tokens_seen": 926023680
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004733308114242481,
      "loss": 2.7402,
      "theoretical_loss": 3.67614728135982,
      "tokens_seen": 926089216
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047331426139587664,
      "loss": 2.6519,
      "theoretical_loss": 3.6761221724777204,
      "tokens_seen": 926154752
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004732977113675052,
      "loss": 2.7026,
      "theoretical_loss": 3.676097065869744,
      "tokens_seen": 926220288
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004732811613391337,
      "loss": 2.7372,
      "theoretical_loss": 3.6760719615355244,
      "tokens_seen": 926285824
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004732646113107622,
      "loss": 2.6784,
      "theoretical_loss": 3.6760468594746953,
      "tokens_seen": 926351360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047324806128239074,
      "loss": 2.5131,
      "theoretical_loss": 3.676021759686889,
      "tokens_seen": 926416896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047323151125401927,
      "loss": 2.6881,
      "theoretical_loss": 3.6759966621717393,
      "tokens_seen": 926482432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004732149612256478,
      "loss": 2.7753,
      "theoretical_loss": 3.6759715669288795,
      "tokens_seen": 926547968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004731984111972763,
      "loss": 2.671,
      "theoretical_loss": 3.6759464739579437,
      "tokens_seen": 926613504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047318186116890483,
      "loss": 2.7375,
      "theoretical_loss": 3.6759213832585647,
      "tokens_seen": 926679040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047316531114053337,
      "loss": 2.6648,
      "theoretical_loss": 3.675896294830377,
      "tokens_seen": 926744576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004731487611121619,
      "loss": 2.7326,
      "theoretical_loss": 3.6758712086730148,
      "tokens_seen": 926810112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004731322110837904,
      "loss": 2.2607,
      "theoretical_loss": 3.6758461247861107,
      "tokens_seen": 926875648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047311566105541893,
      "loss": 2.7374,
      "theoretical_loss": 3.6758210431692993,
      "tokens_seen": 926941184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047309911102704746,
      "loss": 2.5679,
      "theoretical_loss": 3.6757959638222153,
      "tokens_seen": 927006720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000473082560998676,
      "loss": 2.8518,
      "theoretical_loss": 3.675770886744492,
      "tokens_seen": 927072256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004730660109703045,
      "loss": 2.692,
      "theoretical_loss": 3.675745811935764,
      "tokens_seen": 927137792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000473049460941933,
      "loss": 2.5465,
      "theoretical_loss": 3.6757207393956657,
      "tokens_seen": 927203328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047303291091356156,
      "loss": 2.5968,
      "theoretical_loss": 3.675695669123832,
      "tokens_seen": 927268864
    },
    {
      "epoch": 0.33,
      "objective/train/avg_token_score": 0.007149987854063511,
      "objective/train/avg_weight": 0.9915858507156372,
      "objective/train/docs_used": 531599,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0497851371765137,
      "objective/train/std_weight": 0.045658838003873825,
      "objective/train/theoretical_loss": 3.6756706011198963,
      "objective/train/tokens_used": 947794400,
      "theoretical_loss": 3.6756706011198963,
      "tokens_seen": 927334400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004730163608851901,
      "loss": 2.7963,
      "theoretical_loss": 3.6756706011198963,
      "tokens_seen": 927334400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004729998108568186,
      "loss": 2.6427,
      "theoretical_loss": 3.6756455353834934,
      "tokens_seen": 927399936
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004729832608284471,
      "loss": 2.5311,
      "theoretical_loss": 3.675620471914259,
      "tokens_seen": 927465472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047296671080007566,
      "loss": 2.6858,
      "theoretical_loss": 3.6755954107118267,
      "tokens_seen": 927531008
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004729501607717042,
      "loss": 2.5061,
      "theoretical_loss": 3.675570351775832,
      "tokens_seen": 927596544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004729336107433327,
      "loss": 2.5493,
      "theoretical_loss": 3.675545295105909,
      "tokens_seen": 927662080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004729170607149612,
      "loss": 2.6292,
      "theoretical_loss": 3.6755202407016943,
      "tokens_seen": 927727616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047290051068658975,
      "loss": 2.6701,
      "theoretical_loss": 3.675495188562822,
      "tokens_seen": 927793152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004728839606582183,
      "loss": 2.6555,
      "theoretical_loss": 3.675470138688927,
      "tokens_seen": 927858688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004728674106298468,
      "loss": 2.5185,
      "theoretical_loss": 3.6754450910796446,
      "tokens_seen": 927924224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004728508606014753,
      "loss": 2.616,
      "theoretical_loss": 3.6754200457346107,
      "tokens_seen": 927989760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047283431057310385,
      "loss": 2.5448,
      "theoretical_loss": 3.6753950026534605,
      "tokens_seen": 928055296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004728177605447324,
      "loss": 2.6934,
      "theoretical_loss": 3.6753699618358295,
      "tokens_seen": 928120832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047280121051636087,
      "loss": 2.8667,
      "theoretical_loss": 3.675344923281353,
      "tokens_seen": 928186368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004727846604879894,
      "loss": 2.7186,
      "theoretical_loss": 3.6753198869896675,
      "tokens_seen": 928251904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047276811045961794,
      "loss": 2.8509,
      "theoretical_loss": 3.675294852960408,
      "tokens_seen": 928317440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004727515604312465,
      "loss": 2.8216,
      "theoretical_loss": 3.6752698211932104,
      "tokens_seen": 928382976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047273501040287497,
      "loss": 2.4939,
      "theoretical_loss": 3.675244791687711,
      "tokens_seen": 928448512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004727184603745035,
      "loss": 2.648,
      "theoretical_loss": 3.675219764443546,
      "tokens_seen": 928514048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047270191034613204,
      "loss": 2.9236,
      "theoretical_loss": 3.6751947394603506,
      "tokens_seen": 928579584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004726853603177606,
      "loss": 2.8331,
      "theoretical_loss": 3.675169716737762,
      "tokens_seen": 928645120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047266881028938906,
      "loss": 2.6589,
      "theoretical_loss": 3.6751446962754155,
      "tokens_seen": 928710656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004726522602610176,
      "loss": 2.8145,
      "theoretical_loss": 3.6751196780729485,
      "tokens_seen": 928776192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047263571023264614,
      "loss": 2.7272,
      "theoretical_loss": 3.675094662129997,
      "tokens_seen": 928841728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004726191602042747,
      "loss": 2.6231,
      "theoretical_loss": 3.6750696484461978,
      "tokens_seen": 928907264
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047260261017590316,
      "loss": 2.7947,
      "theoretical_loss": 3.6750446370211867,
      "tokens_seen": 928972800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004725860601475317,
      "loss": 2.6095,
      "theoretical_loss": 3.675019627854601,
      "tokens_seen": 929038336
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047256951011916023,
      "loss": 2.547,
      "theoretical_loss": 3.6749946209460775,
      "tokens_seen": 929103872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047255296009078877,
      "loss": 2.4717,
      "theoretical_loss": 3.6749696162952525,
      "tokens_seen": 929169408
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047253641006241725,
      "loss": 2.8135,
      "theoretical_loss": 3.674944613901764,
      "tokens_seen": 929234944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004725198600340458,
      "loss": 2.6212,
      "theoretical_loss": 3.674919613765248,
      "tokens_seen": 929300480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047250331000567433,
      "loss": 2.5524,
      "theoretical_loss": 3.674894615885342,
      "tokens_seen": 929366016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047248675997730287,
      "loss": 2.7009,
      "theoretical_loss": 3.6748696202616835,
      "tokens_seen": 929431552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047247020994893135,
      "loss": 2.4158,
      "theoretical_loss": 3.6748446268939094,
      "tokens_seen": 929497088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004724536599205599,
      "loss": 2.6189,
      "theoretical_loss": 3.674819635781658,
      "tokens_seen": 929562624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004724371098921884,
      "loss": 2.5566,
      "theoretical_loss": 3.674794646924565,
      "tokens_seen": 929628160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047242055986381696,
      "loss": 2.5174,
      "theoretical_loss": 3.674769660322269,
      "tokens_seen": 929693696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047240400983544545,
      "loss": 2.8931,
      "theoretical_loss": 3.6747446759744076,
      "tokens_seen": 929759232
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047238745980707393,
      "loss": 2.8026,
      "theoretical_loss": 3.6747196938806184,
      "tokens_seen": 929824768
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004723709097787024,
      "loss": 2.835,
      "theoretical_loss": 3.6746947140405393,
      "tokens_seen": 929890304
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047235435975033095,
      "loss": 2.4402,
      "theoretical_loss": 3.6746697364538083,
      "tokens_seen": 929955840
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004723378097219595,
      "loss": 2.7382,
      "theoretical_loss": 3.6746447611200628,
      "tokens_seen": 930021376
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000472321259693588,
      "loss": 2.6065,
      "theoretical_loss": 3.6746197880389415,
      "tokens_seen": 930086912
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004723047096652165,
      "loss": 2.8941,
      "theoretical_loss": 3.674594817210082,
      "tokens_seen": 930152448
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047228815963684505,
      "loss": 2.635,
      "theoretical_loss": 3.674569848633123,
      "tokens_seen": 930217984
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004722716096084736,
      "loss": 2.6106,
      "theoretical_loss": 3.674544882307702,
      "tokens_seen": 930283520
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004722550595801021,
      "loss": 2.8913,
      "theoretical_loss": 3.6745199182334582,
      "tokens_seen": 930349056
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004722385095517306,
      "loss": 2.7717,
      "theoretical_loss": 3.6744949564100295,
      "tokens_seen": 930414592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047222195952335914,
      "loss": 2.6624,
      "theoretical_loss": 3.674469996837055,
      "tokens_seen": 930480128
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004722054094949877,
      "loss": 2.4497,
      "theoretical_loss": 3.674445039514173,
      "tokens_seen": 930545664
    },
    {
      "epoch": 0.33,
      "objective/train/avg_token_score": 0.016706664115190506,
      "objective/train/avg_weight": 0.980344295501709,
      "objective/train/docs_used": 533595,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.864992141723633,
      "objective/train/std_weight": 0.07425805181264877,
      "objective/train/theoretical_loss": 3.6744200844410217,
      "objective/train/tokens_used": 951071200,
      "theoretical_loss": 3.6744200844410217,
      "tokens_seen": 930611200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004721888594666162,
      "loss": 2.8589,
      "theoretical_loss": 3.6744200844410217,
      "tokens_seen": 930611200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004721723094382447,
      "loss": 2.6552,
      "theoretical_loss": 3.674395131617241,
      "tokens_seen": 930676736
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047215575940987324,
      "loss": 2.6298,
      "theoretical_loss": 3.674370181042469,
      "tokens_seen": 930742272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004721392093815018,
      "loss": 2.5914,
      "theoretical_loss": 3.6743452327163446,
      "tokens_seen": 930807808
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004721226593531303,
      "loss": 2.8847,
      "theoretical_loss": 3.6743202866385074,
      "tokens_seen": 930873344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004721061093247588,
      "loss": 2.5646,
      "theoretical_loss": 3.6742953428085965,
      "tokens_seen": 930938880
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047208955929638733,
      "loss": 2.5214,
      "theoretical_loss": 3.6742704012262504,
      "tokens_seen": 931004416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047207300926801587,
      "loss": 2.5902,
      "theoretical_loss": 3.6742454618911085,
      "tokens_seen": 931069952
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004720564592396444,
      "loss": 2.7183,
      "theoretical_loss": 3.674220524802811,
      "tokens_seen": 931135488
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004720399092112729,
      "loss": 2.7365,
      "theoretical_loss": 3.6741955899609966,
      "tokens_seen": 931201024
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047202335918290143,
      "loss": 2.8513,
      "theoretical_loss": 3.674170657365305,
      "tokens_seen": 931266560
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047200680915452997,
      "loss": 2.6371,
      "theoretical_loss": 3.674145727015376,
      "tokens_seen": 931332096
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004719902591261585,
      "loss": 2.4924,
      "theoretical_loss": 3.6741207989108493,
      "tokens_seen": 931397632
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047197370909778704,
      "loss": 2.4255,
      "theoretical_loss": 3.674095873051365,
      "tokens_seen": 931463168
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004719571590694155,
      "loss": 2.6293,
      "theoretical_loss": 3.6740709494365618,
      "tokens_seen": 931528704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047194060904104406,
      "loss": 2.7933,
      "theoretical_loss": 3.6740460280660807,
      "tokens_seen": 931594240
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004719240590126726,
      "loss": 2.7589,
      "theoretical_loss": 3.6740211089395616,
      "tokens_seen": 931659776
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047190750898430114,
      "loss": 2.5858,
      "theoretical_loss": 3.673996192056644,
      "tokens_seen": 931725312
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004718909589559296,
      "loss": 2.5282,
      "theoretical_loss": 3.6739712774169693,
      "tokens_seen": 931790848
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047187440892755816,
      "loss": 2.8514,
      "theoretical_loss": 3.673946365020177,
      "tokens_seen": 931856384
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004718578588991867,
      "loss": 2.6559,
      "theoretical_loss": 3.6739214548659076,
      "tokens_seen": 931921920
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047184130887081523,
      "loss": 2.7159,
      "theoretical_loss": 3.6738965469538014,
      "tokens_seen": 931987456
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004718247588424437,
      "loss": 2.6422,
      "theoretical_loss": 3.6738716412834984,
      "tokens_seen": 932052992
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047180820881407225,
      "loss": 2.709,
      "theoretical_loss": 3.6738467378546407,
      "tokens_seen": 932118528
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004717916587857008,
      "loss": 2.7549,
      "theoretical_loss": 3.6738218366668685,
      "tokens_seen": 932184064
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047177510875732933,
      "loss": 2.8281,
      "theoretical_loss": 3.673796937719821,
      "tokens_seen": 932249600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004717585587289578,
      "loss": 2.6975,
      "theoretical_loss": 3.673772041013142,
      "tokens_seen": 932315136
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047174200870058635,
      "loss": 2.5978,
      "theoretical_loss": 3.6737471465464697,
      "tokens_seen": 932380672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004717254586722149,
      "loss": 2.8171,
      "theoretical_loss": 3.673722254319446,
      "tokens_seen": 932446208
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004717089086438434,
      "loss": 2.7657,
      "theoretical_loss": 3.673697364331713,
      "tokens_seen": 932511744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004716923586154719,
      "loss": 2.4394,
      "theoretical_loss": 3.6736724765829107,
      "tokens_seen": 932577280
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047167580858710045,
      "loss": 2.6502,
      "theoretical_loss": 3.6736475910726814,
      "tokens_seen": 932642816
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000471659258558729,
      "loss": 2.5178,
      "theoretical_loss": 3.673622707800665,
      "tokens_seen": 932708352
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004716427085303575,
      "loss": 2.7278,
      "theoretical_loss": 3.6735978267665046,
      "tokens_seen": 932773888
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000471626158501986,
      "loss": 2.7247,
      "theoretical_loss": 3.6735729479698405,
      "tokens_seen": 932839424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047160960847361454,
      "loss": 2.6503,
      "theoretical_loss": 3.6735480714103153,
      "tokens_seen": 932904960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004715930584452431,
      "loss": 2.3709,
      "theoretical_loss": 3.6735231970875697,
      "tokens_seen": 932970496
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004715765084168716,
      "loss": 2.8175,
      "theoretical_loss": 3.6734983250012467,
      "tokens_seen": 933036032
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004715599583885001,
      "loss": 2.8735,
      "theoretical_loss": 3.6734734551509867,
      "tokens_seen": 933101568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047154340836012864,
      "loss": 2.8006,
      "theoretical_loss": 3.6734485875364324,
      "tokens_seen": 933167104
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004715268583317572,
      "loss": 2.6312,
      "theoretical_loss": 3.673423722157226,
      "tokens_seen": 933232640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004715103083033857,
      "loss": 2.6576,
      "theoretical_loss": 3.6733988590130098,
      "tokens_seen": 933298176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004714937582750142,
      "loss": 2.5131,
      "theoretical_loss": 3.673373998103425,
      "tokens_seen": 933363712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047147720824664273,
      "loss": 2.5838,
      "theoretical_loss": 3.6733491394281153,
      "tokens_seen": 933429248
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047146065821827127,
      "loss": 2.5806,
      "theoretical_loss": 3.673324282986721,
      "tokens_seen": 933494784
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004714441081898998,
      "loss": 2.6479,
      "theoretical_loss": 3.6732994287788867,
      "tokens_seen": 933560320
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004714275581615283,
      "loss": 2.6867,
      "theoretical_loss": 3.673274576804254,
      "tokens_seen": 933625856
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004714110081331568,
      "loss": 2.6976,
      "theoretical_loss": 3.6732497270624656,
      "tokens_seen": 933691392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047139445810478526,
      "loss": 2.8459,
      "theoretical_loss": 3.673224879553163,
      "tokens_seen": 933756928
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004713779080764138,
      "loss": 2.666,
      "theoretical_loss": 3.6732000342759914,
      "tokens_seen": 933822464
    },
    {
      "epoch": 0.33,
      "objective/train/avg_token_score": 0.01587812788784504,
      "objective/train/avg_weight": 0.981309175491333,
      "objective/train/docs_used": 535440,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6055450439453125,
      "objective/train/std_weight": 0.08301340043544769,
      "objective/train/theoretical_loss": 3.6731751912305914,
      "objective/train/tokens_used": 954348000,
      "theoretical_loss": 3.6731751912305914,
      "tokens_seen": 933888000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047136135804804233,
      "loss": 2.6679,
      "theoretical_loss": 3.6731751912305914,
      "tokens_seen": 933888000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047134480801967087,
      "loss": 2.8672,
      "theoretical_loss": 3.6731503504166074,
      "tokens_seen": 933953536
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047132825799129936,
      "loss": 2.5915,
      "theoretical_loss": 3.673125511833682,
      "tokens_seen": 934019072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004713117079629279,
      "loss": 2.6528,
      "theoretical_loss": 3.6731006754814572,
      "tokens_seen": 934084608
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047129515793455643,
      "loss": 2.593,
      "theoretical_loss": 3.673075841359578,
      "tokens_seen": 934150144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047127860790618497,
      "loss": 2.8752,
      "theoretical_loss": 3.6730510094676863,
      "tokens_seen": 934215680
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047126205787781345,
      "loss": 2.7884,
      "theoretical_loss": 3.6730261798054262,
      "tokens_seen": 934281216
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000471245507849442,
      "loss": 2.621,
      "theoretical_loss": 3.6730013523724407,
      "tokens_seen": 934346752
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004712289578210705,
      "loss": 2.8877,
      "theoretical_loss": 3.6729765271683736,
      "tokens_seen": 934412288
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047121240779269906,
      "loss": 2.9289,
      "theoretical_loss": 3.6729517041928683,
      "tokens_seen": 934477824
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047119585776432755,
      "loss": 2.7792,
      "theoretical_loss": 3.6729268834455686,
      "tokens_seen": 934543360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004711793077359561,
      "loss": 2.5517,
      "theoretical_loss": 3.6729020649261184,
      "tokens_seen": 934608896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004711627577075846,
      "loss": 2.6857,
      "theoretical_loss": 3.672877248634161,
      "tokens_seen": 934674432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047114620767921316,
      "loss": 2.8437,
      "theoretical_loss": 3.6728524345693403,
      "tokens_seen": 934739968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047112965765084164,
      "loss": 2.6558,
      "theoretical_loss": 3.672827622731301,
      "tokens_seen": 934805504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004711131076224702,
      "loss": 2.8273,
      "theoretical_loss": 3.672802813119687,
      "tokens_seen": 934871040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004710965575940987,
      "loss": 2.765,
      "theoretical_loss": 3.672778005734142,
      "tokens_seen": 934936576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047108000756572726,
      "loss": 2.4407,
      "theoretical_loss": 3.672753200574311,
      "tokens_seen": 935002112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047106345753735574,
      "loss": 2.9301,
      "theoretical_loss": 3.672728397639837,
      "tokens_seen": 935067648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004710469075089843,
      "loss": 2.7497,
      "theoretical_loss": 3.672703596930366,
      "tokens_seen": 935133184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004710303574806128,
      "loss": 2.6999,
      "theoretical_loss": 3.672678798445541,
      "tokens_seen": 935198720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047101380745224135,
      "loss": 2.6681,
      "theoretical_loss": 3.6726540021850083,
      "tokens_seen": 935264256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047099725742386984,
      "loss": 2.7699,
      "theoretical_loss": 3.672629208148411,
      "tokens_seen": 935329792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004709807073954984,
      "loss": 2.6471,
      "theoretical_loss": 3.6726044163353944,
      "tokens_seen": 935395328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004709641573671269,
      "loss": 2.6921,
      "theoretical_loss": 3.6725796267456032,
      "tokens_seen": 935460864
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047094760733875545,
      "loss": 2.5283,
      "theoretical_loss": 3.672554839378683,
      "tokens_seen": 935526400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047093105731038393,
      "loss": 2.8663,
      "theoretical_loss": 3.6725300542342776,
      "tokens_seen": 935591936
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047091450728201247,
      "loss": 2.6015,
      "theoretical_loss": 3.6725052713120325,
      "tokens_seen": 935657472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000470897957253641,
      "loss": 2.4184,
      "theoretical_loss": 3.6724804906115933,
      "tokens_seen": 935723008
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047088140722526954,
      "loss": 2.7158,
      "theoretical_loss": 3.672455712132605,
      "tokens_seen": 935788544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047086485719689803,
      "loss": 2.5212,
      "theoretical_loss": 3.672430935874713,
      "tokens_seen": 935854080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047084830716852656,
      "loss": 2.4176,
      "theoretical_loss": 3.6724061618375625,
      "tokens_seen": 935919616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004708317571401551,
      "loss": 2.4223,
      "theoretical_loss": 3.6723813900207984,
      "tokens_seen": 935985152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047081520711178364,
      "loss": 2.5868,
      "theoretical_loss": 3.6723566204240674,
      "tokens_seen": 936050688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004707986570834121,
      "loss": 2.6945,
      "theoretical_loss": 3.672331853047014,
      "tokens_seen": 936116224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047078210705504066,
      "loss": 2.7674,
      "theoretical_loss": 3.672307087889285,
      "tokens_seen": 936181760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004707655570266692,
      "loss": 2.6575,
      "theoretical_loss": 3.6722823249505256,
      "tokens_seen": 936247296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047074900699829774,
      "loss": 2.6234,
      "theoretical_loss": 3.6722575642303816,
      "tokens_seen": 936312832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004707324569699262,
      "loss": 2.9303,
      "theoretical_loss": 3.672232805728499,
      "tokens_seen": 936378368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047071590694155476,
      "loss": 2.7558,
      "theoretical_loss": 3.6722080494445244,
      "tokens_seen": 936443904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004706993569131833,
      "loss": 2.9449,
      "theoretical_loss": 3.672183295378103,
      "tokens_seen": 936509440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047068280688481183,
      "loss": 2.6287,
      "theoretical_loss": 3.672158543528882,
      "tokens_seen": 936574976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004706662568564403,
      "loss": 2.684,
      "theoretical_loss": 3.6721337938965064,
      "tokens_seen": 936640512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047064970682806885,
      "loss": 2.4409,
      "theoretical_loss": 3.672109046480624,
      "tokens_seen": 936706048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004706331567996974,
      "loss": 2.5729,
      "theoretical_loss": 3.67208430128088,
      "tokens_seen": 936771584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047061660677132593,
      "loss": 2.8505,
      "theoretical_loss": 3.672059558296922,
      "tokens_seen": 936837120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004706000567429544,
      "loss": 2.7953,
      "theoretical_loss": 3.672034817528396,
      "tokens_seen": 936902656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047058350671458295,
      "loss": 2.4693,
      "theoretical_loss": 3.6720100789749486,
      "tokens_seen": 936968192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004705669566862115,
      "loss": 2.8375,
      "theoretical_loss": 3.6719853426362263,
      "tokens_seen": 937033728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047055040665784,
      "loss": 2.7479,
      "theoretical_loss": 3.6719606085118772,
      "tokens_seen": 937099264
    },
    {
      "epoch": 0.33,
      "objective/train/avg_token_score": 0.008387474343180656,
      "objective/train/avg_weight": 0.9901561141014099,
      "objective/train/docs_used": 537222,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5532093048095703,
      "objective/train/std_weight": 0.05495748668909073,
      "objective/train/theoretical_loss": 3.671935876601547,
      "objective/train/tokens_used": 957624800,
      "theoretical_loss": 3.671935876601547,
      "tokens_seen": 937164800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004705338566294685,
      "loss": 2.7058,
      "theoretical_loss": 3.671935876601547,
      "tokens_seen": 937164800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047051730660109704,
      "loss": 2.6477,
      "theoretical_loss": 3.671911146904883,
      "tokens_seen": 937230336
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004705007565727256,
      "loss": 2.7273,
      "theoretical_loss": 3.671886419421533,
      "tokens_seen": 937295872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004704842065443541,
      "loss": 2.7054,
      "theoretical_loss": 3.6718616941511426,
      "tokens_seen": 937361408
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047046765651598266,
      "loss": 2.6684,
      "theoretical_loss": 3.6718369710933607,
      "tokens_seen": 937426944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047045110648761114,
      "loss": 2.6706,
      "theoretical_loss": 3.671812250247834,
      "tokens_seen": 937492480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004704345564592396,
      "loss": 2.6745,
      "theoretical_loss": 3.6717875316142097,
      "tokens_seen": 937558016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004704180064308681,
      "loss": 2.6092,
      "theoretical_loss": 3.6717628151921353,
      "tokens_seen": 937623552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047040145640249664,
      "loss": 2.7183,
      "theoretical_loss": 3.6717381009812593,
      "tokens_seen": 937689088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004703849063741252,
      "loss": 2.5157,
      "theoretical_loss": 3.671713388981228,
      "tokens_seen": 937754624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004703683563457537,
      "loss": 2.2832,
      "theoretical_loss": 3.67168867919169,
      "tokens_seen": 937820160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0004703518063173822,
      "loss": 2.6126,
      "theoretical_loss": 3.671663971612293,
      "tokens_seen": 937885696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00047033525628901074,
      "loss": 2.586,
      "theoretical_loss": 3.6716392662426847,
      "tokens_seen": 937951232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004703187062606393,
      "loss": 2.5083,
      "theoretical_loss": 3.671614563082513,
      "tokens_seen": 938016768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004703021562322678,
      "loss": 2.5003,
      "theoretical_loss": 3.6715898621314262,
      "tokens_seen": 938082304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004702856062038963,
      "loss": 2.4886,
      "theoretical_loss": 3.6715651633890722,
      "tokens_seen": 938147840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00047026905617552484,
      "loss": 2.5774,
      "theoretical_loss": 3.6715404668550997,
      "tokens_seen": 938213376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004702525061471534,
      "loss": 2.6598,
      "theoretical_loss": 3.671515772529157,
      "tokens_seen": 938278912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004702359561187819,
      "loss": 2.7585,
      "theoretical_loss": 3.671491080410892,
      "tokens_seen": 938344448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004702194060904104,
      "loss": 2.634,
      "theoretical_loss": 3.671466390499953,
      "tokens_seen": 938409984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00047020285606203893,
      "loss": 2.6543,
      "theoretical_loss": 3.6714417027959887,
      "tokens_seen": 938475520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00047018630603366747,
      "loss": 2.4163,
      "theoretical_loss": 3.6714170172986487,
      "tokens_seen": 938541056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000470169756005296,
      "loss": 2.7003,
      "theoretical_loss": 3.6713923340075802,
      "tokens_seen": 938606592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004701532059769245,
      "loss": 2.6504,
      "theoretical_loss": 3.6713676529224335,
      "tokens_seen": 938672128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00047013665594855303,
      "loss": 2.4621,
      "theoretical_loss": 3.671342974042856,
      "tokens_seen": 938737664
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00047012010592018157,
      "loss": 2.5439,
      "theoretical_loss": 3.6713182973684972,
      "tokens_seen": 938803200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004701035558918101,
      "loss": 2.7864,
      "theoretical_loss": 3.6712936228990065,
      "tokens_seen": 938868736
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004700870058634386,
      "loss": 2.6448,
      "theoretical_loss": 3.6712689506340332,
      "tokens_seen": 938934272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004700704558350671,
      "loss": 2.7807,
      "theoretical_loss": 3.6712442805732257,
      "tokens_seen": 938999808
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00047005390580669566,
      "loss": 2.7264,
      "theoretical_loss": 3.671219612716234,
      "tokens_seen": 939065344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004700373557783242,
      "loss": 2.7506,
      "theoretical_loss": 3.6711949470627063,
      "tokens_seen": 939130880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004700208057499527,
      "loss": 2.4657,
      "theoretical_loss": 3.6711702836122937,
      "tokens_seen": 939196416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004700042557215812,
      "loss": 2.3535,
      "theoretical_loss": 3.671145622364644,
      "tokens_seen": 939261952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046998770569320976,
      "loss": 2.7612,
      "theoretical_loss": 3.671120963319408,
      "tokens_seen": 939327488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004699711556648383,
      "loss": 2.7476,
      "theoretical_loss": 3.671096306476235,
      "tokens_seen": 939393024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004699546056364668,
      "loss": 2.5996,
      "theoretical_loss": 3.6710716518347746,
      "tokens_seen": 939458560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004699380556080953,
      "loss": 2.8253,
      "theoretical_loss": 3.671046999394677,
      "tokens_seen": 939524096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046992150557972385,
      "loss": 2.7885,
      "theoretical_loss": 3.6710223491555913,
      "tokens_seen": 939589632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004699049555513524,
      "loss": 2.5314,
      "theoretical_loss": 3.6709977011171686,
      "tokens_seen": 939655168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004698884055229809,
      "loss": 2.741,
      "theoretical_loss": 3.6709730552790583,
      "tokens_seen": 939720704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004698718554946094,
      "loss": 2.6279,
      "theoretical_loss": 3.670948411640911,
      "tokens_seen": 939786240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046985530546623795,
      "loss": 2.7821,
      "theoretical_loss": 3.670923770202376,
      "tokens_seen": 939851776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004698387554378665,
      "loss": 2.5482,
      "theoretical_loss": 3.6708991309631043,
      "tokens_seen": 939917312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046982220540949497,
      "loss": 2.621,
      "theoretical_loss": 3.6708744939227462,
      "tokens_seen": 939982848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004698056553811235,
      "loss": 2.7581,
      "theoretical_loss": 3.6708498590809526,
      "tokens_seen": 940048384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046978910535275205,
      "loss": 2.5416,
      "theoretical_loss": 3.6708252264373735,
      "tokens_seen": 940113920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004697725553243806,
      "loss": 2.8358,
      "theoretical_loss": 3.6708005959916594,
      "tokens_seen": 940179456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046975600529600907,
      "loss": 2.7361,
      "theoretical_loss": 3.6707759677434613,
      "tokens_seen": 940244992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004697394552676376,
      "loss": 2.636,
      "theoretical_loss": 3.67075134169243,
      "tokens_seen": 940310528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046972290523926614,
      "loss": 2.8751,
      "theoretical_loss": 3.6707267178382166,
      "tokens_seen": 940376064
    },
    {
      "epoch": 0.34,
      "objective/train/avg_token_score": 0.01680607721209526,
      "objective/train/avg_weight": 0.9802517890930176,
      "objective/train/docs_used": 539286,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8097641468048096,
      "objective/train/std_weight": 0.07812841981649399,
      "objective/train/theoretical_loss": 3.6707020961804715,
      "objective/train/tokens_used": 960901600,
      "theoretical_loss": 3.6707020961804715,
      "tokens_seen": 940441600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004697063552108947,
      "loss": 2.6436,
      "theoretical_loss": 3.6707020961804715,
      "tokens_seen": 940441600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046968980518252316,
      "loss": 2.4611,
      "theoretical_loss": 3.670677476718846,
      "tokens_seen": 940507136
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004696732551541517,
      "loss": 2.7317,
      "theoretical_loss": 3.6706528594529915,
      "tokens_seen": 940572672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046965670512578024,
      "loss": 2.4232,
      "theoretical_loss": 3.6706282443825584,
      "tokens_seen": 940638208
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004696401550974088,
      "loss": 2.7873,
      "theoretical_loss": 3.6706036315071993,
      "tokens_seen": 940703744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046962360506903726,
      "loss": 2.6793,
      "theoretical_loss": 3.670579020826564,
      "tokens_seen": 940769280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004696070550406658,
      "loss": 2.5477,
      "theoretical_loss": 3.670554412340305,
      "tokens_seen": 940834816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046959050501229433,
      "loss": 2.7154,
      "theoretical_loss": 3.670529806048073,
      "tokens_seen": 940900352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046957395498392287,
      "loss": 2.7268,
      "theoretical_loss": 3.6705052019495206,
      "tokens_seen": 940965888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046955740495555135,
      "loss": 2.8049,
      "theoretical_loss": 3.6704806000442987,
      "tokens_seen": 941031424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004695408549271799,
      "loss": 2.7691,
      "theoretical_loss": 3.6704560003320594,
      "tokens_seen": 941096960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046952430489880843,
      "loss": 2.7658,
      "theoretical_loss": 3.6704314028124543,
      "tokens_seen": 941162496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046950775487043697,
      "loss": 2.7778,
      "theoretical_loss": 3.6704068074851355,
      "tokens_seen": 941228032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046949120484206545,
      "loss": 2.7131,
      "theoretical_loss": 3.670382214349755,
      "tokens_seen": 941293568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000469474654813694,
      "loss": 2.6128,
      "theoretical_loss": 3.6703576234059643,
      "tokens_seen": 941359104
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004694581047853225,
      "loss": 2.5488,
      "theoretical_loss": 3.6703330346534155,
      "tokens_seen": 941424640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046944155475695096,
      "loss": 2.6021,
      "theoretical_loss": 3.6703084480917623,
      "tokens_seen": 941490176
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004694250047285795,
      "loss": 2.5844,
      "theoretical_loss": 3.670283863720656,
      "tokens_seen": 941555712
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046940845470020803,
      "loss": 3.114,
      "theoretical_loss": 3.6702592815397486,
      "tokens_seen": 941621248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046939190467183657,
      "loss": 2.8472,
      "theoretical_loss": 3.6702347015486927,
      "tokens_seen": 941686784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046937535464346505,
      "loss": 2.5008,
      "theoretical_loss": 3.670210123747142,
      "tokens_seen": 941752320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004693588046150936,
      "loss": 2.6449,
      "theoretical_loss": 3.6701855481347474,
      "tokens_seen": 941817856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004693422545867221,
      "loss": 2.7135,
      "theoretical_loss": 3.6701609747111625,
      "tokens_seen": 941883392
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046932570455835066,
      "loss": 2.6046,
      "theoretical_loss": 3.6701364034760404,
      "tokens_seen": 941948928
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046930915452997915,
      "loss": 2.817,
      "theoretical_loss": 3.6701118344290333,
      "tokens_seen": 942014464
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004692926045016077,
      "loss": 2.8613,
      "theoretical_loss": 3.6700872675697944,
      "tokens_seen": 942080000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004692760544732362,
      "loss": 2.9669,
      "theoretical_loss": 3.670062702897977,
      "tokens_seen": 942145536
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046925950444486476,
      "loss": 2.8112,
      "theoretical_loss": 3.670038140413234,
      "tokens_seen": 942211072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046924295441649324,
      "loss": 2.808,
      "theoretical_loss": 3.6700135801152176,
      "tokens_seen": 942276608
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004692264043881218,
      "loss": 2.7761,
      "theoretical_loss": 3.6699890220035822,
      "tokens_seen": 942342144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004692098543597503,
      "loss": 2.5845,
      "theoretical_loss": 3.6699644660779818,
      "tokens_seen": 942407680
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046919330433137886,
      "loss": 2.6437,
      "theoretical_loss": 3.669939912338068,
      "tokens_seen": 942473216
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046917675430300734,
      "loss": 2.6596,
      "theoretical_loss": 3.669915360783495,
      "tokens_seen": 942538752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004691602042746359,
      "loss": 2.6513,
      "theoretical_loss": 3.6698908114139175,
      "tokens_seen": 942604288
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004691436542462644,
      "loss": 2.7813,
      "theoretical_loss": 3.6698662642289874,
      "tokens_seen": 942669824
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046912710421789295,
      "loss": 2.7236,
      "theoretical_loss": 3.6698417192283594,
      "tokens_seen": 942735360
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046911055418952143,
      "loss": 2.8577,
      "theoretical_loss": 3.669817176411687,
      "tokens_seen": 942800896
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046909400416114997,
      "loss": 2.6995,
      "theoretical_loss": 3.6697926357786246,
      "tokens_seen": 942866432
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004690774541327785,
      "loss": 2.5691,
      "theoretical_loss": 3.6697680973288254,
      "tokens_seen": 942931968
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046906090410440705,
      "loss": 2.619,
      "theoretical_loss": 3.669743561061944,
      "tokens_seen": 942997504
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046904435407603553,
      "loss": 2.9448,
      "theoretical_loss": 3.6697190269776345,
      "tokens_seen": 943063040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046902780404766407,
      "loss": 2.7483,
      "theoretical_loss": 3.669694495075551,
      "tokens_seen": 943128576
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004690112540192926,
      "loss": 2.8447,
      "theoretical_loss": 3.6696699653553475,
      "tokens_seen": 943194112
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046899470399092114,
      "loss": 2.5245,
      "theoretical_loss": 3.669645437816678,
      "tokens_seen": 943259648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004689781539625496,
      "loss": 2.6492,
      "theoretical_loss": 3.6696209124591985,
      "tokens_seen": 943325184
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046896160393417816,
      "loss": 2.8852,
      "theoretical_loss": 3.669596389282562,
      "tokens_seen": 943390720
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004689450539058067,
      "loss": 2.4509,
      "theoretical_loss": 3.6695718682864236,
      "tokens_seen": 943456256
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046892850387743524,
      "loss": 2.4495,
      "theoretical_loss": 3.6695473494704385,
      "tokens_seen": 943521792
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004689119538490637,
      "loss": 2.5228,
      "theoretical_loss": 3.669522832834261,
      "tokens_seen": 943587328
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046889540382069226,
      "loss": 2.6516,
      "theoretical_loss": 3.6694983183775447,
      "tokens_seen": 943652864
    },
    {
      "epoch": 0.34,
      "objective/train/avg_token_score": 0.017853079363703728,
      "objective/train/avg_weight": 0.9790512919425964,
      "objective/train/docs_used": 541407,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7964439392089844,
      "objective/train/std_weight": 0.06530439108610153,
      "objective/train/theoretical_loss": 3.6694738060999468,
      "objective/train/tokens_used": 964178400,
      "theoretical_loss": 3.6694738060999468,
      "tokens_seen": 943718400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004688788537923208,
      "loss": 2.6702,
      "theoretical_loss": 3.6694738060999468,
      "tokens_seen": 943718400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046886230376394934,
      "loss": 2.9405,
      "theoretical_loss": 3.6694492960011207,
      "tokens_seen": 943783936
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004688457537355778,
      "loss": 2.9254,
      "theoretical_loss": 3.6694247880807223,
      "tokens_seen": 943849472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046882920370720636,
      "loss": 2.8817,
      "theoretical_loss": 3.6694002823384064,
      "tokens_seen": 943915008
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004688126536788349,
      "loss": 2.4442,
      "theoretical_loss": 3.669375778773828,
      "tokens_seen": 943980544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046879610365046343,
      "loss": 2.6393,
      "theoretical_loss": 3.6693512773866424,
      "tokens_seen": 944046080
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004687795536220919,
      "loss": 2.9108,
      "theoretical_loss": 3.6693267781765058,
      "tokens_seen": 944111616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046876300359372045,
      "loss": 2.8077,
      "theoretical_loss": 3.6693022811430724,
      "tokens_seen": 944177152
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000468746453565349,
      "loss": 2.4993,
      "theoretical_loss": 3.669277786285999,
      "tokens_seen": 944242688
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046872990353697753,
      "loss": 2.6443,
      "theoretical_loss": 3.6692532936049407,
      "tokens_seen": 944308224
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000468713353508606,
      "loss": 2.4974,
      "theoretical_loss": 3.6692288030995526,
      "tokens_seen": 944373760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046869680348023455,
      "loss": 2.8117,
      "theoretical_loss": 3.669204314769491,
      "tokens_seen": 944439296
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004686802534518631,
      "loss": 2.9796,
      "theoretical_loss": 3.6691798286144124,
      "tokens_seen": 944504832
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004686637034234916,
      "loss": 2.7885,
      "theoretical_loss": 3.669155344633971,
      "tokens_seen": 944570368
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004686471533951201,
      "loss": 2.5639,
      "theoretical_loss": 3.6691308628278247,
      "tokens_seen": 944635904
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046863060336674864,
      "loss": 2.6681,
      "theoretical_loss": 3.6691063831956283,
      "tokens_seen": 944701440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004686140533383772,
      "loss": 2.906,
      "theoretical_loss": 3.669081905737039,
      "tokens_seen": 944766976
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004685975033100057,
      "loss": 2.8848,
      "theoretical_loss": 3.669057430451712,
      "tokens_seen": 944832512
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004685809532816342,
      "loss": 2.786,
      "theoretical_loss": 3.6690329573393043,
      "tokens_seen": 944898048
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046856440325326274,
      "loss": 2.7857,
      "theoretical_loss": 3.669008486399472,
      "tokens_seen": 944963584
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004685478532248913,
      "loss": 2.8383,
      "theoretical_loss": 3.668984017631872,
      "tokens_seen": 945029120
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004685313031965198,
      "loss": 2.3876,
      "theoretical_loss": 3.66895955103616,
      "tokens_seen": 945094656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004685147531681483,
      "loss": 2.5706,
      "theoretical_loss": 3.6689350866119934,
      "tokens_seen": 945160192
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046849820313977684,
      "loss": 2.8243,
      "theoretical_loss": 3.668910624359029,
      "tokens_seen": 945225728
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004684816531114054,
      "loss": 2.6637,
      "theoretical_loss": 3.668886164276923,
      "tokens_seen": 945291264
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004684651030830338,
      "loss": 2.7492,
      "theoretical_loss": 3.668861706365332,
      "tokens_seen": 945356800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046844855305466234,
      "loss": 2.5279,
      "theoretical_loss": 3.668837250623914,
      "tokens_seen": 945422336
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004684320030262909,
      "loss": 2.6291,
      "theoretical_loss": 3.6688127970523254,
      "tokens_seen": 945487872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046841545299791936,
      "loss": 2.7411,
      "theoretical_loss": 3.6687883456502233,
      "tokens_seen": 945553408
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004683989029695479,
      "loss": 2.8205,
      "theoretical_loss": 3.668763896417265,
      "tokens_seen": 945618944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046838235294117644,
      "loss": 2.5845,
      "theoretical_loss": 3.668739449353108,
      "tokens_seen": 945684480
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000468365802912805,
      "loss": 2.4987,
      "theoretical_loss": 3.6687150044574093,
      "tokens_seen": 945750016
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046834925288443346,
      "loss": 2.7367,
      "theoretical_loss": 3.6686905617298264,
      "tokens_seen": 945815552
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000468332702856062,
      "loss": 2.6955,
      "theoretical_loss": 3.6686661211700167,
      "tokens_seen": 945881088
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046831615282769053,
      "loss": 2.9895,
      "theoretical_loss": 3.668641682777638,
      "tokens_seen": 945946624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046829960279931907,
      "loss": 2.6318,
      "theoretical_loss": 3.6686172465523477,
      "tokens_seen": 946012160
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046828305277094755,
      "loss": 2.6119,
      "theoretical_loss": 3.668592812493803,
      "tokens_seen": 946077696
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004682665027425761,
      "loss": 2.6363,
      "theoretical_loss": 3.668568380601663,
      "tokens_seen": 946143232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046824995271420463,
      "loss": 2.9645,
      "theoretical_loss": 3.668543950875585,
      "tokens_seen": 946208768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046823340268583317,
      "loss": 2.963,
      "theoretical_loss": 3.6685195233152266,
      "tokens_seen": 946274304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046821685265746165,
      "loss": 2.831,
      "theoretical_loss": 3.668495097920246,
      "tokens_seen": 946339840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004682003026290902,
      "loss": 2.5195,
      "theoretical_loss": 3.668470674690302,
      "tokens_seen": 946405376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004681837526007187,
      "loss": 2.6349,
      "theoretical_loss": 3.668446253625051,
      "tokens_seen": 946470912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046816720257234726,
      "loss": 2.5214,
      "theoretical_loss": 3.6684218347241533,
      "tokens_seen": 946536448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046815065254397575,
      "loss": 2.5563,
      "theoretical_loss": 3.6683974179872667,
      "tokens_seen": 946601984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004681341025156043,
      "loss": 2.8459,
      "theoretical_loss": 3.6683730034140485,
      "tokens_seen": 946667520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004681175524872328,
      "loss": 2.6097,
      "theoretical_loss": 3.668348591004159,
      "tokens_seen": 946733056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046810100245886136,
      "loss": 2.7442,
      "theoretical_loss": 3.668324180757255,
      "tokens_seen": 946798592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046808445243048984,
      "loss": 2.6007,
      "theoretical_loss": 3.6682997726729965,
      "tokens_seen": 946864128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004680679024021184,
      "loss": 2.8213,
      "theoretical_loss": 3.6682753667510415,
      "tokens_seen": 946929664
    },
    {
      "epoch": 0.34,
      "objective/train/avg_token_score": 0.010546211153268814,
      "objective/train/avg_weight": 0.9876182675361633,
      "objective/train/docs_used": 543040,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.162337064743042,
      "objective/train/std_weight": 0.0808374211192131,
      "objective/train/theoretical_loss": 3.668250962991049,
      "objective/train/tokens_used": 967455200,
      "theoretical_loss": 3.668250962991049,
      "tokens_seen": 946995200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004680513523737469,
      "loss": 2.6395,
      "theoretical_loss": 3.668250962991049,
      "tokens_seen": 946995200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046803480234537545,
      "loss": 2.5482,
      "theoretical_loss": 3.668226561392678,
      "tokens_seen": 947060736
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046801825231700394,
      "loss": 2.6868,
      "theoretical_loss": 3.668202161955587,
      "tokens_seen": 947126272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004680017022886325,
      "loss": 2.6999,
      "theoretical_loss": 3.6681777646794362,
      "tokens_seen": 947191808
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000467985152260261,
      "loss": 2.6575,
      "theoretical_loss": 3.668153369563883,
      "tokens_seen": 947257344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046796860223188955,
      "loss": 2.6202,
      "theoretical_loss": 3.668128976608588,
      "tokens_seen": 947322880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004679520522035181,
      "loss": 2.787,
      "theoretical_loss": 3.6681045858132104,
      "tokens_seen": 947388416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046793550217514657,
      "loss": 2.8099,
      "theoretical_loss": 3.6680801971774084,
      "tokens_seen": 947453952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004679189521467751,
      "loss": 2.4567,
      "theoretical_loss": 3.668055810700843,
      "tokens_seen": 947519488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046790240211840365,
      "loss": 2.7429,
      "theoretical_loss": 3.6680314263831724,
      "tokens_seen": 947585024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004678858520900322,
      "loss": 2.5864,
      "theoretical_loss": 3.6680070442240567,
      "tokens_seen": 947650560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046786930206166067,
      "loss": 2.5322,
      "theoretical_loss": 3.667982664223156,
      "tokens_seen": 947716096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004678527520332892,
      "loss": 2.649,
      "theoretical_loss": 3.667958286380129,
      "tokens_seen": 947781632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046783620200491774,
      "loss": 2.7393,
      "theoretical_loss": 3.6679339106946363,
      "tokens_seen": 947847168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004678196519765463,
      "loss": 2.6532,
      "theoretical_loss": 3.667909537166338,
      "tokens_seen": 947912704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046780310194817476,
      "loss": 2.7706,
      "theoretical_loss": 3.667885165794893,
      "tokens_seen": 947978240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004677865519198033,
      "loss": 2.8504,
      "theoretical_loss": 3.667860796579962,
      "tokens_seen": 948043776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046777000189143184,
      "loss": 2.7222,
      "theoretical_loss": 3.6678364295212056,
      "tokens_seen": 948109312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004677534518630604,
      "loss": 2.9458,
      "theoretical_loss": 3.6678120646182837,
      "tokens_seen": 948174848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046773690183468886,
      "loss": 2.8647,
      "theoretical_loss": 3.6677877018708562,
      "tokens_seen": 948240384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004677203518063174,
      "loss": 2.644,
      "theoretical_loss": 3.6677633412785835,
      "tokens_seen": 948305920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046770380177794593,
      "loss": 2.5917,
      "theoretical_loss": 3.667738982841126,
      "tokens_seen": 948371456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046768725174957447,
      "loss": 2.7234,
      "theoretical_loss": 3.667714626558145,
      "tokens_seen": 948436992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046767070172120295,
      "loss": 2.5488,
      "theoretical_loss": 3.6676902724293003,
      "tokens_seen": 948502528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004676541516928315,
      "loss": 2.7339,
      "theoretical_loss": 3.6676659204542528,
      "tokens_seen": 948568064
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046763760166446003,
      "loss": 2.6716,
      "theoretical_loss": 3.6676415706326635,
      "tokens_seen": 948633600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046762105163608857,
      "loss": 2.6701,
      "theoretical_loss": 3.6676172229641923,
      "tokens_seen": 948699136
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046760450160771705,
      "loss": 2.5955,
      "theoretical_loss": 3.667592877448501,
      "tokens_seen": 948764672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004675879515793456,
      "loss": 2.5479,
      "theoretical_loss": 3.66756853408525,
      "tokens_seen": 948830208
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004675714015509741,
      "loss": 2.725,
      "theoretical_loss": 3.667544192874101,
      "tokens_seen": 948895744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046755485152260266,
      "loss": 2.5624,
      "theoretical_loss": 3.6675198538147153,
      "tokens_seen": 948961280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046753830149423115,
      "loss": 2.6986,
      "theoretical_loss": 3.6674955169067527,
      "tokens_seen": 949026816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004675217514658597,
      "loss": 2.6658,
      "theoretical_loss": 3.6674711821498756,
      "tokens_seen": 949092352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004675052014374882,
      "loss": 2.6638,
      "theoretical_loss": 3.6674468495437456,
      "tokens_seen": 949157888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046748865140911676,
      "loss": 2.9393,
      "theoretical_loss": 3.667422519088023,
      "tokens_seen": 949223424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004674721013807452,
      "loss": 2.678,
      "theoretical_loss": 3.66739819078237,
      "tokens_seen": 949288960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004674555513523737,
      "loss": 2.7269,
      "theoretical_loss": 3.667373864626448,
      "tokens_seen": 949354496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004674390013240022,
      "loss": 2.9272,
      "theoretical_loss": 3.6673495406199197,
      "tokens_seen": 949420032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046742245129563075,
      "loss": 2.6494,
      "theoretical_loss": 3.667325218762445,
      "tokens_seen": 949485568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004674059012672593,
      "loss": 2.8637,
      "theoretical_loss": 3.667300899053687,
      "tokens_seen": 949551104
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004673893512388878,
      "loss": 3.8566,
      "theoretical_loss": 3.667276581493307,
      "tokens_seen": 949616640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004673728012105163,
      "loss": 2.7205,
      "theoretical_loss": 3.6672522660809674,
      "tokens_seen": 949682176
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046735625118214484,
      "loss": 3.0557,
      "theoretical_loss": 3.6672279528163303,
      "tokens_seen": 949747712
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004673397011537734,
      "loss": 2.6819,
      "theoretical_loss": 3.667203641699057,
      "tokens_seen": 949813248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004673231511254019,
      "loss": 2.9077,
      "theoretical_loss": 3.6671793327288107,
      "tokens_seen": 949878784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004673066010970304,
      "loss": 2.7274,
      "theoretical_loss": 3.667155025905253,
      "tokens_seen": 949944320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046729005106865894,
      "loss": 2.7651,
      "theoretical_loss": 3.6671307212280473,
      "tokens_seen": 950009856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004672735010402875,
      "loss": 2.8845,
      "theoretical_loss": 3.6671064186968545,
      "tokens_seen": 950075392
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000467256951011916,
      "loss": 2.6556,
      "theoretical_loss": 3.6670821183113373,
      "tokens_seen": 950140928
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004672404009835445,
      "loss": 2.802,
      "theoretical_loss": 3.6670578200711597,
      "tokens_seen": 950206464
    },
    {
      "epoch": 0.34,
      "objective/train/avg_token_score": 0.019580410793423653,
      "objective/train/avg_weight": 0.9769836068153381,
      "objective/train/docs_used": 545196,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0863075256347656,
      "objective/train/std_weight": 0.12256739288568497,
      "objective/train/theoretical_loss": 3.667033523975983,
      "objective/train/tokens_used": 970732000,
      "theoretical_loss": 3.667033523975983,
      "tokens_seen": 950272000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046722385095517303,
      "loss": 2.7533,
      "theoretical_loss": 3.667033523975983,
      "tokens_seen": 950272000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046720730092680157,
      "loss": 2.8758,
      "theoretical_loss": 3.667009230025471,
      "tokens_seen": 950337536
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004671907508984301,
      "loss": 2.7972,
      "theoretical_loss": 3.6669849382192847,
      "tokens_seen": 950403072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004671742008700586,
      "loss": 2.7876,
      "theoretical_loss": 3.666960648557089,
      "tokens_seen": 950468608
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046715765084168713,
      "loss": 2.751,
      "theoretical_loss": 3.6669363610385464,
      "tokens_seen": 950534144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046714110081331567,
      "loss": 2.6318,
      "theoretical_loss": 3.6669120756633187,
      "tokens_seen": 950599680
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004671245507849442,
      "loss": 2.8487,
      "theoretical_loss": 3.66688779243107,
      "tokens_seen": 950665216
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004671080007565727,
      "loss": 2.7272,
      "theoretical_loss": 3.6668635113414636,
      "tokens_seen": 950730752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004670914507282012,
      "loss": 3.0414,
      "theoretical_loss": 3.666839232394163,
      "tokens_seen": 950796288
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046707490069982976,
      "loss": 2.8077,
      "theoretical_loss": 3.6668149555888307,
      "tokens_seen": 950861824
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004670583506714583,
      "loss": 2.7065,
      "theoretical_loss": 3.6667906809251303,
      "tokens_seen": 950927360
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004670418006430868,
      "loss": 2.5651,
      "theoretical_loss": 3.666766408402726,
      "tokens_seen": 950992896
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004670252506147153,
      "loss": 2.7685,
      "theoretical_loss": 3.666742138021281,
      "tokens_seen": 951058432
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046700870058634386,
      "loss": 2.4927,
      "theoretical_loss": 3.6667178697804577,
      "tokens_seen": 951123968
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004669921505579724,
      "loss": 2.8202,
      "theoretical_loss": 3.6666936036799216,
      "tokens_seen": 951189504
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004669756005296009,
      "loss": 2.9059,
      "theoretical_loss": 3.666669339719336,
      "tokens_seen": 951255040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004669590505012294,
      "loss": 2.854,
      "theoretical_loss": 3.666645077898364,
      "tokens_seen": 951320576
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046694250047285796,
      "loss": 2.783,
      "theoretical_loss": 3.666620818216671,
      "tokens_seen": 951386112
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004669259504444865,
      "loss": 2.6198,
      "theoretical_loss": 3.66659656067392,
      "tokens_seen": 951451648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000466909400416115,
      "loss": 2.4706,
      "theoretical_loss": 3.666572305269775,
      "tokens_seen": 951517184
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004668928503877435,
      "loss": 2.7065,
      "theoretical_loss": 3.6665480520039004,
      "tokens_seen": 951582720
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046687630035937205,
      "loss": 2.675,
      "theoretical_loss": 3.666523800875961,
      "tokens_seen": 951648256
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004668597503310006,
      "loss": 2.7249,
      "theoretical_loss": 3.66649955188562,
      "tokens_seen": 951713792
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046684320030262907,
      "loss": 2.5889,
      "theoretical_loss": 3.6664753050325425,
      "tokens_seen": 951779328
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004668266502742576,
      "loss": 2.8239,
      "theoretical_loss": 3.666451060316393,
      "tokens_seen": 951844864
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046681010024588615,
      "loss": 2.4912,
      "theoretical_loss": 3.6664268177368364,
      "tokens_seen": 951910400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004667935502175147,
      "loss": 2.6954,
      "theoretical_loss": 3.666402577293537,
      "tokens_seen": 951975936
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046677700018914317,
      "loss": 2.6637,
      "theoretical_loss": 3.666378338986159,
      "tokens_seen": 952041472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004667604501607717,
      "loss": 2.5363,
      "theoretical_loss": 3.666354102814368,
      "tokens_seen": 952107008
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046674390013240024,
      "loss": 2.6093,
      "theoretical_loss": 3.6663298687778276,
      "tokens_seen": 952172544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004667273501040288,
      "loss": 2.9447,
      "theoretical_loss": 3.6663056368762046,
      "tokens_seen": 952238080
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046671080007565726,
      "loss": 2.7346,
      "theoretical_loss": 3.6662814071091625,
      "tokens_seen": 952303616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004666942500472858,
      "loss": 2.5725,
      "theoretical_loss": 3.666257179476367,
      "tokens_seen": 952369152
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046667770001891434,
      "loss": 2.8204,
      "theoretical_loss": 3.6662329539774827,
      "tokens_seen": 952434688
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004666611499905429,
      "loss": 2.5836,
      "theoretical_loss": 3.6662087306121762,
      "tokens_seen": 952500224
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046664459996217136,
      "loss": 2.5917,
      "theoretical_loss": 3.6661845093801118,
      "tokens_seen": 952565760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004666280499337999,
      "loss": 2.7637,
      "theoretical_loss": 3.6661602902809545,
      "tokens_seen": 952631296
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046661149990542844,
      "loss": 2.8157,
      "theoretical_loss": 3.6661360733143704,
      "tokens_seen": 952696832
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000466594949877057,
      "loss": 2.8026,
      "theoretical_loss": 3.6661118584800247,
      "tokens_seen": 952762368
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046657839984868546,
      "loss": 2.6332,
      "theoretical_loss": 3.6660876457775835,
      "tokens_seen": 952827904
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000466561849820314,
      "loss": 2.4466,
      "theoretical_loss": 3.6660634352067123,
      "tokens_seen": 952893440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046654529979194253,
      "loss": 2.733,
      "theoretical_loss": 3.6660392267670767,
      "tokens_seen": 952958976
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046652874976357107,
      "loss": 2.7078,
      "theoretical_loss": 3.6660150204583424,
      "tokens_seen": 953024512
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046651219973519955,
      "loss": 2.8269,
      "theoretical_loss": 3.6659908162801758,
      "tokens_seen": 953090048
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046649564970682804,
      "loss": 2.5611,
      "theoretical_loss": 3.665966614232243,
      "tokens_seen": 953155584
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004664790996784566,
      "loss": 2.5531,
      "theoretical_loss": 3.665942414314209,
      "tokens_seen": 953221120
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046646254965008506,
      "loss": 2.9584,
      "theoretical_loss": 3.665918216525741,
      "tokens_seen": 953286656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004664459996217136,
      "loss": 2.7118,
      "theoretical_loss": 3.6658940208665047,
      "tokens_seen": 953352192
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046642944959334213,
      "loss": 2.7303,
      "theoretical_loss": 3.6658698273361665,
      "tokens_seen": 953417728
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046641289956497067,
      "loss": 2.7048,
      "theoretical_loss": 3.6658456359343927,
      "tokens_seen": 953483264
    },
    {
      "epoch": 0.34,
      "objective/train/avg_token_score": 0.016852226108312607,
      "objective/train/avg_weight": 0.9801610708236694,
      "objective/train/docs_used": 546810,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.345428943634033,
      "objective/train/std_weight": 0.1098736822605133,
      "objective/train/theoretical_loss": 3.66582144666085,
      "objective/train/tokens_used": 974008800,
      "theoretical_loss": 3.66582144666085,
      "tokens_seen": 953548800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046639634953659915,
      "loss": 2.5435,
      "theoretical_loss": 3.66582144666085,
      "tokens_seen": 953548800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004663797995082277,
      "loss": 2.7003,
      "theoretical_loss": 3.6657972595152053,
      "tokens_seen": 953614336
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046636324947985623,
      "loss": 2.7399,
      "theoretical_loss": 3.6657730744971238,
      "tokens_seen": 953679872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046634669945148477,
      "loss": 2.4808,
      "theoretical_loss": 3.6657488916062735,
      "tokens_seen": 953745408
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046633014942311325,
      "loss": 2.9067,
      "theoretical_loss": 3.665724710842321,
      "tokens_seen": 953810944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004663135993947418,
      "loss": 2.917,
      "theoretical_loss": 3.665700532204932,
      "tokens_seen": 953876480
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004662970493663703,
      "loss": 2.684,
      "theoretical_loss": 3.6656763556937744,
      "tokens_seen": 953942016
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046628049933799886,
      "loss": 2.6884,
      "theoretical_loss": 3.6656521813085154,
      "tokens_seen": 954007552
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046626394930962734,
      "loss": 2.7624,
      "theoretical_loss": 3.6656280090488216,
      "tokens_seen": 954073088
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004662473992812559,
      "loss": 2.6569,
      "theoretical_loss": 3.66560383891436,
      "tokens_seen": 954138624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004662308492528844,
      "loss": 2.6574,
      "theoretical_loss": 3.6655796709047976,
      "tokens_seen": 954204160
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046621429922451296,
      "loss": 2.6038,
      "theoretical_loss": 3.6655555050198023,
      "tokens_seen": 954269696
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046619774919614144,
      "loss": 2.444,
      "theoretical_loss": 3.665531341259041,
      "tokens_seen": 954335232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046618119916777,
      "loss": 2.6085,
      "theoretical_loss": 3.665507179622182,
      "tokens_seen": 954400768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004661646491393985,
      "loss": 2.7236,
      "theoretical_loss": 3.6654830201088906,
      "tokens_seen": 954466304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046614809911102705,
      "loss": 2.6933,
      "theoretical_loss": 3.665458862718837,
      "tokens_seen": 954531840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046613154908265554,
      "loss": 2.7545,
      "theoretical_loss": 3.6654347074516873,
      "tokens_seen": 954597376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004661149990542841,
      "loss": 2.642,
      "theoretical_loss": 3.6654105543071096,
      "tokens_seen": 954662912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004660984490259126,
      "loss": 2.5994,
      "theoretical_loss": 3.6653864032847716,
      "tokens_seen": 954728448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046608189899754115,
      "loss": 2.9011,
      "theoretical_loss": 3.665362254384341,
      "tokens_seen": 954793984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046606534896916963,
      "loss": 2.7335,
      "theoretical_loss": 3.6653381076054865,
      "tokens_seen": 954859520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046604879894079817,
      "loss": 2.4626,
      "theoretical_loss": 3.6653139629478755,
      "tokens_seen": 954925056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004660322489124267,
      "loss": 2.7706,
      "theoretical_loss": 3.6652898204111763,
      "tokens_seen": 954990592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046601569888405525,
      "loss": 2.6619,
      "theoretical_loss": 3.6652656799950565,
      "tokens_seen": 955056128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046599914885568373,
      "loss": 2.6996,
      "theoretical_loss": 3.665241541699185,
      "tokens_seen": 955121664
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046598259882731227,
      "loss": 2.8093,
      "theoretical_loss": 3.6652174055232303,
      "tokens_seen": 955187200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004659660487989408,
      "loss": 2.7073,
      "theoretical_loss": 3.66519327146686,
      "tokens_seen": 955252736
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046594949877056934,
      "loss": 2.845,
      "theoretical_loss": 3.665169139529743,
      "tokens_seen": 955318272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004659329487421978,
      "loss": 2.4298,
      "theoretical_loss": 3.665145009711548,
      "tokens_seen": 955383808
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046591639871382636,
      "loss": 2.6826,
      "theoretical_loss": 3.6651208820119434,
      "tokens_seen": 955449344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004658998486854549,
      "loss": 2.763,
      "theoretical_loss": 3.6650967564305983,
      "tokens_seen": 955514880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046588329865708344,
      "loss": 3.0256,
      "theoretical_loss": 3.6650726329671803,
      "tokens_seen": 955580416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004658667486287119,
      "loss": 2.885,
      "theoretical_loss": 3.6650485116213596,
      "tokens_seen": 955645952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046585019860034046,
      "loss": 2.8586,
      "theoretical_loss": 3.6650243923928043,
      "tokens_seen": 955711488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000465833648571969,
      "loss": 2.3896,
      "theoretical_loss": 3.6650002752811837,
      "tokens_seen": 955777024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046581709854359753,
      "loss": 2.2899,
      "theoretical_loss": 3.664976160286167,
      "tokens_seen": 955842560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000465800548515226,
      "loss": 2.5609,
      "theoretical_loss": 3.664952047407423,
      "tokens_seen": 955908096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046578399848685455,
      "loss": 2.7868,
      "theoretical_loss": 3.664927936644621,
      "tokens_seen": 955973632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004657674484584831,
      "loss": 2.8783,
      "theoretical_loss": 3.66490382799743,
      "tokens_seen": 956039168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046575089843011163,
      "loss": 2.4937,
      "theoretical_loss": 3.66487972146552,
      "tokens_seen": 956104704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004657343484017401,
      "loss": 2.6589,
      "theoretical_loss": 3.66485561704856,
      "tokens_seen": 956170240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046571779837336865,
      "loss": 2.6708,
      "theoretical_loss": 3.66483151474622,
      "tokens_seen": 956235776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004657012483449972,
      "loss": 2.8108,
      "theoretical_loss": 3.664807414558169,
      "tokens_seen": 956301312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004656846983166257,
      "loss": 2.7004,
      "theoretical_loss": 3.664783316484076,
      "tokens_seen": 956366848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004656681482882542,
      "loss": 2.8036,
      "theoretical_loss": 3.664759220523613,
      "tokens_seen": 956432384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046565159825988275,
      "loss": 2.6183,
      "theoretical_loss": 3.664735126676448,
      "tokens_seen": 956497920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004656350482315113,
      "loss": 2.7578,
      "theoretical_loss": 3.6647110349422505,
      "tokens_seen": 956563456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004656184982031398,
      "loss": 2.7582,
      "theoretical_loss": 3.6646869453206925,
      "tokens_seen": 956628992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004656019481747683,
      "loss": 2.7521,
      "theoretical_loss": 3.6646628578114417,
      "tokens_seen": 956694528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046558539814639684,
      "loss": 2.9812,
      "theoretical_loss": 3.66463877241417,
      "tokens_seen": 956760064
    },
    {
      "epoch": 0.34,
      "objective/train/avg_token_score": 0.008166112005710602,
      "objective/train/avg_weight": 0.9903982281684875,
      "objective/train/docs_used": 548880,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.897420644760132,
      "objective/train/std_weight": 0.03776023909449577,
      "objective/train/theoretical_loss": 3.664614689128546,
      "objective/train/tokens_used": 977285600,
      "theoretical_loss": 3.664614689128546,
      "tokens_seen": 956825600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004655688481180254,
      "loss": 2.8578,
      "theoretical_loss": 3.664614689128546,
      "tokens_seen": 956825600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004655522980896539,
      "loss": 2.5222,
      "theoretical_loss": 3.6645906079542416,
      "tokens_seen": 956891136
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004655357480612824,
      "loss": 2.7713,
      "theoretical_loss": 3.6645665288909255,
      "tokens_seen": 956956672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004655191980329109,
      "loss": 2.5957,
      "theoretical_loss": 3.66454245193827,
      "tokens_seen": 957022208
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004655026480045394,
      "loss": 2.8176,
      "theoretical_loss": 3.664518377095944,
      "tokens_seen": 957087744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004654860979761679,
      "loss": 2.6814,
      "theoretical_loss": 3.6644943043636182,
      "tokens_seen": 957153280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046546954794779644,
      "loss": 2.6478,
      "theoretical_loss": 3.6644702337409636,
      "tokens_seen": 957218816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000465452997919425,
      "loss": 2.9331,
      "theoretical_loss": 3.664446165227651,
      "tokens_seen": 957284352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004654364478910535,
      "loss": 2.7911,
      "theoretical_loss": 3.6644220988233513,
      "tokens_seen": 957349888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000465419897862682,
      "loss": 2.7521,
      "theoretical_loss": 3.6643980345277347,
      "tokens_seen": 957415424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046540334783431054,
      "loss": 2.6464,
      "theoretical_loss": 3.6643739723404725,
      "tokens_seen": 957480960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004653867978059391,
      "loss": 2.5928,
      "theoretical_loss": 3.664349912261236,
      "tokens_seen": 957546496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004653702477775676,
      "loss": 2.7996,
      "theoretical_loss": 3.6643258542896957,
      "tokens_seen": 957612032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004653536977491961,
      "loss": 2.5358,
      "theoretical_loss": 3.664301798425523,
      "tokens_seen": 957677568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046533714772082463,
      "loss": 2.7151,
      "theoretical_loss": 3.664277744668389,
      "tokens_seen": 957743104
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046532059769245317,
      "loss": 2.8141,
      "theoretical_loss": 3.664253693017965,
      "tokens_seen": 957808640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004653040476640817,
      "loss": 2.6104,
      "theoretical_loss": 3.6642296434739228,
      "tokens_seen": 957874176
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004652874976357102,
      "loss": 2.6956,
      "theoretical_loss": 3.664205596035933,
      "tokens_seen": 957939712
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046527094760733873,
      "loss": 2.8259,
      "theoretical_loss": 3.6641815507036677,
      "tokens_seen": 958005248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046525439757896727,
      "loss": 2.524,
      "theoretical_loss": 3.6641575074767982,
      "tokens_seen": 958070784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004652378475505958,
      "loss": 2.8315,
      "theoretical_loss": 3.664133466354997,
      "tokens_seen": 958136320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004652212975222243,
      "loss": 2.7074,
      "theoretical_loss": 3.664109427337934,
      "tokens_seen": 958201856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004652047474938528,
      "loss": 2.7515,
      "theoretical_loss": 3.664085390425283,
      "tokens_seen": 958267392
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046518819746548136,
      "loss": 2.6886,
      "theoretical_loss": 3.6640613556167145,
      "tokens_seen": 958332928
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004651716474371099,
      "loss": 2.7326,
      "theoretical_loss": 3.664037322911901,
      "tokens_seen": 958398464
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004651550974087384,
      "loss": 2.6756,
      "theoretical_loss": 3.664013292310515,
      "tokens_seen": 958464000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004651385473803669,
      "loss": 2.7284,
      "theoretical_loss": 3.6639892638122276,
      "tokens_seen": 958529536
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046512199735199546,
      "loss": 2.7465,
      "theoretical_loss": 3.6639652374167113,
      "tokens_seen": 958595072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000465105447323624,
      "loss": 2.807,
      "theoretical_loss": 3.663941213123639,
      "tokens_seen": 958660608
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004650888972952525,
      "loss": 2.6223,
      "theoretical_loss": 3.6639171909326818,
      "tokens_seen": 958726144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000465072347266881,
      "loss": 2.5951,
      "theoretical_loss": 3.6638931708435134,
      "tokens_seen": 958791680
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046505579723850956,
      "loss": 2.6226,
      "theoretical_loss": 3.6638691528558054,
      "tokens_seen": 958857216
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004650392472101381,
      "loss": 2.8266,
      "theoretical_loss": 3.663845136969231,
      "tokens_seen": 958922752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004650226971817666,
      "loss": 2.6782,
      "theoretical_loss": 3.6638211231834616,
      "tokens_seen": 958988288
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004650061471533951,
      "loss": 2.9623,
      "theoretical_loss": 3.6637971114981713,
      "tokens_seen": 959053824
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046498959712502365,
      "loss": 2.8135,
      "theoretical_loss": 3.6637731019130317,
      "tokens_seen": 959119360
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004649730470966522,
      "loss": 2.8655,
      "theoretical_loss": 3.6637490944277165,
      "tokens_seen": 959184896
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046495649706828067,
      "loss": 2.6977,
      "theoretical_loss": 3.663725089041898,
      "tokens_seen": 959250432
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004649399470399092,
      "loss": 2.6631,
      "theoretical_loss": 3.6637010857552497,
      "tokens_seen": 959315968
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046492339701153775,
      "loss": 2.569,
      "theoretical_loss": 3.6636770845674445,
      "tokens_seen": 959381504
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004649068469831663,
      "loss": 2.679,
      "theoretical_loss": 3.6636530854781553,
      "tokens_seen": 959447040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046489029695479477,
      "loss": 2.6535,
      "theoretical_loss": 3.6636290884870553,
      "tokens_seen": 959512576
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004648737469264233,
      "loss": 2.6114,
      "theoretical_loss": 3.663605093593818,
      "tokens_seen": 959578112
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046485719689805184,
      "loss": 2.5853,
      "theoretical_loss": 3.6635811007981163,
      "tokens_seen": 959643648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004648406468696804,
      "loss": 2.7599,
      "theoretical_loss": 3.6635571100996245,
      "tokens_seen": 959709184
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046482409684130886,
      "loss": 2.5683,
      "theoretical_loss": 3.6635331214980154,
      "tokens_seen": 959774720
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004648075468129374,
      "loss": 2.7127,
      "theoretical_loss": 3.663509134992962,
      "tokens_seen": 959840256
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046479099678456594,
      "loss": 2.9056,
      "theoretical_loss": 3.6634851505841395,
      "tokens_seen": 959905792
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004647744467561945,
      "loss": 2.6124,
      "theoretical_loss": 3.66346116827122,
      "tokens_seen": 959971328
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046475789672782296,
      "loss": 2.8974,
      "theoretical_loss": 3.6634371880538783,
      "tokens_seen": 960036864
    },
    {
      "epoch": 0.34,
      "objective/train/avg_token_score": 0.00573047436773777,
      "objective/train/avg_weight": 0.993259072303772,
      "objective/train/docs_used": 550512,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.811941385269165,
      "objective/train/std_weight": 0.032185349613428116,
      "objective/train/theoretical_loss": 3.6634132099317886,
      "objective/train/tokens_used": 980562400,
      "theoretical_loss": 3.6634132099317886,
      "tokens_seen": 960102400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004647413466994515,
      "loss": 2.7694,
      "theoretical_loss": 3.6634132099317886,
      "tokens_seen": 960102400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046472479667108004,
      "loss": 2.6697,
      "theoretical_loss": 3.6633892339046232,
      "tokens_seen": 960167936
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046470824664270857,
      "loss": 2.4599,
      "theoretical_loss": 3.663365259972058,
      "tokens_seen": 960233472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046469169661433706,
      "loss": 2.6454,
      "theoretical_loss": 3.6633412881337653,
      "tokens_seen": 960299008
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004646751465859656,
      "loss": 2.6218,
      "theoretical_loss": 3.6633173183894208,
      "tokens_seen": 960364544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046465859655759413,
      "loss": 2.7849,
      "theoretical_loss": 3.6632933507386976,
      "tokens_seen": 960430080
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046464204652922267,
      "loss": 2.5855,
      "theoretical_loss": 3.6632693851812705,
      "tokens_seen": 960495616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046462549650085115,
      "loss": 2.693,
      "theoretical_loss": 3.663245421716814,
      "tokens_seen": 960561152
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004646089464724797,
      "loss": 2.7048,
      "theoretical_loss": 3.663221460345002,
      "tokens_seen": 960626688
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046459239644410823,
      "loss": 2.7921,
      "theoretical_loss": 3.6631975010655093,
      "tokens_seen": 960692224
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046457584641573676,
      "loss": 2.6793,
      "theoretical_loss": 3.6631735438780106,
      "tokens_seen": 960757760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046455929638736525,
      "loss": 2.8464,
      "theoretical_loss": 3.663149588782181,
      "tokens_seen": 960823296
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004645427463589938,
      "loss": 2.8093,
      "theoretical_loss": 3.6631256357776945,
      "tokens_seen": 960888832
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004645261963306222,
      "loss": 2.6962,
      "theoretical_loss": 3.6631016848642255,
      "tokens_seen": 960954368
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046450964630225075,
      "loss": 2.7514,
      "theoretical_loss": 3.6630777360414504,
      "tokens_seen": 961019904
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004644930962738793,
      "loss": 2.6888,
      "theoretical_loss": 3.6630537893090422,
      "tokens_seen": 961085440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046447654624550783,
      "loss": 2.896,
      "theoretical_loss": 3.663029844666678,
      "tokens_seen": 961150976
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004644599962171363,
      "loss": 2.6717,
      "theoretical_loss": 3.663005902114031,
      "tokens_seen": 961216512
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046444344618876485,
      "loss": 2.9026,
      "theoretical_loss": 3.6629819616507775,
      "tokens_seen": 961282048
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004644268961603934,
      "loss": 2.7185,
      "theoretical_loss": 3.662958023276593,
      "tokens_seen": 961347584
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004644103461320219,
      "loss": 2.7187,
      "theoretical_loss": 3.662934086991151,
      "tokens_seen": 961413120
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004643937961036504,
      "loss": 2.5473,
      "theoretical_loss": 3.6629101527941295,
      "tokens_seen": 961478656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046437724607527894,
      "loss": 2.6466,
      "theoretical_loss": 3.662886220685202,
      "tokens_seen": 961544192
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004643606960469075,
      "loss": 2.8066,
      "theoretical_loss": 3.6628622906640445,
      "tokens_seen": 961609728
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000464344146018536,
      "loss": 2.6548,
      "theoretical_loss": 3.6628383627303327,
      "tokens_seen": 961675264
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004643275959901645,
      "loss": 2.8664,
      "theoretical_loss": 3.6628144368837425,
      "tokens_seen": 961740800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046431104596179304,
      "loss": 2.7966,
      "theoretical_loss": 3.662790513123949,
      "tokens_seen": 961806336
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004642944959334216,
      "loss": 2.8258,
      "theoretical_loss": 3.662766591450629,
      "tokens_seen": 961871872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004642779459050501,
      "loss": 2.8181,
      "theoretical_loss": 3.662742671863458,
      "tokens_seen": 961937408
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004642613958766786,
      "loss": 2.9313,
      "theoretical_loss": 3.6627187543621114,
      "tokens_seen": 962002944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046424484584830714,
      "loss": 2.9181,
      "theoretical_loss": 3.6626948389462655,
      "tokens_seen": 962068480
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004642282958199357,
      "loss": 2.5753,
      "theoretical_loss": 3.662670925615597,
      "tokens_seen": 962134016
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004642117457915642,
      "loss": 2.9215,
      "theoretical_loss": 3.6626470143697816,
      "tokens_seen": 962199552
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004641951957631927,
      "loss": 2.773,
      "theoretical_loss": 3.662623105208495,
      "tokens_seen": 962265088
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046417864573482123,
      "loss": 2.6157,
      "theoretical_loss": 3.6625991981314145,
      "tokens_seen": 962330624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046416209570644977,
      "loss": 2.4229,
      "theoretical_loss": 3.6625752931382163,
      "tokens_seen": 962396160
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004641455456780783,
      "loss": 2.6523,
      "theoretical_loss": 3.6625513902285767,
      "tokens_seen": 962461696
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004641289956497068,
      "loss": 2.7721,
      "theoretical_loss": 3.6625274894021724,
      "tokens_seen": 962527232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046411244562133533,
      "loss": 2.4738,
      "theoretical_loss": 3.6625035906586794,
      "tokens_seen": 962592768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046409589559296387,
      "loss": 2.716,
      "theoretical_loss": 3.662479693997775,
      "tokens_seen": 962658304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004640793455645924,
      "loss": 2.6572,
      "theoretical_loss": 3.662455799419136,
      "tokens_seen": 962723840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004640627955362209,
      "loss": 2.7534,
      "theoretical_loss": 3.6624319069224387,
      "tokens_seen": 962789376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004640462455078494,
      "loss": 2.6509,
      "theoretical_loss": 3.66240801650736,
      "tokens_seen": 962854912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046402969547947796,
      "loss": 2.8563,
      "theoretical_loss": 3.662384128173578,
      "tokens_seen": 962920448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004640131454511065,
      "loss": 2.7494,
      "theoretical_loss": 3.6623602419207684,
      "tokens_seen": 962985984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046399659542273504,
      "loss": 2.9575,
      "theoretical_loss": 3.662336357748609,
      "tokens_seen": 963051520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004639800453943635,
      "loss": 2.5785,
      "theoretical_loss": 3.662312475656777,
      "tokens_seen": 963117056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046396349536599206,
      "loss": 2.7199,
      "theoretical_loss": 3.6622885956449496,
      "tokens_seen": 963182592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004639469453376206,
      "loss": 2.7614,
      "theoretical_loss": 3.6622647177128034,
      "tokens_seen": 963248128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046393039530924913,
      "loss": 2.6794,
      "theoretical_loss": 3.662240841860017,
      "tokens_seen": 963313664
    },
    {
      "epoch": 0.34,
      "objective/train/avg_token_score": 0.01566668041050434,
      "objective/train/avg_weight": 0.9815562963485718,
      "objective/train/docs_used": 552388,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8503410816192627,
      "objective/train/std_weight": 0.10024194419384003,
      "objective/train/theoretical_loss": 3.662216968086267,
      "objective/train/tokens_used": 983839200,
      "theoretical_loss": 3.662216968086267,
      "tokens_seen": 963379200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004639138452808776,
      "loss": 2.724,
      "theoretical_loss": 3.662216968086267,
      "tokens_seen": 963379200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046389729525250615,
      "loss": 2.8311,
      "theoretical_loss": 3.6621930963912317,
      "tokens_seen": 963444736
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004638807452241347,
      "loss": 2.8712,
      "theoretical_loss": 3.662169226774588,
      "tokens_seen": 963510272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046386419519576323,
      "loss": 2.5901,
      "theoretical_loss": 3.662145359236014,
      "tokens_seen": 963575808
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004638476451673917,
      "loss": 2.8077,
      "theoretical_loss": 3.6621214937751874,
      "tokens_seen": 963641344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046383109513902025,
      "loss": 2.506,
      "theoretical_loss": 3.662097630391786,
      "tokens_seen": 963706880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004638145451106488,
      "loss": 2.6841,
      "theoretical_loss": 3.662073769085488,
      "tokens_seen": 963772416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004637979950822773,
      "loss": 2.5984,
      "theoretical_loss": 3.6620499098559707,
      "tokens_seen": 963837952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004637814450539058,
      "loss": 2.615,
      "theoretical_loss": 3.6620260527029127,
      "tokens_seen": 963903488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046376489502553435,
      "loss": 2.7441,
      "theoretical_loss": 3.662002197625992,
      "tokens_seen": 963969024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004637483449971629,
      "loss": 2.8227,
      "theoretical_loss": 3.661978344624887,
      "tokens_seen": 964034560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004637317949687914,
      "loss": 2.8719,
      "theoretical_loss": 3.6619544936992754,
      "tokens_seen": 964100096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004637152449404199,
      "loss": 2.9522,
      "theoretical_loss": 3.661930644848836,
      "tokens_seen": 964165632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046369869491204844,
      "loss": 2.8668,
      "theoretical_loss": 3.661906798073247,
      "tokens_seen": 964231168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000463682144883677,
      "loss": 2.7225,
      "theoretical_loss": 3.6618829533721873,
      "tokens_seen": 964296704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004636655948553055,
      "loss": 2.8302,
      "theoretical_loss": 3.661859110745335,
      "tokens_seen": 964362240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000463649044826934,
      "loss": 2.7607,
      "theoretical_loss": 3.661835270192369,
      "tokens_seen": 964427776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046363249479856254,
      "loss": 2.7434,
      "theoretical_loss": 3.6618114317129677,
      "tokens_seen": 964493312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004636159447701911,
      "loss": 3.0195,
      "theoretical_loss": 3.6617875953068095,
      "tokens_seen": 964558848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004635993947418196,
      "loss": 2.669,
      "theoretical_loss": 3.6617637609735745,
      "tokens_seen": 964624384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004635828447134481,
      "loss": 2.7702,
      "theoretical_loss": 3.661739928712941,
      "tokens_seen": 964689920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046356629468507663,
      "loss": 2.7493,
      "theoretical_loss": 3.661716098524588,
      "tokens_seen": 964755456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046354974465670506,
      "loss": 2.6049,
      "theoretical_loss": 3.661692270408194,
      "tokens_seen": 964820992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004635331946283336,
      "loss": 2.9405,
      "theoretical_loss": 3.661668444363438,
      "tokens_seen": 964886528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046351664459996214,
      "loss": 2.5993,
      "theoretical_loss": 3.6616446203900006,
      "tokens_seen": 964952064
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004635000945715907,
      "loss": 2.6091,
      "theoretical_loss": 3.6616207984875606,
      "tokens_seen": 965017600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046348354454321916,
      "loss": 2.7891,
      "theoretical_loss": 3.6615969786557967,
      "tokens_seen": 965083136
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004634669945148477,
      "loss": 2.9137,
      "theoretical_loss": 3.6615731608943882,
      "tokens_seen": 965148672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046345044448647623,
      "loss": 2.7635,
      "theoretical_loss": 3.6615493452030154,
      "tokens_seen": 965214208
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046343389445810477,
      "loss": 2.8074,
      "theoretical_loss": 3.661525531581357,
      "tokens_seen": 965279744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046341734442973325,
      "loss": 2.9663,
      "theoretical_loss": 3.6615017200290936,
      "tokens_seen": 965345280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004634007944013618,
      "loss": 2.6293,
      "theoretical_loss": 3.661477910545904,
      "tokens_seen": 965410816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046338424437299033,
      "loss": 2.8098,
      "theoretical_loss": 3.6614541031314687,
      "tokens_seen": 965476352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046336769434461887,
      "loss": 2.9261,
      "theoretical_loss": 3.6614302977854667,
      "tokens_seen": 965541888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046335114431624735,
      "loss": 2.7948,
      "theoretical_loss": 3.6614064945075784,
      "tokens_seen": 965607424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004633345942878759,
      "loss": 2.7777,
      "theoretical_loss": 3.661382693297484,
      "tokens_seen": 965672960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0004633180442595044,
      "loss": 2.7248,
      "theoretical_loss": 3.661358894154864,
      "tokens_seen": 965738496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046330149423113296,
      "loss": 2.7074,
      "theoretical_loss": 3.661335097079397,
      "tokens_seen": 965804032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046328494420276145,
      "loss": 2.61,
      "theoretical_loss": 3.6613113020707644,
      "tokens_seen": 965869568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00046326839417439,
      "loss": 2.8614,
      "theoretical_loss": 3.6612875091286465,
      "tokens_seen": 965935104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004632518441460185,
      "loss": 2.8664,
      "theoretical_loss": 3.661263718252723,
      "tokens_seen": 966000640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046323529411764706,
      "loss": 2.6323,
      "theoretical_loss": 3.661239929442674,
      "tokens_seen": 966066176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046321874408927554,
      "loss": 2.582,
      "theoretical_loss": 3.6612161426981817,
      "tokens_seen": 966131712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004632021940609041,
      "loss": 2.7244,
      "theoretical_loss": 3.6611923580189245,
      "tokens_seen": 966197248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004631856440325326,
      "loss": 2.7971,
      "theoretical_loss": 3.661168575404585,
      "tokens_seen": 966262784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046316909400416115,
      "loss": 2.6484,
      "theoretical_loss": 3.6611447948548426,
      "tokens_seen": 966328320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046315254397578964,
      "loss": 2.4549,
      "theoretical_loss": 3.6611210163693784,
      "tokens_seen": 966393856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004631359939474182,
      "loss": 2.7906,
      "theoretical_loss": 3.661097239947873,
      "tokens_seen": 966459392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004631194439190467,
      "loss": 2.7027,
      "theoretical_loss": 3.661073465590008,
      "tokens_seen": 966524928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046310289389067525,
      "loss": 2.7147,
      "theoretical_loss": 3.6610496932954635,
      "tokens_seen": 966590464
    },
    {
      "epoch": 0.35,
      "objective/train/avg_token_score": 0.009899495169520378,
      "objective/train/avg_weight": 0.9883555769920349,
      "objective/train/docs_used": 553451,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7601523399353027,
      "objective/train/std_weight": 0.05839155241847038,
      "objective/train/theoretical_loss": 3.6610259230639217,
      "objective/train/tokens_used": 987116000,
      "theoretical_loss": 3.6610259230639217,
      "tokens_seen": 966656000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046308634386230373,
      "loss": 2.8349,
      "theoretical_loss": 3.6610259230639217,
      "tokens_seen": 966656000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046306979383393227,
      "loss": 2.8582,
      "theoretical_loss": 3.661002154895062,
      "tokens_seen": 966721536
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004630532438055608,
      "loss": 2.7928,
      "theoretical_loss": 3.6609783887885676,
      "tokens_seen": 966787072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046303669377718935,
      "loss": 2.7634,
      "theoretical_loss": 3.6609546247441185,
      "tokens_seen": 966852608
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046302014374881783,
      "loss": 2.8771,
      "theoretical_loss": 3.660930862761396,
      "tokens_seen": 966918144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046300359372044637,
      "loss": 2.5798,
      "theoretical_loss": 3.660907102840082,
      "tokens_seen": 966983680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004629870436920749,
      "loss": 2.8734,
      "theoretical_loss": 3.660883344979858,
      "tokens_seen": 967049216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046297049366370344,
      "loss": 2.7227,
      "theoretical_loss": 3.6608595891804057,
      "tokens_seen": 967114752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004629539436353319,
      "loss": 2.8285,
      "theoretical_loss": 3.660835835441406,
      "tokens_seen": 967180288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046293739360696046,
      "loss": 2.5151,
      "theoretical_loss": 3.660812083762541,
      "tokens_seen": 967245824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000462920843578589,
      "loss": 2.8406,
      "theoretical_loss": 3.660788334143492,
      "tokens_seen": 967311360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046290429355021754,
      "loss": 2.7911,
      "theoretical_loss": 3.6607645865839418,
      "tokens_seen": 967376896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000462887743521846,
      "loss": 2.8311,
      "theoretical_loss": 3.6607408410835713,
      "tokens_seen": 967442432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046287119349347456,
      "loss": 2.4832,
      "theoretical_loss": 3.6607170976420638,
      "tokens_seen": 967507968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004628546434651031,
      "loss": 2.7764,
      "theoretical_loss": 3.6606933562590998,
      "tokens_seen": 967573504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046283809343673163,
      "loss": 2.6602,
      "theoretical_loss": 3.660669616934362,
      "tokens_seen": 967639040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004628215434083601,
      "loss": 2.5528,
      "theoretical_loss": 3.660645879667533,
      "tokens_seen": 967704576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046280499337998866,
      "loss": 2.7777,
      "theoretical_loss": 3.6606221444582943,
      "tokens_seen": 967770112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004627884433516172,
      "loss": 2.8687,
      "theoretical_loss": 3.6605984113063292,
      "tokens_seen": 967835648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046277189332324573,
      "loss": 2.8917,
      "theoretical_loss": 3.6605746802113197,
      "tokens_seen": 967901184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004627553432948742,
      "loss": 2.8999,
      "theoretical_loss": 3.6605509511729473,
      "tokens_seen": 967966720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046273879326650275,
      "loss": 2.7652,
      "theoretical_loss": 3.660527224190896,
      "tokens_seen": 968032256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004627222432381313,
      "loss": 2.8946,
      "theoretical_loss": 3.6605034992648475,
      "tokens_seen": 968097792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004627056932097598,
      "loss": 2.7795,
      "theoretical_loss": 3.6604797763944843,
      "tokens_seen": 968163328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004626891431813883,
      "loss": 2.8706,
      "theoretical_loss": 3.6604560555794903,
      "tokens_seen": 968228864
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046267259315301685,
      "loss": 2.7374,
      "theoretical_loss": 3.6604323368195475,
      "tokens_seen": 968294400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004626560431246454,
      "loss": 2.9103,
      "theoretical_loss": 3.660408620114339,
      "tokens_seen": 968359936
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004626394930962739,
      "loss": 2.7129,
      "theoretical_loss": 3.6603849054635473,
      "tokens_seen": 968425472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004626229430679024,
      "loss": 2.8002,
      "theoretical_loss": 3.660361192866856,
      "tokens_seen": 968491008
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046260639303953094,
      "loss": 2.7998,
      "theoretical_loss": 3.660337482323948,
      "tokens_seen": 968556544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004625898430111595,
      "loss": 2.6249,
      "theoretical_loss": 3.660313773834506,
      "tokens_seen": 968622080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000462573292982788,
      "loss": 2.9141,
      "theoretical_loss": 3.660290067398214,
      "tokens_seen": 968687616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046255674295441645,
      "loss": 2.9478,
      "theoretical_loss": 3.660266363014755,
      "tokens_seen": 968753152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000462540192926045,
      "loss": 2.7552,
      "theoretical_loss": 3.6602426606838128,
      "tokens_seen": 968818688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004625236428976735,
      "loss": 2.7541,
      "theoretical_loss": 3.6602189604050697,
      "tokens_seen": 968884224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000462507092869302,
      "loss": 2.8202,
      "theoretical_loss": 3.6601952621782106,
      "tokens_seen": 968949760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046249054284093054,
      "loss": 2.612,
      "theoretical_loss": 3.660171566002918,
      "tokens_seen": 969015296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004624739928125591,
      "loss": 2.8091,
      "theoretical_loss": 3.660147871878876,
      "tokens_seen": 969080832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004624574427841876,
      "loss": 2.684,
      "theoretical_loss": 3.660124179805769,
      "tokens_seen": 969146368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004624408927558161,
      "loss": 2.5672,
      "theoretical_loss": 3.6601004897832796,
      "tokens_seen": 969211904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046242434272744464,
      "loss": 2.6656,
      "theoretical_loss": 3.6600768018110923,
      "tokens_seen": 969277440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004624077926990732,
      "loss": 2.9175,
      "theoretical_loss": 3.660053115888891,
      "tokens_seen": 969342976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004623912426707017,
      "loss": 2.8589,
      "theoretical_loss": 3.6600294320163593,
      "tokens_seen": 969408512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004623746926423302,
      "loss": 2.7524,
      "theoretical_loss": 3.660005750193182,
      "tokens_seen": 969474048
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046235814261395874,
      "loss": 2.7327,
      "theoretical_loss": 3.6599820704190433,
      "tokens_seen": 969539584
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004623415925855873,
      "loss": 2.8251,
      "theoretical_loss": 3.6599583926936265,
      "tokens_seen": 969605120
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004623250425572158,
      "loss": 2.7534,
      "theoretical_loss": 3.6599347170166165,
      "tokens_seen": 969670656
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004623084925288443,
      "loss": 2.7226,
      "theoretical_loss": 3.659911043387698,
      "tokens_seen": 969736192
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046229194250047283,
      "loss": 2.8283,
      "theoretical_loss": 3.6598873718065548,
      "tokens_seen": 969801728
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046227539247210137,
      "loss": 2.8116,
      "theoretical_loss": 3.6598637022728715,
      "tokens_seen": 969867264
    },
    {
      "epoch": 0.35,
      "objective/train/avg_token_score": 0.013452756218612194,
      "objective/train/avg_weight": 0.9841771721839905,
      "objective/train/docs_used": 554109,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.95772123336792,
      "objective/train/std_weight": 0.08248185366392136,
      "objective/train/theoretical_loss": 3.659840034786333,
      "objective/train/tokens_used": 990392800,
      "theoretical_loss": 3.659840034786333,
      "tokens_seen": 969932800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004622588424437299,
      "loss": 2.6931,
      "theoretical_loss": 3.659840034786333,
      "tokens_seen": 969932800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004622422924153584,
      "loss": 2.9229,
      "theoretical_loss": 3.6598163693466232,
      "tokens_seen": 969998336
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046222574238698693,
      "loss": 2.6886,
      "theoretical_loss": 3.6597927059534276,
      "tokens_seen": 970063872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046220919235861547,
      "loss": 2.9037,
      "theoretical_loss": 3.6597690446064313,
      "tokens_seen": 970129408
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000462192642330244,
      "loss": 2.742,
      "theoretical_loss": 3.659745385305318,
      "tokens_seen": 970194944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004621760923018725,
      "loss": 2.9937,
      "theoretical_loss": 3.6597217280497736,
      "tokens_seen": 970260480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000462159542273501,
      "loss": 2.6525,
      "theoretical_loss": 3.659698072839482,
      "tokens_seen": 970326016
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046214299224512956,
      "loss": 2.8287,
      "theoretical_loss": 3.6596744196741295,
      "tokens_seen": 970391552
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004621264422167581,
      "loss": 2.8231,
      "theoretical_loss": 3.6596507685534005,
      "tokens_seen": 970457088
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004621098921883866,
      "loss": 2.9091,
      "theoretical_loss": 3.6596271194769807,
      "tokens_seen": 970522624
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004620933421600151,
      "loss": 2.7101,
      "theoretical_loss": 3.659603472444555,
      "tokens_seen": 970588160
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046207679213164366,
      "loss": 3.0453,
      "theoretical_loss": 3.6595798274558087,
      "tokens_seen": 970653696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004620602421032722,
      "loss": 3.0838,
      "theoretical_loss": 3.6595561845104276,
      "tokens_seen": 970719232
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004620436920749007,
      "loss": 2.8122,
      "theoretical_loss": 3.6595325436080968,
      "tokens_seen": 970784768
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004620271420465292,
      "loss": 2.4264,
      "theoretical_loss": 3.6595089047485017,
      "tokens_seen": 970850304
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046201059201815775,
      "loss": 2.7734,
      "theoretical_loss": 3.6594852679313283,
      "tokens_seen": 970915840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004619940419897863,
      "loss": 2.8071,
      "theoretical_loss": 3.659461633156262,
      "tokens_seen": 970981376
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004619774919614148,
      "loss": 2.7396,
      "theoretical_loss": 3.6594380004229894,
      "tokens_seen": 971046912
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004619609419330433,
      "loss": 2.861,
      "theoretical_loss": 3.659414369731195,
      "tokens_seen": 971112448
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046194439190467185,
      "loss": 3.0376,
      "theoretical_loss": 3.659390741080565,
      "tokens_seen": 971177984
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004619278418763004,
      "loss": 2.8701,
      "theoretical_loss": 3.659367114470786,
      "tokens_seen": 971243520
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046191129184792887,
      "loss": 2.7,
      "theoretical_loss": 3.659343489901544,
      "tokens_seen": 971309056
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004618947418195574,
      "loss": 2.7439,
      "theoretical_loss": 3.659319867372525,
      "tokens_seen": 971374592
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046187819179118594,
      "loss": 2.5486,
      "theoretical_loss": 3.6592962468834145,
      "tokens_seen": 971440128
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004618616417628145,
      "loss": 2.828,
      "theoretical_loss": 3.6592726284339,
      "tokens_seen": 971505664
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046184509173444297,
      "loss": 2.7128,
      "theoretical_loss": 3.6592490120236665,
      "tokens_seen": 971571200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004618285417060715,
      "loss": 2.7248,
      "theoretical_loss": 3.6592253976524014,
      "tokens_seen": 971636736
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046181199167770004,
      "loss": 2.7672,
      "theoretical_loss": 3.65920178531979,
      "tokens_seen": 971702272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004617954416493286,
      "loss": 2.751,
      "theoretical_loss": 3.65917817502552,
      "tokens_seen": 971767808
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046177889162095706,
      "loss": 2.9614,
      "theoretical_loss": 3.6591545667692777,
      "tokens_seen": 971833344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004617623415925856,
      "loss": 3.0477,
      "theoretical_loss": 3.6591309605507494,
      "tokens_seen": 971898880
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046174579156421414,
      "loss": 2.5745,
      "theoretical_loss": 3.6591073563696224,
      "tokens_seen": 971964416
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004617292415358427,
      "loss": 3.0308,
      "theoretical_loss": 3.6590837542255827,
      "tokens_seen": 972029952
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046171269150747116,
      "loss": 2.8491,
      "theoretical_loss": 3.6590601541183174,
      "tokens_seen": 972095488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004616961414790997,
      "loss": 2.9285,
      "theoretical_loss": 3.659036556047514,
      "tokens_seen": 972161024
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046167959145072823,
      "loss": 2.7933,
      "theoretical_loss": 3.659012960012859,
      "tokens_seen": 972226560
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046166304142235677,
      "loss": 2.8893,
      "theoretical_loss": 3.65898936601404,
      "tokens_seen": 972292096
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046164649139398525,
      "loss": 2.9663,
      "theoretical_loss": 3.6589657740507437,
      "tokens_seen": 972357632
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004616299413656138,
      "loss": 3.0494,
      "theoretical_loss": 3.658942184122657,
      "tokens_seen": 972423168
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046161339133724233,
      "loss": 2.7729,
      "theoretical_loss": 3.6589185962294675,
      "tokens_seen": 972488704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046159684130887087,
      "loss": 2.8803,
      "theoretical_loss": 3.658895010370863,
      "tokens_seen": 972554240
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004615802912804993,
      "loss": 2.9556,
      "theoretical_loss": 3.65887142654653,
      "tokens_seen": 972619776
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046156374125212783,
      "loss": 2.8727,
      "theoretical_loss": 3.658847844756157,
      "tokens_seen": 972685312
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004615471912237563,
      "loss": 2.8716,
      "theoretical_loss": 3.6588242649994314,
      "tokens_seen": 972750848
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046153064119538485,
      "loss": 3.0265,
      "theoretical_loss": 3.65880068727604,
      "tokens_seen": 972816384
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004615140911670134,
      "loss": 2.9373,
      "theoretical_loss": 3.658777111585671,
      "tokens_seen": 972881920
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046149754113864193,
      "loss": 3.0931,
      "theoretical_loss": 3.6587535379280123,
      "tokens_seen": 972947456
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046148099111027047,
      "loss": 2.8335,
      "theoretical_loss": 3.6587299663027517,
      "tokens_seen": 973012992
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046146444108189895,
      "loss": 2.9471,
      "theoretical_loss": 3.6587063967095776,
      "tokens_seen": 973078528
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004614478910535275,
      "loss": 2.8795,
      "theoretical_loss": 3.6586828291481766,
      "tokens_seen": 973144064
    },
    {
      "epoch": 0.35,
      "objective/train/avg_token_score": 0.015231223776936531,
      "objective/train/avg_weight": 0.9820674657821655,
      "objective/train/docs_used": 555679,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0831809043884277,
      "objective/train/std_weight": 0.07811808586120605,
      "objective/train/theoretical_loss": 3.6586592636182376,
      "objective/train/tokens_used": 993669600,
      "theoretical_loss": 3.6586592636182376,
      "tokens_seen": 973209600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000461431341025156,
      "loss": 2.7937,
      "theoretical_loss": 3.6586592636182376,
      "tokens_seen": 973209600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046141479099678456,
      "loss": 2.8441,
      "theoretical_loss": 3.658635700119449,
      "tokens_seen": 973275136
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046139824096841305,
      "loss": 2.8308,
      "theoretical_loss": 3.658612138651498,
      "tokens_seen": 973340672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004613816909400416,
      "loss": 2.8144,
      "theoretical_loss": 3.6585885792140744,
      "tokens_seen": 973406208
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004613651409116701,
      "loss": 2.5696,
      "theoretical_loss": 3.658565021806865,
      "tokens_seen": 973471744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046134859088329866,
      "loss": 2.8651,
      "theoretical_loss": 3.658541466429559,
      "tokens_seen": 973537280
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046133204085492714,
      "loss": 3.1259,
      "theoretical_loss": 3.6585179130818446,
      "tokens_seen": 973602816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004613154908265557,
      "loss": 2.9157,
      "theoretical_loss": 3.6584943617634105,
      "tokens_seen": 973668352
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004612989407981842,
      "loss": 2.6402,
      "theoretical_loss": 3.6584708124739453,
      "tokens_seen": 973733888
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046128239076981275,
      "loss": 2.8745,
      "theoretical_loss": 3.658447265213137,
      "tokens_seen": 973799424
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046126584074144124,
      "loss": 3.0328,
      "theoretical_loss": 3.658423719980675,
      "tokens_seen": 973864960
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004612492907130698,
      "loss": 2.9161,
      "theoretical_loss": 3.658400176776248,
      "tokens_seen": 973930496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004612327406846983,
      "loss": 2.4948,
      "theoretical_loss": 3.658376635599545,
      "tokens_seen": 973996032
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046121619065632685,
      "loss": 2.9374,
      "theoretical_loss": 3.658353096450255,
      "tokens_seen": 974061568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046119964062795533,
      "loss": 2.7085,
      "theoretical_loss": 3.6583295593280667,
      "tokens_seen": 974127104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046118309059958387,
      "loss": 2.8877,
      "theoretical_loss": 3.658306024232669,
      "tokens_seen": 974192640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004611665405712124,
      "loss": 2.7161,
      "theoretical_loss": 3.6582824911637513,
      "tokens_seen": 974258176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046114999054284095,
      "loss": 2.6925,
      "theoretical_loss": 3.658258960121003,
      "tokens_seen": 974323712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046113344051446943,
      "loss": 2.9132,
      "theoretical_loss": 3.6582354311041128,
      "tokens_seen": 974389248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046111689048609797,
      "loss": 2.9852,
      "theoretical_loss": 3.658211904112771,
      "tokens_seen": 974454784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004611003404577265,
      "loss": 2.8171,
      "theoretical_loss": 3.658188379146666,
      "tokens_seen": 974520320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046108379042935504,
      "loss": 2.5042,
      "theoretical_loss": 3.658164856205488,
      "tokens_seen": 974585856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004610672404009835,
      "loss": 2.8278,
      "theoretical_loss": 3.658141335288926,
      "tokens_seen": 974651392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046105069037261206,
      "loss": 2.9967,
      "theoretical_loss": 3.6581178163966706,
      "tokens_seen": 974716928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004610341403442406,
      "loss": 2.7183,
      "theoretical_loss": 3.6580942995284103,
      "tokens_seen": 974782464
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046101759031586914,
      "loss": 3.0195,
      "theoretical_loss": 3.6580707846838356,
      "tokens_seen": 974848000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004610010402874976,
      "loss": 2.7389,
      "theoretical_loss": 3.6580472718626353,
      "tokens_seen": 974913536
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046098449025912616,
      "loss": 3.0371,
      "theoretical_loss": 3.6580237610645012,
      "tokens_seen": 974979072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004609679402307547,
      "loss": 2.7964,
      "theoretical_loss": 3.6580002522891215,
      "tokens_seen": 975044608
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046095139020238323,
      "loss": 2.9289,
      "theoretical_loss": 3.657976745536187,
      "tokens_seen": 975110144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004609348401740117,
      "loss": 2.9669,
      "theoretical_loss": 3.657953240805388,
      "tokens_seen": 975175680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046091829014564026,
      "loss": 2.607,
      "theoretical_loss": 3.6579297380964144,
      "tokens_seen": 975241216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004609017401172688,
      "loss": 2.7622,
      "theoretical_loss": 3.657906237408956,
      "tokens_seen": 975306752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046088519008889733,
      "loss": 2.9106,
      "theoretical_loss": 3.6578827387427038,
      "tokens_seen": 975372288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004608686400605258,
      "loss": 2.7365,
      "theoretical_loss": 3.6578592420973477,
      "tokens_seen": 975437824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046085209003215435,
      "loss": 2.9228,
      "theoretical_loss": 3.6578357474725784,
      "tokens_seen": 975503360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004608355400037829,
      "loss": 2.8881,
      "theoretical_loss": 3.6578122548680865,
      "tokens_seen": 975568896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004608189899754114,
      "loss": 3.0305,
      "theoretical_loss": 3.657788764283562,
      "tokens_seen": 975634432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004608024399470399,
      "loss": 2.9507,
      "theoretical_loss": 3.6577652757186963,
      "tokens_seen": 975699968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046078588991866845,
      "loss": 3.0635,
      "theoretical_loss": 3.65774178917318,
      "tokens_seen": 975765504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000460769339890297,
      "loss": 3.1261,
      "theoretical_loss": 3.657718304646703,
      "tokens_seen": 975831040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004607527898619255,
      "loss": 2.861,
      "theoretical_loss": 3.657694822138957,
      "tokens_seen": 975896576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000460736239833554,
      "loss": 2.6589,
      "theoretical_loss": 3.657671341649633,
      "tokens_seen": 975962112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046071968980518254,
      "loss": 2.7454,
      "theoretical_loss": 3.657647863178422,
      "tokens_seen": 976027648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004607031397768111,
      "loss": 2.5276,
      "theoretical_loss": 3.657624386725014,
      "tokens_seen": 976093184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004606865897484396,
      "loss": 2.7743,
      "theoretical_loss": 3.6576009122891016,
      "tokens_seen": 976158720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004606700397200681,
      "loss": 2.5881,
      "theoretical_loss": 3.657577439870375,
      "tokens_seen": 976224256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046065348969169664,
      "loss": 2.6906,
      "theoretical_loss": 3.657553969468526,
      "tokens_seen": 976289792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004606369396633252,
      "loss": 2.8525,
      "theoretical_loss": 3.6575305010832455,
      "tokens_seen": 976355328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004606203896349537,
      "loss": 2.914,
      "theoretical_loss": 3.657507034714225,
      "tokens_seen": 976420864
    },
    {
      "epoch": 0.35,
      "objective/train/avg_token_score": 0.021052435040473938,
      "objective/train/avg_weight": 0.9752611517906189,
      "objective/train/docs_used": 557385,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.703920841217041,
      "objective/train/std_weight": 0.08627495914697647,
      "objective/train/theoretical_loss": 3.6574835703611566,
      "objective/train/tokens_used": 996946400,
      "theoretical_loss": 3.6574835703611566,
      "tokens_seen": 976486400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046060383960658214,
      "loss": 2.7505,
      "theoretical_loss": 3.6574835703611566,
      "tokens_seen": 976486400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004605872895782107,
      "loss": 2.9529,
      "theoretical_loss": 3.657460108023731,
      "tokens_seen": 976551936
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046057073954983916,
      "loss": 2.7526,
      "theoretical_loss": 3.6574366477016396,
      "tokens_seen": 976617472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004605541895214677,
      "loss": 2.8574,
      "theoretical_loss": 3.6574131893945747,
      "tokens_seen": 976683008
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046053763949309624,
      "loss": 2.7735,
      "theoretical_loss": 3.657389733102228,
      "tokens_seen": 976748544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004605210894647248,
      "loss": 2.663,
      "theoretical_loss": 3.6573662788242913,
      "tokens_seen": 976814080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046050453943635326,
      "loss": 2.778,
      "theoretical_loss": 3.6573428265604564,
      "tokens_seen": 976879616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004604879894079818,
      "loss": 3.0234,
      "theoretical_loss": 3.657319376310415,
      "tokens_seen": 976945152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046047143937961034,
      "loss": 3.0168,
      "theoretical_loss": 3.6572959280738595,
      "tokens_seen": 977010688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046045488935123887,
      "loss": 2.7964,
      "theoretical_loss": 3.6572724818504816,
      "tokens_seen": 977076224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046043833932286736,
      "loss": 2.8582,
      "theoretical_loss": 3.657249037639974,
      "tokens_seen": 977141760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004604217892944959,
      "loss": 2.6835,
      "theoretical_loss": 3.6572255954420276,
      "tokens_seen": 977207296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046040523926612443,
      "loss": 2.9834,
      "theoretical_loss": 3.6572021552563365,
      "tokens_seen": 977272832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046038868923775297,
      "loss": 3.1143,
      "theoretical_loss": 3.6571787170825925,
      "tokens_seen": 977338368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046037213920938145,
      "loss": 2.8576,
      "theoretical_loss": 3.657155280920487,
      "tokens_seen": 977403904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046035558918101,
      "loss": 2.9813,
      "theoretical_loss": 3.657131846769713,
      "tokens_seen": 977469440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004603390391526385,
      "loss": 2.6932,
      "theoretical_loss": 3.6571084146299633,
      "tokens_seen": 977534976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046032248912426706,
      "loss": 2.6805,
      "theoretical_loss": 3.657084984500931,
      "tokens_seen": 977600512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046030593909589555,
      "loss": 2.7968,
      "theoretical_loss": 3.6570615563823075,
      "tokens_seen": 977666048
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004602893890675241,
      "loss": 2.8917,
      "theoretical_loss": 3.6570381302737864,
      "tokens_seen": 977731584
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004602728390391526,
      "loss": 2.9975,
      "theoretical_loss": 3.657014706175061,
      "tokens_seen": 977797120
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046025628901078116,
      "loss": 2.7894,
      "theoretical_loss": 3.656991284085823,
      "tokens_seen": 977862656
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046023973898240964,
      "loss": 2.8413,
      "theoretical_loss": 3.6569678640057655,
      "tokens_seen": 977928192
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004602231889540382,
      "loss": 2.7699,
      "theoretical_loss": 3.656944445934583,
      "tokens_seen": 977993728
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004602066389256667,
      "loss": 3.1101,
      "theoretical_loss": 3.6569210298719668,
      "tokens_seen": 978059264
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046019008889729526,
      "loss": 2.9406,
      "theoretical_loss": 3.656897615817611,
      "tokens_seen": 978124800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046017353886892374,
      "loss": 2.8442,
      "theoretical_loss": 3.656874203771208,
      "tokens_seen": 978190336
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004601569888405523,
      "loss": 2.6438,
      "theoretical_loss": 3.6568507937324526,
      "tokens_seen": 978255872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004601404388121808,
      "loss": 2.917,
      "theoretical_loss": 3.6568273857010363,
      "tokens_seen": 978321408
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046012388878380935,
      "loss": 2.3958,
      "theoretical_loss": 3.6568039796766545,
      "tokens_seen": 978386944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046010733875543784,
      "loss": 2.7398,
      "theoretical_loss": 3.656780575658999,
      "tokens_seen": 978452480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004600907887270664,
      "loss": 2.8916,
      "theoretical_loss": 3.6567571736477644,
      "tokens_seen": 978518016
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004600742386986949,
      "loss": 2.976,
      "theoretical_loss": 3.6567337736426433,
      "tokens_seen": 978583552
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046005768867032345,
      "loss": 2.6541,
      "theoretical_loss": 3.656710375643331,
      "tokens_seen": 978649088
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046004113864195193,
      "loss": 2.8149,
      "theoretical_loss": 3.6566869796495194,
      "tokens_seen": 978714624
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00046002458861358047,
      "loss": 2.7651,
      "theoretical_loss": 3.6566635856609038,
      "tokens_seen": 978780160
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000460008038585209,
      "loss": 2.7101,
      "theoretical_loss": 3.6566401936771773,
      "tokens_seen": 978845696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045999148855683754,
      "loss": 2.8889,
      "theoretical_loss": 3.6566168036980344,
      "tokens_seen": 978911232
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004599749385284661,
      "loss": 2.6061,
      "theoretical_loss": 3.6565934157231688,
      "tokens_seen": 978976768
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045995838850009457,
      "loss": 2.8815,
      "theoretical_loss": 3.6565700297522743,
      "tokens_seen": 979042304
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004599418384717231,
      "loss": 2.6509,
      "theoretical_loss": 3.6565466457850455,
      "tokens_seen": 979107840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045992528844335164,
      "loss": 2.8187,
      "theoretical_loss": 3.656523263821177,
      "tokens_seen": 979173376
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004599087384149802,
      "loss": 2.8395,
      "theoretical_loss": 3.656499883860362,
      "tokens_seen": 979238912
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045989218838660866,
      "loss": 2.7495,
      "theoretical_loss": 3.6564765059022957,
      "tokens_seen": 979304448
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004598756383582372,
      "loss": 2.7881,
      "theoretical_loss": 3.6564531299466725,
      "tokens_seen": 979369984
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045985908832986574,
      "loss": 3.1396,
      "theoretical_loss": 3.6564297559931864,
      "tokens_seen": 979435520
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004598425383014943,
      "loss": 2.6971,
      "theoretical_loss": 3.656406384041533,
      "tokens_seen": 979501056
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045982598827312276,
      "loss": 2.6595,
      "theoretical_loss": 3.6563830140914058,
      "tokens_seen": 979566592
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004598094382447513,
      "loss": 2.6316,
      "theoretical_loss": 3.6563596461424996,
      "tokens_seen": 979632128
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045979288821637983,
      "loss": 2.859,
      "theoretical_loss": 3.65633628019451,
      "tokens_seen": 979697664
    },
    {
      "epoch": 0.35,
      "objective/train/avg_token_score": 0.02518784999847412,
      "objective/train/avg_weight": 0.9703923463821411,
      "objective/train/docs_used": 559012,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6982688903808594,
      "objective/train/std_weight": 0.12208882719278336,
      "objective/train/theoretical_loss": 3.6563129162471313,
      "objective/train/tokens_used": 1000223200,
      "theoretical_loss": 3.6563129162471313,
      "tokens_seen": 979763200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045977633818800837,
      "loss": 2.4971,
      "theoretical_loss": 3.6563129162471313,
      "tokens_seen": 979763200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045975978815963685,
      "loss": 2.8122,
      "theoretical_loss": 3.6562895543000584,
      "tokens_seen": 979828736
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004597432381312654,
      "loss": 3.0085,
      "theoretical_loss": 3.656266194352986,
      "tokens_seen": 979894272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045972668810289393,
      "loss": 2.9228,
      "theoretical_loss": 3.65624283640561,
      "tokens_seen": 979959808
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045971013807452247,
      "loss": 2.7258,
      "theoretical_loss": 3.6562194804576245,
      "tokens_seen": 980025344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045969358804615095,
      "loss": 2.9558,
      "theoretical_loss": 3.6561961265087253,
      "tokens_seen": 980090880
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004596770380177795,
      "loss": 2.4982,
      "theoretical_loss": 3.6561727745586072,
      "tokens_seen": 980156416
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000459660487989408,
      "loss": 2.9476,
      "theoretical_loss": 3.6561494246069666,
      "tokens_seen": 980221952
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045964393796103656,
      "loss": 2.7048,
      "theoretical_loss": 3.6561260766534973,
      "tokens_seen": 980287488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045962738793266505,
      "loss": 2.9552,
      "theoretical_loss": 3.6561027306978957,
      "tokens_seen": 980353024
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045961083790429353,
      "loss": 2.5414,
      "theoretical_loss": 3.656079386739857,
      "tokens_seen": 980418560
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000459594287875922,
      "loss": 2.6507,
      "theoretical_loss": 3.656056044779077,
      "tokens_seen": 980484096
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045957773784755055,
      "loss": 2.599,
      "theoretical_loss": 3.656032704815251,
      "tokens_seen": 980549632
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004595611878191791,
      "loss": 2.73,
      "theoretical_loss": 3.656009366848075,
      "tokens_seen": 980615168
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004595446377908076,
      "loss": 2.8439,
      "theoretical_loss": 3.6559860308772447,
      "tokens_seen": 980680704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004595280877624361,
      "loss": 2.8316,
      "theoretical_loss": 3.6559626969024563,
      "tokens_seen": 980746240
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045951153773406465,
      "loss": 2.8367,
      "theoretical_loss": 3.655939364923405,
      "tokens_seen": 980811776
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004594949877056932,
      "loss": 2.5745,
      "theoretical_loss": 3.6559160349397866,
      "tokens_seen": 980877312
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004594784376773217,
      "loss": 2.9827,
      "theoretical_loss": 3.655892706951298,
      "tokens_seen": 980942848
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004594618876489502,
      "loss": 2.6946,
      "theoretical_loss": 3.6558693809576353,
      "tokens_seen": 981008384
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045944533762057874,
      "loss": 2.7279,
      "theoretical_loss": 3.6558460569584943,
      "tokens_seen": 981073920
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004594287875922073,
      "loss": 2.7533,
      "theoretical_loss": 3.6558227349535706,
      "tokens_seen": 981139456
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004594122375638358,
      "loss": 2.8252,
      "theoretical_loss": 3.655799414942561,
      "tokens_seen": 981204992
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004593956875354643,
      "loss": 2.8775,
      "theoretical_loss": 3.655776096925163,
      "tokens_seen": 981270528
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045937913750709284,
      "loss": 2.744,
      "theoretical_loss": 3.6557527809010715,
      "tokens_seen": 981336064
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004593625874787214,
      "loss": 2.8412,
      "theoretical_loss": 3.6557294668699836,
      "tokens_seen": 981401600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004593460374503499,
      "loss": 2.9588,
      "theoretical_loss": 3.655706154831596,
      "tokens_seen": 981467136
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004593294874219784,
      "loss": 2.9567,
      "theoretical_loss": 3.6556828447856047,
      "tokens_seen": 981532672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045931293739360693,
      "loss": 2.855,
      "theoretical_loss": 3.6556595367317075,
      "tokens_seen": 981598208
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045929638736523547,
      "loss": 2.8172,
      "theoretical_loss": 3.6556362306696,
      "tokens_seen": 981663744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000459279837336864,
      "loss": 2.9384,
      "theoretical_loss": 3.6556129265989803,
      "tokens_seen": 981729280
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004592632873084925,
      "loss": 2.5887,
      "theoretical_loss": 3.6555896245195436,
      "tokens_seen": 981794816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045924673728012103,
      "loss": 2.8179,
      "theoretical_loss": 3.655566324430989,
      "tokens_seen": 981860352
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045923018725174957,
      "loss": 2.5546,
      "theoretical_loss": 3.6555430263330115,
      "tokens_seen": 981925888
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004592136372233781,
      "loss": 2.8305,
      "theoretical_loss": 3.6555197302253095,
      "tokens_seen": 981991424
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004591970871950066,
      "loss": 2.824,
      "theoretical_loss": 3.6554964361075797,
      "tokens_seen": 982056960
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004591805371666351,
      "loss": 2.7779,
      "theoretical_loss": 3.6554731439795187,
      "tokens_seen": 982122496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045916398713826366,
      "loss": 2.8114,
      "theoretical_loss": 3.6554498538408255,
      "tokens_seen": 982188032
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004591474371098922,
      "loss": 2.8437,
      "theoretical_loss": 3.655426565691196,
      "tokens_seen": 982253568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004591308870815207,
      "loss": 3.0731,
      "theoretical_loss": 3.6554032795303284,
      "tokens_seen": 982319104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004591143370531492,
      "loss": 2.7952,
      "theoretical_loss": 3.6553799953579196,
      "tokens_seen": 982384640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045909778702477776,
      "loss": 2.7715,
      "theoretical_loss": 3.655356713173667,
      "tokens_seen": 982450176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004590812369964063,
      "loss": 2.8985,
      "theoretical_loss": 3.6553334329772693,
      "tokens_seen": 982515712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004590646869680348,
      "loss": 2.6283,
      "theoretical_loss": 3.6553101547684235,
      "tokens_seen": 982581248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004590481369396633,
      "loss": 2.8098,
      "theoretical_loss": 3.6552868785468275,
      "tokens_seen": 982646784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045903158691129185,
      "loss": 2.8318,
      "theoretical_loss": 3.6552636043121787,
      "tokens_seen": 982712320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004590150368829204,
      "loss": 2.5917,
      "theoretical_loss": 3.655240332064176,
      "tokens_seen": 982777856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004589984868545489,
      "loss": 2.694,
      "theoretical_loss": 3.6552170618025164,
      "tokens_seen": 982843392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004589819368261774,
      "loss": 2.5332,
      "theoretical_loss": 3.655193793526898,
      "tokens_seen": 982908928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045896538679780595,
      "loss": 3.0154,
      "theoretical_loss": 3.6551705272370194,
      "tokens_seen": 982974464
    },
    {
      "debugging/Self-BLEU-5": 0.4286046663919377,
      "debugging/distinct-1-grams": 0.8147567798871364,
      "debugging/distinct-2-grams": 0.9823269374342457,
      "debugging/entropy-1-grams": 6.1671920556004824,
      "debugging/entropy-2-grams": 6.947028138756313,
      "debugging/length": 477.53333333333336,
      "debugging/num_segments": 15,
      "epoch": 0.35,
      "objective/train/avg_token_score": 0.020611366257071495,
      "objective/train/avg_weight": 0.9757350087165833,
      "objective/train/docs_used": 560408,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9668633937835693,
      "objective/train/std_weight": 0.10000527650117874,
      "objective/train/theoretical_loss": 3.6551472629325787,
      "objective/train/tokens_used": 1003500000,
      "theoretical_loss": 3.6551472629325787,
      "tokens_seen": 983040000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004589488367694345,
      "loss": 2.7695,
      "theoretical_loss": 3.6551472629325787,
      "tokens_seen": 983040000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045893228674106297,
      "loss": 2.6503,
      "theoretical_loss": 3.6551240006132737,
      "tokens_seen": 983105536
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004589157367126915,
      "loss": 2.9661,
      "theoretical_loss": 3.655100740278803,
      "tokens_seen": 983171072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045889918668432005,
      "loss": 2.6147,
      "theoretical_loss": 3.655077481928865,
      "tokens_seen": 983236608
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004588826366559486,
      "loss": 2.7843,
      "theoretical_loss": 3.655054225563158,
      "tokens_seen": 983302144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045886608662757707,
      "loss": 2.4415,
      "theoretical_loss": 3.6550309711813807,
      "tokens_seen": 983367680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004588495365992056,
      "loss": 2.6886,
      "theoretical_loss": 3.655007718783231,
      "tokens_seen": 983433216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045883298657083414,
      "loss": 2.938,
      "theoretical_loss": 3.6549844683684083,
      "tokens_seen": 983498752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004588164365424627,
      "loss": 2.6161,
      "theoretical_loss": 3.654961219936611,
      "tokens_seen": 983564288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045879988651409116,
      "loss": 2.9661,
      "theoretical_loss": 3.6549379734875376,
      "tokens_seen": 983629824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004587833364857197,
      "loss": 2.6947,
      "theoretical_loss": 3.654914729020887,
      "tokens_seen": 983695360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045876678645734824,
      "loss": 2.6842,
      "theoretical_loss": 3.654891486536359,
      "tokens_seen": 983760896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004587502364289768,
      "loss": 2.994,
      "theoretical_loss": 3.6548682460336512,
      "tokens_seen": 983826432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045873368640060526,
      "loss": 2.6776,
      "theoretical_loss": 3.6548450075124634,
      "tokens_seen": 983891968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004587171363722338,
      "loss": 2.905,
      "theoretical_loss": 3.6548217709724944,
      "tokens_seen": 983957504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045870058634386233,
      "loss": 2.7552,
      "theoretical_loss": 3.6547985364134434,
      "tokens_seen": 984023040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045868403631549087,
      "loss": 2.5751,
      "theoretical_loss": 3.6547753038350095,
      "tokens_seen": 984088576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045866748628711936,
      "loss": 2.9077,
      "theoretical_loss": 3.6547520732368923,
      "tokens_seen": 984154112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004586509362587479,
      "loss": 2.7914,
      "theoretical_loss": 3.6547288446187913,
      "tokens_seen": 984219648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004586343862303764,
      "loss": 2.8297,
      "theoretical_loss": 3.6547056179804054,
      "tokens_seen": 984285184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045861783620200486,
      "loss": 2.6504,
      "theoretical_loss": 3.654682393321434,
      "tokens_seen": 984350720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004586012861736334,
      "loss": 2.9195,
      "theoretical_loss": 3.654659170641577,
      "tokens_seen": 984416256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045858473614526193,
      "loss": 2.8606,
      "theoretical_loss": 3.654635949940534,
      "tokens_seen": 984481792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045856818611689047,
      "loss": 2.7635,
      "theoretical_loss": 3.6546127312180046,
      "tokens_seen": 984547328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045855163608851896,
      "loss": 2.9582,
      "theoretical_loss": 3.654589514473688,
      "tokens_seen": 984612864
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004585350860601475,
      "loss": 2.9173,
      "theoretical_loss": 3.654566299707285,
      "tokens_seen": 984678400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045851853603177603,
      "loss": 2.5534,
      "theoretical_loss": 3.6545430869184954,
      "tokens_seen": 984743936
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045850198600340457,
      "loss": 2.6518,
      "theoretical_loss": 3.6545198761070186,
      "tokens_seen": 984809472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045848543597503305,
      "loss": 2.9262,
      "theoretical_loss": 3.6544966672725545,
      "tokens_seen": 984875008
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004584688859466616,
      "loss": 2.8676,
      "theoretical_loss": 3.654473460414803,
      "tokens_seen": 984940544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004584523359182901,
      "loss": 2.9096,
      "theoretical_loss": 3.6544502555334653,
      "tokens_seen": 985006080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045843578588991866,
      "loss": 2.8275,
      "theoretical_loss": 3.6544270526282405,
      "tokens_seen": 985071616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045841923586154715,
      "loss": 2.8443,
      "theoretical_loss": 3.6544038516988295,
      "tokens_seen": 985137152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004584026858331757,
      "loss": 2.9302,
      "theoretical_loss": 3.6543806527449325,
      "tokens_seen": 985202688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004583861358048042,
      "loss": 2.9898,
      "theoretical_loss": 3.6543574557662497,
      "tokens_seen": 985268224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045836958577643276,
      "loss": 2.9733,
      "theoretical_loss": 3.6543342607624814,
      "tokens_seen": 985333760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045835303574806124,
      "loss": 2.7816,
      "theoretical_loss": 3.654311067733329,
      "tokens_seen": 985399296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004583364857196898,
      "loss": 2.712,
      "theoretical_loss": 3.654287876678492,
      "tokens_seen": 985464832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004583199356913183,
      "loss": 2.8405,
      "theoretical_loss": 3.6542646875976716,
      "tokens_seen": 985530368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045830338566294686,
      "loss": 2.8188,
      "theoretical_loss": 3.654241500490569,
      "tokens_seen": 985595904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045828683563457534,
      "loss": 2.6734,
      "theoretical_loss": 3.654218315356884,
      "tokens_seen": 985661440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004582702856062039,
      "loss": 2.7352,
      "theoretical_loss": 3.6541951321963175,
      "tokens_seen": 985726976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004582537355778324,
      "loss": 2.761,
      "theoretical_loss": 3.6541719510085713,
      "tokens_seen": 985792512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045823718554946095,
      "loss": 2.8631,
      "theoretical_loss": 3.6541487717933463,
      "tokens_seen": 985858048
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045822063552108944,
      "loss": 2.922,
      "theoretical_loss": 3.654125594550343,
      "tokens_seen": 985923584
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045820408549271797,
      "loss": 2.6673,
      "theoretical_loss": 3.654102419279262,
      "tokens_seen": 985989120
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004581875354643465,
      "loss": 2.8283,
      "theoretical_loss": 3.6540792459798057,
      "tokens_seen": 986054656
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045817098543597505,
      "loss": 2.8253,
      "theoretical_loss": 3.6540560746516744,
      "tokens_seen": 986120192
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045815443540760353,
      "loss": 2.9992,
      "theoretical_loss": 3.65403290529457,
      "tokens_seen": 986185728
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045813788537923207,
      "loss": 2.6251,
      "theoretical_loss": 3.6540097379081935,
      "tokens_seen": 986251264
    },
    {
      "epoch": 0.35,
      "objective/train/avg_token_score": 0.024690205231308937,
      "objective/train/avg_weight": 0.9709514379501343,
      "objective/train/docs_used": 562309,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1787784099578857,
      "objective/train/std_weight": 0.10288284718990326,
      "objective/train/theoretical_loss": 3.653986572492247,
      "objective/train/tokens_used": 1006776800,
      "theoretical_loss": 3.653986572492247,
      "tokens_seen": 986316800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004581213353508606,
      "loss": 3.021,
      "theoretical_loss": 3.653986572492247,
      "tokens_seen": 986316800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045810478532248914,
      "loss": 2.7089,
      "theoretical_loss": 3.653963409046431,
      "tokens_seen": 986382336
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045808823529411763,
      "loss": 2.6736,
      "theoretical_loss": 3.653940247570448,
      "tokens_seen": 986447872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045807168526574616,
      "loss": 2.6973,
      "theoretical_loss": 3.6539170880639986,
      "tokens_seen": 986513408
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004580551352373747,
      "loss": 2.7646,
      "theoretical_loss": 3.6538939305267855,
      "tokens_seen": 986578944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045803858520900324,
      "loss": 2.8861,
      "theoretical_loss": 3.65387077495851,
      "tokens_seen": 986644480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004580220351806317,
      "loss": 2.892,
      "theoretical_loss": 3.653847621358874,
      "tokens_seen": 986710016
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045800548515226026,
      "loss": 2.6027,
      "theoretical_loss": 3.6538244697275792,
      "tokens_seen": 986775552
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004579889351238888,
      "loss": 2.8248,
      "theoretical_loss": 3.653801320064328,
      "tokens_seen": 986841088
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045797238509551734,
      "loss": 2.4272,
      "theoretical_loss": 3.6537781723688223,
      "tokens_seen": 986906624
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004579558350671458,
      "loss": 2.5805,
      "theoretical_loss": 3.6537550266407637,
      "tokens_seen": 986972160
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045793928503877436,
      "loss": 2.891,
      "theoretical_loss": 3.6537318828798546,
      "tokens_seen": 987037696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004579227350104029,
      "loss": 2.621,
      "theoretical_loss": 3.6537087410857976,
      "tokens_seen": 987103232
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045790618498203143,
      "loss": 2.6859,
      "theoretical_loss": 3.6536856012582946,
      "tokens_seen": 987168768
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004578896349536599,
      "loss": 2.8194,
      "theoretical_loss": 3.653662463397048,
      "tokens_seen": 987234304
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045787308492528845,
      "loss": 2.8743,
      "theoretical_loss": 3.65363932750176,
      "tokens_seen": 987299840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000457856534896917,
      "loss": 2.8735,
      "theoretical_loss": 3.653616193572134,
      "tokens_seen": 987365376
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045783998486854553,
      "loss": 2.8738,
      "theoretical_loss": 3.6535930616078716,
      "tokens_seen": 987430912
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000457823434840174,
      "loss": 2.8412,
      "theoretical_loss": 3.6535699316086756,
      "tokens_seen": 987496448
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045780688481180255,
      "loss": 2.8132,
      "theoretical_loss": 3.6535468035742484,
      "tokens_seen": 987561984
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004577903347834311,
      "loss": 2.8249,
      "theoretical_loss": 3.6535236775042934,
      "tokens_seen": 987627520
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004577737847550596,
      "loss": 2.7464,
      "theoretical_loss": 3.6535005533985134,
      "tokens_seen": 987693056
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004577572347266881,
      "loss": 2.6738,
      "theoretical_loss": 3.6534774312566105,
      "tokens_seen": 987758592
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045774068469831664,
      "loss": 2.84,
      "theoretical_loss": 3.6534543110782884,
      "tokens_seen": 987824128
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004577241346699452,
      "loss": 2.8542,
      "theoretical_loss": 3.6534311928632497,
      "tokens_seen": 987889664
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004577075846415737,
      "loss": 2.6815,
      "theoretical_loss": 3.653408076611197,
      "tokens_seen": 987955200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004576910346132022,
      "loss": 2.839,
      "theoretical_loss": 3.6533849623218346,
      "tokens_seen": 988020736
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045767448458483074,
      "loss": 2.573,
      "theoretical_loss": 3.653361849994865,
      "tokens_seen": 988086272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004576579345564593,
      "loss": 2.6645,
      "theoretical_loss": 3.6533387396299912,
      "tokens_seen": 988151808
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004576413845280877,
      "loss": 2.9497,
      "theoretical_loss": 3.653315631226917,
      "tokens_seen": 988217344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045762483449971624,
      "loss": 2.8662,
      "theoretical_loss": 3.6532925247853454,
      "tokens_seen": 988282880
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004576082844713448,
      "loss": 3.0165,
      "theoretical_loss": 3.65326942030498,
      "tokens_seen": 988348416
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045759173444297327,
      "loss": 2.7773,
      "theoretical_loss": 3.6532463177855248,
      "tokens_seen": 988413952
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004575751844146018,
      "loss": 2.5739,
      "theoretical_loss": 3.6532232172266825,
      "tokens_seen": 988479488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045755863438623034,
      "loss": 2.8066,
      "theoretical_loss": 3.653200118628157,
      "tokens_seen": 988545024
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004575420843578589,
      "loss": 2.8299,
      "theoretical_loss": 3.653177021989653,
      "tokens_seen": 988610560
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045752553432948736,
      "loss": 2.76,
      "theoretical_loss": 3.653153927310873,
      "tokens_seen": 988676096
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004575089843011159,
      "loss": 2.483,
      "theoretical_loss": 3.6531308345915208,
      "tokens_seen": 988741632
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045749243427274444,
      "loss": 2.4271,
      "theoretical_loss": 3.653107743831301,
      "tokens_seen": 988807168
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000457475884244373,
      "loss": 2.9494,
      "theoretical_loss": 3.6530846550299176,
      "tokens_seen": 988872704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004574593342160015,
      "loss": 2.5996,
      "theoretical_loss": 3.6530615681870744,
      "tokens_seen": 988938240
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045744278418763,
      "loss": 2.574,
      "theoretical_loss": 3.653038483302475,
      "tokens_seen": 989003776
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045742623415925853,
      "loss": 2.7093,
      "theoretical_loss": 3.653015400375824,
      "tokens_seen": 989069312
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045740968413088707,
      "loss": 2.8399,
      "theoretical_loss": 3.652992319406826,
      "tokens_seen": 989134848
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004573931341025156,
      "loss": 2.4884,
      "theoretical_loss": 3.652969240395185,
      "tokens_seen": 989200384
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004573765840741441,
      "loss": 2.8094,
      "theoretical_loss": 3.652946163340605,
      "tokens_seen": 989265920
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045736003404577263,
      "loss": 2.6757,
      "theoretical_loss": 3.6529230882427903,
      "tokens_seen": 989331456
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045734348401740117,
      "loss": 2.7379,
      "theoretical_loss": 3.652900015101446,
      "tokens_seen": 989396992
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004573269339890297,
      "loss": 2.811,
      "theoretical_loss": 3.6528769439162767,
      "tokens_seen": 989462528
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004573103839606582,
      "loss": 2.7528,
      "theoretical_loss": 3.6528538746869863,
      "tokens_seen": 989528064
    },
    {
      "epoch": 0.35,
      "objective/train/avg_token_score": 0.026258952915668488,
      "objective/train/avg_weight": 0.9691047668457031,
      "objective/train/docs_used": 564240,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1253914833068848,
      "objective/train/std_weight": 0.1120094284415245,
      "objective/train/theoretical_loss": 3.65283080741328,
      "objective/train/tokens_used": 1010053600,
      "theoretical_loss": 3.65283080741328,
      "tokens_seen": 989593600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004572938339322867,
      "loss": 3.003,
      "theoretical_loss": 3.65283080741328,
      "tokens_seen": 989593600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045727728390391526,
      "loss": 2.5398,
      "theoretical_loss": 3.652807742094862,
      "tokens_seen": 989659136
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004572607338755438,
      "loss": 2.9075,
      "theoretical_loss": 3.652784678731438,
      "tokens_seen": 989724672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004572441838471723,
      "loss": 2.731,
      "theoretical_loss": 3.652761617322712,
      "tokens_seen": 989790208
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004572276338188008,
      "loss": 2.8894,
      "theoretical_loss": 3.65273855786839,
      "tokens_seen": 989855744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045721108379042936,
      "loss": 2.91,
      "theoretical_loss": 3.652715500368175,
      "tokens_seen": 989921280
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004571945337620579,
      "loss": 2.9244,
      "theoretical_loss": 3.652692444821774,
      "tokens_seen": 989986816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004571779837336864,
      "loss": 2.8489,
      "theoretical_loss": 3.6526693912288914,
      "tokens_seen": 990052352
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004571614337053149,
      "loss": 2.7003,
      "theoretical_loss": 3.652646339589232,
      "tokens_seen": 990117888
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045714488367694345,
      "loss": 2.9208,
      "theoretical_loss": 3.652623289902502,
      "tokens_seen": 990183424
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000457128333648572,
      "loss": 2.6893,
      "theoretical_loss": 3.6526002421684063,
      "tokens_seen": 990248960
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004571117836202005,
      "loss": 2.855,
      "theoretical_loss": 3.6525771963866496,
      "tokens_seen": 990314496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000457095233591829,
      "loss": 2.5102,
      "theoretical_loss": 3.652554152556938,
      "tokens_seen": 990380032
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045707868356345755,
      "loss": 2.8845,
      "theoretical_loss": 3.652531110678977,
      "tokens_seen": 990445568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004570621335350861,
      "loss": 2.8355,
      "theoretical_loss": 3.6525080707524715,
      "tokens_seen": 990511104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045704558350671457,
      "loss": 2.6626,
      "theoretical_loss": 3.652485032777128,
      "tokens_seen": 990576640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004570290334783431,
      "loss": 2.6861,
      "theoretical_loss": 3.652461996752652,
      "tokens_seen": 990642176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045701248344997165,
      "loss": 2.6115,
      "theoretical_loss": 3.6524389626787492,
      "tokens_seen": 990707712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004569959334216002,
      "loss": 2.6204,
      "theoretical_loss": 3.652415930555125,
      "tokens_seen": 990773248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045697938339322867,
      "loss": 2.7898,
      "theoretical_loss": 3.6523929003814857,
      "tokens_seen": 990838784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004569628333648572,
      "loss": 2.6739,
      "theoretical_loss": 3.652369872157537,
      "tokens_seen": 990904320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045694628333648574,
      "loss": 2.749,
      "theoretical_loss": 3.6523468458829855,
      "tokens_seen": 990969856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004569297333081143,
      "loss": 2.7838,
      "theoretical_loss": 3.6523238215575367,
      "tokens_seen": 991035392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045691318327974276,
      "loss": 2.6668,
      "theoretical_loss": 3.6523007991808965,
      "tokens_seen": 991100928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004568966332513713,
      "loss": 2.7259,
      "theoretical_loss": 3.6522777787527714,
      "tokens_seen": 991166464
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045688008322299984,
      "loss": 2.7099,
      "theoretical_loss": 3.6522547602728683,
      "tokens_seen": 991232000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004568635331946284,
      "loss": 3.0062,
      "theoretical_loss": 3.652231743740893,
      "tokens_seen": 991297536
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045684698316625686,
      "loss": 2.8518,
      "theoretical_loss": 3.652208729156551,
      "tokens_seen": 991363072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004568304331378854,
      "loss": 2.8876,
      "theoretical_loss": 3.6521857165195506,
      "tokens_seen": 991428608
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045681388310951393,
      "loss": 2.6514,
      "theoretical_loss": 3.652162705829597,
      "tokens_seen": 991494144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045679733308114247,
      "loss": 2.7205,
      "theoretical_loss": 3.652139697086397,
      "tokens_seen": 991559680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045678078305277095,
      "loss": 2.8382,
      "theoretical_loss": 3.652116690289658,
      "tokens_seen": 991625216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004567642330243995,
      "loss": 2.8741,
      "theoretical_loss": 3.652093685439085,
      "tokens_seen": 991690752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045674768299602803,
      "loss": 2.8347,
      "theoretical_loss": 3.6520706825343865,
      "tokens_seen": 991756288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045673113296765657,
      "loss": 2.8559,
      "theoretical_loss": 3.6520476815752687,
      "tokens_seen": 991821824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045671458293928505,
      "loss": 2.8718,
      "theoretical_loss": 3.6520246825614384,
      "tokens_seen": 991887360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004566980329109136,
      "loss": 2.8495,
      "theoretical_loss": 3.652001685492603,
      "tokens_seen": 991952896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004566814828825421,
      "loss": 2.8404,
      "theoretical_loss": 3.651978690368469,
      "tokens_seen": 992018432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045666493285417056,
      "loss": 2.7506,
      "theoretical_loss": 3.6519556971887437,
      "tokens_seen": 992083968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004566483828257991,
      "loss": 2.3817,
      "theoretical_loss": 3.651932705953134,
      "tokens_seen": 992149504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045663183279742763,
      "loss": 2.7059,
      "theoretical_loss": 3.651909716661348,
      "tokens_seen": 992215040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004566152827690561,
      "loss": 2.7566,
      "theoretical_loss": 3.651886729313092,
      "tokens_seen": 992280576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045659873274068465,
      "loss": 2.9121,
      "theoretical_loss": 3.6518637439080734,
      "tokens_seen": 992346112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004565821827123132,
      "loss": 2.869,
      "theoretical_loss": 3.651840760446,
      "tokens_seen": 992411648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004565656326839417,
      "loss": 2.8414,
      "theoretical_loss": 3.6518177789265795,
      "tokens_seen": 992477184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004565490826555702,
      "loss": 2.6032,
      "theoretical_loss": 3.651794799349519,
      "tokens_seen": 992542720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045653253262719875,
      "loss": 2.8008,
      "theoretical_loss": 3.651771821714526,
      "tokens_seen": 992608256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004565159825988273,
      "loss": 2.7935,
      "theoretical_loss": 3.6517488460213086,
      "tokens_seen": 992673792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004564994325704558,
      "loss": 2.6136,
      "theoretical_loss": 3.6517258722695747,
      "tokens_seen": 992739328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004564828825420843,
      "loss": 2.7494,
      "theoretical_loss": 3.651702900459031,
      "tokens_seen": 992804864
    },
    {
      "epoch": 0.35,
      "objective/train/avg_token_score": 0.023495366796851158,
      "objective/train/avg_weight": 0.9723411202430725,
      "objective/train/docs_used": 566160,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.49644136428833,
      "objective/train/std_weight": 0.11598529666662216,
      "objective/train/theoretical_loss": 3.6516799305893866,
      "objective/train/tokens_used": 1013330400,
      "theoretical_loss": 3.6516799305893866,
      "tokens_seen": 992870400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045646633251371284,
      "loss": 3.0054,
      "theoretical_loss": 3.6516799305893866,
      "tokens_seen": 992870400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004564497824853414,
      "loss": 2.7157,
      "theoretical_loss": 3.651656962660349,
      "tokens_seen": 992935936
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004564332324569699,
      "loss": 2.8073,
      "theoretical_loss": 3.6516339966716256,
      "tokens_seen": 993001472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004564166824285984,
      "loss": 2.6466,
      "theoretical_loss": 3.6516110326229247,
      "tokens_seen": 993067008
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045640013240022694,
      "loss": 2.993,
      "theoretical_loss": 3.651588070513955,
      "tokens_seen": 993132544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004563835823718555,
      "loss": 2.9212,
      "theoretical_loss": 3.651565110344424,
      "tokens_seen": 993198080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000456367032343484,
      "loss": 2.6561,
      "theoretical_loss": 3.6515421521140405,
      "tokens_seen": 993263616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004563504823151125,
      "loss": 2.5464,
      "theoretical_loss": 3.6515191958225124,
      "tokens_seen": 993329152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045633393228674103,
      "loss": 2.6169,
      "theoretical_loss": 3.651496241469548,
      "tokens_seen": 993394688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045631738225836957,
      "loss": 2.7637,
      "theoretical_loss": 3.6514732890548567,
      "tokens_seen": 993460224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004563008322299981,
      "loss": 2.943,
      "theoretical_loss": 3.6514503385781456,
      "tokens_seen": 993525760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004562842822016266,
      "loss": 2.8748,
      "theoretical_loss": 3.6514273900391236,
      "tokens_seen": 993591296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045626773217325513,
      "loss": 2.7839,
      "theoretical_loss": 3.6514044434374995,
      "tokens_seen": 993656832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045625118214488367,
      "loss": 2.9432,
      "theoretical_loss": 3.6513814987729822,
      "tokens_seen": 993722368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004562346321165122,
      "loss": 2.8289,
      "theoretical_loss": 3.65135855604528,
      "tokens_seen": 993787904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004562180820881407,
      "loss": 2.8963,
      "theoretical_loss": 3.6513356152541023,
      "tokens_seen": 993853440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0004562015320597692,
      "loss": 2.8548,
      "theoretical_loss": 3.651312676399158,
      "tokens_seen": 993918976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00045618498203139776,
      "loss": 2.6991,
      "theoretical_loss": 3.6512897394801547,
      "tokens_seen": 993984512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004561684320030263,
      "loss": 2.7909,
      "theoretical_loss": 3.651266804496803,
      "tokens_seen": 994050048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004561518819746548,
      "loss": 2.8669,
      "theoretical_loss": 3.651243871448811,
      "tokens_seen": 994115584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004561353319462833,
      "loss": 2.7768,
      "theoretical_loss": 3.651220940335888,
      "tokens_seen": 994181120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045611878191791186,
      "loss": 2.7116,
      "theoretical_loss": 3.651198011157743,
      "tokens_seen": 994246656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004561022318895404,
      "loss": 2.7151,
      "theoretical_loss": 3.6511750839140866,
      "tokens_seen": 994312192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004560856818611689,
      "loss": 2.8414,
      "theoretical_loss": 3.651152158604626,
      "tokens_seen": 994377728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004560691318327974,
      "loss": 2.8316,
      "theoretical_loss": 3.651129235229072,
      "tokens_seen": 994443264
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045605258180442596,
      "loss": 2.6767,
      "theoretical_loss": 3.6511063137871336,
      "tokens_seen": 994508800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004560360317760545,
      "loss": 3.0205,
      "theoretical_loss": 3.65108339427852,
      "tokens_seen": 994574336
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000456019481747683,
      "loss": 2.781,
      "theoretical_loss": 3.6510604767029413,
      "tokens_seen": 994639872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004560029317193115,
      "loss": 2.6047,
      "theoretical_loss": 3.6510375610601065,
      "tokens_seen": 994705408
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045598638169094005,
      "loss": 2.7372,
      "theoretical_loss": 3.6510146473497262,
      "tokens_seen": 994770944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004559698316625686,
      "loss": 2.7438,
      "theoretical_loss": 3.6509917355715094,
      "tokens_seen": 994836480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045595328163419713,
      "loss": 2.6612,
      "theoretical_loss": 3.6509688257251662,
      "tokens_seen": 994902016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004559367316058256,
      "loss": 2.8362,
      "theoretical_loss": 3.650945917810406,
      "tokens_seen": 994967552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045592018157745415,
      "loss": 2.8513,
      "theoretical_loss": 3.650923011826939,
      "tokens_seen": 995033088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004559036315490827,
      "loss": 2.7332,
      "theoretical_loss": 3.6509001077744756,
      "tokens_seen": 995098624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004558870815207112,
      "loss": 2.6524,
      "theoretical_loss": 3.6508772056527263,
      "tokens_seen": 995164160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004558705314923397,
      "loss": 2.6519,
      "theoretical_loss": 3.6508543054613996,
      "tokens_seen": 995229696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045585398146396824,
      "loss": 2.9578,
      "theoretical_loss": 3.650831407200206,
      "tokens_seen": 995295232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004558374314355968,
      "loss": 2.7491,
      "theoretical_loss": 3.650808510868858,
      "tokens_seen": 995360768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004558208814072253,
      "loss": 2.6637,
      "theoretical_loss": 3.650785616467063,
      "tokens_seen": 995426304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004558043313788538,
      "loss": 2.765,
      "theoretical_loss": 3.6507627239945326,
      "tokens_seen": 995491840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045578778135048234,
      "loss": 2.8082,
      "theoretical_loss": 3.6507398334509773,
      "tokens_seen": 995557376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004557712313221109,
      "loss": 2.8339,
      "theoretical_loss": 3.650716944836107,
      "tokens_seen": 995622912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004557546812937394,
      "loss": 2.6611,
      "theoretical_loss": 3.650694058149634,
      "tokens_seen": 995688448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004557381312653679,
      "loss": 2.7178,
      "theoretical_loss": 3.6506711733912667,
      "tokens_seen": 995753984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045572158123699644,
      "loss": 2.6568,
      "theoretical_loss": 3.650648290560717,
      "tokens_seen": 995819520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000455705031208625,
      "loss": 2.763,
      "theoretical_loss": 3.6506254096576956,
      "tokens_seen": 995885056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004556884811802535,
      "loss": 2.7019,
      "theoretical_loss": 3.6506025306819128,
      "tokens_seen": 995950592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045567193115188194,
      "loss": 2.6868,
      "theoretical_loss": 3.65057965363308,
      "tokens_seen": 996016128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004556553811235105,
      "loss": 2.8268,
      "theoretical_loss": 3.650556778510908,
      "tokens_seen": 996081664
    },
    {
      "epoch": 0.36,
      "objective/train/avg_token_score": 0.03795424848794937,
      "objective/train/avg_weight": 0.9553444981575012,
      "objective/train/docs_used": 567938,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9436733722686768,
      "objective/train/std_weight": 0.17523138225078583,
      "objective/train/theoretical_loss": 3.6505339053151076,
      "objective/train/tokens_used": 1016607200,
      "theoretical_loss": 3.6505339053151076,
      "tokens_seen": 996147200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045563883109513896,
      "loss": 2.7053,
      "theoretical_loss": 3.6505339053151076,
      "tokens_seen": 996147200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004556222810667675,
      "loss": 2.7808,
      "theoretical_loss": 3.65051103404539,
      "tokens_seen": 996212736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045560573103839604,
      "loss": 2.9489,
      "theoretical_loss": 3.650488164701466,
      "tokens_seen": 996278272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004555891810100246,
      "loss": 2.7671,
      "theoretical_loss": 3.6504652972830476,
      "tokens_seen": 996343808
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045557263098165306,
      "loss": 2.9903,
      "theoretical_loss": 3.6504424317898447,
      "tokens_seen": 996409344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004555560809532816,
      "loss": 2.7847,
      "theoretical_loss": 3.6504195682215697,
      "tokens_seen": 996474880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045553953092491013,
      "loss": 2.9617,
      "theoretical_loss": 3.650396706577934,
      "tokens_seen": 996540416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045552298089653867,
      "loss": 2.7896,
      "theoretical_loss": 3.6503738468586486,
      "tokens_seen": 996605952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045550643086816715,
      "loss": 2.7725,
      "theoretical_loss": 3.6503509890634254,
      "tokens_seen": 996671488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004554898808397957,
      "loss": 2.9218,
      "theoretical_loss": 3.650328133191975,
      "tokens_seen": 996737024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045547333081142423,
      "loss": 2.9881,
      "theoretical_loss": 3.65030527924401,
      "tokens_seen": 996802560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045545678078305277,
      "loss": 2.8438,
      "theoretical_loss": 3.650282427219242,
      "tokens_seen": 996868096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045544023075468125,
      "loss": 2.6781,
      "theoretical_loss": 3.650259577117382,
      "tokens_seen": 996933632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004554236807263098,
      "loss": 2.7044,
      "theoretical_loss": 3.6502367289381423,
      "tokens_seen": 996999168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004554071306979383,
      "loss": 3.0553,
      "theoretical_loss": 3.650213882681235,
      "tokens_seen": 997064704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045539058066956686,
      "loss": 2.7436,
      "theoretical_loss": 3.650191038346372,
      "tokens_seen": 997130240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045537403064119535,
      "loss": 2.8377,
      "theoretical_loss": 3.6501681959332646,
      "tokens_seen": 997195776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004553574806128239,
      "loss": 2.869,
      "theoretical_loss": 3.6501453554416257,
      "tokens_seen": 997261312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004553409305844524,
      "loss": 2.6335,
      "theoretical_loss": 3.650122516871167,
      "tokens_seen": 997326848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045532438055608096,
      "loss": 2.9774,
      "theoretical_loss": 3.650099680221601,
      "tokens_seen": 997392384
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045530783052770944,
      "loss": 2.7596,
      "theoretical_loss": 3.650076845492639,
      "tokens_seen": 997457920
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000455291280499338,
      "loss": 2.7594,
      "theoretical_loss": 3.650054012683994,
      "tokens_seen": 997523456
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004552747304709665,
      "loss": 2.7459,
      "theoretical_loss": 3.6500311817953786,
      "tokens_seen": 997588992
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045525818044259505,
      "loss": 2.5361,
      "theoretical_loss": 3.6500083528265046,
      "tokens_seen": 997654528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045524163041422354,
      "loss": 2.789,
      "theoretical_loss": 3.6499855257770855,
      "tokens_seen": 997720064
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004552250803858521,
      "loss": 2.9548,
      "theoretical_loss": 3.6499627006468325,
      "tokens_seen": 997785600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004552085303574806,
      "loss": 2.5213,
      "theoretical_loss": 3.6499398774354592,
      "tokens_seen": 997851136
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045519198032910915,
      "loss": 2.801,
      "theoretical_loss": 3.6499170561426784,
      "tokens_seen": 997916672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045517543030073763,
      "loss": 2.7199,
      "theoretical_loss": 3.6498942367682017,
      "tokens_seen": 997982208
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045515888027236617,
      "loss": 2.7208,
      "theoretical_loss": 3.6498714193117427,
      "tokens_seen": 998047744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004551423302439947,
      "loss": 2.9585,
      "theoretical_loss": 3.6498486037730142,
      "tokens_seen": 998113280
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045512578021562325,
      "loss": 2.7618,
      "theoretical_loss": 3.6498257901517293,
      "tokens_seen": 998178816
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045510923018725173,
      "loss": 2.8677,
      "theoretical_loss": 3.6498029784476005,
      "tokens_seen": 998244352
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045509268015888027,
      "loss": 2.7368,
      "theoretical_loss": 3.6497801686603415,
      "tokens_seen": 998309888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004550761301305088,
      "loss": 2.6585,
      "theoretical_loss": 3.649757360789664,
      "tokens_seen": 998375424
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045505958010213734,
      "loss": 2.4588,
      "theoretical_loss": 3.649734554835283,
      "tokens_seen": 998440960
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004550430300737658,
      "loss": 2.5921,
      "theoretical_loss": 3.649711750796911,
      "tokens_seen": 998506496
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045502648004539436,
      "loss": 2.9819,
      "theoretical_loss": 3.6496889486742607,
      "tokens_seen": 998572032
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004550099300170229,
      "loss": 2.7594,
      "theoretical_loss": 3.6496661484670465,
      "tokens_seen": 998637568
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045499337998865144,
      "loss": 2.8404,
      "theoretical_loss": 3.6496433501749808,
      "tokens_seen": 998703104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004549768299602799,
      "loss": 2.7701,
      "theoretical_loss": 3.649620553797777,
      "tokens_seen": 998768640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045496027993190846,
      "loss": 2.6938,
      "theoretical_loss": 3.6495977593351503,
      "tokens_seen": 998834176
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000454943729903537,
      "loss": 2.9691,
      "theoretical_loss": 3.6495749667868123,
      "tokens_seen": 998899712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045492717987516553,
      "loss": 2.7822,
      "theoretical_loss": 3.649552176152478,
      "tokens_seen": 998965248
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000454910629846794,
      "loss": 2.5861,
      "theoretical_loss": 3.64952938743186,
      "tokens_seen": 999030784
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045489407981842255,
      "loss": 2.9216,
      "theoretical_loss": 3.6495066006246732,
      "tokens_seen": 999096320
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004548775297900511,
      "loss": 2.53,
      "theoretical_loss": 3.6494838157306306,
      "tokens_seen": 999161856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045486097976167963,
      "loss": 2.9426,
      "theoretical_loss": 3.649461032749447,
      "tokens_seen": 999227392
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004548444297333081,
      "loss": 2.7285,
      "theoretical_loss": 3.6494382516808357,
      "tokens_seen": 999292928
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045482787970493665,
      "loss": 2.9604,
      "theoretical_loss": 3.6494154725245105,
      "tokens_seen": 999358464
    },
    {
      "epoch": 0.36,
      "objective/train/avg_token_score": 0.009048331528902054,
      "objective/train/avg_weight": 0.989374577999115,
      "objective/train/docs_used": 570122,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6023621559143066,
      "objective/train/std_weight": 0.0487808994948864,
      "objective/train/theoretical_loss": 3.649392695280186,
      "objective/train/tokens_used": 1019884000,
      "theoretical_loss": 3.649392695280186,
      "tokens_seen": 999424000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004548113296765652,
      "loss": 2.6358,
      "theoretical_loss": 3.649392695280186,
      "tokens_seen": 999424000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004547947796481937,
      "loss": 2.6239,
      "theoretical_loss": 3.649369919947577,
      "tokens_seen": 999489536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004547782296198222,
      "loss": 2.7333,
      "theoretical_loss": 3.649347146526396,
      "tokens_seen": 999555072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045476167959145075,
      "loss": 2.8762,
      "theoretical_loss": 3.6493243750163584,
      "tokens_seen": 999620608
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004547451295630793,
      "loss": 2.7055,
      "theoretical_loss": 3.649301605417178,
      "tokens_seen": 999686144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004547285795347078,
      "loss": 2.4743,
      "theoretical_loss": 3.64927883772857,
      "tokens_seen": 999751680
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004547120295063363,
      "loss": 2.5068,
      "theoretical_loss": 3.6492560719502483,
      "tokens_seen": 999817216
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004546954794779648,
      "loss": 2.7197,
      "theoretical_loss": 3.6492333080819277,
      "tokens_seen": 999882752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004546789294495933,
      "loss": 2.7414,
      "theoretical_loss": 3.649210546123322,
      "tokens_seen": 999948288
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004546623794212218,
      "loss": 2.6851,
      "theoretical_loss": 3.649187786074147,
      "tokens_seen": 1000013824
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045464582939285035,
      "loss": 2.9412,
      "theoretical_loss": 3.6491650279341163,
      "tokens_seen": 1000079360
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004546292793644789,
      "loss": 2.6363,
      "theoretical_loss": 3.649142271702946,
      "tokens_seen": 1000144896
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004546127293361074,
      "loss": 2.6765,
      "theoretical_loss": 3.6491195173803495,
      "tokens_seen": 1000210432
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004545961793077359,
      "loss": 2.6687,
      "theoretical_loss": 3.6490967649660426,
      "tokens_seen": 1000275968
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045457962927936444,
      "loss": 2.8696,
      "theoretical_loss": 3.64907401445974,
      "tokens_seen": 1000341504
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000454563079250993,
      "loss": 2.5633,
      "theoretical_loss": 3.6490512658611562,
      "tokens_seen": 1000407040
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004545465292226215,
      "loss": 2.8064,
      "theoretical_loss": 3.6490285191700074,
      "tokens_seen": 1000472576
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045452997919425,
      "loss": 2.9631,
      "theoretical_loss": 3.6490057743860076,
      "tokens_seen": 1000538112
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045451342916587854,
      "loss": 2.6564,
      "theoretical_loss": 3.6489830315088727,
      "tokens_seen": 1000603648
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004544968791375071,
      "loss": 3.0575,
      "theoretical_loss": 3.6489602905383176,
      "tokens_seen": 1000669184
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004544803291091356,
      "loss": 2.6157,
      "theoretical_loss": 3.6489375514740576,
      "tokens_seen": 1000734720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004544637790807641,
      "loss": 2.8832,
      "theoretical_loss": 3.648914814315809,
      "tokens_seen": 1000800256
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045444722905239263,
      "loss": 2.7833,
      "theoretical_loss": 3.6488920790632857,
      "tokens_seen": 1000865792
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045443067902402117,
      "loss": 2.924,
      "theoretical_loss": 3.6488693457162045,
      "tokens_seen": 1000931328
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004544141289956497,
      "loss": 2.8113,
      "theoretical_loss": 3.64884661427428,
      "tokens_seen": 1000996864
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004543975789672782,
      "loss": 2.853,
      "theoretical_loss": 3.6488238847372285,
      "tokens_seen": 1001062400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045438102893890673,
      "loss": 2.627,
      "theoretical_loss": 3.6488011571047654,
      "tokens_seen": 1001127936
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045436447891053527,
      "loss": 2.8491,
      "theoretical_loss": 3.6487784313766065,
      "tokens_seen": 1001193472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004543479288821638,
      "loss": 2.7638,
      "theoretical_loss": 3.6487557075524677,
      "tokens_seen": 1001259008
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004543313788537923,
      "loss": 2.5677,
      "theoretical_loss": 3.6487329856320647,
      "tokens_seen": 1001324544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004543148288254208,
      "loss": 2.8312,
      "theoretical_loss": 3.6487102656151134,
      "tokens_seen": 1001390080
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045429827879704936,
      "loss": 3.0102,
      "theoretical_loss": 3.6486875475013303,
      "tokens_seen": 1001455616
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004542817287686779,
      "loss": 2.819,
      "theoretical_loss": 3.6486648312904304,
      "tokens_seen": 1001521152
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004542651787403064,
      "loss": 2.7453,
      "theoretical_loss": 3.648642116982131,
      "tokens_seen": 1001586688
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004542486287119349,
      "loss": 2.7068,
      "theoretical_loss": 3.6486194045761473,
      "tokens_seen": 1001652224
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045423207868356346,
      "loss": 2.7207,
      "theoretical_loss": 3.6485966940721957,
      "tokens_seen": 1001717760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000454215528655192,
      "loss": 2.7188,
      "theoretical_loss": 3.6485739854699935,
      "tokens_seen": 1001783296
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004541989786268205,
      "loss": 2.8651,
      "theoretical_loss": 3.648551278769256,
      "tokens_seen": 1001848832
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000454182428598449,
      "loss": 2.6904,
      "theoretical_loss": 3.6485285739696995,
      "tokens_seen": 1001914368
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045416587857007756,
      "loss": 2.7599,
      "theoretical_loss": 3.648505871071041,
      "tokens_seen": 1001979904
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004541493285417061,
      "loss": 2.6311,
      "theoretical_loss": 3.648483170072997,
      "tokens_seen": 1002045440
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004541327785133346,
      "loss": 2.6169,
      "theoretical_loss": 3.6484604709752846,
      "tokens_seen": 1002110976
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004541162284849631,
      "loss": 2.6872,
      "theoretical_loss": 3.6484377737776192,
      "tokens_seen": 1002176512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045409967845659165,
      "loss": 2.6088,
      "theoretical_loss": 3.6484150784797187,
      "tokens_seen": 1002242048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004540831284282202,
      "loss": 2.6948,
      "theoretical_loss": 3.6483923850812987,
      "tokens_seen": 1002307584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045406657839984867,
      "loss": 2.718,
      "theoretical_loss": 3.6483696935820777,
      "tokens_seen": 1002373120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004540500283714772,
      "loss": 2.8948,
      "theoretical_loss": 3.6483470039817707,
      "tokens_seen": 1002438656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045403347834310575,
      "loss": 2.8599,
      "theoretical_loss": 3.648324316280096,
      "tokens_seen": 1002504192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004540169283147343,
      "loss": 2.6277,
      "theoretical_loss": 3.6483016304767704,
      "tokens_seen": 1002569728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045400037828636277,
      "loss": 2.9152,
      "theoretical_loss": 3.6482789465715104,
      "tokens_seen": 1002635264
    },
    {
      "epoch": 0.36,
      "objective/train/avg_token_score": 0.010111908428370953,
      "objective/train/avg_weight": 0.9881002902984619,
      "objective/train/docs_used": 571707,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.949240207672119,
      "objective/train/std_weight": 0.04590030014514923,
      "objective/train/theoretical_loss": 3.6482562645640337,
      "objective/train/tokens_used": 1023160800,
      "theoretical_loss": 3.6482562645640337,
      "tokens_seen": 1002700800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004539838282579913,
      "loss": 2.9915,
      "theoretical_loss": 3.6482562645640337,
      "tokens_seen": 1002700800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045396727822961984,
      "loss": 2.9099,
      "theoretical_loss": 3.648233584454057,
      "tokens_seen": 1002766336
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004539507282012484,
      "loss": 2.5263,
      "theoretical_loss": 3.6482109062412986,
      "tokens_seen": 1002831872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045393417817287686,
      "loss": 2.7365,
      "theoretical_loss": 3.6481882299254744,
      "tokens_seen": 1002897408
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004539176281445054,
      "loss": 2.6113,
      "theoretical_loss": 3.648165555506303,
      "tokens_seen": 1002962944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045390107811613394,
      "loss": 2.6917,
      "theoretical_loss": 3.648142882983501,
      "tokens_seen": 1003028480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004538845280877625,
      "loss": 2.7851,
      "theoretical_loss": 3.6481202123567864,
      "tokens_seen": 1003094016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045386797805939096,
      "loss": 2.6478,
      "theoretical_loss": 3.6480975436258767,
      "tokens_seen": 1003159552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004538514280310195,
      "loss": 2.3414,
      "theoretical_loss": 3.6480748767904894,
      "tokens_seen": 1003225088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045383487800264804,
      "loss": 2.8098,
      "theoretical_loss": 3.6480522118503425,
      "tokens_seen": 1003290624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004538183279742766,
      "loss": 2.7197,
      "theoretical_loss": 3.6480295488051535,
      "tokens_seen": 1003356160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045380177794590506,
      "loss": 2.5775,
      "theoretical_loss": 3.64800688765464,
      "tokens_seen": 1003421696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004537852279175336,
      "loss": 2.7606,
      "theoretical_loss": 3.64798422839852,
      "tokens_seen": 1003487232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045376867788916213,
      "loss": 2.8392,
      "theoretical_loss": 3.647961571036512,
      "tokens_seen": 1003552768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045375212786079067,
      "loss": 2.7796,
      "theoretical_loss": 3.647938915568333,
      "tokens_seen": 1003618304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045373557783241915,
      "loss": 2.6413,
      "theoretical_loss": 3.647916261993702,
      "tokens_seen": 1003683840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045371902780404764,
      "loss": 3.0259,
      "theoretical_loss": 3.6478936103123365,
      "tokens_seen": 1003749376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004537024777756761,
      "loss": 2.7191,
      "theoretical_loss": 3.6478709605239548,
      "tokens_seen": 1003814912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045368592774730466,
      "loss": 2.8428,
      "theoretical_loss": 3.647848312628275,
      "tokens_seen": 1003880448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004536693777189332,
      "loss": 2.457,
      "theoretical_loss": 3.6478256666250157,
      "tokens_seen": 1003945984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045365282769056173,
      "loss": 2.7354,
      "theoretical_loss": 3.6478030225138953,
      "tokens_seen": 1004011520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004536362776621902,
      "loss": 2.9139,
      "theoretical_loss": 3.647780380294632,
      "tokens_seen": 1004077056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045361972763381875,
      "loss": 2.5217,
      "theoretical_loss": 3.647757739966944,
      "tokens_seen": 1004142592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004536031776054473,
      "loss": 2.8505,
      "theoretical_loss": 3.6477351015305506,
      "tokens_seen": 1004208128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045358662757707583,
      "loss": 2.7846,
      "theoretical_loss": 3.6477124649851698,
      "tokens_seen": 1004273664
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004535700775487043,
      "loss": 2.4388,
      "theoretical_loss": 3.64768983033052,
      "tokens_seen": 1004339200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045355352752033285,
      "loss": 2.5798,
      "theoretical_loss": 3.647667197566321,
      "tokens_seen": 1004404736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004535369774919614,
      "loss": 2.5702,
      "theoretical_loss": 3.6476445666922905,
      "tokens_seen": 1004470272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004535204274635899,
      "loss": 2.6374,
      "theoretical_loss": 3.6476219377081476,
      "tokens_seen": 1004535808
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045350387743521846,
      "loss": 2.7904,
      "theoretical_loss": 3.6475993106136118,
      "tokens_seen": 1004601344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045348732740684694,
      "loss": 2.5964,
      "theoretical_loss": 3.647576685408401,
      "tokens_seen": 1004666880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004534707773784755,
      "loss": 2.9311,
      "theoretical_loss": 3.6475540620922353,
      "tokens_seen": 1004732416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000453454227350104,
      "loss": 3.0192,
      "theoretical_loss": 3.6475314406648334,
      "tokens_seen": 1004797952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045343767732173256,
      "loss": 2.711,
      "theoretical_loss": 3.647508821125914,
      "tokens_seen": 1004863488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045342112729336104,
      "loss": 2.6979,
      "theoretical_loss": 3.6474862034751965,
      "tokens_seen": 1004929024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004534045772649896,
      "loss": 2.6619,
      "theoretical_loss": 3.6474635877124,
      "tokens_seen": 1004994560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004533880272366181,
      "loss": 2.6999,
      "theoretical_loss": 3.6474409738372446,
      "tokens_seen": 1005060096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045337147720824665,
      "loss": 2.9582,
      "theoretical_loss": 3.647418361849449,
      "tokens_seen": 1005125632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045335492717987514,
      "loss": 2.78,
      "theoretical_loss": 3.647395751748733,
      "tokens_seen": 1005191168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004533383771515037,
      "loss": 3.0048,
      "theoretical_loss": 3.6473731435348156,
      "tokens_seen": 1005256704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004533218271231322,
      "loss": 2.8297,
      "theoretical_loss": 3.6473505372074166,
      "tokens_seen": 1005322240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045330527709476075,
      "loss": 2.7811,
      "theoretical_loss": 3.647327932766256,
      "tokens_seen": 1005387776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045328872706638923,
      "loss": 2.9109,
      "theoretical_loss": 3.647305330211053,
      "tokens_seen": 1005453312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045327217703801777,
      "loss": 2.859,
      "theoretical_loss": 3.6472827295415273,
      "tokens_seen": 1005518848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004532556270096463,
      "loss": 2.9247,
      "theoretical_loss": 3.647260130757399,
      "tokens_seen": 1005584384
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045323907698127485,
      "loss": 2.6924,
      "theoretical_loss": 3.647237533858388,
      "tokens_seen": 1005649920
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045322252695290333,
      "loss": 2.6967,
      "theoretical_loss": 3.647214938844214,
      "tokens_seen": 1005715456
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045320597692453187,
      "loss": 2.8435,
      "theoretical_loss": 3.647192345714597,
      "tokens_seen": 1005780992
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004531894268961604,
      "loss": 2.713,
      "theoretical_loss": 3.6471697544692567,
      "tokens_seen": 1005846528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045317287686778894,
      "loss": 2.9197,
      "theoretical_loss": 3.6471471651079135,
      "tokens_seen": 1005912064
    },
    {
      "epoch": 0.36,
      "objective/train/avg_token_score": 0.015042291954159737,
      "objective/train/avg_weight": 0.982290506362915,
      "objective/train/docs_used": 573416,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.902454376220703,
      "objective/train/std_weight": 0.06674180179834366,
      "objective/train/theoretical_loss": 3.6471245776302883,
      "objective/train/tokens_used": 1026437600,
      "theoretical_loss": 3.6471245776302883,
      "tokens_seen": 1005977600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004531563268394174,
      "loss": 2.8468,
      "theoretical_loss": 3.6471245776302883,
      "tokens_seen": 1005977600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045313977681104596,
      "loss": 2.8305,
      "theoretical_loss": 3.6471019920361,
      "tokens_seen": 1006043136
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004531232267826745,
      "loss": 2.7046,
      "theoretical_loss": 3.6470794083250695,
      "tokens_seen": 1006108672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045310667675430304,
      "loss": 2.7797,
      "theoretical_loss": 3.6470568264969176,
      "tokens_seen": 1006174208
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004530901267259315,
      "loss": 2.7982,
      "theoretical_loss": 3.647034246551364,
      "tokens_seen": 1006239744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045307357669756006,
      "loss": 2.8851,
      "theoretical_loss": 3.6470116684881297,
      "tokens_seen": 1006305280
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004530570266691886,
      "loss": 2.8775,
      "theoretical_loss": 3.646989092306934,
      "tokens_seen": 1006370816
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045304047664081713,
      "loss": 2.8858,
      "theoretical_loss": 3.646966518007499,
      "tokens_seen": 1006436352
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004530239266124456,
      "loss": 2.9844,
      "theoretical_loss": 3.6469439455895447,
      "tokens_seen": 1006501888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045300737658407415,
      "loss": 2.9247,
      "theoretical_loss": 3.646921375052792,
      "tokens_seen": 1006567424
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004529908265557027,
      "loss": 2.7672,
      "theoretical_loss": 3.646898806396961,
      "tokens_seen": 1006632960
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045297427652733123,
      "loss": 2.7656,
      "theoretical_loss": 3.646876239621773,
      "tokens_seen": 1006698496
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004529577264989597,
      "loss": 2.8357,
      "theoretical_loss": 3.646853674726949,
      "tokens_seen": 1006764032
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045294117647058825,
      "loss": 2.7634,
      "theoretical_loss": 3.64683111171221,
      "tokens_seen": 1006829568
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004529246264422168,
      "loss": 2.9668,
      "theoretical_loss": 3.6468085505772763,
      "tokens_seen": 1006895104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004529080764138453,
      "loss": 2.6423,
      "theoretical_loss": 3.6467859913218694,
      "tokens_seen": 1006960640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004528915263854738,
      "loss": 2.6966,
      "theoretical_loss": 3.646763433945711,
      "tokens_seen": 1007026176
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045287497635710235,
      "loss": 2.6303,
      "theoretical_loss": 3.6467408784485214,
      "tokens_seen": 1007091712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004528584263287309,
      "loss": 2.69,
      "theoretical_loss": 3.6467183248300215,
      "tokens_seen": 1007157248
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004528418763003594,
      "loss": 2.6483,
      "theoretical_loss": 3.6466957730899336,
      "tokens_seen": 1007222784
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004528253262719879,
      "loss": 3.0001,
      "theoretical_loss": 3.646673223227979,
      "tokens_seen": 1007288320
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045280877624361644,
      "loss": 2.5828,
      "theoretical_loss": 3.6466506752438788,
      "tokens_seen": 1007353856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000452792226215245,
      "loss": 2.6246,
      "theoretical_loss": 3.646628129137354,
      "tokens_seen": 1007419392
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004527756761868735,
      "loss": 2.5669,
      "theoretical_loss": 3.6466055849081265,
      "tokens_seen": 1007484928
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000452759126158502,
      "loss": 2.9206,
      "theoretical_loss": 3.6465830425559185,
      "tokens_seen": 1007550464
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045274257613013054,
      "loss": 2.8698,
      "theoretical_loss": 3.6465605020804506,
      "tokens_seen": 1007616000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045272602610175897,
      "loss": 2.7005,
      "theoretical_loss": 3.646537963481445,
      "tokens_seen": 1007681536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004527094760733875,
      "loss": 2.7673,
      "theoretical_loss": 3.646515426758624,
      "tokens_seen": 1007747072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045269292604501604,
      "loss": 2.6573,
      "theoretical_loss": 3.646492891911708,
      "tokens_seen": 1007812608
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004526763760166446,
      "loss": 2.9548,
      "theoretical_loss": 3.6464703589404204,
      "tokens_seen": 1007878144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045265982598827306,
      "loss": 3.1034,
      "theoretical_loss": 3.646447827844482,
      "tokens_seen": 1007943680
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004526432759599016,
      "loss": 2.793,
      "theoretical_loss": 3.646425298623616,
      "tokens_seen": 1008009216
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045262672593153014,
      "loss": 2.5489,
      "theoretical_loss": 3.6464027712775433,
      "tokens_seen": 1008074752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004526101759031587,
      "loss": 2.4603,
      "theoretical_loss": 3.646380245805986,
      "tokens_seen": 1008140288
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045259362587478716,
      "loss": 2.6564,
      "theoretical_loss": 3.6463577222086676,
      "tokens_seen": 1008205824
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004525770758464157,
      "loss": 2.9085,
      "theoretical_loss": 3.6463352004853093,
      "tokens_seen": 1008271360
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045256052581804423,
      "loss": 2.6687,
      "theoretical_loss": 3.646312680635633,
      "tokens_seen": 1008336896
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045254397578967277,
      "loss": 2.8197,
      "theoretical_loss": 3.646290162659363,
      "tokens_seen": 1008402432
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045252742576130125,
      "loss": 2.6189,
      "theoretical_loss": 3.646267646556219,
      "tokens_seen": 1008467968
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004525108757329298,
      "loss": 2.5945,
      "theoretical_loss": 3.6462451323259257,
      "tokens_seen": 1008533504
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045249432570455833,
      "loss": 2.8691,
      "theoretical_loss": 3.6462226199682046,
      "tokens_seen": 1008599040
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045247777567618687,
      "loss": 2.7079,
      "theoretical_loss": 3.646200109482778,
      "tokens_seen": 1008664576
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045246122564781535,
      "loss": 2.806,
      "theoretical_loss": 3.646177600869369,
      "tokens_seen": 1008730112
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004524446756194439,
      "loss": 2.7526,
      "theoretical_loss": 3.646155094127701,
      "tokens_seen": 1008795648
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004524281255910724,
      "loss": 2.7696,
      "theoretical_loss": 3.646132589257496,
      "tokens_seen": 1008861184
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045241157556270096,
      "loss": 2.7738,
      "theoretical_loss": 3.6461100862584765,
      "tokens_seen": 1008926720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045239502553432945,
      "loss": 2.6261,
      "theoretical_loss": 3.646087585130366,
      "tokens_seen": 1008992256
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000452378475505958,
      "loss": 2.9447,
      "theoretical_loss": 3.6460650858728876,
      "tokens_seen": 1009057792
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004523619254775865,
      "loss": 3.0866,
      "theoretical_loss": 3.6460425884857637,
      "tokens_seen": 1009123328
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045234537544921506,
      "loss": 2.7438,
      "theoretical_loss": 3.6460200929687177,
      "tokens_seen": 1009188864
    },
    {
      "epoch": 0.36,
      "objective/train/avg_token_score": 0.008118830621242523,
      "objective/train/avg_weight": 0.9904751777648926,
      "objective/train/docs_used": 575328,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.643214225769043,
      "objective/train/std_weight": 0.043537985533475876,
      "objective/train/theoretical_loss": 3.6459975993214724,
      "objective/train/tokens_used": 1029714400,
      "theoretical_loss": 3.6459975993214724,
      "tokens_seen": 1009254400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045232882542084354,
      "loss": 2.6509,
      "theoretical_loss": 3.6459975993214724,
      "tokens_seen": 1009254400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004523122753924721,
      "loss": 2.8568,
      "theoretical_loss": 3.6459751075437516,
      "tokens_seen": 1009319936
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004522957253641006,
      "loss": 2.6883,
      "theoretical_loss": 3.645952617635278,
      "tokens_seen": 1009385472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045227917533572916,
      "loss": 2.6479,
      "theoretical_loss": 3.645930129595775,
      "tokens_seen": 1009451008
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045226262530735764,
      "loss": 2.8125,
      "theoretical_loss": 3.645907643424966,
      "tokens_seen": 1009516544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004522460752789862,
      "loss": 2.6395,
      "theoretical_loss": 3.6458851591225754,
      "tokens_seen": 1009582080
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004522295252506147,
      "loss": 2.5597,
      "theoretical_loss": 3.645862676688325,
      "tokens_seen": 1009647616
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045221297522224325,
      "loss": 2.6259,
      "theoretical_loss": 3.645840196121939,
      "tokens_seen": 1009713152
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045219642519387173,
      "loss": 2.787,
      "theoretical_loss": 3.6458177174231414,
      "tokens_seen": 1009778688
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045217987516550027,
      "loss": 2.7962,
      "theoretical_loss": 3.6457952405916556,
      "tokens_seen": 1009844224
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004521633251371288,
      "loss": 2.5847,
      "theoretical_loss": 3.6457727656272043,
      "tokens_seen": 1009909760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045214677510875735,
      "loss": 2.7521,
      "theoretical_loss": 3.645750292529513,
      "tokens_seen": 1009975296
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045213022508038583,
      "loss": 2.8923,
      "theoretical_loss": 3.6457278212983053,
      "tokens_seen": 1010040832
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045211367505201437,
      "loss": 2.5711,
      "theoretical_loss": 3.645705351933304,
      "tokens_seen": 1010106368
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004520971250236429,
      "loss": 2.586,
      "theoretical_loss": 3.6456828844342337,
      "tokens_seen": 1010171904
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045208057499527144,
      "loss": 2.6254,
      "theoretical_loss": 3.6456604188008184,
      "tokens_seen": 1010237440
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004520640249668999,
      "loss": 2.8083,
      "theoretical_loss": 3.6456379550327815,
      "tokens_seen": 1010302976
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045204747493852846,
      "loss": 2.8454,
      "theoretical_loss": 3.6456154931298483,
      "tokens_seen": 1010368512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000452030924910157,
      "loss": 2.4983,
      "theoretical_loss": 3.6455930330917425,
      "tokens_seen": 1010434048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045201437488178554,
      "loss": 2.7345,
      "theoretical_loss": 3.645570574918188,
      "tokens_seen": 1010499584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000451997824853414,
      "loss": 2.7064,
      "theoretical_loss": 3.645548118608909,
      "tokens_seen": 1010565120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045198127482504256,
      "loss": 2.8239,
      "theoretical_loss": 3.6455256641636304,
      "tokens_seen": 1010630656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004519647247966711,
      "loss": 2.6667,
      "theoretical_loss": 3.6455032115820765,
      "tokens_seen": 1010696192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045194817476829964,
      "loss": 2.58,
      "theoretical_loss": 3.645480760863972,
      "tokens_seen": 1010761728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045193162473992817,
      "loss": 2.9637,
      "theoretical_loss": 3.6454583120090405,
      "tokens_seen": 1010827264
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045191507471155666,
      "loss": 2.6844,
      "theoretical_loss": 3.6454358650170073,
      "tokens_seen": 1010892800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004518985246831852,
      "loss": 3.0124,
      "theoretical_loss": 3.645413419887597,
      "tokens_seen": 1010958336
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045188197465481373,
      "loss": 2.6426,
      "theoretical_loss": 3.6453909766205346,
      "tokens_seen": 1011023872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045186542462644227,
      "loss": 2.8293,
      "theoretical_loss": 3.6453685352155443,
      "tokens_seen": 1011089408
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045184887459807075,
      "loss": 2.8597,
      "theoretical_loss": 3.645346095672352,
      "tokens_seen": 1011154944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004518323245696993,
      "loss": 2.7299,
      "theoretical_loss": 3.6453236579906805,
      "tokens_seen": 1011220480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004518157745413278,
      "loss": 2.686,
      "theoretical_loss": 3.6453012221702563,
      "tokens_seen": 1011286016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045179922451295636,
      "loss": 2.7561,
      "theoretical_loss": 3.6452787882108044,
      "tokens_seen": 1011351552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045178267448458485,
      "loss": 2.8923,
      "theoretical_loss": 3.64525635611205,
      "tokens_seen": 1011417088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004517661244562134,
      "loss": 2.6979,
      "theoretical_loss": 3.645233925873717,
      "tokens_seen": 1011482624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004517495744278418,
      "loss": 2.7981,
      "theoretical_loss": 3.6452114974955316,
      "tokens_seen": 1011548160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045173302439947035,
      "loss": 2.7483,
      "theoretical_loss": 3.645189070977219,
      "tokens_seen": 1011613696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004517164743710989,
      "loss": 2.7658,
      "theoretical_loss": 3.645166646318504,
      "tokens_seen": 1011679232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045169992434272743,
      "loss": 2.7769,
      "theoretical_loss": 3.6451442235191127,
      "tokens_seen": 1011744768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004516833743143559,
      "loss": 2.7755,
      "theoretical_loss": 3.64512180257877,
      "tokens_seen": 1011810304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045166682428598445,
      "loss": 2.777,
      "theoretical_loss": 3.645099383497201,
      "tokens_seen": 1011875840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000451650274257613,
      "loss": 2.929,
      "theoretical_loss": 3.6450769662741322,
      "tokens_seen": 1011941376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004516337242292415,
      "loss": 2.5635,
      "theoretical_loss": 3.6450545509092884,
      "tokens_seen": 1012006912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045161717420087,
      "loss": 2.7608,
      "theoretical_loss": 3.6450321374023957,
      "tokens_seen": 1012072448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045160062417249854,
      "loss": 2.6996,
      "theoretical_loss": 3.6450097257531793,
      "tokens_seen": 1012137984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004515840741441271,
      "loss": 2.772,
      "theoretical_loss": 3.6449873159613655,
      "tokens_seen": 1012203520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004515675241157556,
      "loss": 2.9137,
      "theoretical_loss": 3.64496490802668,
      "tokens_seen": 1012269056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004515509740873841,
      "loss": 2.63,
      "theoretical_loss": 3.644942501948848,
      "tokens_seen": 1012334592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045153442405901264,
      "loss": 2.926,
      "theoretical_loss": 3.6449200977275966,
      "tokens_seen": 1012400128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004515178740306412,
      "loss": 2.5526,
      "theoretical_loss": 3.6448976953626513,
      "tokens_seen": 1012465664
    },
    {
      "epoch": 0.36,
      "objective/train/avg_token_score": 0.031021516770124435,
      "objective/train/avg_weight": 0.9634996056556702,
      "objective/train/docs_used": 577263,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.864872932434082,
      "objective/train/std_weight": 0.11461728066205978,
      "objective/train/theoretical_loss": 3.6448752948537377,
      "objective/train/tokens_used": 1032991200,
      "theoretical_loss": 3.6448752948537377,
      "tokens_seen": 1012531200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004515013240022697,
      "loss": 2.7512,
      "theoretical_loss": 3.6448752948537377,
      "tokens_seen": 1012531200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004514847739738982,
      "loss": 2.7645,
      "theoretical_loss": 3.6448528962005824,
      "tokens_seen": 1012596736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045146822394552674,
      "loss": 2.8856,
      "theoretical_loss": 3.6448304994029113,
      "tokens_seen": 1012662272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004514516739171553,
      "loss": 2.6909,
      "theoretical_loss": 3.644808104460451,
      "tokens_seen": 1012727808
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004514351238887838,
      "loss": 3.0078,
      "theoretical_loss": 3.6447857113729274,
      "tokens_seen": 1012793344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004514185738604123,
      "loss": 2.7666,
      "theoretical_loss": 3.6447633201400667,
      "tokens_seen": 1012858880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045140202383204083,
      "loss": 2.7096,
      "theoretical_loss": 3.6447409307615963,
      "tokens_seen": 1012924416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045138547380366937,
      "loss": 2.5965,
      "theoretical_loss": 3.644718543237241,
      "tokens_seen": 1012989952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004513689237752979,
      "loss": 2.9263,
      "theoretical_loss": 3.644696157566729,
      "tokens_seen": 1013055488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004513523737469264,
      "loss": 2.5618,
      "theoretical_loss": 3.644673773749786,
      "tokens_seen": 1013121024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045133582371855493,
      "loss": 2.7266,
      "theoretical_loss": 3.644651391786139,
      "tokens_seen": 1013186560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045131927369018347,
      "loss": 2.9818,
      "theoretical_loss": 3.6446290116755145,
      "tokens_seen": 1013252096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000451302723661812,
      "loss": 2.7844,
      "theoretical_loss": 3.644606633417639,
      "tokens_seen": 1013317632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004512861736334405,
      "loss": 2.9313,
      "theoretical_loss": 3.6445842570122395,
      "tokens_seen": 1013383168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000451269623605069,
      "loss": 2.5514,
      "theoretical_loss": 3.644561882459043,
      "tokens_seen": 1013448704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045125307357669756,
      "loss": 2.4472,
      "theoretical_loss": 3.6445395097577764,
      "tokens_seen": 1013514240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004512365235483261,
      "loss": 2.8376,
      "theoretical_loss": 3.6445171389081663,
      "tokens_seen": 1013579776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004512199735199546,
      "loss": 2.7798,
      "theoretical_loss": 3.64449476990994,
      "tokens_seen": 1013645312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004512034234915831,
      "loss": 2.8761,
      "theoretical_loss": 3.6444724027628252,
      "tokens_seen": 1013710848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045118687346321166,
      "loss": 3.0784,
      "theoretical_loss": 3.644450037466548,
      "tokens_seen": 1013776384
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004511703234348402,
      "loss": 2.9122,
      "theoretical_loss": 3.6444276740208363,
      "tokens_seen": 1013841920
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004511537734064687,
      "loss": 2.8655,
      "theoretical_loss": 3.6444053124254174,
      "tokens_seen": 1013907456
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004511372233780972,
      "loss": 2.7514,
      "theoretical_loss": 3.644382952680018,
      "tokens_seen": 1013972992
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045112067334972575,
      "loss": 2.6196,
      "theoretical_loss": 3.6443605947843665,
      "tokens_seen": 1014038528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004511041233213543,
      "loss": 2.6814,
      "theoretical_loss": 3.644338238738189,
      "tokens_seen": 1014104064
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004510875732929828,
      "loss": 2.9844,
      "theoretical_loss": 3.6443158845412142,
      "tokens_seen": 1014169600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004510710232646113,
      "loss": 2.6697,
      "theoretical_loss": 3.6442935321931693,
      "tokens_seen": 1014235136
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045105447323623985,
      "loss": 2.8923,
      "theoretical_loss": 3.6442711816937816,
      "tokens_seen": 1014300672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004510379232078684,
      "loss": 2.9739,
      "theoretical_loss": 3.644248833042779,
      "tokens_seen": 1014366208
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045102137317949687,
      "loss": 2.7314,
      "theoretical_loss": 3.644226486239889,
      "tokens_seen": 1014431744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004510048231511254,
      "loss": 2.7253,
      "theoretical_loss": 3.64420414128484,
      "tokens_seen": 1014497280
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045098827312275395,
      "loss": 2.6588,
      "theoretical_loss": 3.6441817981773594,
      "tokens_seen": 1014562816
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004509717230943825,
      "loss": 2.8967,
      "theoretical_loss": 3.6441594569171754,
      "tokens_seen": 1014628352
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045095517306601097,
      "loss": 2.552,
      "theoretical_loss": 3.6441371175040156,
      "tokens_seen": 1014693888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004509386230376395,
      "loss": 2.9268,
      "theoretical_loss": 3.644114779937608,
      "tokens_seen": 1014759424
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045092207300926804,
      "loss": 2.7675,
      "theoretical_loss": 3.644092444217681,
      "tokens_seen": 1014824960
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004509055229808966,
      "loss": 2.8567,
      "theoretical_loss": 3.6440701103439626,
      "tokens_seen": 1014890496
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045088897295252506,
      "loss": 2.8381,
      "theoretical_loss": 3.644047778316181,
      "tokens_seen": 1014956032
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004508724229241536,
      "loss": 3.0865,
      "theoretical_loss": 3.644025448134064,
      "tokens_seen": 1015021568
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045085587289578214,
      "loss": 2.7454,
      "theoretical_loss": 3.644003119797341,
      "tokens_seen": 1015087104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004508393228674107,
      "loss": 2.7472,
      "theoretical_loss": 3.6439807933057393,
      "tokens_seen": 1015152640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045082277283903916,
      "loss": 2.8686,
      "theoretical_loss": 3.643958468658988,
      "tokens_seen": 1015218176
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004508062228106677,
      "loss": 2.7026,
      "theoretical_loss": 3.6439361458568147,
      "tokens_seen": 1015283712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045078967278229623,
      "loss": 2.878,
      "theoretical_loss": 3.6439138248989495,
      "tokens_seen": 1015349248
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045077312275392477,
      "loss": 2.7716,
      "theoretical_loss": 3.6438915057851196,
      "tokens_seen": 1015414784
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004507565727255532,
      "loss": 2.7123,
      "theoretical_loss": 3.643869188515054,
      "tokens_seen": 1015480320
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045074002269718174,
      "loss": 2.9137,
      "theoretical_loss": 3.6438468730884814,
      "tokens_seen": 1015545856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004507234726688103,
      "loss": 2.7255,
      "theoretical_loss": 3.643824559505131,
      "tokens_seen": 1015611392
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045070692264043876,
      "loss": 2.8515,
      "theoretical_loss": 3.6438022477647314,
      "tokens_seen": 1015676928
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004506903726120673,
      "loss": 2.707,
      "theoretical_loss": 3.643779937867011,
      "tokens_seen": 1015742464
    },
    {
      "epoch": 0.36,
      "objective/train/avg_token_score": 0.020931856706738472,
      "objective/train/avg_weight": 0.9753589034080505,
      "objective/train/docs_used": 579051,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.275315761566162,
      "objective/train/std_weight": 0.11164892464876175,
      "objective/train/theoretical_loss": 3.6437576298116996,
      "objective/train/tokens_used": 1036268000,
      "theoretical_loss": 3.6437576298116996,
      "tokens_seen": 1015808000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045067382258369583,
      "loss": 2.8686,
      "theoretical_loss": 3.6437576298116996,
      "tokens_seen": 1015808000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045065727255532437,
      "loss": 2.6741,
      "theoretical_loss": 3.6437353235985253,
      "tokens_seen": 1015873536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045064072252695285,
      "loss": 2.6387,
      "theoretical_loss": 3.6437130192272176,
      "tokens_seen": 1015939072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004506241724985814,
      "loss": 2.8104,
      "theoretical_loss": 3.6436907166975057,
      "tokens_seen": 1016004608
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045060762247020993,
      "loss": 2.7259,
      "theoretical_loss": 3.6436684160091186,
      "tokens_seen": 1016070144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045059107244183847,
      "loss": 2.8166,
      "theoretical_loss": 3.6436461171617864,
      "tokens_seen": 1016135680
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045057452241346695,
      "loss": 2.9153,
      "theoretical_loss": 3.643623820155237,
      "tokens_seen": 1016201216
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004505579723850955,
      "loss": 2.7634,
      "theoretical_loss": 3.6436015249892,
      "tokens_seen": 1016266752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000450541422356724,
      "loss": 2.7393,
      "theoretical_loss": 3.6435792316634057,
      "tokens_seen": 1016332288
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045052487232835256,
      "loss": 2.8097,
      "theoretical_loss": 3.6435569401775822,
      "tokens_seen": 1016397824
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045050832229998105,
      "loss": 2.7151,
      "theoretical_loss": 3.6435346505314605,
      "tokens_seen": 1016463360
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004504917722716096,
      "loss": 2.6393,
      "theoretical_loss": 3.643512362724769,
      "tokens_seen": 1016528896
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004504752222432381,
      "loss": 2.7056,
      "theoretical_loss": 3.6434900767572382,
      "tokens_seen": 1016594432
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045045867221486666,
      "loss": 2.9792,
      "theoretical_loss": 3.6434677926285977,
      "tokens_seen": 1016659968
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045044212218649514,
      "loss": 2.7466,
      "theoretical_loss": 3.6434455103385766,
      "tokens_seen": 1016725504
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004504255721581237,
      "loss": 2.8371,
      "theoretical_loss": 3.6434232298869045,
      "tokens_seen": 1016791040
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004504090221297522,
      "loss": 2.9505,
      "theoretical_loss": 3.6434009512733123,
      "tokens_seen": 1016856576
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045039247210138075,
      "loss": 2.7241,
      "theoretical_loss": 3.643378674497529,
      "tokens_seen": 1016922112
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045037592207300924,
      "loss": 2.9552,
      "theoretical_loss": 3.643356399559286,
      "tokens_seen": 1016987648
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004503593720446378,
      "loss": 3.0334,
      "theoretical_loss": 3.6433341264583117,
      "tokens_seen": 1017053184
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004503428220162663,
      "loss": 2.5243,
      "theoretical_loss": 3.6433118551943364,
      "tokens_seen": 1017118720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045032627198789485,
      "loss": 2.8399,
      "theoretical_loss": 3.643289585767091,
      "tokens_seen": 1017184256
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045030972195952333,
      "loss": 2.9352,
      "theoretical_loss": 3.6432673181763056,
      "tokens_seen": 1017249792
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045029317193115187,
      "loss": 2.7091,
      "theoretical_loss": 3.64324505242171,
      "tokens_seen": 1017315328
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004502766219027804,
      "loss": 2.8292,
      "theoretical_loss": 3.643222788503034,
      "tokens_seen": 1017380864
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045026007187440895,
      "loss": 2.5773,
      "theoretical_loss": 3.64320052642001,
      "tokens_seen": 1017446400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045024352184603743,
      "loss": 2.7239,
      "theoretical_loss": 3.643178266172366,
      "tokens_seen": 1017511936
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045022697181766597,
      "loss": 2.922,
      "theoretical_loss": 3.643156007759834,
      "tokens_seen": 1017577472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004502104217892945,
      "loss": 2.5072,
      "theoretical_loss": 3.6431337511821438,
      "tokens_seen": 1017643008
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045019387176092304,
      "loss": 2.7446,
      "theoretical_loss": 3.643111496439027,
      "tokens_seen": 1017708544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004501773217325515,
      "loss": 2.5073,
      "theoretical_loss": 3.6430892435302136,
      "tokens_seen": 1017774080
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045016077170418006,
      "loss": 2.6441,
      "theoretical_loss": 3.6430669924554335,
      "tokens_seen": 1017839616
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004501442216758086,
      "loss": 2.6816,
      "theoretical_loss": 3.6430447432144186,
      "tokens_seen": 1017905152
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045012767164743714,
      "loss": 2.8545,
      "theoretical_loss": 3.6430224958068997,
      "tokens_seen": 1017970688
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004501111216190656,
      "loss": 2.8307,
      "theoretical_loss": 3.643000250232607,
      "tokens_seen": 1018036224
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045009457159069416,
      "loss": 2.504,
      "theoretical_loss": 3.642978006491272,
      "tokens_seen": 1018101760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004500780215623227,
      "loss": 3.1483,
      "theoretical_loss": 3.6429557645826254,
      "tokens_seen": 1018167296
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045006147153395123,
      "loss": 3.1062,
      "theoretical_loss": 3.6429335245063985,
      "tokens_seen": 1018232832
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004500449215055797,
      "loss": 2.9291,
      "theoretical_loss": 3.642911286262322,
      "tokens_seen": 1018298368
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00045002837147720826,
      "loss": 2.542,
      "theoretical_loss": 3.642889049850128,
      "tokens_seen": 1018363904
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004500118214488368,
      "loss": 2.6648,
      "theoretical_loss": 3.6428668152695467,
      "tokens_seen": 1018429440
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044999527142046533,
      "loss": 2.5179,
      "theoretical_loss": 3.6428445825203095,
      "tokens_seen": 1018494976
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004499787213920938,
      "loss": 2.5641,
      "theoretical_loss": 3.642822351602148,
      "tokens_seen": 1018560512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044996217136372235,
      "loss": 2.9452,
      "theoretical_loss": 3.6428001225147937,
      "tokens_seen": 1018626048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004499456213353509,
      "loss": 2.7318,
      "theoretical_loss": 3.642777895257978,
      "tokens_seen": 1018691584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004499290713069794,
      "loss": 2.8407,
      "theoretical_loss": 3.6427556698314323,
      "tokens_seen": 1018757120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004499125212786079,
      "loss": 3.1309,
      "theoretical_loss": 3.6427334462348884,
      "tokens_seen": 1018822656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044989597125023645,
      "loss": 2.8975,
      "theoretical_loss": 3.6427112244680773,
      "tokens_seen": 1018888192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000449879421221865,
      "loss": 2.757,
      "theoretical_loss": 3.642689004530731,
      "tokens_seen": 1018953728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004498628711934935,
      "loss": 2.9237,
      "theoretical_loss": 3.642666786422582,
      "tokens_seen": 1019019264
    },
    {
      "epoch": 0.36,
      "objective/train/avg_token_score": 0.03485102951526642,
      "objective/train/avg_weight": 0.959013819694519,
      "objective/train/docs_used": 580918,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.7227768898010254,
      "objective/train/std_weight": 0.16938525438308716,
      "objective/train/theoretical_loss": 3.6426445701433607,
      "objective/train/tokens_used": 1039544800,
      "theoretical_loss": 3.6426445701433607,
      "tokens_seen": 1019084800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000449846321165122,
      "loss": 2.7244,
      "theoretical_loss": 3.6426445701433607,
      "tokens_seen": 1019084800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044982977113675054,
      "loss": 2.5923,
      "theoretical_loss": 3.6426223556928,
      "tokens_seen": 1019150336
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004498132211083791,
      "loss": 2.6087,
      "theoretical_loss": 3.6426001430706316,
      "tokens_seen": 1019215872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004497966710800076,
      "loss": 2.7988,
      "theoretical_loss": 3.642577932276587,
      "tokens_seen": 1019281408
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044978012105163605,
      "loss": 2.8376,
      "theoretical_loss": 3.642555723310399,
      "tokens_seen": 1019346944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004497635710232646,
      "loss": 2.3459,
      "theoretical_loss": 3.642533516171799,
      "tokens_seen": 1019412480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044974702099489307,
      "loss": 2.7256,
      "theoretical_loss": 3.6425113108605194,
      "tokens_seen": 1019478016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004497304709665216,
      "loss": 2.7585,
      "theoretical_loss": 3.642489107376292,
      "tokens_seen": 1019543552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044971392093815014,
      "loss": 2.7961,
      "theoretical_loss": 3.6424669057188495,
      "tokens_seen": 1019609088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004496973709097787,
      "loss": 2.9431,
      "theoretical_loss": 3.6424447058879243,
      "tokens_seen": 1019674624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044968082088140716,
      "loss": 2.9602,
      "theoretical_loss": 3.642422507883248,
      "tokens_seen": 1019740160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004496642708530357,
      "loss": 2.7529,
      "theoretical_loss": 3.642400311704554,
      "tokens_seen": 1019805696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044964772082466424,
      "loss": 2.7772,
      "theoretical_loss": 3.6423781173515746,
      "tokens_seen": 1019871232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004496311707962928,
      "loss": 2.9253,
      "theoretical_loss": 3.6423559248240416,
      "tokens_seen": 1019936768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044961462076792126,
      "loss": 2.7393,
      "theoretical_loss": 3.6423337341216877,
      "tokens_seen": 1020002304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004495980707395498,
      "loss": 2.7565,
      "theoretical_loss": 3.6423115452442465,
      "tokens_seen": 1020067840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044958152071117834,
      "loss": 2.8384,
      "theoretical_loss": 3.6422893581914497,
      "tokens_seen": 1020133376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004495649706828069,
      "loss": 2.9282,
      "theoretical_loss": 3.64226717296303,
      "tokens_seen": 1020198912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044954842065443536,
      "loss": 2.8448,
      "theoretical_loss": 3.642244989558721,
      "tokens_seen": 1020264448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004495318706260639,
      "loss": 2.7625,
      "theoretical_loss": 3.6422228079782553,
      "tokens_seen": 1020329984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044951532059769243,
      "loss": 2.7512,
      "theoretical_loss": 3.6422006282213655,
      "tokens_seen": 1020395520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044949877056932097,
      "loss": 2.7315,
      "theoretical_loss": 3.642178450287785,
      "tokens_seen": 1020461056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004494822205409495,
      "loss": 2.6818,
      "theoretical_loss": 3.642156274177246,
      "tokens_seen": 1020526592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000449465670512578,
      "loss": 2.6152,
      "theoretical_loss": 3.642134099889483,
      "tokens_seen": 1020592128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044944912048420653,
      "loss": 2.7691,
      "theoretical_loss": 3.642111927424228,
      "tokens_seen": 1020657664
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044943257045583506,
      "loss": 2.9441,
      "theoretical_loss": 3.642089756781214,
      "tokens_seen": 1020723200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004494160204274636,
      "loss": 2.6065,
      "theoretical_loss": 3.642067587960175,
      "tokens_seen": 1020788736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004493994703990921,
      "loss": 2.8207,
      "theoretical_loss": 3.6420454209608444,
      "tokens_seen": 1020854272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004493829203707206,
      "loss": 2.9091,
      "theoretical_loss": 3.6420232557829553,
      "tokens_seen": 1020919808
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044936637034234916,
      "loss": 2.7715,
      "theoretical_loss": 3.6420010924262405,
      "tokens_seen": 1020985344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004493498203139777,
      "loss": 2.7556,
      "theoretical_loss": 3.6419789308904345,
      "tokens_seen": 1021050880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004493332702856062,
      "loss": 2.8436,
      "theoretical_loss": 3.6419567711752703,
      "tokens_seen": 1021116416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004493167202572347,
      "loss": 3.0042,
      "theoretical_loss": 3.6419346132804815,
      "tokens_seen": 1021181952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044930017022886326,
      "loss": 3.0314,
      "theoretical_loss": 3.641912457205802,
      "tokens_seen": 1021247488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004492836202004918,
      "loss": 2.9375,
      "theoretical_loss": 3.6418903029509653,
      "tokens_seen": 1021313024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004492670701721203,
      "loss": 2.8527,
      "theoretical_loss": 3.6418681505157053,
      "tokens_seen": 1021378560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004492505201437488,
      "loss": 2.7169,
      "theoretical_loss": 3.641845999899755,
      "tokens_seen": 1021444096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044923397011537735,
      "loss": 2.8772,
      "theoretical_loss": 3.64182385110285,
      "tokens_seen": 1021509632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004492174200870059,
      "loss": 2.7897,
      "theoretical_loss": 3.6418017041247226,
      "tokens_seen": 1021575168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004492008700586344,
      "loss": 2.9484,
      "theoretical_loss": 3.6417795589651076,
      "tokens_seen": 1021640704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0004491843200302629,
      "loss": 2.5928,
      "theoretical_loss": 3.6417574156237387,
      "tokens_seen": 1021706240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044916777000189145,
      "loss": 3.0494,
      "theoretical_loss": 3.64173527410035,
      "tokens_seen": 1021771776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044915121997352,
      "loss": 2.5059,
      "theoretical_loss": 3.641713134394676,
      "tokens_seen": 1021837312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00044913466994514847,
      "loss": 2.9552,
      "theoretical_loss": 3.6416909965064512,
      "tokens_seen": 1021902848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000449118119916777,
      "loss": 2.7558,
      "theoretical_loss": 3.6416688604354084,
      "tokens_seen": 1021968384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044910156988840554,
      "loss": 2.7989,
      "theoretical_loss": 3.6416467261812837,
      "tokens_seen": 1022033920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004490850198600341,
      "loss": 2.7799,
      "theoretical_loss": 3.64162459374381,
      "tokens_seen": 1022099456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044906846983166257,
      "loss": 2.8175,
      "theoretical_loss": 3.641602463122723,
      "tokens_seen": 1022164992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004490519198032911,
      "loss": 2.8611,
      "theoretical_loss": 3.641580334317756,
      "tokens_seen": 1022230528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044903536977491964,
      "loss": 2.9001,
      "theoretical_loss": 3.6415582073286448,
      "tokens_seen": 1022296064
    },
    {
      "epoch": 0.37,
      "objective/train/avg_token_score": 0.01443505473434925,
      "objective/train/avg_weight": 0.9830074310302734,
      "objective/train/docs_used": 582721,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9416632652282715,
      "objective/train/std_weight": 0.0846373662352562,
      "objective/train/theoretical_loss": 3.6415360821551226,
      "objective/train/tokens_used": 1042821600,
      "theoretical_loss": 3.6415360821551226,
      "tokens_seen": 1022361600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004490188197465482,
      "loss": 2.7765,
      "theoretical_loss": 3.6415360821551226,
      "tokens_seen": 1022361600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044900226971817666,
      "loss": 2.87,
      "theoretical_loss": 3.641513958796925,
      "tokens_seen": 1022427136
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004489857196898052,
      "loss": 2.7846,
      "theoretical_loss": 3.6414918372537866,
      "tokens_seen": 1022492672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044896916966143374,
      "loss": 2.7372,
      "theoretical_loss": 3.6414697175254425,
      "tokens_seen": 1022558208
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004489526196330623,
      "loss": 2.977,
      "theoretical_loss": 3.6414475996116265,
      "tokens_seen": 1022623744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044893606960469076,
      "loss": 2.7975,
      "theoretical_loss": 3.641425483512074,
      "tokens_seen": 1022689280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004489195195763193,
      "loss": 2.7713,
      "theoretical_loss": 3.6414033692265204,
      "tokens_seen": 1022754816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044890296954794783,
      "loss": 2.9164,
      "theoretical_loss": 3.6413812567547,
      "tokens_seen": 1022820352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044888641951957637,
      "loss": 2.862,
      "theoretical_loss": 3.641359146096348,
      "tokens_seen": 1022885888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044886986949120485,
      "loss": 2.8052,
      "theoretical_loss": 3.6413370372512,
      "tokens_seen": 1022951424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004488533194628334,
      "loss": 2.9991,
      "theoretical_loss": 3.6413149302189907,
      "tokens_seen": 1023016960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044883676943446193,
      "loss": 2.4842,
      "theoretical_loss": 3.6412928249994554,
      "tokens_seen": 1023082496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044882021940609047,
      "loss": 2.8885,
      "theoretical_loss": 3.6412707215923295,
      "tokens_seen": 1023148032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004488036693777189,
      "loss": 2.6213,
      "theoretical_loss": 3.6412486199973477,
      "tokens_seen": 1023213568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044878711934934743,
      "loss": 2.8346,
      "theoretical_loss": 3.641226520214246,
      "tokens_seen": 1023279104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004487705693209759,
      "loss": 2.6958,
      "theoretical_loss": 3.64120442224276,
      "tokens_seen": 1023344640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044875401929260445,
      "loss": 2.755,
      "theoretical_loss": 3.6411823260826246,
      "tokens_seen": 1023410176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000448737469264233,
      "loss": 2.5468,
      "theoretical_loss": 3.6411602317335756,
      "tokens_seen": 1023475712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044872091923586153,
      "loss": 2.8147,
      "theoretical_loss": 3.6411381391953492,
      "tokens_seen": 1023541248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044870436920749,
      "loss": 2.4536,
      "theoretical_loss": 3.64111604846768,
      "tokens_seen": 1023606784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044868781917911855,
      "loss": 2.8195,
      "theoretical_loss": 3.6410939595503047,
      "tokens_seen": 1023672320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004486712691507471,
      "loss": 2.631,
      "theoretical_loss": 3.641071872442958,
      "tokens_seen": 1023737856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004486547191223756,
      "loss": 2.9765,
      "theoretical_loss": 3.641049787145376,
      "tokens_seen": 1023803392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004486381690940041,
      "loss": 2.9324,
      "theoretical_loss": 3.641027703657296,
      "tokens_seen": 1023868928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044862161906563265,
      "loss": 2.5901,
      "theoretical_loss": 3.6410056219784526,
      "tokens_seen": 1023934464
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004486050690372612,
      "loss": 2.9526,
      "theoretical_loss": 3.6409835421085814,
      "tokens_seen": 1024000000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004485885190088897,
      "loss": 2.7771,
      "theoretical_loss": 3.640961464047419,
      "tokens_seen": 1024065536
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004485719689805182,
      "loss": 2.5215,
      "theoretical_loss": 3.640939387794702,
      "tokens_seen": 1024131072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044855541895214674,
      "loss": 2.5827,
      "theoretical_loss": 3.640917313350166,
      "tokens_seen": 1024196608
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004485388689237753,
      "loss": 2.6378,
      "theoretical_loss": 3.6408952407135473,
      "tokens_seen": 1024262144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004485223188954038,
      "loss": 2.837,
      "theoretical_loss": 3.640873169884582,
      "tokens_seen": 1024327680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004485057688670323,
      "loss": 2.9502,
      "theoretical_loss": 3.6408511008630073,
      "tokens_seen": 1024393216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044848921883866084,
      "loss": 2.6713,
      "theoretical_loss": 3.6408290336485583,
      "tokens_seen": 1024458752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004484726688102894,
      "loss": 2.7516,
      "theoretical_loss": 3.6408069682409723,
      "tokens_seen": 1024524288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004484561187819179,
      "loss": 2.6242,
      "theoretical_loss": 3.640784904639985,
      "tokens_seen": 1024589824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004484395687535464,
      "loss": 2.5317,
      "theoretical_loss": 3.640762842845334,
      "tokens_seen": 1024655360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044842301872517493,
      "loss": 2.6414,
      "theoretical_loss": 3.6407407828567555,
      "tokens_seen": 1024720896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044840646869680347,
      "loss": 2.6058,
      "theoretical_loss": 3.6407187246739854,
      "tokens_seen": 1024786432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000448389918668432,
      "loss": 3.0199,
      "theoretical_loss": 3.640696668296761,
      "tokens_seen": 1024851968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004483733686400605,
      "loss": 2.6234,
      "theoretical_loss": 3.6406746137248196,
      "tokens_seen": 1024917504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044835681861168903,
      "loss": 2.6957,
      "theoretical_loss": 3.6406525609578972,
      "tokens_seen": 1024983040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044834026858331757,
      "loss": 2.5787,
      "theoretical_loss": 3.6406305099957312,
      "tokens_seen": 1025048576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004483237185549461,
      "loss": 2.872,
      "theoretical_loss": 3.6406084608380582,
      "tokens_seen": 1025114112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004483071685265746,
      "loss": 2.7094,
      "theoretical_loss": 3.6405864134846153,
      "tokens_seen": 1025179648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004482906184982031,
      "loss": 2.8348,
      "theoretical_loss": 3.6405643679351396,
      "tokens_seen": 1025245184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044827406846983166,
      "loss": 2.7951,
      "theoretical_loss": 3.6405423241893677,
      "tokens_seen": 1025310720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004482575184414602,
      "loss": 2.6056,
      "theoretical_loss": 3.6405202822470377,
      "tokens_seen": 1025376256
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004482409684130887,
      "loss": 2.8685,
      "theoretical_loss": 3.6404982421078858,
      "tokens_seen": 1025441792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004482244183847172,
      "loss": 2.8909,
      "theoretical_loss": 3.6404762037716503,
      "tokens_seen": 1025507328
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044820786835634576,
      "loss": 2.895,
      "theoretical_loss": 3.640454167238068,
      "tokens_seen": 1025572864
    },
    {
      "epoch": 0.37,
      "objective/train/avg_token_score": 0.02644398808479309,
      "objective/train/avg_weight": 0.9688830375671387,
      "objective/train/docs_used": 584025,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.020437717437744,
      "objective/train/std_weight": 0.13429781794548035,
      "objective/train/theoretical_loss": 3.6404321325068754,
      "objective/train/tokens_used": 1046098400,
      "theoretical_loss": 3.6404321325068754,
      "tokens_seen": 1025638400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004481913183279743,
      "loss": 2.9217,
      "theoretical_loss": 3.6404321325068754,
      "tokens_seen": 1025638400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004481747682996028,
      "loss": 2.9743,
      "theoretical_loss": 3.6404100995778115,
      "tokens_seen": 1025703936
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004481582182712313,
      "loss": 2.8286,
      "theoretical_loss": 3.640388068450613,
      "tokens_seen": 1025769472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044814166824285986,
      "loss": 2.6948,
      "theoretical_loss": 3.640366039125017,
      "tokens_seen": 1025835008
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004481251182144884,
      "loss": 2.6281,
      "theoretical_loss": 3.640344011600762,
      "tokens_seen": 1025900544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004481085681861169,
      "loss": 2.4894,
      "theoretical_loss": 3.640321985877585,
      "tokens_seen": 1025966080
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004480920181577454,
      "loss": 2.7657,
      "theoretical_loss": 3.6402999619552245,
      "tokens_seen": 1026031616
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044807546812937395,
      "loss": 2.9771,
      "theoretical_loss": 3.640277939833417,
      "tokens_seen": 1026097152
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004480589181010025,
      "loss": 2.6363,
      "theoretical_loss": 3.640255919511901,
      "tokens_seen": 1026162688
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044804236807263097,
      "loss": 2.8131,
      "theoretical_loss": 3.640233900990415,
      "tokens_seen": 1026228224
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004480258180442595,
      "loss": 2.549,
      "theoretical_loss": 3.6402118842686955,
      "tokens_seen": 1026293760
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044800926801588805,
      "loss": 2.644,
      "theoretical_loss": 3.6401898693464814,
      "tokens_seen": 1026359296
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004479927179875166,
      "loss": 2.7244,
      "theoretical_loss": 3.6401678562235107,
      "tokens_seen": 1026424832
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004479761679591451,
      "loss": 2.5725,
      "theoretical_loss": 3.6401458448995214,
      "tokens_seen": 1026490368
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004479596179307736,
      "loss": 2.731,
      "theoretical_loss": 3.6401238353742515,
      "tokens_seen": 1026555904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044794306790240214,
      "loss": 2.5967,
      "theoretical_loss": 3.6401018276474395,
      "tokens_seen": 1026621440
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004479265178740307,
      "loss": 2.8215,
      "theoretical_loss": 3.6400798217188233,
      "tokens_seen": 1026686976
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004479099678456592,
      "loss": 2.7018,
      "theoretical_loss": 3.6400578175881417,
      "tokens_seen": 1026752512
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004478934178172877,
      "loss": 2.6379,
      "theoretical_loss": 3.640035815255132,
      "tokens_seen": 1026818048
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044787686778891624,
      "loss": 2.8473,
      "theoretical_loss": 3.640013814719534,
      "tokens_seen": 1026883584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004478603177605448,
      "loss": 2.7902,
      "theoretical_loss": 3.6399918159810847,
      "tokens_seen": 1026949120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004478437677321733,
      "loss": 2.7789,
      "theoretical_loss": 3.639969819039524,
      "tokens_seen": 1027014656
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004478272177038018,
      "loss": 2.8513,
      "theoretical_loss": 3.63994782389459,
      "tokens_seen": 1027080192
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004478106676754303,
      "loss": 2.8946,
      "theoretical_loss": 3.639925830546021,
      "tokens_seen": 1027145728
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044779411764705876,
      "loss": 2.9093,
      "theoretical_loss": 3.639903838993556,
      "tokens_seen": 1027211264
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004477775676186873,
      "loss": 2.7357,
      "theoretical_loss": 3.6398818492369336,
      "tokens_seen": 1027276800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044776101759031584,
      "loss": 2.9708,
      "theoretical_loss": 3.639859861275893,
      "tokens_seen": 1027342336
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004477444675619444,
      "loss": 2.9552,
      "theoretical_loss": 3.639837875110172,
      "tokens_seen": 1027407872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044772791753357286,
      "loss": 2.6619,
      "theoretical_loss": 3.6398158907395106,
      "tokens_seen": 1027473408
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004477113675052014,
      "loss": 2.7296,
      "theoretical_loss": 3.6397939081636475,
      "tokens_seen": 1027538944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044769481747682994,
      "loss": 2.751,
      "theoretical_loss": 3.6397719273823212,
      "tokens_seen": 1027604480
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044767826744845847,
      "loss": 2.7749,
      "theoretical_loss": 3.6397499483952718,
      "tokens_seen": 1027670016
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044766171742008696,
      "loss": 2.9026,
      "theoretical_loss": 3.639727971202237,
      "tokens_seen": 1027735552
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004476451673917155,
      "loss": 2.838,
      "theoretical_loss": 3.6397059958029576,
      "tokens_seen": 1027801088
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044762861736334403,
      "loss": 2.8722,
      "theoretical_loss": 3.639684022197171,
      "tokens_seen": 1027866624
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044761206733497257,
      "loss": 2.7181,
      "theoretical_loss": 3.639662050384618,
      "tokens_seen": 1027932160
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044759551730660105,
      "loss": 2.9556,
      "theoretical_loss": 3.6396400803650377,
      "tokens_seen": 1027997696
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004475789672782296,
      "loss": 2.9613,
      "theoretical_loss": 3.639618112138169,
      "tokens_seen": 1028063232
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004475624172498581,
      "loss": 2.6004,
      "theoretical_loss": 3.639596145703752,
      "tokens_seen": 1028128768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044754586722148666,
      "loss": 2.8699,
      "theoretical_loss": 3.6395741810615254,
      "tokens_seen": 1028194304
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044752931719311515,
      "loss": 2.5597,
      "theoretical_loss": 3.6395522182112288,
      "tokens_seen": 1028259840
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004475127671647437,
      "loss": 2.7344,
      "theoretical_loss": 3.6395302571526025,
      "tokens_seen": 1028325376
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004474962171363722,
      "loss": 3.0201,
      "theoretical_loss": 3.639508297885386,
      "tokens_seen": 1028390912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044747966710800076,
      "loss": 2.7336,
      "theoretical_loss": 3.6394863404093183,
      "tokens_seen": 1028456448
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044746311707962924,
      "loss": 2.8151,
      "theoretical_loss": 3.6394643847241404,
      "tokens_seen": 1028521984
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004474465670512578,
      "loss": 2.6774,
      "theoretical_loss": 3.6394424308295914,
      "tokens_seen": 1028587520
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004474300170228863,
      "loss": 2.7631,
      "theoretical_loss": 3.6394204787254107,
      "tokens_seen": 1028653056
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044741346699451486,
      "loss": 2.8197,
      "theoretical_loss": 3.6393985284113395,
      "tokens_seen": 1028718592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044739691696614334,
      "loss": 2.8594,
      "theoretical_loss": 3.6393765798871165,
      "tokens_seen": 1028784128
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004473803669377719,
      "loss": 2.9361,
      "theoretical_loss": 3.6393546331524824,
      "tokens_seen": 1028849664
    },
    {
      "epoch": 0.37,
      "objective/train/avg_token_score": 0.014482278376817703,
      "objective/train/avg_weight": 0.9829506874084473,
      "objective/train/docs_used": 586068,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.879977226257324,
      "objective/train/std_weight": 0.08211760222911835,
      "objective/train/theoretical_loss": 3.639332688207178,
      "objective/train/tokens_used": 1049375200,
      "theoretical_loss": 3.639332688207178,
      "tokens_seen": 1028915200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004473638169094004,
      "loss": 2.8984,
      "theoretical_loss": 3.639332688207178,
      "tokens_seen": 1028915200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044734726688102895,
      "loss": 2.6683,
      "theoretical_loss": 3.639310745050942,
      "tokens_seen": 1028980736
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044733071685265744,
      "loss": 2.5985,
      "theoretical_loss": 3.639288803683515,
      "tokens_seen": 1029046272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000447314166824286,
      "loss": 3.013,
      "theoretical_loss": 3.639266864104638,
      "tokens_seen": 1029111808
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004472976167959145,
      "loss": 2.7277,
      "theoretical_loss": 3.639244926314051,
      "tokens_seen": 1029177344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044728106676754305,
      "loss": 2.5503,
      "theoretical_loss": 3.6392229903114943,
      "tokens_seen": 1029242880
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044726451673917153,
      "loss": 2.7173,
      "theoretical_loss": 3.6392010560967085,
      "tokens_seen": 1029308416
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044724796671080007,
      "loss": 2.6778,
      "theoretical_loss": 3.6391791236694333,
      "tokens_seen": 1029373952
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004472314166824286,
      "loss": 2.6528,
      "theoretical_loss": 3.63915719302941,
      "tokens_seen": 1029439488
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044721486665405714,
      "loss": 2.8245,
      "theoretical_loss": 3.6391352641763794,
      "tokens_seen": 1029505024
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044719831662568563,
      "loss": 3.0813,
      "theoretical_loss": 3.6391133371100817,
      "tokens_seen": 1029570560
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044718176659731417,
      "loss": 2.8937,
      "theoretical_loss": 3.6390914118302575,
      "tokens_seen": 1029636096
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004471652165689427,
      "loss": 2.8663,
      "theoretical_loss": 3.639069488336648,
      "tokens_seen": 1029701632
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044714866654057124,
      "loss": 2.68,
      "theoretical_loss": 3.6390475666289936,
      "tokens_seen": 1029767168
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004471321165121997,
      "loss": 2.928,
      "theoretical_loss": 3.639025646707035,
      "tokens_seen": 1029832704
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044711556648382826,
      "loss": 2.7939,
      "theoretical_loss": 3.639003728570514,
      "tokens_seen": 1029898240
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004470990164554568,
      "loss": 2.6712,
      "theoretical_loss": 3.638981812219171,
      "tokens_seen": 1029963776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044708246642708534,
      "loss": 2.9153,
      "theoretical_loss": 3.6389598976527466,
      "tokens_seen": 1030029312
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004470659163987138,
      "loss": 2.7458,
      "theoretical_loss": 3.6389379848709824,
      "tokens_seen": 1030094848
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044704936637034236,
      "loss": 2.4944,
      "theoretical_loss": 3.6389160738736193,
      "tokens_seen": 1030160384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004470328163419709,
      "loss": 2.673,
      "theoretical_loss": 3.638894164660399,
      "tokens_seen": 1030225920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044701626631359943,
      "loss": 2.6532,
      "theoretical_loss": 3.6388722572310623,
      "tokens_seen": 1030291456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004469997162852279,
      "loss": 2.7545,
      "theoretical_loss": 3.6388503515853503,
      "tokens_seen": 1030356992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044698316625685645,
      "loss": 2.7611,
      "theoretical_loss": 3.6388284477230046,
      "tokens_seen": 1030422528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000446966616228485,
      "loss": 2.89,
      "theoretical_loss": 3.6388065456437673,
      "tokens_seen": 1030488064
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044695006620011353,
      "loss": 2.654,
      "theoretical_loss": 3.6387846453473784,
      "tokens_seen": 1030553600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000446933516171742,
      "loss": 2.6306,
      "theoretical_loss": 3.6387627468335806,
      "tokens_seen": 1030619136
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044691696614337055,
      "loss": 2.7632,
      "theoretical_loss": 3.638740850102115,
      "tokens_seen": 1030684672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004469004161149991,
      "loss": 2.7262,
      "theoretical_loss": 3.6387189551527226,
      "tokens_seen": 1030750208
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004468838660866276,
      "loss": 2.7519,
      "theoretical_loss": 3.6386970619851464,
      "tokens_seen": 1030815744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004468673160582561,
      "loss": 2.9333,
      "theoretical_loss": 3.6386751705991274,
      "tokens_seen": 1030881280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044685076602988465,
      "loss": 2.9294,
      "theoretical_loss": 3.6386532809944074,
      "tokens_seen": 1030946816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044683421600151313,
      "loss": 2.4517,
      "theoretical_loss": 3.638631393170728,
      "tokens_seen": 1031012352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004468176659731416,
      "loss": 2.709,
      "theoretical_loss": 3.638609507127832,
      "tokens_seen": 1031077888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044680111594477015,
      "loss": 2.6348,
      "theoretical_loss": 3.6385876228654603,
      "tokens_seen": 1031143424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004467845659163987,
      "loss": 2.9613,
      "theoretical_loss": 3.6385657403833553,
      "tokens_seen": 1031208960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004467680158880272,
      "loss": 2.8589,
      "theoretical_loss": 3.6385438596812594,
      "tokens_seen": 1031274496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004467514658596557,
      "loss": 2.6463,
      "theoretical_loss": 3.638521980758914,
      "tokens_seen": 1031340032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044673491583128425,
      "loss": 2.855,
      "theoretical_loss": 3.638500103616062,
      "tokens_seen": 1031405568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004467183658029128,
      "loss": 2.6961,
      "theoretical_loss": 3.638478228252445,
      "tokens_seen": 1031471104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004467018157745413,
      "loss": 2.5717,
      "theoretical_loss": 3.6384563546678055,
      "tokens_seen": 1031536640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004466852657461698,
      "loss": 2.8536,
      "theoretical_loss": 3.6384344828618858,
      "tokens_seen": 1031602176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044666871571779834,
      "loss": 2.7068,
      "theoretical_loss": 3.6384126128344283,
      "tokens_seen": 1031667712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004466521656894269,
      "loss": 2.6217,
      "theoretical_loss": 3.6383907445851755,
      "tokens_seen": 1031733248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004466356156610554,
      "loss": 2.8262,
      "theoretical_loss": 3.6383688781138694,
      "tokens_seen": 1031798784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004466190656326839,
      "loss": 2.8457,
      "theoretical_loss": 3.638347013420254,
      "tokens_seen": 1031864320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044660251560431244,
      "loss": 2.7325,
      "theoretical_loss": 3.6383251505040697,
      "tokens_seen": 1031929856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000446585965575941,
      "loss": 2.8402,
      "theoretical_loss": 3.6383032893650604,
      "tokens_seen": 1031995392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004465694155475695,
      "loss": 2.9033,
      "theoretical_loss": 3.6382814300029693,
      "tokens_seen": 1032060928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000446552865519198,
      "loss": 2.9315,
      "theoretical_loss": 3.638259572417538,
      "tokens_seen": 1032126464
    },
    {
      "epoch": 0.37,
      "objective/train/avg_token_score": 0.01551381777971983,
      "objective/train/avg_weight": 0.9817401766777039,
      "objective/train/docs_used": 588066,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0801987648010254,
      "objective/train/std_weight": 0.0833304226398468,
      "objective/train/theoretical_loss": 3.6382377166085096,
      "objective/train/tokens_used": 1052652000,
      "theoretical_loss": 3.6382377166085096,
      "tokens_seen": 1032192000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044653631549082653,
      "loss": 2.8589,
      "theoretical_loss": 3.6382377166085096,
      "tokens_seen": 1032192000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044651976546245507,
      "loss": 2.8628,
      "theoretical_loss": 3.6382158625756276,
      "tokens_seen": 1032257536
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004465032154340836,
      "loss": 2.6548,
      "theoretical_loss": 3.6381940103186343,
      "tokens_seen": 1032323072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004464866654057121,
      "loss": 2.6356,
      "theoretical_loss": 3.638172159837273,
      "tokens_seen": 1032388608
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044647011537734063,
      "loss": 2.6387,
      "theoretical_loss": 3.6381503111312865,
      "tokens_seen": 1032454144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044645356534896917,
      "loss": 2.6975,
      "theoretical_loss": 3.6381284642004177,
      "tokens_seen": 1032519680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004464370153205977,
      "loss": 2.6076,
      "theoretical_loss": 3.63810661904441,
      "tokens_seen": 1032585216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004464204652922262,
      "loss": 2.7857,
      "theoretical_loss": 3.638084775663007,
      "tokens_seen": 1032650752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004464039152638547,
      "loss": 2.7023,
      "theoretical_loss": 3.6380629340559505,
      "tokens_seen": 1032716288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044638736523548326,
      "loss": 2.7896,
      "theoretical_loss": 3.6380410942229853,
      "tokens_seen": 1032781824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004463708152071118,
      "loss": 2.808,
      "theoretical_loss": 3.6380192561638545,
      "tokens_seen": 1032847360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004463542651787403,
      "loss": 2.6148,
      "theoretical_loss": 3.6379974198783005,
      "tokens_seen": 1032912896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004463377151503688,
      "loss": 2.8995,
      "theoretical_loss": 3.6379755853660676,
      "tokens_seen": 1032978432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044632116512199736,
      "loss": 2.7435,
      "theoretical_loss": 3.637953752626899,
      "tokens_seen": 1033043968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004463046150936259,
      "loss": 2.6706,
      "theoretical_loss": 3.6379319216605386,
      "tokens_seen": 1033109504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004462880650652544,
      "loss": 2.5675,
      "theoretical_loss": 3.6379100924667296,
      "tokens_seen": 1033175040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004462715150368829,
      "loss": 2.5694,
      "theoretical_loss": 3.637888265045216,
      "tokens_seen": 1033240576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044625496500851145,
      "loss": 2.7498,
      "theoretical_loss": 3.637866439395741,
      "tokens_seen": 1033306112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044623841498014,
      "loss": 2.8033,
      "theoretical_loss": 3.6378446155180484,
      "tokens_seen": 1033371648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004462218649517685,
      "loss": 2.7814,
      "theoretical_loss": 3.6378227934118827,
      "tokens_seen": 1033437184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000446205314923397,
      "loss": 2.6923,
      "theoretical_loss": 3.6378009730769874,
      "tokens_seen": 1033502720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044618876489502555,
      "loss": 2.9575,
      "theoretical_loss": 3.6377791545131064,
      "tokens_seen": 1033568256
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004461722148666541,
      "loss": 2.5693,
      "theoretical_loss": 3.6377573377199837,
      "tokens_seen": 1033633792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044615566483828257,
      "loss": 2.7478,
      "theoretical_loss": 3.6377355226973633,
      "tokens_seen": 1033699328
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004461391148099111,
      "loss": 2.6223,
      "theoretical_loss": 3.6377137094449887,
      "tokens_seen": 1033764864
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044612256478153965,
      "loss": 2.812,
      "theoretical_loss": 3.6376918979626054,
      "tokens_seen": 1033830400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004461060147531682,
      "loss": 2.792,
      "theoretical_loss": 3.6376700882499566,
      "tokens_seen": 1033895936
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044608946472479667,
      "loss": 2.6544,
      "theoretical_loss": 3.637648280306786,
      "tokens_seen": 1033961472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004460729146964252,
      "loss": 2.8496,
      "theoretical_loss": 3.6376264741328397,
      "tokens_seen": 1034027008
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044605636466805374,
      "loss": 2.7071,
      "theoretical_loss": 3.6376046697278603,
      "tokens_seen": 1034092544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004460398146396823,
      "loss": 2.8375,
      "theoretical_loss": 3.6375828670915933,
      "tokens_seen": 1034158080
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044602326461131076,
      "loss": 2.6973,
      "theoretical_loss": 3.6375610662237827,
      "tokens_seen": 1034223616
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004460067145829393,
      "loss": 2.57,
      "theoretical_loss": 3.637539267124173,
      "tokens_seen": 1034289152
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044599016455456784,
      "loss": 2.6826,
      "theoretical_loss": 3.637517469792508,
      "tokens_seen": 1034354688
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004459736145261964,
      "loss": 2.8938,
      "theoretical_loss": 3.6374956742285343,
      "tokens_seen": 1034420224
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044595706449782486,
      "loss": 2.8176,
      "theoretical_loss": 3.637473880431995,
      "tokens_seen": 1034485760
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004459405144694534,
      "loss": 2.5236,
      "theoretical_loss": 3.6374520884026347,
      "tokens_seen": 1034551296
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044592396444108193,
      "loss": 2.8624,
      "theoretical_loss": 3.637430298140199,
      "tokens_seen": 1034616832
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044590741441271047,
      "loss": 2.6862,
      "theoretical_loss": 3.6374085096444326,
      "tokens_seen": 1034682368
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044589086438433896,
      "loss": 2.5154,
      "theoretical_loss": 3.63738672291508,
      "tokens_seen": 1034747904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004458743143559675,
      "loss": 2.7645,
      "theoretical_loss": 3.6373649379518866,
      "tokens_seen": 1034813440
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044585776432759603,
      "loss": 2.9136,
      "theoretical_loss": 3.637343154754597,
      "tokens_seen": 1034878976
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044584121429922446,
      "loss": 2.7771,
      "theoretical_loss": 3.637321373322956,
      "tokens_seen": 1034944512
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000445824664270853,
      "loss": 2.6661,
      "theoretical_loss": 3.6372995936567096,
      "tokens_seen": 1035010048
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044580811424248153,
      "loss": 2.8357,
      "theoretical_loss": 3.637277815755602,
      "tokens_seen": 1035075584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044579156421411,
      "loss": 2.6482,
      "theoretical_loss": 3.6372560396193787,
      "tokens_seen": 1035141120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044577501418573856,
      "loss": 2.6466,
      "theoretical_loss": 3.6372342652477854,
      "tokens_seen": 1035206656
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004457584641573671,
      "loss": 2.7117,
      "theoretical_loss": 3.637212492640567,
      "tokens_seen": 1035272192
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044574191412899563,
      "loss": 2.6499,
      "theoretical_loss": 3.6371907217974684,
      "tokens_seen": 1035337728
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004457253641006241,
      "loss": 2.9556,
      "theoretical_loss": 3.637168952718236,
      "tokens_seen": 1035403264
    },
    {
      "epoch": 0.37,
      "objective/train/avg_token_score": 0.018012594431638718,
      "objective/train/avg_weight": 0.9788045883178711,
      "objective/train/docs_used": 589983,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8577940464019775,
      "objective/train/std_weight": 0.06897003948688507,
      "objective/train/theoretical_loss": 3.6371471854026147,
      "objective/train/tokens_used": 1055928800,
      "theoretical_loss": 3.6371471854026147,
      "tokens_seen": 1035468800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044570881407225265,
      "loss": 2.7166,
      "theoretical_loss": 3.6371471854026147,
      "tokens_seen": 1035468800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004456922640438812,
      "loss": 2.8138,
      "theoretical_loss": 3.63712541985035,
      "tokens_seen": 1035534336
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004456757140155097,
      "loss": 2.6021,
      "theoretical_loss": 3.6371036560611874,
      "tokens_seen": 1035599872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004456591639871382,
      "loss": 2.6733,
      "theoretical_loss": 3.637081894034873,
      "tokens_seen": 1035665408
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044564261395876675,
      "loss": 2.5653,
      "theoretical_loss": 3.6370601337711514,
      "tokens_seen": 1035730944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004456260639303953,
      "loss": 2.6669,
      "theoretical_loss": 3.63703837526977,
      "tokens_seen": 1035796480
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004456095139020238,
      "loss": 2.7525,
      "theoretical_loss": 3.637016618530473,
      "tokens_seen": 1035862016
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004455929638736523,
      "loss": 3.0047,
      "theoretical_loss": 3.6369948635530074,
      "tokens_seen": 1035927552
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044557641384528084,
      "loss": 2.7848,
      "theoretical_loss": 3.6369731103371183,
      "tokens_seen": 1035993088
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004455598638169094,
      "loss": 2.7703,
      "theoretical_loss": 3.636951358882552,
      "tokens_seen": 1036058624
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004455433137885379,
      "loss": 2.6356,
      "theoretical_loss": 3.6369296091890537,
      "tokens_seen": 1036124160
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004455267637601664,
      "loss": 2.4476,
      "theoretical_loss": 3.636907861256371,
      "tokens_seen": 1036189696
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044551021373179494,
      "loss": 2.8281,
      "theoretical_loss": 3.6368861150842484,
      "tokens_seen": 1036255232
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004454936637034235,
      "loss": 2.9028,
      "theoretical_loss": 3.6368643706724333,
      "tokens_seen": 1036320768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000445477113675052,
      "loss": 2.6625,
      "theoretical_loss": 3.636842628020671,
      "tokens_seen": 1036386304
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044546056364668055,
      "loss": 2.7241,
      "theoretical_loss": 3.6368208871287084,
      "tokens_seen": 1036451840
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044544401361830904,
      "loss": 2.5938,
      "theoretical_loss": 3.6367991479962916,
      "tokens_seen": 1036517376
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044542746358993757,
      "loss": 2.6647,
      "theoretical_loss": 3.6367774106231665,
      "tokens_seen": 1036582912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004454109135615661,
      "loss": 2.6392,
      "theoretical_loss": 3.6367556750090797,
      "tokens_seen": 1036648448
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044539436353319465,
      "loss": 2.8433,
      "theoretical_loss": 3.6367339411537785,
      "tokens_seen": 1036713984
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044537781350482313,
      "loss": 2.8544,
      "theoretical_loss": 3.6367122090570088,
      "tokens_seen": 1036779520
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044536126347645167,
      "loss": 2.7903,
      "theoretical_loss": 3.6366904787185166,
      "tokens_seen": 1036845056
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004453447134480802,
      "loss": 2.6013,
      "theoretical_loss": 3.6366687501380492,
      "tokens_seen": 1036910592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044532816341970874,
      "loss": 2.8143,
      "theoretical_loss": 3.636647023315353,
      "tokens_seen": 1036976128
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044531161339133723,
      "loss": 2.7322,
      "theoretical_loss": 3.6366252982501757,
      "tokens_seen": 1037041664
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044529506336296576,
      "loss": 2.9581,
      "theoretical_loss": 3.6366035749422627,
      "tokens_seen": 1037107200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004452785133345943,
      "loss": 2.5406,
      "theoretical_loss": 3.6365818533913608,
      "tokens_seen": 1037172736
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044526196330622284,
      "loss": 2.7364,
      "theoretical_loss": 3.6365601335972184,
      "tokens_seen": 1037238272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004452454132778513,
      "loss": 2.6102,
      "theoretical_loss": 3.636538415559581,
      "tokens_seen": 1037303808
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044522886324947986,
      "loss": 2.7125,
      "theoretical_loss": 3.636516699278196,
      "tokens_seen": 1037369344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004452123132211084,
      "loss": 2.8183,
      "theoretical_loss": 3.6364949847528107,
      "tokens_seen": 1037434880
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044519576319273694,
      "loss": 2.666,
      "theoretical_loss": 3.636473271983172,
      "tokens_seen": 1037500416
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004451792131643654,
      "loss": 2.66,
      "theoretical_loss": 3.636451560969027,
      "tokens_seen": 1037565952
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044516266313599396,
      "loss": 2.8007,
      "theoretical_loss": 3.636429851710123,
      "tokens_seen": 1037631488
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004451461131076225,
      "loss": 2.5728,
      "theoretical_loss": 3.6364081442062073,
      "tokens_seen": 1037697024
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044512956307925103,
      "loss": 2.7252,
      "theoretical_loss": 3.636386438457027,
      "tokens_seen": 1037762560
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004451130130508795,
      "loss": 2.8122,
      "theoretical_loss": 3.6363647344623296,
      "tokens_seen": 1037828096
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044509646302250805,
      "loss": 2.9146,
      "theoretical_loss": 3.636343032221862,
      "tokens_seen": 1037893632
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004450799129941366,
      "loss": 2.6261,
      "theoretical_loss": 3.6363213317353726,
      "tokens_seen": 1037959168
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044506336296576513,
      "loss": 2.6218,
      "theoretical_loss": 3.6362996330026087,
      "tokens_seen": 1038024704
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004450468129373936,
      "loss": 2.7737,
      "theoretical_loss": 3.6362779360233173,
      "tokens_seen": 1038090240
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044503026290902215,
      "loss": 2.5698,
      "theoretical_loss": 3.636256240797246,
      "tokens_seen": 1038155776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004450137128806507,
      "loss": 2.7989,
      "theoretical_loss": 3.6362345473241433,
      "tokens_seen": 1038221312
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004449971628522792,
      "loss": 2.5992,
      "theoretical_loss": 3.636212855603756,
      "tokens_seen": 1038286848
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004449806128239077,
      "loss": 2.8224,
      "theoretical_loss": 3.6361911656358323,
      "tokens_seen": 1038352384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044496406279553624,
      "loss": 2.5413,
      "theoretical_loss": 3.63616947742012,
      "tokens_seen": 1038417920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004449475127671648,
      "loss": 2.8652,
      "theoretical_loss": 3.6361477909563673,
      "tokens_seen": 1038483456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004449309627387933,
      "loss": 3.0363,
      "theoretical_loss": 3.6361261062443218,
      "tokens_seen": 1038548992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004449144127104218,
      "loss": 2.6195,
      "theoretical_loss": 3.636104423283731,
      "tokens_seen": 1038614528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044489786268205034,
      "loss": 2.6447,
      "theoretical_loss": 3.6360827420743442,
      "tokens_seen": 1038680064
    },
    {
      "epoch": 0.37,
      "objective/train/avg_token_score": 0.01140693947672844,
      "objective/train/avg_weight": 0.9865914583206177,
      "objective/train/docs_used": 592008,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.668220281600952,
      "objective/train/std_weight": 0.06103510037064552,
      "objective/train/theoretical_loss": 3.6360610626159087,
      "objective/train/tokens_used": 1059205600,
      "theoretical_loss": 3.6360610626159087,
      "tokens_seen": 1038745600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004448813126536789,
      "loss": 2.7762,
      "theoretical_loss": 3.6360610626159087,
      "tokens_seen": 1038745600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004448647626253073,
      "loss": 2.7088,
      "theoretical_loss": 3.6360393849081722,
      "tokens_seen": 1038811136
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044484821259693584,
      "loss": 2.7006,
      "theoretical_loss": 3.636017708950883,
      "tokens_seen": 1038876672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004448316625685644,
      "loss": 2.7597,
      "theoretical_loss": 3.6359960347437905,
      "tokens_seen": 1038942208
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044481511254019287,
      "loss": 2.6816,
      "theoretical_loss": 3.6359743622866416,
      "tokens_seen": 1039007744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004447985625118214,
      "loss": 2.6967,
      "theoretical_loss": 3.6359526915791855,
      "tokens_seen": 1039073280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044478201248344994,
      "loss": 2.4661,
      "theoretical_loss": 3.63593102262117,
      "tokens_seen": 1039138816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004447654624550785,
      "loss": 2.682,
      "theoretical_loss": 3.6359093554123447,
      "tokens_seen": 1039204352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044474891242670696,
      "loss": 2.7746,
      "theoretical_loss": 3.6358876899524564,
      "tokens_seen": 1039269888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004447323623983355,
      "loss": 2.6834,
      "theoretical_loss": 3.635866026241255,
      "tokens_seen": 1039335424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044471581236996404,
      "loss": 2.7256,
      "theoretical_loss": 3.6358443642784883,
      "tokens_seen": 1039400960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004446992623415926,
      "loss": 2.5643,
      "theoretical_loss": 3.6358227040639055,
      "tokens_seen": 1039466496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044468271231322106,
      "loss": 2.7071,
      "theoretical_loss": 3.635801045597255,
      "tokens_seen": 1039532032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004446661622848496,
      "loss": 2.6139,
      "theoretical_loss": 3.6357793888782863,
      "tokens_seen": 1039597568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044464961225647813,
      "loss": 2.7092,
      "theoretical_loss": 3.635757733906747,
      "tokens_seen": 1039663104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044463306222810667,
      "loss": 2.8059,
      "theoretical_loss": 3.6357360806823866,
      "tokens_seen": 1039728640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044461651219973515,
      "loss": 2.9983,
      "theoretical_loss": 3.635714429204954,
      "tokens_seen": 1039794176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004445999621713637,
      "loss": 2.7112,
      "theoretical_loss": 3.6356927794741978,
      "tokens_seen": 1039859712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044458341214299223,
      "loss": 2.8501,
      "theoretical_loss": 3.6356711314898678,
      "tokens_seen": 1039925248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044456686211462077,
      "loss": 2.6067,
      "theoretical_loss": 3.6356494852517125,
      "tokens_seen": 1039990784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044455031208624925,
      "loss": 2.8233,
      "theoretical_loss": 3.635627840759481,
      "tokens_seen": 1040056320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004445337620578778,
      "loss": 2.6683,
      "theoretical_loss": 3.6356061980129235,
      "tokens_seen": 1040121856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004445172120295063,
      "loss": 2.6227,
      "theoretical_loss": 3.6355845570117875,
      "tokens_seen": 1040187392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044450066200113486,
      "loss": 2.6778,
      "theoretical_loss": 3.6355629177558235,
      "tokens_seen": 1040252928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044448411197276335,
      "loss": 2.8748,
      "theoretical_loss": 3.63554128024478,
      "tokens_seen": 1040318464
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004444675619443919,
      "loss": 2.5405,
      "theoretical_loss": 3.635519644478408,
      "tokens_seen": 1040384000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004444510119160204,
      "loss": 2.7589,
      "theoretical_loss": 3.6354980104564545,
      "tokens_seen": 1040449536
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044443446188764896,
      "loss": 2.5702,
      "theoretical_loss": 3.635476378178671,
      "tokens_seen": 1040515072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044441791185927744,
      "loss": 2.6446,
      "theoretical_loss": 3.6354547476448067,
      "tokens_seen": 1040580608
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000444401361830906,
      "loss": 2.5296,
      "theoretical_loss": 3.63543311885461,
      "tokens_seen": 1040646144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004443848118025345,
      "loss": 2.4037,
      "theoretical_loss": 3.6354114918078317,
      "tokens_seen": 1040711680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044436826177416305,
      "loss": 2.7014,
      "theoretical_loss": 3.6353898665042212,
      "tokens_seen": 1040777216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044435171174579154,
      "loss": 2.6734,
      "theoretical_loss": 3.6353682429435286,
      "tokens_seen": 1040842752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004443351617174201,
      "loss": 2.739,
      "theoretical_loss": 3.635346621125503,
      "tokens_seen": 1040908288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004443186116890486,
      "loss": 2.8087,
      "theoretical_loss": 3.6353250010498943,
      "tokens_seen": 1040973824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044430206166067715,
      "loss": 2.6592,
      "theoretical_loss": 3.635303382716453,
      "tokens_seen": 1041039360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044428551163230563,
      "loss": 2.6043,
      "theoretical_loss": 3.635281766124929,
      "tokens_seen": 1041104896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044426896160393417,
      "loss": 2.5147,
      "theoretical_loss": 3.6352601512750713,
      "tokens_seen": 1041170432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004442524115755627,
      "loss": 2.9126,
      "theoretical_loss": 3.6352385381666315,
      "tokens_seen": 1041235968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044423586154719125,
      "loss": 2.5348,
      "theoretical_loss": 3.635216926799358,
      "tokens_seen": 1041301504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044421931151881973,
      "loss": 2.7521,
      "theoretical_loss": 3.6351953171730025,
      "tokens_seen": 1041367040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044420276149044827,
      "loss": 2.6833,
      "theoretical_loss": 3.635173709287314,
      "tokens_seen": 1041432576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004441862114620768,
      "loss": 2.5895,
      "theoretical_loss": 3.635152103142044,
      "tokens_seen": 1041498112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044416966143370534,
      "loss": 2.6333,
      "theoretical_loss": 3.6351304987369417,
      "tokens_seen": 1041563648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004441531114053338,
      "loss": 2.7498,
      "theoretical_loss": 3.635108896071759,
      "tokens_seen": 1041629184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044413656137696236,
      "loss": 2.4804,
      "theoretical_loss": 3.635087295146244,
      "tokens_seen": 1041694720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004441200113485909,
      "loss": 2.4628,
      "theoretical_loss": 3.6350656959601486,
      "tokens_seen": 1041760256
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044410346132021944,
      "loss": 2.7502,
      "theoretical_loss": 3.635044098513223,
      "tokens_seen": 1041825792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004440869112918479,
      "loss": 2.9939,
      "theoretical_loss": 3.6350225028052185,
      "tokens_seen": 1041891328
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044407036126347646,
      "loss": 2.8038,
      "theoretical_loss": 3.6350009088358846,
      "tokens_seen": 1041956864
    },
    {
      "epoch": 0.37,
      "objective/train/avg_token_score": 0.020296555012464523,
      "objective/train/avg_weight": 0.9761171936988831,
      "objective/train/docs_used": 593661,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.689988136291504,
      "objective/train/std_weight": 0.07269823551177979,
      "objective/train/theoretical_loss": 3.634979316604973,
      "objective/train/tokens_used": 1062482400,
      "theoretical_loss": 3.634979316604973,
      "tokens_seen": 1042022400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000444053811235105,
      "loss": 2.7375,
      "theoretical_loss": 3.634979316604973,
      "tokens_seen": 1042022400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044403726120673353,
      "loss": 2.8463,
      "theoretical_loss": 3.6349577261122334,
      "tokens_seen": 1042087936
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000444020711178362,
      "loss": 2.7722,
      "theoretical_loss": 3.6349361373574176,
      "tokens_seen": 1042153472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044400416114999055,
      "loss": 2.8555,
      "theoretical_loss": 3.6349145503402758,
      "tokens_seen": 1042219008
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004439876111216191,
      "loss": 2.4635,
      "theoretical_loss": 3.6348929650605584,
      "tokens_seen": 1042284544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044397106109324763,
      "loss": 2.6687,
      "theoretical_loss": 3.6348713815180176,
      "tokens_seen": 1042350080
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044395451106487617,
      "loss": 2.7449,
      "theoretical_loss": 3.634849799712404,
      "tokens_seen": 1042415616
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044393796103650465,
      "loss": 2.7489,
      "theoretical_loss": 3.6348282196434676,
      "tokens_seen": 1042481152
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004439214110081332,
      "loss": 2.7324,
      "theoretical_loss": 3.63480664131096,
      "tokens_seen": 1042546688
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004439048609797617,
      "loss": 2.7847,
      "theoretical_loss": 3.6347850647146336,
      "tokens_seen": 1042612224
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044388831095139015,
      "loss": 2.789,
      "theoretical_loss": 3.6347634898542385,
      "tokens_seen": 1042677760
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004438717609230187,
      "loss": 2.7765,
      "theoretical_loss": 3.634741916729525,
      "tokens_seen": 1042743296
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044385521089464723,
      "loss": 2.7705,
      "theoretical_loss": 3.6347203453402464,
      "tokens_seen": 1042808832
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004438386608662757,
      "loss": 2.9783,
      "theoretical_loss": 3.634698775686153,
      "tokens_seen": 1042874368
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044382211083790425,
      "loss": 2.7055,
      "theoretical_loss": 3.6346772077669955,
      "tokens_seen": 1042939904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004438055608095328,
      "loss": 2.7327,
      "theoretical_loss": 3.634655641582527,
      "tokens_seen": 1043005440
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004437890107811613,
      "loss": 2.5479,
      "theoretical_loss": 3.6346340771324974,
      "tokens_seen": 1043070976
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004437724607527898,
      "loss": 2.6815,
      "theoretical_loss": 3.6346125144166592,
      "tokens_seen": 1043136512
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044375591072441835,
      "loss": 2.7103,
      "theoretical_loss": 3.634590953434764,
      "tokens_seen": 1043202048
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004437393606960469,
      "loss": 2.4422,
      "theoretical_loss": 3.6345693941865624,
      "tokens_seen": 1043267584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004437228106676754,
      "loss": 2.4775,
      "theoretical_loss": 3.634547836671808,
      "tokens_seen": 1043333120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004437062606393039,
      "loss": 2.6693,
      "theoretical_loss": 3.6345262808902508,
      "tokens_seen": 1043398656
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044368971061093244,
      "loss": 2.6115,
      "theoretical_loss": 3.6345047268416435,
      "tokens_seen": 1043464192
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000443673160582561,
      "loss": 2.699,
      "theoretical_loss": 3.6344831745257373,
      "tokens_seen": 1043529728
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004436566105541895,
      "loss": 2.7823,
      "theoretical_loss": 3.6344616239422844,
      "tokens_seen": 1043595264
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000443640060525818,
      "loss": 2.7067,
      "theoretical_loss": 3.634440075091037,
      "tokens_seen": 1043660800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044362351049744654,
      "loss": 2.7103,
      "theoretical_loss": 3.6344185279717474,
      "tokens_seen": 1043726336
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004436069604690751,
      "loss": 2.8632,
      "theoretical_loss": 3.634396982584167,
      "tokens_seen": 1043791872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004435904104407036,
      "loss": 2.5879,
      "theoretical_loss": 3.634375438928048,
      "tokens_seen": 1043857408
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004435738604123321,
      "loss": 2.5726,
      "theoretical_loss": 3.6343538970031433,
      "tokens_seen": 1043922944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044355731038396063,
      "loss": 2.7363,
      "theoretical_loss": 3.634332356809204,
      "tokens_seen": 1043988480
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044354076035558917,
      "loss": 2.4161,
      "theoretical_loss": 3.6343108183459827,
      "tokens_seen": 1044054016
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004435242103272177,
      "loss": 2.7348,
      "theoretical_loss": 3.634289281613232,
      "tokens_seen": 1044119552
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004435076602988462,
      "loss": 2.9361,
      "theoretical_loss": 3.6342677466107043,
      "tokens_seen": 1044185088
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044349111027047473,
      "loss": 2.6748,
      "theoretical_loss": 3.634246213338152,
      "tokens_seen": 1044250624
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044347456024210327,
      "loss": 2.6761,
      "theoretical_loss": 3.6342246817953274,
      "tokens_seen": 1044316160
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004434580102137318,
      "loss": 2.7576,
      "theoretical_loss": 3.634203151981983,
      "tokens_seen": 1044381696
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004434414601853603,
      "loss": 2.6056,
      "theoretical_loss": 3.6341816238978715,
      "tokens_seen": 1044447232
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004434249101569888,
      "loss": 2.818,
      "theoretical_loss": 3.634160097542745,
      "tokens_seen": 1044512768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044340836012861736,
      "loss": 2.4478,
      "theoretical_loss": 3.634138572916357,
      "tokens_seen": 1044578304
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004433918101002459,
      "loss": 2.8139,
      "theoretical_loss": 3.6341170500184603,
      "tokens_seen": 1044643840
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004433752600718744,
      "loss": 2.7649,
      "theoretical_loss": 3.6340955288488064,
      "tokens_seen": 1044709376
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004433587100435029,
      "loss": 2.8106,
      "theoretical_loss": 3.6340740094071493,
      "tokens_seen": 1044774912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044334216001513146,
      "loss": 2.8107,
      "theoretical_loss": 3.6340524916932413,
      "tokens_seen": 1044840448
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044332560998676,
      "loss": 2.7233,
      "theoretical_loss": 3.6340309757068354,
      "tokens_seen": 1044905984
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004433090599583885,
      "loss": 2.7086,
      "theoretical_loss": 3.634009461447685,
      "tokens_seen": 1044971520
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000443292509930017,
      "loss": 2.7689,
      "theoretical_loss": 3.6339879489155433,
      "tokens_seen": 1045037056
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044327595990164556,
      "loss": 2.8854,
      "theoretical_loss": 3.6339664381101624,
      "tokens_seen": 1045102592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004432594098732741,
      "loss": 2.5571,
      "theoretical_loss": 3.633944929031296,
      "tokens_seen": 1045168128
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004432428598449026,
      "loss": 2.8742,
      "theoretical_loss": 3.6339234216786975,
      "tokens_seen": 1045233664
    },
    {
      "epoch": 0.37,
      "objective/train/avg_token_score": 0.013435124419629574,
      "objective/train/avg_weight": 0.9841884970664978,
      "objective/train/docs_used": 595739,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.042588233947754,
      "objective/train/std_weight": 0.08420182764530182,
      "objective/train/theoretical_loss": 3.6339019160521198,
      "objective/train/tokens_used": 1065759200,
      "theoretical_loss": 3.6339019160521198,
      "tokens_seen": 1045299200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004432263098165311,
      "loss": 2.8202,
      "theoretical_loss": 3.6339019160521198,
      "tokens_seen": 1045299200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044320975978815965,
      "loss": 2.7785,
      "theoretical_loss": 3.633880412151316,
      "tokens_seen": 1045364736
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004431932097597882,
      "loss": 2.6158,
      "theoretical_loss": 3.6338589099760394,
      "tokens_seen": 1045430272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004431766597314167,
      "loss": 2.7827,
      "theoretical_loss": 3.633837409526044,
      "tokens_seen": 1045495808
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004431601097030452,
      "loss": 2.4831,
      "theoretical_loss": 3.6338159108010832,
      "tokens_seen": 1045561344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044314355967467375,
      "loss": 2.8052,
      "theoretical_loss": 3.63379441380091,
      "tokens_seen": 1045626880
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004431270096463023,
      "loss": 2.5054,
      "theoretical_loss": 3.6337729185252776,
      "tokens_seen": 1045692416
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044311045961793077,
      "loss": 2.732,
      "theoretical_loss": 3.6337514249739407,
      "tokens_seen": 1045757952
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004430939095895593,
      "loss": 2.6115,
      "theoretical_loss": 3.633729933146652,
      "tokens_seen": 1045823488
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044307735956118784,
      "loss": 2.6312,
      "theoretical_loss": 3.6337084430431656,
      "tokens_seen": 1045889024
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004430608095328164,
      "loss": 2.6278,
      "theoretical_loss": 3.633686954663235,
      "tokens_seen": 1045954560
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044304425950444486,
      "loss": 2.7007,
      "theoretical_loss": 3.6336654680066145,
      "tokens_seen": 1046020096
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004430277094760734,
      "loss": 2.7857,
      "theoretical_loss": 3.6336439830730574,
      "tokens_seen": 1046085632
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044301115944770194,
      "loss": 2.7132,
      "theoretical_loss": 3.6336224998623177,
      "tokens_seen": 1046151168
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004429946094193305,
      "loss": 2.4245,
      "theoretical_loss": 3.6336010183741494,
      "tokens_seen": 1046216704
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044297805939095896,
      "loss": 2.6836,
      "theoretical_loss": 3.6335795386083065,
      "tokens_seen": 1046282240
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004429615093625875,
      "loss": 2.8485,
      "theoretical_loss": 3.6335580605645434,
      "tokens_seen": 1046347776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044294495933421604,
      "loss": 2.4773,
      "theoretical_loss": 3.633536584242613,
      "tokens_seen": 1046413312
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004429284093058446,
      "loss": 2.8265,
      "theoretical_loss": 3.6335151096422713,
      "tokens_seen": 1046478848
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044291185927747306,
      "loss": 2.4246,
      "theoretical_loss": 3.6334936367632706,
      "tokens_seen": 1046544384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044289530924910154,
      "loss": 2.8536,
      "theoretical_loss": 3.633472165605366,
      "tokens_seen": 1046609920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004428787592207301,
      "loss": 2.4276,
      "theoretical_loss": 3.633450696168312,
      "tokens_seen": 1046675456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044286220919235856,
      "loss": 2.5636,
      "theoretical_loss": 3.6334292284518632,
      "tokens_seen": 1046740992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004428456591639871,
      "loss": 2.8377,
      "theoretical_loss": 3.6334077624557732,
      "tokens_seen": 1046806528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044282910913561564,
      "loss": 2.7424,
      "theoretical_loss": 3.633386298179797,
      "tokens_seen": 1046872064
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004428125591072442,
      "loss": 2.8393,
      "theoretical_loss": 3.6333648356236883,
      "tokens_seen": 1046937600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044279600907887266,
      "loss": 2.9755,
      "theoretical_loss": 3.6333433747872026,
      "tokens_seen": 1047003136
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004427794590505012,
      "loss": 2.8134,
      "theoretical_loss": 3.633321915670094,
      "tokens_seen": 1047068672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044276290902212973,
      "loss": 2.5343,
      "theoretical_loss": 3.6333004582721173,
      "tokens_seen": 1047134208
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044274635899375827,
      "loss": 2.8678,
      "theoretical_loss": 3.633279002593027,
      "tokens_seen": 1047199744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044272980896538675,
      "loss": 2.7047,
      "theoretical_loss": 3.6332575486325784,
      "tokens_seen": 1047265280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004427132589370153,
      "loss": 2.6296,
      "theoretical_loss": 3.6332360963905255,
      "tokens_seen": 1047330816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044269670890864383,
      "loss": 2.5482,
      "theoretical_loss": 3.633214645866624,
      "tokens_seen": 1047396352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044268015888027237,
      "loss": 2.6532,
      "theoretical_loss": 3.6331931970606277,
      "tokens_seen": 1047461888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044266360885190085,
      "loss": 2.5166,
      "theoretical_loss": 3.6331717499722926,
      "tokens_seen": 1047527424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004426470588235294,
      "loss": 2.759,
      "theoretical_loss": 3.633150304601373,
      "tokens_seen": 1047592960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004426305087951579,
      "loss": 2.8682,
      "theoretical_loss": 3.6331288609476244,
      "tokens_seen": 1047658496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044261395876678646,
      "loss": 2.5101,
      "theoretical_loss": 3.633107419010802,
      "tokens_seen": 1047724032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044259740873841494,
      "loss": 2.8466,
      "theoretical_loss": 3.6330859787906604,
      "tokens_seen": 1047789568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004425808587100435,
      "loss": 2.7622,
      "theoretical_loss": 3.6330645402869552,
      "tokens_seen": 1047855104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000442564308681672,
      "loss": 2.851,
      "theoretical_loss": 3.6330431034994417,
      "tokens_seen": 1047920640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044254775865330056,
      "loss": 2.5648,
      "theoretical_loss": 3.6330216684278747,
      "tokens_seen": 1047986176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044253120862492904,
      "loss": 2.6661,
      "theoretical_loss": 3.63300023507201,
      "tokens_seen": 1048051712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004425146585965576,
      "loss": 2.8767,
      "theoretical_loss": 3.6329788034316026,
      "tokens_seen": 1048117248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004424981085681861,
      "loss": 2.6751,
      "theoretical_loss": 3.6329573735064087,
      "tokens_seen": 1048182784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044248155853981465,
      "loss": 2.9813,
      "theoretical_loss": 3.6329359452961834,
      "tokens_seen": 1048248320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044246500851144314,
      "loss": 2.8456,
      "theoretical_loss": 3.6329145188006815,
      "tokens_seen": 1048313856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004424484584830717,
      "loss": 2.7327,
      "theoretical_loss": 3.6328930940196598,
      "tokens_seen": 1048379392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004424319084547002,
      "loss": 2.4165,
      "theoretical_loss": 3.6328716709528734,
      "tokens_seen": 1048444928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044241535842632875,
      "loss": 2.6258,
      "theoretical_loss": 3.632850249600078,
      "tokens_seen": 1048510464
    },
    {
      "epoch": 0.37,
      "objective/train/avg_token_score": 0.022597815841436386,
      "objective/train/avg_weight": 0.9734179377555847,
      "objective/train/docs_used": 597756,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.797966957092285,
      "objective/train/std_weight": 0.0981125459074974,
      "objective/train/theoretical_loss": 3.632828829961029,
      "objective/train/tokens_used": 1069036000,
      "theoretical_loss": 3.632828829961029,
      "tokens_seen": 1048576000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044239880839795723,
      "loss": 2.7547,
      "theoretical_loss": 3.632828829961029,
      "tokens_seen": 1048576000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044238225836958577,
      "loss": 2.7674,
      "theoretical_loss": 3.632807412035483,
      "tokens_seen": 1048641536
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004423657083412143,
      "loss": 2.6572,
      "theoretical_loss": 3.6327859958231956,
      "tokens_seen": 1048707072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044234915831284285,
      "loss": 2.975,
      "theoretical_loss": 3.632764581323922,
      "tokens_seen": 1048772608
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044233260828447133,
      "loss": 2.4823,
      "theoretical_loss": 3.632743168537419,
      "tokens_seen": 1048838144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044231605825609987,
      "loss": 2.7159,
      "theoretical_loss": 3.6327217574634423,
      "tokens_seen": 1048903680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004422995082277284,
      "loss": 2.7558,
      "theoretical_loss": 3.632700348101748,
      "tokens_seen": 1048969216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044228295819935694,
      "loss": 2.6703,
      "theoretical_loss": 3.632678940452092,
      "tokens_seen": 1049034752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004422664081709854,
      "loss": 2.616,
      "theoretical_loss": 3.6326575345142307,
      "tokens_seen": 1049100288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044224985814261396,
      "loss": 2.735,
      "theoretical_loss": 3.63263613028792,
      "tokens_seen": 1049165824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004422333081142425,
      "loss": 2.6076,
      "theoretical_loss": 3.632614727772917,
      "tokens_seen": 1049231360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044221675808587104,
      "loss": 2.5678,
      "theoretical_loss": 3.6325933269689763,
      "tokens_seen": 1049296896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004422002080574995,
      "loss": 2.7116,
      "theoretical_loss": 3.6325719278758557,
      "tokens_seen": 1049362432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044218365802912806,
      "loss": 2.8882,
      "theoretical_loss": 3.6325505304933117,
      "tokens_seen": 1049427968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004421671080007566,
      "loss": 2.7731,
      "theoretical_loss": 3.6325291348211,
      "tokens_seen": 1049493504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044215055797238513,
      "loss": 2.7587,
      "theoretical_loss": 3.632507740858977,
      "tokens_seen": 1049559040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004421340079440136,
      "loss": 2.4768,
      "theoretical_loss": 3.6324863486067,
      "tokens_seen": 1049624576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044211745791564215,
      "loss": 2.6262,
      "theoretical_loss": 3.632464958064025,
      "tokens_seen": 1049690112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004421009078872707,
      "loss": 2.808,
      "theoretical_loss": 3.6324435692307087,
      "tokens_seen": 1049755648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044208435785889923,
      "loss": 2.51,
      "theoretical_loss": 3.6324221821065077,
      "tokens_seen": 1049821184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0004420678078305277,
      "loss": 2.6677,
      "theoretical_loss": 3.6324007966911798,
      "tokens_seen": 1049886720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00044205125780215625,
      "loss": 2.5544,
      "theoretical_loss": 3.6323794129844806,
      "tokens_seen": 1049952256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004420347077737848,
      "loss": 2.8079,
      "theoretical_loss": 3.6323580309861674,
      "tokens_seen": 1050017792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004420181577454133,
      "loss": 2.575,
      "theoretical_loss": 3.632336650695997,
      "tokens_seen": 1050083328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004420016077170418,
      "loss": 2.6289,
      "theoretical_loss": 3.6323152721137264,
      "tokens_seen": 1050148864
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044198505768867035,
      "loss": 2.627,
      "theoretical_loss": 3.6322938952391124,
      "tokens_seen": 1050214400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004419685076602989,
      "loss": 2.6092,
      "theoretical_loss": 3.632272520071912,
      "tokens_seen": 1050279936
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004419519576319274,
      "loss": 2.9186,
      "theoretical_loss": 3.632251146611883,
      "tokens_seen": 1050345472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004419354076035559,
      "loss": 2.5339,
      "theoretical_loss": 3.632229774858782,
      "tokens_seen": 1050411008
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004419188575751844,
      "loss": 2.8796,
      "theoretical_loss": 3.6322084048123666,
      "tokens_seen": 1050476544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044190230754681287,
      "loss": 2.7971,
      "theoretical_loss": 3.632187036472393,
      "tokens_seen": 1050542080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004418857575184414,
      "loss": 2.6369,
      "theoretical_loss": 3.63216566983862,
      "tokens_seen": 1050607616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044186920749006995,
      "loss": 2.4871,
      "theoretical_loss": 3.632144304910804,
      "tokens_seen": 1050673152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004418526574616985,
      "loss": 2.8632,
      "theoretical_loss": 3.6321229416887024,
      "tokens_seen": 1050738688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044183610743332697,
      "loss": 3.0431,
      "theoretical_loss": 3.6321015801720726,
      "tokens_seen": 1050804224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004418195574049555,
      "loss": 2.8058,
      "theoretical_loss": 3.632080220360673,
      "tokens_seen": 1050869760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044180300737658404,
      "loss": 2.7526,
      "theoretical_loss": 3.63205886225426,
      "tokens_seen": 1050935296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004417864573482126,
      "loss": 2.7593,
      "theoretical_loss": 3.6320375058525913,
      "tokens_seen": 1051000832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044176990731984106,
      "loss": 2.7071,
      "theoretical_loss": 3.632016151155425,
      "tokens_seen": 1051066368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004417533572914696,
      "loss": 2.8155,
      "theoretical_loss": 3.631994798162519,
      "tokens_seen": 1051131904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044173680726309814,
      "loss": 2.6889,
      "theoretical_loss": 3.631973446873631,
      "tokens_seen": 1051197440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004417202572347267,
      "loss": 2.5694,
      "theoretical_loss": 3.6319520972885178,
      "tokens_seen": 1051262976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044170370720635516,
      "loss": 2.9114,
      "theoretical_loss": 3.6319307494069384,
      "tokens_seen": 1051328512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004416871571779837,
      "loss": 2.5414,
      "theoretical_loss": 3.63190940322865,
      "tokens_seen": 1051394048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044167060714961223,
      "loss": 2.7856,
      "theoretical_loss": 3.631888058753411,
      "tokens_seen": 1051459584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044165405712124077,
      "loss": 2.7098,
      "theoretical_loss": 3.6318667159809794,
      "tokens_seen": 1051525120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044163750709286926,
      "loss": 2.774,
      "theoretical_loss": 3.631845374911113,
      "tokens_seen": 1051590656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004416209570644978,
      "loss": 2.7472,
      "theoretical_loss": 3.63182403554357,
      "tokens_seen": 1051656192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044160440703612633,
      "loss": 2.601,
      "theoretical_loss": 3.6318026978781077,
      "tokens_seen": 1051721728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044158785700775487,
      "loss": 2.7166,
      "theoretical_loss": 3.631781361914485,
      "tokens_seen": 1051787264
    },
    {
      "epoch": 0.38,
      "objective/train/avg_token_score": 0.009286642074584961,
      "objective/train/avg_weight": 0.9890668988227844,
      "objective/train/docs_used": 599577,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.511042356491089,
      "objective/train/std_weight": 0.061566129326820374,
      "objective/train/theoretical_loss": 3.631760027652461,
      "objective/train/tokens_used": 1072312800,
      "theoretical_loss": 3.631760027652461,
      "tokens_seen": 1051852800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044157130697938335,
      "loss": 2.6373,
      "theoretical_loss": 3.631760027652461,
      "tokens_seen": 1051852800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004415547569510119,
      "loss": 2.5593,
      "theoretical_loss": 3.631738695091793,
      "tokens_seen": 1051918336
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004415382069226404,
      "loss": 2.801,
      "theoretical_loss": 3.631717364232239,
      "tokens_seen": 1051983872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044152165689426896,
      "loss": 3.0528,
      "theoretical_loss": 3.6316960350735585,
      "tokens_seen": 1052049408
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044150510686589745,
      "loss": 2.6151,
      "theoretical_loss": 3.631674707615509,
      "tokens_seen": 1052114944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000441488556837526,
      "loss": 2.8576,
      "theoretical_loss": 3.6316533818578494,
      "tokens_seen": 1052180480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004414720068091545,
      "loss": 2.9531,
      "theoretical_loss": 3.6316320578003385,
      "tokens_seen": 1052246016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044145545678078306,
      "loss": 2.4884,
      "theoretical_loss": 3.631610735442734,
      "tokens_seen": 1052311552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004414389067524116,
      "loss": 2.7671,
      "theoretical_loss": 3.631589414784795,
      "tokens_seen": 1052377088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004414223567240401,
      "loss": 2.7584,
      "theoretical_loss": 3.631568095826281,
      "tokens_seen": 1052442624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004414058066956686,
      "loss": 2.8336,
      "theoretical_loss": 3.631546778566949,
      "tokens_seen": 1052508160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044138925666729716,
      "loss": 2.7663,
      "theoretical_loss": 3.631525463006559,
      "tokens_seen": 1052573696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004413727066389257,
      "loss": 2.5366,
      "theoretical_loss": 3.6315041491448703,
      "tokens_seen": 1052639232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004413561566105542,
      "loss": 2.6451,
      "theoretical_loss": 3.6314828369816405,
      "tokens_seen": 1052704768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004413396065821827,
      "loss": 2.7942,
      "theoretical_loss": 3.6314615265166292,
      "tokens_seen": 1052770304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044132305655381125,
      "loss": 2.714,
      "theoretical_loss": 3.6314402177495957,
      "tokens_seen": 1052835840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004413065065254398,
      "loss": 2.7652,
      "theoretical_loss": 3.6314189106802974,
      "tokens_seen": 1052901376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044128995649706827,
      "loss": 2.8026,
      "theoretical_loss": 3.6313976053084955,
      "tokens_seen": 1052966912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004412734064686968,
      "loss": 2.9647,
      "theoretical_loss": 3.6313763016339484,
      "tokens_seen": 1053032448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044125685644032535,
      "loss": 2.6581,
      "theoretical_loss": 3.6313549996564145,
      "tokens_seen": 1053097984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004412403064119539,
      "loss": 2.9755,
      "theoretical_loss": 3.631333699375654,
      "tokens_seen": 1053163520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044122375638358237,
      "loss": 2.7313,
      "theoretical_loss": 3.6313124007914253,
      "tokens_seen": 1053229056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004412072063552109,
      "loss": 2.637,
      "theoretical_loss": 3.6312911039034885,
      "tokens_seen": 1053294592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044119065632683944,
      "loss": 2.9313,
      "theoretical_loss": 3.6312698087116027,
      "tokens_seen": 1053360128
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000441174106298468,
      "loss": 2.9303,
      "theoretical_loss": 3.6312485152155274,
      "tokens_seen": 1053425664
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044115755627009646,
      "loss": 2.7637,
      "theoretical_loss": 3.6312272234150216,
      "tokens_seen": 1053491200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000441141006241725,
      "loss": 2.6959,
      "theoretical_loss": 3.6312059333098454,
      "tokens_seen": 1053556736
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044112445621335354,
      "loss": 2.6479,
      "theoretical_loss": 3.631184644899758,
      "tokens_seen": 1053622272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004411079061849821,
      "loss": 2.8052,
      "theoretical_loss": 3.6311633581845193,
      "tokens_seen": 1053687808
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044109135615661056,
      "loss": 2.9902,
      "theoretical_loss": 3.631142073163889,
      "tokens_seen": 1053753344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004410748061282391,
      "loss": 2.6837,
      "theoretical_loss": 3.631120789837626,
      "tokens_seen": 1053818880
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044105825609986764,
      "loss": 2.6225,
      "theoretical_loss": 3.6310995082054913,
      "tokens_seen": 1053884416
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004410417060714962,
      "loss": 2.6535,
      "theoretical_loss": 3.631078228267244,
      "tokens_seen": 1053949952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044102515604312466,
      "loss": 2.6454,
      "theoretical_loss": 3.6310569500226437,
      "tokens_seen": 1054015488
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004410086060147532,
      "loss": 2.8551,
      "theoretical_loss": 3.631035673471451,
      "tokens_seen": 1054081024
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044099205598638173,
      "loss": 2.8147,
      "theoretical_loss": 3.6310143986134253,
      "tokens_seen": 1054146560
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044097550595801027,
      "loss": 2.6169,
      "theoretical_loss": 3.6309931254483274,
      "tokens_seen": 1054212096
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044095895592963875,
      "loss": 2.7944,
      "theoretical_loss": 3.6309718539759164,
      "tokens_seen": 1054277632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004409424059012673,
      "loss": 2.7444,
      "theoretical_loss": 3.630950584195953,
      "tokens_seen": 1054343168
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004409258558728957,
      "loss": 2.7589,
      "theoretical_loss": 3.6309293161081966,
      "tokens_seen": 1054408704
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044090930584452426,
      "loss": 2.6996,
      "theoretical_loss": 3.6309080497124087,
      "tokens_seen": 1054474240
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004408927558161528,
      "loss": 2.8254,
      "theoretical_loss": 3.6308867850083484,
      "tokens_seen": 1054539776
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044087620578778133,
      "loss": 2.726,
      "theoretical_loss": 3.6308655219957764,
      "tokens_seen": 1054605312
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004408596557594098,
      "loss": 2.7748,
      "theoretical_loss": 3.6308442606744533,
      "tokens_seen": 1054670848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044084310573103835,
      "loss": 2.6916,
      "theoretical_loss": 3.6308230010441394,
      "tokens_seen": 1054736384
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004408265557026669,
      "loss": 2.7048,
      "theoretical_loss": 3.6308017431045947,
      "tokens_seen": 1054801920
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044081000567429543,
      "loss": 2.8036,
      "theoretical_loss": 3.6307804868555804,
      "tokens_seen": 1054867456
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004407934556459239,
      "loss": 2.4696,
      "theoretical_loss": 3.6307592322968563,
      "tokens_seen": 1054932992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044077690561755245,
      "loss": 2.5812,
      "theoretical_loss": 3.6307379794281838,
      "tokens_seen": 1054998528
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000440760355589181,
      "loss": 2.7798,
      "theoretical_loss": 3.6307167282493227,
      "tokens_seen": 1055064064
    },
    {
      "epoch": 0.38,
      "objective/train/avg_token_score": 0.0391039177775383,
      "objective/train/avg_weight": 0.953972578048706,
      "objective/train/docs_used": 601442,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7431955337524414,
      "objective/train/std_weight": 0.17007966339588165,
      "objective/train/theoretical_loss": 3.630695478760034,
      "objective/train/tokens_used": 1075589600,
      "theoretical_loss": 3.630695478760034,
      "tokens_seen": 1055129600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004407438055608095,
      "loss": 2.5104,
      "theoretical_loss": 3.630695478760034,
      "tokens_seen": 1055129600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000440727255532438,
      "loss": 2.8012,
      "theoretical_loss": 3.6306742309600786,
      "tokens_seen": 1055195136
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044071070550406654,
      "loss": 2.6084,
      "theoretical_loss": 3.6306529848492177,
      "tokens_seen": 1055260672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004406941554756951,
      "loss": 2.8462,
      "theoretical_loss": 3.6306317404272113,
      "tokens_seen": 1055326208
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004406776054473236,
      "loss": 2.7836,
      "theoretical_loss": 3.6306104976938203,
      "tokens_seen": 1055391744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004406610554189521,
      "loss": 2.791,
      "theoretical_loss": 3.6305892566488067,
      "tokens_seen": 1055457280
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044064450539058064,
      "loss": 2.5724,
      "theoretical_loss": 3.6305680172919303,
      "tokens_seen": 1055522816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004406279553622092,
      "loss": 2.8303,
      "theoretical_loss": 3.6305467796229527,
      "tokens_seen": 1055588352
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004406114053338377,
      "loss": 2.8072,
      "theoretical_loss": 3.6305255436416353,
      "tokens_seen": 1055653888
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004405948553054662,
      "loss": 2.6115,
      "theoretical_loss": 3.6305043093477383,
      "tokens_seen": 1055719424
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044057830527709474,
      "loss": 2.8067,
      "theoretical_loss": 3.630483076741023,
      "tokens_seen": 1055784960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004405617552487233,
      "loss": 2.6006,
      "theoretical_loss": 3.6304618458212516,
      "tokens_seen": 1055850496
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004405452052203518,
      "loss": 2.7539,
      "theoretical_loss": 3.630440616588185,
      "tokens_seen": 1055916032
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004405286551919803,
      "loss": 2.5831,
      "theoretical_loss": 3.630419389041584,
      "tokens_seen": 1055981568
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044051210516360883,
      "loss": 2.8899,
      "theoretical_loss": 3.6303981631812103,
      "tokens_seen": 1056047104
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044049555513523737,
      "loss": 2.7986,
      "theoretical_loss": 3.6303769390068257,
      "tokens_seen": 1056112640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004404790051068659,
      "loss": 2.7356,
      "theoretical_loss": 3.6303557165181912,
      "tokens_seen": 1056178176
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004404624550784944,
      "loss": 2.7817,
      "theoretical_loss": 3.630334495715068,
      "tokens_seen": 1056243712
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044044590505012293,
      "loss": 2.809,
      "theoretical_loss": 3.630313276597218,
      "tokens_seen": 1056309248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044042935502175147,
      "loss": 2.5658,
      "theoretical_loss": 3.630292059164404,
      "tokens_seen": 1056374784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044041280499338,
      "loss": 2.6271,
      "theoretical_loss": 3.630270843416385,
      "tokens_seen": 1056440320
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004403962549650085,
      "loss": 2.5701,
      "theoretical_loss": 3.6302496293529254,
      "tokens_seen": 1056505856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000440379704936637,
      "loss": 2.7639,
      "theoretical_loss": 3.6302284169737855,
      "tokens_seen": 1056571392
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044036315490826556,
      "loss": 2.7662,
      "theoretical_loss": 3.630207206278727,
      "tokens_seen": 1056636928
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004403466048798941,
      "loss": 2.6531,
      "theoretical_loss": 3.630185997267513,
      "tokens_seen": 1056702464
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004403300548515226,
      "loss": 2.8286,
      "theoretical_loss": 3.6301647899399043,
      "tokens_seen": 1056768000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004403135048231511,
      "loss": 2.7143,
      "theoretical_loss": 3.630143584295663,
      "tokens_seen": 1056833536
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044029695479477966,
      "loss": 2.7301,
      "theoretical_loss": 3.6301223803345515,
      "tokens_seen": 1056899072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004402804047664082,
      "loss": 2.7607,
      "theoretical_loss": 3.6301011780563313,
      "tokens_seen": 1056964608
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004402638547380367,
      "loss": 2.7927,
      "theoretical_loss": 3.630079977460765,
      "tokens_seen": 1057030144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004402473047096652,
      "loss": 2.809,
      "theoretical_loss": 3.630058778547615,
      "tokens_seen": 1057095680
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044023075468129375,
      "loss": 2.7433,
      "theoretical_loss": 3.630037581316642,
      "tokens_seen": 1057161216
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004402142046529223,
      "loss": 2.6327,
      "theoretical_loss": 3.63001638576761,
      "tokens_seen": 1057226752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004401976546245508,
      "loss": 2.5757,
      "theoretical_loss": 3.6299951919002806,
      "tokens_seen": 1057292288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004401811045961793,
      "loss": 2.4753,
      "theoretical_loss": 3.6299739997144163,
      "tokens_seen": 1057357824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044016455456780785,
      "loss": 2.9634,
      "theoretical_loss": 3.6299528092097795,
      "tokens_seen": 1057423360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004401480045394364,
      "loss": 2.8903,
      "theoretical_loss": 3.629931620386132,
      "tokens_seen": 1057488896
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044013145451106487,
      "loss": 2.7672,
      "theoretical_loss": 3.6299104332432366,
      "tokens_seen": 1057554432
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004401149044826934,
      "loss": 2.5103,
      "theoretical_loss": 3.6298892477808566,
      "tokens_seen": 1057619968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044009835445432195,
      "loss": 2.8514,
      "theoretical_loss": 3.6298680639987535,
      "tokens_seen": 1057685504
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004400818044259505,
      "loss": 2.3832,
      "theoretical_loss": 3.6298468818966905,
      "tokens_seen": 1057751040
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044006525439757897,
      "loss": 2.7644,
      "theoretical_loss": 3.6298257014744304,
      "tokens_seen": 1057816576
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004400487043692075,
      "loss": 2.733,
      "theoretical_loss": 3.6298045227317357,
      "tokens_seen": 1057882112
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00044003215434083604,
      "loss": 2.5425,
      "theoretical_loss": 3.6297833456683692,
      "tokens_seen": 1057947648
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004400156043124646,
      "loss": 2.6183,
      "theoretical_loss": 3.629762170284094,
      "tokens_seen": 1058013184
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043999905428409306,
      "loss": 2.5648,
      "theoretical_loss": 3.629740996578672,
      "tokens_seen": 1058078720
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004399825042557216,
      "loss": 2.7041,
      "theoretical_loss": 3.6297198245518674,
      "tokens_seen": 1058144256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043996595422735014,
      "loss": 2.6886,
      "theoretical_loss": 3.629698654203442,
      "tokens_seen": 1058209792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043994940419897857,
      "loss": 2.5941,
      "theoretical_loss": 3.6296774855331604,
      "tokens_seen": 1058275328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004399328541706071,
      "loss": 2.8172,
      "theoretical_loss": 3.6296563185407837,
      "tokens_seen": 1058340864
    },
    {
      "epoch": 0.38,
      "objective/train/avg_token_score": 0.012255735695362091,
      "objective/train/avg_weight": 0.9856202602386475,
      "objective/train/docs_used": 602965,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.730323314666748,
      "objective/train/std_weight": 0.06556746363639832,
      "objective/train/theoretical_loss": 3.6296351532260767,
      "objective/train/tokens_used": 1078866400,
      "theoretical_loss": 3.6296351532260767,
      "tokens_seen": 1058406400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043991630414223564,
      "loss": 2.6044,
      "theoretical_loss": 3.6296351532260767,
      "tokens_seen": 1058406400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004398997541138642,
      "loss": 2.8879,
      "theoretical_loss": 3.629613989588801,
      "tokens_seen": 1058471936
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043988320408549266,
      "loss": 2.6419,
      "theoretical_loss": 3.6295928276287217,
      "tokens_seen": 1058537472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004398666540571212,
      "loss": 2.755,
      "theoretical_loss": 3.6295716673456004,
      "tokens_seen": 1058603008
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043985010402874974,
      "loss": 2.5801,
      "theoretical_loss": 3.629550508739201,
      "tokens_seen": 1058668544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004398335540003783,
      "loss": 2.6203,
      "theoretical_loss": 3.6295293518092873,
      "tokens_seen": 1058734080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043981700397200676,
      "loss": 2.5321,
      "theoretical_loss": 3.629508196555622,
      "tokens_seen": 1058799616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004398004539436353,
      "loss": 2.6068,
      "theoretical_loss": 3.629487042977969,
      "tokens_seen": 1058865152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043978390391526383,
      "loss": 2.9114,
      "theoretical_loss": 3.6294658910760913,
      "tokens_seen": 1058930688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043976735388689237,
      "loss": 2.7666,
      "theoretical_loss": 3.6294447408497525,
      "tokens_seen": 1058996224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043975080385852085,
      "loss": 2.7669,
      "theoretical_loss": 3.6294235922987177,
      "tokens_seen": 1059061760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004397342538301494,
      "loss": 2.6686,
      "theoretical_loss": 3.6294024454227483,
      "tokens_seen": 1059127296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043971770380177793,
      "loss": 2.67,
      "theoretical_loss": 3.6293813002216093,
      "tokens_seen": 1059192832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043970115377340647,
      "loss": 2.8028,
      "theoretical_loss": 3.629360156695064,
      "tokens_seen": 1059258368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043968460374503495,
      "loss": 2.6827,
      "theoretical_loss": 3.629339014842876,
      "tokens_seen": 1059323904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004396680537166635,
      "loss": 2.8404,
      "theoretical_loss": 3.6293178746648103,
      "tokens_seen": 1059389440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000439651503688292,
      "loss": 2.6785,
      "theoretical_loss": 3.6292967361606294,
      "tokens_seen": 1059454976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043963495365992056,
      "loss": 2.6769,
      "theoretical_loss": 3.6292755993300982,
      "tokens_seen": 1059520512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043961840363154905,
      "loss": 2.7703,
      "theoretical_loss": 3.62925446417298,
      "tokens_seen": 1059586048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004396018536031776,
      "loss": 2.6909,
      "theoretical_loss": 3.629233330689039,
      "tokens_seen": 1059651584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004395853035748061,
      "loss": 2.8712,
      "theoretical_loss": 3.6292121988780393,
      "tokens_seen": 1059717120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043956875354643466,
      "loss": 2.6114,
      "theoretical_loss": 3.6291910687397455,
      "tokens_seen": 1059782656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043955220351806314,
      "loss": 2.7041,
      "theoretical_loss": 3.629169940273921,
      "tokens_seen": 1059848192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004395356534896917,
      "loss": 2.5955,
      "theoretical_loss": 3.6291488134803305,
      "tokens_seen": 1059913728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004395191034613202,
      "loss": 2.6625,
      "theoretical_loss": 3.629127688358738,
      "tokens_seen": 1059979264
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043950255343294876,
      "loss": 3.0016,
      "theoretical_loss": 3.6291065649089083,
      "tokens_seen": 1060044800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043948600340457724,
      "loss": 2.7037,
      "theoretical_loss": 3.629085443130605,
      "tokens_seen": 1060110336
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004394694533762058,
      "loss": 2.9583,
      "theoretical_loss": 3.6290643230235924,
      "tokens_seen": 1060175872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004394529033478343,
      "loss": 2.6234,
      "theoretical_loss": 3.6290432045876364,
      "tokens_seen": 1060241408
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043943635331946285,
      "loss": 2.9317,
      "theoretical_loss": 3.6290220878225004,
      "tokens_seen": 1060306944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043941980329109133,
      "loss": 2.5849,
      "theoretical_loss": 3.629000972727949,
      "tokens_seen": 1060372480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043940325326271987,
      "loss": 2.4916,
      "theoretical_loss": 3.6289798593037466,
      "tokens_seen": 1060438016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004393867032343484,
      "loss": 2.7555,
      "theoretical_loss": 3.6289587475496585,
      "tokens_seen": 1060503552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043937015320597695,
      "loss": 2.5172,
      "theoretical_loss": 3.628937637465449,
      "tokens_seen": 1060569088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043935360317760543,
      "loss": 2.7785,
      "theoretical_loss": 3.628916529050883,
      "tokens_seen": 1060634624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043933705314923397,
      "loss": 2.6195,
      "theoretical_loss": 3.628895422305725,
      "tokens_seen": 1060700160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004393205031208625,
      "loss": 2.2099,
      "theoretical_loss": 3.6288743172297404,
      "tokens_seen": 1060765696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043930395309249104,
      "loss": 2.9365,
      "theoretical_loss": 3.628853213822693,
      "tokens_seen": 1060831232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004392874030641195,
      "loss": 2.7903,
      "theoretical_loss": 3.6288321120843494,
      "tokens_seen": 1060896768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043927085303574806,
      "loss": 2.7468,
      "theoretical_loss": 3.6288110120144728,
      "tokens_seen": 1060962304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004392543030073766,
      "loss": 2.6928,
      "theoretical_loss": 3.6287899136128297,
      "tokens_seen": 1061027840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043923775297900514,
      "loss": 2.7061,
      "theoretical_loss": 3.6287688168791843,
      "tokens_seen": 1061093376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004392212029506336,
      "loss": 2.7422,
      "theoretical_loss": 3.6287477218133017,
      "tokens_seen": 1061158912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043920465292226216,
      "loss": 2.8008,
      "theoretical_loss": 3.6287266284149475,
      "tokens_seen": 1061224448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004391881028938907,
      "loss": 2.5366,
      "theoretical_loss": 3.6287055366838867,
      "tokens_seen": 1061289984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043917155286551924,
      "loss": 2.6917,
      "theoretical_loss": 3.628684446619885,
      "tokens_seen": 1061355520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004391550028371477,
      "loss": 2.9612,
      "theoretical_loss": 3.628663358222707,
      "tokens_seen": 1061421056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043913845280877626,
      "loss": 2.8072,
      "theoretical_loss": 3.6286422714921187,
      "tokens_seen": 1061486592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004391219027804048,
      "loss": 2.7577,
      "theoretical_loss": 3.6286211864278846,
      "tokens_seen": 1061552128
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043910535275203333,
      "loss": 2.6177,
      "theoretical_loss": 3.628600103029771,
      "tokens_seen": 1061617664
    },
    {
      "epoch": 0.38,
      "objective/train/avg_token_score": 0.007020127028226852,
      "objective/train/avg_weight": 0.9917437434196472,
      "objective/train/docs_used": 604364,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.954558849334717,
      "objective/train/std_weight": 0.05792878568172455,
      "objective/train/theoretical_loss": 3.6285790212975435,
      "objective/train/tokens_used": 1082143200,
      "theoretical_loss": 3.6285790212975435,
      "tokens_seen": 1061683200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004390888027236618,
      "loss": 2.5972,
      "theoretical_loss": 3.6285790212975435,
      "tokens_seen": 1061683200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043907225269529035,
      "loss": 2.6793,
      "theoretical_loss": 3.6285579412309668,
      "tokens_seen": 1061748736
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004390557026669189,
      "loss": 2.4683,
      "theoretical_loss": 3.628536862829807,
      "tokens_seen": 1061814272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004390391526385474,
      "loss": 2.8281,
      "theoretical_loss": 3.62851578609383,
      "tokens_seen": 1061879808
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004390226026101759,
      "loss": 2.7958,
      "theoretical_loss": 3.6284947110228014,
      "tokens_seen": 1061945344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043900605258180445,
      "loss": 2.7188,
      "theoretical_loss": 3.6284736376164872,
      "tokens_seen": 1062010880
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000438989502553433,
      "loss": 2.556,
      "theoretical_loss": 3.6284525658746523,
      "tokens_seen": 1062076416
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004389729525250614,
      "loss": 2.6591,
      "theoretical_loss": 3.628431495797063,
      "tokens_seen": 1062141952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043895640249668995,
      "loss": 2.8174,
      "theoretical_loss": 3.6284104273834847,
      "tokens_seen": 1062207488
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004389398524683185,
      "loss": 2.6326,
      "theoretical_loss": 3.628389360633684,
      "tokens_seen": 1062273024
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043892330243994703,
      "loss": 2.4031,
      "theoretical_loss": 3.6283682955474275,
      "tokens_seen": 1062338560
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004389067524115755,
      "loss": 2.7164,
      "theoretical_loss": 3.62834723212448,
      "tokens_seen": 1062404096
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043889020238320405,
      "loss": 2.4754,
      "theoretical_loss": 3.628326170364608,
      "tokens_seen": 1062469632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004388736523548326,
      "loss": 2.8002,
      "theoretical_loss": 3.6283051102675774,
      "tokens_seen": 1062535168
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004388571023264611,
      "loss": 2.3891,
      "theoretical_loss": 3.6282840518331545,
      "tokens_seen": 1062600704
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004388405522980896,
      "loss": 2.507,
      "theoretical_loss": 3.628262995061106,
      "tokens_seen": 1062666240
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043882400226971814,
      "loss": 2.8368,
      "theoretical_loss": 3.628241939951198,
      "tokens_seen": 1062731776
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004388074522413467,
      "loss": 2.6839,
      "theoretical_loss": 3.628220886503196,
      "tokens_seen": 1062797312
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004387909022129752,
      "loss": 2.8188,
      "theoretical_loss": 3.628199834716867,
      "tokens_seen": 1062862848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004387743521846037,
      "loss": 2.6427,
      "theoretical_loss": 3.6281787845919777,
      "tokens_seen": 1062928384
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043875780215623224,
      "loss": 2.7665,
      "theoretical_loss": 3.628157736128294,
      "tokens_seen": 1062993920
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004387412521278608,
      "loss": 2.583,
      "theoretical_loss": 3.6281366893255824,
      "tokens_seen": 1063059456
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004387247020994893,
      "loss": 2.8556,
      "theoretical_loss": 3.6281156441836098,
      "tokens_seen": 1063124992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004387081520711178,
      "loss": 2.4802,
      "theoretical_loss": 3.628094600702143,
      "tokens_seen": 1063190528
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043869160204274634,
      "loss": 2.7581,
      "theoretical_loss": 3.6280735588809474,
      "tokens_seen": 1063256064
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004386750520143749,
      "loss": 2.8235,
      "theoretical_loss": 3.628052518719791,
      "tokens_seen": 1063321600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004386585019860034,
      "loss": 2.7793,
      "theoretical_loss": 3.62803148021844,
      "tokens_seen": 1063387136
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004386419519576319,
      "loss": 2.6796,
      "theoretical_loss": 3.628010443376662,
      "tokens_seen": 1063452672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043862540192926043,
      "loss": 2.7736,
      "theoretical_loss": 3.6279894081942223,
      "tokens_seen": 1063518208
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043860885190088897,
      "loss": 2.8832,
      "theoretical_loss": 3.6279683746708886,
      "tokens_seen": 1063583744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004385923018725175,
      "loss": 2.887,
      "theoretical_loss": 3.627947342806428,
      "tokens_seen": 1063649280
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000438575751844146,
      "loss": 2.5929,
      "theoretical_loss": 3.627926312600607,
      "tokens_seen": 1063714816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043855920181577453,
      "loss": 2.9522,
      "theoretical_loss": 3.6279052840531936,
      "tokens_seen": 1063780352
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043854265178740307,
      "loss": 2.9375,
      "theoretical_loss": 3.627884257163953,
      "tokens_seen": 1063845888
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004385261017590316,
      "loss": 2.8586,
      "theoretical_loss": 3.6278632319326545,
      "tokens_seen": 1063911424
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004385095517306601,
      "loss": 2.8395,
      "theoretical_loss": 3.6278422083590636,
      "tokens_seen": 1063976960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004384930017022886,
      "loss": 2.3857,
      "theoretical_loss": 3.627821186442948,
      "tokens_seen": 1064042496
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043847645167391716,
      "loss": 2.7929,
      "theoretical_loss": 3.627800166184076,
      "tokens_seen": 1064108032
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004384599016455457,
      "loss": 2.5499,
      "theoretical_loss": 3.6277791475822134,
      "tokens_seen": 1064173568
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004384433516171742,
      "loss": 2.7814,
      "theoretical_loss": 3.627758130637128,
      "tokens_seen": 1064239104
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004384268015888027,
      "loss": 2.7125,
      "theoretical_loss": 3.6277371153485873,
      "tokens_seen": 1064304640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043841025156043126,
      "loss": 2.7014,
      "theoretical_loss": 3.6277161017163593,
      "tokens_seen": 1064370176
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004383937015320598,
      "loss": 2.7426,
      "theoretical_loss": 3.62769508974021,
      "tokens_seen": 1064435712
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004383771515036883,
      "loss": 2.5238,
      "theoretical_loss": 3.6276740794199087,
      "tokens_seen": 1064501248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004383606014753168,
      "loss": 2.7908,
      "theoretical_loss": 3.6276530707552217,
      "tokens_seen": 1064566784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043834405144694535,
      "loss": 2.8158,
      "theoretical_loss": 3.627632063745917,
      "tokens_seen": 1064632320
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004383275014185739,
      "loss": 2.7077,
      "theoretical_loss": 3.6276110583917625,
      "tokens_seen": 1064697856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004383109513902024,
      "loss": 2.5011,
      "theoretical_loss": 3.6275900546925257,
      "tokens_seen": 1064763392
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004382944013618309,
      "loss": 2.4433,
      "theoretical_loss": 3.6275690526479747,
      "tokens_seen": 1064828928
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043827785133345945,
      "loss": 2.728,
      "theoretical_loss": 3.627548052257877,
      "tokens_seen": 1064894464
    },
    {
      "epoch": 0.38,
      "objective/train/avg_token_score": 0.03025345504283905,
      "objective/train/avg_weight": 0.9643911719322205,
      "objective/train/docs_used": 606079,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4460737705230713,
      "objective/train/std_weight": 0.11676499992609024,
      "objective/train/theoretical_loss": 3.6275270535220008,
      "objective/train/tokens_used": 1085420000,
      "theoretical_loss": 3.6275270535220008,
      "tokens_seen": 1064960000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000438261301305088,
      "loss": 2.3854,
      "theoretical_loss": 3.6275270535220008,
      "tokens_seen": 1064960000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043824475127671647,
      "loss": 2.6749,
      "theoretical_loss": 3.627506056440113,
      "tokens_seen": 1065025536
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000438228201248345,
      "loss": 2.6714,
      "theoretical_loss": 3.627485061011983,
      "tokens_seen": 1065091072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043821165121997355,
      "loss": 2.8448,
      "theoretical_loss": 3.6274640672373772,
      "tokens_seen": 1065156608
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004381951011916021,
      "loss": 2.7382,
      "theoretical_loss": 3.6274430751160653,
      "tokens_seen": 1065222144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043817855116323057,
      "loss": 2.7425,
      "theoretical_loss": 3.6274220846478142,
      "tokens_seen": 1065287680
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004381620011348591,
      "loss": 2.427,
      "theoretical_loss": 3.627401095832393,
      "tokens_seen": 1065353216
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043814545110648764,
      "loss": 2.6895,
      "theoretical_loss": 3.627380108669569,
      "tokens_seen": 1065418752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004381289010781162,
      "loss": 2.7568,
      "theoretical_loss": 3.6273591231591107,
      "tokens_seen": 1065484288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043811235104974466,
      "loss": 2.4482,
      "theoretical_loss": 3.627338139300787,
      "tokens_seen": 1065549824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004380958010213732,
      "loss": 2.6805,
      "theoretical_loss": 3.627317157094365,
      "tokens_seen": 1065615360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043807925099300174,
      "loss": 2.4077,
      "theoretical_loss": 3.6272961765396143,
      "tokens_seen": 1065680896
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004380627009646303,
      "loss": 2.5673,
      "theoretical_loss": 3.627275197636303,
      "tokens_seen": 1065746432
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043804615093625876,
      "loss": 2.6666,
      "theoretical_loss": 3.627254220384199,
      "tokens_seen": 1065811968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004380296009078873,
      "loss": 2.5769,
      "theoretical_loss": 3.6272332447830715,
      "tokens_seen": 1065877504
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043801305087951583,
      "loss": 2.5606,
      "theoretical_loss": 3.6272122708326884,
      "tokens_seen": 1065943040
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043799650085114437,
      "loss": 2.797,
      "theoretical_loss": 3.6271912985328187,
      "tokens_seen": 1066008576
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004379799508227728,
      "loss": 2.6628,
      "theoretical_loss": 3.6271703278832312,
      "tokens_seen": 1066074112
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043796340079440134,
      "loss": 2.6667,
      "theoretical_loss": 3.6271493588836945,
      "tokens_seen": 1066139648
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004379468507660298,
      "loss": 2.554,
      "theoretical_loss": 3.6271283915339776,
      "tokens_seen": 1066205184
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043793030073765836,
      "loss": 2.79,
      "theoretical_loss": 3.6271074258338487,
      "tokens_seen": 1066270720
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004379137507092869,
      "loss": 2.7366,
      "theoretical_loss": 3.6270864617830765,
      "tokens_seen": 1066336256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043789720068091543,
      "loss": 2.7408,
      "theoretical_loss": 3.627065499381431,
      "tokens_seen": 1066401792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004378806506525439,
      "loss": 2.7257,
      "theoretical_loss": 3.62704453862868,
      "tokens_seen": 1066467328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043786410062417245,
      "loss": 2.7296,
      "theoretical_loss": 3.627023579524593,
      "tokens_seen": 1066532864
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000437847550595801,
      "loss": 2.7874,
      "theoretical_loss": 3.627002622068939,
      "tokens_seen": 1066598400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043783100056742953,
      "loss": 2.6415,
      "theoretical_loss": 3.6269816662614875,
      "tokens_seen": 1066663936
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000437814450539058,
      "loss": 2.8438,
      "theoretical_loss": 3.626960712102006,
      "tokens_seen": 1066729472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043779790051068655,
      "loss": 2.6107,
      "theoretical_loss": 3.6269397595902664,
      "tokens_seen": 1066795008
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004377813504823151,
      "loss": 2.7691,
      "theoretical_loss": 3.6269188087260353,
      "tokens_seen": 1066860544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004377648004539436,
      "loss": 2.6282,
      "theoretical_loss": 3.6268978595090835,
      "tokens_seen": 1066926080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004377482504255721,
      "loss": 2.8954,
      "theoretical_loss": 3.6268769119391795,
      "tokens_seen": 1066991616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043773170039720065,
      "loss": 2.552,
      "theoretical_loss": 3.6268559660160933,
      "tokens_seen": 1067057152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004377151503688292,
      "loss": 2.7511,
      "theoretical_loss": 3.626835021739594,
      "tokens_seen": 1067122688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004376986003404577,
      "loss": 2.8435,
      "theoretical_loss": 3.626814079109451,
      "tokens_seen": 1067188224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004376820503120862,
      "loss": 2.8517,
      "theoretical_loss": 3.6267931381254335,
      "tokens_seen": 1067253760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043766550028371474,
      "loss": 2.7712,
      "theoretical_loss": 3.626772198787312,
      "tokens_seen": 1067319296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004376489502553433,
      "loss": 2.6778,
      "theoretical_loss": 3.6267512610948547,
      "tokens_seen": 1067384832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004376324002269718,
      "loss": 2.5528,
      "theoretical_loss": 3.6267303250478324,
      "tokens_seen": 1067450368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004376158501986003,
      "loss": 2.9441,
      "theoretical_loss": 3.626709390646014,
      "tokens_seen": 1067515904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043759930017022884,
      "loss": 2.7286,
      "theoretical_loss": 3.6266884578891703,
      "tokens_seen": 1067581440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004375827501418574,
      "loss": 2.8518,
      "theoretical_loss": 3.6266675267770703,
      "tokens_seen": 1067646976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004375662001134859,
      "loss": 2.9908,
      "theoretical_loss": 3.6266465973094832,
      "tokens_seen": 1067712512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004375496500851144,
      "loss": 2.859,
      "theoretical_loss": 3.62662566948618,
      "tokens_seen": 1067778048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043753310005674293,
      "loss": 2.5416,
      "theoretical_loss": 3.62660474330693,
      "tokens_seen": 1067843584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043751655002837147,
      "loss": 2.7342,
      "theoretical_loss": 3.626583818771504,
      "tokens_seen": 1067909120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004375,
      "loss": 2.7818,
      "theoretical_loss": 3.6265628958796707,
      "tokens_seen": 1067974656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043748344997162855,
      "loss": 2.647,
      "theoretical_loss": 3.626541974631201,
      "tokens_seen": 1068040192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043746689994325703,
      "loss": 2.5793,
      "theoretical_loss": 3.626521055025864,
      "tokens_seen": 1068105728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043745034991488557,
      "loss": 2.7039,
      "theoretical_loss": 3.6265001370634318,
      "tokens_seen": 1068171264
    },
    {
      "epoch": 0.38,
      "objective/train/avg_token_score": 0.013915282674133778,
      "objective/train/avg_weight": 0.9836201667785645,
      "objective/train/docs_used": 607761,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1923723220825195,
      "objective/train/std_weight": 0.09058886021375656,
      "objective/train/theoretical_loss": 3.626479220743673,
      "objective/train/tokens_used": 1088696800,
      "theoretical_loss": 3.626479220743673,
      "tokens_seen": 1068236800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004374337998865141,
      "loss": 2.817,
      "theoretical_loss": 3.626479220743673,
      "tokens_seen": 1068236800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043741724985814264,
      "loss": 2.642,
      "theoretical_loss": 3.6264583060663584,
      "tokens_seen": 1068302336
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004374006998297711,
      "loss": 2.8804,
      "theoretical_loss": 3.6264373930312575,
      "tokens_seen": 1068367872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043738414980139966,
      "loss": 2.4796,
      "theoretical_loss": 3.6264164816381417,
      "tokens_seen": 1068433408
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004373675997730282,
      "loss": 2.8256,
      "theoretical_loss": 3.626395571886781,
      "tokens_seen": 1068498944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043735104974465674,
      "loss": 2.8932,
      "theoretical_loss": 3.626374663776946,
      "tokens_seen": 1068564480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004373344997162852,
      "loss": 2.7966,
      "theoretical_loss": 3.6263537573084066,
      "tokens_seen": 1068630016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043731794968791376,
      "loss": 2.5767,
      "theoretical_loss": 3.6263328524809335,
      "tokens_seen": 1068695552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004373013996595423,
      "loss": 2.6129,
      "theoretical_loss": 3.6263119492942977,
      "tokens_seen": 1068761088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043728484963117083,
      "loss": 2.8618,
      "theoretical_loss": 3.6262910477482695,
      "tokens_seen": 1068826624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004372682996027993,
      "loss": 2.8681,
      "theoretical_loss": 3.62627014784262,
      "tokens_seen": 1068892160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043725174957442786,
      "loss": 2.4164,
      "theoretical_loss": 3.6262492495771186,
      "tokens_seen": 1068957696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004372351995460564,
      "loss": 2.9333,
      "theoretical_loss": 3.6262283529515376,
      "tokens_seen": 1069023232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043721864951768493,
      "loss": 2.5962,
      "theoretical_loss": 3.6262074579656467,
      "tokens_seen": 1069088768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004372020994893134,
      "loss": 2.7714,
      "theoretical_loss": 3.626186564619217,
      "tokens_seen": 1069154304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043718554946094195,
      "loss": 2.5451,
      "theoretical_loss": 3.6261656729120197,
      "tokens_seen": 1069219840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004371689994325705,
      "loss": 2.637,
      "theoretical_loss": 3.6261447828438254,
      "tokens_seen": 1069285376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000437152449404199,
      "loss": 2.451,
      "theoretical_loss": 3.6261238944144054,
      "tokens_seen": 1069350912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004371358993758275,
      "loss": 2.7699,
      "theoretical_loss": 3.6261030076235303,
      "tokens_seen": 1069416448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043711934934745605,
      "loss": 2.5661,
      "theoretical_loss": 3.6260821224709714,
      "tokens_seen": 1069481984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004371027993190846,
      "loss": 2.6821,
      "theoretical_loss": 3.6260612389565,
      "tokens_seen": 1069547520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004370862492907131,
      "loss": 2.7178,
      "theoretical_loss": 3.626040357079887,
      "tokens_seen": 1069613056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004370696992623416,
      "loss": 2.6231,
      "theoretical_loss": 3.626019476840904,
      "tokens_seen": 1069678592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043705314923397014,
      "loss": 2.6858,
      "theoretical_loss": 3.6259985982393217,
      "tokens_seen": 1069744128
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004370365992055987,
      "loss": 2.755,
      "theoretical_loss": 3.6259777212749116,
      "tokens_seen": 1069809664
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004370200491772272,
      "loss": 2.8908,
      "theoretical_loss": 3.625956845947445,
      "tokens_seen": 1069875200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043700349914885565,
      "loss": 2.6783,
      "theoretical_loss": 3.625935972256694,
      "tokens_seen": 1069940736
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004369869491204842,
      "loss": 2.7699,
      "theoretical_loss": 3.6259151002024286,
      "tokens_seen": 1070006272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043697039909211267,
      "loss": 2.8509,
      "theoretical_loss": 3.6258942297844214,
      "tokens_seen": 1070071808
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004369538490637412,
      "loss": 2.7025,
      "theoretical_loss": 3.6258733610024434,
      "tokens_seen": 1070137344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043693729903536974,
      "loss": 2.7681,
      "theoretical_loss": 3.6258524938562666,
      "tokens_seen": 1070202880
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004369207490069983,
      "loss": 2.5899,
      "theoretical_loss": 3.6258316283456624,
      "tokens_seen": 1070268416
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043690419897862676,
      "loss": 2.7631,
      "theoretical_loss": 3.625810764470402,
      "tokens_seen": 1070333952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004368876489502553,
      "loss": 2.7103,
      "theoretical_loss": 3.6257899022302578,
      "tokens_seen": 1070399488
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043687109892188384,
      "loss": 2.6052,
      "theoretical_loss": 3.6257690416250012,
      "tokens_seen": 1070465024
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004368545488935124,
      "loss": 2.6946,
      "theoretical_loss": 3.6257481826544047,
      "tokens_seen": 1070530560
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043683799886514086,
      "loss": 2.8897,
      "theoretical_loss": 3.6257273253182394,
      "tokens_seen": 1070596096
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004368214488367694,
      "loss": 2.8869,
      "theoretical_loss": 3.6257064696162766,
      "tokens_seen": 1070661632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043680489880839794,
      "loss": 2.5304,
      "theoretical_loss": 3.62568561554829,
      "tokens_seen": 1070727168
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004367883487800265,
      "loss": 2.8192,
      "theoretical_loss": 3.6256647631140497,
      "tokens_seen": 1070792704
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043677179875165496,
      "loss": 2.586,
      "theoretical_loss": 3.625643912313329,
      "tokens_seen": 1070858240
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004367552487232835,
      "loss": 2.5281,
      "theoretical_loss": 3.6256230631458997,
      "tokens_seen": 1070923776
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043673869869491203,
      "loss": 2.8457,
      "theoretical_loss": 3.6256022156115337,
      "tokens_seen": 1070989312
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043672214866654057,
      "loss": 2.6066,
      "theoretical_loss": 3.6255813697100034,
      "tokens_seen": 1071054848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043670559863816905,
      "loss": 2.8211,
      "theoretical_loss": 3.625560525441081,
      "tokens_seen": 1071120384
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004366890486097976,
      "loss": 2.61,
      "theoretical_loss": 3.625539682804538,
      "tokens_seen": 1071185920
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043667249858142613,
      "loss": 2.7522,
      "theoretical_loss": 3.6255188418001483,
      "tokens_seen": 1071251456
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043665594855305466,
      "loss": 2.5019,
      "theoretical_loss": 3.6254980024276824,
      "tokens_seen": 1071316992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043663939852468315,
      "loss": 2.8768,
      "theoretical_loss": 3.625477164686914,
      "tokens_seen": 1071382528
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004366228484963117,
      "loss": 2.7751,
      "theoretical_loss": 3.625456328577615,
      "tokens_seen": 1071448064
    },
    {
      "epoch": 0.38,
      "objective/train/avg_token_score": 0.01613588258624077,
      "objective/train/avg_weight": 0.981002151966095,
      "objective/train/docs_used": 609849,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.732696533203125,
      "objective/train/std_weight": 0.10503644496202469,
      "objective/train/theoretical_loss": 3.6254354940995586,
      "objective/train/tokens_used": 1091973600,
      "theoretical_loss": 3.6254354940995586,
      "tokens_seen": 1071513600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004366062984679402,
      "loss": 2.8427,
      "theoretical_loss": 3.6254354940995586,
      "tokens_seen": 1071513600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043658974843956876,
      "loss": 2.5432,
      "theoretical_loss": 3.625414661252516,
      "tokens_seen": 1071579136
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043657319841119724,
      "loss": 2.6863,
      "theoretical_loss": 3.625393830036261,
      "tokens_seen": 1071644672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004365566483828258,
      "loss": 2.7783,
      "theoretical_loss": 3.6253730004505655,
      "tokens_seen": 1071710208
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004365400983544543,
      "loss": 2.7934,
      "theoretical_loss": 3.6253521724952025,
      "tokens_seen": 1071775744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043652354832608286,
      "loss": 2.6188,
      "theoretical_loss": 3.6253313461699452,
      "tokens_seen": 1071841280
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043650699829771134,
      "loss": 2.7058,
      "theoretical_loss": 3.6253105214745647,
      "tokens_seen": 1071906816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004364904482693399,
      "loss": 2.691,
      "theoretical_loss": 3.625289698408836,
      "tokens_seen": 1071972352
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004364738982409684,
      "loss": 2.8204,
      "theoretical_loss": 3.6252688769725303,
      "tokens_seen": 1072037888
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043645734821259695,
      "loss": 2.6001,
      "theoretical_loss": 3.625248057165421,
      "tokens_seen": 1072103424
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043644079818422544,
      "loss": 2.7454,
      "theoretical_loss": 3.6252272389872813,
      "tokens_seen": 1072168960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000436424248155854,
      "loss": 2.5049,
      "theoretical_loss": 3.625206422437884,
      "tokens_seen": 1072234496
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004364076981274825,
      "loss": 2.7657,
      "theoretical_loss": 3.625185607517002,
      "tokens_seen": 1072300032
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043639114809911105,
      "loss": 2.8277,
      "theoretical_loss": 3.6251647942244087,
      "tokens_seen": 1072365568
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043637459807073953,
      "loss": 2.8278,
      "theoretical_loss": 3.625143982559877,
      "tokens_seen": 1072431104
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043635804804236807,
      "loss": 2.8436,
      "theoretical_loss": 3.62512317252318,
      "tokens_seen": 1072496640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004363414980139966,
      "loss": 2.7765,
      "theoretical_loss": 3.625102364114091,
      "tokens_seen": 1072562176
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043632494798562514,
      "loss": 2.6584,
      "theoretical_loss": 3.6250815573323836,
      "tokens_seen": 1072627712
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043630839795725363,
      "loss": 2.4697,
      "theoretical_loss": 3.6250607521778306,
      "tokens_seen": 1072693248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043629184792888217,
      "loss": 2.7556,
      "theoretical_loss": 3.6250399486502056,
      "tokens_seen": 1072758784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004362752979005107,
      "loss": 2.5415,
      "theoretical_loss": 3.6250191467492816,
      "tokens_seen": 1072824320
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043625874787213924,
      "loss": 2.4916,
      "theoretical_loss": 3.6249983464748325,
      "tokens_seen": 1072889856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004362421978437677,
      "loss": 2.8464,
      "theoretical_loss": 3.624977547826632,
      "tokens_seen": 1072955392
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043622564781539626,
      "loss": 2.5951,
      "theoretical_loss": 3.6249567508044533,
      "tokens_seen": 1073020928
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004362090977870248,
      "loss": 2.6934,
      "theoretical_loss": 3.62493595540807,
      "tokens_seen": 1073086464
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043619254775865334,
      "loss": 2.8139,
      "theoretical_loss": 3.6249151616372552,
      "tokens_seen": 1073152000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004361759977302818,
      "loss": 2.7664,
      "theoretical_loss": 3.6248943694917832,
      "tokens_seen": 1073217536
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043615944770191036,
      "loss": 2.7596,
      "theoretical_loss": 3.624873578971428,
      "tokens_seen": 1073283072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004361428976735389,
      "loss": 2.7968,
      "theoretical_loss": 3.6248527900759626,
      "tokens_seen": 1073348608
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043612634764516743,
      "loss": 2.5566,
      "theoretical_loss": 3.6248320028051615,
      "tokens_seen": 1073414144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004361097976167959,
      "loss": 2.6096,
      "theoretical_loss": 3.624811217158798,
      "tokens_seen": 1073479680
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043609324758842445,
      "loss": 2.7264,
      "theoretical_loss": 3.6247904331366465,
      "tokens_seen": 1073545216
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000436076697560053,
      "loss": 2.6324,
      "theoretical_loss": 3.6247696507384806,
      "tokens_seen": 1073610752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043606014753168153,
      "loss": 2.7014,
      "theoretical_loss": 3.6247488699640735,
      "tokens_seen": 1073676288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043604359750331,
      "loss": 2.7346,
      "theoretical_loss": 3.6247280908132007,
      "tokens_seen": 1073741824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043602704747493855,
      "loss": 2.7165,
      "theoretical_loss": 3.6247073132856364,
      "tokens_seen": 1073807360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043601049744656703,
      "loss": 2.7603,
      "theoretical_loss": 3.624686537381153,
      "tokens_seen": 1073872896
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004359939474181955,
      "loss": 2.8635,
      "theoretical_loss": 3.624665763099526,
      "tokens_seen": 1073938432
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043597739738982405,
      "loss": 2.8137,
      "theoretical_loss": 3.6246449904405287,
      "tokens_seen": 1074003968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004359608473614526,
      "loss": 2.6389,
      "theoretical_loss": 3.6246242194039366,
      "tokens_seen": 1074069504
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043594429733308113,
      "loss": 2.833,
      "theoretical_loss": 3.6246034499895234,
      "tokens_seen": 1074135040
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004359277473047096,
      "loss": 2.6244,
      "theoretical_loss": 3.6245826821970626,
      "tokens_seen": 1074200576
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043591119727633815,
      "loss": 2.8731,
      "theoretical_loss": 3.6245619160263303,
      "tokens_seen": 1074266112
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004358946472479667,
      "loss": 2.8322,
      "theoretical_loss": 3.624541151477099,
      "tokens_seen": 1074331648
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004358780972195952,
      "loss": 2.827,
      "theoretical_loss": 3.6245203885491453,
      "tokens_seen": 1074397184
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004358615471912237,
      "loss": 2.6792,
      "theoretical_loss": 3.6244996272422414,
      "tokens_seen": 1074462720
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043584499716285225,
      "loss": 2.6031,
      "theoretical_loss": 3.624478867556164,
      "tokens_seen": 1074528256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004358284471344808,
      "loss": 2.8008,
      "theoretical_loss": 3.6244581094906865,
      "tokens_seen": 1074593792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004358118971061093,
      "loss": 2.5932,
      "theoretical_loss": 3.6244373530455842,
      "tokens_seen": 1074659328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004357953470777378,
      "loss": 2.4676,
      "theoretical_loss": 3.624416598220631,
      "tokens_seen": 1074724864
    },
    {
      "epoch": 0.38,
      "objective/train/avg_token_score": 0.011467177420854568,
      "objective/train/avg_weight": 0.9865047335624695,
      "objective/train/docs_used": 611858,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9882326126098633,
      "objective/train/std_weight": 0.06434530764818192,
      "objective/train/theoretical_loss": 3.624395845015602,
      "objective/train/tokens_used": 1095250400,
      "theoretical_loss": 3.624395845015602,
      "tokens_seen": 1074790400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043577879704936634,
      "loss": 2.8133,
      "theoretical_loss": 3.624395845015602,
      "tokens_seen": 1074790400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004357622470209949,
      "loss": 2.9461,
      "theoretical_loss": 3.624375093430273,
      "tokens_seen": 1074855936
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004357456969926234,
      "loss": 2.5262,
      "theoretical_loss": 3.624354343464417,
      "tokens_seen": 1074921472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004357291469642519,
      "loss": 2.6662,
      "theoretical_loss": 3.62433359511781,
      "tokens_seen": 1074987008
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043571259693588044,
      "loss": 2.5293,
      "theoretical_loss": 3.6243128483902276,
      "tokens_seen": 1075052544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000435696046907509,
      "loss": 2.5232,
      "theoretical_loss": 3.6242921032814435,
      "tokens_seen": 1075118080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004356794968791375,
      "loss": 2.7149,
      "theoretical_loss": 3.6242713597912335,
      "tokens_seen": 1075183616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000435662946850766,
      "loss": 2.4735,
      "theoretical_loss": 3.6242506179193716,
      "tokens_seen": 1075249152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043564639682239453,
      "loss": 2.7749,
      "theoretical_loss": 3.624229877665634,
      "tokens_seen": 1075314688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043562984679402307,
      "loss": 2.722,
      "theoretical_loss": 3.624209139029796,
      "tokens_seen": 1075380224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004356132967656516,
      "loss": 2.6194,
      "theoretical_loss": 3.624188402011632,
      "tokens_seen": 1075445760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004355967467372801,
      "loss": 2.7608,
      "theoretical_loss": 3.624167666610918,
      "tokens_seen": 1075511296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043558019670890863,
      "loss": 2.7068,
      "theoretical_loss": 3.6241469328274283,
      "tokens_seen": 1075576832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043556364668053717,
      "loss": 2.6788,
      "theoretical_loss": 3.6241262006609394,
      "tokens_seen": 1075642368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004355470966521657,
      "loss": 2.7475,
      "theoretical_loss": 3.6241054701112256,
      "tokens_seen": 1075707904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004355305466237942,
      "loss": 2.5747,
      "theoretical_loss": 3.6240847411780637,
      "tokens_seen": 1075773440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004355139965954227,
      "loss": 2.8111,
      "theoretical_loss": 3.6240640138612275,
      "tokens_seen": 1075838976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043549744656705126,
      "loss": 2.6028,
      "theoretical_loss": 3.6240432881604936,
      "tokens_seen": 1075904512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004354808965386798,
      "loss": 2.5343,
      "theoretical_loss": 3.6240225640756374,
      "tokens_seen": 1075970048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004354643465103083,
      "loss": 2.8208,
      "theoretical_loss": 3.624001841606434,
      "tokens_seen": 1076035584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004354477964819368,
      "loss": 2.6603,
      "theoretical_loss": 3.62398112075266,
      "tokens_seen": 1076101120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043543124645356536,
      "loss": 2.6563,
      "theoretical_loss": 3.6239604015140907,
      "tokens_seen": 1076166656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004354146964251939,
      "loss": 2.9344,
      "theoretical_loss": 3.6239396838905016,
      "tokens_seen": 1076232192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004353981463968224,
      "loss": 2.4182,
      "theoretical_loss": 3.623918967881668,
      "tokens_seen": 1076297728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004353815963684509,
      "loss": 2.5584,
      "theoretical_loss": 3.6238982534873667,
      "tokens_seen": 1076363264
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043536504634007945,
      "loss": 2.6237,
      "theoretical_loss": 3.6238775407073733,
      "tokens_seen": 1076428800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000435348496311708,
      "loss": 2.6366,
      "theoretical_loss": 3.623856829541464,
      "tokens_seen": 1076494336
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004353319462833365,
      "loss": 2.5262,
      "theoretical_loss": 3.623836119989414,
      "tokens_seen": 1076559872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000435315396254965,
      "loss": 2.7386,
      "theoretical_loss": 3.6238154120509996,
      "tokens_seen": 1076625408
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043529884622659355,
      "loss": 2.7017,
      "theoretical_loss": 3.623794705725997,
      "tokens_seen": 1076690944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004352822961982221,
      "loss": 2.694,
      "theoretical_loss": 3.6237740010141826,
      "tokens_seen": 1076756480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043526574616985057,
      "loss": 2.6011,
      "theoretical_loss": 3.6237532979153317,
      "tokens_seen": 1076822016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004352491961414791,
      "loss": 2.5989,
      "theoretical_loss": 3.6237325964292215,
      "tokens_seen": 1076887552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043523264611310765,
      "loss": 2.8157,
      "theoretical_loss": 3.6237118965556276,
      "tokens_seen": 1076953088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004352160960847362,
      "loss": 2.6974,
      "theoretical_loss": 3.6236911982943267,
      "tokens_seen": 1077018624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043519954605636467,
      "loss": 2.6317,
      "theoretical_loss": 3.623670501645094,
      "tokens_seen": 1077084160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004351829960279932,
      "loss": 2.4997,
      "theoretical_loss": 3.623649806607707,
      "tokens_seen": 1077149696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043516644599962174,
      "loss": 2.9376,
      "theoretical_loss": 3.6236291131819427,
      "tokens_seen": 1077215232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004351498959712503,
      "loss": 2.6131,
      "theoretical_loss": 3.623608421367576,
      "tokens_seen": 1077280768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043513334594287876,
      "loss": 2.7146,
      "theoretical_loss": 3.623587731164384,
      "tokens_seen": 1077346304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004351167959145073,
      "loss": 2.644,
      "theoretical_loss": 3.6235670425721436,
      "tokens_seen": 1077411840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043510024588613584,
      "loss": 2.876,
      "theoretical_loss": 3.623546355590631,
      "tokens_seen": 1077477376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004350836958577644,
      "loss": 2.6364,
      "theoretical_loss": 3.623525670219623,
      "tokens_seen": 1077542912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043506714582939286,
      "loss": 2.6044,
      "theoretical_loss": 3.6235049864588964,
      "tokens_seen": 1077608448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004350505958010214,
      "loss": 2.472,
      "theoretical_loss": 3.623484304308228,
      "tokens_seen": 1077673984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004350340457726498,
      "loss": 2.7054,
      "theoretical_loss": 3.6234636237673934,
      "tokens_seen": 1077739520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043501749574427836,
      "loss": 2.5317,
      "theoretical_loss": 3.623442944836171,
      "tokens_seen": 1077805056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0004350009457159069,
      "loss": 2.7274,
      "theoretical_loss": 3.623422267514337,
      "tokens_seen": 1077870592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00043498439568753544,
      "loss": 2.8742,
      "theoretical_loss": 3.6234015918016684,
      "tokens_seen": 1077936128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000434967845659164,
      "loss": 2.6083,
      "theoretical_loss": 3.6233809176979417,
      "tokens_seen": 1078001664
    },
    {
      "epoch": 0.39,
      "objective/train/avg_token_score": 0.024254929274320602,
      "objective/train/avg_weight": 0.9714846611022949,
      "objective/train/docs_used": 613931,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.557846784591675,
      "objective/train/std_weight": 0.09766264259815216,
      "objective/train/theoretical_loss": 3.6233602452029348,
      "objective/train/tokens_used": 1098527200,
      "theoretical_loss": 3.6233602452029348,
      "tokens_seen": 1078067200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043495129563079246,
      "loss": 2.49,
      "theoretical_loss": 3.6233602452029348,
      "tokens_seen": 1078067200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000434934745602421,
      "loss": 2.63,
      "theoretical_loss": 3.6233395743164243,
      "tokens_seen": 1078132736
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043491819557404953,
      "loss": 2.5981,
      "theoretical_loss": 3.623318905038187,
      "tokens_seen": 1078198272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043490164554567807,
      "loss": 2.775,
      "theoretical_loss": 3.6232982373679996,
      "tokens_seen": 1078263808
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043488509551730656,
      "loss": 2.6901,
      "theoretical_loss": 3.6232775713056404,
      "tokens_seen": 1078329344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004348685454889351,
      "loss": 2.6555,
      "theoretical_loss": 3.623256906850886,
      "tokens_seen": 1078394880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043485199546056363,
      "loss": 2.8077,
      "theoretical_loss": 3.6232362440035137,
      "tokens_seen": 1078460416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043483544543219217,
      "loss": 2.562,
      "theoretical_loss": 3.623215582763301,
      "tokens_seen": 1078525952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043481889540382065,
      "loss": 2.4753,
      "theoretical_loss": 3.623194923130025,
      "tokens_seen": 1078591488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004348023453754492,
      "loss": 2.7457,
      "theoretical_loss": 3.6231742651034633,
      "tokens_seen": 1078657024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004347857953470777,
      "loss": 2.7259,
      "theoretical_loss": 3.623153608683393,
      "tokens_seen": 1078722560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043476924531870626,
      "loss": 2.774,
      "theoretical_loss": 3.623132953869592,
      "tokens_seen": 1078788096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043475269529033475,
      "loss": 2.5184,
      "theoretical_loss": 3.6231123006618375,
      "tokens_seen": 1078853632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004347361452619633,
      "loss": 2.6007,
      "theoretical_loss": 3.6230916490599077,
      "tokens_seen": 1078919168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004347195952335918,
      "loss": 2.7267,
      "theoretical_loss": 3.623070999063579,
      "tokens_seen": 1078984704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043470304520522036,
      "loss": 2.6794,
      "theoretical_loss": 3.6230503506726306,
      "tokens_seen": 1079050240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043468649517684884,
      "loss": 2.9374,
      "theoretical_loss": 3.623029703886839,
      "tokens_seen": 1079115776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004346699451484774,
      "loss": 2.748,
      "theoretical_loss": 3.623009058705982,
      "tokens_seen": 1079181312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004346533951201059,
      "loss": 2.6727,
      "theoretical_loss": 3.622988415129838,
      "tokens_seen": 1079246848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043463684509173446,
      "loss": 2.8153,
      "theoretical_loss": 3.6229677731581846,
      "tokens_seen": 1079312384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043462029506336294,
      "loss": 2.8212,
      "theoretical_loss": 3.6229471327907996,
      "tokens_seen": 1079377920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004346037450349915,
      "loss": 2.732,
      "theoretical_loss": 3.6229264940274613,
      "tokens_seen": 1079443456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043458719500662,
      "loss": 2.8348,
      "theoretical_loss": 3.6229058568679466,
      "tokens_seen": 1079508992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043457064497824855,
      "loss": 2.6419,
      "theoretical_loss": 3.6228852213120346,
      "tokens_seen": 1079574528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043455409494987704,
      "loss": 2.8149,
      "theoretical_loss": 3.622864587359503,
      "tokens_seen": 1079640064
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004345375449215056,
      "loss": 2.7328,
      "theoretical_loss": 3.62284395501013,
      "tokens_seen": 1079705600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004345209948931341,
      "loss": 2.5449,
      "theoretical_loss": 3.622823324263694,
      "tokens_seen": 1079771136
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043450444486476265,
      "loss": 2.6161,
      "theoretical_loss": 3.6228026951199723,
      "tokens_seen": 1079836672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043448789483639113,
      "loss": 2.6052,
      "theoretical_loss": 3.6227820675787443,
      "tokens_seen": 1079902208
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043447134480801967,
      "loss": 2.7119,
      "theoretical_loss": 3.6227614416397866,
      "tokens_seen": 1079967744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004344547947796482,
      "loss": 2.7531,
      "theoretical_loss": 3.6227408173028794,
      "tokens_seen": 1080033280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043443824475127674,
      "loss": 2.708,
      "theoretical_loss": 3.6227201945678003,
      "tokens_seen": 1080098816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043442169472290523,
      "loss": 2.569,
      "theoretical_loss": 3.6226995734343275,
      "tokens_seen": 1080164352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043440514469453377,
      "loss": 2.5359,
      "theoretical_loss": 3.6226789539022395,
      "tokens_seen": 1080229888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004343885946661623,
      "loss": 2.7695,
      "theoretical_loss": 3.6226583359713147,
      "tokens_seen": 1080295424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043437204463779084,
      "loss": 2.7498,
      "theoretical_loss": 3.622637719641332,
      "tokens_seen": 1080360960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004343554946094193,
      "loss": 2.5767,
      "theoretical_loss": 3.6226171049120697,
      "tokens_seen": 1080426496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043433894458104786,
      "loss": 2.6948,
      "theoretical_loss": 3.6225964917833067,
      "tokens_seen": 1080492032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004343223945526764,
      "loss": 2.5944,
      "theoretical_loss": 3.6225758802548214,
      "tokens_seen": 1080557568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043430584452430494,
      "loss": 2.6837,
      "theoretical_loss": 3.622555270326393,
      "tokens_seen": 1080623104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004342892944959334,
      "loss": 2.6194,
      "theoretical_loss": 3.6225346619977996,
      "tokens_seen": 1080688640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043427274446756196,
      "loss": 2.8033,
      "theoretical_loss": 3.6225140552688204,
      "tokens_seen": 1080754176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004342561944391905,
      "loss": 2.5233,
      "theoretical_loss": 3.6224934501392343,
      "tokens_seen": 1080819712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043423964441081903,
      "loss": 2.7419,
      "theoretical_loss": 3.62247284660882,
      "tokens_seen": 1080885248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004342230943824475,
      "loss": 2.6512,
      "theoretical_loss": 3.6224522446773566,
      "tokens_seen": 1080950784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043420654435407605,
      "loss": 2.7819,
      "theoretical_loss": 3.6224316443446227,
      "tokens_seen": 1081016320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004341899943257046,
      "loss": 2.7837,
      "theoretical_loss": 3.6224110456103977,
      "tokens_seen": 1081081856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043417344429733313,
      "loss": 2.7168,
      "theoretical_loss": 3.622390448474461,
      "tokens_seen": 1081147392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004341568942689616,
      "loss": 2.3567,
      "theoretical_loss": 3.6223698529365906,
      "tokens_seen": 1081212928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043414034424059015,
      "loss": 2.6164,
      "theoretical_loss": 3.622349258996567,
      "tokens_seen": 1081278464
    },
    {
      "epoch": 0.39,
      "objective/train/avg_token_score": 0.013294922187924385,
      "objective/train/avg_weight": 0.9843488931655884,
      "objective/train/docs_used": 615851,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0584399700164795,
      "objective/train/std_weight": 0.07720416784286499,
      "objective/train/theoretical_loss": 3.6223286666541683,
      "objective/train/tokens_used": 1101804000,
      "theoretical_loss": 3.6223286666541683,
      "tokens_seen": 1081344000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004341237942122187,
      "loss": 2.7515,
      "theoretical_loss": 3.6223286666541683,
      "tokens_seen": 1081344000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004341072441838472,
      "loss": 2.9317,
      "theoretical_loss": 3.6223080759091744,
      "tokens_seen": 1081409536
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004340906941554757,
      "loss": 2.9872,
      "theoretical_loss": 3.6222874867613646,
      "tokens_seen": 1081475072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043407414412710424,
      "loss": 2.4569,
      "theoretical_loss": 3.6222668992105174,
      "tokens_seen": 1081540608
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004340575940987327,
      "loss": 2.6608,
      "theoretical_loss": 3.6222463132564133,
      "tokens_seen": 1081606144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004340410440703612,
      "loss": 2.6369,
      "theoretical_loss": 3.6222257288988313,
      "tokens_seen": 1081671680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043402449404198975,
      "loss": 2.6926,
      "theoretical_loss": 3.62220514613755,
      "tokens_seen": 1081737216
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004340079440136183,
      "loss": 2.5918,
      "theoretical_loss": 3.6221845649723505,
      "tokens_seen": 1081802752
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043399139398524677,
      "loss": 2.7336,
      "theoretical_loss": 3.6221639854030117,
      "tokens_seen": 1081868288
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004339748439568753,
      "loss": 2.7833,
      "theoretical_loss": 3.6221434074293124,
      "tokens_seen": 1081933824
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043395829392850385,
      "loss": 2.6872,
      "theoretical_loss": 3.6221228310510334,
      "tokens_seen": 1081999360
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004339417439001324,
      "loss": 2.7879,
      "theoretical_loss": 3.6221022562679535,
      "tokens_seen": 1082064896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043392519387176087,
      "loss": 2.8301,
      "theoretical_loss": 3.6220816830798532,
      "tokens_seen": 1082130432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004339086438433894,
      "loss": 2.8236,
      "theoretical_loss": 3.6220611114865116,
      "tokens_seen": 1082195968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043389209381501794,
      "loss": 2.604,
      "theoretical_loss": 3.622040541487709,
      "tokens_seen": 1082261504
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004338755437866465,
      "loss": 2.6034,
      "theoretical_loss": 3.6220199730832245,
      "tokens_seen": 1082327040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043385899375827496,
      "loss": 2.8369,
      "theoretical_loss": 3.621999406272839,
      "tokens_seen": 1082392576
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004338424437299035,
      "loss": 2.6165,
      "theoretical_loss": 3.6219788410563316,
      "tokens_seen": 1082458112
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043382589370153204,
      "loss": 2.2334,
      "theoretical_loss": 3.6219582774334835,
      "tokens_seen": 1082523648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004338093436731606,
      "loss": 2.5937,
      "theoretical_loss": 3.6219377154040733,
      "tokens_seen": 1082589184
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043379279364478906,
      "loss": 2.8343,
      "theoretical_loss": 3.621917154967882,
      "tokens_seen": 1082654720
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004337762436164176,
      "loss": 2.7885,
      "theoretical_loss": 3.6218965961246887,
      "tokens_seen": 1082720256
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043375969358804613,
      "loss": 2.5266,
      "theoretical_loss": 3.621876038874275,
      "tokens_seen": 1082785792
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043374314355967467,
      "loss": 2.6943,
      "theoretical_loss": 3.6218554832164203,
      "tokens_seen": 1082851328
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043372659353130315,
      "loss": 2.7084,
      "theoretical_loss": 3.621834929150905,
      "tokens_seen": 1082916864
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004337100435029317,
      "loss": 2.7145,
      "theoretical_loss": 3.6218143766775093,
      "tokens_seen": 1082982400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043369349347456023,
      "loss": 2.8798,
      "theoretical_loss": 3.621793825796013,
      "tokens_seen": 1083047936
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043367694344618877,
      "loss": 2.5459,
      "theoretical_loss": 3.6217732765061976,
      "tokens_seen": 1083113472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043366039341781725,
      "loss": 2.689,
      "theoretical_loss": 3.621752728807843,
      "tokens_seen": 1083179008
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004336438433894458,
      "loss": 2.5794,
      "theoretical_loss": 3.62173218270073,
      "tokens_seen": 1083244544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004336272933610743,
      "loss": 2.8733,
      "theoretical_loss": 3.621711638184638,
      "tokens_seen": 1083310080
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043361074333270286,
      "loss": 2.8268,
      "theoretical_loss": 3.6216910952593486,
      "tokens_seen": 1083375616
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043359419330433135,
      "loss": 2.9034,
      "theoretical_loss": 3.621670553924642,
      "tokens_seen": 1083441152
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004335776432759599,
      "loss": 2.871,
      "theoretical_loss": 3.6216500141802994,
      "tokens_seen": 1083506688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004335610932475884,
      "loss": 2.7953,
      "theoretical_loss": 3.6216294760261007,
      "tokens_seen": 1083572224
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043354454321921696,
      "loss": 2.6027,
      "theoretical_loss": 3.621608939461827,
      "tokens_seen": 1083637760
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043352799319084544,
      "loss": 2.8701,
      "theoretical_loss": 3.6215884044872593,
      "tokens_seen": 1083703296
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000433511443162474,
      "loss": 2.6042,
      "theoretical_loss": 3.621567871102178,
      "tokens_seen": 1083768832
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004334948931341025,
      "loss": 2.9149,
      "theoretical_loss": 3.6215473393063644,
      "tokens_seen": 1083834368
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043347834310573105,
      "loss": 2.649,
      "theoretical_loss": 3.621526809099599,
      "tokens_seen": 1083899904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004334617930773596,
      "loss": 2.785,
      "theoretical_loss": 3.621506280481663,
      "tokens_seen": 1083965440
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004334452430489881,
      "loss": 2.5076,
      "theoretical_loss": 3.621485753452337,
      "tokens_seen": 1084030976
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004334286930206166,
      "loss": 2.9725,
      "theoretical_loss": 3.6214652280114024,
      "tokens_seen": 1084096512
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043341214299224515,
      "loss": 2.8838,
      "theoretical_loss": 3.62144470415864,
      "tokens_seen": 1084162048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004333955929638737,
      "loss": 2.6603,
      "theoretical_loss": 3.6214241818938318,
      "tokens_seen": 1084227584
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043337904293550217,
      "loss": 2.8294,
      "theoretical_loss": 3.621403661216758,
      "tokens_seen": 1084293120
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004333624929071307,
      "loss": 2.6165,
      "theoretical_loss": 3.6213831421272005,
      "tokens_seen": 1084358656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043334594287875925,
      "loss": 2.3548,
      "theoretical_loss": 3.6213626246249393,
      "tokens_seen": 1084424192
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004333293928503878,
      "loss": 2.8339,
      "theoretical_loss": 3.621342108709757,
      "tokens_seen": 1084489728
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043331284282201627,
      "loss": 2.6281,
      "theoretical_loss": 3.6213215943814347,
      "tokens_seen": 1084555264
    },
    {
      "epoch": 0.39,
      "objective/train/avg_token_score": 0.014173408038914204,
      "objective/train/avg_weight": 0.9833377003669739,
      "objective/train/docs_used": 617294,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.745293140411377,
      "objective/train/std_weight": 0.06035545840859413,
      "objective/train/theoretical_loss": 3.621301081639753,
      "objective/train/tokens_used": 1105080800,
      "theoretical_loss": 3.621301081639753,
      "tokens_seen": 1084620800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004332962927936448,
      "loss": 2.5664,
      "theoretical_loss": 3.621301081639753,
      "tokens_seen": 1084620800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043327974276527334,
      "loss": 2.6851,
      "theoretical_loss": 3.621280570484494,
      "tokens_seen": 1084686336
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004332631927369019,
      "loss": 2.7177,
      "theoretical_loss": 3.6212600609154393,
      "tokens_seen": 1084751872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043324664270853036,
      "loss": 3.0356,
      "theoretical_loss": 3.62123955293237,
      "tokens_seen": 1084817408
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004332300926801589,
      "loss": 2.8463,
      "theoretical_loss": 3.621219046535068,
      "tokens_seen": 1084882944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043321354265178744,
      "loss": 2.5672,
      "theoretical_loss": 3.6211985417233157,
      "tokens_seen": 1084948480
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000433196992623416,
      "loss": 2.8132,
      "theoretical_loss": 3.6211780384968923,
      "tokens_seen": 1085014016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043318044259504446,
      "loss": 2.8076,
      "theoretical_loss": 3.6211575368555815,
      "tokens_seen": 1085079552
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000433163892566673,
      "loss": 2.9865,
      "theoretical_loss": 3.6211370367991647,
      "tokens_seen": 1085145088
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043314734253830153,
      "loss": 2.7687,
      "theoretical_loss": 3.6211165383274233,
      "tokens_seen": 1085210624
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043313079250993007,
      "loss": 2.7689,
      "theoretical_loss": 3.621096041440139,
      "tokens_seen": 1085276160
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043311424248155856,
      "loss": 2.8316,
      "theoretical_loss": 3.621075546137094,
      "tokens_seen": 1085341696
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004330976924531871,
      "loss": 2.6648,
      "theoretical_loss": 3.621055052418071,
      "tokens_seen": 1085407232
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043308114242481563,
      "loss": 2.8285,
      "theoretical_loss": 3.62103456028285,
      "tokens_seen": 1085472768
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043306459239644406,
      "loss": 2.8148,
      "theoretical_loss": 3.6210140697312143,
      "tokens_seen": 1085538304
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004330480423680726,
      "loss": 2.8113,
      "theoretical_loss": 3.6209935807629465,
      "tokens_seen": 1085603840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043303149233970113,
      "loss": 2.8412,
      "theoretical_loss": 3.620973093377827,
      "tokens_seen": 1085669376
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004330149423113296,
      "loss": 2.6965,
      "theoretical_loss": 3.620952607575639,
      "tokens_seen": 1085734912
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043299839228295816,
      "loss": 2.8275,
      "theoretical_loss": 3.6209321233561647,
      "tokens_seen": 1085800448
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004329818422545867,
      "loss": 2.4969,
      "theoretical_loss": 3.6209116407191857,
      "tokens_seen": 1085865984
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043296529222621523,
      "loss": 2.6958,
      "theoretical_loss": 3.620891159664485,
      "tokens_seen": 1085931520
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004329487421978437,
      "loss": 2.6371,
      "theoretical_loss": 3.620870680191844,
      "tokens_seen": 1085997056
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043293219216947225,
      "loss": 2.7111,
      "theoretical_loss": 3.620850202301046,
      "tokens_seen": 1086062592
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004329156421411008,
      "loss": 2.8128,
      "theoretical_loss": 3.620829725991873,
      "tokens_seen": 1086128128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004328990921127293,
      "loss": 2.654,
      "theoretical_loss": 3.620809251264107,
      "tokens_seen": 1086193664
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004328825420843578,
      "loss": 2.7584,
      "theoretical_loss": 3.6207887781175305,
      "tokens_seen": 1086259200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043286599205598635,
      "loss": 2.7735,
      "theoretical_loss": 3.620768306551927,
      "tokens_seen": 1086324736
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004328494420276149,
      "loss": 2.8219,
      "theoretical_loss": 3.620747836567078,
      "tokens_seen": 1086390272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004328328919992434,
      "loss": 2.8845,
      "theoretical_loss": 3.6207273681627665,
      "tokens_seen": 1086455808
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004328163419708719,
      "loss": 2.8795,
      "theoretical_loss": 3.620706901338775,
      "tokens_seen": 1086521344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043279979194250044,
      "loss": 2.7015,
      "theoretical_loss": 3.6206864360948865,
      "tokens_seen": 1086586880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000432783241914129,
      "loss": 2.9378,
      "theoretical_loss": 3.620665972430883,
      "tokens_seen": 1086652416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004327666918857575,
      "loss": 2.554,
      "theoretical_loss": 3.620645510346548,
      "tokens_seen": 1086717952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000432750141857386,
      "loss": 2.534,
      "theoretical_loss": 3.620625049841664,
      "tokens_seen": 1086783488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043273359182901454,
      "loss": 2.7113,
      "theoretical_loss": 3.620604590916014,
      "tokens_seen": 1086849024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004327170418006431,
      "loss": 2.6568,
      "theoretical_loss": 3.6205841335693805,
      "tokens_seen": 1086914560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004327004917722716,
      "loss": 2.7288,
      "theoretical_loss": 3.620563677801547,
      "tokens_seen": 1086980096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004326839417439001,
      "loss": 2.6159,
      "theoretical_loss": 3.6205432236122963,
      "tokens_seen": 1087045632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043266739171552864,
      "loss": 2.861,
      "theoretical_loss": 3.620522771001411,
      "tokens_seen": 1087111168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043265084168715717,
      "loss": 3.0317,
      "theoretical_loss": 3.6205023199686743,
      "tokens_seen": 1087176704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004326342916587857,
      "loss": 2.612,
      "theoretical_loss": 3.6204818705138697,
      "tokens_seen": 1087242240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004326177416304142,
      "loss": 2.6517,
      "theoretical_loss": 3.6204614226367804,
      "tokens_seen": 1087307776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043260119160204273,
      "loss": 2.7742,
      "theoretical_loss": 3.620440976337189,
      "tokens_seen": 1087373312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043258464157367127,
      "loss": 2.8364,
      "theoretical_loss": 3.6204205316148794,
      "tokens_seen": 1087438848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004325680915452998,
      "loss": 2.5643,
      "theoretical_loss": 3.6204000884696343,
      "tokens_seen": 1087504384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004325515415169283,
      "loss": 2.5641,
      "theoretical_loss": 3.6203796469012373,
      "tokens_seen": 1087569920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043253499148855683,
      "loss": 2.9207,
      "theoretical_loss": 3.6203592069094714,
      "tokens_seen": 1087635456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043251844146018536,
      "loss": 2.7468,
      "theoretical_loss": 3.620338768494121,
      "tokens_seen": 1087700992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004325018914318139,
      "loss": 2.8298,
      "theoretical_loss": 3.6203183316549685,
      "tokens_seen": 1087766528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004324853414034424,
      "loss": 2.8051,
      "theoretical_loss": 3.620297896391798,
      "tokens_seen": 1087832064
    },
    {
      "epoch": 0.39,
      "objective/train/avg_token_score": 0.010958181694149971,
      "objective/train/avg_weight": 0.9870997667312622,
      "objective/train/docs_used": 619111,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8227803707122803,
      "objective/train/std_weight": 0.060610074549913406,
      "objective/train/theoretical_loss": 3.6202774627043923,
      "objective/train/tokens_used": 1108357600,
      "theoretical_loss": 3.6202774627043923,
      "tokens_seen": 1087897600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004324687913750709,
      "loss": 2.5694,
      "theoretical_loss": 3.6202774627043923,
      "tokens_seen": 1087897600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043245224134669946,
      "loss": 2.6718,
      "theoretical_loss": 3.620257030592536,
      "tokens_seen": 1087963136
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000432435691318328,
      "loss": 2.6943,
      "theoretical_loss": 3.6202366000560118,
      "tokens_seen": 1088028672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004324191412899565,
      "loss": 2.6527,
      "theoretical_loss": 3.6202161710946044,
      "tokens_seen": 1088094208
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000432402591261585,
      "loss": 2.9172,
      "theoretical_loss": 3.620195743708096,
      "tokens_seen": 1088159744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043238604123321356,
      "loss": 2.557,
      "theoretical_loss": 3.6201753178962717,
      "tokens_seen": 1088225280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004323694912048421,
      "loss": 2.6927,
      "theoretical_loss": 3.620154893658915,
      "tokens_seen": 1088290816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004323529411764706,
      "loss": 2.7608,
      "theoretical_loss": 3.620134470995809,
      "tokens_seen": 1088356352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004323363911480991,
      "loss": 2.9069,
      "theoretical_loss": 3.620114049906739,
      "tokens_seen": 1088421888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043231984111972765,
      "loss": 2.6498,
      "theoretical_loss": 3.6200936303914872,
      "tokens_seen": 1088487424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004323032910913562,
      "loss": 2.6737,
      "theoretical_loss": 3.6200732124498387,
      "tokens_seen": 1088552960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004322867410629847,
      "loss": 2.7005,
      "theoretical_loss": 3.6200527960815774,
      "tokens_seen": 1088618496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004322701910346132,
      "loss": 2.7122,
      "theoretical_loss": 3.620032381286487,
      "tokens_seen": 1088684032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043225364100624175,
      "loss": 2.5907,
      "theoretical_loss": 3.6200119680643517,
      "tokens_seen": 1088749568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004322370909778703,
      "loss": 2.665,
      "theoretical_loss": 3.6199915564149556,
      "tokens_seen": 1088815104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043222054094949877,
      "loss": 2.817,
      "theoretical_loss": 3.6199711463380835,
      "tokens_seen": 1088880640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004322039909211273,
      "loss": 2.906,
      "theoretical_loss": 3.6199507378335185,
      "tokens_seen": 1088946176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043218744089275584,
      "loss": 2.7229,
      "theoretical_loss": 3.6199303309010458,
      "tokens_seen": 1089011712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004321708908643844,
      "loss": 2.7533,
      "theoretical_loss": 3.6199099255404494,
      "tokens_seen": 1089077248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043215434083601287,
      "loss": 2.8876,
      "theoretical_loss": 3.6198895217515132,
      "tokens_seen": 1089142784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004321377908076414,
      "loss": 2.773,
      "theoretical_loss": 3.6198691195340222,
      "tokens_seen": 1089208320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043212124077926994,
      "loss": 2.7342,
      "theoretical_loss": 3.6198487188877606,
      "tokens_seen": 1089273856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004321046907508985,
      "loss": 2.6774,
      "theoretical_loss": 3.619828319812513,
      "tokens_seen": 1089339392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004320881407225269,
      "loss": 2.5984,
      "theoretical_loss": 3.6198079223080635,
      "tokens_seen": 1089404928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043207159069415544,
      "loss": 2.7633,
      "theoretical_loss": 3.619787526374197,
      "tokens_seen": 1089470464
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000432055040665784,
      "loss": 2.6543,
      "theoretical_loss": 3.6197671320106983,
      "tokens_seen": 1089536000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043203849063741247,
      "loss": 2.8339,
      "theoretical_loss": 3.6197467392173515,
      "tokens_seen": 1089601536
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000432021940609041,
      "loss": 2.7184,
      "theoretical_loss": 3.619726347993942,
      "tokens_seen": 1089667072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043200539058066954,
      "loss": 2.9083,
      "theoretical_loss": 3.619705958340254,
      "tokens_seen": 1089732608
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004319888405522981,
      "loss": 2.6176,
      "theoretical_loss": 3.619685570256072,
      "tokens_seen": 1089798144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043197229052392656,
      "loss": 2.3476,
      "theoretical_loss": 3.6196651837411817,
      "tokens_seen": 1089863680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004319557404955551,
      "loss": 2.8226,
      "theoretical_loss": 3.6196447987953677,
      "tokens_seen": 1089929216
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043193919046718364,
      "loss": 2.7776,
      "theoretical_loss": 3.619624415418414,
      "tokens_seen": 1089994752
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004319226404388122,
      "loss": 2.8245,
      "theoretical_loss": 3.6196040336101065,
      "tokens_seen": 1090060288
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043190609041044066,
      "loss": 2.8332,
      "theoretical_loss": 3.6195836533702295,
      "tokens_seen": 1090125824
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004318895403820692,
      "loss": 2.5901,
      "theoretical_loss": 3.619563274698569,
      "tokens_seen": 1090191360
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043187299035369773,
      "loss": 2.6497,
      "theoretical_loss": 3.619542897594909,
      "tokens_seen": 1090256896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043185644032532627,
      "loss": 2.6992,
      "theoretical_loss": 3.619522522059035,
      "tokens_seen": 1090322432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043183989029695475,
      "loss": 2.6684,
      "theoretical_loss": 3.619502148090733,
      "tokens_seen": 1090387968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004318233402685833,
      "loss": 2.7584,
      "theoretical_loss": 3.6194817756897866,
      "tokens_seen": 1090453504
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043180679024021183,
      "loss": 2.6808,
      "theoretical_loss": 3.6194614048559828,
      "tokens_seen": 1090519040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043179024021184037,
      "loss": 2.7299,
      "theoretical_loss": 3.619441035589105,
      "tokens_seen": 1090584576
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043177369018346885,
      "loss": 2.5845,
      "theoretical_loss": 3.6194206678889396,
      "tokens_seen": 1090650112
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004317571401550974,
      "loss": 2.7559,
      "theoretical_loss": 3.619400301755272,
      "tokens_seen": 1090715648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004317405901267259,
      "loss": 2.6578,
      "theoretical_loss": 3.6193799371878876,
      "tokens_seen": 1090781184
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043172404009835446,
      "loss": 2.2421,
      "theoretical_loss": 3.619359574186572,
      "tokens_seen": 1090846720
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043170749006998295,
      "loss": 2.6764,
      "theoretical_loss": 3.6193392127511097,
      "tokens_seen": 1090912256
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004316909400416115,
      "loss": 2.7851,
      "theoretical_loss": 3.619318852881287,
      "tokens_seen": 1090977792
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043167439001324,
      "loss": 2.3765,
      "theoretical_loss": 3.61929849457689,
      "tokens_seen": 1091043328
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043165783998486856,
      "loss": 2.6425,
      "theoretical_loss": 3.619278137837703,
      "tokens_seen": 1091108864
    },
    {
      "epoch": 0.39,
      "objective/train/avg_token_score": 0.02256636694073677,
      "objective/train/avg_weight": 0.9734399914741516,
      "objective/train/docs_used": 620966,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7899389266967773,
      "objective/train/std_weight": 0.10971695929765701,
      "objective/train/theoretical_loss": 3.619257782663513,
      "objective/train/tokens_used": 1111634400,
      "theoretical_loss": 3.619257782663513,
      "tokens_seen": 1091174400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043164128995649704,
      "loss": 2.9046,
      "theoretical_loss": 3.619257782663513,
      "tokens_seen": 1091174400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004316247399281256,
      "loss": 2.7514,
      "theoretical_loss": 3.6192374290541047,
      "tokens_seen": 1091239936
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004316081898997541,
      "loss": 2.7035,
      "theoretical_loss": 3.619217077009264,
      "tokens_seen": 1091305472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043159163987138265,
      "loss": 2.6462,
      "theoretical_loss": 3.619196726528777,
      "tokens_seen": 1091371008
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043157508984301114,
      "loss": 2.5398,
      "theoretical_loss": 3.6191763776124297,
      "tokens_seen": 1091436544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004315585398146397,
      "loss": 2.6601,
      "theoretical_loss": 3.619156030260008,
      "tokens_seen": 1091502080
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004315419897862682,
      "loss": 2.7355,
      "theoretical_loss": 3.619135684471297,
      "tokens_seen": 1091567616
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043152543975789675,
      "loss": 2.8567,
      "theoretical_loss": 3.6191153402460836,
      "tokens_seen": 1091633152
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043150888972952523,
      "loss": 2.8467,
      "theoretical_loss": 3.6190949975841527,
      "tokens_seen": 1091698688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043149233970115377,
      "loss": 2.8404,
      "theoretical_loss": 3.6190746564852914,
      "tokens_seen": 1091764224
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004314757896727823,
      "loss": 2.8467,
      "theoretical_loss": 3.6190543169492857,
      "tokens_seen": 1091829760
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043145923964441085,
      "loss": 2.8551,
      "theoretical_loss": 3.6190339789759207,
      "tokens_seen": 1091895296
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043144268961603933,
      "loss": 2.7169,
      "theoretical_loss": 3.6190136425649837,
      "tokens_seen": 1091960832
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043142613958766787,
      "loss": 2.7261,
      "theoretical_loss": 3.6189933077162606,
      "tokens_seen": 1092026368
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004314095895592964,
      "loss": 2.9243,
      "theoretical_loss": 3.6189729744295374,
      "tokens_seen": 1092091904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043139303953092494,
      "loss": 2.6565,
      "theoretical_loss": 3.618952642704601,
      "tokens_seen": 1092157440
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004313764895025534,
      "loss": 2.8229,
      "theoretical_loss": 3.6189323125412365,
      "tokens_seen": 1092222976
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043135993947418196,
      "loss": 2.7089,
      "theoretical_loss": 3.618911983939231,
      "tokens_seen": 1092288512
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004313433894458105,
      "loss": 2.8881,
      "theoretical_loss": 3.618891656898372,
      "tokens_seen": 1092354048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043132683941743904,
      "loss": 2.8933,
      "theoretical_loss": 3.618871331418444,
      "tokens_seen": 1092419584
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004313102893890675,
      "loss": 2.7231,
      "theoretical_loss": 3.6188510074992344,
      "tokens_seen": 1092485120
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043129373936069606,
      "loss": 2.7468,
      "theoretical_loss": 3.61883068514053,
      "tokens_seen": 1092550656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004312771893323246,
      "loss": 2.8236,
      "theoretical_loss": 3.6188103643421172,
      "tokens_seen": 1092616192
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043126063930395313,
      "loss": 2.6654,
      "theoretical_loss": 3.6187900451037827,
      "tokens_seen": 1092681728
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004312440892755816,
      "loss": 2.7625,
      "theoretical_loss": 3.618769727425313,
      "tokens_seen": 1092747264
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043122753924721015,
      "loss": 2.6534,
      "theoretical_loss": 3.618749411306495,
      "tokens_seen": 1092812800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004312109892188387,
      "loss": 2.6709,
      "theoretical_loss": 3.618729096747115,
      "tokens_seen": 1092878336
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043119443919046723,
      "loss": 2.7404,
      "theoretical_loss": 3.6187087837469605,
      "tokens_seen": 1092943872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004311778891620957,
      "loss": 2.6235,
      "theoretical_loss": 3.6186884723058172,
      "tokens_seen": 1093009408
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043116133913372425,
      "loss": 2.6715,
      "theoretical_loss": 3.6186681624234733,
      "tokens_seen": 1093074944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004311447891053528,
      "loss": 2.7822,
      "theoretical_loss": 3.6186478540997156,
      "tokens_seen": 1093140480
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004311282390769813,
      "loss": 2.7566,
      "theoretical_loss": 3.61862754733433,
      "tokens_seen": 1093206016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004311116890486098,
      "loss": 2.8109,
      "theoretical_loss": 3.618607242127104,
      "tokens_seen": 1093271552
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004310951390202383,
      "loss": 2.5516,
      "theoretical_loss": 3.6185869384778253,
      "tokens_seen": 1093337088
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004310785889918668,
      "loss": 2.7506,
      "theoretical_loss": 3.61856663638628,
      "tokens_seen": 1093402624
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004310620389634953,
      "loss": 2.6241,
      "theoretical_loss": 3.6185463358522565,
      "tokens_seen": 1093468160
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043104548893512385,
      "loss": 2.4259,
      "theoretical_loss": 3.61852603687554,
      "tokens_seen": 1093533696
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004310289389067524,
      "loss": 2.7485,
      "theoretical_loss": 3.61850573945592,
      "tokens_seen": 1093599232
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004310123888783809,
      "loss": 2.5878,
      "theoretical_loss": 3.618485443593182,
      "tokens_seen": 1093664768
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004309958388500094,
      "loss": 2.7345,
      "theoretical_loss": 3.618465149287114,
      "tokens_seen": 1093730304
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043097928882163795,
      "loss": 2.8862,
      "theoretical_loss": 3.6184448565375034,
      "tokens_seen": 1093795840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004309627387932665,
      "loss": 2.6904,
      "theoretical_loss": 3.6184245653441374,
      "tokens_seen": 1093861376
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000430946188764895,
      "loss": 2.7862,
      "theoretical_loss": 3.6184042757068036,
      "tokens_seen": 1093926912
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004309296387365235,
      "loss": 2.6805,
      "theoretical_loss": 3.6183839876252892,
      "tokens_seen": 1093992448
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043091308870815204,
      "loss": 2.6669,
      "theoretical_loss": 3.618363701099382,
      "tokens_seen": 1094057984
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004308965386797806,
      "loss": 2.4029,
      "theoretical_loss": 3.6183434161288694,
      "tokens_seen": 1094123520
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004308799886514091,
      "loss": 2.5139,
      "theoretical_loss": 3.618323132713539,
      "tokens_seen": 1094189056
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004308634386230376,
      "loss": 2.6988,
      "theoretical_loss": 3.6183028508531785,
      "tokens_seen": 1094254592
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043084688859466614,
      "loss": 2.6648,
      "theoretical_loss": 3.618282570547575,
      "tokens_seen": 1094320128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004308303385662947,
      "loss": 2.7941,
      "theoretical_loss": 3.6182622917965173,
      "tokens_seen": 1094385664
    },
    {
      "epoch": 0.39,
      "objective/train/avg_token_score": 0.01205773837864399,
      "objective/train/avg_weight": 0.9858139157295227,
      "objective/train/docs_used": 623021,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4904532432556152,
      "objective/train/std_weight": 0.043908391147851944,
      "objective/train/theoretical_loss": 3.618242014599793,
      "objective/train/tokens_used": 1114911200,
      "theoretical_loss": 3.618242014599793,
      "tokens_seen": 1094451200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004308137885379232,
      "loss": 2.5755,
      "theoretical_loss": 3.618242014599793,
      "tokens_seen": 1094451200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004307972385095517,
      "loss": 2.8199,
      "theoretical_loss": 3.618221738957189,
      "tokens_seen": 1094516736
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043078068848118023,
      "loss": 2.8336,
      "theoretical_loss": 3.618201464868493,
      "tokens_seen": 1094582272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043076413845280877,
      "loss": 2.6229,
      "theoretical_loss": 3.6181811923334943,
      "tokens_seen": 1094647808
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004307475884244373,
      "loss": 2.5787,
      "theoretical_loss": 3.61816092135198,
      "tokens_seen": 1094713344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004307310383960658,
      "loss": 2.5947,
      "theoretical_loss": 3.6181406519237376,
      "tokens_seen": 1094778880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043071448836769433,
      "loss": 2.6391,
      "theoretical_loss": 3.618120384048556,
      "tokens_seen": 1094844416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043069793833932287,
      "loss": 2.6347,
      "theoretical_loss": 3.618100117726223,
      "tokens_seen": 1094909952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004306813883109514,
      "loss": 2.7383,
      "theoretical_loss": 3.6180798529565266,
      "tokens_seen": 1094975488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004306648382825799,
      "loss": 2.7464,
      "theoretical_loss": 3.6180595897392545,
      "tokens_seen": 1095041024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004306482882542084,
      "loss": 2.6234,
      "theoretical_loss": 3.6180393280741954,
      "tokens_seen": 1095106560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043063173822583696,
      "loss": 2.535,
      "theoretical_loss": 3.6180190679611375,
      "tokens_seen": 1095172096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004306151881974655,
      "loss": 2.5822,
      "theoretical_loss": 3.617998809399869,
      "tokens_seen": 1095237632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000430598638169094,
      "loss": 2.8227,
      "theoretical_loss": 3.6179785523901775,
      "tokens_seen": 1095303168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004305820881407225,
      "loss": 2.9193,
      "theoretical_loss": 3.6179582969318527,
      "tokens_seen": 1095368704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043056553811235106,
      "loss": 2.8744,
      "theoretical_loss": 3.617938043024682,
      "tokens_seen": 1095434240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004305489880839796,
      "loss": 2.6835,
      "theoretical_loss": 3.6179177906684536,
      "tokens_seen": 1095499776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004305324380556081,
      "loss": 2.6226,
      "theoretical_loss": 3.617897539862957,
      "tokens_seen": 1095565312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004305158880272366,
      "loss": 2.5856,
      "theoretical_loss": 3.6178772906079804,
      "tokens_seen": 1095630848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043049933799886516,
      "loss": 2.9882,
      "theoretical_loss": 3.617857042903312,
      "tokens_seen": 1095696384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004304827879704937,
      "loss": 2.7251,
      "theoretical_loss": 3.6178367967487404,
      "tokens_seen": 1095761920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004304662379421222,
      "loss": 2.6913,
      "theoretical_loss": 3.617816552144054,
      "tokens_seen": 1095827456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004304496879137507,
      "loss": 2.6716,
      "theoretical_loss": 3.6177963090890417,
      "tokens_seen": 1095892992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043043313788537925,
      "loss": 2.8136,
      "theoretical_loss": 3.6177760675834927,
      "tokens_seen": 1095958528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004304165878570078,
      "loss": 2.7135,
      "theoretical_loss": 3.617755827627195,
      "tokens_seen": 1096024064
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004304000378286363,
      "loss": 2.8071,
      "theoretical_loss": 3.6177355892199383,
      "tokens_seen": 1096089600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004303834878002648,
      "loss": 2.7275,
      "theoretical_loss": 3.6177153523615107,
      "tokens_seen": 1096155136
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043036693777189335,
      "loss": 2.6926,
      "theoretical_loss": 3.617695117051701,
      "tokens_seen": 1096220672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004303503877435219,
      "loss": 2.4534,
      "theoretical_loss": 3.6176748832902987,
      "tokens_seen": 1096286208
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043033383771515037,
      "loss": 2.5809,
      "theoretical_loss": 3.6176546510770926,
      "tokens_seen": 1096351744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004303172876867789,
      "loss": 2.8946,
      "theoretical_loss": 3.6176344204118713,
      "tokens_seen": 1096417280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043030073765840744,
      "loss": 2.5629,
      "theoretical_loss": 3.617614191294424,
      "tokens_seen": 1096482816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000430284187630036,
      "loss": 2.6215,
      "theoretical_loss": 3.6175939637245405,
      "tokens_seen": 1096548352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043026763760166446,
      "loss": 2.7587,
      "theoretical_loss": 3.6175737377020085,
      "tokens_seen": 1096613888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000430251087573293,
      "loss": 2.5447,
      "theoretical_loss": 3.6175535132266186,
      "tokens_seen": 1096679424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043023453754492154,
      "loss": 2.7537,
      "theoretical_loss": 3.6175332902981587,
      "tokens_seen": 1096744960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004302179875165501,
      "loss": 2.6665,
      "theoretical_loss": 3.6175130689164194,
      "tokens_seen": 1096810496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043020143748817856,
      "loss": 2.9153,
      "theoretical_loss": 3.6174928490811893,
      "tokens_seen": 1096876032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004301848874598071,
      "loss": 2.7849,
      "theoretical_loss": 3.617472630792258,
      "tokens_seen": 1096941568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043016833743143564,
      "loss": 2.6276,
      "theoretical_loss": 3.617452414049414,
      "tokens_seen": 1097007104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004301517874030642,
      "loss": 2.4464,
      "theoretical_loss": 3.617432198852448,
      "tokens_seen": 1097072640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043013523737469266,
      "loss": 2.6695,
      "theoretical_loss": 3.6174119852011484,
      "tokens_seen": 1097138176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043011868734632114,
      "loss": 2.7697,
      "theoretical_loss": 3.6173917730953056,
      "tokens_seen": 1097203712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004301021373179496,
      "loss": 2.5366,
      "theoretical_loss": 3.617371562534708,
      "tokens_seen": 1097269248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043008558728957816,
      "loss": 2.8572,
      "theoretical_loss": 3.6173513535191466,
      "tokens_seen": 1097334784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004300690372612067,
      "loss": 2.9126,
      "theoretical_loss": 3.61733114604841,
      "tokens_seen": 1097400320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043005248723283524,
      "loss": 2.9046,
      "theoretical_loss": 3.6173109401222883,
      "tokens_seen": 1097465856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004300359372044637,
      "loss": 2.8963,
      "theoretical_loss": 3.6172907357405704,
      "tokens_seen": 1097531392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00043001938717609226,
      "loss": 2.6125,
      "theoretical_loss": 3.617270532903047,
      "tokens_seen": 1097596928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004300028371477208,
      "loss": 2.6486,
      "theoretical_loss": 3.617250331609508,
      "tokens_seen": 1097662464
    },
    {
      "epoch": 0.39,
      "objective/train/avg_token_score": 0.012495676055550575,
      "objective/train/avg_weight": 0.9853193163871765,
      "objective/train/docs_used": 624881,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.907550811767578,
      "objective/train/std_weight": 0.08662146329879761,
      "objective/train/theoretical_loss": 3.617230131859743,
      "objective/train/tokens_used": 1118188000,
      "theoretical_loss": 3.617230131859743,
      "tokens_seen": 1097728000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042998628711934933,
      "loss": 2.8396,
      "theoretical_loss": 3.617230131859743,
      "tokens_seen": 1097728000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004299697370909778,
      "loss": 2.5036,
      "theoretical_loss": 3.617209933653541,
      "tokens_seen": 1097793536
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042995318706260635,
      "loss": 2.59,
      "theoretical_loss": 3.6171897369906922,
      "tokens_seen": 1097859072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004299366370342349,
      "loss": 2.8178,
      "theoretical_loss": 3.6171695418709877,
      "tokens_seen": 1097924608
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042992008700586343,
      "loss": 2.6063,
      "theoretical_loss": 3.617149348294217,
      "tokens_seen": 1097990144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004299035369774919,
      "loss": 2.5875,
      "theoretical_loss": 3.6171291562601695,
      "tokens_seen": 1098055680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042988698694912045,
      "loss": 2.765,
      "theoretical_loss": 3.6171089657686357,
      "tokens_seen": 1098121216
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000429870436920749,
      "loss": 2.7842,
      "theoretical_loss": 3.617088776819406,
      "tokens_seen": 1098186752
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004298538868923775,
      "loss": 2.7824,
      "theoretical_loss": 3.61706858941227,
      "tokens_seen": 1098252288
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000429837336864006,
      "loss": 2.6613,
      "theoretical_loss": 3.6170484035470176,
      "tokens_seen": 1098317824
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042982078683563454,
      "loss": 2.5709,
      "theoretical_loss": 3.61702821922344,
      "tokens_seen": 1098383360
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004298042368072631,
      "loss": 2.724,
      "theoretical_loss": 3.6170080364413275,
      "tokens_seen": 1098448896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004297876867788916,
      "loss": 2.6708,
      "theoretical_loss": 3.6169878552004695,
      "tokens_seen": 1098514432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004297711367505201,
      "loss": 2.8149,
      "theoretical_loss": 3.616967675500657,
      "tokens_seen": 1098579968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042975458672214864,
      "loss": 2.8136,
      "theoretical_loss": 3.6169474973416804,
      "tokens_seen": 1098645504
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004297380366937772,
      "loss": 2.7571,
      "theoretical_loss": 3.6169273207233292,
      "tokens_seen": 1098711040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004297214866654057,
      "loss": 2.4577,
      "theoretical_loss": 3.6169071456453956,
      "tokens_seen": 1098776576
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004297049366370342,
      "loss": 2.684,
      "theoretical_loss": 3.6168869721076686,
      "tokens_seen": 1098842112
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042968838660866274,
      "loss": 2.663,
      "theoretical_loss": 3.61686680010994,
      "tokens_seen": 1098907648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004296718365802913,
      "loss": 2.8735,
      "theoretical_loss": 3.616846629651999,
      "tokens_seen": 1098973184
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004296552865519198,
      "loss": 2.8015,
      "theoretical_loss": 3.6168264607336376,
      "tokens_seen": 1099038720
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004296387365235483,
      "loss": 2.6439,
      "theoretical_loss": 3.616806293354646,
      "tokens_seen": 1099104256
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042962218649517683,
      "loss": 2.7169,
      "theoretical_loss": 3.6167861275148145,
      "tokens_seen": 1099169792
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042960563646680537,
      "loss": 2.7987,
      "theoretical_loss": 3.6167659632139344,
      "tokens_seen": 1099235328
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004295890864384339,
      "loss": 2.8021,
      "theoretical_loss": 3.6167458004517963,
      "tokens_seen": 1099300864
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004295725364100624,
      "loss": 2.4914,
      "theoretical_loss": 3.616725639228191,
      "tokens_seen": 1099366400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042955598638169093,
      "loss": 2.5571,
      "theoretical_loss": 3.61670547954291,
      "tokens_seen": 1099431936
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042953943635331947,
      "loss": 2.7847,
      "theoretical_loss": 3.6166853213957433,
      "tokens_seen": 1099497472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000429522886324948,
      "loss": 2.8254,
      "theoretical_loss": 3.616665164786482,
      "tokens_seen": 1099563008
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004295063362965765,
      "loss": 2.648,
      "theoretical_loss": 3.616645009714918,
      "tokens_seen": 1099628544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000429489786268205,
      "loss": 2.6426,
      "theoretical_loss": 3.6166248561808416,
      "tokens_seen": 1099694080
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042947323623983356,
      "loss": 2.7715,
      "theoretical_loss": 3.616604704184044,
      "tokens_seen": 1099759616
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004294566862114621,
      "loss": 2.6553,
      "theoretical_loss": 3.616584553724316,
      "tokens_seen": 1099825152
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042944013618309064,
      "loss": 2.7151,
      "theoretical_loss": 3.6165644048014505,
      "tokens_seen": 1099890688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004294235861547191,
      "loss": 2.514,
      "theoretical_loss": 3.616544257415236,
      "tokens_seen": 1099956224
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042940703612634766,
      "loss": 2.5279,
      "theoretical_loss": 3.6165241115654663,
      "tokens_seen": 1100021760
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004293904860979762,
      "loss": 2.7526,
      "theoretical_loss": 3.616503967251931,
      "tokens_seen": 1100087296
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042937393606960473,
      "loss": 2.8082,
      "theoretical_loss": 3.616483824474422,
      "tokens_seen": 1100152832
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004293573860412332,
      "loss": 2.7141,
      "theoretical_loss": 3.6164636832327313,
      "tokens_seen": 1100218368
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042934083601286175,
      "loss": 2.815,
      "theoretical_loss": 3.616443543526649,
      "tokens_seen": 1100283904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004293242859844903,
      "loss": 2.759,
      "theoretical_loss": 3.6164234053559676,
      "tokens_seen": 1100349440
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042930773595611883,
      "loss": 2.8818,
      "theoretical_loss": 3.616403268720479,
      "tokens_seen": 1100414976
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004292911859277473,
      "loss": 2.6402,
      "theoretical_loss": 3.616383133619973,
      "tokens_seen": 1100480512
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042927463589937585,
      "loss": 2.5741,
      "theoretical_loss": 3.6163630000542435,
      "tokens_seen": 1100546048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004292580858710044,
      "loss": 2.6975,
      "theoretical_loss": 3.6163428680230796,
      "tokens_seen": 1100611584
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004292415358426329,
      "loss": 2.7838,
      "theoretical_loss": 3.616322737526275,
      "tokens_seen": 1100677120
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004292249858142614,
      "loss": 2.4501,
      "theoretical_loss": 3.6163026085636205,
      "tokens_seen": 1100742656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042920843578588995,
      "loss": 2.6096,
      "theoretical_loss": 3.6162824811349084,
      "tokens_seen": 1100808192
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004291918857575185,
      "loss": 2.866,
      "theoretical_loss": 3.61626235523993,
      "tokens_seen": 1100873728
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000429175335729147,
      "loss": 2.7073,
      "theoretical_loss": 3.616242230878477,
      "tokens_seen": 1100939264
    },
    {
      "epoch": 0.39,
      "objective/train/avg_token_score": 0.009765907190740108,
      "objective/train/avg_weight": 0.9885014891624451,
      "objective/train/docs_used": 626669,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.959967613220215,
      "objective/train/std_weight": 0.06800971925258636,
      "objective/train/theoretical_loss": 3.6162221080503416,
      "objective/train/tokens_used": 1121464800,
      "theoretical_loss": 3.6162221080503416,
      "tokens_seen": 1101004800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004291587857007755,
      "loss": 2.7011,
      "theoretical_loss": 3.6162221080503416,
      "tokens_seen": 1101004800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000429142235672404,
      "loss": 2.5699,
      "theoretical_loss": 3.6162019867553155,
      "tokens_seen": 1101070336
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042912568564403247,
      "loss": 2.6745,
      "theoretical_loss": 3.616181866993191,
      "tokens_seen": 1101135872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000429109135615661,
      "loss": 2.5318,
      "theoretical_loss": 3.6161617487637603,
      "tokens_seen": 1101201408
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042909258558728955,
      "loss": 2.6188,
      "theoretical_loss": 3.6161416320668147,
      "tokens_seen": 1101266944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004290760355589181,
      "loss": 2.6821,
      "theoretical_loss": 3.6161215169021466,
      "tokens_seen": 1101332480
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042905948553054657,
      "loss": 2.7821,
      "theoretical_loss": 3.616101403269548,
      "tokens_seen": 1101398016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004290429355021751,
      "loss": 2.4128,
      "theoretical_loss": 3.616081291168811,
      "tokens_seen": 1101463552
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042902638547380364,
      "loss": 2.5214,
      "theoretical_loss": 3.6160611805997283,
      "tokens_seen": 1101529088
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004290098354454322,
      "loss": 2.8408,
      "theoretical_loss": 3.616041071562092,
      "tokens_seen": 1101594624
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042899328541706066,
      "loss": 2.6695,
      "theoretical_loss": 3.616020964055694,
      "tokens_seen": 1101660160
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004289767353886892,
      "loss": 2.6769,
      "theoretical_loss": 3.616000858080327,
      "tokens_seen": 1101725696
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042896018536031774,
      "loss": 2.2207,
      "theoretical_loss": 3.615980753635783,
      "tokens_seen": 1101791232
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004289436353319463,
      "loss": 2.7103,
      "theoretical_loss": 3.615960650721855,
      "tokens_seen": 1101856768
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042892708530357476,
      "loss": 2.7155,
      "theoretical_loss": 3.6159405493383345,
      "tokens_seen": 1101922304
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004289105352752033,
      "loss": 2.4721,
      "theoretical_loss": 3.6159204494850146,
      "tokens_seen": 1101987840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042889398524683183,
      "loss": 2.8062,
      "theoretical_loss": 3.615900351161688,
      "tokens_seen": 1102053376
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042887743521846037,
      "loss": 2.7285,
      "theoretical_loss": 3.6158802543681468,
      "tokens_seen": 1102118912
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042886088519008886,
      "loss": 2.6086,
      "theoretical_loss": 3.615860159104184,
      "tokens_seen": 1102184448
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004288443351617174,
      "loss": 2.644,
      "theoretical_loss": 3.615840065369592,
      "tokens_seen": 1102249984
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042882778513334593,
      "loss": 2.6383,
      "theoretical_loss": 3.6158199731641636,
      "tokens_seen": 1102315520
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042881123510497447,
      "loss": 2.6125,
      "theoretical_loss": 3.6157998824876914,
      "tokens_seen": 1102381056
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042879468507660295,
      "loss": 2.8708,
      "theoretical_loss": 3.6157797933399682,
      "tokens_seen": 1102446592
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004287781350482315,
      "loss": 2.6425,
      "theoretical_loss": 3.6157597057207873,
      "tokens_seen": 1102512128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042876158501986,
      "loss": 2.6893,
      "theoretical_loss": 3.6157396196299403,
      "tokens_seen": 1102577664
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042874503499148856,
      "loss": 2.7941,
      "theoretical_loss": 3.615719535067221,
      "tokens_seen": 1102643200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042872848496311705,
      "loss": 2.8433,
      "theoretical_loss": 3.6156994520324224,
      "tokens_seen": 1102708736
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004287119349347456,
      "loss": 2.6919,
      "theoretical_loss": 3.6156793705253376,
      "tokens_seen": 1102774272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004286953849063741,
      "loss": 2.7049,
      "theoretical_loss": 3.615659290545759,
      "tokens_seen": 1102839808
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042867883487800266,
      "loss": 2.531,
      "theoretical_loss": 3.6156392120934795,
      "tokens_seen": 1102905344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042866228484963114,
      "loss": 2.5189,
      "theoretical_loss": 3.6156191351682936,
      "tokens_seen": 1102970880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004286457348212597,
      "loss": 2.6985,
      "theoretical_loss": 3.615599059769993,
      "tokens_seen": 1103036416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004286291847928882,
      "loss": 2.692,
      "theoretical_loss": 3.615578985898371,
      "tokens_seen": 1103101952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042861263476451676,
      "loss": 2.4507,
      "theoretical_loss": 3.615558913553221,
      "tokens_seen": 1103167488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042859608473614524,
      "loss": 2.5819,
      "theoretical_loss": 3.615538842734337,
      "tokens_seen": 1103233024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004285795347077738,
      "loss": 2.5725,
      "theoretical_loss": 3.615518773441511,
      "tokens_seen": 1103298560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004285629846794023,
      "loss": 2.5295,
      "theoretical_loss": 3.615498705674537,
      "tokens_seen": 1103364096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042854643465103085,
      "loss": 2.5529,
      "theoretical_loss": 3.615478639433209,
      "tokens_seen": 1103429632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042852988462265933,
      "loss": 2.8151,
      "theoretical_loss": 3.6154585747173194,
      "tokens_seen": 1103495168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042851333459428787,
      "loss": 2.759,
      "theoretical_loss": 3.615438511526662,
      "tokens_seen": 1103560704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004284967845659164,
      "loss": 2.6474,
      "theoretical_loss": 3.6154184498610307,
      "tokens_seen": 1103626240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042848023453754495,
      "loss": 2.7961,
      "theoretical_loss": 3.6153983897202187,
      "tokens_seen": 1103691776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042846368450917343,
      "loss": 2.6651,
      "theoretical_loss": 3.6153783311040186,
      "tokens_seen": 1103757312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042844713448080197,
      "loss": 2.9282,
      "theoretical_loss": 3.615358274012226,
      "tokens_seen": 1103822848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004284305844524305,
      "loss": 2.8171,
      "theoretical_loss": 3.6153382184446325,
      "tokens_seen": 1103888384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042841403442405904,
      "loss": 2.6859,
      "theoretical_loss": 3.6153181644010335,
      "tokens_seen": 1103953920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004283974843956875,
      "loss": 2.6268,
      "theoretical_loss": 3.6152981118812217,
      "tokens_seen": 1104019456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042838093436731606,
      "loss": 2.64,
      "theoretical_loss": 3.6152780608849913,
      "tokens_seen": 1104084992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004283643843389446,
      "loss": 2.8767,
      "theoretical_loss": 3.615258011412136,
      "tokens_seen": 1104150528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042834783431057314,
      "loss": 2.8614,
      "theoretical_loss": 3.6152379634624494,
      "tokens_seen": 1104216064
    },
    {
      "epoch": 0.39,
      "objective/train/avg_token_score": 0.01492348127067089,
      "objective/train/avg_weight": 0.982431948184967,
      "objective/train/docs_used": 628420,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.147702932357788,
      "objective/train/std_weight": 0.07864157110452652,
      "objective/train/theoretical_loss": 3.615217917035726,
      "objective/train/tokens_used": 1124741600,
      "theoretical_loss": 3.615217917035726,
      "tokens_seen": 1104281600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004283312842822016,
      "loss": 2.7769,
      "theoretical_loss": 3.615217917035726,
      "tokens_seen": 1104281600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042831473425383016,
      "loss": 2.7143,
      "theoretical_loss": 3.615197872131759,
      "tokens_seen": 1104347136
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004282981842254587,
      "loss": 2.6533,
      "theoretical_loss": 3.615177828750343,
      "tokens_seen": 1104412672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042828163419708724,
      "loss": 2.9865,
      "theoretical_loss": 3.6151577868912717,
      "tokens_seen": 1104478208
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004282650841687157,
      "loss": 2.6031,
      "theoretical_loss": 3.615137746554339,
      "tokens_seen": 1104543744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042824853414034426,
      "loss": 2.7158,
      "theoretical_loss": 3.6151177077393397,
      "tokens_seen": 1104609280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004282319841119728,
      "loss": 2.7248,
      "theoretical_loss": 3.615097670446067,
      "tokens_seen": 1104674816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042821543408360133,
      "loss": 2.84,
      "theoretical_loss": 3.615077634674316,
      "tokens_seen": 1104740352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004281988840552298,
      "loss": 2.6037,
      "theoretical_loss": 3.6150576004238797,
      "tokens_seen": 1104805888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042818233402685835,
      "loss": 2.7066,
      "theoretical_loss": 3.6150375676945536,
      "tokens_seen": 1104871424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004281657839984869,
      "loss": 2.6315,
      "theoretical_loss": 3.6150175364861314,
      "tokens_seen": 1104936960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004281492339701153,
      "loss": 2.7599,
      "theoretical_loss": 3.6149975067984075,
      "tokens_seen": 1105002496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042813268394174386,
      "loss": 2.7227,
      "theoretical_loss": 3.614977478631176,
      "tokens_seen": 1105068032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004281161339133724,
      "loss": 2.6205,
      "theoretical_loss": 3.6149574519842322,
      "tokens_seen": 1105133568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042809958388500093,
      "loss": 2.8727,
      "theoretical_loss": 3.6149374268573697,
      "tokens_seen": 1105199104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004280830338566294,
      "loss": 2.4647,
      "theoretical_loss": 3.6149174032503835,
      "tokens_seen": 1105264640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042806648382825795,
      "loss": 2.8422,
      "theoretical_loss": 3.6148973811630674,
      "tokens_seen": 1105330176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004280499337998865,
      "loss": 2.937,
      "theoretical_loss": 3.6148773605952167,
      "tokens_seen": 1105395712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042803338377151503,
      "loss": 2.6538,
      "theoretical_loss": 3.6148573415466254,
      "tokens_seen": 1105461248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004280168337431435,
      "loss": 2.522,
      "theoretical_loss": 3.6148373240170892,
      "tokens_seen": 1105526784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042800028371477205,
      "loss": 2.69,
      "theoretical_loss": 3.614817308006401,
      "tokens_seen": 1105592320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004279837336864006,
      "loss": 2.8332,
      "theoretical_loss": 3.614797293514358,
      "tokens_seen": 1105657856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004279671836580291,
      "loss": 2.6047,
      "theoretical_loss": 3.614777280540753,
      "tokens_seen": 1105723392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004279506336296576,
      "loss": 2.5747,
      "theoretical_loss": 3.6147572690853815,
      "tokens_seen": 1105788928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00042793408360128614,
      "loss": 2.842,
      "theoretical_loss": 3.614737259148038,
      "tokens_seen": 1105854464
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0004279175335729147,
      "loss": 2.8476,
      "theoretical_loss": 3.614717250728518,
      "tokens_seen": 1105920000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004279009835445432,
      "loss": 2.8253,
      "theoretical_loss": 3.6146972438266163,
      "tokens_seen": 1105985536
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004278844335161717,
      "loss": 2.6302,
      "theoretical_loss": 3.6146772384421273,
      "tokens_seen": 1106051072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042786788348780024,
      "loss": 2.6288,
      "theoretical_loss": 3.6146572345748464,
      "tokens_seen": 1106116608
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004278513334594288,
      "loss": 2.2949,
      "theoretical_loss": 3.614637232224569,
      "tokens_seen": 1106182144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004278347834310573,
      "loss": 2.7983,
      "theoretical_loss": 3.6146172313910894,
      "tokens_seen": 1106247680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004278182334026858,
      "loss": 2.685,
      "theoretical_loss": 3.6145972320742037,
      "tokens_seen": 1106313216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042780168337431434,
      "loss": 2.7602,
      "theoretical_loss": 3.614577234273706,
      "tokens_seen": 1106378752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004277851333459429,
      "loss": 2.9248,
      "theoretical_loss": 3.6145572379893927,
      "tokens_seen": 1106444288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004277685833175714,
      "loss": 2.5934,
      "theoretical_loss": 3.614537243221058,
      "tokens_seen": 1106509824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004277520332891999,
      "loss": 2.5021,
      "theoretical_loss": 3.6145172499684977,
      "tokens_seen": 1106575360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042773548326082843,
      "loss": 2.9372,
      "theoretical_loss": 3.6144972582315065,
      "tokens_seen": 1106640896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042771893323245697,
      "loss": 2.689,
      "theoretical_loss": 3.614477268009881,
      "tokens_seen": 1106706432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004277023832040855,
      "loss": 2.7185,
      "theoretical_loss": 3.6144572793034158,
      "tokens_seen": 1106771968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000427685833175714,
      "loss": 2.7188,
      "theoretical_loss": 3.6144372921119063,
      "tokens_seen": 1106837504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042766928314734253,
      "loss": 2.5651,
      "theoretical_loss": 3.614417306435148,
      "tokens_seen": 1106903040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042765273311897107,
      "loss": 2.3605,
      "theoretical_loss": 3.614397322272937,
      "tokens_seen": 1106968576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004276361830905996,
      "loss": 2.8136,
      "theoretical_loss": 3.6143773396250687,
      "tokens_seen": 1107034112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004276196330622281,
      "loss": 2.5023,
      "theoretical_loss": 3.614357358491338,
      "tokens_seen": 1107099648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004276030830338566,
      "loss": 2.5301,
      "theoretical_loss": 3.614337378871541,
      "tokens_seen": 1107165184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042758653300548516,
      "loss": 2.6466,
      "theoretical_loss": 3.6143174007654735,
      "tokens_seen": 1107230720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004275699829771137,
      "loss": 2.8662,
      "theoretical_loss": 3.614297424172931,
      "tokens_seen": 1107296256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004275534329487422,
      "loss": 2.7399,
      "theoretical_loss": 3.6142774490937093,
      "tokens_seen": 1107361792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004275368829203707,
      "loss": 2.7811,
      "theoretical_loss": 3.6142574755276047,
      "tokens_seen": 1107427328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042752033289199926,
      "loss": 2.7683,
      "theoretical_loss": 3.6142375034744125,
      "tokens_seen": 1107492864
    },
    {
      "epoch": 0.4,
      "objective/train/avg_token_score": 0.01359342597424984,
      "objective/train/avg_weight": 0.9840013980865479,
      "objective/train/docs_used": 629445,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9303736686706543,
      "objective/train/std_weight": 0.06392809748649597,
      "objective/train/theoretical_loss": 3.614217532933929,
      "objective/train/tokens_used": 1128018400,
      "theoretical_loss": 3.614217532933929,
      "tokens_seen": 1107558400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004275037828636278,
      "loss": 2.7619,
      "theoretical_loss": 3.614217532933929,
      "tokens_seen": 1107558400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004274872328352563,
      "loss": 3.0295,
      "theoretical_loss": 3.6141975639059494,
      "tokens_seen": 1107623936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004274706828068848,
      "loss": 2.8468,
      "theoretical_loss": 3.6141775963902703,
      "tokens_seen": 1107689472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042745413277851335,
      "loss": 2.7423,
      "theoretical_loss": 3.6141576303866874,
      "tokens_seen": 1107755008
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004274375827501419,
      "loss": 2.6899,
      "theoretical_loss": 3.6141376658949973,
      "tokens_seen": 1107820544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004274210327217704,
      "loss": 2.6368,
      "theoretical_loss": 3.614117702914995,
      "tokens_seen": 1107886080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004274044826933989,
      "loss": 2.8682,
      "theoretical_loss": 3.614097741446478,
      "tokens_seen": 1107951616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042738793266502745,
      "loss": 2.7838,
      "theoretical_loss": 3.6140777814892413,
      "tokens_seen": 1108017152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000427371382636656,
      "loss": 2.5592,
      "theoretical_loss": 3.6140578230430815,
      "tokens_seen": 1108082688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042735483260828447,
      "loss": 2.4224,
      "theoretical_loss": 3.6140378661077954,
      "tokens_seen": 1108148224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000427338282579913,
      "loss": 2.6766,
      "theoretical_loss": 3.6140179106831787,
      "tokens_seen": 1108213760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042732173255154155,
      "loss": 2.7225,
      "theoretical_loss": 3.6139979567690275,
      "tokens_seen": 1108279296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004273051825231701,
      "loss": 2.6468,
      "theoretical_loss": 3.613978004365139,
      "tokens_seen": 1108344832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042728863249479857,
      "loss": 2.7037,
      "theoretical_loss": 3.6139580534713085,
      "tokens_seen": 1108410368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004272720824664271,
      "loss": 2.4942,
      "theoretical_loss": 3.613938104087333,
      "tokens_seen": 1108475904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042725553243805564,
      "loss": 2.7608,
      "theoretical_loss": 3.6139181562130096,
      "tokens_seen": 1108541440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004272389824096842,
      "loss": 2.7526,
      "theoretical_loss": 3.6138982098481334,
      "tokens_seen": 1108606976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042722243238131266,
      "loss": 2.833,
      "theoretical_loss": 3.6138782649925023,
      "tokens_seen": 1108672512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004272058823529412,
      "loss": 2.5924,
      "theoretical_loss": 3.613858321645912,
      "tokens_seen": 1108738048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042718933232456974,
      "loss": 2.6053,
      "theoretical_loss": 3.61383837980816,
      "tokens_seen": 1108803584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042717278229619817,
      "loss": 2.6015,
      "theoretical_loss": 3.613818439479042,
      "tokens_seen": 1108869120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004271562322678267,
      "loss": 2.5801,
      "theoretical_loss": 3.6137985006583557,
      "tokens_seen": 1108934656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042713968223945524,
      "loss": 2.5939,
      "theoretical_loss": 3.613778563345897,
      "tokens_seen": 1109000192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004271231322110837,
      "loss": 2.801,
      "theoretical_loss": 3.613758627541463,
      "tokens_seen": 1109065728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042710658218271226,
      "loss": 2.6325,
      "theoretical_loss": 3.6137386932448505,
      "tokens_seen": 1109131264
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004270900321543408,
      "loss": 2.8322,
      "theoretical_loss": 3.613718760455857,
      "tokens_seen": 1109196800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042707348212596934,
      "loss": 2.7511,
      "theoretical_loss": 3.613698829174278,
      "tokens_seen": 1109262336
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004270569320975978,
      "loss": 2.9082,
      "theoretical_loss": 3.613678899399912,
      "tokens_seen": 1109327872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042704038206922636,
      "loss": 2.6177,
      "theoretical_loss": 3.6136589711325557,
      "tokens_seen": 1109393408
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004270238320408549,
      "loss": 2.6467,
      "theoretical_loss": 3.613639044372005,
      "tokens_seen": 1109458944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042700728201248343,
      "loss": 2.5678,
      "theoretical_loss": 3.613619119118058,
      "tokens_seen": 1109524480
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042699073198411197,
      "loss": 2.6747,
      "theoretical_loss": 3.6135991953705116,
      "tokens_seen": 1109590016
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042697418195574045,
      "loss": 2.7123,
      "theoretical_loss": 3.6135792731291625,
      "tokens_seen": 1109655552
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000426957631927369,
      "loss": 2.8426,
      "theoretical_loss": 3.6135593523938088,
      "tokens_seen": 1109721088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042694108189899753,
      "loss": 2.263,
      "theoretical_loss": 3.613539433164247,
      "tokens_seen": 1109786624
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042692453187062607,
      "loss": 2.5,
      "theoretical_loss": 3.613519515440274,
      "tokens_seen": 1109852160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042690798184225455,
      "loss": 2.6986,
      "theoretical_loss": 3.6134995992216883,
      "tokens_seen": 1109917696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004268914318138831,
      "loss": 2.8229,
      "theoretical_loss": 3.6134796845082864,
      "tokens_seen": 1109983232
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004268748817855116,
      "loss": 2.7771,
      "theoretical_loss": 3.613459771299866,
      "tokens_seen": 1110048768
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042685833175714016,
      "loss": 2.7293,
      "theoretical_loss": 3.613439859596224,
      "tokens_seen": 1110114304
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042684178172876865,
      "loss": 2.8418,
      "theoretical_loss": 3.613419949397159,
      "tokens_seen": 1110179840
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004268252317003972,
      "loss": 2.8764,
      "theoretical_loss": 3.6134000407024676,
      "tokens_seen": 1110245376
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004268086816720257,
      "loss": 2.7868,
      "theoretical_loss": 3.6133801335119475,
      "tokens_seen": 1110310912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042679213164365426,
      "loss": 2.7715,
      "theoretical_loss": 3.613360227825396,
      "tokens_seen": 1110376448
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042677558161528274,
      "loss": 2.6762,
      "theoretical_loss": 3.6133403236426114,
      "tokens_seen": 1110441984
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004267590315869113,
      "loss": 2.7463,
      "theoretical_loss": 3.6133204209633907,
      "tokens_seen": 1110507520
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004267424815585398,
      "loss": 2.7826,
      "theoretical_loss": 3.613300519787532,
      "tokens_seen": 1110573056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042672593153016836,
      "loss": 2.618,
      "theoretical_loss": 3.613280620114833,
      "tokens_seen": 1110638592
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042670938150179684,
      "loss": 2.588,
      "theoretical_loss": 3.6132607219450916,
      "tokens_seen": 1110704128
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004266928314734254,
      "loss": 2.4912,
      "theoretical_loss": 3.6132408252781048,
      "tokens_seen": 1110769664
    },
    {
      "epoch": 0.4,
      "objective/train/avg_token_score": 0.0256155114620924,
      "objective/train/avg_weight": 0.9698538780212402,
      "objective/train/docs_used": 631502,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.249163627624512,
      "objective/train/std_weight": 0.12881620228290558,
      "objective/train/theoretical_loss": 3.6132209301136715,
      "objective/train/tokens_used": 1131295200,
      "theoretical_loss": 3.6132209301136715,
      "tokens_seen": 1110835200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004266762814450539,
      "loss": 2.5343,
      "theoretical_loss": 3.6132209301136715,
      "tokens_seen": 1110835200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042665973141668245,
      "loss": 2.8165,
      "theoretical_loss": 3.6132010364515894,
      "tokens_seen": 1110900736
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042664318138831093,
      "loss": 2.6496,
      "theoretical_loss": 3.613181144291656,
      "tokens_seen": 1110966272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042662663135993947,
      "loss": 2.5352,
      "theoretical_loss": 3.6131612536336695,
      "tokens_seen": 1111031808
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000426610081331568,
      "loss": 2.5632,
      "theoretical_loss": 3.6131413644774275,
      "tokens_seen": 1111097344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042659353130319655,
      "loss": 2.378,
      "theoretical_loss": 3.613121476822729,
      "tokens_seen": 1111162880
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042657698127482503,
      "loss": 2.7465,
      "theoretical_loss": 3.613101590669371,
      "tokens_seen": 1111228416
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042656043124645357,
      "loss": 2.6602,
      "theoretical_loss": 3.6130817060171525,
      "tokens_seen": 1111293952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004265438812180821,
      "loss": 2.6466,
      "theoretical_loss": 3.6130618228658715,
      "tokens_seen": 1111359488
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042652733118971064,
      "loss": 2.7282,
      "theoretical_loss": 3.6130419412153256,
      "tokens_seen": 1111425024
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004265107811613391,
      "loss": 2.6833,
      "theoretical_loss": 3.6130220610653137,
      "tokens_seen": 1111490560
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042649423113296766,
      "loss": 2.8408,
      "theoretical_loss": 3.613002182415634,
      "tokens_seen": 1111556096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004264776811045962,
      "loss": 2.7446,
      "theoretical_loss": 3.6129823052660845,
      "tokens_seen": 1111621632
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042646113107622474,
      "loss": 2.6151,
      "theoretical_loss": 3.6129624296164637,
      "tokens_seen": 1111687168
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004264445810478532,
      "loss": 2.4997,
      "theoretical_loss": 3.61294255546657,
      "tokens_seen": 1111752704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042642803101948176,
      "loss": 2.5952,
      "theoretical_loss": 3.612922682816202,
      "tokens_seen": 1111818240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004264114809911103,
      "loss": 2.6361,
      "theoretical_loss": 3.6129028116651583,
      "tokens_seen": 1111883776
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042639493096273883,
      "loss": 2.5458,
      "theoretical_loss": 3.612882942013237,
      "tokens_seen": 1111949312
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004263783809343673,
      "loss": 2.5764,
      "theoretical_loss": 3.612863073860237,
      "tokens_seen": 1112014848
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042636183090599586,
      "loss": 2.5735,
      "theoretical_loss": 3.6128432072059566,
      "tokens_seen": 1112080384
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004263452808776244,
      "loss": 2.8846,
      "theoretical_loss": 3.6128233420501945,
      "tokens_seen": 1112145920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042632873084925293,
      "loss": 2.8663,
      "theoretical_loss": 3.61280347839275,
      "tokens_seen": 1112211456
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004263121808208814,
      "loss": 2.6921,
      "theoretical_loss": 3.6127836162334206,
      "tokens_seen": 1112276992
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042629563079250995,
      "loss": 2.8157,
      "theoretical_loss": 3.612763755572006,
      "tokens_seen": 1112342528
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004262790807641385,
      "loss": 2.7012,
      "theoretical_loss": 3.6127438964083045,
      "tokens_seen": 1112408064
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000426262530735767,
      "loss": 2.6615,
      "theoretical_loss": 3.6127240387421153,
      "tokens_seen": 1112473600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004262459807073955,
      "loss": 2.8315,
      "theoretical_loss": 3.6127041825732373,
      "tokens_seen": 1112539136
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042622943067902405,
      "loss": 2.6674,
      "theoretical_loss": 3.612684327901469,
      "tokens_seen": 1112604672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004262128806506526,
      "loss": 2.4711,
      "theoretical_loss": 3.6126644747266097,
      "tokens_seen": 1112670208
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004261963306222811,
      "loss": 2.7252,
      "theoretical_loss": 3.6126446230484577,
      "tokens_seen": 1112735744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042617978059390955,
      "loss": 2.7761,
      "theoretical_loss": 3.612624772866813,
      "tokens_seen": 1112801280
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004261632305655381,
      "loss": 2.7043,
      "theoretical_loss": 3.6126049241814746,
      "tokens_seen": 1112866816
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042614668053716657,
      "loss": 2.7399,
      "theoretical_loss": 3.6125850769922407,
      "tokens_seen": 1112932352
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004261301305087951,
      "loss": 2.6763,
      "theoretical_loss": 3.612565231298911,
      "tokens_seen": 1112997888
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042611358048042365,
      "loss": 2.7196,
      "theoretical_loss": 3.612545387101285,
      "tokens_seen": 1113063424
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004260970304520522,
      "loss": 2.6336,
      "theoretical_loss": 3.6125255443991615,
      "tokens_seen": 1113128960
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042608048042368067,
      "loss": 2.7468,
      "theoretical_loss": 3.61250570319234,
      "tokens_seen": 1113194496
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004260639303953092,
      "loss": 2.8512,
      "theoretical_loss": 3.6124858634806194,
      "tokens_seen": 1113260032
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042604738036693774,
      "loss": 2.818,
      "theoretical_loss": 3.612466025263799,
      "tokens_seen": 1113325568
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004260308303385663,
      "loss": 2.7515,
      "theoretical_loss": 3.612446188541679,
      "tokens_seen": 1113391104
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042601428031019476,
      "loss": 2.6321,
      "theoretical_loss": 3.612426353314058,
      "tokens_seen": 1113456640
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004259977302818233,
      "loss": 2.5135,
      "theoretical_loss": 3.6124065195807358,
      "tokens_seen": 1113522176
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042598118025345184,
      "loss": 2.5086,
      "theoretical_loss": 3.6123866873415116,
      "tokens_seen": 1113587712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004259646302250804,
      "loss": 2.6224,
      "theoretical_loss": 3.612366856596185,
      "tokens_seen": 1113653248
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042594808019670886,
      "loss": 2.4301,
      "theoretical_loss": 3.612347027344556,
      "tokens_seen": 1113718784
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004259315301683374,
      "loss": 2.6706,
      "theoretical_loss": 3.6123271995864235,
      "tokens_seen": 1113784320
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042591498013996594,
      "loss": 2.7724,
      "theoretical_loss": 3.6123073733215882,
      "tokens_seen": 1113849856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004258984301115945,
      "loss": 2.7325,
      "theoretical_loss": 3.6122875485498485,
      "tokens_seen": 1113915392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042588188008322296,
      "loss": 2.6603,
      "theoretical_loss": 3.612267725271005,
      "tokens_seen": 1113980928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004258653300548515,
      "loss": 2.5857,
      "theoretical_loss": 3.6122479034848576,
      "tokens_seen": 1114046464
    },
    {
      "epoch": 0.4,
      "objective/train/avg_token_score": 0.037937842309474945,
      "objective/train/avg_weight": 0.9553614258766174,
      "objective/train/docs_used": 633386,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6363205909729004,
      "objective/train/std_weight": 0.2054106742143631,
      "objective/train/theoretical_loss": 3.612228083191205,
      "objective/train/tokens_used": 1134572000,
      "theoretical_loss": 3.612228083191205,
      "tokens_seen": 1114112000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042584878002648003,
      "loss": 2.5687,
      "theoretical_loss": 3.612228083191205,
      "tokens_seen": 1114112000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042583222999810857,
      "loss": 2.923,
      "theoretical_loss": 3.6122082643898485,
      "tokens_seen": 1114177536
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042581567996973705,
      "loss": 2.7146,
      "theoretical_loss": 3.6121884470805865,
      "tokens_seen": 1114243072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004257991299413656,
      "loss": 2.7912,
      "theoretical_loss": 3.61216863126322,
      "tokens_seen": 1114308608
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042578257991299413,
      "loss": 2.6838,
      "theoretical_loss": 3.612148816937548,
      "tokens_seen": 1114374144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042576602988462267,
      "loss": 2.8461,
      "theoretical_loss": 3.6121290041033722,
      "tokens_seen": 1114439680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042574947985625115,
      "loss": 2.8298,
      "theoretical_loss": 3.6121091927604905,
      "tokens_seen": 1114505216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004257329298278797,
      "loss": 2.7777,
      "theoretical_loss": 3.612089382908705,
      "tokens_seen": 1114570752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004257163797995082,
      "loss": 2.5581,
      "theoretical_loss": 3.6120695745478137,
      "tokens_seen": 1114636288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042569982977113676,
      "loss": 3.0284,
      "theoretical_loss": 3.612049767677618,
      "tokens_seen": 1114701824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042568327974276524,
      "loss": 2.875,
      "theoretical_loss": 3.6120299622979184,
      "tokens_seen": 1114767360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004256667297143938,
      "loss": 2.6959,
      "theoretical_loss": 3.6120101584085145,
      "tokens_seen": 1114832896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004256501796860223,
      "loss": 2.7006,
      "theoretical_loss": 3.6119903560092066,
      "tokens_seen": 1114898432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042563362965765086,
      "loss": 2.7118,
      "theoretical_loss": 3.611970555099795,
      "tokens_seen": 1114963968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042561707962927934,
      "loss": 2.6289,
      "theoretical_loss": 3.6119507556800805,
      "tokens_seen": 1115029504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004256005296009079,
      "loss": 2.5571,
      "theoretical_loss": 3.611930957749863,
      "tokens_seen": 1115095040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004255839795725364,
      "loss": 2.9942,
      "theoretical_loss": 3.611911161308943,
      "tokens_seen": 1115160576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042556742954416495,
      "loss": 2.6818,
      "theoretical_loss": 3.6118913663571206,
      "tokens_seen": 1115226112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042555087951579344,
      "loss": 2.6624,
      "theoretical_loss": 3.6118715728941972,
      "tokens_seen": 1115291648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000425534329487422,
      "loss": 2.7204,
      "theoretical_loss": 3.611851780919973,
      "tokens_seen": 1115357184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004255177794590505,
      "loss": 2.5708,
      "theoretical_loss": 3.611831990434248,
      "tokens_seen": 1115422720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042550122943067905,
      "loss": 2.8215,
      "theoretical_loss": 3.611812201436823,
      "tokens_seen": 1115488256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004254846794023076,
      "loss": 2.7894,
      "theoretical_loss": 3.611792413927499,
      "tokens_seen": 1115553792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042546812937393607,
      "loss": 2.4659,
      "theoretical_loss": 3.6117726279060767,
      "tokens_seen": 1115619328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004254515793455646,
      "loss": 2.8064,
      "theoretical_loss": 3.611752843372356,
      "tokens_seen": 1115684864
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042543502931719315,
      "loss": 2.8911,
      "theoretical_loss": 3.611733060326139,
      "tokens_seen": 1115750400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004254184792888217,
      "loss": 2.8622,
      "theoretical_loss": 3.6117132787672257,
      "tokens_seen": 1115815936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042540192926045017,
      "loss": 2.7281,
      "theoretical_loss": 3.611693498695417,
      "tokens_seen": 1115881472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004253853792320787,
      "loss": 2.5073,
      "theoretical_loss": 3.6116737201105136,
      "tokens_seen": 1115947008
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042536882920370724,
      "loss": 2.8198,
      "theoretical_loss": 3.6116539430123167,
      "tokens_seen": 1116012544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004253522791753358,
      "loss": 2.6821,
      "theoretical_loss": 3.611634167400627,
      "tokens_seen": 1116078080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042533572914696426,
      "loss": 2.5232,
      "theoretical_loss": 3.6116143932752465,
      "tokens_seen": 1116143616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004253191791185928,
      "loss": 2.6534,
      "theoretical_loss": 3.611594620635975,
      "tokens_seen": 1116209152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042530262909022134,
      "loss": 2.5708,
      "theoretical_loss": 3.6115748494826136,
      "tokens_seen": 1116274688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004252860790618499,
      "loss": 2.4824,
      "theoretical_loss": 3.6115550798149645,
      "tokens_seen": 1116340224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042526952903347836,
      "loss": 2.8726,
      "theoretical_loss": 3.6115353116328275,
      "tokens_seen": 1116405760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004252529790051069,
      "loss": 2.8996,
      "theoretical_loss": 3.6115155449360046,
      "tokens_seen": 1116471296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042523642897673543,
      "loss": 2.8477,
      "theoretical_loss": 3.611495779724297,
      "tokens_seen": 1116536832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042521987894836397,
      "loss": 2.7804,
      "theoretical_loss": 3.6114760159975052,
      "tokens_seen": 1116602368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004252033289199924,
      "loss": 2.7777,
      "theoretical_loss": 3.6114562537554318,
      "tokens_seen": 1116667904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042518677889162094,
      "loss": 2.8525,
      "theoretical_loss": 3.611436492997877,
      "tokens_seen": 1116733440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004251702288632494,
      "loss": 2.5965,
      "theoretical_loss": 3.611416733724643,
      "tokens_seen": 1116798976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042515367883487796,
      "loss": 2.659,
      "theoretical_loss": 3.6113969759355307,
      "tokens_seen": 1116864512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004251371288065065,
      "loss": 2.6385,
      "theoretical_loss": 3.6113772196303415,
      "tokens_seen": 1116930048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042512057877813503,
      "loss": 2.7253,
      "theoretical_loss": 3.6113574648088766,
      "tokens_seen": 1116995584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004251040287497635,
      "loss": 2.8181,
      "theoretical_loss": 3.611337711470939,
      "tokens_seen": 1117061120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042508747872139205,
      "loss": 2.7838,
      "theoretical_loss": 3.6113179596163283,
      "tokens_seen": 1117126656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004250709286930206,
      "loss": 2.7801,
      "theoretical_loss": 3.611298209244848,
      "tokens_seen": 1117192192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042505437866464913,
      "loss": 2.8328,
      "theoretical_loss": 3.611278460356298,
      "tokens_seen": 1117257728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004250378286362776,
      "loss": 2.7528,
      "theoretical_loss": 3.611258712950481,
      "tokens_seen": 1117323264
    },
    {
      "epoch": 0.4,
      "objective/train/avg_token_score": 0.016785919666290283,
      "objective/train/avg_weight": 0.9802464842796326,
      "objective/train/docs_used": 634684,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.164474010467529,
      "objective/train/std_weight": 0.10775027424097061,
      "objective/train/theoretical_loss": 3.611238967027199,
      "objective/train/tokens_used": 1137848800,
      "theoretical_loss": 3.611238967027199,
      "tokens_seen": 1117388800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042502127860790615,
      "loss": 2.7561,
      "theoretical_loss": 3.611238967027199,
      "tokens_seen": 1117388800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004250047285795347,
      "loss": 2.8785,
      "theoretical_loss": 3.6112192225862527,
      "tokens_seen": 1117454336
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004249881785511632,
      "loss": 2.4779,
      "theoretical_loss": 3.6111994796274445,
      "tokens_seen": 1117519872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004249716285227917,
      "loss": 2.8723,
      "theoretical_loss": 3.611179738150576,
      "tokens_seen": 1117585408
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042495507849442025,
      "loss": 2.8587,
      "theoretical_loss": 3.6111599981554496,
      "tokens_seen": 1117650944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004249385284660488,
      "loss": 2.5789,
      "theoretical_loss": 3.611140259641867,
      "tokens_seen": 1117716480
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004249219784376773,
      "loss": 2.8784,
      "theoretical_loss": 3.61112052260963,
      "tokens_seen": 1117782016
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004249054284093058,
      "loss": 2.7693,
      "theoretical_loss": 3.6111007870585405,
      "tokens_seen": 1117847552
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042488887838093434,
      "loss": 2.5937,
      "theoretical_loss": 3.6110810529884003,
      "tokens_seen": 1117913088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004248723283525629,
      "loss": 2.8969,
      "theoretical_loss": 3.6110613203990116,
      "tokens_seen": 1117978624
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004248557783241914,
      "loss": 2.8041,
      "theoretical_loss": 3.6110415892901773,
      "tokens_seen": 1118044160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004248392282958199,
      "loss": 2.8357,
      "theoretical_loss": 3.6110218596616988,
      "tokens_seen": 1118109696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042482267826744844,
      "loss": 2.973,
      "theoretical_loss": 3.611002131513378,
      "tokens_seen": 1118175232
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000424806128239077,
      "loss": 2.7101,
      "theoretical_loss": 3.610982404845018,
      "tokens_seen": 1118240768
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004247895782107055,
      "loss": 2.9301,
      "theoretical_loss": 3.6109626796564203,
      "tokens_seen": 1118306304
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000424773028182334,
      "loss": 2.6483,
      "theoretical_loss": 3.6109429559473876,
      "tokens_seen": 1118371840
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042475647815396253,
      "loss": 2.5615,
      "theoretical_loss": 3.6109232337177217,
      "tokens_seen": 1118437376
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042473992812559107,
      "loss": 2.7503,
      "theoretical_loss": 3.6109035129672256,
      "tokens_seen": 1118502912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004247233780972196,
      "loss": 2.9269,
      "theoretical_loss": 3.6108837936957014,
      "tokens_seen": 1118568448
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004247068280688481,
      "loss": 2.7678,
      "theoretical_loss": 3.6108640759029518,
      "tokens_seen": 1118633984
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042469027804047663,
      "loss": 2.8542,
      "theoretical_loss": 3.6108443595887785,
      "tokens_seen": 1118699520
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042467372801210517,
      "loss": 2.7606,
      "theoretical_loss": 3.6108246447529853,
      "tokens_seen": 1118765056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004246571779837337,
      "loss": 2.8019,
      "theoretical_loss": 3.610804931395373,
      "tokens_seen": 1118830592
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004246406279553622,
      "loss": 2.9633,
      "theoretical_loss": 3.6107852195157464,
      "tokens_seen": 1118896128
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004246240779269907,
      "loss": 2.9181,
      "theoretical_loss": 3.6107655091139064,
      "tokens_seen": 1118961664
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042460752789861926,
      "loss": 2.9212,
      "theoretical_loss": 3.6107458001896564,
      "tokens_seen": 1119027200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004245909778702478,
      "loss": 2.4702,
      "theoretical_loss": 3.6107260927427984,
      "tokens_seen": 1119092736
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004245744278418763,
      "loss": 2.5891,
      "theoretical_loss": 3.610706386773136,
      "tokens_seen": 1119158272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004245578778135048,
      "loss": 2.805,
      "theoretical_loss": 3.610686682280472,
      "tokens_seen": 1119223808
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042454132778513336,
      "loss": 2.7279,
      "theoretical_loss": 3.6106669792646087,
      "tokens_seen": 1119289344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004245247777567619,
      "loss": 2.6963,
      "theoretical_loss": 3.610647277725349,
      "tokens_seen": 1119354880
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004245082277283904,
      "loss": 2.7891,
      "theoretical_loss": 3.6106275776624956,
      "tokens_seen": 1119420416
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004244916777000189,
      "loss": 2.7696,
      "theoretical_loss": 3.610607879075853,
      "tokens_seen": 1119485952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042447512767164746,
      "loss": 2.6446,
      "theoretical_loss": 3.6105881819652215,
      "tokens_seen": 1119551488
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000424458577643276,
      "loss": 2.8084,
      "theoretical_loss": 3.6105684863304064,
      "tokens_seen": 1119617024
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004244420276149045,
      "loss": 2.9014,
      "theoretical_loss": 3.6105487921712096,
      "tokens_seen": 1119682560
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000424425477586533,
      "loss": 2.7963,
      "theoretical_loss": 3.610529099487435,
      "tokens_seen": 1119748096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042440892755816155,
      "loss": 2.6688,
      "theoretical_loss": 3.6105094082788844,
      "tokens_seen": 1119813632
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004243923775297901,
      "loss": 2.6267,
      "theoretical_loss": 3.610489718545362,
      "tokens_seen": 1119879168
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042437582750141857,
      "loss": 2.7844,
      "theoretical_loss": 3.6104700302866712,
      "tokens_seen": 1119944704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004243592774730471,
      "loss": 2.9385,
      "theoretical_loss": 3.610450343502614,
      "tokens_seen": 1120010240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042434272744467565,
      "loss": 2.6647,
      "theoretical_loss": 3.6104306581929952,
      "tokens_seen": 1120075776
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004243261774163042,
      "loss": 2.7522,
      "theoretical_loss": 3.6104109743576176,
      "tokens_seen": 1120141312
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042430962738793267,
      "loss": 2.9119,
      "theoretical_loss": 3.610391291996284,
      "tokens_seen": 1120206848
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004242930773595612,
      "loss": 2.8938,
      "theoretical_loss": 3.610371611108798,
      "tokens_seen": 1120272384
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042427652733118974,
      "loss": 2.8669,
      "theoretical_loss": 3.6103519316949626,
      "tokens_seen": 1120337920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004242599773028183,
      "loss": 2.886,
      "theoretical_loss": 3.6103322537545828,
      "tokens_seen": 1120403456
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042424342727444676,
      "loss": 2.8388,
      "theoretical_loss": 3.610312577287461,
      "tokens_seen": 1120468992
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042422687724607525,
      "loss": 2.6537,
      "theoretical_loss": 3.6102929022934003,
      "tokens_seen": 1120534528
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004242103272177038,
      "loss": 2.7896,
      "theoretical_loss": 3.610273228772205,
      "tokens_seen": 1120600064
    },
    {
      "epoch": 0.4,
      "objective/train/avg_token_score": 0.01670733280479908,
      "objective/train/avg_weight": 0.9803562164306641,
      "objective/train/docs_used": 636549,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1677310466766357,
      "objective/train/std_weight": 0.0908128097653389,
      "objective/train/theoretical_loss": 3.610253556723679,
      "objective/train/tokens_used": 1141125600,
      "theoretical_loss": 3.610253556723679,
      "tokens_seen": 1120665600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042419377718933227,
      "loss": 2.8913,
      "theoretical_loss": 3.610253556723679,
      "tokens_seen": 1120665600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004241772271609608,
      "loss": 2.8982,
      "theoretical_loss": 3.610233886147625,
      "tokens_seen": 1120731136
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042416067713258934,
      "loss": 2.8198,
      "theoretical_loss": 3.6102142170438474,
      "tokens_seen": 1120796672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004241441271042179,
      "loss": 2.7724,
      "theoretical_loss": 3.61019454941215,
      "tokens_seen": 1120862208
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042412757707584636,
      "loss": 2.718,
      "theoretical_loss": 3.610174883252336,
      "tokens_seen": 1120927744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004241110270474749,
      "loss": 2.6279,
      "theoretical_loss": 3.610155218564209,
      "tokens_seen": 1120993280
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042409447701910344,
      "loss": 2.7736,
      "theoretical_loss": 3.610135555347574,
      "tokens_seen": 1121058816
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000424077926990732,
      "loss": 2.6458,
      "theoretical_loss": 3.610115893602234,
      "tokens_seen": 1121124352
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042406137696236046,
      "loss": 2.6066,
      "theoretical_loss": 3.610096233327993,
      "tokens_seen": 1121189888
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000424044826933989,
      "loss": 2.9894,
      "theoretical_loss": 3.6100765745246552,
      "tokens_seen": 1121255424
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042402827690561754,
      "loss": 2.9383,
      "theoretical_loss": 3.6100569171920247,
      "tokens_seen": 1121320960
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004240117268772461,
      "loss": 2.6946,
      "theoretical_loss": 3.6100372613299045,
      "tokens_seen": 1121386496
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042399517684887456,
      "loss": 2.5557,
      "theoretical_loss": 3.6100176069381003,
      "tokens_seen": 1121452032
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004239786268205031,
      "loss": 2.5978,
      "theoretical_loss": 3.609997954016415,
      "tokens_seen": 1121517568
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042396207679213163,
      "loss": 2.8387,
      "theoretical_loss": 3.6099783025646532,
      "tokens_seen": 1121583104
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042394552676376017,
      "loss": 2.745,
      "theoretical_loss": 3.609958652582619,
      "tokens_seen": 1121648640
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042392897673538865,
      "loss": 2.8323,
      "theoretical_loss": 3.6099390040701165,
      "tokens_seen": 1121714176
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004239124267070172,
      "loss": 2.8748,
      "theoretical_loss": 3.60991935702695,
      "tokens_seen": 1121779712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042389587667864573,
      "loss": 2.7981,
      "theoretical_loss": 3.6098997114529237,
      "tokens_seen": 1121845248
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042387932665027426,
      "loss": 2.4913,
      "theoretical_loss": 3.609880067347842,
      "tokens_seen": 1121910784
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042386277662190275,
      "loss": 2.7616,
      "theoretical_loss": 3.60986042471151,
      "tokens_seen": 1121976320
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004238462265935313,
      "loss": 2.6006,
      "theoretical_loss": 3.6098407835437305,
      "tokens_seen": 1122041856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004238296765651598,
      "loss": 2.4964,
      "theoretical_loss": 3.609821143844309,
      "tokens_seen": 1122107392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042381312653678836,
      "loss": 2.7342,
      "theoretical_loss": 3.6098015056130506,
      "tokens_seen": 1122172928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042379657650841684,
      "loss": 2.6995,
      "theoretical_loss": 3.6097818688497583,
      "tokens_seen": 1122238464
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004237800264800454,
      "loss": 2.6769,
      "theoretical_loss": 3.609762233554238,
      "tokens_seen": 1122304000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004237634764516739,
      "loss": 2.9571,
      "theoretical_loss": 3.6097425997262933,
      "tokens_seen": 1122369536
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042374692642330246,
      "loss": 2.7786,
      "theoretical_loss": 3.6097229673657294,
      "tokens_seen": 1122435072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042373037639493094,
      "loss": 2.7864,
      "theoretical_loss": 3.6097033364723505,
      "tokens_seen": 1122500608
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004237138263665595,
      "loss": 2.6731,
      "theoretical_loss": 3.6096837070459618,
      "tokens_seen": 1122566144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000423697276338188,
      "loss": 2.7118,
      "theoretical_loss": 3.609664079086368,
      "tokens_seen": 1122631680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042368072630981655,
      "loss": 2.6672,
      "theoretical_loss": 3.6096444525933737,
      "tokens_seen": 1122697216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042366417628144504,
      "loss": 2.9042,
      "theoretical_loss": 3.6096248275667833,
      "tokens_seen": 1122762752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004236476262530736,
      "loss": 2.7325,
      "theoretical_loss": 3.6096052040064026,
      "tokens_seen": 1122828288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004236310762247021,
      "loss": 2.6686,
      "theoretical_loss": 3.609585581912036,
      "tokens_seen": 1122893824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042361452619633065,
      "loss": 2.7126,
      "theoretical_loss": 3.6095659612834883,
      "tokens_seen": 1122959360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042359797616795913,
      "loss": 2.7004,
      "theoretical_loss": 3.6095463421205642,
      "tokens_seen": 1123024896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042358142613958767,
      "loss": 2.9892,
      "theoretical_loss": 3.609526724423069,
      "tokens_seen": 1123090432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004235648761112162,
      "loss": 2.7331,
      "theoretical_loss": 3.609507108190808,
      "tokens_seen": 1123155968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042354832608284474,
      "loss": 2.7708,
      "theoretical_loss": 3.609487493423586,
      "tokens_seen": 1123221504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042353177605447323,
      "loss": 2.7632,
      "theoretical_loss": 3.609467880121209,
      "tokens_seen": 1123287040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042351522602610177,
      "loss": 2.6162,
      "theoretical_loss": 3.6094482682834803,
      "tokens_seen": 1123352576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004234986759977303,
      "loss": 2.9032,
      "theoretical_loss": 3.6094286579102066,
      "tokens_seen": 1123418112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042348212596935884,
      "loss": 2.7161,
      "theoretical_loss": 3.6094090490011927,
      "tokens_seen": 1123483648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004234655759409873,
      "loss": 2.7106,
      "theoretical_loss": 3.6093894415562437,
      "tokens_seen": 1123549184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042344902591261586,
      "loss": 2.6984,
      "theoretical_loss": 3.6093698355751647,
      "tokens_seen": 1123614720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004234324758842444,
      "loss": 2.6045,
      "theoretical_loss": 3.609350231057762,
      "tokens_seen": 1123680256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042341592585587294,
      "loss": 2.9235,
      "theoretical_loss": 3.60933062800384,
      "tokens_seen": 1123745792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004233993758275014,
      "loss": 2.68,
      "theoretical_loss": 3.609311026413204,
      "tokens_seen": 1123811328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042338282579912996,
      "loss": 2.9298,
      "theoretical_loss": 3.609291426285661,
      "tokens_seen": 1123876864
    },
    {
      "epoch": 0.4,
      "objective/train/avg_token_score": 0.012252209708094597,
      "objective/train/avg_weight": 0.9855793118476868,
      "objective/train/docs_used": 638453,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8035435676574707,
      "objective/train/std_weight": 0.07573553919792175,
      "objective/train/theoretical_loss": 3.609271827621014,
      "objective/train/tokens_used": 1144402400,
      "theoretical_loss": 3.609271827621014,
      "tokens_seen": 1123942400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004233662757707585,
      "loss": 2.6711,
      "theoretical_loss": 3.609271827621014,
      "tokens_seen": 1123942400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042334972574238703,
      "loss": 2.867,
      "theoretical_loss": 3.609252230419071,
      "tokens_seen": 1124007936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004233331757140155,
      "loss": 2.6621,
      "theoretical_loss": 3.6092326346796355,
      "tokens_seen": 1124073472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042331662568564405,
      "loss": 2.9773,
      "theoretical_loss": 3.6092130404025147,
      "tokens_seen": 1124139008
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004233000756572726,
      "loss": 2.7483,
      "theoretical_loss": 3.6091934475875136,
      "tokens_seen": 1124204544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042328352562890113,
      "loss": 2.8304,
      "theoretical_loss": 3.6091738562344378,
      "tokens_seen": 1124270080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004232669756005296,
      "loss": 2.8343,
      "theoretical_loss": 3.6091542663430927,
      "tokens_seen": 1124335616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042325042557215815,
      "loss": 2.657,
      "theoretical_loss": 3.6091346779132847,
      "tokens_seen": 1124401152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004232338755437866,
      "loss": 2.6994,
      "theoretical_loss": 3.6091150909448193,
      "tokens_seen": 1124466688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004232173255154151,
      "loss": 2.6317,
      "theoretical_loss": 3.6090955054375025,
      "tokens_seen": 1124532224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042320077548704365,
      "loss": 3.081,
      "theoretical_loss": 3.6090759213911396,
      "tokens_seen": 1124597760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004231842254586722,
      "loss": 2.8344,
      "theoretical_loss": 3.6090563388055372,
      "tokens_seen": 1124663296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004231676754303007,
      "loss": 2.7284,
      "theoretical_loss": 3.6090367576805007,
      "tokens_seen": 1124728832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004231511254019292,
      "loss": 2.6798,
      "theoretical_loss": 3.6090171780158364,
      "tokens_seen": 1124794368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042313457537355775,
      "loss": 2.6662,
      "theoretical_loss": 3.60899759981135,
      "tokens_seen": 1124859904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004231180253451863,
      "loss": 2.5576,
      "theoretical_loss": 3.6089780230668476,
      "tokens_seen": 1124925440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042310147531681477,
      "loss": 2.703,
      "theoretical_loss": 3.6089584477821357,
      "tokens_seen": 1124990976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004230849252884433,
      "loss": 2.718,
      "theoretical_loss": 3.60893887395702,
      "tokens_seen": 1125056512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042306837526007185,
      "loss": 2.7512,
      "theoretical_loss": 3.608919301591307,
      "tokens_seen": 1125122048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004230518252317004,
      "loss": 2.7915,
      "theoretical_loss": 3.608899730684802,
      "tokens_seen": 1125187584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042303527520332887,
      "loss": 2.8765,
      "theoretical_loss": 3.6088801612373125,
      "tokens_seen": 1125253120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004230187251749574,
      "loss": 2.6047,
      "theoretical_loss": 3.608860593248644,
      "tokens_seen": 1125318656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042300217514658594,
      "loss": 2.7298,
      "theoretical_loss": 3.608841026718603,
      "tokens_seen": 1125384192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004229856251182145,
      "loss": 2.7439,
      "theoretical_loss": 3.608821461646995,
      "tokens_seen": 1125449728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000422969075089843,
      "loss": 2.8613,
      "theoretical_loss": 3.6088018980336285,
      "tokens_seen": 1125515264
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004229525250614715,
      "loss": 2.795,
      "theoretical_loss": 3.608782335878308,
      "tokens_seen": 1125580800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042293597503310004,
      "loss": 2.7322,
      "theoretical_loss": 3.6087627751808404,
      "tokens_seen": 1125646336
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004229194250047286,
      "loss": 2.7515,
      "theoretical_loss": 3.608743215941032,
      "tokens_seen": 1125711872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004229028749763571,
      "loss": 2.7426,
      "theoretical_loss": 3.6087236581586906,
      "tokens_seen": 1125777408
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004228863249479856,
      "loss": 2.6726,
      "theoretical_loss": 3.608704101833621,
      "tokens_seen": 1125842944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042286977491961413,
      "loss": 2.7289,
      "theoretical_loss": 3.608684546965631,
      "tokens_seen": 1125908480
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042285322489124267,
      "loss": 2.662,
      "theoretical_loss": 3.6086649935545267,
      "tokens_seen": 1125974016
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004228366748628712,
      "loss": 2.7074,
      "theoretical_loss": 3.608645441600115,
      "tokens_seen": 1126039552
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004228201248344997,
      "loss": 2.8105,
      "theoretical_loss": 3.6086258911022018,
      "tokens_seen": 1126105088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042280357480612823,
      "loss": 2.6188,
      "theoretical_loss": 3.6086063420605954,
      "tokens_seen": 1126170624
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042278702477775677,
      "loss": 2.6637,
      "theoretical_loss": 3.6085867944751016,
      "tokens_seen": 1126236160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004227704747493853,
      "loss": 2.7937,
      "theoretical_loss": 3.608567248345527,
      "tokens_seen": 1126301696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004227539247210138,
      "loss": 2.6733,
      "theoretical_loss": 3.6085477036716793,
      "tokens_seen": 1126367232
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004227373746926423,
      "loss": 2.9452,
      "theoretical_loss": 3.6085281604533646,
      "tokens_seen": 1126432768
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042272082466427086,
      "loss": 2.5864,
      "theoretical_loss": 3.60850861869039,
      "tokens_seen": 1126498304
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004227042746358994,
      "loss": 2.8785,
      "theoretical_loss": 3.608489078382563,
      "tokens_seen": 1126563840
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004226877246075279,
      "loss": 2.7282,
      "theoretical_loss": 3.60846953952969,
      "tokens_seen": 1126629376
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004226711745791564,
      "loss": 2.6336,
      "theoretical_loss": 3.6084500021315784,
      "tokens_seen": 1126694912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042265462455078496,
      "loss": 2.6821,
      "theoretical_loss": 3.608430466188035,
      "tokens_seen": 1126760448
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004226380745224135,
      "loss": 2.8153,
      "theoretical_loss": 3.6084109316988666,
      "tokens_seen": 1126825984
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000422621524494042,
      "loss": 2.8189,
      "theoretical_loss": 3.6083913986638816,
      "tokens_seen": 1126891520
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004226049744656705,
      "loss": 3.1116,
      "theoretical_loss": 3.6083718670828864,
      "tokens_seen": 1126957056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042258842443729905,
      "loss": 2.5162,
      "theoretical_loss": 3.6083523369556874,
      "tokens_seen": 1127022592
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004225718744089276,
      "loss": 2.5945,
      "theoretical_loss": 3.608332808282093,
      "tokens_seen": 1127088128
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004225553243805561,
      "loss": 2.8669,
      "theoretical_loss": 3.6083132810619105,
      "tokens_seen": 1127153664
    },
    {
      "epoch": 0.4,
      "objective/train/avg_token_score": 0.008184703066945076,
      "objective/train/avg_weight": 0.9903714656829834,
      "objective/train/docs_used": 640230,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9015262126922607,
      "objective/train/std_weight": 0.04961738362908363,
      "objective/train/theoretical_loss": 3.6082937552949463,
      "objective/train/tokens_used": 1147679200,
      "theoretical_loss": 3.6082937552949463,
      "tokens_seen": 1127219200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004225387743521846,
      "loss": 2.8297,
      "theoretical_loss": 3.6082937552949463,
      "tokens_seen": 1127219200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042252222432381315,
      "loss": 2.812,
      "theoretical_loss": 3.608274230981009,
      "tokens_seen": 1127284736
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004225056742954417,
      "loss": 2.7535,
      "theoretical_loss": 3.608254708119905,
      "tokens_seen": 1127350272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042248912426707017,
      "loss": 2.6942,
      "theoretical_loss": 3.6082351867114424,
      "tokens_seen": 1127415808
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004224725742386987,
      "loss": 2.7009,
      "theoretical_loss": 3.608215666755428,
      "tokens_seen": 1127481344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042245602421032725,
      "loss": 2.8653,
      "theoretical_loss": 3.60819614825167,
      "tokens_seen": 1127546880
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004224394741819558,
      "loss": 2.8283,
      "theoretical_loss": 3.6081766311999757,
      "tokens_seen": 1127612416
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042242292415358427,
      "loss": 2.8828,
      "theoretical_loss": 3.608157115600153,
      "tokens_seen": 1127677952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004224063741252128,
      "loss": 2.532,
      "theoretical_loss": 3.6081376014520083,
      "tokens_seen": 1127743488
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042238982409684134,
      "loss": 2.5762,
      "theoretical_loss": 3.608118088755351,
      "tokens_seen": 1127809024
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004223732740684699,
      "loss": 2.7405,
      "theoretical_loss": 3.6080985775099874,
      "tokens_seen": 1127874560
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042235672404009836,
      "loss": 2.9253,
      "theoretical_loss": 3.6080790677157255,
      "tokens_seen": 1127940096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004223401740117269,
      "loss": 2.7299,
      "theoretical_loss": 3.608059559372374,
      "tokens_seen": 1128005632
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042232362398335544,
      "loss": 2.7293,
      "theoretical_loss": 3.60804005247974,
      "tokens_seen": 1128071168
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000422307073954984,
      "loss": 2.3791,
      "theoretical_loss": 3.6080205470376314,
      "tokens_seen": 1128136704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042229052392661246,
      "loss": 2.7025,
      "theoretical_loss": 3.608001043045856,
      "tokens_seen": 1128202240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000422273973898241,
      "loss": 3.0057,
      "theoretical_loss": 3.607981540504222,
      "tokens_seen": 1128267776
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004222574238698694,
      "loss": 2.5071,
      "theoretical_loss": 3.607962039412537,
      "tokens_seen": 1128333312
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042224087384149796,
      "loss": 2.7469,
      "theoretical_loss": 3.607942539770609,
      "tokens_seen": 1128398848
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004222243238131265,
      "loss": 2.7998,
      "theoretical_loss": 3.607923041578247,
      "tokens_seen": 1128464384
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042220777378475504,
      "loss": 2.7222,
      "theoretical_loss": 3.6079035448352577,
      "tokens_seen": 1128529920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004221912237563835,
      "loss": 2.7254,
      "theoretical_loss": 3.6078840495414495,
      "tokens_seen": 1128595456
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042217467372801206,
      "loss": 2.7161,
      "theoretical_loss": 3.6078645556966316,
      "tokens_seen": 1128660992
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004221581236996406,
      "loss": 2.783,
      "theoretical_loss": 3.607845063300611,
      "tokens_seen": 1128726528
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042214157367126913,
      "loss": 2.9191,
      "theoretical_loss": 3.6078255723531965,
      "tokens_seen": 1128792064
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004221250236428976,
      "loss": 2.5616,
      "theoretical_loss": 3.6078060828541956,
      "tokens_seen": 1128857600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042210847361452616,
      "loss": 2.5272,
      "theoretical_loss": 3.607786594803418,
      "tokens_seen": 1128923136
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004220919235861547,
      "loss": 2.7606,
      "theoretical_loss": 3.6077671082006706,
      "tokens_seen": 1128988672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042207537355778323,
      "loss": 2.8097,
      "theoretical_loss": 3.6077476230457624,
      "tokens_seen": 1129054208
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004220588235294117,
      "loss": 2.7677,
      "theoretical_loss": 3.6077281393385023,
      "tokens_seen": 1129119744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042204227350104025,
      "loss": 2.7619,
      "theoretical_loss": 3.6077086570786974,
      "tokens_seen": 1129185280
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004220257234726688,
      "loss": 2.5788,
      "theoretical_loss": 3.6076891762661574,
      "tokens_seen": 1129250816
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004220091734442973,
      "loss": 2.7227,
      "theoretical_loss": 3.6076696969006905,
      "tokens_seen": 1129316352
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004219926234159258,
      "loss": 2.7114,
      "theoretical_loss": 3.6076502189821045,
      "tokens_seen": 1129381888
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042197607338755435,
      "loss": 2.6111,
      "theoretical_loss": 3.6076307425102088,
      "tokens_seen": 1129447424
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004219595233591829,
      "loss": 2.806,
      "theoretical_loss": 3.6076112674848115,
      "tokens_seen": 1129512960
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004219429733308114,
      "loss": 2.5522,
      "theoretical_loss": 3.6075917939057223,
      "tokens_seen": 1129578496
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004219264233024399,
      "loss": 2.6969,
      "theoretical_loss": 3.6075723217727487,
      "tokens_seen": 1129644032
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042190987327406844,
      "loss": 2.6247,
      "theoretical_loss": 3.6075528510857,
      "tokens_seen": 1129709568
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000421893323245697,
      "loss": 2.91,
      "theoretical_loss": 3.6075333818443847,
      "tokens_seen": 1129775104
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004218767732173255,
      "loss": 2.3776,
      "theoretical_loss": 3.6075139140486114,
      "tokens_seen": 1129840640
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000421860223188954,
      "loss": 2.7694,
      "theoretical_loss": 3.6074944476981896,
      "tokens_seen": 1129906176
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042184367316058254,
      "loss": 2.5492,
      "theoretical_loss": 3.6074749827929278,
      "tokens_seen": 1129971712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004218271231322111,
      "loss": 2.8075,
      "theoretical_loss": 3.6074555193326345,
      "tokens_seen": 1130037248
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004218105731038396,
      "loss": 2.7834,
      "theoretical_loss": 3.6074360573171194,
      "tokens_seen": 1130102784
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004217940230754681,
      "loss": 2.5801,
      "theoretical_loss": 3.607416596746191,
      "tokens_seen": 1130168320
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042177747304709664,
      "loss": 2.8315,
      "theoretical_loss": 3.6073971376196585,
      "tokens_seen": 1130233856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004217609230187252,
      "loss": 2.8036,
      "theoretical_loss": 3.607377679937331,
      "tokens_seen": 1130299392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004217443729903537,
      "loss": 2.7872,
      "theoretical_loss": 3.6073582236990176,
      "tokens_seen": 1130364928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004217278229619822,
      "loss": 2.8011,
      "theoretical_loss": 3.607338768904527,
      "tokens_seen": 1130430464
    },
    {
      "epoch": 0.4,
      "objective/train/avg_token_score": 0.009804421104490757,
      "objective/train/avg_weight": 0.9884589910507202,
      "objective/train/docs_used": 641368,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7378711700439453,
      "objective/train/std_weight": 0.05072376877069473,
      "objective/train/theoretical_loss": 3.607319315553669,
      "objective/train/tokens_used": 1150956000,
      "theoretical_loss": 3.607319315553669,
      "tokens_seen": 1130496000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042171127293361073,
      "loss": 2.7225,
      "theoretical_loss": 3.607319315553669,
      "tokens_seen": 1130496000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042169472290523927,
      "loss": 2.6582,
      "theoretical_loss": 3.607299863646252,
      "tokens_seen": 1130561536
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004216781728768678,
      "loss": 2.777,
      "theoretical_loss": 3.607280413182086,
      "tokens_seen": 1130627072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004216616228484963,
      "loss": 2.7947,
      "theoretical_loss": 3.60726096416098,
      "tokens_seen": 1130692608
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042164507282012483,
      "loss": 2.6179,
      "theoretical_loss": 3.607241516582743,
      "tokens_seen": 1130758144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042162852279175337,
      "loss": 2.6569,
      "theoretical_loss": 3.607222070447185,
      "tokens_seen": 1130823680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004216119727633819,
      "loss": 2.535,
      "theoretical_loss": 3.607202625754115,
      "tokens_seen": 1130889216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004215954227350104,
      "loss": 2.677,
      "theoretical_loss": 3.6071831825033422,
      "tokens_seen": 1130954752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004215788727066389,
      "loss": 2.628,
      "theoretical_loss": 3.607163740694676,
      "tokens_seen": 1131020288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042156232267826746,
      "loss": 2.8175,
      "theoretical_loss": 3.607144300327927,
      "tokens_seen": 1131085824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000421545772649896,
      "loss": 2.7282,
      "theoretical_loss": 3.6071248614029026,
      "tokens_seen": 1131151360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004215292226215245,
      "loss": 2.7592,
      "theoretical_loss": 3.607105423919415,
      "tokens_seen": 1131216896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000421512672593153,
      "loss": 2.6073,
      "theoretical_loss": 3.607085987877271,
      "tokens_seen": 1131282432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042149612256478156,
      "loss": 2.7272,
      "theoretical_loss": 3.607066553276283,
      "tokens_seen": 1131347968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004214795725364101,
      "loss": 2.7424,
      "theoretical_loss": 3.607047120116259,
      "tokens_seen": 1131413504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042146302250803863,
      "loss": 2.5811,
      "theoretical_loss": 3.6070276883970083,
      "tokens_seen": 1131479040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004214464724796671,
      "loss": 2.8457,
      "theoretical_loss": 3.607008258118342,
      "tokens_seen": 1131544576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042142992245129565,
      "loss": 2.705,
      "theoretical_loss": 3.606988829280069,
      "tokens_seen": 1131610112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004214133724229242,
      "loss": 2.4592,
      "theoretical_loss": 3.606969401881999,
      "tokens_seen": 1131675648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042139682239455273,
      "loss": 2.7791,
      "theoretical_loss": 3.6069499759239427,
      "tokens_seen": 1131741184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004213802723661812,
      "loss": 2.8412,
      "theoretical_loss": 3.6069305514057097,
      "tokens_seen": 1131806720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042136372233780975,
      "loss": 2.6464,
      "theoretical_loss": 3.606911128327109,
      "tokens_seen": 1131872256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004213471723094383,
      "loss": 2.7091,
      "theoretical_loss": 3.606891706687952,
      "tokens_seen": 1131937792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004213306222810668,
      "loss": 2.5972,
      "theoretical_loss": 3.6068722864880476,
      "tokens_seen": 1132003328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004213140722526953,
      "loss": 2.6754,
      "theoretical_loss": 3.6068528677272065,
      "tokens_seen": 1132068864
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042129752222432384,
      "loss": 2.707,
      "theoretical_loss": 3.6068334504052375,
      "tokens_seen": 1132134400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004212809721959524,
      "loss": 2.552,
      "theoretical_loss": 3.6068140345219524,
      "tokens_seen": 1132199936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004212644221675808,
      "loss": 2.5216,
      "theoretical_loss": 3.6067946200771606,
      "tokens_seen": 1132265472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042124787213920935,
      "loss": 2.6407,
      "theoretical_loss": 3.6067752070706725,
      "tokens_seen": 1132331008
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004212313221108379,
      "loss": 2.6569,
      "theoretical_loss": 3.6067557955022975,
      "tokens_seen": 1132396544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042121477208246637,
      "loss": 2.7761,
      "theoretical_loss": 3.606736385371846,
      "tokens_seen": 1132462080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004211982220540949,
      "loss": 2.7337,
      "theoretical_loss": 3.6067169766791296,
      "tokens_seen": 1132527616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042118167202572345,
      "loss": 2.6186,
      "theoretical_loss": 3.6066975694239574,
      "tokens_seen": 1132593152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000421165121997352,
      "loss": 2.8327,
      "theoretical_loss": 3.6066781636061402,
      "tokens_seen": 1132658688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042114857196898047,
      "loss": 2.6995,
      "theoretical_loss": 3.606658759225488,
      "tokens_seen": 1132724224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000421132021940609,
      "loss": 2.8011,
      "theoretical_loss": 3.6066393562818115,
      "tokens_seen": 1132789760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042111547191223754,
      "loss": 2.7087,
      "theoretical_loss": 3.6066199547749216,
      "tokens_seen": 1132855296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004210989218838661,
      "loss": 2.7209,
      "theoretical_loss": 3.606600554704628,
      "tokens_seen": 1132920832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042108237185549456,
      "loss": 2.7018,
      "theoretical_loss": 3.6065811560707415,
      "tokens_seen": 1132986368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004210658218271231,
      "loss": 2.7783,
      "theoretical_loss": 3.6065617588730725,
      "tokens_seen": 1133051904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042104927179875164,
      "loss": 2.6078,
      "theoretical_loss": 3.606542363111432,
      "tokens_seen": 1133117440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004210327217703802,
      "loss": 2.7015,
      "theoretical_loss": 3.6065229687856304,
      "tokens_seen": 1133182976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042101617174200866,
      "loss": 2.65,
      "theoretical_loss": 3.606503575895479,
      "tokens_seen": 1133248512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004209996217136372,
      "loss": 2.6267,
      "theoretical_loss": 3.6064841844407876,
      "tokens_seen": 1133314048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042098307168526573,
      "loss": 2.5469,
      "theoretical_loss": 3.6064647944213672,
      "tokens_seen": 1133379584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042096652165689427,
      "loss": 2.7211,
      "theoretical_loss": 3.6064454058370288,
      "tokens_seen": 1133445120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042094997162852275,
      "loss": 2.6169,
      "theoretical_loss": 3.6064260186875834,
      "tokens_seen": 1133510656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004209334216001513,
      "loss": 2.7198,
      "theoretical_loss": 3.606406632972841,
      "tokens_seen": 1133576192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042091687157177983,
      "loss": 2.6333,
      "theoretical_loss": 3.606387248692614,
      "tokens_seen": 1133641728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042090032154340837,
      "loss": 2.8995,
      "theoretical_loss": 3.6063678658467113,
      "tokens_seen": 1133707264
    },
    {
      "epoch": 0.4,
      "objective/train/avg_token_score": 0.01233475748449564,
      "objective/train/avg_weight": 0.9854807257652283,
      "objective/train/docs_used": 643613,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8653640747070312,
      "objective/train/std_weight": 0.07495511323213577,
      "objective/train/theoretical_loss": 3.6063484844349456,
      "objective/train/tokens_used": 1154232800,
      "theoretical_loss": 3.6063484844349456,
      "tokens_seen": 1133772800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042088377151503685,
      "loss": 2.7127,
      "theoretical_loss": 3.6063484844349456,
      "tokens_seen": 1133772800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004208672214866654,
      "loss": 2.4902,
      "theoretical_loss": 3.606329104457127,
      "tokens_seen": 1133838336
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0004208506714582939,
      "loss": 2.8216,
      "theoretical_loss": 3.606309725913067,
      "tokens_seen": 1133903872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00042083412142992246,
      "loss": 2.6561,
      "theoretical_loss": 3.6062903488025766,
      "tokens_seen": 1133969408
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042081757140155095,
      "loss": 2.7499,
      "theoretical_loss": 3.6062709731254663,
      "tokens_seen": 1134034944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004208010213731795,
      "loss": 2.7035,
      "theoretical_loss": 3.6062515988815482,
      "tokens_seen": 1134100480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000420784471344808,
      "loss": 2.894,
      "theoretical_loss": 3.6062322260706328,
      "tokens_seen": 1134166016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042076792131643656,
      "loss": 2.8063,
      "theoretical_loss": 3.606212854692531,
      "tokens_seen": 1134231552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042075137128806504,
      "loss": 2.6598,
      "theoretical_loss": 3.606193484747055,
      "tokens_seen": 1134297088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004207348212596936,
      "loss": 2.5221,
      "theoretical_loss": 3.606174116234016,
      "tokens_seen": 1134362624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004207182712313221,
      "loss": 2.7638,
      "theoretical_loss": 3.606154749153225,
      "tokens_seen": 1134428160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042070172120295065,
      "loss": 2.6009,
      "theoretical_loss": 3.6061353835044927,
      "tokens_seen": 1134493696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042068517117457914,
      "loss": 2.6511,
      "theoretical_loss": 3.606116019287631,
      "tokens_seen": 1134559232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004206686211462077,
      "loss": 2.7273,
      "theoretical_loss": 3.6060966565024524,
      "tokens_seen": 1134624768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004206520711178362,
      "loss": 2.4477,
      "theoretical_loss": 3.6060772951487667,
      "tokens_seen": 1134690304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042063552108946475,
      "loss": 2.5632,
      "theoretical_loss": 3.606057935226386,
      "tokens_seen": 1134755840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042061897106109323,
      "loss": 2.8712,
      "theoretical_loss": 3.6060385767351226,
      "tokens_seen": 1134821376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042060242103272177,
      "loss": 2.6527,
      "theoretical_loss": 3.6060192196747867,
      "tokens_seen": 1134886912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004205858710043503,
      "loss": 2.8329,
      "theoretical_loss": 3.6059998640451907,
      "tokens_seen": 1134952448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042056932097597885,
      "loss": 2.6197,
      "theoretical_loss": 3.605980509846147,
      "tokens_seen": 1135017984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042055277094760733,
      "loss": 2.5264,
      "theoretical_loss": 3.6059611570774655,
      "tokens_seen": 1135083520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042053622091923587,
      "loss": 2.7332,
      "theoretical_loss": 3.6059418057389596,
      "tokens_seen": 1135149056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004205196708908644,
      "loss": 2.7278,
      "theoretical_loss": 3.6059224558304397,
      "tokens_seen": 1135214592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042050312086249294,
      "loss": 2.7024,
      "theoretical_loss": 3.6059031073517183,
      "tokens_seen": 1135280128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004204865708341214,
      "loss": 2.954,
      "theoretical_loss": 3.6058837603026075,
      "tokens_seen": 1135345664
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042047002080574996,
      "loss": 2.7661,
      "theoretical_loss": 3.6058644146829186,
      "tokens_seen": 1135411200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004204534707773785,
      "loss": 2.6564,
      "theoretical_loss": 3.605845070492464,
      "tokens_seen": 1135476736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042043692074900704,
      "loss": 2.7233,
      "theoretical_loss": 3.6058257277310544,
      "tokens_seen": 1135542272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004204203707206355,
      "loss": 2.7941,
      "theoretical_loss": 3.6058063863985033,
      "tokens_seen": 1135607808
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042040382069226406,
      "loss": 3.3147,
      "theoretical_loss": 3.605787046494622,
      "tokens_seen": 1135673344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004203872706638926,
      "loss": 2.721,
      "theoretical_loss": 3.605767708019222,
      "tokens_seen": 1135738880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042037072063552113,
      "loss": 2.7502,
      "theoretical_loss": 3.6057483709721168,
      "tokens_seen": 1135804416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004203541706071496,
      "loss": 2.9833,
      "theoretical_loss": 3.605729035353117,
      "tokens_seen": 1135869952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042033762057877816,
      "loss": 2.8417,
      "theoretical_loss": 3.605709701162036,
      "tokens_seen": 1135935488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004203210705504067,
      "loss": 2.8412,
      "theoretical_loss": 3.6056903683986854,
      "tokens_seen": 1136001024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042030452052203523,
      "loss": 2.6587,
      "theoretical_loss": 3.6056710370628773,
      "tokens_seen": 1136066560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042028797049366366,
      "loss": 2.7707,
      "theoretical_loss": 3.605651707154424,
      "tokens_seen": 1136132096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004202714204652922,
      "loss": 2.8282,
      "theoretical_loss": 3.6056323786731372,
      "tokens_seen": 1136197632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042025487043692073,
      "loss": 2.7553,
      "theoretical_loss": 3.6056130516188305,
      "tokens_seen": 1136263168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004202383204085492,
      "loss": 2.7839,
      "theoretical_loss": 3.605593725991316,
      "tokens_seen": 1136328704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042022177038017776,
      "loss": 2.6274,
      "theoretical_loss": 3.605574401790405,
      "tokens_seen": 1136394240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004202052203518063,
      "loss": 2.794,
      "theoretical_loss": 3.605555079015911,
      "tokens_seen": 1136459776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042018867032343483,
      "loss": 2.6893,
      "theoretical_loss": 3.6055357576676457,
      "tokens_seen": 1136525312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004201721202950633,
      "loss": 2.8637,
      "theoretical_loss": 3.605516437745422,
      "tokens_seen": 1136590848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042015557026669185,
      "loss": 2.8003,
      "theoretical_loss": 3.6054971192490526,
      "tokens_seen": 1136656384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004201390202383204,
      "loss": 2.7967,
      "theoretical_loss": 3.60547780217835,
      "tokens_seen": 1136721920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004201224702099489,
      "loss": 2.627,
      "theoretical_loss": 3.6054584865331263,
      "tokens_seen": 1136787456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004201059201815774,
      "loss": 2.4157,
      "theoretical_loss": 3.605439172313195,
      "tokens_seen": 1136852992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042008937015320595,
      "loss": 2.8096,
      "theoretical_loss": 3.605419859518368,
      "tokens_seen": 1136918528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004200728201248345,
      "loss": 2.5648,
      "theoretical_loss": 3.6054005481484586,
      "tokens_seen": 1136984064
    },
    {
      "epoch": 0.41,
      "objective/train/avg_token_score": 0.014539276249706745,
      "objective/train/avg_weight": 0.9828861951828003,
      "objective/train/docs_used": 645222,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4794723987579346,
      "objective/train/std_weight": 0.08218029886484146,
      "objective/train/theoretical_loss": 3.605381238203279,
      "objective/train/tokens_used": 1157509600,
      "theoretical_loss": 3.605381238203279,
      "tokens_seen": 1137049600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000420056270096463,
      "loss": 2.5939,
      "theoretical_loss": 3.605381238203279,
      "tokens_seen": 1137049600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004200397200680915,
      "loss": 2.8943,
      "theoretical_loss": 3.6053619296826422,
      "tokens_seen": 1137115136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00042002317003972004,
      "loss": 2.855,
      "theoretical_loss": 3.6053426225863614,
      "tokens_seen": 1137180672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004200066200113486,
      "loss": 2.5799,
      "theoretical_loss": 3.605323316914249,
      "tokens_seen": 1137246208
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004199900699829771,
      "loss": 2.6873,
      "theoretical_loss": 3.6053040126661173,
      "tokens_seen": 1137311744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004199735199546056,
      "loss": 2.8188,
      "theoretical_loss": 3.6052847098417806,
      "tokens_seen": 1137377280
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041995696992623414,
      "loss": 2.7294,
      "theoretical_loss": 3.605265408441051,
      "tokens_seen": 1137442816
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004199404198978627,
      "loss": 2.7851,
      "theoretical_loss": 3.6052461084637413,
      "tokens_seen": 1137508352
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004199238698694912,
      "loss": 2.635,
      "theoretical_loss": 3.6052268099096656,
      "tokens_seen": 1137573888
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004199073198411197,
      "loss": 2.8137,
      "theoretical_loss": 3.605207512778636,
      "tokens_seen": 1137639424
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041989076981274824,
      "loss": 2.5166,
      "theoretical_loss": 3.6051882170704657,
      "tokens_seen": 1137704960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041987421978437677,
      "loss": 2.5868,
      "theoretical_loss": 3.605168922784968,
      "tokens_seen": 1137770496
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004198576697560053,
      "loss": 2.6054,
      "theoretical_loss": 3.605149629921956,
      "tokens_seen": 1137836032
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004198411197276338,
      "loss": 2.5051,
      "theoretical_loss": 3.605130338481243,
      "tokens_seen": 1137901568
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041982456969926233,
      "loss": 2.7257,
      "theoretical_loss": 3.6051110484626423,
      "tokens_seen": 1137967104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041980801967089087,
      "loss": 2.7584,
      "theoretical_loss": 3.605091759865967,
      "tokens_seen": 1138032640
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004197914696425194,
      "loss": 2.8568,
      "theoretical_loss": 3.60507247269103,
      "tokens_seen": 1138098176
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004197749196141479,
      "loss": 2.6163,
      "theoretical_loss": 3.6050531869376456,
      "tokens_seen": 1138163712
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004197583695857764,
      "loss": 2.7337,
      "theoretical_loss": 3.605033902605627,
      "tokens_seen": 1138229248
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041974181955740496,
      "loss": 2.7506,
      "theoretical_loss": 3.6050146196947868,
      "tokens_seen": 1138294784
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004197252695290335,
      "loss": 2.7823,
      "theoretical_loss": 3.6049953382049393,
      "tokens_seen": 1138360320
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000419708719500662,
      "loss": 3.0195,
      "theoretical_loss": 3.6049760581358976,
      "tokens_seen": 1138425856
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004196921694722905,
      "loss": 2.7592,
      "theoretical_loss": 3.604956779487475,
      "tokens_seen": 1138491392
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041967561944391906,
      "loss": 2.7609,
      "theoretical_loss": 3.604937502259485,
      "tokens_seen": 1138556928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004196590694155476,
      "loss": 2.5657,
      "theoretical_loss": 3.6049182264517423,
      "tokens_seen": 1138622464
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004196425193871761,
      "loss": 2.7653,
      "theoretical_loss": 3.60489895206406,
      "tokens_seen": 1138688000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004196259693588046,
      "loss": 2.528,
      "theoretical_loss": 3.6048796790962507,
      "tokens_seen": 1138753536
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041960941933043316,
      "loss": 2.6957,
      "theoretical_loss": 3.604860407548129,
      "tokens_seen": 1138819072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004195928693020617,
      "loss": 2.6701,
      "theoretical_loss": 3.604841137419509,
      "tokens_seen": 1138884608
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004195763192736902,
      "loss": 2.5787,
      "theoretical_loss": 3.6048218687102036,
      "tokens_seen": 1138950144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004195597692453187,
      "loss": 2.5984,
      "theoretical_loss": 3.6048026014200274,
      "tokens_seen": 1139015680
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041954321921694725,
      "loss": 2.7674,
      "theoretical_loss": 3.6047833355487935,
      "tokens_seen": 1139081216
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004195266691885758,
      "loss": 2.8875,
      "theoretical_loss": 3.6047640710963162,
      "tokens_seen": 1139146752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004195101191602043,
      "loss": 2.874,
      "theoretical_loss": 3.604744808062409,
      "tokens_seen": 1139212288
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004194935691318328,
      "loss": 2.5282,
      "theoretical_loss": 3.604725546446887,
      "tokens_seen": 1139277824
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041947701910346135,
      "loss": 2.902,
      "theoretical_loss": 3.604706286249563,
      "tokens_seen": 1139343360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004194604690750899,
      "loss": 2.5959,
      "theoretical_loss": 3.604687027470251,
      "tokens_seen": 1139408896
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041944391904671837,
      "loss": 2.7197,
      "theoretical_loss": 3.604667770108766,
      "tokens_seen": 1139474432
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004194273690183469,
      "loss": 2.7757,
      "theoretical_loss": 3.6046485141649214,
      "tokens_seen": 1139539968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041941081898997544,
      "loss": 2.6825,
      "theoretical_loss": 3.604629259638531,
      "tokens_seen": 1139605504
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000419394268961604,
      "loss": 3.0472,
      "theoretical_loss": 3.60461000652941,
      "tokens_seen": 1139671040
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041937771893323247,
      "loss": 2.73,
      "theoretical_loss": 3.6045907548373712,
      "tokens_seen": 1139736576
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000419361168904861,
      "loss": 2.7653,
      "theoretical_loss": 3.60457150456223,
      "tokens_seen": 1139802112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041934461887648954,
      "loss": 2.703,
      "theoretical_loss": 3.6045522557038003,
      "tokens_seen": 1139867648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004193280688481181,
      "loss": 2.5531,
      "theoretical_loss": 3.604533008261896,
      "tokens_seen": 1139933184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004193115188197465,
      "loss": 2.6956,
      "theoretical_loss": 3.6045137622363326,
      "tokens_seen": 1139998720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041929496879137504,
      "loss": 2.8386,
      "theoretical_loss": 3.6044945176269225,
      "tokens_seen": 1140064256
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041927841876300353,
      "loss": 2.6851,
      "theoretical_loss": 3.604475274433482,
      "tokens_seen": 1140129792
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041926186873463207,
      "loss": 2.6603,
      "theoretical_loss": 3.6044560326558246,
      "tokens_seen": 1140195328
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004192453187062606,
      "loss": 2.6369,
      "theoretical_loss": 3.6044367922937646,
      "tokens_seen": 1140260864
    },
    {
      "epoch": 0.41,
      "objective/train/avg_token_score": 0.008652465417981148,
      "objective/train/avg_weight": 0.9898143410682678,
      "objective/train/docs_used": 646474,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.7525014877319336,
      "objective/train/std_weight": 0.07145651429891586,
      "objective/train/theoretical_loss": 3.604417553347117,
      "objective/train/tokens_used": 1160786400,
      "theoretical_loss": 3.604417553347117,
      "tokens_seen": 1140326400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041922876867788914,
      "loss": 2.8225,
      "theoretical_loss": 3.604417553347117,
      "tokens_seen": 1140326400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004192122186495176,
      "loss": 2.7064,
      "theoretical_loss": 3.604398315815697,
      "tokens_seen": 1140391936
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041919566862114616,
      "loss": 2.6168,
      "theoretical_loss": 3.6043790796993176,
      "tokens_seen": 1140457472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004191791185927747,
      "loss": 2.5471,
      "theoretical_loss": 3.604359844997794,
      "tokens_seen": 1140523008
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041916256856440324,
      "loss": 2.7313,
      "theoretical_loss": 3.604340611710941,
      "tokens_seen": 1140588544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004191460185360317,
      "loss": 2.5957,
      "theoretical_loss": 3.604321379838574,
      "tokens_seen": 1140654080
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041912946850766026,
      "loss": 2.6618,
      "theoretical_loss": 3.6043021493805067,
      "tokens_seen": 1140719616
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004191129184792888,
      "loss": 2.5585,
      "theoretical_loss": 3.604282920336554,
      "tokens_seen": 1140785152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041909636845091733,
      "loss": 2.6439,
      "theoretical_loss": 3.604263692706531,
      "tokens_seen": 1140850688
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004190798184225458,
      "loss": 2.4649,
      "theoretical_loss": 3.6042444664902518,
      "tokens_seen": 1140916224
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041906326839417435,
      "loss": 2.6997,
      "theoretical_loss": 3.604225241687532,
      "tokens_seen": 1140981760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004190467183658029,
      "loss": 2.7641,
      "theoretical_loss": 3.6042060182981865,
      "tokens_seen": 1141047296
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041903016833743143,
      "loss": 2.8619,
      "theoretical_loss": 3.60418679632203,
      "tokens_seen": 1141112832
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004190136183090599,
      "loss": 2.6955,
      "theoretical_loss": 3.6041675757588774,
      "tokens_seen": 1141178368
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041899706828068845,
      "loss": 2.6376,
      "theoretical_loss": 3.604148356608544,
      "tokens_seen": 1141243904
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000418980518252317,
      "loss": 2.6791,
      "theoretical_loss": 3.604129138870844,
      "tokens_seen": 1141309440
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004189639682239455,
      "loss": 2.7249,
      "theoretical_loss": 3.604109922545594,
      "tokens_seen": 1141374976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041894741819557406,
      "loss": 2.663,
      "theoretical_loss": 3.6040907076326074,
      "tokens_seen": 1141440512
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041893086816720255,
      "loss": 2.7981,
      "theoretical_loss": 3.6040714941317002,
      "tokens_seen": 1141506048
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004189143181388311,
      "loss": 2.4706,
      "theoretical_loss": 3.6040522820426877,
      "tokens_seen": 1141571584
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004188977681104596,
      "loss": 2.54,
      "theoretical_loss": 3.6040330713653845,
      "tokens_seen": 1141637120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041888121808208816,
      "loss": 2.7426,
      "theoretical_loss": 3.604013862099607,
      "tokens_seen": 1141702656
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041886466805371664,
      "loss": 2.7557,
      "theoretical_loss": 3.6039946542451684,
      "tokens_seen": 1141768192
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004188481180253452,
      "loss": 2.7018,
      "theoretical_loss": 3.6039754478018864,
      "tokens_seen": 1141833728
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004188315679969737,
      "loss": 2.6131,
      "theoretical_loss": 3.6039562427695744,
      "tokens_seen": 1141899264
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041881501796860225,
      "loss": 2.6443,
      "theoretical_loss": 3.603937039148049,
      "tokens_seen": 1141964800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041879846794023074,
      "loss": 2.8143,
      "theoretical_loss": 3.603917836937125,
      "tokens_seen": 1142030336
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004187819179118593,
      "loss": 2.5321,
      "theoretical_loss": 3.603898636136618,
      "tokens_seen": 1142095872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004187653678834878,
      "loss": 2.6217,
      "theoretical_loss": 3.603879436746344,
      "tokens_seen": 1142161408
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041874881785511635,
      "loss": 2.7529,
      "theoretical_loss": 3.6038602387661176,
      "tokens_seen": 1142226944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041873226782674483,
      "loss": 2.7788,
      "theoretical_loss": 3.603841042195755,
      "tokens_seen": 1142292480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041871571779837337,
      "loss": 2.8433,
      "theoretical_loss": 3.6038218470350714,
      "tokens_seen": 1142358016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004186991677700019,
      "loss": 2.6519,
      "theoretical_loss": 3.603802653283883,
      "tokens_seen": 1142423552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041868261774163045,
      "loss": 2.6858,
      "theoretical_loss": 3.6037834609420045,
      "tokens_seen": 1142489088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041866606771325893,
      "loss": 2.8089,
      "theoretical_loss": 3.6037642700092523,
      "tokens_seen": 1142554624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041864951768488747,
      "loss": 2.9367,
      "theoretical_loss": 3.603745080485442,
      "tokens_seen": 1142620160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000418632967656516,
      "loss": 2.6348,
      "theoretical_loss": 3.603725892370389,
      "tokens_seen": 1142685696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041861641762814454,
      "loss": 2.5768,
      "theoretical_loss": 3.60370670566391,
      "tokens_seen": 1142751232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000418599867599773,
      "loss": 2.7307,
      "theoretical_loss": 3.6036875203658196,
      "tokens_seen": 1142816768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041858331757140156,
      "loss": 2.6286,
      "theoretical_loss": 3.6036683364759345,
      "tokens_seen": 1142882304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004185667675430301,
      "loss": 2.8558,
      "theoretical_loss": 3.6036491539940707,
      "tokens_seen": 1142947840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041855021751465864,
      "loss": 2.8816,
      "theoretical_loss": 3.6036299729200434,
      "tokens_seen": 1143013376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004185336674862871,
      "loss": 2.8325,
      "theoretical_loss": 3.603610793253669,
      "tokens_seen": 1143078912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041851711745791566,
      "loss": 2.8419,
      "theoretical_loss": 3.6035916149947638,
      "tokens_seen": 1143144448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004185005674295442,
      "loss": 2.9285,
      "theoretical_loss": 3.6035724381431433,
      "tokens_seen": 1143209984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041848401740117273,
      "loss": 2.7653,
      "theoretical_loss": 3.6035532626986235,
      "tokens_seen": 1143275520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004184674673728012,
      "loss": 2.5126,
      "theoretical_loss": 3.6035340886610214,
      "tokens_seen": 1143341056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041845091734442975,
      "loss": 2.7913,
      "theoretical_loss": 3.6035149160301523,
      "tokens_seen": 1143406592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004184343673160583,
      "loss": 2.7179,
      "theoretical_loss": 3.6034957448058327,
      "tokens_seen": 1143472128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041841781728768683,
      "loss": 2.7486,
      "theoretical_loss": 3.6034765749878783,
      "tokens_seen": 1143537664
    },
    {
      "epoch": 0.41,
      "objective/train/avg_token_score": 0.01388220302760601,
      "objective/train/avg_weight": 0.9836693406105042,
      "objective/train/docs_used": 648095,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.020151138305664,
      "objective/train/std_weight": 0.05976041033864021,
      "objective/train/theoretical_loss": 3.603457406576106,
      "objective/train/tokens_used": 1164063200,
      "theoretical_loss": 3.603457406576106,
      "tokens_seen": 1143603200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004184012672593153,
      "loss": 2.8372,
      "theoretical_loss": 3.603457406576106,
      "tokens_seen": 1143603200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041838471723094385,
      "loss": 2.7757,
      "theoretical_loss": 3.603438239570332,
      "tokens_seen": 1143668736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004183681672025724,
      "loss": 2.745,
      "theoretical_loss": 3.6034190739703718,
      "tokens_seen": 1143734272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004183516171742009,
      "loss": 2.7926,
      "theoretical_loss": 3.6033999097760434,
      "tokens_seen": 1143799808
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004183350671458294,
      "loss": 2.909,
      "theoretical_loss": 3.6033807469871615,
      "tokens_seen": 1143865344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004183185171174579,
      "loss": 2.304,
      "theoretical_loss": 3.603361585603543,
      "tokens_seen": 1143930880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004183019670890864,
      "loss": 2.7476,
      "theoretical_loss": 3.603342425625005,
      "tokens_seen": 1143996416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004182854170607149,
      "loss": 2.5281,
      "theoretical_loss": 3.603323267051363,
      "tokens_seen": 1144061952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041826886703234345,
      "loss": 2.8372,
      "theoretical_loss": 3.6033041098824343,
      "tokens_seen": 1144127488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000418252317003972,
      "loss": 2.8305,
      "theoretical_loss": 3.603284954118035,
      "tokens_seen": 1144193024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041823576697560047,
      "loss": 2.6393,
      "theoretical_loss": 3.6032657997579824,
      "tokens_seen": 1144258560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000418219216947229,
      "loss": 2.9098,
      "theoretical_loss": 3.6032466468020923,
      "tokens_seen": 1144324096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041820266691885755,
      "loss": 2.4442,
      "theoretical_loss": 3.6032274952501817,
      "tokens_seen": 1144389632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004181861168904861,
      "loss": 2.5855,
      "theoretical_loss": 3.603208345102067,
      "tokens_seen": 1144455168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041816956686211457,
      "loss": 2.8128,
      "theoretical_loss": 3.6031891963575653,
      "tokens_seen": 1144520704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004181530168337431,
      "loss": 2.6289,
      "theoretical_loss": 3.6031700490164926,
      "tokens_seen": 1144586240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041813646680537164,
      "loss": 2.8009,
      "theoretical_loss": 3.6031509030786673,
      "tokens_seen": 1144651776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004181199167770002,
      "loss": 2.6809,
      "theoretical_loss": 3.6031317585439044,
      "tokens_seen": 1144717312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041810336674862866,
      "loss": 2.7031,
      "theoretical_loss": 3.603112615412022,
      "tokens_seen": 1144782848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004180868167202572,
      "loss": 2.4811,
      "theoretical_loss": 3.603093473682837,
      "tokens_seen": 1144848384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041807026669188574,
      "loss": 2.7174,
      "theoretical_loss": 3.6030743333561652,
      "tokens_seen": 1144913920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004180537166635143,
      "loss": 2.7646,
      "theoretical_loss": 3.6030551944318248,
      "tokens_seen": 1144979456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041803716663514276,
      "loss": 2.6841,
      "theoretical_loss": 3.603036056909632,
      "tokens_seen": 1145044992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004180206166067713,
      "loss": 2.8424,
      "theoretical_loss": 3.6030169207894045,
      "tokens_seen": 1145110528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041800406657839983,
      "loss": 2.7943,
      "theoretical_loss": 3.6029977860709588,
      "tokens_seen": 1145176064
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041798751655002837,
      "loss": 2.6112,
      "theoretical_loss": 3.6029786527541123,
      "tokens_seen": 1145241600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041797096652165686,
      "loss": 2.9723,
      "theoretical_loss": 3.6029595208386818,
      "tokens_seen": 1145307136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004179544164932854,
      "loss": 2.8326,
      "theoretical_loss": 3.6029403903244845,
      "tokens_seen": 1145372672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041793786646491393,
      "loss": 2.8761,
      "theoretical_loss": 3.602921261211338,
      "tokens_seen": 1145438208
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041792131643654247,
      "loss": 2.7037,
      "theoretical_loss": 3.6029021334990596,
      "tokens_seen": 1145503744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041790476640817095,
      "loss": 2.8681,
      "theoretical_loss": 3.602883007187466,
      "tokens_seen": 1145569280
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004178882163797995,
      "loss": 2.8425,
      "theoretical_loss": 3.6028638822763748,
      "tokens_seen": 1145634816
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000417871666351428,
      "loss": 2.4719,
      "theoretical_loss": 3.6028447587656034,
      "tokens_seen": 1145700352
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041785511632305656,
      "loss": 2.4299,
      "theoretical_loss": 3.60282563665497,
      "tokens_seen": 1145765888
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041783856629468505,
      "loss": 2.7341,
      "theoretical_loss": 3.60280651594429,
      "tokens_seen": 1145831424
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004178220162663136,
      "loss": 2.564,
      "theoretical_loss": 3.6027873966333823,
      "tokens_seen": 1145896960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004178054662379421,
      "loss": 2.4636,
      "theoretical_loss": 3.6027682787220643,
      "tokens_seen": 1145962496
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041778891620957066,
      "loss": 2.8613,
      "theoretical_loss": 3.6027491622101535,
      "tokens_seen": 1146028032
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041777236618119914,
      "loss": 2.749,
      "theoretical_loss": 3.602730047097467,
      "tokens_seen": 1146093568
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004177558161528277,
      "loss": 2.7871,
      "theoretical_loss": 3.6027109333838228,
      "tokens_seen": 1146159104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004177392661244562,
      "loss": 2.6386,
      "theoretical_loss": 3.602691821069038,
      "tokens_seen": 1146224640
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041772271609608476,
      "loss": 2.7439,
      "theoretical_loss": 3.6026727101529312,
      "tokens_seen": 1146290176
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041770616606771324,
      "loss": 2.7389,
      "theoretical_loss": 3.602653600635319,
      "tokens_seen": 1146355712
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004176896160393418,
      "loss": 2.6729,
      "theoretical_loss": 3.6026344925160196,
      "tokens_seen": 1146421248
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004176730660109703,
      "loss": 2.678,
      "theoretical_loss": 3.602615385794851,
      "tokens_seen": 1146486784
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041765651598259885,
      "loss": 2.7469,
      "theoretical_loss": 3.6025962804716305,
      "tokens_seen": 1146552320
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041763996595422734,
      "loss": 2.724,
      "theoretical_loss": 3.6025771765461765,
      "tokens_seen": 1146617856
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041762341592585587,
      "loss": 2.6919,
      "theoretical_loss": 3.6025580740183063,
      "tokens_seen": 1146683392
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004176068658974844,
      "loss": 2.7018,
      "theoretical_loss": 3.602538972887838,
      "tokens_seen": 1146748928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041759031586911295,
      "loss": 2.6184,
      "theoretical_loss": 3.6025198731545895,
      "tokens_seen": 1146814464
    },
    {
      "epoch": 0.41,
      "objective/train/avg_token_score": 0.022877871990203857,
      "objective/train/avg_weight": 0.9730834364891052,
      "objective/train/docs_used": 649861,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7307114601135254,
      "objective/train/std_weight": 0.11231692135334015,
      "objective/train/theoretical_loss": 3.602500774818379,
      "objective/train/tokens_used": 1167340000,
      "theoretical_loss": 3.602500774818379,
      "tokens_seen": 1146880000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041757376584074143,
      "loss": 2.6687,
      "theoretical_loss": 3.602500774818379,
      "tokens_seen": 1146880000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041755721581236997,
      "loss": 2.6898,
      "theoretical_loss": 3.602481677879024,
      "tokens_seen": 1146945536
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004175406657839985,
      "loss": 2.6253,
      "theoretical_loss": 3.6024625823363428,
      "tokens_seen": 1147011072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041752411575562704,
      "loss": 2.7565,
      "theoretical_loss": 3.602443488190154,
      "tokens_seen": 1147076608
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041750756572725553,
      "loss": 2.6307,
      "theoretical_loss": 3.602424395440275,
      "tokens_seen": 1147142144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041749101569888406,
      "loss": 2.8826,
      "theoretical_loss": 3.602405304086524,
      "tokens_seen": 1147207680
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004174744656705126,
      "loss": 2.8188,
      "theoretical_loss": 3.6023862141287193,
      "tokens_seen": 1147273216
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041745791564214114,
      "loss": 2.6989,
      "theoretical_loss": 3.6023671255666785,
      "tokens_seen": 1147338752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004174413656137697,
      "loss": 2.8365,
      "theoretical_loss": 3.602348038400221,
      "tokens_seen": 1147404288
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041742481558539816,
      "loss": 2.7401,
      "theoretical_loss": 3.6023289526291644,
      "tokens_seen": 1147469824
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004174082655570267,
      "loss": 2.6755,
      "theoretical_loss": 3.602309868253327,
      "tokens_seen": 1147535360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041739171552865524,
      "loss": 2.6351,
      "theoretical_loss": 3.602290785272527,
      "tokens_seen": 1147600896
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004173751655002838,
      "loss": 2.8553,
      "theoretical_loss": 3.6022717036865832,
      "tokens_seen": 1147666432
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041735861547191226,
      "loss": 2.6342,
      "theoretical_loss": 3.6022526234953136,
      "tokens_seen": 1147731968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041734206544354074,
      "loss": 2.936,
      "theoretical_loss": 3.602233544698537,
      "tokens_seen": 1147797504
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004173255154151692,
      "loss": 2.7358,
      "theoretical_loss": 3.6022144672960716,
      "tokens_seen": 1147863040
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041730896538679776,
      "loss": 2.6059,
      "theoretical_loss": 3.602195391287736,
      "tokens_seen": 1147928576
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004172924153584263,
      "loss": 2.793,
      "theoretical_loss": 3.602176316673349,
      "tokens_seen": 1147994112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041727586533005484,
      "loss": 2.7918,
      "theoretical_loss": 3.6021572434527283,
      "tokens_seen": 1148059648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004172593153016833,
      "loss": 3.1251,
      "theoretical_loss": 3.6021381716256933,
      "tokens_seen": 1148125184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041724276527331186,
      "loss": 2.731,
      "theoretical_loss": 3.6021191011920624,
      "tokens_seen": 1148190720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004172262152449404,
      "loss": 2.5805,
      "theoretical_loss": 3.6021000321516543,
      "tokens_seen": 1148256256
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041720966521656893,
      "loss": 2.6139,
      "theoretical_loss": 3.602080964504288,
      "tokens_seen": 1148321792
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004171931151881974,
      "loss": 2.8482,
      "theoretical_loss": 3.602061898249781,
      "tokens_seen": 1148387328
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041717656515982595,
      "loss": 2.494,
      "theoretical_loss": 3.602042833387954,
      "tokens_seen": 1148452864
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004171600151314545,
      "loss": 2.6981,
      "theoretical_loss": 3.602023769918625,
      "tokens_seen": 1148518400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041714346510308303,
      "loss": 2.7647,
      "theoretical_loss": 3.602004707841612,
      "tokens_seen": 1148583936
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004171269150747115,
      "loss": 2.9834,
      "theoretical_loss": 3.6019856471567344,
      "tokens_seen": 1148649472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041711036504634005,
      "loss": 2.7412,
      "theoretical_loss": 3.601966587863812,
      "tokens_seen": 1148715008
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004170938150179686,
      "loss": 2.7679,
      "theoretical_loss": 3.601947529962662,
      "tokens_seen": 1148780544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004170772649895971,
      "loss": 2.6673,
      "theoretical_loss": 3.6019284734531056,
      "tokens_seen": 1148846080
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004170607149612256,
      "loss": 2.8149,
      "theoretical_loss": 3.6019094183349596,
      "tokens_seen": 1148911616
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041704416493285414,
      "loss": 2.8459,
      "theoretical_loss": 3.6018903646080447,
      "tokens_seen": 1148977152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004170276149044827,
      "loss": 2.4377,
      "theoretical_loss": 3.601871312272179,
      "tokens_seen": 1149042688
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004170110648761112,
      "loss": 2.7464,
      "theoretical_loss": 3.601852261327182,
      "tokens_seen": 1149108224
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004169945148477397,
      "loss": 2.7063,
      "theoretical_loss": 3.601833211772873,
      "tokens_seen": 1149173760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041697796481936824,
      "loss": 2.659,
      "theoretical_loss": 3.601814163609071,
      "tokens_seen": 1149239296
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004169614147909968,
      "loss": 2.8486,
      "theoretical_loss": 3.6017951168355946,
      "tokens_seen": 1149304832
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004169448647626253,
      "loss": 2.7988,
      "theoretical_loss": 3.6017760714522637,
      "tokens_seen": 1149370368
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004169283147342538,
      "loss": 2.9184,
      "theoretical_loss": 3.601757027458898,
      "tokens_seen": 1149435904
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041691176470588234,
      "loss": 2.7285,
      "theoretical_loss": 3.6017379848553164,
      "tokens_seen": 1149501440
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004168952146775109,
      "loss": 3.0326,
      "theoretical_loss": 3.6017189436413375,
      "tokens_seen": 1149566976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004168786646491394,
      "loss": 2.8006,
      "theoretical_loss": 3.6016999038167814,
      "tokens_seen": 1149632512
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004168621146207679,
      "loss": 2.8312,
      "theoretical_loss": 3.601680865381468,
      "tokens_seen": 1149698048
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041684556459239643,
      "loss": 2.8471,
      "theoretical_loss": 3.6016618283352164,
      "tokens_seen": 1149763584
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041682901456402497,
      "loss": 2.6291,
      "theoretical_loss": 3.601642792677845,
      "tokens_seen": 1149829120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004168124645356535,
      "loss": 2.6,
      "theoretical_loss": 3.6016237584091755,
      "tokens_seen": 1149894656
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000416795914507282,
      "loss": 2.7473,
      "theoretical_loss": 3.6016047255290253,
      "tokens_seen": 1149960192
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041677936447891053,
      "loss": 2.8225,
      "theoretical_loss": 3.601585694037215,
      "tokens_seen": 1150025728
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041676281445053907,
      "loss": 2.8614,
      "theoretical_loss": 3.601566663933564,
      "tokens_seen": 1150091264
    },
    {
      "epoch": 0.41,
      "objective/train/avg_token_score": 0.02063121646642685,
      "objective/train/avg_weight": 0.9757132530212402,
      "objective/train/docs_used": 651918,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7943034172058105,
      "objective/train/std_weight": 0.12852656841278076,
      "objective/train/theoretical_loss": 3.601547635217892,
      "objective/train/tokens_used": 1170616800,
      "theoretical_loss": 3.601547635217892,
      "tokens_seen": 1150156800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004167462644221676,
      "loss": 2.6491,
      "theoretical_loss": 3.601547635217892,
      "tokens_seen": 1150156800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004167297143937961,
      "loss": 2.6242,
      "theoretical_loss": 3.6015286078900193,
      "tokens_seen": 1150222336
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004167131643654246,
      "loss": 2.8781,
      "theoretical_loss": 3.601509581949765,
      "tokens_seen": 1150287872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041669661433705316,
      "loss": 2.8565,
      "theoretical_loss": 3.6014905573969482,
      "tokens_seen": 1150353408
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004166800643086817,
      "loss": 2.8667,
      "theoretical_loss": 3.60147153423139,
      "tokens_seen": 1150418944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004166635142803102,
      "loss": 2.4718,
      "theoretical_loss": 3.6014525124529095,
      "tokens_seen": 1150484480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004166469642519387,
      "loss": 2.5693,
      "theoretical_loss": 3.601433492061326,
      "tokens_seen": 1150550016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041663041422356726,
      "loss": 2.8443,
      "theoretical_loss": 3.6014144730564612,
      "tokens_seen": 1150615552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004166138641951958,
      "loss": 2.9222,
      "theoretical_loss": 3.6013954554381336,
      "tokens_seen": 1150681088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004165973141668243,
      "loss": 2.8601,
      "theoretical_loss": 3.601376439206163,
      "tokens_seen": 1150746624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004165807641384528,
      "loss": 2.6761,
      "theoretical_loss": 3.60135742436037,
      "tokens_seen": 1150812160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041656421411008135,
      "loss": 2.8235,
      "theoretical_loss": 3.6013384109005746,
      "tokens_seen": 1150877696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004165476640817099,
      "loss": 2.6898,
      "theoretical_loss": 3.601319398826597,
      "tokens_seen": 1150943232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004165311140533384,
      "loss": 2.7467,
      "theoretical_loss": 3.601300388138257,
      "tokens_seen": 1151008768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004165145640249669,
      "loss": 2.9144,
      "theoretical_loss": 3.601281378835374,
      "tokens_seen": 1151074304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041649801399659545,
      "loss": 3.0789,
      "theoretical_loss": 3.6012623709177696,
      "tokens_seen": 1151139840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000416481463968224,
      "loss": 2.8926,
      "theoretical_loss": 3.6012433643852626,
      "tokens_seen": 1151205376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041646491393985247,
      "loss": 2.5138,
      "theoretical_loss": 3.6012243592376745,
      "tokens_seen": 1151270912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000416448363911481,
      "loss": 2.7495,
      "theoretical_loss": 3.601205355474825,
      "tokens_seen": 1151336448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041643181388310955,
      "loss": 2.8998,
      "theoretical_loss": 3.6011863530965345,
      "tokens_seen": 1151401984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004164152638547381,
      "loss": 2.7903,
      "theoretical_loss": 3.601167352102623,
      "tokens_seen": 1151467520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041639871382636657,
      "loss": 2.9015,
      "theoretical_loss": 3.6011483524929107,
      "tokens_seen": 1151533056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004163821637979951,
      "loss": 2.7012,
      "theoretical_loss": 3.601129354267219,
      "tokens_seen": 1151598592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041636561376962364,
      "loss": 2.5256,
      "theoretical_loss": 3.6011103574253673,
      "tokens_seen": 1151664128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041634906374125207,
      "loss": 2.6557,
      "theoretical_loss": 3.6010913619671765,
      "tokens_seen": 1151729664
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004163325137128806,
      "loss": 2.4733,
      "theoretical_loss": 3.601072367892467,
      "tokens_seen": 1151795200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041631596368450915,
      "loss": 2.8663,
      "theoretical_loss": 3.6010533752010594,
      "tokens_seen": 1151860736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004162994136561377,
      "loss": 2.6321,
      "theoretical_loss": 3.6010343838927743,
      "tokens_seen": 1151926272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041628286362776617,
      "loss": 2.7662,
      "theoretical_loss": 3.6010153939674323,
      "tokens_seen": 1151991808
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004162663135993947,
      "loss": 2.7625,
      "theoretical_loss": 3.600996405424854,
      "tokens_seen": 1152057344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041624976357102324,
      "loss": 2.6655,
      "theoretical_loss": 3.6009774182648595,
      "tokens_seen": 1152122880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004162332135426518,
      "loss": 2.8556,
      "theoretical_loss": 3.6009584324872703,
      "tokens_seen": 1152188416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041621666351428026,
      "loss": 2.8123,
      "theoretical_loss": 3.600939448091907,
      "tokens_seen": 1152253952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004162001134859088,
      "loss": 2.7134,
      "theoretical_loss": 3.60092046507859,
      "tokens_seen": 1152319488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041618356345753734,
      "loss": 2.7753,
      "theoretical_loss": 3.6009014834471404,
      "tokens_seen": 1152385024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004161670134291659,
      "loss": 2.8164,
      "theoretical_loss": 3.6008825031973792,
      "tokens_seen": 1152450560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041615046340079436,
      "loss": 2.9455,
      "theoretical_loss": 3.6008635243291263,
      "tokens_seen": 1152516096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004161339133724229,
      "loss": 2.616,
      "theoretical_loss": 3.600844546842204,
      "tokens_seen": 1152581632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041611736334405143,
      "loss": 2.6514,
      "theoretical_loss": 3.6008255707364323,
      "tokens_seen": 1152647168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041610081331567997,
      "loss": 2.5939,
      "theoretical_loss": 3.600806596011632,
      "tokens_seen": 1152712704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041608426328730846,
      "loss": 2.7505,
      "theoretical_loss": 3.6007876226676245,
      "tokens_seen": 1152778240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000416067713258937,
      "loss": 2.8673,
      "theoretical_loss": 3.6007686507042314,
      "tokens_seen": 1152843776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041605116323056553,
      "loss": 2.6664,
      "theoretical_loss": 3.6007496801212726,
      "tokens_seen": 1152909312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041603461320219407,
      "loss": 2.6242,
      "theoretical_loss": 3.60073071091857,
      "tokens_seen": 1152974848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041601806317382255,
      "loss": 2.6419,
      "theoretical_loss": 3.6007117430959443,
      "tokens_seen": 1153040384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004160015131454511,
      "loss": 2.6632,
      "theoretical_loss": 3.6006927766532173,
      "tokens_seen": 1153105920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004159849631170796,
      "loss": 2.7141,
      "theoretical_loss": 3.6006738115902097,
      "tokens_seen": 1153171456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041596841308870816,
      "loss": 2.6268,
      "theoretical_loss": 3.600654847906742,
      "tokens_seen": 1153236992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041595186306033665,
      "loss": 2.7693,
      "theoretical_loss": 3.600635885602637,
      "tokens_seen": 1153302528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004159353130319652,
      "loss": 2.9031,
      "theoretical_loss": 3.6006169246777153,
      "tokens_seen": 1153368064
    },
    {
      "epoch": 0.41,
      "objective/train/avg_token_score": 0.01345257367938757,
      "objective/train/avg_weight": 0.9841749668121338,
      "objective/train/docs_used": 653958,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.550262928009033,
      "objective/train/std_weight": 0.07069134712219238,
      "objective/train/theoretical_loss": 3.6005979651317976,
      "objective/train/tokens_used": 1173893600,
      "theoretical_loss": 3.6005979651317976,
      "tokens_seen": 1153433600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004159187630035937,
      "loss": 2.5348,
      "theoretical_loss": 3.6005979651317976,
      "tokens_seen": 1153433600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041590221297522226,
      "loss": 3.0117,
      "theoretical_loss": 3.6005790069647063,
      "tokens_seen": 1153499136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041588566294685074,
      "loss": 2.5818,
      "theoretical_loss": 3.600560050176262,
      "tokens_seen": 1153564672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004158691129184793,
      "loss": 2.721,
      "theoretical_loss": 3.6005410947662866,
      "tokens_seen": 1153630208
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004158525628901078,
      "loss": 2.8325,
      "theoretical_loss": 3.6005221407346015,
      "tokens_seen": 1153695744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041583601286173636,
      "loss": 2.5955,
      "theoretical_loss": 3.600503188081028,
      "tokens_seen": 1153761280
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041581946283336484,
      "loss": 2.7461,
      "theoretical_loss": 3.6004842368053875,
      "tokens_seen": 1153826816
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004158029128049934,
      "loss": 2.6467,
      "theoretical_loss": 3.600465286907502,
      "tokens_seen": 1153892352
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004157863627766219,
      "loss": 2.7928,
      "theoretical_loss": 3.6004463383871927,
      "tokens_seen": 1153957888
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041576981274825045,
      "loss": 2.7489,
      "theoretical_loss": 3.600427391244282,
      "tokens_seen": 1154023424
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041575326271987893,
      "loss": 2.7003,
      "theoretical_loss": 3.6004084454785907,
      "tokens_seen": 1154088960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041573671269150747,
      "loss": 2.8027,
      "theoretical_loss": 3.600389501089941,
      "tokens_seen": 1154154496
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000415720162663136,
      "loss": 2.6744,
      "theoretical_loss": 3.600370558078154,
      "tokens_seen": 1154220032
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041570361263476455,
      "loss": 3.0041,
      "theoretical_loss": 3.600351616443052,
      "tokens_seen": 1154285568
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041568706260639303,
      "loss": 2.8025,
      "theoretical_loss": 3.6003326761844567,
      "tokens_seen": 1154351104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041567051257802157,
      "loss": 2.6261,
      "theoretical_loss": 3.60031373730219,
      "tokens_seen": 1154416640
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004156539625496501,
      "loss": 2.3473,
      "theoretical_loss": 3.6002947997960737,
      "tokens_seen": 1154482176
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041563741252127864,
      "loss": 2.8244,
      "theoretical_loss": 3.6002758636659293,
      "tokens_seen": 1154547712
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004156208624929071,
      "loss": 2.6764,
      "theoretical_loss": 3.6002569289115796,
      "tokens_seen": 1154613248
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041560431246453566,
      "loss": 2.8096,
      "theoretical_loss": 3.6002379955328454,
      "tokens_seen": 1154678784
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004155877624361642,
      "loss": 2.6818,
      "theoretical_loss": 3.6002190635295497,
      "tokens_seen": 1154744320
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041557121240779274,
      "loss": 3.0267,
      "theoretical_loss": 3.600200132901514,
      "tokens_seen": 1154809856
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004155546623794212,
      "loss": 2.7642,
      "theoretical_loss": 3.6001812036485608,
      "tokens_seen": 1154875392
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041553811235104976,
      "loss": 2.8117,
      "theoretical_loss": 3.600162275770512,
      "tokens_seen": 1154940928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004155215623226783,
      "loss": 2.5475,
      "theoretical_loss": 3.6001433492671895,
      "tokens_seen": 1155006464
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041550501229430684,
      "loss": 2.5845,
      "theoretical_loss": 3.600124424138415,
      "tokens_seen": 1155072000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004154884622659353,
      "loss": 2.8331,
      "theoretical_loss": 3.600105500384012,
      "tokens_seen": 1155137536
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041547191223756386,
      "loss": 2.7822,
      "theoretical_loss": 3.6000865780038023,
      "tokens_seen": 1155203072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004154553622091924,
      "loss": 2.7143,
      "theoretical_loss": 3.600067656997607,
      "tokens_seen": 1155268608
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041543881218082093,
      "loss": 2.5919,
      "theoretical_loss": 3.60004873736525,
      "tokens_seen": 1155334144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004154222621524494,
      "loss": 2.907,
      "theoretical_loss": 3.6000298191065525,
      "tokens_seen": 1155399680
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041540571212407795,
      "loss": 2.791,
      "theoretical_loss": 3.600010902221337,
      "tokens_seen": 1155465216
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004153891620957065,
      "loss": 2.7825,
      "theoretical_loss": 3.5999919867094268,
      "tokens_seen": 1155530752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004153726120673349,
      "loss": 2.9235,
      "theoretical_loss": 3.599973072570643,
      "tokens_seen": 1155596288
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041535606203896346,
      "loss": 2.8158,
      "theoretical_loss": 3.5999541598048097,
      "tokens_seen": 1155661824
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000415339512010592,
      "loss": 2.921,
      "theoretical_loss": 3.5999352484117475,
      "tokens_seen": 1155727360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004153229619822205,
      "loss": 2.7029,
      "theoretical_loss": 3.59991633839128,
      "tokens_seen": 1155792896
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000415306411953849,
      "loss": 2.8561,
      "theoretical_loss": 3.59989742974323,
      "tokens_seen": 1155858432
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041528986192547755,
      "loss": 2.7522,
      "theoretical_loss": 3.599878522467419,
      "tokens_seen": 1155923968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004152733118971061,
      "loss": 2.7402,
      "theoretical_loss": 3.5998596165636707,
      "tokens_seen": 1155989504
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004152567618687346,
      "loss": 2.9762,
      "theoretical_loss": 3.599840712031807,
      "tokens_seen": 1156055040
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004152402118403631,
      "loss": 2.6195,
      "theoretical_loss": 3.599821808871652,
      "tokens_seen": 1156120576
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041522366181199165,
      "loss": 2.6554,
      "theoretical_loss": 3.5998029070830264,
      "tokens_seen": 1156186112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004152071117836202,
      "loss": 2.604,
      "theoretical_loss": 3.5997840066657543,
      "tokens_seen": 1156251648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041519056175524867,
      "loss": 2.7873,
      "theoretical_loss": 3.5997651076196577,
      "tokens_seen": 1156317184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004151740117268772,
      "loss": 2.7945,
      "theoretical_loss": 3.59974620994456,
      "tokens_seen": 1156382720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041515746169850574,
      "loss": 2.9198,
      "theoretical_loss": 3.599727313640284,
      "tokens_seen": 1156448256
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004151409116701343,
      "loss": 2.7898,
      "theoretical_loss": 3.599708418706653,
      "tokens_seen": 1156513792
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041512436164176277,
      "loss": 2.7741,
      "theoretical_loss": 3.599689525143489,
      "tokens_seen": 1156579328
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004151078116133913,
      "loss": 2.802,
      "theoretical_loss": 3.599670632950615,
      "tokens_seen": 1156644864
    },
    {
      "epoch": 0.41,
      "objective/train/avg_token_score": 0.014211218804121017,
      "objective/train/avg_weight": 0.9832714796066284,
      "objective/train/docs_used": 655833,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8430676460266113,
      "objective/train/std_weight": 0.07536811381578445,
      "objective/train/theoretical_loss": 3.599651742127855,
      "objective/train/tokens_used": 1177170400,
      "theoretical_loss": 3.599651742127855,
      "tokens_seen": 1156710400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041509126158501984,
      "loss": 2.5748,
      "theoretical_loss": 3.599651742127855,
      "tokens_seen": 1156710400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004150747115566484,
      "loss": 2.5703,
      "theoretical_loss": 3.5996328526750307,
      "tokens_seen": 1156775936
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041505816152827686,
      "loss": 2.6706,
      "theoretical_loss": 3.5996139645919656,
      "tokens_seen": 1156841472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004150416114999054,
      "loss": 2.4397,
      "theoretical_loss": 3.5995950778784835,
      "tokens_seen": 1156907008
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041502506147153394,
      "loss": 2.6169,
      "theoretical_loss": 3.599576192534407,
      "tokens_seen": 1156972544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004150085114431625,
      "loss": 2.6676,
      "theoretical_loss": 3.5995573085595587,
      "tokens_seen": 1157038080
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000414991961414791,
      "loss": 2.6974,
      "theoretical_loss": 3.599538425953763,
      "tokens_seen": 1157103616
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004149754113864195,
      "loss": 2.5405,
      "theoretical_loss": 3.5995195447168418,
      "tokens_seen": 1157169152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041495886135804803,
      "loss": 2.8334,
      "theoretical_loss": 3.5995006648486196,
      "tokens_seen": 1157234688
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041494231132967657,
      "loss": 2.7326,
      "theoretical_loss": 3.5994817863489184,
      "tokens_seen": 1157300224
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004149257613013051,
      "loss": 2.7041,
      "theoretical_loss": 3.5994629092175625,
      "tokens_seen": 1157365760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004149092112729336,
      "loss": 2.888,
      "theoretical_loss": 3.5994440334543754,
      "tokens_seen": 1157431296
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041489266124456213,
      "loss": 2.8347,
      "theoretical_loss": 3.5994251590591793,
      "tokens_seen": 1157496832
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041487611121619067,
      "loss": 2.8648,
      "theoretical_loss": 3.5994062860317984,
      "tokens_seen": 1157562368
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004148595611878192,
      "loss": 2.6747,
      "theoretical_loss": 3.5993874143720563,
      "tokens_seen": 1157627904
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004148430111594477,
      "loss": 2.6157,
      "theoretical_loss": 3.599368544079776,
      "tokens_seen": 1157693440
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004148264611310762,
      "loss": 2.7888,
      "theoretical_loss": 3.5993496751547815,
      "tokens_seen": 1157758976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041480991110270476,
      "loss": 2.8264,
      "theoretical_loss": 3.599330807596896,
      "tokens_seen": 1157824512
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004147933610743333,
      "loss": 2.6251,
      "theoretical_loss": 3.5993119414059427,
      "tokens_seen": 1157890048
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004147768110459618,
      "loss": 2.6412,
      "theoretical_loss": 3.5992930765817466,
      "tokens_seen": 1157955584
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004147602610175903,
      "loss": 2.7165,
      "theoretical_loss": 3.5992742131241293,
      "tokens_seen": 1158021120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041474371098921886,
      "loss": 2.5313,
      "theoretical_loss": 3.599255351032917,
      "tokens_seen": 1158086656
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004147271609608474,
      "loss": 2.5337,
      "theoretical_loss": 3.5992364903079306,
      "tokens_seen": 1158152192
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004147106109324759,
      "loss": 2.7503,
      "theoretical_loss": 3.599217630948996,
      "tokens_seen": 1158217728
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004146940609041044,
      "loss": 2.8649,
      "theoretical_loss": 3.599198772955936,
      "tokens_seen": 1158283264
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041467751087573295,
      "loss": 2.9822,
      "theoretical_loss": 3.5991799163285743,
      "tokens_seen": 1158348800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004146609608473615,
      "loss": 2.5365,
      "theoretical_loss": 3.5991610610667353,
      "tokens_seen": 1158414336
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041464441081899,
      "loss": 2.7125,
      "theoretical_loss": 3.599142207170243,
      "tokens_seen": 1158479872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004146278607906185,
      "loss": 2.5064,
      "theoretical_loss": 3.5991233546389205,
      "tokens_seen": 1158545408
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041461131076224705,
      "loss": 2.724,
      "theoretical_loss": 3.599104503472592,
      "tokens_seen": 1158610944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004145947607338756,
      "loss": 2.7637,
      "theoretical_loss": 3.599085653671082,
      "tokens_seen": 1158676480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041457821070550407,
      "loss": 2.6004,
      "theoretical_loss": 3.599066805234214,
      "tokens_seen": 1158742016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004145616606771326,
      "loss": 2.604,
      "theoretical_loss": 3.5990479581618118,
      "tokens_seen": 1158807552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041454511064876115,
      "loss": 2.4625,
      "theoretical_loss": 3.5990291124537004,
      "tokens_seen": 1158873088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004145285606203897,
      "loss": 2.8058,
      "theoretical_loss": 3.5990102681097027,
      "tokens_seen": 1158938624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041451201059201817,
      "loss": 2.8138,
      "theoretical_loss": 3.5989914251296438,
      "tokens_seen": 1159004160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004144954605636467,
      "loss": 2.8817,
      "theoretical_loss": 3.5989725835133473,
      "tokens_seen": 1159069696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041447891053527524,
      "loss": 2.6622,
      "theoretical_loss": 3.5989537432606378,
      "tokens_seen": 1159135232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004144623605069038,
      "loss": 2.8055,
      "theoretical_loss": 3.5989349043713394,
      "tokens_seen": 1159200768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041444581047853226,
      "loss": 2.9857,
      "theoretical_loss": 3.598916066845276,
      "tokens_seen": 1159266304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004144292604501608,
      "loss": 2.8052,
      "theoretical_loss": 3.5988972306822724,
      "tokens_seen": 1159331840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041441271042178934,
      "loss": 2.6871,
      "theoretical_loss": 3.5988783958821524,
      "tokens_seen": 1159397376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004143961603934179,
      "loss": 2.7391,
      "theoretical_loss": 3.598859562444741,
      "tokens_seen": 1159462912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004143796103650463,
      "loss": 2.6683,
      "theoretical_loss": 3.5988407303698624,
      "tokens_seen": 1159528448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041436306033667484,
      "loss": 2.6201,
      "theoretical_loss": 3.5988218996573407,
      "tokens_seen": 1159593984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004143465103083033,
      "loss": 2.6783,
      "theoretical_loss": 3.5988030703070004,
      "tokens_seen": 1159659520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041432996027993186,
      "loss": 2.7004,
      "theoretical_loss": 3.5987842423186662,
      "tokens_seen": 1159725056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004143134102515604,
      "loss": 2.5603,
      "theoretical_loss": 3.5987654156921627,
      "tokens_seen": 1159790592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041429686022318894,
      "loss": 2.7172,
      "theoretical_loss": 3.5987465904273144,
      "tokens_seen": 1159856128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004142803101948174,
      "loss": 2.7101,
      "theoretical_loss": 3.598727766523945,
      "tokens_seen": 1159921664
    },
    {
      "epoch": 0.41,
      "objective/train/avg_token_score": 0.010517098009586334,
      "objective/train/avg_weight": 0.9876223206520081,
      "objective/train/docs_used": 657914,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.866899013519287,
      "objective/train/std_weight": 0.06436856091022491,
      "objective/train/theoretical_loss": 3.5987089439818805,
      "objective/train/tokens_used": 1180447200,
      "theoretical_loss": 3.5987089439818805,
      "tokens_seen": 1159987200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041426376016644596,
      "loss": 2.6319,
      "theoretical_loss": 3.5987089439818805,
      "tokens_seen": 1159987200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004142472101380745,
      "loss": 2.6232,
      "theoretical_loss": 3.5986901228009454,
      "tokens_seen": 1160052736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041423066010970303,
      "loss": 2.6369,
      "theoretical_loss": 3.5986713029809634,
      "tokens_seen": 1160118272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004142141100813315,
      "loss": 2.8344,
      "theoretical_loss": 3.59865248452176,
      "tokens_seen": 1160183808
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041419756005296005,
      "loss": 2.7671,
      "theoretical_loss": 3.59863366742316,
      "tokens_seen": 1160249344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004141810100245886,
      "loss": 2.6532,
      "theoretical_loss": 3.598614851684987,
      "tokens_seen": 1160314880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041416445999621713,
      "loss": 2.5178,
      "theoretical_loss": 3.5985960373070673,
      "tokens_seen": 1160380416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004141479099678456,
      "loss": 2.7196,
      "theoretical_loss": 3.598577224289225,
      "tokens_seen": 1160445952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041413135993947415,
      "loss": 2.82,
      "theoretical_loss": 3.5985584126312853,
      "tokens_seen": 1160511488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004141148099111027,
      "loss": 2.6776,
      "theoretical_loss": 3.598539602333073,
      "tokens_seen": 1160577024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004140982598827312,
      "loss": 2.3997,
      "theoretical_loss": 3.598520793394413,
      "tokens_seen": 1160642560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004140817098543597,
      "loss": 2.8988,
      "theoretical_loss": 3.59850198581513,
      "tokens_seen": 1160708096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041406515982598825,
      "loss": 2.8579,
      "theoretical_loss": 3.59848317959505,
      "tokens_seen": 1160773632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004140486097976168,
      "loss": 2.5545,
      "theoretical_loss": 3.5984643747339966,
      "tokens_seen": 1160839168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004140320597692453,
      "loss": 2.651,
      "theoretical_loss": 3.5984455712317955,
      "tokens_seen": 1160904704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004140155097408738,
      "loss": 2.6901,
      "theoretical_loss": 3.598426769088272,
      "tokens_seen": 1160970240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041399895971250234,
      "loss": 2.6536,
      "theoretical_loss": 3.5984079683032517,
      "tokens_seen": 1161035776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004139824096841309,
      "loss": 2.52,
      "theoretical_loss": 3.5983891688765586,
      "tokens_seen": 1161101312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004139658596557594,
      "loss": 2.7493,
      "theoretical_loss": 3.598370370808019,
      "tokens_seen": 1161166848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004139493096273879,
      "loss": 2.6195,
      "theoretical_loss": 3.598351574097457,
      "tokens_seen": 1161232384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041393275959901644,
      "loss": 2.6542,
      "theoretical_loss": 3.598332778744699,
      "tokens_seen": 1161297920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000413916209570645,
      "loss": 2.7328,
      "theoretical_loss": 3.59831398474957,
      "tokens_seen": 1161363456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004138996595422735,
      "loss": 2.6498,
      "theoretical_loss": 3.5982951921118946,
      "tokens_seen": 1161428992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000413883109513902,
      "loss": 2.7077,
      "theoretical_loss": 3.598276400831499,
      "tokens_seen": 1161494528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041386655948553053,
      "loss": 2.8086,
      "theoretical_loss": 3.5982576109082083,
      "tokens_seen": 1161560064
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041385000945715907,
      "loss": 2.3978,
      "theoretical_loss": 3.598238822341848,
      "tokens_seen": 1161625600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004138334594287876,
      "loss": 2.6501,
      "theoretical_loss": 3.5982200351322433,
      "tokens_seen": 1161691136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004138169094004161,
      "loss": 2.71,
      "theoretical_loss": 3.59820124927922,
      "tokens_seen": 1161756672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041380035937204463,
      "loss": 2.6464,
      "theoretical_loss": 3.598182464782604,
      "tokens_seen": 1161822208
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00041378380934367317,
      "loss": 2.6041,
      "theoretical_loss": 3.5981636816422196,
      "tokens_seen": 1161887744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0004137672593153017,
      "loss": 2.6539,
      "theoretical_loss": 3.5981448998578935,
      "tokens_seen": 1161953280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004137507092869302,
      "loss": 2.8597,
      "theoretical_loss": 3.598126119429451,
      "tokens_seen": 1162018816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004137341592585587,
      "loss": 2.6728,
      "theoretical_loss": 3.5981073403567176,
      "tokens_seen": 1162084352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041371760923018726,
      "loss": 2.8536,
      "theoretical_loss": 3.5980885626395196,
      "tokens_seen": 1162149888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004137010592018158,
      "loss": 2.9081,
      "theoretical_loss": 3.598069786277682,
      "tokens_seen": 1162215424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004136845091734443,
      "loss": 2.7562,
      "theoretical_loss": 3.598051011271031,
      "tokens_seen": 1162280960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004136679591450728,
      "loss": 2.8312,
      "theoretical_loss": 3.5980322376193916,
      "tokens_seen": 1162346496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041365140911670136,
      "loss": 2.4889,
      "theoretical_loss": 3.598013465322591,
      "tokens_seen": 1162412032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004136348590883299,
      "loss": 2.5382,
      "theoretical_loss": 3.5979946943804535,
      "tokens_seen": 1162477568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004136183090599584,
      "loss": 2.7167,
      "theoretical_loss": 3.5979759247928067,
      "tokens_seen": 1162543104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004136017590315869,
      "loss": 2.6861,
      "theoretical_loss": 3.597957156559475,
      "tokens_seen": 1162608640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041358520900321546,
      "loss": 2.6462,
      "theoretical_loss": 3.5979383896802846,
      "tokens_seen": 1162674176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000413568658974844,
      "loss": 2.6736,
      "theoretical_loss": 3.597919624155062,
      "tokens_seen": 1162739712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004135521089464725,
      "loss": 2.7094,
      "theoretical_loss": 3.597900859983633,
      "tokens_seen": 1162805248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000413535558918101,
      "loss": 2.6832,
      "theoretical_loss": 3.5978820971658236,
      "tokens_seen": 1162870784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041351900888972955,
      "loss": 2.8177,
      "theoretical_loss": 3.5978633357014598,
      "tokens_seen": 1162936320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004135024588613581,
      "loss": 2.6919,
      "theoretical_loss": 3.5978445755903685,
      "tokens_seen": 1163001856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004134859088329866,
      "loss": 2.6896,
      "theoretical_loss": 3.5978258168323745,
      "tokens_seen": 1163067392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004134693588046151,
      "loss": 2.6529,
      "theoretical_loss": 3.5978070594273044,
      "tokens_seen": 1163132928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041345280877624365,
      "loss": 2.7093,
      "theoretical_loss": 3.597788303374985,
      "tokens_seen": 1163198464
    },
    {
      "epoch": 0.42,
      "objective/train/avg_token_score": 0.01010025292634964,
      "objective/train/avg_weight": 0.9881318211555481,
      "objective/train/docs_used": 659757,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6964457035064697,
      "objective/train/std_weight": 0.08749702572822571,
      "objective/train/theoretical_loss": 3.5977695486752426,
      "objective/train/tokens_used": 1183724000,
      "theoretical_loss": 3.5977695486752426,
      "tokens_seen": 1163264000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004134362587478722,
      "loss": 2.6169,
      "theoretical_loss": 3.5977695486752426,
      "tokens_seen": 1163264000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004134197087195007,
      "loss": 2.8308,
      "theoretical_loss": 3.597750795327902,
      "tokens_seen": 1163329536
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041340315869112915,
      "loss": 2.5799,
      "theoretical_loss": 3.5977320433327913,
      "tokens_seen": 1163395072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004133866086627577,
      "loss": 2.7405,
      "theoretical_loss": 3.597713292689736,
      "tokens_seen": 1163460608
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041337005863438617,
      "loss": 2.6653,
      "theoretical_loss": 3.5976945433985623,
      "tokens_seen": 1163526144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004133535086060147,
      "loss": 2.7035,
      "theoretical_loss": 3.5976757954590965,
      "tokens_seen": 1163591680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041333695857764325,
      "loss": 2.3863,
      "theoretical_loss": 3.597657048871166,
      "tokens_seen": 1163657216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004133204085492718,
      "loss": 2.5107,
      "theoretical_loss": 3.5976383036345965,
      "tokens_seen": 1163722752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041330385852090027,
      "loss": 2.5947,
      "theoretical_loss": 3.597619559749214,
      "tokens_seen": 1163788288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004132873084925288,
      "loss": 2.5512,
      "theoretical_loss": 3.5976008172148464,
      "tokens_seen": 1163853824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041327075846415734,
      "loss": 2.6949,
      "theoretical_loss": 3.597582076031319,
      "tokens_seen": 1163919360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004132542084357859,
      "loss": 2.7354,
      "theoretical_loss": 3.5975633361984594,
      "tokens_seen": 1163984896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041323765840741436,
      "loss": 2.6359,
      "theoretical_loss": 3.5975445977160927,
      "tokens_seen": 1164050432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004132211083790429,
      "loss": 2.4811,
      "theoretical_loss": 3.597525860584047,
      "tokens_seen": 1164115968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041320455835067144,
      "loss": 2.7625,
      "theoretical_loss": 3.597507124802149,
      "tokens_seen": 1164181504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004131880083223,
      "loss": 2.6698,
      "theoretical_loss": 3.5974883903702244,
      "tokens_seen": 1164247040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041317145829392846,
      "loss": 3.078,
      "theoretical_loss": 3.5974696572881006,
      "tokens_seen": 1164312576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000413154908265557,
      "loss": 2.6409,
      "theoretical_loss": 3.597450925555604,
      "tokens_seen": 1164378112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041313835823718554,
      "loss": 2.5578,
      "theoretical_loss": 3.597432195172562,
      "tokens_seen": 1164443648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004131218082088141,
      "loss": 2.7181,
      "theoretical_loss": 3.597413466138801,
      "tokens_seen": 1164509184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041310525818044256,
      "loss": 2.5544,
      "theoretical_loss": 3.597394738454148,
      "tokens_seen": 1164574720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004130887081520711,
      "loss": 2.8743,
      "theoretical_loss": 3.59737601211843,
      "tokens_seen": 1164640256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041307215812369963,
      "loss": 2.8081,
      "theoretical_loss": 3.597357287131474,
      "tokens_seen": 1164705792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041305560809532817,
      "loss": 2.5724,
      "theoretical_loss": 3.597338563493106,
      "tokens_seen": 1164771328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041303905806695665,
      "loss": 2.575,
      "theoretical_loss": 3.5973198412031544,
      "tokens_seen": 1164836864
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004130225080385852,
      "loss": 2.7856,
      "theoretical_loss": 3.5973011202614456,
      "tokens_seen": 1164902400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041300595801021373,
      "loss": 2.8861,
      "theoretical_loss": 3.5972824006678064,
      "tokens_seen": 1164967936
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041298940798184227,
      "loss": 2.5578,
      "theoretical_loss": 3.5972636824220645,
      "tokens_seen": 1165033472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041297285795347075,
      "loss": 2.7593,
      "theoretical_loss": 3.5972449655240473,
      "tokens_seen": 1165099008
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004129563079250993,
      "loss": 2.8758,
      "theoretical_loss": 3.5972262499735805,
      "tokens_seen": 1165164544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004129397578967278,
      "loss": 2.8046,
      "theoretical_loss": 3.597207535770493,
      "tokens_seen": 1165230080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041292320786835636,
      "loss": 2.7139,
      "theoretical_loss": 3.5971888229146103,
      "tokens_seen": 1165295616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041290665783998484,
      "loss": 2.5844,
      "theoretical_loss": 3.5971701114057613,
      "tokens_seen": 1165361152
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004128901078116134,
      "loss": 2.962,
      "theoretical_loss": 3.5971514012437718,
      "tokens_seen": 1165426688
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004128735577832419,
      "loss": 2.729,
      "theoretical_loss": 3.5971326924284703,
      "tokens_seen": 1165492224
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041285700775487046,
      "loss": 2.7219,
      "theoretical_loss": 3.5971139849596843,
      "tokens_seen": 1165557760
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041284045772649894,
      "loss": 2.7049,
      "theoretical_loss": 3.59709527883724,
      "tokens_seen": 1165623296
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004128239076981275,
      "loss": 2.5693,
      "theoretical_loss": 3.5970765740609654,
      "tokens_seen": 1165688832
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000412807357669756,
      "loss": 2.843,
      "theoretical_loss": 3.5970578706306884,
      "tokens_seen": 1165754368
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041279080764138455,
      "loss": 2.8337,
      "theoretical_loss": 3.597039168546236,
      "tokens_seen": 1165819904
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041277425761301304,
      "loss": 2.7035,
      "theoretical_loss": 3.597020467807436,
      "tokens_seen": 1165885440
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004127577075846416,
      "loss": 2.7917,
      "theoretical_loss": 3.597001768414115,
      "tokens_seen": 1165950976
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004127411575562701,
      "loss": 2.8121,
      "theoretical_loss": 3.596983070366102,
      "tokens_seen": 1166016512
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041272460752789865,
      "loss": 2.6388,
      "theoretical_loss": 3.5969643736632237,
      "tokens_seen": 1166082048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041270805749952713,
      "loss": 2.7639,
      "theoretical_loss": 3.5969456783053078,
      "tokens_seen": 1166147584
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041269150747115567,
      "loss": 2.7053,
      "theoretical_loss": 3.5969269842921823,
      "tokens_seen": 1166213120
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004126749574427842,
      "loss": 2.4576,
      "theoretical_loss": 3.5969082916236745,
      "tokens_seen": 1166278656
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041265840741441275,
      "loss": 2.5949,
      "theoretical_loss": 3.5968896002996127,
      "tokens_seen": 1166344192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041264185738604123,
      "loss": 2.6543,
      "theoretical_loss": 3.596870910319824,
      "tokens_seen": 1166409728
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041262530735766977,
      "loss": 2.9176,
      "theoretical_loss": 3.596852221684137,
      "tokens_seen": 1166475264
    },
    {
      "epoch": 0.42,
      "objective/train/avg_token_score": 0.023470789194107056,
      "objective/train/avg_weight": 0.9723694920539856,
      "objective/train/docs_used": 661442,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.020427227020264,
      "objective/train/std_weight": 0.10709500312805176,
      "objective/train/theoretical_loss": 3.596833534392379,
      "objective/train/tokens_used": 1187000800,
      "theoretical_loss": 3.596833534392379,
      "tokens_seen": 1166540800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004126087573292983,
      "loss": 2.9665,
      "theoretical_loss": 3.596833534392379,
      "tokens_seen": 1166540800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041259220730092684,
      "loss": 2.783,
      "theoretical_loss": 3.596814848444378,
      "tokens_seen": 1166606336
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004125756572725553,
      "loss": 2.3937,
      "theoretical_loss": 3.5967961638399615,
      "tokens_seen": 1166671872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041255910724418386,
      "loss": 2.6118,
      "theoretical_loss": 3.596777480578958,
      "tokens_seen": 1166737408
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004125425572158124,
      "loss": 2.7085,
      "theoretical_loss": 3.596758798661195,
      "tokens_seen": 1166802944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041252600718744094,
      "loss": 2.7518,
      "theoretical_loss": 3.5967401180865006,
      "tokens_seen": 1166868480
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004125094571590694,
      "loss": 2.6506,
      "theoretical_loss": 3.5967214388547033,
      "tokens_seen": 1166934016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041249290713069796,
      "loss": 2.7619,
      "theoretical_loss": 3.5967027609656306,
      "tokens_seen": 1166999552
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004124763571023265,
      "loss": 2.7733,
      "theoretical_loss": 3.5966840844191106,
      "tokens_seen": 1167065088
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041245980707395503,
      "loss": 2.9445,
      "theoretical_loss": 3.596665409214972,
      "tokens_seen": 1167130624
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004124432570455835,
      "loss": 2.8809,
      "theoretical_loss": 3.5966467353530422,
      "tokens_seen": 1167196160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000412426707017212,
      "loss": 2.6883,
      "theoretical_loss": 3.5966280628331493,
      "tokens_seen": 1167261696
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041241015698884054,
      "loss": 2.9483,
      "theoretical_loss": 3.5966093916551225,
      "tokens_seen": 1167327232
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000412393606960469,
      "loss": 2.7148,
      "theoretical_loss": 3.5965907218187896,
      "tokens_seen": 1167392768
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041237705693209756,
      "loss": 2.6973,
      "theoretical_loss": 3.5965720533239782,
      "tokens_seen": 1167458304
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004123605069037261,
      "loss": 2.7883,
      "theoretical_loss": 3.596553386170518,
      "tokens_seen": 1167523840
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041234395687535463,
      "loss": 2.5487,
      "theoretical_loss": 3.5965347203582354,
      "tokens_seen": 1167589376
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004123274068469831,
      "loss": 2.6686,
      "theoretical_loss": 3.59651605588696,
      "tokens_seen": 1167654912
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041231085681861165,
      "loss": 2.837,
      "theoretical_loss": 3.596497392756521,
      "tokens_seen": 1167720448
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004122943067902402,
      "loss": 2.5083,
      "theoretical_loss": 3.596478730966745,
      "tokens_seen": 1167785984
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041227775676186873,
      "loss": 2.7549,
      "theoretical_loss": 3.5964600705174616,
      "tokens_seen": 1167851520
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004122612067334972,
      "loss": 2.677,
      "theoretical_loss": 3.596441411408499,
      "tokens_seen": 1167917056
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041224465670512575,
      "loss": 2.9683,
      "theoretical_loss": 3.5964227536396858,
      "tokens_seen": 1167982592
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004122281066767543,
      "loss": 2.6948,
      "theoretical_loss": 3.5964040972108506,
      "tokens_seen": 1168048128
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004122115566483828,
      "loss": 2.8222,
      "theoretical_loss": 3.5963854421218215,
      "tokens_seen": 1168113664
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004121950066200113,
      "loss": 2.6598,
      "theoretical_loss": 3.596366788372428,
      "tokens_seen": 1168179200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041217845659163985,
      "loss": 2.9596,
      "theoretical_loss": 3.596348135962498,
      "tokens_seen": 1168244736
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004121619065632684,
      "loss": 2.6812,
      "theoretical_loss": 3.5963294848918608,
      "tokens_seen": 1168310272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004121453565348969,
      "loss": 2.4948,
      "theoretical_loss": 3.5963108351603443,
      "tokens_seen": 1168375808
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004121288065065254,
      "loss": 2.6419,
      "theoretical_loss": 3.596292186767778,
      "tokens_seen": 1168441344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041211225647815394,
      "loss": 2.7783,
      "theoretical_loss": 3.596273539713991,
      "tokens_seen": 1168506880
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004120957064497825,
      "loss": 2.8153,
      "theoretical_loss": 3.5962548939988106,
      "tokens_seen": 1168572416
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000412079156421411,
      "loss": 2.7176,
      "theoretical_loss": 3.5962362496220672,
      "tokens_seen": 1168637952
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004120626063930395,
      "loss": 2.6608,
      "theoretical_loss": 3.5962176065835885,
      "tokens_seen": 1168703488
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041204605636466804,
      "loss": 2.6423,
      "theoretical_loss": 3.5961989648832042,
      "tokens_seen": 1168769024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004120295063362966,
      "loss": 2.8577,
      "theoretical_loss": 3.596180324520743,
      "tokens_seen": 1168834560
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004120129563079251,
      "loss": 2.5854,
      "theoretical_loss": 3.5961616854960337,
      "tokens_seen": 1168900096
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004119964062795536,
      "loss": 2.6056,
      "theoretical_loss": 3.596143047808906,
      "tokens_seen": 1168965632
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041197985625118213,
      "loss": 2.7698,
      "theoretical_loss": 3.5961244114591877,
      "tokens_seen": 1169031168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041196330622281067,
      "loss": 2.614,
      "theoretical_loss": 3.596105776446709,
      "tokens_seen": 1169096704
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004119467561944392,
      "loss": 2.557,
      "theoretical_loss": 3.596087142771298,
      "tokens_seen": 1169162240
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004119302061660677,
      "loss": 2.5638,
      "theoretical_loss": 3.5960685104327847,
      "tokens_seen": 1169227776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041191365613769623,
      "loss": 2.8393,
      "theoretical_loss": 3.596049879430998,
      "tokens_seen": 1169293312
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041189710610932477,
      "loss": 2.5683,
      "theoretical_loss": 3.596031249765767,
      "tokens_seen": 1169358848
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004118805560809533,
      "loss": 2.8628,
      "theoretical_loss": 3.5960126214369206,
      "tokens_seen": 1169424384
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004118640060525818,
      "loss": 2.6356,
      "theoretical_loss": 3.5959939944442887,
      "tokens_seen": 1169489920
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004118474560242103,
      "loss": 2.6969,
      "theoretical_loss": 3.5959753687877005,
      "tokens_seen": 1169555456
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041183090599583886,
      "loss": 2.8433,
      "theoretical_loss": 3.5959567444669847,
      "tokens_seen": 1169620992
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004118143559674674,
      "loss": 2.7171,
      "theoretical_loss": 3.595938121481971,
      "tokens_seen": 1169686528
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004117978059390959,
      "loss": 2.5846,
      "theoretical_loss": 3.5959194998324886,
      "tokens_seen": 1169752064
    },
    {
      "epoch": 0.42,
      "objective/train/avg_token_score": 0.015245710499584675,
      "objective/train/avg_weight": 0.9820719361305237,
      "objective/train/docs_used": 663194,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.037665367126465,
      "objective/train/std_weight": 0.06628498435020447,
      "objective/train/theoretical_loss": 3.595900879518368,
      "objective/train/tokens_used": 1190277600,
      "theoretical_loss": 3.595900879518368,
      "tokens_seen": 1169817600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004117812559107244,
      "loss": 2.985,
      "theoretical_loss": 3.595900879518368,
      "tokens_seen": 1169817600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041176470588235296,
      "loss": 2.66,
      "theoretical_loss": 3.5958822605394367,
      "tokens_seen": 1169883136
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004117481558539815,
      "loss": 2.6538,
      "theoretical_loss": 3.5958636428955257,
      "tokens_seen": 1169948672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041173160582561,
      "loss": 2.6595,
      "theoretical_loss": 3.595845026586464,
      "tokens_seen": 1170014208
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004117150557972385,
      "loss": 2.776,
      "theoretical_loss": 3.595826411612081,
      "tokens_seen": 1170079744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041169850576886706,
      "loss": 3.0152,
      "theoretical_loss": 3.5958077979722063,
      "tokens_seen": 1170145280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004116819557404956,
      "loss": 2.9958,
      "theoretical_loss": 3.59578918566667,
      "tokens_seen": 1170210816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004116654057121241,
      "loss": 2.7915,
      "theoretical_loss": 3.595770574695301,
      "tokens_seen": 1170276352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004116488556837526,
      "loss": 2.6241,
      "theoretical_loss": 3.5957519650579295,
      "tokens_seen": 1170341888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041163230565538115,
      "loss": 2.7745,
      "theoretical_loss": 3.5957333567543848,
      "tokens_seen": 1170407424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004116157556270097,
      "loss": 2.5694,
      "theoretical_loss": 3.595714749784497,
      "tokens_seen": 1170472960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041159920559863817,
      "loss": 2.7296,
      "theoretical_loss": 3.5956961441480955,
      "tokens_seen": 1170538496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004115826555702667,
      "loss": 2.8848,
      "theoretical_loss": 3.5956775398450103,
      "tokens_seen": 1170604032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041156610554189525,
      "loss": 2.8249,
      "theoretical_loss": 3.595658936875071,
      "tokens_seen": 1170669568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004115495555135238,
      "loss": 2.8671,
      "theoretical_loss": 3.595640335238108,
      "tokens_seen": 1170735104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041153300548515227,
      "loss": 2.8322,
      "theoretical_loss": 3.5956217349339505,
      "tokens_seen": 1170800640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004115164554567808,
      "loss": 2.5579,
      "theoretical_loss": 3.5956031359624285,
      "tokens_seen": 1170866176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041149990542840934,
      "loss": 2.9272,
      "theoretical_loss": 3.595584538323372,
      "tokens_seen": 1170931712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004114833554000379,
      "loss": 2.7028,
      "theoretical_loss": 3.595565942016612,
      "tokens_seen": 1170997248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041146680537166636,
      "loss": 2.6707,
      "theoretical_loss": 3.5955473470419763,
      "tokens_seen": 1171062784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004114502553432949,
      "loss": 2.572,
      "theoretical_loss": 3.5955287533992975,
      "tokens_seen": 1171128320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041143370531492333,
      "loss": 2.4546,
      "theoretical_loss": 3.5955101610884035,
      "tokens_seen": 1171193856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041141715528655187,
      "loss": 2.5238,
      "theoretical_loss": 3.5954915701091257,
      "tokens_seen": 1171259392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004114006052581804,
      "loss": 2.8683,
      "theoretical_loss": 3.5954729804612935,
      "tokens_seen": 1171324928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041138405522980894,
      "loss": 2.6997,
      "theoretical_loss": 3.5954543921447377,
      "tokens_seen": 1171390464
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004113675052014374,
      "loss": 2.7264,
      "theoretical_loss": 3.595435805159288,
      "tokens_seen": 1171456000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041135095517306596,
      "loss": 2.7008,
      "theoretical_loss": 3.5954172195047747,
      "tokens_seen": 1171521536
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004113344051446945,
      "loss": 2.8011,
      "theoretical_loss": 3.5953986351810285,
      "tokens_seen": 1171587072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041131785511632304,
      "loss": 2.6033,
      "theoretical_loss": 3.5953800521878794,
      "tokens_seen": 1171652608
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004113013050879515,
      "loss": 2.8084,
      "theoretical_loss": 3.595361470525157,
      "tokens_seen": 1171718144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041128475505958006,
      "loss": 2.8012,
      "theoretical_loss": 3.5953428901926925,
      "tokens_seen": 1171783680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004112682050312086,
      "loss": 2.8952,
      "theoretical_loss": 3.595324311190316,
      "tokens_seen": 1171849216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041125165500283714,
      "loss": 2.5669,
      "theoretical_loss": 3.595305733517858,
      "tokens_seen": 1171914752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004112351049744656,
      "loss": 2.7971,
      "theoretical_loss": 3.5952871571751492,
      "tokens_seen": 1171980288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041121855494609416,
      "loss": 2.6698,
      "theoretical_loss": 3.5952685821620194,
      "tokens_seen": 1172045824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004112020049177227,
      "loss": 2.6528,
      "theoretical_loss": 3.5952500084782995,
      "tokens_seen": 1172111360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041118545488935123,
      "loss": 2.6494,
      "theoretical_loss": 3.59523143612382,
      "tokens_seen": 1172176896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004111689048609797,
      "loss": 3.0167,
      "theoretical_loss": 3.5952128650984116,
      "tokens_seen": 1172242432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041115235483260825,
      "loss": 2.6985,
      "theoretical_loss": 3.595194295401905,
      "tokens_seen": 1172307968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004111358048042368,
      "loss": 2.5702,
      "theoretical_loss": 3.59517572703413,
      "tokens_seen": 1172373504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041111925477586533,
      "loss": 2.7383,
      "theoretical_loss": 3.595157159994918,
      "tokens_seen": 1172439040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004111027047474938,
      "loss": 2.6736,
      "theoretical_loss": 3.5951385942840997,
      "tokens_seen": 1172504576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041108615471912235,
      "loss": 2.8209,
      "theoretical_loss": 3.595120029901506,
      "tokens_seen": 1172570112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004110696046907509,
      "loss": 2.5966,
      "theoretical_loss": 3.595101466846966,
      "tokens_seen": 1172635648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004110530546623794,
      "loss": 2.8517,
      "theoretical_loss": 3.595082905120313,
      "tokens_seen": 1172701184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004110365046340079,
      "loss": 2.7516,
      "theoretical_loss": 3.595064344721376,
      "tokens_seen": 1172766720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041101995460563644,
      "loss": 2.7037,
      "theoretical_loss": 3.5950457856499867,
      "tokens_seen": 1172832256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000411003404577265,
      "loss": 2.7555,
      "theoretical_loss": 3.595027227905976,
      "tokens_seen": 1172897792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004109868545488935,
      "loss": 2.7521,
      "theoretical_loss": 3.5950086714891736,
      "tokens_seen": 1172963328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041097030452052206,
      "loss": 2.6844,
      "theoretical_loss": 3.594990116399412,
      "tokens_seen": 1173028864
    },
    {
      "epoch": 0.42,
      "objective/train/avg_token_score": 0.015103803016245365,
      "objective/train/avg_weight": 0.9822337031364441,
      "objective/train/docs_used": 665258,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.019533157348633,
      "objective/train/std_weight": 0.06946699321269989,
      "objective/train/theoretical_loss": 3.594971562636521,
      "objective/train/tokens_used": 1193554400,
      "theoretical_loss": 3.594971562636521,
      "tokens_seen": 1173094400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041095375449215054,
      "loss": 2.9087,
      "theoretical_loss": 3.594971562636521,
      "tokens_seen": 1173094400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004109372044637791,
      "loss": 2.7099,
      "theoretical_loss": 3.5949530102003324,
      "tokens_seen": 1173159936
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004109206544354076,
      "loss": 2.3604,
      "theoretical_loss": 3.594934459090677,
      "tokens_seen": 1173225472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041090410440703615,
      "loss": 2.934,
      "theoretical_loss": 3.5949159093073857,
      "tokens_seen": 1173291008
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041088755437866464,
      "loss": 2.9203,
      "theoretical_loss": 3.5948973608502897,
      "tokens_seen": 1173356544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004108710043502932,
      "loss": 2.7989,
      "theoretical_loss": 3.5948788137192196,
      "tokens_seen": 1173422080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004108544543219217,
      "loss": 2.6193,
      "theoretical_loss": 3.594860267914008,
      "tokens_seen": 1173487616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041083790429355025,
      "loss": 2.5464,
      "theoretical_loss": 3.5948417234344845,
      "tokens_seen": 1173553152
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041082135426517873,
      "loss": 2.5997,
      "theoretical_loss": 3.594823180280481,
      "tokens_seen": 1173618688
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041080480423680727,
      "loss": 2.6678,
      "theoretical_loss": 3.5948046384518286,
      "tokens_seen": 1173684224
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004107882542084358,
      "loss": 2.594,
      "theoretical_loss": 3.594786097948359,
      "tokens_seen": 1173749760
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041077170418006434,
      "loss": 2.6189,
      "theoretical_loss": 3.5947675587699024,
      "tokens_seen": 1173815296
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041075515415169283,
      "loss": 2.7247,
      "theoretical_loss": 3.594749020916292,
      "tokens_seen": 1173880832
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041073860412332137,
      "loss": 2.7841,
      "theoretical_loss": 3.594730484387357,
      "tokens_seen": 1173946368
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004107220540949499,
      "loss": 2.6426,
      "theoretical_loss": 3.5947119491829307,
      "tokens_seen": 1174011904
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041070550406657844,
      "loss": 2.6975,
      "theoretical_loss": 3.594693415302843,
      "tokens_seen": 1174077440
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004106889540382069,
      "loss": 2.7529,
      "theoretical_loss": 3.5946748827469266,
      "tokens_seen": 1174142976
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041067240400983546,
      "loss": 2.6219,
      "theoretical_loss": 3.5946563515150123,
      "tokens_seen": 1174208512
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000410655853981464,
      "loss": 2.8196,
      "theoretical_loss": 3.5946378216069315,
      "tokens_seen": 1174274048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041063930395309254,
      "loss": 2.5406,
      "theoretical_loss": 3.5946192930225163,
      "tokens_seen": 1174339584
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000410622753924721,
      "loss": 2.7432,
      "theoretical_loss": 3.594600765761598,
      "tokens_seen": 1174405120
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041060620389634956,
      "loss": 2.455,
      "theoretical_loss": 3.5945822398240077,
      "tokens_seen": 1174470656
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004105896538679781,
      "loss": 2.9067,
      "theoretical_loss": 3.5945637152095786,
      "tokens_seen": 1174536192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041057310383960663,
      "loss": 2.6438,
      "theoretical_loss": 3.59454519191814,
      "tokens_seen": 1174601728
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004105565538112351,
      "loss": 2.6625,
      "theoretical_loss": 3.594526669949526,
      "tokens_seen": 1174667264
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041054000378286365,
      "loss": 2.6919,
      "theoretical_loss": 3.594508149303567,
      "tokens_seen": 1174732800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004105234537544922,
      "loss": 2.9043,
      "theoretical_loss": 3.594489629980095,
      "tokens_seen": 1174798336
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041050690372612073,
      "loss": 2.7874,
      "theoretical_loss": 3.594471111978942,
      "tokens_seen": 1174863872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004104903536977492,
      "loss": 2.5027,
      "theoretical_loss": 3.5944525952999395,
      "tokens_seen": 1174929408
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041047380366937775,
      "loss": 2.6983,
      "theoretical_loss": 3.5944340799429195,
      "tokens_seen": 1174994944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004104572536410062,
      "loss": 2.6529,
      "theoretical_loss": 3.594415565907714,
      "tokens_seen": 1175060480
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004104407036126347,
      "loss": 2.6146,
      "theoretical_loss": 3.5943970531941556,
      "tokens_seen": 1175126016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041042415358426325,
      "loss": 2.7514,
      "theoretical_loss": 3.5943785418020746,
      "tokens_seen": 1175191552
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004104076035558918,
      "loss": 2.7456,
      "theoretical_loss": 3.594360031731304,
      "tokens_seen": 1175257088
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004103910535275203,
      "loss": 2.755,
      "theoretical_loss": 3.5943415229816758,
      "tokens_seen": 1175322624
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004103745034991488,
      "loss": 2.8392,
      "theoretical_loss": 3.594323015553022,
      "tokens_seen": 1175388160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041035795347077735,
      "loss": 2.5371,
      "theoretical_loss": 3.5943045094451747,
      "tokens_seen": 1175453696
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004103414034424059,
      "loss": 2.8411,
      "theoretical_loss": 3.594286004657966,
      "tokens_seen": 1175519232
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041032485341403437,
      "loss": 2.676,
      "theoretical_loss": 3.5942675011912275,
      "tokens_seen": 1175584768
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004103083033856629,
      "loss": 2.4827,
      "theoretical_loss": 3.594248999044792,
      "tokens_seen": 1175650304
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041029175335729145,
      "loss": 2.801,
      "theoretical_loss": 3.594230498218492,
      "tokens_seen": 1175715840
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041027520332892,
      "loss": 2.5725,
      "theoretical_loss": 3.594211998712159,
      "tokens_seen": 1175781376
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041025865330054847,
      "loss": 2.9999,
      "theoretical_loss": 3.5941935005256256,
      "tokens_seen": 1175846912
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000410242103272177,
      "loss": 2.529,
      "theoretical_loss": 3.594175003658724,
      "tokens_seen": 1175912448
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041022555324380554,
      "loss": 2.8101,
      "theoretical_loss": 3.5941565081112863,
      "tokens_seen": 1175977984
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004102090032154341,
      "loss": 2.6185,
      "theoretical_loss": 3.5941380138831454,
      "tokens_seen": 1176043520
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041019245318706256,
      "loss": 2.7572,
      "theoretical_loss": 3.5941195209741332,
      "tokens_seen": 1176109056
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004101759031586911,
      "loss": 2.6718,
      "theoretical_loss": 3.5941010293840825,
      "tokens_seen": 1176174592
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041015935313031964,
      "loss": 2.7833,
      "theoretical_loss": 3.5940825391128257,
      "tokens_seen": 1176240128
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004101428031019482,
      "loss": 2.8,
      "theoretical_loss": 3.5940640501601946,
      "tokens_seen": 1176305664
    },
    {
      "epoch": 0.42,
      "objective/train/avg_token_score": 0.026933737099170685,
      "objective/train/avg_weight": 0.968317449092865,
      "objective/train/docs_used": 667248,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.968960762023926,
      "objective/train/std_weight": 0.11369124799966812,
      "objective/train/theoretical_loss": 3.5940455625260226,
      "objective/train/tokens_used": 1196831200,
      "theoretical_loss": 3.5940455625260226,
      "tokens_seen": 1176371200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041012625307357666,
      "loss": 2.5444,
      "theoretical_loss": 3.5940455625260226,
      "tokens_seen": 1176371200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004101097030452052,
      "loss": 2.6788,
      "theoretical_loss": 3.594027076210142,
      "tokens_seen": 1176436736
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041009315301683373,
      "loss": 2.6961,
      "theoretical_loss": 3.5940085912123845,
      "tokens_seen": 1176502272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041007660298846227,
      "loss": 2.8218,
      "theoretical_loss": 3.593990107532584,
      "tokens_seen": 1176567808
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041006005296009075,
      "loss": 2.4477,
      "theoretical_loss": 3.593971625170572,
      "tokens_seen": 1176633344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004100435029317193,
      "loss": 2.7272,
      "theoretical_loss": 3.593953144126182,
      "tokens_seen": 1176698880
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041002695290334783,
      "loss": 2.8014,
      "theoretical_loss": 3.5939346643992467,
      "tokens_seen": 1176764416
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00041001040287497637,
      "loss": 2.7669,
      "theoretical_loss": 3.5939161859895985,
      "tokens_seen": 1176829952
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040999385284660485,
      "loss": 2.7782,
      "theoretical_loss": 3.59389770889707,
      "tokens_seen": 1176895488
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004099773028182334,
      "loss": 2.6847,
      "theoretical_loss": 3.593879233121494,
      "tokens_seen": 1176961024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004099607527898619,
      "loss": 3.1044,
      "theoretical_loss": 3.5938607586627036,
      "tokens_seen": 1177026560
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040994420276149046,
      "loss": 2.5944,
      "theoretical_loss": 3.5938422855205316,
      "tokens_seen": 1177092096
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040992765273311895,
      "loss": 2.6776,
      "theoretical_loss": 3.5938238136948106,
      "tokens_seen": 1177157632
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004099111027047475,
      "loss": 2.6755,
      "theoretical_loss": 3.5938053431853736,
      "tokens_seen": 1177223168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000409894552676376,
      "loss": 2.5372,
      "theoretical_loss": 3.5937868739920535,
      "tokens_seen": 1177288704
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040987800264800456,
      "loss": 2.5178,
      "theoretical_loss": 3.593768406114684,
      "tokens_seen": 1177354240
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040986145261963304,
      "loss": 2.8314,
      "theoretical_loss": 3.593749939553097,
      "tokens_seen": 1177419776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004098449025912616,
      "loss": 2.637,
      "theoretical_loss": 3.593731474307126,
      "tokens_seen": 1177485312
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004098283525628901,
      "loss": 2.7503,
      "theoretical_loss": 3.5937130103766046,
      "tokens_seen": 1177550848
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040981180253451865,
      "loss": 2.7799,
      "theoretical_loss": 3.5936945477613644,
      "tokens_seen": 1177616384
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040979525250614714,
      "loss": 2.7558,
      "theoretical_loss": 3.59367608646124,
      "tokens_seen": 1177681920
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004097787024777757,
      "loss": 2.5223,
      "theoretical_loss": 3.593657626476064,
      "tokens_seen": 1177747456
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004097621524494042,
      "loss": 2.998,
      "theoretical_loss": 3.5936391678056694,
      "tokens_seen": 1177812992
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040974560242103275,
      "loss": 2.706,
      "theoretical_loss": 3.5936207104498896,
      "tokens_seen": 1177878528
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040972905239266123,
      "loss": 2.6586,
      "theoretical_loss": 3.5936022544085575,
      "tokens_seen": 1177944064
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040971250236428977,
      "loss": 2.7939,
      "theoretical_loss": 3.593583799681507,
      "tokens_seen": 1178009600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004096959523359183,
      "loss": 2.6522,
      "theoretical_loss": 3.593565346268571,
      "tokens_seen": 1178075136
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040967940230754685,
      "loss": 2.6522,
      "theoretical_loss": 3.593546894169583,
      "tokens_seen": 1178140672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040966285227917533,
      "loss": 2.7595,
      "theoretical_loss": 3.5935284433843764,
      "tokens_seen": 1178206208
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040964630225080387,
      "loss": 2.5834,
      "theoretical_loss": 3.593509993912784,
      "tokens_seen": 1178271744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004096297522224324,
      "loss": 2.6665,
      "theoretical_loss": 3.593491545754641,
      "tokens_seen": 1178337280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040961320219406094,
      "loss": 2.6663,
      "theoretical_loss": 3.5934730989097776,
      "tokens_seen": 1178402816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004095966521656894,
      "loss": 2.7548,
      "theoretical_loss": 3.5934546533780303,
      "tokens_seen": 1178468352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040958010213731796,
      "loss": 2.4831,
      "theoretical_loss": 3.593436209159232,
      "tokens_seen": 1178533888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004095635521089465,
      "loss": 2.872,
      "theoretical_loss": 3.5934177662532147,
      "tokens_seen": 1178599424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040954700208057504,
      "loss": 2.6129,
      "theoretical_loss": 3.5933993246598135,
      "tokens_seen": 1178664960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004095304520522035,
      "loss": 2.7582,
      "theoretical_loss": 3.5933808843788615,
      "tokens_seen": 1178730496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040951390202383206,
      "loss": 2.7734,
      "theoretical_loss": 3.5933624454101922,
      "tokens_seen": 1178796032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004094973519954606,
      "loss": 2.7472,
      "theoretical_loss": 3.5933440077536396,
      "tokens_seen": 1178861568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040948080196708913,
      "loss": 2.7909,
      "theoretical_loss": 3.5933255714090375,
      "tokens_seen": 1178927104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040946425193871756,
      "loss": 2.8663,
      "theoretical_loss": 3.593307136376219,
      "tokens_seen": 1178992640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004094477019103461,
      "loss": 2.6572,
      "theoretical_loss": 3.5932887026550184,
      "tokens_seen": 1179058176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040943115188197464,
      "loss": 2.6399,
      "theoretical_loss": 3.5932702702452692,
      "tokens_seen": 1179123712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004094146018536031,
      "loss": 2.8172,
      "theoretical_loss": 3.5932518391468053,
      "tokens_seen": 1179189248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040939805182523166,
      "loss": 2.7888,
      "theoretical_loss": 3.5932334093594607,
      "tokens_seen": 1179254784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004093815017968602,
      "loss": 2.5091,
      "theoretical_loss": 3.5932149808830687,
      "tokens_seen": 1179320320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040936495176848873,
      "loss": 2.6509,
      "theoretical_loss": 3.593196553717464,
      "tokens_seen": 1179385856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004093484017401172,
      "loss": 2.7661,
      "theoretical_loss": 3.5931781278624806,
      "tokens_seen": 1179451392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040933185171174576,
      "loss": 2.9592,
      "theoretical_loss": 3.5931597033179514,
      "tokens_seen": 1179516928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004093153016833743,
      "loss": 2.593,
      "theoretical_loss": 3.593141280083711,
      "tokens_seen": 1179582464
    },
    {
      "epoch": 0.42,
      "objective/train/avg_token_score": 0.012930875644087791,
      "objective/train/avg_weight": 0.9847822189331055,
      "objective/train/docs_used": 669029,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.680131435394287,
      "objective/train/std_weight": 0.07169842720031738,
      "objective/train/theoretical_loss": 3.5931228581595938,
      "objective/train/tokens_used": 1200108000,
      "theoretical_loss": 3.5931228581595938,
      "tokens_seen": 1179648000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040929875165500283,
      "loss": 2.6193,
      "theoretical_loss": 3.5931228581595938,
      "tokens_seen": 1179648000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004092822016266313,
      "loss": 2.7686,
      "theoretical_loss": 3.593104437545433,
      "tokens_seen": 1179713536
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040926565159825985,
      "loss": 2.8169,
      "theoretical_loss": 3.5930860182410638,
      "tokens_seen": 1179779072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004092491015698884,
      "loss": 2.7707,
      "theoretical_loss": 3.5930676002463198,
      "tokens_seen": 1179844608
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004092325515415169,
      "loss": 2.3769,
      "theoretical_loss": 3.5930491835610345,
      "tokens_seen": 1179910144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004092160015131454,
      "loss": 3.0176,
      "theoretical_loss": 3.593030768185043,
      "tokens_seen": 1179975680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040919945148477395,
      "loss": 2.786,
      "theoretical_loss": 3.593012354118179,
      "tokens_seen": 1180041216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004091829014564025,
      "loss": 2.8733,
      "theoretical_loss": 3.5929939413602767,
      "tokens_seen": 1180106752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000409166351428031,
      "loss": 2.7839,
      "theoretical_loss": 3.592975529911171,
      "tokens_seen": 1180172288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004091498013996595,
      "loss": 2.6991,
      "theoretical_loss": 3.5929571197706958,
      "tokens_seen": 1180237824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040913325137128804,
      "loss": 2.7857,
      "theoretical_loss": 3.592938710938685,
      "tokens_seen": 1180303360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004091167013429166,
      "loss": 2.7989,
      "theoretical_loss": 3.5929203034149735,
      "tokens_seen": 1180368896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004091001513145451,
      "loss": 2.6385,
      "theoretical_loss": 3.5929018971993956,
      "tokens_seen": 1180434432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004090836012861736,
      "loss": 2.9289,
      "theoretical_loss": 3.5928834922917856,
      "tokens_seen": 1180499968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040906705125780214,
      "loss": 2.9556,
      "theoretical_loss": 3.592865088691978,
      "tokens_seen": 1180565504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004090505012294307,
      "loss": 2.7436,
      "theoretical_loss": 3.592846686399807,
      "tokens_seen": 1180631040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004090339512010592,
      "loss": 2.8984,
      "theoretical_loss": 3.592828285415108,
      "tokens_seen": 1180696576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004090174011726877,
      "loss": 2.8403,
      "theoretical_loss": 3.5928098857377146,
      "tokens_seen": 1180762112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040900085114431624,
      "loss": 2.8706,
      "theoretical_loss": 3.592791487367462,
      "tokens_seen": 1180827648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004089843011159448,
      "loss": 2.7454,
      "theoretical_loss": 3.592773090304184,
      "tokens_seen": 1180893184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004089677510875733,
      "loss": 3.0594,
      "theoretical_loss": 3.592754694547716,
      "tokens_seen": 1180958720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004089512010592018,
      "loss": 2.9071,
      "theoretical_loss": 3.5927363000978927,
      "tokens_seen": 1181024256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040893465103083033,
      "loss": 2.758,
      "theoretical_loss": 3.592717906954548,
      "tokens_seen": 1181089792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040891810100245887,
      "loss": 2.5859,
      "theoretical_loss": 3.592699515117517,
      "tokens_seen": 1181155328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004089015509740874,
      "loss": 2.6443,
      "theoretical_loss": 3.592681124586635,
      "tokens_seen": 1181220864
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004088850009457159,
      "loss": 2.3794,
      "theoretical_loss": 3.592662735361736,
      "tokens_seen": 1181286400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040886845091734443,
      "loss": 2.3421,
      "theoretical_loss": 3.5926443474426555,
      "tokens_seen": 1181351936
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040885190088897296,
      "loss": 2.7947,
      "theoretical_loss": 3.5926259608292277,
      "tokens_seen": 1181417472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004088353508606015,
      "loss": 2.6404,
      "theoretical_loss": 3.592607575521288,
      "tokens_seen": 1181483008
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040881880083223,
      "loss": 2.7285,
      "theoretical_loss": 3.592589191518671,
      "tokens_seen": 1181548544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004088022508038585,
      "loss": 2.5847,
      "theoretical_loss": 3.5925708088212116,
      "tokens_seen": 1181614080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040878570077548706,
      "loss": 2.685,
      "theoretical_loss": 3.592552427428745,
      "tokens_seen": 1181679616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004087691507471156,
      "loss": 2.5705,
      "theoretical_loss": 3.592534047341106,
      "tokens_seen": 1181745152
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004087526007187441,
      "loss": 2.7027,
      "theoretical_loss": 3.5925156685581294,
      "tokens_seen": 1181810688
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004087360506903726,
      "loss": 2.5882,
      "theoretical_loss": 3.592497291079651,
      "tokens_seen": 1181876224
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040871950066200116,
      "loss": 2.8297,
      "theoretical_loss": 3.592478914905505,
      "tokens_seen": 1181941760
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004087029506336297,
      "loss": 2.7533,
      "theoretical_loss": 3.592460540035527,
      "tokens_seen": 1182007296
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004086864006052582,
      "loss": 2.628,
      "theoretical_loss": 3.592442166469552,
      "tokens_seen": 1182072832
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004086698505768867,
      "loss": 2.7497,
      "theoretical_loss": 3.5924237942074155,
      "tokens_seen": 1182138368
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040865330054851525,
      "loss": 2.6051,
      "theoretical_loss": 3.5924054232489517,
      "tokens_seen": 1182203904
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004086367505201438,
      "loss": 2.6753,
      "theoretical_loss": 3.5923870535939972,
      "tokens_seen": 1182269440
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004086202004917723,
      "loss": 2.8482,
      "theoretical_loss": 3.592368685242387,
      "tokens_seen": 1182334976
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004086036504634008,
      "loss": 2.6437,
      "theoretical_loss": 3.5923503181939553,
      "tokens_seen": 1182400512
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040858710043502935,
      "loss": 2.7627,
      "theoretical_loss": 3.5923319524485384,
      "tokens_seen": 1182466048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004085705504066579,
      "loss": 2.6428,
      "theoretical_loss": 3.592313588005971,
      "tokens_seen": 1182531584
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040855400037828637,
      "loss": 2.5966,
      "theoretical_loss": 3.5922952248660893,
      "tokens_seen": 1182597120
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004085374503499149,
      "loss": 2.6996,
      "theoretical_loss": 3.592276863028728,
      "tokens_seen": 1182662656
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040852090032154344,
      "loss": 2.6718,
      "theoretical_loss": 3.5922585024937232,
      "tokens_seen": 1182728192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000408504350293172,
      "loss": 2.5699,
      "theoretical_loss": 3.5922401432609092,
      "tokens_seen": 1182793728
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004084878002648004,
      "loss": 2.7493,
      "theoretical_loss": 3.592221785330123,
      "tokens_seen": 1182859264
    },
    {
      "epoch": 0.42,
      "objective/train/avg_token_score": 0.006908494979143143,
      "objective/train/avg_weight": 0.9918748140335083,
      "objective/train/docs_used": 670515,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6606884002685547,
      "objective/train/std_weight": 0.040807824581861496,
      "objective/train/theoretical_loss": 3.5922034287011995,
      "objective/train/tokens_used": 1203384800,
      "theoretical_loss": 3.5922034287011995,
      "tokens_seen": 1182924800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040847125023642895,
      "loss": 2.6663,
      "theoretical_loss": 3.5922034287011995,
      "tokens_seen": 1182924800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004084547002080575,
      "loss": 2.7833,
      "theoretical_loss": 3.5921850733739733,
      "tokens_seen": 1182990336
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040843815017968597,
      "loss": 2.8577,
      "theoretical_loss": 3.592166719348282,
      "tokens_seen": 1183055872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004084216001513145,
      "loss": 2.7419,
      "theoretical_loss": 3.5921483666239595,
      "tokens_seen": 1183121408
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040840505012294305,
      "loss": 2.5687,
      "theoretical_loss": 3.592130015200842,
      "tokens_seen": 1183186944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004083885000945716,
      "loss": 2.4238,
      "theoretical_loss": 3.592111665078765,
      "tokens_seen": 1183252480
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040837195006620007,
      "loss": 2.954,
      "theoretical_loss": 3.592093316257565,
      "tokens_seen": 1183318016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004083554000378286,
      "loss": 2.5236,
      "theoretical_loss": 3.5920749687370765,
      "tokens_seen": 1183383552
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040833885000945714,
      "loss": 2.6918,
      "theoretical_loss": 3.5920566225171364,
      "tokens_seen": 1183449088
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004083222999810857,
      "loss": 2.7034,
      "theoretical_loss": 3.5920382775975797,
      "tokens_seen": 1183514624
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040830574995271416,
      "loss": 2.5499,
      "theoretical_loss": 3.592019933978243,
      "tokens_seen": 1183580160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004082891999243427,
      "loss": 2.7477,
      "theoretical_loss": 3.592001591658961,
      "tokens_seen": 1183645696
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040827264989597124,
      "loss": 2.7999,
      "theoretical_loss": 3.5919832506395712,
      "tokens_seen": 1183711232
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004082560998675998,
      "loss": 2.6488,
      "theoretical_loss": 3.591964910919908,
      "tokens_seen": 1183776768
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040823954983922826,
      "loss": 2.5483,
      "theoretical_loss": 3.5919465724998076,
      "tokens_seen": 1183842304
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004082229998108568,
      "loss": 2.6936,
      "theoretical_loss": 3.591928235379107,
      "tokens_seen": 1183907840
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040820644978248533,
      "loss": 2.7073,
      "theoretical_loss": 3.5919098995576415,
      "tokens_seen": 1183973376
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040818989975411387,
      "loss": 2.6739,
      "theoretical_loss": 3.5918915650352474,
      "tokens_seen": 1184038912
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040817334972574235,
      "loss": 2.7318,
      "theoretical_loss": 3.59187323181176,
      "tokens_seen": 1184104448
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004081567996973709,
      "loss": 2.6447,
      "theoretical_loss": 3.591854899887016,
      "tokens_seen": 1184169984
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040814024966899943,
      "loss": 2.7108,
      "theoretical_loss": 3.5918365692608516,
      "tokens_seen": 1184235520
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040812369964062797,
      "loss": 2.7152,
      "theoretical_loss": 3.5918182399331036,
      "tokens_seen": 1184301056
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040810714961225645,
      "loss": 2.8772,
      "theoretical_loss": 3.5917999119036064,
      "tokens_seen": 1184366592
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000408090599583885,
      "loss": 2.6516,
      "theoretical_loss": 3.5917815851721975,
      "tokens_seen": 1184432128
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004080740495555135,
      "loss": 2.7198,
      "theoretical_loss": 3.5917632597387126,
      "tokens_seen": 1184497664
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040805749952714206,
      "loss": 2.8128,
      "theoretical_loss": 3.591744935602989,
      "tokens_seen": 1184563200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040804094949877055,
      "loss": 2.52,
      "theoretical_loss": 3.5917266127648615,
      "tokens_seen": 1184628736
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004080243994703991,
      "loss": 2.3821,
      "theoretical_loss": 3.591708291224168,
      "tokens_seen": 1184694272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004080078494420276,
      "loss": 2.9977,
      "theoretical_loss": 3.5916899709807435,
      "tokens_seen": 1184759808
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040799129941365616,
      "loss": 2.547,
      "theoretical_loss": 3.5916716520344254,
      "tokens_seen": 1184825344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040797474938528464,
      "loss": 2.5479,
      "theoretical_loss": 3.591653334385049,
      "tokens_seen": 1184890880
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004079581993569132,
      "loss": 2.573,
      "theoretical_loss": 3.591635018032452,
      "tokens_seen": 1184956416
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004079416493285417,
      "loss": 2.5556,
      "theoretical_loss": 3.59161670297647,
      "tokens_seen": 1185021952
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040792509930017025,
      "loss": 2.5717,
      "theoretical_loss": 3.59159838921694,
      "tokens_seen": 1185087488
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040790854927179874,
      "loss": 2.5421,
      "theoretical_loss": 3.591580076753698,
      "tokens_seen": 1185153024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004078919992434273,
      "loss": 2.4247,
      "theoretical_loss": 3.5915617655865812,
      "tokens_seen": 1185218560
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004078754492150558,
      "loss": 2.6322,
      "theoretical_loss": 3.591543455715426,
      "tokens_seen": 1185284096
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040785889918668435,
      "loss": 2.6563,
      "theoretical_loss": 3.591525147140069,
      "tokens_seen": 1185349632
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040784234915831283,
      "loss": 2.7744,
      "theoretical_loss": 3.591506839860347,
      "tokens_seen": 1185415168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040782579912994137,
      "loss": 2.6664,
      "theoretical_loss": 3.5914885338760962,
      "tokens_seen": 1185480704
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004078092491015699,
      "loss": 2.8654,
      "theoretical_loss": 3.591470229187154,
      "tokens_seen": 1185546240
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040779269907319845,
      "loss": 2.5109,
      "theoretical_loss": 3.5914519257933564,
      "tokens_seen": 1185611776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040777614904482693,
      "loss": 2.6771,
      "theoretical_loss": 3.5914336236945408,
      "tokens_seen": 1185677312
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040775959901645547,
      "loss": 2.7468,
      "theoretical_loss": 3.5914153228905437,
      "tokens_seen": 1185742848
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000407743048988084,
      "loss": 2.854,
      "theoretical_loss": 3.5913970233812016,
      "tokens_seen": 1185808384
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040772649895971254,
      "loss": 2.8135,
      "theoretical_loss": 3.5913787251663525,
      "tokens_seen": 1185873920
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000407709948931341,
      "loss": 2.8882,
      "theoretical_loss": 3.5913604282458325,
      "tokens_seen": 1185939456
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040769339890296956,
      "loss": 2.7244,
      "theoretical_loss": 3.591342132619478,
      "tokens_seen": 1186004992
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004076768488745981,
      "loss": 2.8619,
      "theoretical_loss": 3.5913238382871273,
      "tokens_seen": 1186070528
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040766029884622664,
      "loss": 2.5021,
      "theoretical_loss": 3.5913055452486162,
      "tokens_seen": 1186136064
    },
    {
      "epoch": 0.42,
      "objective/train/avg_token_score": 0.01173190027475357,
      "objective/train/avg_weight": 0.986225426197052,
      "objective/train/docs_used": 672504,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.112558364868164,
      "objective/train/std_weight": 0.07041691988706589,
      "objective/train/theoretical_loss": 3.5912872535037828,
      "objective/train/tokens_used": 1206661600,
      "theoretical_loss": 3.5912872535037828,
      "tokens_seen": 1186201600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004076437488178551,
      "loss": 2.5901,
      "theoretical_loss": 3.5912872535037828,
      "tokens_seen": 1186201600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040762719878948366,
      "loss": 2.9032,
      "theoretical_loss": 3.591268963052463,
      "tokens_seen": 1186267136
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004076106487611122,
      "loss": 2.8754,
      "theoretical_loss": 3.591250673894494,
      "tokens_seen": 1186332672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040759409873274073,
      "loss": 2.9098,
      "theoretical_loss": 3.5912323860297137,
      "tokens_seen": 1186398208
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004075775487043692,
      "loss": 2.7738,
      "theoretical_loss": 3.591214099457959,
      "tokens_seen": 1186463744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040756099867599776,
      "loss": 2.7733,
      "theoretical_loss": 3.5911958141790667,
      "tokens_seen": 1186529280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004075444486476263,
      "loss": 2.6101,
      "theoretical_loss": 3.5911775301928746,
      "tokens_seen": 1186594816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040752789861925483,
      "loss": 2.7654,
      "theoretical_loss": 3.5911592474992187,
      "tokens_seen": 1186660352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040751134859088326,
      "loss": 2.8566,
      "theoretical_loss": 3.591140966097938,
      "tokens_seen": 1186725888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004074947985625118,
      "loss": 2.7533,
      "theoretical_loss": 3.591122685988868,
      "tokens_seen": 1186791424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004074782485341403,
      "loss": 2.8329,
      "theoretical_loss": 3.5911044071718474,
      "tokens_seen": 1186856960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004074616985057688,
      "loss": 2.673,
      "theoretical_loss": 3.5910861296467127,
      "tokens_seen": 1186922496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040744514847739736,
      "loss": 2.6485,
      "theoretical_loss": 3.5910678534133016,
      "tokens_seen": 1186988032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004074285984490259,
      "loss": 2.7341,
      "theoretical_loss": 3.591049578471452,
      "tokens_seen": 1187053568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004074120484206544,
      "loss": 2.4927,
      "theoretical_loss": 3.591031304821,
      "tokens_seen": 1187119104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004073954983922829,
      "loss": 2.5669,
      "theoretical_loss": 3.5910130324617846,
      "tokens_seen": 1187184640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040737894836391145,
      "loss": 2.9035,
      "theoretical_loss": 3.5909947613936417,
      "tokens_seen": 1187250176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040736239833554,
      "loss": 2.7406,
      "theoretical_loss": 3.59097649161641,
      "tokens_seen": 1187315712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040734584830716847,
      "loss": 2.7332,
      "theoretical_loss": 3.5909582231299266,
      "tokens_seen": 1187381248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000407329298278797,
      "loss": 2.9405,
      "theoretical_loss": 3.5909399559340294,
      "tokens_seen": 1187446784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040731274825042555,
      "loss": 2.8866,
      "theoretical_loss": 3.590921690028556,
      "tokens_seen": 1187512320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004072961982220541,
      "loss": 2.805,
      "theoretical_loss": 3.5909034254133436,
      "tokens_seen": 1187577856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040727964819368257,
      "loss": 2.8282,
      "theoretical_loss": 3.59088516208823,
      "tokens_seen": 1187643392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004072630981653111,
      "loss": 2.6583,
      "theoretical_loss": 3.590866900053053,
      "tokens_seen": 1187708928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040724654813693964,
      "loss": 2.8316,
      "theoretical_loss": 3.5908486393076506,
      "tokens_seen": 1187774464
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004072299981085682,
      "loss": 2.7906,
      "theoretical_loss": 3.5908303798518597,
      "tokens_seen": 1187840000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040721344808019666,
      "loss": 2.8547,
      "theoretical_loss": 3.590812121685519,
      "tokens_seen": 1187905536
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004071968980518252,
      "loss": 2.8222,
      "theoretical_loss": 3.590793864808466,
      "tokens_seen": 1187971072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040718034802345374,
      "loss": 2.8813,
      "theoretical_loss": 3.590775609220538,
      "tokens_seen": 1188036608
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004071637979950823,
      "loss": 2.71,
      "theoretical_loss": 3.590757354921574,
      "tokens_seen": 1188102144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040714724796671076,
      "loss": 2.9867,
      "theoretical_loss": 3.590739101911411,
      "tokens_seen": 1188167680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004071306979383393,
      "loss": 2.6862,
      "theoretical_loss": 3.5907208501898866,
      "tokens_seen": 1188233216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040711414790996784,
      "loss": 2.9832,
      "theoretical_loss": 3.59070259975684,
      "tokens_seen": 1188298752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040709759788159637,
      "loss": 2.7431,
      "theoretical_loss": 3.5906843506121087,
      "tokens_seen": 1188364288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040708104785322486,
      "loss": 2.7071,
      "theoretical_loss": 3.59066610275553,
      "tokens_seen": 1188429824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004070644978248534,
      "loss": 2.9832,
      "theoretical_loss": 3.5906478561869424,
      "tokens_seen": 1188495360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040704794779648193,
      "loss": 2.6669,
      "theoretical_loss": 3.590629610906184,
      "tokens_seen": 1188560896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040703139776811047,
      "loss": 2.8643,
      "theoretical_loss": 3.5906113669130937,
      "tokens_seen": 1188626432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040701484773973895,
      "loss": 2.8474,
      "theoretical_loss": 3.5905931242075084,
      "tokens_seen": 1188691968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004069982977113675,
      "loss": 2.6262,
      "theoretical_loss": 3.5905748827892667,
      "tokens_seen": 1188757504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000406981747682996,
      "loss": 2.4771,
      "theoretical_loss": 3.5905566426582074,
      "tokens_seen": 1188823040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040696519765462456,
      "loss": 2.6622,
      "theoretical_loss": 3.5905384038141674,
      "tokens_seen": 1188888576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004069486476262531,
      "loss": 2.6766,
      "theoretical_loss": 3.5905201662569866,
      "tokens_seen": 1188954112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004069320975978816,
      "loss": 2.77,
      "theoretical_loss": 3.5905019299865018,
      "tokens_seen": 1189019648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004069155475695101,
      "loss": 2.9232,
      "theoretical_loss": 3.590483695002552,
      "tokens_seen": 1189085184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040689899754113866,
      "loss": 2.5458,
      "theoretical_loss": 3.5904654613049756,
      "tokens_seen": 1189150720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004068824475127672,
      "loss": 2.4873,
      "theoretical_loss": 3.5904472288936105,
      "tokens_seen": 1189216256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004068658974843957,
      "loss": 2.7772,
      "theoretical_loss": 3.5904289977682957,
      "tokens_seen": 1189281792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004068493474560242,
      "loss": 2.643,
      "theoretical_loss": 3.590410767928869,
      "tokens_seen": 1189347328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040683279742765276,
      "loss": 2.7212,
      "theoretical_loss": 3.5903925393751694,
      "tokens_seen": 1189412864
    },
    {
      "epoch": 0.42,
      "objective/train/avg_token_score": 0.007362178526818752,
      "objective/train/avg_weight": 0.9913352727890015,
      "objective/train/docs_used": 674473,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7501187324523926,
      "objective/train/std_weight": 0.05132012814283371,
      "objective/train/theoretical_loss": 3.590374312107035,
      "objective/train/tokens_used": 1209938400,
      "theoretical_loss": 3.590374312107035,
      "tokens_seen": 1189478400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004068162473992813,
      "loss": 2.7874,
      "theoretical_loss": 3.590374312107035,
      "tokens_seen": 1189478400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004067996973709098,
      "loss": 2.6273,
      "theoretical_loss": 3.590356086124305,
      "tokens_seen": 1189543936
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004067831473425383,
      "loss": 2.9484,
      "theoretical_loss": 3.590337861426817,
      "tokens_seen": 1189609472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040676659731416685,
      "loss": 2.6169,
      "theoretical_loss": 3.59031963801441,
      "tokens_seen": 1189675008
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004067500472857954,
      "loss": 2.6966,
      "theoretical_loss": 3.5903014158869224,
      "tokens_seen": 1189740544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004067334972574239,
      "loss": 2.9584,
      "theoretical_loss": 3.5902831950441936,
      "tokens_seen": 1189806080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0004067169472290524,
      "loss": 2.599,
      "theoretical_loss": 3.590264975486061,
      "tokens_seen": 1189871616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00040670039720068095,
      "loss": 2.6419,
      "theoretical_loss": 3.5902467572123644,
      "tokens_seen": 1189937152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004066838471723095,
      "loss": 2.5033,
      "theoretical_loss": 3.5902285402229426,
      "tokens_seen": 1190002688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040666729714393797,
      "loss": 2.7279,
      "theoretical_loss": 3.590210324517633,
      "tokens_seen": 1190068224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004066507471155665,
      "loss": 2.9156,
      "theoretical_loss": 3.5901921100962753,
      "tokens_seen": 1190133760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040663419708719504,
      "loss": 2.6526,
      "theoretical_loss": 3.590173896958709,
      "tokens_seen": 1190199296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004066176470588236,
      "loss": 2.776,
      "theoretical_loss": 3.5901556851047713,
      "tokens_seen": 1190264832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040660109703045207,
      "loss": 2.6778,
      "theoretical_loss": 3.590137474534302,
      "tokens_seen": 1190330368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004065845470020806,
      "loss": 2.839,
      "theoretical_loss": 3.5901192652471403,
      "tokens_seen": 1190395904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040656799697370914,
      "loss": 2.8508,
      "theoretical_loss": 3.5901010572431247,
      "tokens_seen": 1190461440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004065514469453377,
      "loss": 2.5815,
      "theoretical_loss": 3.5900828505220943,
      "tokens_seen": 1190526976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040653489691696616,
      "loss": 2.8488,
      "theoretical_loss": 3.5900646450838876,
      "tokens_seen": 1190592512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040651834688859464,
      "loss": 2.7928,
      "theoretical_loss": 3.590046440928344,
      "tokens_seen": 1190658048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040650179686022313,
      "loss": 2.8088,
      "theoretical_loss": 3.590028238055303,
      "tokens_seen": 1190723584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040648524683185167,
      "loss": 2.8384,
      "theoretical_loss": 3.5900100364646024,
      "tokens_seen": 1190789120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004064686968034802,
      "loss": 2.8044,
      "theoretical_loss": 3.5899918361560825,
      "tokens_seen": 1190854656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040645214677510874,
      "loss": 2.7715,
      "theoretical_loss": 3.5899736371295825,
      "tokens_seen": 1190920192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004064355967467372,
      "loss": 2.9796,
      "theoretical_loss": 3.5899554393849398,
      "tokens_seen": 1190985728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040641904671836576,
      "loss": 2.7621,
      "theoretical_loss": 3.589937242921996,
      "tokens_seen": 1191051264
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004064024966899943,
      "loss": 2.7276,
      "theoretical_loss": 3.5899190477405885,
      "tokens_seen": 1191116800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040638594666162284,
      "loss": 2.6725,
      "theoretical_loss": 3.5899008538405575,
      "tokens_seen": 1191182336
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004063693966332513,
      "loss": 2.8647,
      "theoretical_loss": 3.5898826612217416,
      "tokens_seen": 1191247872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040635284660487986,
      "loss": 2.6336,
      "theoretical_loss": 3.589864469883981,
      "tokens_seen": 1191313408
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004063362965765084,
      "loss": 2.3876,
      "theoretical_loss": 3.589846279827114,
      "tokens_seen": 1191378944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040631974654813693,
      "loss": 2.9075,
      "theoretical_loss": 3.5898280910509808,
      "tokens_seen": 1191444480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004063031965197654,
      "loss": 2.8295,
      "theoretical_loss": 3.5898099035554196,
      "tokens_seen": 1191510016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040628664649139395,
      "loss": 2.6297,
      "theoretical_loss": 3.5897917173402716,
      "tokens_seen": 1191575552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004062700964630225,
      "loss": 2.6019,
      "theoretical_loss": 3.589773532405375,
      "tokens_seen": 1191641088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040625354643465103,
      "loss": 2.7656,
      "theoretical_loss": 3.589755348750569,
      "tokens_seen": 1191706624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004062369964062795,
      "loss": 2.7168,
      "theoretical_loss": 3.5897371663756945,
      "tokens_seen": 1191772160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040622044637790805,
      "loss": 3.0322,
      "theoretical_loss": 3.58971898528059,
      "tokens_seen": 1191837696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004062038963495366,
      "loss": 2.5494,
      "theoretical_loss": 3.5897008054650947,
      "tokens_seen": 1191903232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004061873463211651,
      "loss": 2.862,
      "theoretical_loss": 3.589682626929049,
      "tokens_seen": 1191968768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004061707962927936,
      "loss": 2.6778,
      "theoretical_loss": 3.5896644496722927,
      "tokens_seen": 1192034304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040615424626442215,
      "loss": 2.6254,
      "theoretical_loss": 3.5896462736946644,
      "tokens_seen": 1192099840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004061376962360507,
      "loss": 2.7371,
      "theoretical_loss": 3.5896280989960045,
      "tokens_seen": 1192165376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004061211462076792,
      "loss": 2.8648,
      "theoretical_loss": 3.589609925576153,
      "tokens_seen": 1192230912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004061045961793077,
      "loss": 2.4284,
      "theoretical_loss": 3.5895917534349486,
      "tokens_seen": 1192296448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040608804615093624,
      "loss": 2.6231,
      "theoretical_loss": 3.589573582572232,
      "tokens_seen": 1192361984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004060714961225648,
      "loss": 2.8056,
      "theoretical_loss": 3.589555412987843,
      "tokens_seen": 1192427520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004060549460941933,
      "loss": 2.7924,
      "theoretical_loss": 3.5895372446816203,
      "tokens_seen": 1192493056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004060383960658218,
      "loss": 2.6701,
      "theoretical_loss": 3.5895190776534047,
      "tokens_seen": 1192558592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040602184603745034,
      "loss": 2.8923,
      "theoretical_loss": 3.5895009119030368,
      "tokens_seen": 1192624128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004060052960090789,
      "loss": 2.9482,
      "theoretical_loss": 3.5894827474303543,
      "tokens_seen": 1192689664
    },
    {
      "epoch": 0.43,
      "objective/train/avg_token_score": 0.03687932714819908,
      "objective/train/avg_weight": 0.9566222429275513,
      "objective/train/docs_used": 676257,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.768890857696533,
      "objective/train/std_weight": 0.1592606008052826,
      "objective/train/theoretical_loss": 3.5894645842351993,
      "objective/train/tokens_used": 1213215200,
      "theoretical_loss": 3.5894645842351993,
      "tokens_seen": 1192755200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004059887459807074,
      "loss": 2.6921,
      "theoretical_loss": 3.5894645842351993,
      "tokens_seen": 1192755200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004059721959523359,
      "loss": 2.5737,
      "theoretical_loss": 3.589446422317411,
      "tokens_seen": 1192820736
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040595564592396443,
      "loss": 2.7125,
      "theoretical_loss": 3.5894282616768285,
      "tokens_seen": 1192886272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040593909589559297,
      "loss": 2.5927,
      "theoretical_loss": 3.5894101023132934,
      "tokens_seen": 1192951808
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004059225458672215,
      "loss": 2.8644,
      "theoretical_loss": 3.5893919442266444,
      "tokens_seen": 1193017344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040590599583885,
      "loss": 2.8746,
      "theoretical_loss": 3.5893737874167226,
      "tokens_seen": 1193082880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040588944581047853,
      "loss": 2.6132,
      "theoretical_loss": 3.5893556318833673,
      "tokens_seen": 1193148416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040587289578210707,
      "loss": 2.8984,
      "theoretical_loss": 3.5893374776264197,
      "tokens_seen": 1193213952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004058563457537356,
      "loss": 2.7232,
      "theoretical_loss": 3.589319324645719,
      "tokens_seen": 1193279488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004058397957253641,
      "loss": 2.8203,
      "theoretical_loss": 3.589301172941105,
      "tokens_seen": 1193345024
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004058232456969926,
      "loss": 2.7496,
      "theoretical_loss": 3.5892830225124186,
      "tokens_seen": 1193410560
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040580669566862116,
      "loss": 3.0133,
      "theoretical_loss": 3.5892648733595007,
      "tokens_seen": 1193476096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004057901456402497,
      "loss": 2.68,
      "theoretical_loss": 3.589246725482191,
      "tokens_seen": 1193541632
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004057735956118782,
      "loss": 2.623,
      "theoretical_loss": 3.589228578880329,
      "tokens_seen": 1193607168
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004057570455835067,
      "loss": 2.5235,
      "theoretical_loss": 3.589210433553756,
      "tokens_seen": 1193672704
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040574049555513526,
      "loss": 2.8356,
      "theoretical_loss": 3.5891922895023125,
      "tokens_seen": 1193738240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004057239455267638,
      "loss": 2.6202,
      "theoretical_loss": 3.5891741467258385,
      "tokens_seen": 1193803776
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004057073954983923,
      "loss": 2.7073,
      "theoretical_loss": 3.589156005224174,
      "tokens_seen": 1193869312
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004056908454700208,
      "loss": 2.5897,
      "theoretical_loss": 3.58913786499716,
      "tokens_seen": 1193934848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040567429544164935,
      "loss": 2.7334,
      "theoretical_loss": 3.589119726044637,
      "tokens_seen": 1194000384
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004056577454132779,
      "loss": 2.642,
      "theoretical_loss": 3.5891015883664457,
      "tokens_seen": 1194065920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004056411953849064,
      "loss": 2.6788,
      "theoretical_loss": 3.5890834519624257,
      "tokens_seen": 1194131456
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004056246453565349,
      "loss": 2.5387,
      "theoretical_loss": 3.5890653168324187,
      "tokens_seen": 1194196992
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040560809532816345,
      "loss": 2.7178,
      "theoretical_loss": 3.5890471829762647,
      "tokens_seen": 1194262528
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000405591545299792,
      "loss": 2.401,
      "theoretical_loss": 3.5890290503938047,
      "tokens_seen": 1194328064
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040557499527142047,
      "loss": 2.4512,
      "theoretical_loss": 3.5890109190848785,
      "tokens_seen": 1194393600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000405558445243049,
      "loss": 2.8893,
      "theoretical_loss": 3.5889927890493274,
      "tokens_seen": 1194459136
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004055418952146775,
      "loss": 2.6407,
      "theoretical_loss": 3.5889746602869925,
      "tokens_seen": 1194524672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000405525345186306,
      "loss": 2.6228,
      "theoretical_loss": 3.588956532797714,
      "tokens_seen": 1194590208
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004055087951579345,
      "loss": 3.0062,
      "theoretical_loss": 3.5889384065813323,
      "tokens_seen": 1194655744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040549224512956305,
      "loss": 2.5287,
      "theoretical_loss": 3.588920281637689,
      "tokens_seen": 1194721280
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004054756951011916,
      "loss": 2.4063,
      "theoretical_loss": 3.5889021579666247,
      "tokens_seen": 1194786816
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040545914507282007,
      "loss": 2.7456,
      "theoretical_loss": 3.58888403556798,
      "tokens_seen": 1194852352
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004054425950444486,
      "loss": 2.8074,
      "theoretical_loss": 3.5888659144415964,
      "tokens_seen": 1194917888
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040542604501607715,
      "loss": 2.6859,
      "theoretical_loss": 3.5888477945873136,
      "tokens_seen": 1194983424
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004054094949877057,
      "loss": 2.4768,
      "theoretical_loss": 3.5888296760049743,
      "tokens_seen": 1195048960
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040539294495933417,
      "loss": 2.8443,
      "theoretical_loss": 3.5888115586944176,
      "tokens_seen": 1195114496
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004053763949309627,
      "loss": 2.7698,
      "theoretical_loss": 3.5887934426554855,
      "tokens_seen": 1195180032
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040535984490259124,
      "loss": 2.6935,
      "theoretical_loss": 3.588775327888019,
      "tokens_seen": 1195245568
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004053432948742198,
      "loss": 3.0128,
      "theoretical_loss": 3.588757214391859,
      "tokens_seen": 1195311104
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040532674484584826,
      "loss": 2.79,
      "theoretical_loss": 3.588739102166847,
      "tokens_seen": 1195376640
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004053101948174768,
      "loss": 2.5918,
      "theoretical_loss": 3.5887209912128233,
      "tokens_seen": 1195442176
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040529364478910534,
      "loss": 2.9214,
      "theoretical_loss": 3.5887028815296294,
      "tokens_seen": 1195507712
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004052770947607339,
      "loss": 2.8156,
      "theoretical_loss": 3.588684773117107,
      "tokens_seen": 1195573248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040526054473236236,
      "loss": 2.8521,
      "theoretical_loss": 3.5886666659750963,
      "tokens_seen": 1195638784
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004052439947039909,
      "loss": 2.8039,
      "theoretical_loss": 3.5886485601034397,
      "tokens_seen": 1195704320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040522744467561943,
      "loss": 2.849,
      "theoretical_loss": 3.5886304555019772,
      "tokens_seen": 1195769856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040521089464724797,
      "loss": 2.8298,
      "theoretical_loss": 3.588612352170551,
      "tokens_seen": 1195835392
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040519434461887646,
      "loss": 2.6943,
      "theoretical_loss": 3.5885942501090025,
      "tokens_seen": 1195900928
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000405177794590505,
      "loss": 2.7567,
      "theoretical_loss": 3.588576149317172,
      "tokens_seen": 1195966464
    },
    {
      "epoch": 0.43,
      "objective/train/avg_token_score": 0.020835787057876587,
      "objective/train/avg_weight": 0.9754777550697327,
      "objective/train/docs_used": 678049,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.866452217102051,
      "objective/train/std_weight": 0.07386878877878189,
      "objective/train/theoretical_loss": 3.588558049794902,
      "objective/train/tokens_used": 1216492000,
      "theoretical_loss": 3.588558049794902,
      "tokens_seen": 1196032000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040516124456213353,
      "loss": 2.8399,
      "theoretical_loss": 3.588558049794902,
      "tokens_seen": 1196032000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040514469453376207,
      "loss": 2.8249,
      "theoretical_loss": 3.588539951542033,
      "tokens_seen": 1196097536
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040512814450539055,
      "loss": 2.7864,
      "theoretical_loss": 3.588521854558407,
      "tokens_seen": 1196163072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004051115944770191,
      "loss": 2.5661,
      "theoretical_loss": 3.5885037588438657,
      "tokens_seen": 1196228608
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004050950444486476,
      "loss": 2.8327,
      "theoretical_loss": 3.5884856643982497,
      "tokens_seen": 1196294144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040507849442027616,
      "loss": 2.6844,
      "theoretical_loss": 3.5884675712214014,
      "tokens_seen": 1196359680
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040506194439190465,
      "loss": 2.6654,
      "theoretical_loss": 3.588449479313162,
      "tokens_seen": 1196425216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004050453943635332,
      "loss": 2.5876,
      "theoretical_loss": 3.5884313886733725,
      "tokens_seen": 1196490752
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004050288443351617,
      "loss": 2.6003,
      "theoretical_loss": 3.5884132993018754,
      "tokens_seen": 1196556288
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040501229430679026,
      "loss": 2.5613,
      "theoretical_loss": 3.588395211198512,
      "tokens_seen": 1196621824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040499574427841874,
      "loss": 2.753,
      "theoretical_loss": 3.5883771243631237,
      "tokens_seen": 1196687360
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004049791942500473,
      "loss": 2.7101,
      "theoretical_loss": 3.588359038795552,
      "tokens_seen": 1196752896
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004049626442216758,
      "loss": 2.7128,
      "theoretical_loss": 3.5883409544956395,
      "tokens_seen": 1196818432
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040494609419330436,
      "loss": 2.5168,
      "theoretical_loss": 3.5883228714632276,
      "tokens_seen": 1196883968
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040492954416493284,
      "loss": 2.4495,
      "theoretical_loss": 3.5883047896981575,
      "tokens_seen": 1196949504
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004049129941365614,
      "loss": 2.6572,
      "theoretical_loss": 3.588286709200271,
      "tokens_seen": 1197015040
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004048964441081899,
      "loss": 2.684,
      "theoretical_loss": 3.588268629969411,
      "tokens_seen": 1197080576
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040487989407981845,
      "loss": 2.6282,
      "theoretical_loss": 3.5882505520054186,
      "tokens_seen": 1197146112
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040486334405144694,
      "loss": 2.8388,
      "theoretical_loss": 3.5882324753081356,
      "tokens_seen": 1197211648
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040484679402307547,
      "loss": 2.6952,
      "theoretical_loss": 3.588214399877404,
      "tokens_seen": 1197277184
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000404830243994704,
      "loss": 2.8657,
      "theoretical_loss": 3.588196325713066,
      "tokens_seen": 1197342720
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040481369396633255,
      "loss": 2.4558,
      "theoretical_loss": 3.588178252814963,
      "tokens_seen": 1197408256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040479714393796103,
      "loss": 2.4258,
      "theoretical_loss": 3.5881601811829373,
      "tokens_seen": 1197473792
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040478059390958957,
      "loss": 2.7427,
      "theoretical_loss": 3.5881421108168317,
      "tokens_seen": 1197539328
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004047640438812181,
      "loss": 2.8167,
      "theoretical_loss": 3.588124041716487,
      "tokens_seen": 1197604864
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040474749385284664,
      "loss": 2.6207,
      "theoretical_loss": 3.5881059738817456,
      "tokens_seen": 1197670400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040473094382447513,
      "loss": 2.7086,
      "theoretical_loss": 3.58808790731245,
      "tokens_seen": 1197735936
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040471439379610366,
      "loss": 2.8529,
      "theoretical_loss": 3.588069842008442,
      "tokens_seen": 1197801472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004046978437677322,
      "loss": 2.558,
      "theoretical_loss": 3.5880517779695644,
      "tokens_seen": 1197867008
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040468129373936074,
      "loss": 2.8852,
      "theoretical_loss": 3.5880337151956585,
      "tokens_seen": 1197932544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004046647437109892,
      "loss": 2.7526,
      "theoretical_loss": 3.588015653686567,
      "tokens_seen": 1197998080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040464819368261776,
      "loss": 2.7628,
      "theoretical_loss": 3.5879975934421324,
      "tokens_seen": 1198063616
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004046316436542463,
      "loss": 2.5076,
      "theoretical_loss": 3.587979534462196,
      "tokens_seen": 1198129152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040461509362587484,
      "loss": 2.7731,
      "theoretical_loss": 3.5879614767466013,
      "tokens_seen": 1198194688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004045985435975033,
      "loss": 2.6219,
      "theoretical_loss": 3.58794342029519,
      "tokens_seen": 1198260224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040458199356913186,
      "loss": 2.7588,
      "theoretical_loss": 3.5879253651078047,
      "tokens_seen": 1198325760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004045654435407604,
      "loss": 2.5333,
      "theoretical_loss": 3.5879073111842876,
      "tokens_seen": 1198391296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004045488935123888,
      "loss": 2.8961,
      "theoretical_loss": 3.5878892585244806,
      "tokens_seen": 1198456832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040453234348401736,
      "loss": 2.6263,
      "theoretical_loss": 3.5878712071282273,
      "tokens_seen": 1198522368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004045157934556459,
      "loss": 2.7097,
      "theoretical_loss": 3.5878531569953696,
      "tokens_seen": 1198587904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040449924342727444,
      "loss": 2.6454,
      "theoretical_loss": 3.58783510812575,
      "tokens_seen": 1198653440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004044826933989029,
      "loss": 2.7776,
      "theoretical_loss": 3.5878170605192103,
      "tokens_seen": 1198718976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040446614337053146,
      "loss": 2.7398,
      "theoretical_loss": 3.5877990141755944,
      "tokens_seen": 1198784512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040444959334216,
      "loss": 2.8954,
      "theoretical_loss": 3.5877809690947444,
      "tokens_seen": 1198850048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040443304331378853,
      "loss": 2.6385,
      "theoretical_loss": 3.5877629252765026,
      "tokens_seen": 1198915584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000404416493285417,
      "loss": 2.6518,
      "theoretical_loss": 3.587744882720712,
      "tokens_seen": 1198981120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040439994325704555,
      "loss": 2.7233,
      "theoretical_loss": 3.587726841427215,
      "tokens_seen": 1199046656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004043833932286741,
      "loss": 2.6762,
      "theoretical_loss": 3.5877088013958547,
      "tokens_seen": 1199112192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040436684320030263,
      "loss": 2.8228,
      "theoretical_loss": 3.5876907626264734,
      "tokens_seen": 1199177728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004043502931719311,
      "loss": 2.5468,
      "theoretical_loss": 3.5876727251189138,
      "tokens_seen": 1199243264
    },
    {
      "epoch": 0.43,
      "objective/train/avg_token_score": 0.017939018085598946,
      "objective/train/avg_weight": 0.978886604309082,
      "objective/train/docs_used": 679914,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6579363346099854,
      "objective/train/std_weight": 0.06393834948539734,
      "objective/train/theoretical_loss": 3.5876546888730187,
      "objective/train/tokens_used": 1219768800,
      "theoretical_loss": 3.5876546888730187,
      "tokens_seen": 1199308800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040433374314355965,
      "loss": 2.555,
      "theoretical_loss": 3.5876546888730187,
      "tokens_seen": 1199308800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004043171931151882,
      "loss": 2.6574,
      "theoretical_loss": 3.5876366538886315,
      "tokens_seen": 1199374336
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004043006430868167,
      "loss": 2.6105,
      "theoretical_loss": 3.5876186201655944,
      "tokens_seen": 1199439872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004042840930584452,
      "loss": 2.7541,
      "theoretical_loss": 3.5876005877037507,
      "tokens_seen": 1199505408
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040426754303007374,
      "loss": 2.5393,
      "theoretical_loss": 3.587582556502943,
      "tokens_seen": 1199570944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004042509930017023,
      "loss": 2.6514,
      "theoretical_loss": 3.5875645265630145,
      "tokens_seen": 1199636480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004042344429733308,
      "loss": 2.7999,
      "theoretical_loss": 3.5875464978838076,
      "tokens_seen": 1199702016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004042178929449593,
      "loss": 2.9084,
      "theoretical_loss": 3.587528470465166,
      "tokens_seen": 1199767552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040420134291658784,
      "loss": 2.6422,
      "theoretical_loss": 3.5875104443069326,
      "tokens_seen": 1199833088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004041847928882164,
      "loss": 2.5983,
      "theoretical_loss": 3.5874924194089504,
      "tokens_seen": 1199898624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004041682428598449,
      "loss": 2.6971,
      "theoretical_loss": 3.5874743957710615,
      "tokens_seen": 1199964160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004041516928314734,
      "loss": 2.6457,
      "theoretical_loss": 3.5874563733931106,
      "tokens_seen": 1200029696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040413514280310194,
      "loss": 2.6653,
      "theoretical_loss": 3.5874383522749396,
      "tokens_seen": 1200095232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004041185927747305,
      "loss": 2.7508,
      "theoretical_loss": 3.587420332416392,
      "tokens_seen": 1200160768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000404102042746359,
      "loss": 2.788,
      "theoretical_loss": 3.5874023138173117,
      "tokens_seen": 1200226304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004040854927179875,
      "loss": 2.8331,
      "theoretical_loss": 3.5873842964775404,
      "tokens_seen": 1200291840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040406894268961603,
      "loss": 2.8617,
      "theoretical_loss": 3.587366280396923,
      "tokens_seen": 1200357376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040405239266124457,
      "loss": 2.7844,
      "theoretical_loss": 3.5873482655753013,
      "tokens_seen": 1200422912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004040358426328731,
      "loss": 2.8818,
      "theoretical_loss": 3.5873302520125194,
      "tokens_seen": 1200488448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004040192926045016,
      "loss": 2.9499,
      "theoretical_loss": 3.587312239708421,
      "tokens_seen": 1200553984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040400274257613013,
      "loss": 2.5888,
      "theoretical_loss": 3.587294228662848,
      "tokens_seen": 1200619520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040398619254775867,
      "loss": 2.6406,
      "theoretical_loss": 3.5872762188756453,
      "tokens_seen": 1200685056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004039696425193872,
      "loss": 2.9891,
      "theoretical_loss": 3.587258210346656,
      "tokens_seen": 1200750592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004039530924910157,
      "loss": 2.8672,
      "theoretical_loss": 3.5872402030757224,
      "tokens_seen": 1200816128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004039365424626442,
      "loss": 2.7714,
      "theoretical_loss": 3.5872221970626894,
      "tokens_seen": 1200881664
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040391999243427276,
      "loss": 2.6341,
      "theoretical_loss": 3.5872041923074,
      "tokens_seen": 1200947200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004039034424059013,
      "loss": 2.8571,
      "theoretical_loss": 3.587186188809697,
      "tokens_seen": 1201012736
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004038868923775298,
      "loss": 2.8282,
      "theoretical_loss": 3.587168186569425,
      "tokens_seen": 1201078272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004038703423491583,
      "loss": 2.5316,
      "theoretical_loss": 3.5871501855864274,
      "tokens_seen": 1201143808
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040385379232078686,
      "loss": 2.8037,
      "theoretical_loss": 3.587132185860547,
      "tokens_seen": 1201209344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004038372422924154,
      "loss": 2.8142,
      "theoretical_loss": 3.587114187391628,
      "tokens_seen": 1201274880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004038206922640439,
      "loss": 2.9845,
      "theoretical_loss": 3.587096190179514,
      "tokens_seen": 1201340416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004038041422356724,
      "loss": 2.7024,
      "theoretical_loss": 3.5870781942240493,
      "tokens_seen": 1201405952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040378759220730095,
      "loss": 2.852,
      "theoretical_loss": 3.5870601995250766,
      "tokens_seen": 1201471488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004037710421789295,
      "loss": 2.664,
      "theoretical_loss": 3.58704220608244,
      "tokens_seen": 1201537024
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000403754492150558,
      "loss": 2.6049,
      "theoretical_loss": 3.587024213895983,
      "tokens_seen": 1201602560
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004037379421221865,
      "loss": 2.5996,
      "theoretical_loss": 3.58700622296555,
      "tokens_seen": 1201668096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040372139209381505,
      "loss": 2.7502,
      "theoretical_loss": 3.5869882332909846,
      "tokens_seen": 1201733632
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004037048420654436,
      "loss": 2.7635,
      "theoretical_loss": 3.5869702448721306,
      "tokens_seen": 1201799168
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040368829203707207,
      "loss": 2.7815,
      "theoretical_loss": 3.5869522577088317,
      "tokens_seen": 1201864704
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004036717420087006,
      "loss": 2.4943,
      "theoretical_loss": 3.5869342718009323,
      "tokens_seen": 1201930240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040365519198032915,
      "loss": 2.6934,
      "theoretical_loss": 3.586916287148276,
      "tokens_seen": 1201995776
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004036386419519577,
      "loss": 2.6918,
      "theoretical_loss": 3.5868983037507065,
      "tokens_seen": 1202061312
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040362209192358617,
      "loss": 2.5058,
      "theoretical_loss": 3.586880321608068,
      "tokens_seen": 1202126848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004036055418952147,
      "loss": 2.7083,
      "theoretical_loss": 3.586862340720205,
      "tokens_seen": 1202192384
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040358899186684324,
      "loss": 2.6358,
      "theoretical_loss": 3.5868443610869605,
      "tokens_seen": 1202257920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040357244183847167,
      "loss": 2.7563,
      "theoretical_loss": 3.58682638270818,
      "tokens_seen": 1202323456
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004035558918101002,
      "loss": 2.7825,
      "theoretical_loss": 3.5868084055837066,
      "tokens_seen": 1202388992
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040353934178172875,
      "loss": 2.7155,
      "theoretical_loss": 3.5867904297133846,
      "tokens_seen": 1202454528
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040352279175335723,
      "loss": 2.7048,
      "theoretical_loss": 3.5867724550970577,
      "tokens_seen": 1202520064
    },
    {
      "epoch": 0.43,
      "objective/train/avg_token_score": 0.009649103507399559,
      "objective/train/avg_weight": 0.9886401891708374,
      "objective/train/docs_used": 681814,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3006649017333984,
      "objective/train/std_weight": 0.06972628086805344,
      "objective/train/theoretical_loss": 3.5867544817345713,
      "objective/train/tokens_used": 1223045600,
      "theoretical_loss": 3.5867544817345713,
      "tokens_seen": 1202585600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040350624172498577,
      "loss": 2.8637,
      "theoretical_loss": 3.5867544817345713,
      "tokens_seen": 1202585600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004034896916966143,
      "loss": 2.7562,
      "theoretical_loss": 3.5867365096257684,
      "tokens_seen": 1202651136
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040347314166824284,
      "loss": 2.9959,
      "theoretical_loss": 3.5867185387704943,
      "tokens_seen": 1202716672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004034565916398713,
      "loss": 2.7694,
      "theoretical_loss": 3.586700569168592,
      "tokens_seen": 1202782208
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040344004161149986,
      "loss": 2.8166,
      "theoretical_loss": 3.5866826008199073,
      "tokens_seen": 1202847744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004034234915831284,
      "loss": 2.6976,
      "theoretical_loss": 3.5866646337242836,
      "tokens_seen": 1202913280
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040340694155475694,
      "loss": 3.1595,
      "theoretical_loss": 3.586646667881565,
      "tokens_seen": 1202978816
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004033903915263854,
      "loss": 2.6227,
      "theoretical_loss": 3.586628703291596,
      "tokens_seen": 1203044352
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040337384149801396,
      "loss": 2.8111,
      "theoretical_loss": 3.5866107399542217,
      "tokens_seen": 1203109888
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004033572914696425,
      "loss": 2.8755,
      "theoretical_loss": 3.5865927778692868,
      "tokens_seen": 1203175424
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040334074144127103,
      "loss": 2.8519,
      "theoretical_loss": 3.5865748170366345,
      "tokens_seen": 1203240960
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004033241914128995,
      "loss": 2.8349,
      "theoretical_loss": 3.5865568574561095,
      "tokens_seen": 1203306496
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040330764138452805,
      "loss": 2.6812,
      "theoretical_loss": 3.5865388991275573,
      "tokens_seen": 1203372032
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004032910913561566,
      "loss": 2.5362,
      "theoretical_loss": 3.5865209420508215,
      "tokens_seen": 1203437568
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040327454132778513,
      "loss": 2.868,
      "theoretical_loss": 3.5865029862257476,
      "tokens_seen": 1203503104
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004032579912994136,
      "loss": 2.8139,
      "theoretical_loss": 3.5864850316521792,
      "tokens_seen": 1203568640
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040324144127104215,
      "loss": 2.7631,
      "theoretical_loss": 3.586467078329961,
      "tokens_seen": 1203634176
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004032248912426707,
      "loss": 2.7128,
      "theoretical_loss": 3.586449126258939,
      "tokens_seen": 1203699712
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004032083412142992,
      "loss": 2.8169,
      "theoretical_loss": 3.586431175438956,
      "tokens_seen": 1203765248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004031917911859277,
      "loss": 2.8155,
      "theoretical_loss": 3.5864132258698582,
      "tokens_seen": 1203830784
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040317524115755625,
      "loss": 2.538,
      "theoretical_loss": 3.58639527755149,
      "tokens_seen": 1203896320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004031586911291848,
      "loss": 2.5538,
      "theoretical_loss": 3.5863773304836952,
      "tokens_seen": 1203961856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004031421411008133,
      "loss": 2.5722,
      "theoretical_loss": 3.58635938466632,
      "tokens_seen": 1204027392
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004031255910724418,
      "loss": 2.8791,
      "theoretical_loss": 3.586341440099208,
      "tokens_seen": 1204092928
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040310904104407034,
      "loss": 2.5981,
      "theoretical_loss": 3.586323496782205,
      "tokens_seen": 1204158464
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004030924910156989,
      "loss": 2.8761,
      "theoretical_loss": 3.5863055547151554,
      "tokens_seen": 1204224000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004030759409873274,
      "loss": 2.8306,
      "theoretical_loss": 3.586287613897904,
      "tokens_seen": 1204289536
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004030593909589559,
      "loss": 2.7694,
      "theoretical_loss": 3.586269674330296,
      "tokens_seen": 1204355072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040304284093058444,
      "loss": 2.9008,
      "theoretical_loss": 3.5862517360121773,
      "tokens_seen": 1204420608
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000403026290902213,
      "loss": 2.6908,
      "theoretical_loss": 3.5862337989433906,
      "tokens_seen": 1204486144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004030097408738415,
      "loss": 2.7689,
      "theoretical_loss": 3.5862158631237833,
      "tokens_seen": 1204551680
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004030097408738415,
      "loss": 4.2568,
      "theoretical_loss": 3.5861979285531986,
      "tokens_seen": 1204617216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040299319084547005,
      "loss": 2.9737,
      "theoretical_loss": 3.5861799952314826,
      "tokens_seen": 1204682752
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040299319084547005,
      "loss": 3.0847,
      "theoretical_loss": 3.58616206315848,
      "tokens_seen": 1204748288
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040297664081709853,
      "loss": 4.3956,
      "theoretical_loss": 3.586144132334036,
      "tokens_seen": 1204813824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040296009078872707,
      "loss": 2.7736,
      "theoretical_loss": 3.5861262027579963,
      "tokens_seen": 1204879360
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004029435407603556,
      "loss": 2.6572,
      "theoretical_loss": 3.5861082744302055,
      "tokens_seen": 1204944896
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040292699073198415,
      "loss": 2.6296,
      "theoretical_loss": 3.586090347350509,
      "tokens_seen": 1205010432
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040291044070361263,
      "loss": 2.7935,
      "theoretical_loss": 3.586072421518752,
      "tokens_seen": 1205075968
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040289389067524117,
      "loss": 2.759,
      "theoretical_loss": 3.5860544969347794,
      "tokens_seen": 1205141504
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004028773406468697,
      "loss": 2.6991,
      "theoretical_loss": 3.586036573598437,
      "tokens_seen": 1205207040
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040286079061849824,
      "loss": 2.9021,
      "theoretical_loss": 3.58601865150957,
      "tokens_seen": 1205272576
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004028442405901267,
      "loss": 2.5129,
      "theoretical_loss": 3.586000730668024,
      "tokens_seen": 1205338112
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040282769056175526,
      "loss": 2.8672,
      "theoretical_loss": 3.585982811073644,
      "tokens_seen": 1205403648
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004028111405333838,
      "loss": 2.661,
      "theoretical_loss": 3.5859648927262753,
      "tokens_seen": 1205469184
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040279459050501234,
      "loss": 2.816,
      "theoretical_loss": 3.5859469756257636,
      "tokens_seen": 1205534720
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004027780404766408,
      "loss": 2.773,
      "theoretical_loss": 3.5859290597719546,
      "tokens_seen": 1205600256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040276149044826936,
      "loss": 2.7931,
      "theoretical_loss": 3.585911145164693,
      "tokens_seen": 1205665792
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004027449404198979,
      "loss": 2.7636,
      "theoretical_loss": 3.5858932318038255,
      "tokens_seen": 1205731328
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040272839039152644,
      "loss": 2.8731,
      "theoretical_loss": 3.585875319689196,
      "tokens_seen": 1205796864
    },
    {
      "epoch": 0.43,
      "objective/train/avg_token_score": 0.010162741877138615,
      "objective/train/avg_weight": 0.9880374670028687,
      "objective/train/docs_used": 683756,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0627450942993164,
      "objective/train/std_weight": 0.07086824625730515,
      "objective/train/theoretical_loss": 3.585857408820652,
      "objective/train/tokens_used": 1226322400,
      "theoretical_loss": 3.585857408820652,
      "tokens_seen": 1205862400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004027118403631549,
      "loss": 2.829,
      "theoretical_loss": 3.585857408820652,
      "tokens_seen": 1205862400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040269529033478346,
      "loss": 2.634,
      "theoretical_loss": 3.5858394991980376,
      "tokens_seen": 1205927936
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000402678740306412,
      "loss": 2.8714,
      "theoretical_loss": 3.585821590821199,
      "tokens_seen": 1205993472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040266219027804053,
      "loss": 2.5223,
      "theoretical_loss": 3.5858036836899814,
      "tokens_seen": 1206059008
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000402645640249669,
      "loss": 3.0105,
      "theoretical_loss": 3.5857857778042312,
      "tokens_seen": 1206124544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040262909022129755,
      "loss": 2.6885,
      "theoretical_loss": 3.5857678731637943,
      "tokens_seen": 1206190080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004026125401929261,
      "loss": 2.6528,
      "theoretical_loss": 3.585749969768515,
      "tokens_seen": 1206255616
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004025959901645545,
      "loss": 2.8879,
      "theoretical_loss": 3.585732067618241,
      "tokens_seen": 1206321152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040257944013618306,
      "loss": 2.934,
      "theoretical_loss": 3.5857141667128163,
      "tokens_seen": 1206386688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004025628901078116,
      "loss": 2.7753,
      "theoretical_loss": 3.5856962670520884,
      "tokens_seen": 1206452224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004025463400794401,
      "loss": 2.8716,
      "theoretical_loss": 3.5856783686359015,
      "tokens_seen": 1206517760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004025297900510686,
      "loss": 2.4711,
      "theoretical_loss": 3.5856604714641023,
      "tokens_seen": 1206583296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040251324002269715,
      "loss": 2.8466,
      "theoretical_loss": 3.5856425755365366,
      "tokens_seen": 1206648832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004024966899943257,
      "loss": 3.0283,
      "theoretical_loss": 3.585624680853051,
      "tokens_seen": 1206714368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004024801399659542,
      "loss": 2.8449,
      "theoretical_loss": 3.5856067874134903,
      "tokens_seen": 1206779904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004024635899375827,
      "loss": 2.6403,
      "theoretical_loss": 3.5855888952177013,
      "tokens_seen": 1206845440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040244703990921125,
      "loss": 2.6893,
      "theoretical_loss": 3.585571004265529,
      "tokens_seen": 1206910976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004024304898808398,
      "loss": 2.7671,
      "theoretical_loss": 3.585553114556821,
      "tokens_seen": 1206976512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040241393985246827,
      "loss": 2.8144,
      "theoretical_loss": 3.585535226091422,
      "tokens_seen": 1207042048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004023973898240968,
      "loss": 2.7,
      "theoretical_loss": 3.585517338869179,
      "tokens_seen": 1207107584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040238083979572534,
      "loss": 2.7734,
      "theoretical_loss": 3.5854994528899375,
      "tokens_seen": 1207173120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004023642897673539,
      "loss": 3.0279,
      "theoretical_loss": 3.5854815681535435,
      "tokens_seen": 1207238656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040234773973898237,
      "loss": 2.8224,
      "theoretical_loss": 3.585463684659844,
      "tokens_seen": 1207304192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004023311897106109,
      "loss": 2.827,
      "theoretical_loss": 3.5854458024086853,
      "tokens_seen": 1207369728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040231463968223944,
      "loss": 2.7323,
      "theoretical_loss": 3.585427921399912,
      "tokens_seen": 1207435264
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000402298089653868,
      "loss": 2.7003,
      "theoretical_loss": 3.585410041633372,
      "tokens_seen": 1207500800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040228153962549646,
      "loss": 2.7204,
      "theoretical_loss": 3.585392163108911,
      "tokens_seen": 1207566336
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000402264989597125,
      "loss": 2.7883,
      "theoretical_loss": 3.5853742858263753,
      "tokens_seen": 1207631872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040224843956875354,
      "loss": 2.4668,
      "theoretical_loss": 3.5853564097856117,
      "tokens_seen": 1207697408
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004022318895403821,
      "loss": 2.6616,
      "theoretical_loss": 3.585338534986466,
      "tokens_seen": 1207762944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040221533951201056,
      "loss": 2.7267,
      "theoretical_loss": 3.585320661428784,
      "tokens_seen": 1207828480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004021987894836391,
      "loss": 2.5533,
      "theoretical_loss": 3.585302789112413,
      "tokens_seen": 1207894016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040218223945526763,
      "loss": 2.7583,
      "theoretical_loss": 3.5852849180372006,
      "tokens_seen": 1207959552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040216568942689617,
      "loss": 2.7168,
      "theoretical_loss": 3.585267048202991,
      "tokens_seen": 1208025088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040214913939852465,
      "loss": 2.6405,
      "theoretical_loss": 3.5852491796096317,
      "tokens_seen": 1208090624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004021325893701532,
      "loss": 2.8321,
      "theoretical_loss": 3.5852313122569694,
      "tokens_seen": 1208156160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040211603934178173,
      "loss": 2.6689,
      "theoretical_loss": 3.5852134461448504,
      "tokens_seen": 1208221696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040209948931341027,
      "loss": 2.8024,
      "theoretical_loss": 3.5851955812731218,
      "tokens_seen": 1208287232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040208293928503875,
      "loss": 2.5574,
      "theoretical_loss": 3.585177717641629,
      "tokens_seen": 1208352768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004020663892566673,
      "loss": 2.572,
      "theoretical_loss": 3.58515985525022,
      "tokens_seen": 1208418304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004020498392282958,
      "loss": 2.8424,
      "theoretical_loss": 3.585141994098741,
      "tokens_seen": 1208483840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040203328919992436,
      "loss": 2.6827,
      "theoretical_loss": 3.585124134187038,
      "tokens_seen": 1208549376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040201673917155284,
      "loss": 2.7212,
      "theoretical_loss": 3.585106275514959,
      "tokens_seen": 1208614912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004020001891431814,
      "loss": 2.8281,
      "theoretical_loss": 3.58508841808235,
      "tokens_seen": 1208680448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004019836391148099,
      "loss": 2.8848,
      "theoretical_loss": 3.5850705618890575,
      "tokens_seen": 1208745984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040196708908643846,
      "loss": 2.6033,
      "theoretical_loss": 3.5850527069349285,
      "tokens_seen": 1208811520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040195053905806694,
      "loss": 2.5703,
      "theoretical_loss": 3.5850348532198106,
      "tokens_seen": 1208877056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004019339890296955,
      "loss": 2.8953,
      "theoretical_loss": 3.58501700074355,
      "tokens_seen": 1208942592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000401917439001324,
      "loss": 2.7314,
      "theoretical_loss": 3.5849991495059927,
      "tokens_seen": 1209008128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040190088897295255,
      "loss": 2.8423,
      "theoretical_loss": 3.584981299506987,
      "tokens_seen": 1209073664
    },
    {
      "epoch": 0.43,
      "objective/train/avg_token_score": 0.006638803984969854,
      "objective/train/avg_weight": 0.9921845197677612,
      "objective/train/docs_used": 685102,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.069593906402588,
      "objective/train/std_weight": 0.055420469492673874,
      "objective/train/theoretical_loss": 3.58496345074638,
      "objective/train/tokens_used": 1229599200,
      "theoretical_loss": 3.58496345074638,
      "tokens_seen": 1209139200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040188433894458104,
      "loss": 2.787,
      "theoretical_loss": 3.58496345074638,
      "tokens_seen": 1209139200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004018677889162096,
      "loss": 2.8267,
      "theoretical_loss": 3.5849456032240172,
      "tokens_seen": 1209204736
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004018512388878381,
      "loss": 2.8753,
      "theoretical_loss": 3.584927756939747,
      "tokens_seen": 1209270272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040183468885946665,
      "loss": 2.7004,
      "theoretical_loss": 3.5849099118934156,
      "tokens_seen": 1209335808
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040181813883109513,
      "loss": 2.6386,
      "theoretical_loss": 3.58489206808487,
      "tokens_seen": 1209401344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040180158880272367,
      "loss": 2.7116,
      "theoretical_loss": 3.5848742255139583,
      "tokens_seen": 1209466880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004017850387743522,
      "loss": 2.8433,
      "theoretical_loss": 3.5848563841805268,
      "tokens_seen": 1209532416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040176848874598075,
      "loss": 2.7097,
      "theoretical_loss": 3.5848385440844224,
      "tokens_seen": 1209597952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040175193871760923,
      "loss": 2.8474,
      "theoretical_loss": 3.584820705225493,
      "tokens_seen": 1209663488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040173538868923777,
      "loss": 2.6762,
      "theoretical_loss": 3.584802867603585,
      "tokens_seen": 1209729024
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004017188386608663,
      "loss": 2.645,
      "theoretical_loss": 3.584785031218546,
      "tokens_seen": 1209794560
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040170228863249484,
      "loss": 2.8152,
      "theoretical_loss": 3.584767196070223,
      "tokens_seen": 1209860096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004016857386041233,
      "loss": 2.8135,
      "theoretical_loss": 3.5847493621584636,
      "tokens_seen": 1209925632
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040166918857575186,
      "loss": 2.5776,
      "theoretical_loss": 3.5847315294831152,
      "tokens_seen": 1209991168
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004016526385473804,
      "loss": 2.6263,
      "theoretical_loss": 3.584713698044025,
      "tokens_seen": 1210056704
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040163608851900894,
      "loss": 2.6089,
      "theoretical_loss": 3.58469586784104,
      "tokens_seen": 1210122240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004016195384906374,
      "loss": 2.6965,
      "theoretical_loss": 3.5846780388740074,
      "tokens_seen": 1210187776
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004016029884622659,
      "loss": 2.8222,
      "theoretical_loss": 3.5846602111427757,
      "tokens_seen": 1210253312
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040158643843389444,
      "loss": 2.7843,
      "theoretical_loss": 3.5846423846471915,
      "tokens_seen": 1210318848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004015698884055229,
      "loss": 2.6506,
      "theoretical_loss": 3.5846245593871027,
      "tokens_seen": 1210384384
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040155333837715146,
      "loss": 2.4166,
      "theoretical_loss": 3.584606735362356,
      "tokens_seen": 1210449920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040153678834878,
      "loss": 2.8099,
      "theoretical_loss": 3.5845889125727997,
      "tokens_seen": 1210515456
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040152023832040854,
      "loss": 2.746,
      "theoretical_loss": 3.5845710910182804,
      "tokens_seen": 1210580992
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000401503688292037,
      "loss": 2.6937,
      "theoretical_loss": 3.584553270698647,
      "tokens_seen": 1210646528
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040148713826366556,
      "loss": 2.645,
      "theoretical_loss": 3.584535451613746,
      "tokens_seen": 1210712064
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004014705882352941,
      "loss": 2.8747,
      "theoretical_loss": 3.5845176337634257,
      "tokens_seen": 1210777600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040145403820692263,
      "loss": 2.4868,
      "theoretical_loss": 3.5844998171475333,
      "tokens_seen": 1210843136
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004014374881785511,
      "loss": 2.9643,
      "theoretical_loss": 3.5844820017659167,
      "tokens_seen": 1210908672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040142093815017965,
      "loss": 2.7288,
      "theoretical_loss": 3.5844641876184236,
      "tokens_seen": 1210974208
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004014043881218082,
      "loss": 2.773,
      "theoretical_loss": 3.5844463747049016,
      "tokens_seen": 1211039744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040138783809343673,
      "loss": 2.9521,
      "theoretical_loss": 3.5844285630251984,
      "tokens_seen": 1211105280
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004013712880650652,
      "loss": 2.8437,
      "theoretical_loss": 3.5844107525791618,
      "tokens_seen": 1211170816
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040135473803669375,
      "loss": 2.732,
      "theoretical_loss": 3.58439294336664,
      "tokens_seen": 1211236352
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004013381880083223,
      "loss": 2.7583,
      "theoretical_loss": 3.5843751353874804,
      "tokens_seen": 1211301888
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004013216379799508,
      "loss": 2.6664,
      "theoretical_loss": 3.584357328641531,
      "tokens_seen": 1211367424
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004013050879515793,
      "loss": 2.6384,
      "theoretical_loss": 3.58433952312864,
      "tokens_seen": 1211432960
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040128853792320785,
      "loss": 2.7378,
      "theoretical_loss": 3.5843217188486545,
      "tokens_seen": 1211498496
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004012719878948364,
      "loss": 2.7635,
      "theoretical_loss": 3.584303915801423,
      "tokens_seen": 1211564032
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004012554378664649,
      "loss": 2.8823,
      "theoretical_loss": 3.584286113986794,
      "tokens_seen": 1211629568
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004012388878380934,
      "loss": 2.7154,
      "theoretical_loss": 3.584268313404614,
      "tokens_seen": 1211695104
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040122233780972194,
      "loss": 2.7668,
      "theoretical_loss": 3.584250514054732,
      "tokens_seen": 1211760640
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004012057877813505,
      "loss": 3.0308,
      "theoretical_loss": 3.584232715936997,
      "tokens_seen": 1211826176
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000401189237752979,
      "loss": 2.6019,
      "theoretical_loss": 3.584214919051255,
      "tokens_seen": 1211891712
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004011726877246075,
      "loss": 2.9601,
      "theoretical_loss": 3.584197123397356,
      "tokens_seen": 1211957248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040115613769623604,
      "loss": 2.6206,
      "theoretical_loss": 3.5841793289751465,
      "tokens_seen": 1212022784
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004011395876678646,
      "loss": 2.9661,
      "theoretical_loss": 3.5841615357844763,
      "tokens_seen": 1212088320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004011230376394931,
      "loss": 2.9199,
      "theoretical_loss": 3.584143743825192,
      "tokens_seen": 1212153856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004011064876111216,
      "loss": 2.8054,
      "theoretical_loss": 3.5841259530971428,
      "tokens_seen": 1212219392
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040108993758275013,
      "loss": 2.5797,
      "theoretical_loss": 3.584108163600177,
      "tokens_seen": 1212284928
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040107338755437867,
      "loss": 2.6791,
      "theoretical_loss": 3.584090375334142,
      "tokens_seen": 1212350464
    },
    {
      "epoch": 0.43,
      "objective/train/avg_token_score": 0.0061058844439685345,
      "objective/train/avg_weight": 0.9928159117698669,
      "objective/train/docs_used": 686738,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1542887687683105,
      "objective/train/std_weight": 0.044166449457407,
      "objective/train/theoretical_loss": 3.5840725882988873,
      "objective/train/tokens_used": 1232876000,
      "theoretical_loss": 3.5840725882988873,
      "tokens_seen": 1212416000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004010568375260072,
      "loss": 2.7414,
      "theoretical_loss": 3.5840725882988873,
      "tokens_seen": 1212416000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004010402874976357,
      "loss": 2.8428,
      "theoretical_loss": 3.5840548024942604,
      "tokens_seen": 1212481536
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040102373746926423,
      "loss": 3.0719,
      "theoretical_loss": 3.5840370179201093,
      "tokens_seen": 1212547072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040100718744089277,
      "loss": 2.7242,
      "theoretical_loss": 3.5840192345762834,
      "tokens_seen": 1212612608
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004009906374125213,
      "loss": 2.6831,
      "theoretical_loss": 3.5840014524626307,
      "tokens_seen": 1212678144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004009740873841498,
      "loss": 2.7488,
      "theoretical_loss": 3.5839836715789994,
      "tokens_seen": 1212743680
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004009575373557783,
      "loss": 2.7138,
      "theoretical_loss": 3.5839658919252377,
      "tokens_seen": 1212809216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040094098732740686,
      "loss": 2.4714,
      "theoretical_loss": 3.5839481135011946,
      "tokens_seen": 1212874752
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004009244372990354,
      "loss": 2.6926,
      "theoretical_loss": 3.583930336306719,
      "tokens_seen": 1212940288
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004009078872706639,
      "loss": 2.9572,
      "theoretical_loss": 3.583912560341658,
      "tokens_seen": 1213005824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004008913372422924,
      "loss": 2.7229,
      "theoretical_loss": 3.583894785605862,
      "tokens_seen": 1213071360
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040087478721392096,
      "loss": 2.845,
      "theoretical_loss": 3.583877012099178,
      "tokens_seen": 1213136896
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004008582371855495,
      "loss": 2.6488,
      "theoretical_loss": 3.5838592398214555,
      "tokens_seen": 1213202432
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000400841687157178,
      "loss": 2.8457,
      "theoretical_loss": 3.583841468772543,
      "tokens_seen": 1213267968
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004008251371288065,
      "loss": 2.7144,
      "theoretical_loss": 3.5838236989522887,
      "tokens_seen": 1213333504
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040080858710043506,
      "loss": 2.8686,
      "theoretical_loss": 3.5838059303605423,
      "tokens_seen": 1213399040
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004007920370720636,
      "loss": 2.783,
      "theoretical_loss": 3.583788162997151,
      "tokens_seen": 1213464576
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004007754870436921,
      "loss": 2.5009,
      "theoretical_loss": 3.5837703968619654,
      "tokens_seen": 1213530112
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004007589370153206,
      "loss": 3.0596,
      "theoretical_loss": 3.583752631954833,
      "tokens_seen": 1213595648
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040074238698694915,
      "loss": 2.8661,
      "theoretical_loss": 3.5837348682756027,
      "tokens_seen": 1213661184
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004007258369585777,
      "loss": 2.8352,
      "theoretical_loss": 3.5837171058241237,
      "tokens_seen": 1213726720
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040070928693020617,
      "loss": 2.6173,
      "theoretical_loss": 3.5836993446002445,
      "tokens_seen": 1213792256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004006927369018347,
      "loss": 2.7373,
      "theoretical_loss": 3.5836815846038146,
      "tokens_seen": 1213857792
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040067618687346325,
      "loss": 2.6451,
      "theoretical_loss": 3.5836638258346825,
      "tokens_seen": 1213923328
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004006596368450918,
      "loss": 2.5509,
      "theoretical_loss": 3.5836460682926967,
      "tokens_seen": 1213988864
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040064308681672027,
      "loss": 2.6202,
      "theoretical_loss": 3.583628311977707,
      "tokens_seen": 1214054400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040062653678834875,
      "loss": 2.844,
      "theoretical_loss": 3.583610556889562,
      "tokens_seen": 1214119936
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040060998675997724,
      "loss": 2.6294,
      "theoretical_loss": 3.5835928030281106,
      "tokens_seen": 1214185472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040059343673160577,
      "loss": 2.652,
      "theoretical_loss": 3.583575050393202,
      "tokens_seen": 1214251008
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004005768867032343,
      "loss": 2.4033,
      "theoretical_loss": 3.583557298984685,
      "tokens_seen": 1214316544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040056033667486285,
      "loss": 2.6712,
      "theoretical_loss": 3.583539548802409,
      "tokens_seen": 1214382080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040054378664649133,
      "loss": 2.8206,
      "theoretical_loss": 3.583521799846223,
      "tokens_seen": 1214447616
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040052723661811987,
      "loss": 2.5024,
      "theoretical_loss": 3.5835040521159764,
      "tokens_seen": 1214513152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004005106865897484,
      "loss": 2.8817,
      "theoretical_loss": 3.583486305611518,
      "tokens_seen": 1214578688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040049413656137694,
      "loss": 2.7031,
      "theoretical_loss": 3.5834685603326975,
      "tokens_seen": 1214644224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004004775865330055,
      "loss": 2.6625,
      "theoretical_loss": 3.5834508162793637,
      "tokens_seen": 1214709760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040046103650463396,
      "loss": 2.722,
      "theoretical_loss": 3.583433073451366,
      "tokens_seen": 1214775296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004004444864762625,
      "loss": 2.5363,
      "theoretical_loss": 3.5834153318485535,
      "tokens_seen": 1214840832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040042793644789104,
      "loss": 2.6779,
      "theoretical_loss": 3.5833975914707756,
      "tokens_seen": 1214906368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004004113864195196,
      "loss": 2.7961,
      "theoretical_loss": 3.583379852317882,
      "tokens_seen": 1214971904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040039483639114806,
      "loss": 2.8235,
      "theoretical_loss": 3.5833621143897214,
      "tokens_seen": 1215037440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004003782863627766,
      "loss": 2.4814,
      "theoretical_loss": 3.583344377686144,
      "tokens_seen": 1215102976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040036173633440514,
      "loss": 2.7894,
      "theoretical_loss": 3.5833266422069983,
      "tokens_seen": 1215168512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004003451863060337,
      "loss": 2.4608,
      "theoretical_loss": 3.5833089079521345,
      "tokens_seen": 1215234048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040032863627766216,
      "loss": 2.9011,
      "theoretical_loss": 3.583291174921402,
      "tokens_seen": 1215299584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004003120862492907,
      "loss": 2.5814,
      "theoretical_loss": 3.5832734431146496,
      "tokens_seen": 1215365120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040029553622091923,
      "loss": 2.539,
      "theoretical_loss": 3.5832557125317273,
      "tokens_seen": 1215430656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040027898619254777,
      "loss": 2.8634,
      "theoretical_loss": 3.5832379831724848,
      "tokens_seen": 1215496192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040026243616417625,
      "loss": 2.9015,
      "theoretical_loss": 3.583220255036772,
      "tokens_seen": 1215561728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004002458861358048,
      "loss": 2.9069,
      "theoretical_loss": 3.5832025281244375,
      "tokens_seen": 1215627264
    },
    {
      "epoch": 0.43,
      "objective/train/avg_token_score": 0.020750168710947037,
      "objective/train/avg_weight": 0.975578784942627,
      "objective/train/docs_used": 688520,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.6475157737731934,
      "objective/train/std_weight": 0.12923723459243774,
      "objective/train/theoretical_loss": 3.5831848024353317,
      "objective/train/tokens_used": 1236152800,
      "theoretical_loss": 3.5831848024353317,
      "tokens_seen": 1215692800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040022933610743333,
      "loss": 2.764,
      "theoretical_loss": 3.5831848024353317,
      "tokens_seen": 1215692800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040021278607906187,
      "loss": 2.8946,
      "theoretical_loss": 3.5831670779693034,
      "tokens_seen": 1215758336
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040019623605069035,
      "loss": 2.8321,
      "theoretical_loss": 3.5831493547262037,
      "tokens_seen": 1215823872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004001796860223189,
      "loss": 2.9291,
      "theoretical_loss": 3.5831316327058813,
      "tokens_seen": 1215889408
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004001631359939474,
      "loss": 2.8232,
      "theoretical_loss": 3.5831139119081863,
      "tokens_seen": 1215954944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040014658596557596,
      "loss": 2.8176,
      "theoretical_loss": 3.583096192332968,
      "tokens_seen": 1216020480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040013003593720444,
      "loss": 2.7334,
      "theoretical_loss": 3.583078473980077,
      "tokens_seen": 1216086016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000400113485908833,
      "loss": 2.4886,
      "theoretical_loss": 3.5830607568493624,
      "tokens_seen": 1216151552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004000969358804615,
      "loss": 2.4196,
      "theoretical_loss": 3.583043040940674,
      "tokens_seen": 1216217088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040008038585209006,
      "loss": 2.6469,
      "theoretical_loss": 3.5830253262538623,
      "tokens_seen": 1216282624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040006383582371854,
      "loss": 2.722,
      "theoretical_loss": 3.5830076127887764,
      "tokens_seen": 1216348160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004000472857953471,
      "loss": 2.7963,
      "theoretical_loss": 3.582989900545267,
      "tokens_seen": 1216413696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0004000307357669756,
      "loss": 2.7759,
      "theoretical_loss": 3.5829721895231836,
      "tokens_seen": 1216479232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00040001418573860415,
      "loss": 2.5234,
      "theoretical_loss": 3.5829544797223765,
      "tokens_seen": 1216544768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00039999763571023264,
      "loss": 2.9685,
      "theoretical_loss": 3.582936771142695,
      "tokens_seen": 1216610304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0003999810856818612,
      "loss": 2.6622,
      "theoretical_loss": 3.58291906378399,
      "tokens_seen": 1216675840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0003999645356534897,
      "loss": 2.7524,
      "theoretical_loss": 3.582901357646111,
      "tokens_seen": 1216741376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00039994798562511825,
      "loss": 2.643,
      "theoretical_loss": 3.582883652728908,
      "tokens_seen": 1216806912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00039993143559674673,
      "loss": 3.1272,
      "theoretical_loss": 3.5828659490322323,
      "tokens_seen": 1216872448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00039991488556837527,
      "loss": 2.6074,
      "theoretical_loss": 3.5828482465559324,
      "tokens_seen": 1216937984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0003998983355400038,
      "loss": 2.7624,
      "theoretical_loss": 3.582830545299859,
      "tokens_seen": 1217003520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00039988178551163235,
      "loss": 2.7913,
      "theoretical_loss": 3.582812845263863,
      "tokens_seen": 1217069056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00039986523548326083,
      "loss": 2.773,
      "theoretical_loss": 3.5827951464477934,
      "tokens_seen": 1217134592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00039984868545488937,
      "loss": 2.4826,
      "theoretical_loss": 3.582777448851502,
      "tokens_seen": 1217200128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0003998321354265179,
      "loss": 3.018,
      "theoretical_loss": 3.5827597524748374,
      "tokens_seen": 1217265664
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00039981558539814644,
      "loss": 2.6425,
      "theoretical_loss": 3.5827420573176507,
      "tokens_seen": 1217331200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0003997990353697749,
      "loss": 2.6703,
      "theoretical_loss": 3.582724363379792,
      "tokens_seen": 1217396736
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00039978248534140346,
      "loss": 2.7059,
      "theoretical_loss": 3.5827066706611124,
      "tokens_seen": 1217462272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000399765935313032,
      "loss": 2.6212,
      "theoretical_loss": 3.5826889791614613,
      "tokens_seen": 1217527808
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00039974938528466054,
      "loss": 2.9627,
      "theoretical_loss": 3.5826712888806895,
      "tokens_seen": 1217593344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000399732835256289,
      "loss": 2.626,
      "theoretical_loss": 3.5826535998186477,
      "tokens_seen": 1217658880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00039971628522791756,
      "loss": 2.7965,
      "theoretical_loss": 3.582635911975186,
      "tokens_seen": 1217724416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0003996997351995461,
      "loss": 2.893,
      "theoretical_loss": 3.5826182253501546,
      "tokens_seen": 1217789952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00039968318517117463,
      "loss": 2.8687,
      "theoretical_loss": 3.582600539943405,
      "tokens_seen": 1217855488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0003996666351428031,
      "loss": 2.6238,
      "theoretical_loss": 3.5825828557547865,
      "tokens_seen": 1217921024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039965008511443165,
      "loss": 2.8294,
      "theoretical_loss": 3.5825651727841503,
      "tokens_seen": 1217986560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003996335350860601,
      "loss": 2.8043,
      "theoretical_loss": 3.582547491031347,
      "tokens_seen": 1218052096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003996169850576886,
      "loss": 2.6222,
      "theoretical_loss": 3.582529810496227,
      "tokens_seen": 1218117632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039960043502931716,
      "loss": 2.6309,
      "theoretical_loss": 3.5825121311786416,
      "tokens_seen": 1218183168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003995838850009457,
      "loss": 2.5881,
      "theoretical_loss": 3.58249445307844,
      "tokens_seen": 1218248704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003995673349725742,
      "loss": 2.7717,
      "theoretical_loss": 3.5824767761954748,
      "tokens_seen": 1218314240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003995507849442027,
      "loss": 2.7432,
      "theoretical_loss": 3.5824591005295954,
      "tokens_seen": 1218379776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039953423491583125,
      "loss": 2.6722,
      "theoretical_loss": 3.5824414260806527,
      "tokens_seen": 1218445312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003995176848874598,
      "loss": 2.6918,
      "theoretical_loss": 3.5824237528484977,
      "tokens_seen": 1218510848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003995011348590883,
      "loss": 2.8902,
      "theoretical_loss": 3.582406080832981,
      "tokens_seen": 1218576384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003994845848307168,
      "loss": 2.9736,
      "theoretical_loss": 3.5823884100339543,
      "tokens_seen": 1218641920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039946803480234535,
      "loss": 2.7765,
      "theoretical_loss": 3.582370740451267,
      "tokens_seen": 1218707456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003994514847739739,
      "loss": 3.0998,
      "theoretical_loss": 3.582353072084771,
      "tokens_seen": 1218772992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039943493474560237,
      "loss": 2.6634,
      "theoretical_loss": 3.582335404934317,
      "tokens_seen": 1218838528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003994183847172309,
      "loss": 2.7226,
      "theoretical_loss": 3.5823177389997554,
      "tokens_seen": 1218904064
    },
    {
      "epoch": 0.44,
      "objective/train/avg_token_score": 0.025673676282167435,
      "objective/train/avg_weight": 0.9697913527488708,
      "objective/train/docs_used": 690313,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0010156631469727,
      "objective/train/std_weight": 0.09369528293609619,
      "objective/train/theoretical_loss": 3.5823000742809374,
      "objective/train/tokens_used": 1239429600,
      "theoretical_loss": 3.5823000742809374,
      "tokens_seen": 1218969600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039940183468885945,
      "loss": 2.9494,
      "theoretical_loss": 3.5823000742809374,
      "tokens_seen": 1218969600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000399385284660488,
      "loss": 2.5091,
      "theoretical_loss": 3.5822824107777143,
      "tokens_seen": 1219035136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039936873463211647,
      "loss": 2.671,
      "theoretical_loss": 3.582264748489937,
      "tokens_seen": 1219100672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000399352184603745,
      "loss": 2.5937,
      "theoretical_loss": 3.582247087417457,
      "tokens_seen": 1219166208
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039933563457537354,
      "loss": 2.8243,
      "theoretical_loss": 3.582229427560124,
      "tokens_seen": 1219231744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003993190845470021,
      "loss": 2.9642,
      "theoretical_loss": 3.5822117689177904,
      "tokens_seen": 1219297280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039930253451863056,
      "loss": 2.9325,
      "theoretical_loss": 3.582194111490307,
      "tokens_seen": 1219362816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003992859844902591,
      "loss": 2.5995,
      "theoretical_loss": 3.582176455277524,
      "tokens_seen": 1219428352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039926943446188764,
      "loss": 2.598,
      "theoretical_loss": 3.582158800279294,
      "tokens_seen": 1219493888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003992528844335162,
      "loss": 2.7874,
      "theoretical_loss": 3.5821411464954673,
      "tokens_seen": 1219559424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039923633440514466,
      "loss": 2.699,
      "theoretical_loss": 3.5821234939258955,
      "tokens_seen": 1219624960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003992197843767732,
      "loss": 2.8547,
      "theoretical_loss": 3.5821058425704293,
      "tokens_seen": 1219690496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039920323434840173,
      "loss": 2.5158,
      "theoretical_loss": 3.5820881924289205,
      "tokens_seen": 1219756032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039918668432003027,
      "loss": 2.8125,
      "theoretical_loss": 3.5820705435012203,
      "tokens_seen": 1219821568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039917013429165875,
      "loss": 2.5548,
      "theoretical_loss": 3.58205289578718,
      "tokens_seen": 1219887104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003991535842632873,
      "loss": 2.7294,
      "theoretical_loss": 3.5820352492866507,
      "tokens_seen": 1219952640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039913703423491583,
      "loss": 2.6148,
      "theoretical_loss": 3.5820176039994838,
      "tokens_seen": 1220018176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039912048420654437,
      "loss": 2.6246,
      "theoretical_loss": 3.5819999599255317,
      "tokens_seen": 1220083712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039910393417817285,
      "loss": 2.8253,
      "theoretical_loss": 3.581982317064644,
      "tokens_seen": 1220149248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003990873841498014,
      "loss": 2.5442,
      "theoretical_loss": 3.581964675416674,
      "tokens_seen": 1220214784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003990708341214299,
      "loss": 2.8707,
      "theoretical_loss": 3.5819470349814715,
      "tokens_seen": 1220280320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039905428409305846,
      "loss": 2.8555,
      "theoretical_loss": 3.5819293957588894,
      "tokens_seen": 1220345856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039903773406468695,
      "loss": 2.803,
      "theoretical_loss": 3.5819117577487782,
      "tokens_seen": 1220411392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003990211840363155,
      "loss": 2.6759,
      "theoretical_loss": 3.5818941209509902,
      "tokens_seen": 1220476928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000399004634007944,
      "loss": 2.8513,
      "theoretical_loss": 3.5818764853653766,
      "tokens_seen": 1220542464
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039898808397957256,
      "loss": 2.72,
      "theoretical_loss": 3.581858850991789,
      "tokens_seen": 1220608000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003989715339512011,
      "loss": 2.7225,
      "theoretical_loss": 3.581841217830079,
      "tokens_seen": 1220673536
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003989549839228296,
      "loss": 2.8348,
      "theoretical_loss": 3.581823585880099,
      "tokens_seen": 1220739072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003989384338944581,
      "loss": 2.6098,
      "theoretical_loss": 3.5818059551416996,
      "tokens_seen": 1220804608
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039892188386608666,
      "loss": 2.8008,
      "theoretical_loss": 3.581788325614733,
      "tokens_seen": 1220870144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003989053338377152,
      "loss": 2.397,
      "theoretical_loss": 3.581770697299051,
      "tokens_seen": 1220935680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003988887838093437,
      "loss": 2.2688,
      "theoretical_loss": 3.5817530701945053,
      "tokens_seen": 1221001216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003988722337809722,
      "loss": 2.7334,
      "theoretical_loss": 3.5817354443009473,
      "tokens_seen": 1221066752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039885568375260075,
      "loss": 2.5952,
      "theoretical_loss": 3.5817178196182295,
      "tokens_seen": 1221132288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003988391337242293,
      "loss": 2.8569,
      "theoretical_loss": 3.581700196146204,
      "tokens_seen": 1221197824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039882258369585777,
      "loss": 2.9381,
      "theoretical_loss": 3.5816825738847213,
      "tokens_seen": 1221263360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003988060336674863,
      "loss": 2.7688,
      "theoretical_loss": 3.581664952833634,
      "tokens_seen": 1221328896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039878948363911485,
      "loss": 2.7527,
      "theoretical_loss": 3.581647332992795,
      "tokens_seen": 1221394432
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003987729336107434,
      "loss": 3.0289,
      "theoretical_loss": 3.5816297143620544,
      "tokens_seen": 1221459968
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039875638358237187,
      "loss": 2.721,
      "theoretical_loss": 3.581612096941266,
      "tokens_seen": 1221525504
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003987398335540004,
      "loss": 2.7835,
      "theoretical_loss": 3.5815944807302804,
      "tokens_seen": 1221591040
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039872328352562894,
      "loss": 2.667,
      "theoretical_loss": 3.58157686572895,
      "tokens_seen": 1221656576
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003987067334972575,
      "loss": 2.8455,
      "theoretical_loss": 3.5815592519371275,
      "tokens_seen": 1221722112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039869018346888596,
      "loss": 2.3222,
      "theoretical_loss": 3.581541639354664,
      "tokens_seen": 1221787648
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003986736334405145,
      "loss": 2.8655,
      "theoretical_loss": 3.5815240279814127,
      "tokens_seen": 1221853184
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039865708341214293,
      "loss": 2.7467,
      "theoretical_loss": 3.5815064178172245,
      "tokens_seen": 1221918720
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039864053338377147,
      "loss": 2.5499,
      "theoretical_loss": 3.5814888088619528,
      "tokens_seen": 1221984256
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003986239833554,
      "loss": 2.603,
      "theoretical_loss": 3.5814712011154484,
      "tokens_seen": 1222049792
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039860743332702854,
      "loss": 2.8338,
      "theoretical_loss": 3.5814535945775647,
      "tokens_seen": 1222115328
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000398590883298657,
      "loss": 2.8438,
      "theoretical_loss": 3.5814359892481535,
      "tokens_seen": 1222180864
    },
    {
      "epoch": 0.44,
      "objective/train/avg_token_score": 0.011004425585269928,
      "objective/train/avg_weight": 0.9870535135269165,
      "objective/train/docs_used": 692066,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9164443016052246,
      "objective/train/std_weight": 0.05700500309467316,
      "objective/train/theoretical_loss": 3.5814183851270673,
      "objective/train/tokens_used": 1242706400,
      "theoretical_loss": 3.5814183851270673,
      "tokens_seen": 1222246400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039857433327028556,
      "loss": 2.8535,
      "theoretical_loss": 3.5814183851270673,
      "tokens_seen": 1222246400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003985577832419141,
      "loss": 2.6617,
      "theoretical_loss": 3.581400782214158,
      "tokens_seen": 1222311936
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039854123321354264,
      "loss": 2.8643,
      "theoretical_loss": 3.5813831805092775,
      "tokens_seen": 1222377472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003985246831851711,
      "loss": 2.6546,
      "theoretical_loss": 3.581365580012279,
      "tokens_seen": 1222443008
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039850813315679966,
      "loss": 2.8831,
      "theoretical_loss": 3.581347980723015,
      "tokens_seen": 1222508544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003984915831284282,
      "loss": 2.8435,
      "theoretical_loss": 3.5813303826413367,
      "tokens_seen": 1222574080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039847503310005674,
      "loss": 2.8219,
      "theoretical_loss": 3.5813127857670977,
      "tokens_seen": 1222639616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003984584830716852,
      "loss": 2.6589,
      "theoretical_loss": 3.5812951901001497,
      "tokens_seen": 1222705152
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039844193304331376,
      "loss": 2.8032,
      "theoretical_loss": 3.581277595640346,
      "tokens_seen": 1222770688
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003984253830149423,
      "loss": 2.8018,
      "theoretical_loss": 3.581260002387538,
      "tokens_seen": 1222836224
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039840883298657083,
      "loss": 2.9355,
      "theoretical_loss": 3.581242410341579,
      "tokens_seen": 1222901760
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003983922829581993,
      "loss": 2.815,
      "theoretical_loss": 3.5812248195023213,
      "tokens_seen": 1222967296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039837573292982785,
      "loss": 2.6167,
      "theoretical_loss": 3.5812072298696176,
      "tokens_seen": 1223032832
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003983591829014564,
      "loss": 2.8387,
      "theoretical_loss": 3.58118964144332,
      "tokens_seen": 1223098368
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039834263287308493,
      "loss": 2.7735,
      "theoretical_loss": 3.581172054223282,
      "tokens_seen": 1223163904
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003983260828447134,
      "loss": 2.7421,
      "theoretical_loss": 3.5811544682093555,
      "tokens_seen": 1223229440
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039830953281634195,
      "loss": 2.5184,
      "theoretical_loss": 3.5811368834013937,
      "tokens_seen": 1223294976
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003982929827879705,
      "loss": 2.79,
      "theoretical_loss": 3.5811192997992487,
      "tokens_seen": 1223360512
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000398276432759599,
      "loss": 2.6412,
      "theoretical_loss": 3.581101717402773,
      "tokens_seen": 1223426048
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003982598827312275,
      "loss": 2.5785,
      "theoretical_loss": 3.581084136211821,
      "tokens_seen": 1223491584
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039824333270285604,
      "loss": 3.1037,
      "theoretical_loss": 3.5810665562262436,
      "tokens_seen": 1223557120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003982267826744846,
      "loss": 2.6286,
      "theoretical_loss": 3.581048977445895,
      "tokens_seen": 1223622656
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003982102326461131,
      "loss": 2.685,
      "theoretical_loss": 3.581031399870627,
      "tokens_seen": 1223688192
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003981936826177416,
      "loss": 3.0489,
      "theoretical_loss": 3.5810138235002933,
      "tokens_seen": 1223753728
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039817713258937014,
      "loss": 2.7105,
      "theoretical_loss": 3.580996248334746,
      "tokens_seen": 1223819264
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003981605825609987,
      "loss": 2.6651,
      "theoretical_loss": 3.580978674373838,
      "tokens_seen": 1223884800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003981440325326272,
      "loss": 2.4055,
      "theoretical_loss": 3.580961101617423,
      "tokens_seen": 1223950336
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003981274825042557,
      "loss": 2.5976,
      "theoretical_loss": 3.580943530065354,
      "tokens_seen": 1224015872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039811093247588424,
      "loss": 2.5133,
      "theoretical_loss": 3.580925959717483,
      "tokens_seen": 1224081408
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003980943824475128,
      "loss": 2.5121,
      "theoretical_loss": 3.580908390573663,
      "tokens_seen": 1224146944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003980778324191413,
      "loss": 2.7275,
      "theoretical_loss": 3.580890822633749,
      "tokens_seen": 1224212480
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003980612823907698,
      "loss": 2.7707,
      "theoretical_loss": 3.5808732558975915,
      "tokens_seen": 1224278016
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039804473236239833,
      "loss": 2.7636,
      "theoretical_loss": 3.580855690365045,
      "tokens_seen": 1224343552
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039802818233402687,
      "loss": 2.4531,
      "theoretical_loss": 3.580838126035962,
      "tokens_seen": 1224409088
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003980116323056554,
      "loss": 2.8238,
      "theoretical_loss": 3.580820562910197,
      "tokens_seen": 1224474624
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003979950822772839,
      "loss": 2.9529,
      "theoretical_loss": 3.5808030009876015,
      "tokens_seen": 1224540160
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039797853224891243,
      "loss": 2.6047,
      "theoretical_loss": 3.580785440268029,
      "tokens_seen": 1224605696
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039796198222054097,
      "loss": 2.774,
      "theoretical_loss": 3.5807678807513335,
      "tokens_seen": 1224671232
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003979454321921695,
      "loss": 2.7892,
      "theoretical_loss": 3.5807503224373676,
      "tokens_seen": 1224736768
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000397928882163798,
      "loss": 2.7185,
      "theoretical_loss": 3.580732765325985,
      "tokens_seen": 1224802304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003979123321354265,
      "loss": 3.0892,
      "theoretical_loss": 3.5807152094170385,
      "tokens_seen": 1224867840
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039789578210705506,
      "loss": 2.8147,
      "theoretical_loss": 3.5806976547103817,
      "tokens_seen": 1224933376
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003978792320786836,
      "loss": 2.8195,
      "theoretical_loss": 3.5806801012058678,
      "tokens_seen": 1224998912
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003978626820503121,
      "loss": 2.7048,
      "theoretical_loss": 3.5806625489033506,
      "tokens_seen": 1225064448
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003978461320219406,
      "loss": 2.6895,
      "theoretical_loss": 3.580644997802683,
      "tokens_seen": 1225129984
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039782958199356916,
      "loss": 2.8113,
      "theoretical_loss": 3.580627447903719,
      "tokens_seen": 1225195520
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003978130319651977,
      "loss": 3.0085,
      "theoretical_loss": 3.580609899206311,
      "tokens_seen": 1225261056
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003977964819368262,
      "loss": 2.6854,
      "theoretical_loss": 3.5805923517103135,
      "tokens_seen": 1225326592
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003977799319084547,
      "loss": 2.5218,
      "theoretical_loss": 3.58057480541558,
      "tokens_seen": 1225392128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039776338188008325,
      "loss": 2.5533,
      "theoretical_loss": 3.580557260321963,
      "tokens_seen": 1225457664
    },
    {
      "epoch": 0.44,
      "objective/train/avg_token_score": 0.008024219423532486,
      "objective/train/avg_weight": 0.9905560612678528,
      "objective/train/docs_used": 694202,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5069117546081543,
      "objective/train/std_weight": 0.05336223542690277,
      "objective/train/theoretical_loss": 3.5805397164293167,
      "objective/train/tokens_used": 1245983200,
      "theoretical_loss": 3.5805397164293167,
      "tokens_seen": 1225523200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003977468318517118,
      "loss": 2.5998,
      "theoretical_loss": 3.5805397164293167,
      "tokens_seen": 1225523200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003977302818233403,
      "loss": 2.9102,
      "theoretical_loss": 3.580522173737495,
      "tokens_seen": 1225588736
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003977137317949688,
      "loss": 2.8776,
      "theoretical_loss": 3.580504632246351,
      "tokens_seen": 1225654272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039769718176659735,
      "loss": 2.6563,
      "theoretical_loss": 3.5804870919557388,
      "tokens_seen": 1225719808
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003976806317382258,
      "loss": 2.785,
      "theoretical_loss": 3.5804695528655115,
      "tokens_seen": 1225785344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003976640817098543,
      "loss": 2.8208,
      "theoretical_loss": 3.5804520149755232,
      "tokens_seen": 1225850880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039764753168148285,
      "loss": 2.8063,
      "theoretical_loss": 3.5804344782856274,
      "tokens_seen": 1225916416
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003976309816531114,
      "loss": 2.6245,
      "theoretical_loss": 3.580416942795678,
      "tokens_seen": 1225981952
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003976144316247399,
      "loss": 2.7566,
      "theoretical_loss": 3.5803994085055284,
      "tokens_seen": 1226047488
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003975978815963684,
      "loss": 2.6966,
      "theoretical_loss": 3.580381875415033,
      "tokens_seen": 1226113024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039758133156799695,
      "loss": 2.5811,
      "theoretical_loss": 3.5803643435240446,
      "tokens_seen": 1226178560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003975647815396255,
      "loss": 2.4642,
      "theoretical_loss": 3.580346812832418,
      "tokens_seen": 1226244096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039754823151125397,
      "loss": 2.6057,
      "theoretical_loss": 3.5803292833400064,
      "tokens_seen": 1226309632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003975316814828825,
      "loss": 2.6852,
      "theoretical_loss": 3.5803117550466643,
      "tokens_seen": 1226375168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039751513145451105,
      "loss": 2.8449,
      "theoretical_loss": 3.580294227952246,
      "tokens_seen": 1226440704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003974985814261396,
      "loss": 2.662,
      "theoretical_loss": 3.5802767020566035,
      "tokens_seen": 1226506240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039748203139776807,
      "loss": 2.7792,
      "theoretical_loss": 3.5802591773595926,
      "tokens_seen": 1226571776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003974654813693966,
      "loss": 2.5132,
      "theoretical_loss": 3.5802416538610666,
      "tokens_seen": 1226637312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039744893134102514,
      "loss": 2.7664,
      "theoretical_loss": 3.58022413156088,
      "tokens_seen": 1226702848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003974323813126537,
      "loss": 2.6446,
      "theoretical_loss": 3.580206610458886,
      "tokens_seen": 1226768384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039741583128428216,
      "loss": 2.7807,
      "theoretical_loss": 3.5801890905549394,
      "tokens_seen": 1226833920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003973992812559107,
      "loss": 2.7723,
      "theoretical_loss": 3.5801715718488936,
      "tokens_seen": 1226899456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039738273122753924,
      "loss": 2.6561,
      "theoretical_loss": 3.5801540543406034,
      "tokens_seen": 1226964992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003973661811991678,
      "loss": 2.6701,
      "theoretical_loss": 3.580136538029923,
      "tokens_seen": 1227030528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039734963117079626,
      "loss": 2.6498,
      "theoretical_loss": 3.5801190229167057,
      "tokens_seen": 1227096064
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003973330811424248,
      "loss": 2.6717,
      "theoretical_loss": 3.580101509000806,
      "tokens_seen": 1227161600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039731653111405333,
      "loss": 2.9407,
      "theoretical_loss": 3.5800839962820787,
      "tokens_seen": 1227227136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039729998108568187,
      "loss": 2.6313,
      "theoretical_loss": 3.580066484760378,
      "tokens_seen": 1227292672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039728343105731035,
      "loss": 2.7771,
      "theoretical_loss": 3.5800489744355573,
      "tokens_seen": 1227358208
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003972668810289389,
      "loss": 2.9394,
      "theoretical_loss": 3.5800314653074716,
      "tokens_seen": 1227423744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039725033100056743,
      "loss": 2.5501,
      "theoretical_loss": 3.580013957375975,
      "tokens_seen": 1227489280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039723378097219597,
      "loss": 2.8196,
      "theoretical_loss": 3.5799964506409214,
      "tokens_seen": 1227554816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039721723094382445,
      "loss": 2.6582,
      "theoretical_loss": 3.5799789451021664,
      "tokens_seen": 1227620352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000397200680915453,
      "loss": 2.6163,
      "theoretical_loss": 3.579961440759564,
      "tokens_seen": 1227685888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003971841308870815,
      "loss": 2.5649,
      "theoretical_loss": 3.579943937612967,
      "tokens_seen": 1227751424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039716758085871006,
      "loss": 2.6379,
      "theoretical_loss": 3.579926435662232,
      "tokens_seen": 1227816960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039715103083033855,
      "loss": 2.4319,
      "theoretical_loss": 3.5799089349072126,
      "tokens_seen": 1227882496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003971344808019671,
      "loss": 2.7285,
      "theoretical_loss": 3.579891435347763,
      "tokens_seen": 1227948032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003971179307735956,
      "loss": 2.566,
      "theoretical_loss": 3.5798739369837382,
      "tokens_seen": 1228013568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039710138074522416,
      "loss": 2.9056,
      "theoretical_loss": 3.5798564398149924,
      "tokens_seen": 1228079104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039708483071685264,
      "loss": 2.7024,
      "theoretical_loss": 3.5798389438413807,
      "tokens_seen": 1228144640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003970682806884812,
      "loss": 2.7044,
      "theoretical_loss": 3.5798214490627567,
      "tokens_seen": 1228210176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003970517306601097,
      "loss": 2.9401,
      "theoretical_loss": 3.5798039554789765,
      "tokens_seen": 1228275712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039703518063173825,
      "loss": 2.5606,
      "theoretical_loss": 3.579786463089893,
      "tokens_seen": 1228341248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039701863060336674,
      "loss": 2.735,
      "theoretical_loss": 3.5797689718953625,
      "tokens_seen": 1228406784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003970020805749953,
      "loss": 2.6718,
      "theoretical_loss": 3.5797514818952383,
      "tokens_seen": 1228472320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003969855305466238,
      "loss": 2.6692,
      "theoretical_loss": 3.579733993089377,
      "tokens_seen": 1228537856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039696898051825235,
      "loss": 2.4889,
      "theoretical_loss": 3.579716505477631,
      "tokens_seen": 1228603392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039695243048988083,
      "loss": 2.6606,
      "theoretical_loss": 3.5796990190598565,
      "tokens_seen": 1228668928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039693588046150937,
      "loss": 2.6113,
      "theoretical_loss": 3.5796815338359087,
      "tokens_seen": 1228734464
    },
    {
      "epoch": 0.44,
      "objective/train/avg_token_score": 0.02064388617873192,
      "objective/train/avg_weight": 0.9757285714149475,
      "objective/train/docs_used": 696236,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.777797222137451,
      "objective/train/std_weight": 0.10844160616397858,
      "objective/train/theoretical_loss": 3.5796640498056407,
      "objective/train/tokens_used": 1249260000,
      "theoretical_loss": 3.5796640498056407,
      "tokens_seen": 1228800000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003969193304331379,
      "loss": 2.7487,
      "theoretical_loss": 3.5796640498056407,
      "tokens_seen": 1228800000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039690278040476645,
      "loss": 2.7633,
      "theoretical_loss": 3.5796465669689095,
      "tokens_seen": 1228865536
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039688623037639493,
      "loss": 2.7818,
      "theoretical_loss": 3.579629085325568,
      "tokens_seen": 1228931072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039686968034802347,
      "loss": 2.7219,
      "theoretical_loss": 3.5796116048754727,
      "tokens_seen": 1228996608
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000396853130319652,
      "loss": 2.572,
      "theoretical_loss": 3.5795941256184776,
      "tokens_seen": 1229062144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039683658029128054,
      "loss": 2.4969,
      "theoretical_loss": 3.579576647554438,
      "tokens_seen": 1229127680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000396820030262909,
      "loss": 2.6811,
      "theoretical_loss": 3.5795591706832086,
      "tokens_seen": 1229193216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039680348023453756,
      "loss": 2.828,
      "theoretical_loss": 3.5795416950046444,
      "tokens_seen": 1229258752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003967869302061661,
      "loss": 2.7849,
      "theoretical_loss": 3.579524220518601,
      "tokens_seen": 1229324288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039677038017779464,
      "loss": 2.7954,
      "theoretical_loss": 3.579506747224933,
      "tokens_seen": 1229389824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003967538301494231,
      "loss": 2.753,
      "theoretical_loss": 3.5794892751234952,
      "tokens_seen": 1229455360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039673728012105166,
      "loss": 2.5771,
      "theoretical_loss": 3.579471804214143,
      "tokens_seen": 1229520896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003967207300926802,
      "loss": 2.7966,
      "theoretical_loss": 3.579454334496732,
      "tokens_seen": 1229586432
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039670418006430873,
      "loss": 2.7716,
      "theoretical_loss": 3.579436865971117,
      "tokens_seen": 1229651968
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039668763003593716,
      "loss": 2.8697,
      "theoretical_loss": 3.5794193986371527,
      "tokens_seen": 1229717504
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003966710800075657,
      "loss": 2.7583,
      "theoretical_loss": 3.579401932494695,
      "tokens_seen": 1229783040
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003966545299791942,
      "loss": 2.751,
      "theoretical_loss": 3.579384467543599,
      "tokens_seen": 1229848576
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003966379799508227,
      "loss": 3.0012,
      "theoretical_loss": 3.5793670037837195,
      "tokens_seen": 1229914112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039662142992245126,
      "loss": 2.5697,
      "theoretical_loss": 3.5793495412149126,
      "tokens_seen": 1229979648
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003966048798940798,
      "loss": 2.7629,
      "theoretical_loss": 3.579332079837033,
      "tokens_seen": 1230045184
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003965883298657083,
      "loss": 2.8131,
      "theoretical_loss": 3.579314619649936,
      "tokens_seen": 1230110720
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003965717798373368,
      "loss": 2.8194,
      "theoretical_loss": 3.5792971606534767,
      "tokens_seen": 1230176256
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039655522980896536,
      "loss": 2.7626,
      "theoretical_loss": 3.5792797028475114,
      "tokens_seen": 1230241792
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003965386797805939,
      "loss": 2.6637,
      "theoretical_loss": 3.579262246231895,
      "tokens_seen": 1230307328
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003965221297522224,
      "loss": 2.6485,
      "theoretical_loss": 3.5792447908064826,
      "tokens_seen": 1230372864
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003965055797238509,
      "loss": 2.6254,
      "theoretical_loss": 3.5792273365711305,
      "tokens_seen": 1230438400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039648902969547945,
      "loss": 2.8063,
      "theoretical_loss": 3.5792098835256936,
      "tokens_seen": 1230503936
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000396472479667108,
      "loss": 2.7596,
      "theoretical_loss": 3.5791924316700277,
      "tokens_seen": 1230569472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003964559296387365,
      "loss": 2.9445,
      "theoretical_loss": 3.5791749810039875,
      "tokens_seen": 1230635008
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000396439379610365,
      "loss": 2.7063,
      "theoretical_loss": 3.5791575315274295,
      "tokens_seen": 1230700544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039642282958199355,
      "loss": 2.7241,
      "theoretical_loss": 3.579140083240209,
      "tokens_seen": 1230766080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003964062795536221,
      "loss": 2.6482,
      "theoretical_loss": 3.5791226361421815,
      "tokens_seen": 1230831616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003963897295252506,
      "loss": 2.7489,
      "theoretical_loss": 3.579105190233203,
      "tokens_seen": 1230897152
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003963731794968791,
      "loss": 2.6753,
      "theoretical_loss": 3.579087745513129,
      "tokens_seen": 1230962688
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039635662946850764,
      "loss": 2.9446,
      "theoretical_loss": 3.579070301981815,
      "tokens_seen": 1231028224
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003963400794401362,
      "loss": 2.5143,
      "theoretical_loss": 3.5790528596391167,
      "tokens_seen": 1231093760
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003963235294117647,
      "loss": 2.6904,
      "theoretical_loss": 3.5790354184848896,
      "tokens_seen": 1231159296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003963069793833932,
      "loss": 2.6646,
      "theoretical_loss": 3.5790179785189906,
      "tokens_seen": 1231224832
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039629042935502174,
      "loss": 2.7035,
      "theoretical_loss": 3.579000539741274,
      "tokens_seen": 1231290368
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003962738793266503,
      "loss": 2.5315,
      "theoretical_loss": 3.5789831021515965,
      "tokens_seen": 1231355904
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003962573292982788,
      "loss": 2.7177,
      "theoretical_loss": 3.5789656657498137,
      "tokens_seen": 1231421440
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003962407792699073,
      "loss": 2.6047,
      "theoretical_loss": 3.578948230535782,
      "tokens_seen": 1231486976
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039622422924153584,
      "loss": 2.7656,
      "theoretical_loss": 3.5789307965093564,
      "tokens_seen": 1231552512
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003962076792131644,
      "loss": 2.7352,
      "theoretical_loss": 3.5789133636703934,
      "tokens_seen": 1231618048
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003961911291847929,
      "loss": 2.6206,
      "theoretical_loss": 3.5788959320187486,
      "tokens_seen": 1231683584
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003961745791564214,
      "loss": 2.6275,
      "theoretical_loss": 3.5788785015542786,
      "tokens_seen": 1231749120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039615802912804993,
      "loss": 2.6429,
      "theoretical_loss": 3.578861072276838,
      "tokens_seen": 1231814656
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039614147909967847,
      "loss": 2.67,
      "theoretical_loss": 3.5788436441862843,
      "tokens_seen": 1231880192
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000396124929071307,
      "loss": 2.489,
      "theoretical_loss": 3.578826217282473,
      "tokens_seen": 1231945728
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003961083790429355,
      "loss": 2.642,
      "theoretical_loss": 3.57880879156526,
      "tokens_seen": 1232011264
    },
    {
      "epoch": 0.44,
      "objective/train/avg_token_score": 0.02417842298746109,
      "objective/train/avg_weight": 0.9715310335159302,
      "objective/train/docs_used": 697487,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1986165046691895,
      "objective/train/std_weight": 0.12792585790157318,
      "objective/train/theoretical_loss": 3.5787913670345013,
      "objective/train/tokens_used": 1252536800,
      "theoretical_loss": 3.5787913670345013,
      "tokens_seen": 1232076800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039609182901456403,
      "loss": 2.7494,
      "theoretical_loss": 3.5787913670345013,
      "tokens_seen": 1232076800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039607527898619256,
      "loss": 2.7754,
      "theoretical_loss": 3.578773943690054,
      "tokens_seen": 1232142336
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003960587289578211,
      "loss": 2.8816,
      "theoretical_loss": 3.5787565215317727,
      "tokens_seen": 1232207872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003960421789294496,
      "loss": 2.7271,
      "theoretical_loss": 3.578739100559515,
      "tokens_seen": 1232273408
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003960256289010781,
      "loss": 2.5553,
      "theoretical_loss": 3.578721680773136,
      "tokens_seen": 1232338944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039600907887270666,
      "loss": 2.6351,
      "theoretical_loss": 3.5787042621724927,
      "tokens_seen": 1232404480
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003959925288443352,
      "loss": 2.738,
      "theoretical_loss": 3.5786868447574403,
      "tokens_seen": 1232470016
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003959759788159637,
      "loss": 2.5616,
      "theoretical_loss": 3.5786694285278364,
      "tokens_seen": 1232535552
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003959594287875922,
      "loss": 2.5546,
      "theoretical_loss": 3.578652013483537,
      "tokens_seen": 1232601088
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039594287875922076,
      "loss": 2.886,
      "theoretical_loss": 3.5786345996243973,
      "tokens_seen": 1232666624
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003959263287308493,
      "loss": 2.7273,
      "theoretical_loss": 3.5786171869502743,
      "tokens_seen": 1232732160
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003959097787024778,
      "loss": 2.4139,
      "theoretical_loss": 3.578599775461025,
      "tokens_seen": 1232797696
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003958932286741063,
      "loss": 2.7279,
      "theoretical_loss": 3.578582365156505,
      "tokens_seen": 1232863232
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039587667864573485,
      "loss": 2.6726,
      "theoretical_loss": 3.5785649560365718,
      "tokens_seen": 1232928768
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003958601286173634,
      "loss": 2.539,
      "theoretical_loss": 3.5785475481010796,
      "tokens_seen": 1232994304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003958435785889919,
      "loss": 2.707,
      "theoretical_loss": 3.5785301413498876,
      "tokens_seen": 1233059840
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003958270285606204,
      "loss": 2.628,
      "theoretical_loss": 3.57851273578285,
      "tokens_seen": 1233125376
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039581047853224895,
      "loss": 2.7781,
      "theoretical_loss": 3.578495331399825,
      "tokens_seen": 1233190912
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003957939285038775,
      "loss": 2.6943,
      "theoretical_loss": 3.5784779282006682,
      "tokens_seen": 1233256448
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039577737847550597,
      "loss": 2.7408,
      "theoretical_loss": 3.5784605261852365,
      "tokens_seen": 1233321984
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003957608284471345,
      "loss": 2.6587,
      "theoretical_loss": 3.5784431253533864,
      "tokens_seen": 1233387520
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039574427841876304,
      "loss": 2.7057,
      "theoretical_loss": 3.5784257257049745,
      "tokens_seen": 1233453056
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003957277283903916,
      "loss": 2.6507,
      "theoretical_loss": 3.578408327239857,
      "tokens_seen": 1233518592
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039571117836202,
      "loss": 2.7308,
      "theoretical_loss": 3.5783909299578918,
      "tokens_seen": 1233584128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039569462833364855,
      "loss": 2.8443,
      "theoretical_loss": 3.578373533858935,
      "tokens_seen": 1233649664
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039567807830527703,
      "loss": 2.5404,
      "theoretical_loss": 3.578356138942842,
      "tokens_seen": 1233715200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039566152827690557,
      "loss": 2.6257,
      "theoretical_loss": 3.5783387452094715,
      "tokens_seen": 1233780736
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003956449782485341,
      "loss": 2.5185,
      "theoretical_loss": 3.578321352658679,
      "tokens_seen": 1233846272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039562842822016264,
      "loss": 2.8802,
      "theoretical_loss": 3.578303961290322,
      "tokens_seen": 1233911808
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039561187819179113,
      "loss": 2.411,
      "theoretical_loss": 3.5782865711042575,
      "tokens_seen": 1233977344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039559532816341967,
      "loss": 2.7292,
      "theoretical_loss": 3.5782691821003407,
      "tokens_seen": 1234042880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003955787781350482,
      "loss": 2.5251,
      "theoretical_loss": 3.578251794278431,
      "tokens_seen": 1234108416
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039556222810667674,
      "loss": 2.7702,
      "theoretical_loss": 3.578234407638383,
      "tokens_seen": 1234173952
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003955456780783052,
      "loss": 2.7797,
      "theoretical_loss": 3.5782170221800556,
      "tokens_seen": 1234239488
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039552912804993376,
      "loss": 2.4887,
      "theoretical_loss": 3.5781996379033036,
      "tokens_seen": 1234305024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003955125780215623,
      "loss": 2.8556,
      "theoretical_loss": 3.5781822548079854,
      "tokens_seen": 1234370560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039549602799319084,
      "loss": 2.4966,
      "theoretical_loss": 3.578164872893958,
      "tokens_seen": 1234436096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003954794779648193,
      "loss": 2.8138,
      "theoretical_loss": 3.5781474921610776,
      "tokens_seen": 1234501632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039546292793644786,
      "loss": 2.7654,
      "theoretical_loss": 3.578130112609202,
      "tokens_seen": 1234567168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003954463779080764,
      "loss": 2.8081,
      "theoretical_loss": 3.5781127342381884,
      "tokens_seen": 1234632704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039542982787970493,
      "loss": 2.4296,
      "theoretical_loss": 3.578095357047893,
      "tokens_seen": 1234698240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003954132778513334,
      "loss": 2.7207,
      "theoretical_loss": 3.5780779810381733,
      "tokens_seen": 1234763776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039539672782296195,
      "loss": 2.5418,
      "theoretical_loss": 3.578060606208887,
      "tokens_seen": 1234829312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003953801777945905,
      "loss": 2.525,
      "theoretical_loss": 3.5780432325598905,
      "tokens_seen": 1234894848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039536362776621903,
      "loss": 2.8602,
      "theoretical_loss": 3.5780258600910413,
      "tokens_seen": 1234960384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003953470777378475,
      "loss": 2.8053,
      "theoretical_loss": 3.5780084888021966,
      "tokens_seen": 1235025920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039533052770947605,
      "loss": 2.8372,
      "theoretical_loss": 3.5779911186932134,
      "tokens_seen": 1235091456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003953139776811046,
      "loss": 2.6973,
      "theoretical_loss": 3.5779737497639497,
      "tokens_seen": 1235156992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003952974276527331,
      "loss": 2.5481,
      "theoretical_loss": 3.5779563820142624,
      "tokens_seen": 1235222528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003952808776243616,
      "loss": 2.629,
      "theoretical_loss": 3.577939015444008,
      "tokens_seen": 1235288064
    },
    {
      "epoch": 0.44,
      "objective/train/avg_token_score": 0.017072327435016632,
      "objective/train/avg_weight": 0.9799002408981323,
      "objective/train/docs_used": 699233,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.246771812438965,
      "objective/train/std_weight": 0.1003691703081131,
      "objective/train/theoretical_loss": 3.577921650053045,
      "objective/train/tokens_used": 1255813600,
      "theoretical_loss": 3.577921650053045,
      "tokens_seen": 1235353600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039526432759599015,
      "loss": 2.731,
      "theoretical_loss": 3.577921650053045,
      "tokens_seen": 1235353600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003952477775676187,
      "loss": 2.9427,
      "theoretical_loss": 3.57790428584123,
      "tokens_seen": 1235419136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003952312275392472,
      "loss": 2.5822,
      "theoretical_loss": 3.577886922808421,
      "tokens_seen": 1235484672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003952146775108757,
      "loss": 2.6455,
      "theoretical_loss": 3.5778695609544755,
      "tokens_seen": 1235550208
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039519812748250424,
      "loss": 2.496,
      "theoretical_loss": 3.57785220027925,
      "tokens_seen": 1235615744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003951815774541328,
      "loss": 2.6943,
      "theoretical_loss": 3.577834840782603,
      "tokens_seen": 1235681280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003951650274257613,
      "loss": 2.5978,
      "theoretical_loss": 3.5778174824643907,
      "tokens_seen": 1235746816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003951484773973898,
      "loss": 2.5177,
      "theoretical_loss": 3.577800125324472,
      "tokens_seen": 1235812352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039513192736901834,
      "loss": 2.5828,
      "theoretical_loss": 3.577782769362704,
      "tokens_seen": 1235877888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003951153773406469,
      "loss": 2.6238,
      "theoretical_loss": 3.577765414578944,
      "tokens_seen": 1235943424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003950988273122754,
      "loss": 2.7306,
      "theoretical_loss": 3.5777480609730494,
      "tokens_seen": 1236008960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003950822772839039,
      "loss": 2.7373,
      "theoretical_loss": 3.577730708544878,
      "tokens_seen": 1236074496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039506572725553243,
      "loss": 2.9308,
      "theoretical_loss": 3.577713357294288,
      "tokens_seen": 1236140032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039504917722716097,
      "loss": 2.4719,
      "theoretical_loss": 3.5776960072211366,
      "tokens_seen": 1236205568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003950326271987895,
      "loss": 2.6892,
      "theoretical_loss": 3.577678658325281,
      "tokens_seen": 1236271104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000395016077170418,
      "loss": 2.9052,
      "theoretical_loss": 3.5776613106065804,
      "tokens_seen": 1236336640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039499952714204653,
      "loss": 2.2913,
      "theoretical_loss": 3.5776439640648907,
      "tokens_seen": 1236402176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039498297711367507,
      "loss": 2.608,
      "theoretical_loss": 3.5776266187000707,
      "tokens_seen": 1236467712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003949664270853036,
      "loss": 2.5109,
      "theoretical_loss": 3.5776092745119783,
      "tokens_seen": 1236533248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039494987705693214,
      "loss": 2.8597,
      "theoretical_loss": 3.5775919315004705,
      "tokens_seen": 1236598784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003949333270285606,
      "loss": 2.6402,
      "theoretical_loss": 3.5775745896654056,
      "tokens_seen": 1236664320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039491677700018916,
      "loss": 2.6072,
      "theoretical_loss": 3.5775572490066425,
      "tokens_seen": 1236729856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003949002269718177,
      "loss": 2.625,
      "theoretical_loss": 3.577539909524037,
      "tokens_seen": 1236795392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039488367694344624,
      "loss": 2.5606,
      "theoretical_loss": 3.5775225712174485,
      "tokens_seen": 1236860928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003948671269150747,
      "loss": 2.6584,
      "theoretical_loss": 3.5775052340867344,
      "tokens_seen": 1236926464
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039485057688670326,
      "loss": 2.7295,
      "theoretical_loss": 3.577487898131753,
      "tokens_seen": 1236992000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003948340268583318,
      "loss": 2.7586,
      "theoretical_loss": 3.577470563352362,
      "tokens_seen": 1237057536
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039481747682996033,
      "loss": 2.705,
      "theoretical_loss": 3.5774532297484196,
      "tokens_seen": 1237123072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003948009268015888,
      "loss": 2.6614,
      "theoretical_loss": 3.5774358973197833,
      "tokens_seen": 1237188608
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039478437677321735,
      "loss": 2.7342,
      "theoretical_loss": 3.577418566066312,
      "tokens_seen": 1237254144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003947678267448459,
      "loss": 2.8071,
      "theoretical_loss": 3.577401235987863,
      "tokens_seen": 1237319680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039475127671647443,
      "loss": 2.7156,
      "theoretical_loss": 3.577383907084295,
      "tokens_seen": 1237385216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003947347266881029,
      "loss": 2.8478,
      "theoretical_loss": 3.5773665793554654,
      "tokens_seen": 1237450752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003947181766597314,
      "loss": 2.563,
      "theoretical_loss": 3.577349252801233,
      "tokens_seen": 1237516288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003947016266313599,
      "loss": 2.6061,
      "theoretical_loss": 3.577331927421456,
      "tokens_seen": 1237581824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003946850766029884,
      "loss": 2.7035,
      "theoretical_loss": 3.5773146032159926,
      "tokens_seen": 1237647360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039466852657461696,
      "loss": 2.8022,
      "theoretical_loss": 3.5772972801847005,
      "tokens_seen": 1237712896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003946519765462455,
      "loss": 2.6083,
      "theoretical_loss": 3.5772799583274386,
      "tokens_seen": 1237778432
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000394635426517874,
      "loss": 2.7933,
      "theoretical_loss": 3.577262637644065,
      "tokens_seen": 1237843968
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003946188764895025,
      "loss": 2.374,
      "theoretical_loss": 3.5772453181344375,
      "tokens_seen": 1237909504
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039460232646113105,
      "loss": 2.7955,
      "theoretical_loss": 3.5772279997984144,
      "tokens_seen": 1237975040
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003945857764327596,
      "loss": 2.6461,
      "theoretical_loss": 3.5772106826358554,
      "tokens_seen": 1238040576
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039456922640438807,
      "loss": 2.9347,
      "theoretical_loss": 3.5771933666466174,
      "tokens_seen": 1238106112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003945526763760166,
      "loss": 2.5922,
      "theoretical_loss": 3.5771760518305595,
      "tokens_seen": 1238171648
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039453612634764515,
      "loss": 2.6844,
      "theoretical_loss": 3.57715873818754,
      "tokens_seen": 1238237184
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003945195763192737,
      "loss": 2.872,
      "theoretical_loss": 3.577141425717417,
      "tokens_seen": 1238302720
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039450302629090217,
      "loss": 2.7985,
      "theoretical_loss": 3.5771241144200494,
      "tokens_seen": 1238368256
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003944864762625307,
      "loss": 2.6854,
      "theoretical_loss": 3.577106804295296,
      "tokens_seen": 1238433792
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039446992623415924,
      "loss": 2.7462,
      "theoretical_loss": 3.5770894953430146,
      "tokens_seen": 1238499328
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003944533762057878,
      "loss": 2.7315,
      "theoretical_loss": 3.577072187563064,
      "tokens_seen": 1238564864
    },
    {
      "epoch": 0.44,
      "objective/train/avg_token_score": 0.02050260826945305,
      "objective/train/avg_weight": 0.9758704304695129,
      "objective/train/docs_used": 700968,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.690286636352539,
      "objective/train/std_weight": 0.10567745566368103,
      "objective/train/theoretical_loss": 3.577054880955303,
      "objective/train/tokens_used": 1259090400,
      "theoretical_loss": 3.577054880955303,
      "tokens_seen": 1238630400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039443682617741626,
      "loss": 2.74,
      "theoretical_loss": 3.577054880955303,
      "tokens_seen": 1238630400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003944202761490448,
      "loss": 2.5526,
      "theoretical_loss": 3.57703757551959,
      "tokens_seen": 1238695936
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039440372612067334,
      "loss": 2.7179,
      "theoretical_loss": 3.5770202712557833,
      "tokens_seen": 1238761472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003943871760923019,
      "loss": 2.5145,
      "theoretical_loss": 3.5770029681637423,
      "tokens_seen": 1238827008
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039437062606393036,
      "loss": 2.6322,
      "theoretical_loss": 3.576985666243325,
      "tokens_seen": 1238892544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003943540760355589,
      "loss": 2.6418,
      "theoretical_loss": 3.5769683654943907,
      "tokens_seen": 1238958080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039433752600718743,
      "loss": 2.6473,
      "theoretical_loss": 3.5769510659167976,
      "tokens_seen": 1239023616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039432097597881597,
      "loss": 2.5897,
      "theoretical_loss": 3.576933767510405,
      "tokens_seen": 1239089152
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039430442595044446,
      "loss": 2.7943,
      "theoretical_loss": 3.5769164702750706,
      "tokens_seen": 1239154688
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000394287875922073,
      "loss": 2.7037,
      "theoretical_loss": 3.576899174210655,
      "tokens_seen": 1239220224
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039427132589370153,
      "loss": 2.5966,
      "theoretical_loss": 3.576881879317015,
      "tokens_seen": 1239285760
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039425477586533007,
      "loss": 2.4363,
      "theoretical_loss": 3.57686458559401,
      "tokens_seen": 1239351296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039423822583695855,
      "loss": 2.7432,
      "theoretical_loss": 3.5768472930415003,
      "tokens_seen": 1239416832
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003942216758085871,
      "loss": 2.6408,
      "theoretical_loss": 3.576830001659343,
      "tokens_seen": 1239482368
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003942051257802156,
      "loss": 2.608,
      "theoretical_loss": 3.576812711447398,
      "tokens_seen": 1239547904
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039418857575184416,
      "loss": 2.6723,
      "theoretical_loss": 3.5767954224055236,
      "tokens_seen": 1239613440
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039417202572347265,
      "loss": 2.9461,
      "theoretical_loss": 3.57677813453358,
      "tokens_seen": 1239678976
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003941554756951012,
      "loss": 2.5036,
      "theoretical_loss": 3.576760847831425,
      "tokens_seen": 1239744512
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003941389256667297,
      "loss": 2.6052,
      "theoretical_loss": 3.576743562298917,
      "tokens_seen": 1239810048
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039412237563835826,
      "loss": 2.4881,
      "theoretical_loss": 3.576726277935917,
      "tokens_seen": 1239875584
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039410582560998674,
      "loss": 2.5706,
      "theoretical_loss": 3.5767089947422828,
      "tokens_seen": 1239941120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003940892755816153,
      "loss": 2.8356,
      "theoretical_loss": 3.5766917127178735,
      "tokens_seen": 1240006656
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003940727255532438,
      "loss": 2.6885,
      "theoretical_loss": 3.576674431862549,
      "tokens_seen": 1240072192
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039405617552487236,
      "loss": 2.645,
      "theoretical_loss": 3.5766571521761676,
      "tokens_seen": 1240137728
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039403962549650084,
      "loss": 2.7859,
      "theoretical_loss": 3.5766398736585883,
      "tokens_seen": 1240203264
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003940230754681294,
      "loss": 2.6952,
      "theoretical_loss": 3.5766225963096714,
      "tokens_seen": 1240268800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003940065254397579,
      "loss": 2.4959,
      "theoretical_loss": 3.576605320129275,
      "tokens_seen": 1240334336
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039398997541138645,
      "loss": 2.5723,
      "theoretical_loss": 3.576588045117259,
      "tokens_seen": 1240399872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039397342538301494,
      "loss": 2.9617,
      "theoretical_loss": 3.576570771273482,
      "tokens_seen": 1240465408
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003939568753546435,
      "loss": 2.8222,
      "theoretical_loss": 3.5765534985978045,
      "tokens_seen": 1240530944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000393940325326272,
      "loss": 2.7534,
      "theoretical_loss": 3.5765362270900845,
      "tokens_seen": 1240596480
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039392377529790055,
      "loss": 2.7181,
      "theoretical_loss": 3.576518956750182,
      "tokens_seen": 1240662016
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039390722526952903,
      "loss": 2.7622,
      "theoretical_loss": 3.576501687577956,
      "tokens_seen": 1240727552
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039389067524115757,
      "loss": 2.7188,
      "theoretical_loss": 3.5764844195732666,
      "tokens_seen": 1240793088
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003938741252127861,
      "loss": 2.5172,
      "theoretical_loss": 3.576467152735972,
      "tokens_seen": 1240858624
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039385757518441464,
      "loss": 2.7591,
      "theoretical_loss": 3.5764498870659325,
      "tokens_seen": 1240924160
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039384102515604313,
      "loss": 2.6326,
      "theoretical_loss": 3.576432622563008,
      "tokens_seen": 1240989696
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039382447512767167,
      "loss": 2.7155,
      "theoretical_loss": 3.576415359227057,
      "tokens_seen": 1241055232
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003938079250993002,
      "loss": 2.6386,
      "theoretical_loss": 3.5763980970579388,
      "tokens_seen": 1241120768
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039379137507092874,
      "loss": 2.5698,
      "theoretical_loss": 3.576380836055514,
      "tokens_seen": 1241186304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003937748250425572,
      "loss": 2.558,
      "theoretical_loss": 3.5763635762196415,
      "tokens_seen": 1241251840
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039375827501418576,
      "loss": 2.7463,
      "theoretical_loss": 3.576346317550181,
      "tokens_seen": 1241317376
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039374172498581424,
      "loss": 2.9392,
      "theoretical_loss": 3.5763290600469926,
      "tokens_seen": 1241382912
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039372517495744273,
      "loss": 2.6483,
      "theoretical_loss": 3.5763118037099346,
      "tokens_seen": 1241448448
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039370862492907127,
      "loss": 2.7348,
      "theoretical_loss": 3.5762945485388675,
      "tokens_seen": 1241513984
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003936920749006998,
      "loss": 2.5429,
      "theoretical_loss": 3.576277294533652,
      "tokens_seen": 1241579520
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039367552487232834,
      "loss": 2.7161,
      "theoretical_loss": 3.5762600416941455,
      "tokens_seen": 1241645056
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003936589748439568,
      "loss": 2.7803,
      "theoretical_loss": 3.57624279002021,
      "tokens_seen": 1241710592
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039364242481558536,
      "loss": 2.7279,
      "theoretical_loss": 3.5762255395117037,
      "tokens_seen": 1241776128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003936258747872139,
      "loss": 2.5061,
      "theoretical_loss": 3.576208290168487,
      "tokens_seen": 1241841664
    },
    {
      "epoch": 0.44,
      "objective/train/avg_token_score": 0.01642107591032982,
      "objective/train/avg_weight": 0.9806733727455139,
      "objective/train/docs_used": 702465,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0523881912231445,
      "objective/train/std_weight": 0.08076360076665878,
      "objective/train/theoretical_loss": 3.5761910419904193,
      "objective/train/tokens_used": 1262367200,
      "theoretical_loss": 3.5761910419904193,
      "tokens_seen": 1241907200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039360932475884244,
      "loss": 2.7443,
      "theoretical_loss": 3.5761910419904193,
      "tokens_seen": 1241907200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003935927747304709,
      "loss": 2.789,
      "theoretical_loss": 3.576173794977361,
      "tokens_seen": 1241972736
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039357622470209946,
      "loss": 2.6761,
      "theoretical_loss": 3.5761565491291716,
      "tokens_seen": 1242038272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000393559674673728,
      "loss": 2.8576,
      "theoretical_loss": 3.576139304445711,
      "tokens_seen": 1242103808
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039354312464535653,
      "loss": 2.6988,
      "theoretical_loss": 3.5761220609268385,
      "tokens_seen": 1242169344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000393526574616985,
      "loss": 2.8542,
      "theoretical_loss": 3.5761048185724156,
      "tokens_seen": 1242234880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039351002458861355,
      "loss": 3.0162,
      "theoretical_loss": 3.5760875773823004,
      "tokens_seen": 1242300416
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003934934745602421,
      "loss": 2.7382,
      "theoretical_loss": 3.5760703373563545,
      "tokens_seen": 1242365952
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039347692453187063,
      "loss": 2.8753,
      "theoretical_loss": 3.5760530984944365,
      "tokens_seen": 1242431488
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003934603745034991,
      "loss": 2.9469,
      "theoretical_loss": 3.5760358607964076,
      "tokens_seen": 1242497024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039344382447512765,
      "loss": 2.7354,
      "theoretical_loss": 3.576018624262127,
      "tokens_seen": 1242562560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003934272744467562,
      "loss": 2.7055,
      "theoretical_loss": 3.5760013888914544,
      "tokens_seen": 1242628096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003934107244183847,
      "loss": 2.6515,
      "theoretical_loss": 3.5759841546842512,
      "tokens_seen": 1242693632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003933941743900132,
      "loss": 2.5343,
      "theoretical_loss": 3.575966921640376,
      "tokens_seen": 1242759168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039337762436164175,
      "loss": 2.4988,
      "theoretical_loss": 3.5759496897596907,
      "tokens_seen": 1242824704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003933610743332703,
      "loss": 2.7879,
      "theoretical_loss": 3.575932459042054,
      "tokens_seen": 1242890240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003933445243048988,
      "loss": 2.612,
      "theoretical_loss": 3.5759152294873267,
      "tokens_seen": 1242955776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003933279742765273,
      "loss": 2.9024,
      "theoretical_loss": 3.5758980010953687,
      "tokens_seen": 1243021312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039331142424815584,
      "loss": 2.8133,
      "theoretical_loss": 3.5758807738660403,
      "tokens_seen": 1243086848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003932948742197844,
      "loss": 2.727,
      "theoretical_loss": 3.5758635477992025,
      "tokens_seen": 1243152384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003932783241914129,
      "loss": 2.6151,
      "theoretical_loss": 3.575846322894715,
      "tokens_seen": 1243217920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003932617741630414,
      "loss": 2.8309,
      "theoretical_loss": 3.575829099152437,
      "tokens_seen": 1243283456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039324522413466994,
      "loss": 3.0584,
      "theoretical_loss": 3.5758118765722307,
      "tokens_seen": 1243348992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003932286741062985,
      "loss": 2.5507,
      "theoretical_loss": 3.5757946551539552,
      "tokens_seen": 1243414528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000393212124077927,
      "loss": 2.7277,
      "theoretical_loss": 3.5757774348974714,
      "tokens_seen": 1243480064
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003931955740495555,
      "loss": 2.5578,
      "theoretical_loss": 3.5757602158026396,
      "tokens_seen": 1243545600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039317902402118403,
      "loss": 2.5083,
      "theoretical_loss": 3.5757429978693205,
      "tokens_seen": 1243611136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039316247399281257,
      "loss": 2.6932,
      "theoretical_loss": 3.5757257810973737,
      "tokens_seen": 1243676672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003931459239644411,
      "loss": 2.6763,
      "theoretical_loss": 3.5757085654866607,
      "tokens_seen": 1243742208
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003931293739360696,
      "loss": 2.7041,
      "theoretical_loss": 3.575691351037041,
      "tokens_seen": 1243807744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039311282390769813,
      "loss": 2.7038,
      "theoretical_loss": 3.5756741377483765,
      "tokens_seen": 1243873280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039309627387932667,
      "loss": 2.6553,
      "theoretical_loss": 3.5756569256205264,
      "tokens_seen": 1243938816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003930797238509552,
      "loss": 2.7026,
      "theoretical_loss": 3.5756397146533514,
      "tokens_seen": 1244004352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003930631738225837,
      "loss": 2.5687,
      "theoretical_loss": 3.575622504846713,
      "tokens_seen": 1244069888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003930466237942122,
      "loss": 2.8812,
      "theoretical_loss": 3.5756052962004707,
      "tokens_seen": 1244135424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039303007376584076,
      "loss": 2.847,
      "theoretical_loss": 3.575588088714486,
      "tokens_seen": 1244200960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003930135237374693,
      "loss": 2.7724,
      "theoretical_loss": 3.57557088238862,
      "tokens_seen": 1244266496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003929969737090978,
      "loss": 2.806,
      "theoretical_loss": 3.5755536772227314,
      "tokens_seen": 1244332032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003929804236807263,
      "loss": 2.8168,
      "theoretical_loss": 3.5755364732166828,
      "tokens_seen": 1244397568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039296387365235486,
      "loss": 2.563,
      "theoretical_loss": 3.5755192703703345,
      "tokens_seen": 1244463104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003929473236239834,
      "loss": 2.6326,
      "theoretical_loss": 3.5755020686835466,
      "tokens_seen": 1244528640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003929307735956119,
      "loss": 2.624,
      "theoretical_loss": 3.5754848681561806,
      "tokens_seen": 1244594176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003929142235672404,
      "loss": 2.6741,
      "theoretical_loss": 3.575467668788097,
      "tokens_seen": 1244659712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039289767353886895,
      "loss": 2.4638,
      "theoretical_loss": 3.575450470579157,
      "tokens_seen": 1244725248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003928811235104975,
      "loss": 2.8678,
      "theoretical_loss": 3.575433273529221,
      "tokens_seen": 1244790784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000392864573482126,
      "loss": 2.5427,
      "theoretical_loss": 3.5754160776381494,
      "tokens_seen": 1244856320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003928480234537545,
      "loss": 2.576,
      "theoretical_loss": 3.5753988829058043,
      "tokens_seen": 1244921856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039283147342538305,
      "loss": 2.9335,
      "theoretical_loss": 3.575381689332046,
      "tokens_seen": 1244987392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003928149233970116,
      "loss": 2.7973,
      "theoretical_loss": 3.5753644969167353,
      "tokens_seen": 1245052928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039279837336864007,
      "loss": 2.6704,
      "theoretical_loss": 3.575347305659734,
      "tokens_seen": 1245118464
    },
    {
      "epoch": 0.44,
      "objective/train/avg_token_score": 0.015754736959934235,
      "objective/train/avg_weight": 0.9814580678939819,
      "objective/train/docs_used": 704525,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.921267032623291,
      "objective/train/std_weight": 0.09757089614868164,
      "objective/train/theoretical_loss": 3.5753301155609014,
      "objective/train/tokens_used": 1265644000,
      "theoretical_loss": 3.5753301155609014,
      "tokens_seen": 1245184000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003927818233402686,
      "loss": 2.8762,
      "theoretical_loss": 3.5753301155609014,
      "tokens_seen": 1245184000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039276527331189704,
      "loss": 2.9242,
      "theoretical_loss": 3.5753129266201005,
      "tokens_seen": 1245249536
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003927487232835256,
      "loss": 2.3996,
      "theoretical_loss": 3.575295738837191,
      "tokens_seen": 1245315072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003927321732551541,
      "loss": 2.628,
      "theoretical_loss": 3.5752785522120347,
      "tokens_seen": 1245380608
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039271562322678265,
      "loss": 2.6922,
      "theoretical_loss": 3.5752613667444924,
      "tokens_seen": 1245446144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039269907319841113,
      "loss": 2.6919,
      "theoretical_loss": 3.5752441824344254,
      "tokens_seen": 1245511680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039268252317003967,
      "loss": 2.6984,
      "theoretical_loss": 3.5752269992816945,
      "tokens_seen": 1245577216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003926659731416682,
      "loss": 2.8833,
      "theoretical_loss": 3.5752098172861606,
      "tokens_seen": 1245642752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039264942311329675,
      "loss": 2.5849,
      "theoretical_loss": 3.575192636447686,
      "tokens_seen": 1245708288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039263287308492523,
      "loss": 2.6012,
      "theoretical_loss": 3.5751754567661314,
      "tokens_seen": 1245773824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039261632305655377,
      "loss": 2.7858,
      "theoretical_loss": 3.575158278241357,
      "tokens_seen": 1245839360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0003925997730281823,
      "loss": 2.7207,
      "theoretical_loss": 3.5751411008732257,
      "tokens_seen": 1245904896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00039258322299981084,
      "loss": 2.7206,
      "theoretical_loss": 3.575123924661598,
      "tokens_seen": 1245970432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003925666729714393,
      "loss": 2.6873,
      "theoretical_loss": 3.575106749606335,
      "tokens_seen": 1246035968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039255012294306786,
      "loss": 2.6229,
      "theoretical_loss": 3.575089575707299,
      "tokens_seen": 1246101504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003925335729146964,
      "loss": 2.6849,
      "theoretical_loss": 3.5750724029643495,
      "tokens_seen": 1246167040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039251702288632494,
      "loss": 2.6605,
      "theoretical_loss": 3.57505523137735,
      "tokens_seen": 1246232576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003925004728579535,
      "loss": 2.7558,
      "theoretical_loss": 3.5750380609461607,
      "tokens_seen": 1246298112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039248392282958196,
      "loss": 2.6649,
      "theoretical_loss": 3.5750208916706434,
      "tokens_seen": 1246363648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003924673728012105,
      "loss": 2.5129,
      "theoretical_loss": 3.575003723550659,
      "tokens_seen": 1246429184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039245082277283903,
      "loss": 2.7694,
      "theoretical_loss": 3.5749865565860697,
      "tokens_seen": 1246494720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039243427274446757,
      "loss": 2.5139,
      "theoretical_loss": 3.5749693907767366,
      "tokens_seen": 1246560256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039241772271609606,
      "loss": 2.7096,
      "theoretical_loss": 3.5749522261225217,
      "tokens_seen": 1246625792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003924011726877246,
      "loss": 2.5176,
      "theoretical_loss": 3.574935062623286,
      "tokens_seen": 1246691328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039238462265935313,
      "loss": 2.5731,
      "theoretical_loss": 3.5749179002788916,
      "tokens_seen": 1246756864
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039236807263098167,
      "loss": 2.8186,
      "theoretical_loss": 3.574900739089199,
      "tokens_seen": 1246822400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039235152260261015,
      "loss": 2.6962,
      "theoretical_loss": 3.574883579054071,
      "tokens_seen": 1246887936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003923349725742387,
      "loss": 2.6669,
      "theoretical_loss": 3.574866420173369,
      "tokens_seen": 1246953472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003923184225458672,
      "loss": 2.5951,
      "theoretical_loss": 3.5748492624469543,
      "tokens_seen": 1247019008
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039230187251749576,
      "loss": 2.7986,
      "theoretical_loss": 3.5748321058746892,
      "tokens_seen": 1247084544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039228532248912425,
      "loss": 2.83,
      "theoretical_loss": 3.5748149504564344,
      "tokens_seen": 1247150080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003922687724607528,
      "loss": 2.5497,
      "theoretical_loss": 3.574797796192053,
      "tokens_seen": 1247215616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003922522224323813,
      "loss": 2.7795,
      "theoretical_loss": 3.5747806430814055,
      "tokens_seen": 1247281152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039223567240400986,
      "loss": 2.5035,
      "theoretical_loss": 3.5747634911243544,
      "tokens_seen": 1247346688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039221912237563834,
      "loss": 2.43,
      "theoretical_loss": 3.574746340320761,
      "tokens_seen": 1247412224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003922025723472669,
      "loss": 2.6756,
      "theoretical_loss": 3.574729190670488,
      "tokens_seen": 1247477760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003921860223188954,
      "loss": 2.5278,
      "theoretical_loss": 3.574712042173396,
      "tokens_seen": 1247543296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039216947229052396,
      "loss": 2.6177,
      "theoretical_loss": 3.5746948948293484,
      "tokens_seen": 1247608832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039215292226215244,
      "loss": 2.7204,
      "theoretical_loss": 3.5746777486382064,
      "tokens_seen": 1247674368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000392136372233781,
      "loss": 2.5932,
      "theoretical_loss": 3.574660603599831,
      "tokens_seen": 1247739904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003921198222054095,
      "loss": 2.6634,
      "theoretical_loss": 3.574643459714086,
      "tokens_seen": 1247805440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039210327217703805,
      "loss": 2.6759,
      "theoretical_loss": 3.574626316980832,
      "tokens_seen": 1247870976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039208672214866654,
      "loss": 2.5383,
      "theoretical_loss": 3.574609175399931,
      "tokens_seen": 1247936512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039207017212029507,
      "loss": 2.7183,
      "theoretical_loss": 3.5745920349712463,
      "tokens_seen": 1248002048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003920536220919236,
      "loss": 2.7206,
      "theoretical_loss": 3.574574895694638,
      "tokens_seen": 1248067584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039203707206355215,
      "loss": 2.7434,
      "theoretical_loss": 3.57455775756997,
      "tokens_seen": 1248133120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039202052203518063,
      "loss": 2.7742,
      "theoretical_loss": 3.5745406205971033,
      "tokens_seen": 1248198656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039200397200680917,
      "loss": 2.7106,
      "theoretical_loss": 3.5745234847759004,
      "tokens_seen": 1248264192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003919874219784377,
      "loss": 2.702,
      "theoretical_loss": 3.5745063501062235,
      "tokens_seen": 1248329728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039197087195006624,
      "loss": 2.781,
      "theoretical_loss": 3.574489216587935,
      "tokens_seen": 1248395264
    },
    {
      "epoch": 0.45,
      "objective/train/avg_token_score": 0.009189441800117493,
      "objective/train/avg_weight": 0.9891865849494934,
      "objective/train/docs_used": 706749,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6693453788757324,
      "objective/train/std_weight": 0.05263233557343483,
      "objective/train/theoretical_loss": 3.574472084220896,
      "objective/train/tokens_used": 1268920800,
      "theoretical_loss": 3.574472084220896,
      "tokens_seen": 1248460800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039195432192169473,
      "loss": 2.73,
      "theoretical_loss": 3.574472084220896,
      "tokens_seen": 1248460800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039193777189332326,
      "loss": 2.6228,
      "theoretical_loss": 3.5744549530049703,
      "tokens_seen": 1248526336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003919212218649518,
      "loss": 2.4826,
      "theoretical_loss": 3.574437822940019,
      "tokens_seen": 1248591872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039190467183658034,
      "loss": 2.7089,
      "theoretical_loss": 3.574420694025905,
      "tokens_seen": 1248657408
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003918881218082088,
      "loss": 2.8786,
      "theoretical_loss": 3.5744035662624896,
      "tokens_seen": 1248722944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039187157177983736,
      "loss": 2.5552,
      "theoretical_loss": 3.5743864396496363,
      "tokens_seen": 1248788480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003918550217514659,
      "loss": 2.6524,
      "theoretical_loss": 3.574369314187207,
      "tokens_seen": 1248854016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039183847172309444,
      "loss": 2.7161,
      "theoretical_loss": 3.5743521898750643,
      "tokens_seen": 1248919552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003918219216947229,
      "loss": 2.8329,
      "theoretical_loss": 3.5743350667130698,
      "tokens_seen": 1248985088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039180537166635146,
      "loss": 2.7105,
      "theoretical_loss": 3.574317944701086,
      "tokens_seen": 1249050624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039178882163798,
      "loss": 2.5039,
      "theoretical_loss": 3.5743008238389766,
      "tokens_seen": 1249116160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003917722716096084,
      "loss": 2.5621,
      "theoretical_loss": 3.5742837041266027,
      "tokens_seen": 1249181696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039175572158123696,
      "loss": 2.631,
      "theoretical_loss": 3.5742665855638274,
      "tokens_seen": 1249247232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003917391715528655,
      "loss": 3.0492,
      "theoretical_loss": 3.5742494681505126,
      "tokens_seen": 1249312768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000391722621524494,
      "loss": 2.6783,
      "theoretical_loss": 3.574232351886522,
      "tokens_seen": 1249378304
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003917060714961225,
      "loss": 2.7326,
      "theoretical_loss": 3.5742152367717166,
      "tokens_seen": 1249443840
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039168952146775106,
      "loss": 2.4936,
      "theoretical_loss": 3.5741981228059605,
      "tokens_seen": 1249509376
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003916729714393796,
      "loss": 2.8182,
      "theoretical_loss": 3.5741810099891156,
      "tokens_seen": 1249574912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003916564214110081,
      "loss": 2.5367,
      "theoretical_loss": 3.5741638983210438,
      "tokens_seen": 1249640448
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003916398713826366,
      "loss": 2.5028,
      "theoretical_loss": 3.574146787801609,
      "tokens_seen": 1249705984
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039162332135426515,
      "loss": 2.5514,
      "theoretical_loss": 3.5741296784306735,
      "tokens_seen": 1249771520
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003916067713258937,
      "loss": 2.7869,
      "theoretical_loss": 3.5741125702081,
      "tokens_seen": 1249837056
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003915902212975222,
      "loss": 2.9483,
      "theoretical_loss": 3.57409546313375,
      "tokens_seen": 1249902592
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003915736712691507,
      "loss": 2.6173,
      "theoretical_loss": 3.5740783572074877,
      "tokens_seen": 1249968128
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039155712124077925,
      "loss": 2.6005,
      "theoretical_loss": 3.5740612524291757,
      "tokens_seen": 1250033664
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003915405712124078,
      "loss": 2.4833,
      "theoretical_loss": 3.5740441487986767,
      "tokens_seen": 1250099200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039152402118403627,
      "loss": 2.7545,
      "theoretical_loss": 3.574027046315853,
      "tokens_seen": 1250164736
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003915074711556648,
      "loss": 2.5292,
      "theoretical_loss": 3.574009944980568,
      "tokens_seen": 1250230272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039149092112729334,
      "loss": 2.8332,
      "theoretical_loss": 3.573992844792684,
      "tokens_seen": 1250295808
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003914743710989219,
      "loss": 2.5375,
      "theoretical_loss": 3.5739757457520644,
      "tokens_seen": 1250361344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039145782107055037,
      "loss": 2.6181,
      "theoretical_loss": 3.5739586478585714,
      "tokens_seen": 1250426880
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003914412710421789,
      "loss": 2.5466,
      "theoretical_loss": 3.5739415511120693,
      "tokens_seen": 1250492416
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039142472101380744,
      "loss": 2.8318,
      "theoretical_loss": 3.573924455512419,
      "tokens_seen": 1250557952
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000391408170985436,
      "loss": 2.3878,
      "theoretical_loss": 3.5739073610594856,
      "tokens_seen": 1250623488
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039139162095706446,
      "loss": 2.6825,
      "theoretical_loss": 3.5738902677531312,
      "tokens_seen": 1250689024
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000391375070928693,
      "loss": 2.7638,
      "theoretical_loss": 3.573873175593218,
      "tokens_seen": 1250754560
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039135852090032154,
      "loss": 2.6305,
      "theoretical_loss": 3.5738560845796106,
      "tokens_seen": 1250820096
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003913419708719501,
      "loss": 2.5774,
      "theoretical_loss": 3.5738389947121707,
      "tokens_seen": 1250885632
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039132542084357856,
      "loss": 2.5075,
      "theoretical_loss": 3.573821905990762,
      "tokens_seen": 1250951168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003913088708152071,
      "loss": 2.5955,
      "theoretical_loss": 3.573804818415248,
      "tokens_seen": 1251016704
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039129232078683563,
      "loss": 2.7624,
      "theoretical_loss": 3.573787731985491,
      "tokens_seen": 1251082240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039127577075846417,
      "loss": 2.6017,
      "theoretical_loss": 3.5737706467013552,
      "tokens_seen": 1251147776
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039125922073009265,
      "loss": 2.2499,
      "theoretical_loss": 3.5737535625627026,
      "tokens_seen": 1251213312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003912426707017212,
      "loss": 2.6965,
      "theoretical_loss": 3.573736479569397,
      "tokens_seen": 1251278848
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039122612067334973,
      "loss": 2.7197,
      "theoretical_loss": 3.5737193977213018,
      "tokens_seen": 1251344384
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039120957064497827,
      "loss": 2.8639,
      "theoretical_loss": 3.57370231701828,
      "tokens_seen": 1251409920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039119302061660675,
      "loss": 2.8178,
      "theoretical_loss": 3.573685237460195,
      "tokens_seen": 1251475456
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003911764705882353,
      "loss": 2.9406,
      "theoretical_loss": 3.5736681590469104,
      "tokens_seen": 1251540992
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003911599205598638,
      "loss": 2.6761,
      "theoretical_loss": 3.5736510817782885,
      "tokens_seen": 1251606528
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039114337053149236,
      "loss": 2.7828,
      "theoretical_loss": 3.5736340056541933,
      "tokens_seen": 1251672064
    },
    {
      "epoch": 0.45,
      "objective/train/avg_token_score": 0.013426965102553368,
      "objective/train/avg_weight": 0.984242856502533,
      "objective/train/docs_used": 708687,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9367380142211914,
      "objective/train/std_weight": 0.06406966596841812,
      "objective/train/theoretical_loss": 3.5736169306744885,
      "objective/train/tokens_used": 1272197600,
      "theoretical_loss": 3.5736169306744885,
      "tokens_seen": 1251737600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039112682050312085,
      "loss": 2.7316,
      "theoretical_loss": 3.5736169306744885,
      "tokens_seen": 1251737600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003911102704747494,
      "loss": 2.6877,
      "theoretical_loss": 3.5735998568390372,
      "tokens_seen": 1251803136
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003910937204463779,
      "loss": 2.6317,
      "theoretical_loss": 3.573582784147703,
      "tokens_seen": 1251868672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039107717041800646,
      "loss": 2.7043,
      "theoretical_loss": 3.573565712600349,
      "tokens_seen": 1251934208
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039106062038963494,
      "loss": 2.6074,
      "theoretical_loss": 3.5735486421968385,
      "tokens_seen": 1251999744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003910440703612635,
      "loss": 2.526,
      "theoretical_loss": 3.5735315729370356,
      "tokens_seen": 1252065280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000391027520332892,
      "loss": 2.6801,
      "theoretical_loss": 3.5735145048208032,
      "tokens_seen": 1252130816
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039101097030452055,
      "loss": 2.6635,
      "theoretical_loss": 3.5734974378480056,
      "tokens_seen": 1252196352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039099442027614904,
      "loss": 2.6704,
      "theoretical_loss": 3.5734803720185058,
      "tokens_seen": 1252261888
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003909778702477776,
      "loss": 2.6,
      "theoretical_loss": 3.573463307332168,
      "tokens_seen": 1252327424
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003909613202194061,
      "loss": 2.7484,
      "theoretical_loss": 3.5734462437888546,
      "tokens_seen": 1252392960
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039094477019103465,
      "loss": 2.7318,
      "theoretical_loss": 3.57342918138843,
      "tokens_seen": 1252458496
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003909282201626632,
      "loss": 2.452,
      "theoretical_loss": 3.573412120130758,
      "tokens_seen": 1252524032
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039091167013429167,
      "loss": 2.6257,
      "theoretical_loss": 3.5733950600157023,
      "tokens_seen": 1252589568
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003908951201059202,
      "loss": 2.6622,
      "theoretical_loss": 3.573378001043126,
      "tokens_seen": 1252655104
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039087857007754875,
      "loss": 2.7277,
      "theoretical_loss": 3.5733609432128937,
      "tokens_seen": 1252720640
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003908620200491773,
      "loss": 2.3788,
      "theoretical_loss": 3.573343886524868,
      "tokens_seen": 1252786176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039084547002080577,
      "loss": 2.4037,
      "theoretical_loss": 3.5733268309789135,
      "tokens_seen": 1252851712
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003908289199924343,
      "loss": 2.6513,
      "theoretical_loss": 3.5733097765748942,
      "tokens_seen": 1252917248
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039081236996406284,
      "loss": 2.5026,
      "theoretical_loss": 3.5732927233126732,
      "tokens_seen": 1252982784
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039079581993569127,
      "loss": 2.543,
      "theoretical_loss": 3.573275671192115,
      "tokens_seen": 1253048320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003907792699073198,
      "loss": 2.5543,
      "theoretical_loss": 3.5732586202130827,
      "tokens_seen": 1253113856
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039076271987894835,
      "loss": 2.3787,
      "theoretical_loss": 3.5732415703754405,
      "tokens_seen": 1253179392
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039074616985057683,
      "loss": 2.7151,
      "theoretical_loss": 3.573224521679053,
      "tokens_seen": 1253244928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039072961982220537,
      "loss": 2.4272,
      "theoretical_loss": 3.5732074741237834,
      "tokens_seen": 1253310464
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003907130697938339,
      "loss": 2.8907,
      "theoretical_loss": 3.5731904277094957,
      "tokens_seen": 1253376000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039069651976546244,
      "loss": 2.7226,
      "theoretical_loss": 3.573173382436054,
      "tokens_seen": 1253441536
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003906799697370909,
      "loss": 2.5922,
      "theoretical_loss": 3.5731563383033227,
      "tokens_seen": 1253507072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039066341970871946,
      "loss": 2.7805,
      "theoretical_loss": 3.573139295311165,
      "tokens_seen": 1253572608
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000390646869680348,
      "loss": 2.5017,
      "theoretical_loss": 3.573122253459446,
      "tokens_seen": 1253638144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039063031965197654,
      "loss": 2.5541,
      "theoretical_loss": 3.5731052127480285,
      "tokens_seen": 1253703680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000390613769623605,
      "loss": 2.7485,
      "theoretical_loss": 3.573088173176778,
      "tokens_seen": 1253769216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039059721959523356,
      "loss": 2.6176,
      "theoretical_loss": 3.5730711347455575,
      "tokens_seen": 1253834752
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003905806695668621,
      "loss": 2.615,
      "theoretical_loss": 3.5730540974542313,
      "tokens_seen": 1253900288
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039056411953849063,
      "loss": 2.6352,
      "theoretical_loss": 3.573037061302664,
      "tokens_seen": 1253965824
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003905475695101191,
      "loss": 2.3596,
      "theoretical_loss": 3.5730200262907195,
      "tokens_seen": 1254031360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039053101948174765,
      "loss": 2.7127,
      "theoretical_loss": 3.5730029924182625,
      "tokens_seen": 1254096896
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003905144694533762,
      "loss": 2.6859,
      "theoretical_loss": 3.5729859596851563,
      "tokens_seen": 1254162432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039049791942500473,
      "loss": 2.6641,
      "theoretical_loss": 3.5729689280912664,
      "tokens_seen": 1254227968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003904813693966332,
      "loss": 2.6809,
      "theoretical_loss": 3.5729518976364556,
      "tokens_seen": 1254293504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039046481936826175,
      "loss": 2.686,
      "theoretical_loss": 3.5729348683205897,
      "tokens_seen": 1254359040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003904482693398903,
      "loss": 2.8465,
      "theoretical_loss": 3.572917840143532,
      "tokens_seen": 1254424576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003904317193115188,
      "loss": 2.5237,
      "theoretical_loss": 3.5729008131051474,
      "tokens_seen": 1254490112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003904151692831473,
      "loss": 2.4636,
      "theoretical_loss": 3.5728837872053,
      "tokens_seen": 1254555648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039039861925477585,
      "loss": 2.5454,
      "theoretical_loss": 3.572866762443854,
      "tokens_seen": 1254621184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003903820692264044,
      "loss": 2.7861,
      "theoretical_loss": 3.5728497388206746,
      "tokens_seen": 1254686720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003903655191980329,
      "loss": 2.607,
      "theoretical_loss": 3.5728327163356255,
      "tokens_seen": 1254752256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003903489691696614,
      "loss": 2.6204,
      "theoretical_loss": 3.572815694988572,
      "tokens_seen": 1254817792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039033241914128994,
      "loss": 2.4203,
      "theoretical_loss": 3.572798674779377,
      "tokens_seen": 1254883328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003903158691129185,
      "loss": 2.4314,
      "theoretical_loss": 3.5727816557079066,
      "tokens_seen": 1254948864
    },
    {
      "epoch": 0.45,
      "objective/train/avg_token_score": 0.039900243282318115,
      "objective/train/avg_weight": 0.9531037211418152,
      "objective/train/docs_used": 710568,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5821361541748047,
      "objective/train/std_weight": 0.14320991933345795,
      "objective/train/theoretical_loss": 3.572764637774024,
      "objective/train/tokens_used": 1275474400,
      "theoretical_loss": 3.572764637774024,
      "tokens_seen": 1255014400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000390299319084547,
      "loss": 2.5107,
      "theoretical_loss": 3.572764637774024,
      "tokens_seen": 1255014400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003902827690561755,
      "loss": 2.6749,
      "theoretical_loss": 3.5727476209775952,
      "tokens_seen": 1255079936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039026621902780404,
      "loss": 2.517,
      "theoretical_loss": 3.572730605318484,
      "tokens_seen": 1255145472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003902496689994326,
      "loss": 2.5704,
      "theoretical_loss": 3.572713590796555,
      "tokens_seen": 1255211008
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003902331189710611,
      "loss": 2.738,
      "theoretical_loss": 3.5726965774116737,
      "tokens_seen": 1255276544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003902165689426896,
      "loss": 2.602,
      "theoretical_loss": 3.5726795651637033,
      "tokens_seen": 1255342080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039020001891431813,
      "loss": 2.7804,
      "theoretical_loss": 3.572662554052509,
      "tokens_seen": 1255407616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039018346888594667,
      "loss": 2.6926,
      "theoretical_loss": 3.5726455440779565,
      "tokens_seen": 1255473152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003901669188575752,
      "loss": 2.6713,
      "theoretical_loss": 3.572628535239909,
      "tokens_seen": 1255538688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003901503688292037,
      "loss": 2.5656,
      "theoretical_loss": 3.572611527538232,
      "tokens_seen": 1255604224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039013381880083223,
      "loss": 2.5123,
      "theoretical_loss": 3.5725945209727907,
      "tokens_seen": 1255669760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039011726877246077,
      "loss": 2.5868,
      "theoretical_loss": 3.5725775155434496,
      "tokens_seen": 1255735296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003901007187440893,
      "loss": 2.6777,
      "theoretical_loss": 3.5725605112500727,
      "tokens_seen": 1255800832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003900841687157178,
      "loss": 2.6613,
      "theoretical_loss": 3.572543508092526,
      "tokens_seen": 1255866368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003900676186873463,
      "loss": 2.5201,
      "theoretical_loss": 3.5725265060706732,
      "tokens_seen": 1255931904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00039005106865897486,
      "loss": 2.5061,
      "theoretical_loss": 3.5725095051843803,
      "tokens_seen": 1255997440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003900345186306034,
      "loss": 2.5917,
      "theoretical_loss": 3.5724925054335115,
      "tokens_seen": 1256062976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003900179686022319,
      "loss": 2.5019,
      "theoretical_loss": 3.5724755068179324,
      "tokens_seen": 1256128512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003900014185738604,
      "loss": 2.7694,
      "theoretical_loss": 3.5724585093375074,
      "tokens_seen": 1256194048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038998486854548896,
      "loss": 2.3973,
      "theoretical_loss": 3.5724415129921017,
      "tokens_seen": 1256259584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003899683185171175,
      "loss": 2.5788,
      "theoretical_loss": 3.5724245177815797,
      "tokens_seen": 1256325120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000389951768488746,
      "loss": 2.5695,
      "theoretical_loss": 3.5724075237058077,
      "tokens_seen": 1256390656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003899352184603745,
      "loss": 2.6339,
      "theoretical_loss": 3.5723905307646495,
      "tokens_seen": 1256456192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038991866843200306,
      "loss": 2.6999,
      "theoretical_loss": 3.5723735389579705,
      "tokens_seen": 1256521728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003899021184036316,
      "loss": 2.6545,
      "theoretical_loss": 3.5723565482856365,
      "tokens_seen": 1256587264
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003898855683752601,
      "loss": 2.7767,
      "theoretical_loss": 3.5723395587475117,
      "tokens_seen": 1256652800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003898690183468886,
      "loss": 2.5829,
      "theoretical_loss": 3.572322570343462,
      "tokens_seen": 1256718336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038985246831851715,
      "loss": 2.6169,
      "theoretical_loss": 3.572305583073352,
      "tokens_seen": 1256783872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003898359182901457,
      "loss": 2.778,
      "theoretical_loss": 3.5722885969370473,
      "tokens_seen": 1256849408
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003898193682617742,
      "loss": 2.431,
      "theoretical_loss": 3.5722716119344122,
      "tokens_seen": 1256914944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038980281823340266,
      "loss": 2.659,
      "theoretical_loss": 3.5722546280653136,
      "tokens_seen": 1256980480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003897862682050312,
      "loss": 2.8886,
      "theoretical_loss": 3.572237645329615,
      "tokens_seen": 1257046016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003897697181766597,
      "loss": 2.5519,
      "theoretical_loss": 3.572220663727183,
      "tokens_seen": 1257111552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003897531681482882,
      "loss": 2.7887,
      "theoretical_loss": 3.5722036832578823,
      "tokens_seen": 1257177088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038973661811991675,
      "loss": 2.6482,
      "theoretical_loss": 3.572186703921578,
      "tokens_seen": 1257242624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003897200680915453,
      "loss": 2.5119,
      "theoretical_loss": 3.572169725718136,
      "tokens_seen": 1257308160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003897035180631738,
      "loss": 2.9573,
      "theoretical_loss": 3.5721527486474214,
      "tokens_seen": 1257373696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003896869680348023,
      "loss": 2.6994,
      "theoretical_loss": 3.5721357727092995,
      "tokens_seen": 1257439232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038967041800643085,
      "loss": 2.6788,
      "theoretical_loss": 3.5721187979036353,
      "tokens_seen": 1257504768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003896538679780594,
      "loss": 2.6908,
      "theoretical_loss": 3.5721018242302955,
      "tokens_seen": 1257570304
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038963731794968787,
      "loss": 2.8335,
      "theoretical_loss": 3.572084851689145,
      "tokens_seen": 1257635840
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003896207679213164,
      "loss": 2.2856,
      "theoretical_loss": 3.5720678802800485,
      "tokens_seen": 1257701376
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038960421789294494,
      "loss": 2.7717,
      "theoretical_loss": 3.572050910002872,
      "tokens_seen": 1257766912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003895876678645735,
      "loss": 2.4794,
      "theoretical_loss": 3.572033940857482,
      "tokens_seen": 1257832448
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038957111783620197,
      "loss": 2.8013,
      "theoretical_loss": 3.572016972843743,
      "tokens_seen": 1257897984
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003895545678078305,
      "loss": 2.5172,
      "theoretical_loss": 3.5720000059615202,
      "tokens_seen": 1257963520
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038953801777945904,
      "loss": 2.6367,
      "theoretical_loss": 3.5719830402106805,
      "tokens_seen": 1258029056
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003895214677510876,
      "loss": 2.6387,
      "theoretical_loss": 3.5719660755910883,
      "tokens_seen": 1258094592
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038950491772271606,
      "loss": 2.5744,
      "theoretical_loss": 3.57194911210261,
      "tokens_seen": 1258160128
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003894883676943446,
      "loss": 2.5836,
      "theoretical_loss": 3.571932149745111,
      "tokens_seen": 1258225664
    },
    {
      "epoch": 0.45,
      "objective/train/avg_token_score": 0.00616793055087328,
      "objective/train/avg_weight": 0.99274080991745,
      "objective/train/docs_used": 712328,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.894456386566162,
      "objective/train/std_weight": 0.036608241498470306,
      "objective/train/theoretical_loss": 3.571915188518457,
      "objective/train/tokens_used": 1278751200,
      "theoretical_loss": 3.571915188518457,
      "tokens_seen": 1258291200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038947181766597314,
      "loss": 2.9564,
      "theoretical_loss": 3.571915188518457,
      "tokens_seen": 1258291200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003894552676376017,
      "loss": 2.6303,
      "theoretical_loss": 3.571898228422514,
      "tokens_seen": 1258356736
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038943871760923016,
      "loss": 2.5326,
      "theoretical_loss": 3.571881269457147,
      "tokens_seen": 1258422272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003894221675808587,
      "loss": 2.9081,
      "theoretical_loss": 3.571864311622223,
      "tokens_seen": 1258487808
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038940561755248723,
      "loss": 2.5775,
      "theoretical_loss": 3.5718473549176064,
      "tokens_seen": 1258553344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038938906752411577,
      "loss": 2.3805,
      "theoretical_loss": 3.5718303993431637,
      "tokens_seen": 1258618880
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038937251749574425,
      "loss": 2.3403,
      "theoretical_loss": 3.571813444898761,
      "tokens_seen": 1258684416
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003893559674673728,
      "loss": 2.562,
      "theoretical_loss": 3.5717964915842635,
      "tokens_seen": 1258749952
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038933941743900133,
      "loss": 2.8104,
      "theoretical_loss": 3.5717795393995377,
      "tokens_seen": 1258815488
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038932286741062987,
      "loss": 2.8165,
      "theoretical_loss": 3.5717625883444493,
      "tokens_seen": 1258881024
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038930631738225835,
      "loss": 2.6869,
      "theoretical_loss": 3.5717456384188635,
      "tokens_seen": 1258946560
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003892897673538869,
      "loss": 2.6316,
      "theoretical_loss": 3.5717286896226472,
      "tokens_seen": 1259012096
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003892732173255154,
      "loss": 2.6455,
      "theoretical_loss": 3.5717117419556663,
      "tokens_seen": 1259077632
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038925666729714396,
      "loss": 2.6148,
      "theoretical_loss": 3.5716947954177867,
      "tokens_seen": 1259143168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038924011726877244,
      "loss": 2.6024,
      "theoretical_loss": 3.5716778500088737,
      "tokens_seen": 1259208704
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000389223567240401,
      "loss": 2.4718,
      "theoretical_loss": 3.5716609057287942,
      "tokens_seen": 1259274240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003892070172120295,
      "loss": 2.8348,
      "theoretical_loss": 3.571643962577414,
      "tokens_seen": 1259339776
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038919046718365806,
      "loss": 2.8478,
      "theoretical_loss": 3.5716270205545992,
      "tokens_seen": 1259405312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038917391715528654,
      "loss": 2.7711,
      "theoretical_loss": 3.571610079660215,
      "tokens_seen": 1259470848
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003891573671269151,
      "loss": 2.66,
      "theoretical_loss": 3.571593139894129,
      "tokens_seen": 1259536384
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003891408170985436,
      "loss": 2.7115,
      "theoretical_loss": 3.571576201256207,
      "tokens_seen": 1259601920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038912426707017215,
      "loss": 2.7906,
      "theoretical_loss": 3.571559263746315,
      "tokens_seen": 1259667456
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038910771704180064,
      "loss": 2.7563,
      "theoretical_loss": 3.5715423273643188,
      "tokens_seen": 1259732992
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003890911670134292,
      "loss": 2.6145,
      "theoretical_loss": 3.5715253921100847,
      "tokens_seen": 1259798528
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003890746169850577,
      "loss": 2.745,
      "theoretical_loss": 3.5715084579834793,
      "tokens_seen": 1259864064
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038905806695668625,
      "loss": 2.8978,
      "theoretical_loss": 3.571491524984369,
      "tokens_seen": 1259929600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038904151692831473,
      "loss": 2.6293,
      "theoretical_loss": 3.5714745931126193,
      "tokens_seen": 1259995136
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038902496689994327,
      "loss": 2.784,
      "theoretical_loss": 3.5714576623680974,
      "tokens_seen": 1260060672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003890084168715718,
      "loss": 2.5491,
      "theoretical_loss": 3.571440732750669,
      "tokens_seen": 1260126208
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038899186684320035,
      "loss": 2.8879,
      "theoretical_loss": 3.571423804260201,
      "tokens_seen": 1260191744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038897531681482883,
      "loss": 2.6135,
      "theoretical_loss": 3.5714068768965594,
      "tokens_seen": 1260257280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038895876678645737,
      "loss": 2.5774,
      "theoretical_loss": 3.5713899506596105,
      "tokens_seen": 1260322816
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003889422167580859,
      "loss": 2.6699,
      "theoretical_loss": 3.571373025549221,
      "tokens_seen": 1260388352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038892566672971444,
      "loss": 2.8022,
      "theoretical_loss": 3.5713561015652573,
      "tokens_seen": 1260453888
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003889091167013429,
      "loss": 2.5582,
      "theoretical_loss": 3.5713391787075857,
      "tokens_seen": 1260519424
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038889256667297146,
      "loss": 2.6758,
      "theoretical_loss": 3.571322256976073,
      "tokens_seen": 1260584960
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003888760166446,
      "loss": 2.6498,
      "theoretical_loss": 3.571305336370586,
      "tokens_seen": 1260650496
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038885946661622854,
      "loss": 2.6225,
      "theoretical_loss": 3.57128841689099,
      "tokens_seen": 1260716032
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000388842916587857,
      "loss": 2.5508,
      "theoretical_loss": 3.5712714985371528,
      "tokens_seen": 1260781568
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003888263665594855,
      "loss": 2.5221,
      "theoretical_loss": 3.5712545813089402,
      "tokens_seen": 1260847104
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000388809816531114,
      "loss": 2.75,
      "theoretical_loss": 3.57123766520622,
      "tokens_seen": 1260912640
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003887932665027425,
      "loss": 2.8161,
      "theoretical_loss": 3.571220750228857,
      "tokens_seen": 1260978176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038877671647437106,
      "loss": 2.6546,
      "theoretical_loss": 3.571203836376719,
      "tokens_seen": 1261043712
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003887601664459996,
      "loss": 2.6545,
      "theoretical_loss": 3.5711869236496727,
      "tokens_seen": 1261109248
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003887436164176281,
      "loss": 2.725,
      "theoretical_loss": 3.5711700120475847,
      "tokens_seen": 1261174784
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003887270663892566,
      "loss": 2.6372,
      "theoretical_loss": 3.5711531015703217,
      "tokens_seen": 1261240320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038871051636088516,
      "loss": 2.7112,
      "theoretical_loss": 3.5711361922177502,
      "tokens_seen": 1261305856
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003886939663325137,
      "loss": 2.6267,
      "theoretical_loss": 3.571119283989737,
      "tokens_seen": 1261371392
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003886774163041422,
      "loss": 2.6817,
      "theoretical_loss": 3.5711023768861487,
      "tokens_seen": 1261436928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003886608662757707,
      "loss": 2.7289,
      "theoretical_loss": 3.571085470906853,
      "tokens_seen": 1261502464
    },
    {
      "epoch": 0.45,
      "objective/train/avg_token_score": 0.02035050094127655,
      "objective/train/avg_weight": 0.9760775566101074,
      "objective/train/docs_used": 713944,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2140555381774902,
      "objective/train/std_weight": 0.10421819984912872,
      "objective/train/theoretical_loss": 3.571068566051716,
      "objective/train/tokens_used": 1282028000,
      "theoretical_loss": 3.571068566051716,
      "tokens_seen": 1261568000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038864431624739925,
      "loss": 2.6265,
      "theoretical_loss": 3.571068566051716,
      "tokens_seen": 1261568000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003886277662190278,
      "loss": 2.9393,
      "theoretical_loss": 3.571051662320605,
      "tokens_seen": 1261633536
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003886112161906563,
      "loss": 2.579,
      "theoretical_loss": 3.5710347597133865,
      "tokens_seen": 1261699072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003885946661622848,
      "loss": 2.7155,
      "theoretical_loss": 3.5710178582299275,
      "tokens_seen": 1261764608
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038857811613391335,
      "loss": 2.856,
      "theoretical_loss": 3.571000957870094,
      "tokens_seen": 1261830144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003885615661055419,
      "loss": 2.6229,
      "theoretical_loss": 3.570984058633755,
      "tokens_seen": 1261895680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038854501607717037,
      "loss": 2.8412,
      "theoretical_loss": 3.5709671605207762,
      "tokens_seen": 1261961216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003885284660487989,
      "loss": 2.6979,
      "theoretical_loss": 3.5709502635310244,
      "tokens_seen": 1262026752
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038851191602042745,
      "loss": 2.582,
      "theoretical_loss": 3.5709333676643675,
      "tokens_seen": 1262092288
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000388495365992056,
      "loss": 2.5906,
      "theoretical_loss": 3.570916472920671,
      "tokens_seen": 1262157824
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003884788159636845,
      "loss": 2.6657,
      "theoretical_loss": 3.570899579299804,
      "tokens_seen": 1262223360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000388462265935313,
      "loss": 2.627,
      "theoretical_loss": 3.570882686801632,
      "tokens_seen": 1262288896
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038844571590694154,
      "loss": 2.5026,
      "theoretical_loss": 3.570865795426023,
      "tokens_seen": 1262354432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003884291658785701,
      "loss": 2.5665,
      "theoretical_loss": 3.570848905172843,
      "tokens_seen": 1262419968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003884126158501986,
      "loss": 2.5637,
      "theoretical_loss": 3.570832016041961,
      "tokens_seen": 1262485504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003883960658218271,
      "loss": 2.7998,
      "theoretical_loss": 3.5708151280332423,
      "tokens_seen": 1262551040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038837951579345564,
      "loss": 2.6211,
      "theoretical_loss": 3.5707982411465555,
      "tokens_seen": 1262616576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003883629657650842,
      "loss": 2.8994,
      "theoretical_loss": 3.570781355381767,
      "tokens_seen": 1262682112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003883464157367127,
      "loss": 2.4944,
      "theoretical_loss": 3.570764470738744,
      "tokens_seen": 1262747648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003883298657083412,
      "loss": 2.5412,
      "theoretical_loss": 3.5707475872173546,
      "tokens_seen": 1262813184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038831331567996973,
      "loss": 2.7762,
      "theoretical_loss": 3.570730704817465,
      "tokens_seen": 1262878720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038829676565159827,
      "loss": 2.8328,
      "theoretical_loss": 3.5707138235389433,
      "tokens_seen": 1262944256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003882802156232268,
      "loss": 2.8685,
      "theoretical_loss": 3.570696943381656,
      "tokens_seen": 1263009792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003882636655948553,
      "loss": 2.7897,
      "theoretical_loss": 3.5706800643454715,
      "tokens_seen": 1263075328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038824711556648383,
      "loss": 2.5747,
      "theoretical_loss": 3.5706631864302567,
      "tokens_seen": 1263140864
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038823056553811237,
      "loss": 2.6462,
      "theoretical_loss": 3.5706463096358787,
      "tokens_seen": 1263206400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003882140155097409,
      "loss": 2.8177,
      "theoretical_loss": 3.570629433962206,
      "tokens_seen": 1263271936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003881974654813694,
      "loss": 2.6582,
      "theoretical_loss": 3.5706125594091045,
      "tokens_seen": 1263337472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003881809154529979,
      "loss": 2.5739,
      "theoretical_loss": 3.5705956859764427,
      "tokens_seen": 1263403008
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038816436542462646,
      "loss": 2.8223,
      "theoretical_loss": 3.5705788136640875,
      "tokens_seen": 1263468544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000388147815396255,
      "loss": 2.764,
      "theoretical_loss": 3.5705619424719073,
      "tokens_seen": 1263534080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003881312653678835,
      "loss": 2.7156,
      "theoretical_loss": 3.5705450723997685,
      "tokens_seen": 1263599616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000388114715339512,
      "loss": 2.848,
      "theoretical_loss": 3.5705282034475396,
      "tokens_seen": 1263665152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038809816531114056,
      "loss": 2.6572,
      "theoretical_loss": 3.5705113356150875,
      "tokens_seen": 1263730688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003880816152827691,
      "loss": 2.5605,
      "theoretical_loss": 3.5704944689022806,
      "tokens_seen": 1263796224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003880650652543976,
      "loss": 2.6096,
      "theoretical_loss": 3.570477603308986,
      "tokens_seen": 1263861760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003880485152260261,
      "loss": 2.6239,
      "theoretical_loss": 3.570460738835071,
      "tokens_seen": 1263927296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038803196519765466,
      "loss": 2.6063,
      "theoretical_loss": 3.5704438754804038,
      "tokens_seen": 1263992832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003880154151692832,
      "loss": 2.6668,
      "theoretical_loss": 3.570427013244852,
      "tokens_seen": 1264058368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003879988651409117,
      "loss": 2.7427,
      "theoretical_loss": 3.570410152128283,
      "tokens_seen": 1264123904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003879823151125402,
      "loss": 2.281,
      "theoretical_loss": 3.570393292130565,
      "tokens_seen": 1264189440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038796576508416875,
      "loss": 2.6927,
      "theoretical_loss": 3.5703764332515657,
      "tokens_seen": 1264254976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003879492150557973,
      "loss": 2.7565,
      "theoretical_loss": 3.5703595754911523,
      "tokens_seen": 1264320512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038793266502742577,
      "loss": 2.8117,
      "theoretical_loss": 3.5703427188491936,
      "tokens_seen": 1264386048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003879161149990543,
      "loss": 2.3204,
      "theoretical_loss": 3.570325863325557,
      "tokens_seen": 1264451584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038789956497068285,
      "loss": 2.7851,
      "theoretical_loss": 3.5703090089201095,
      "tokens_seen": 1264517120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003878830149423114,
      "loss": 2.6345,
      "theoretical_loss": 3.5702921556327203,
      "tokens_seen": 1264582656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038786646491393987,
      "loss": 2.8775,
      "theoretical_loss": 3.5702753034632564,
      "tokens_seen": 1264648192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038784991488556835,
      "loss": 2.6334,
      "theoretical_loss": 3.5702584524115863,
      "tokens_seen": 1264713728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038783336485719684,
      "loss": 2.6375,
      "theoretical_loss": 3.5702416024775774,
      "tokens_seen": 1264779264
    },
    {
      "epoch": 0.45,
      "objective/train/avg_token_score": 0.009163270704448223,
      "objective/train/avg_weight": 0.989212691783905,
      "objective/train/docs_used": 715923,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.005680561065674,
      "objective/train/std_weight": 0.0668521523475647,
      "objective/train/theoretical_loss": 3.5702247536610976,
      "objective/train/tokens_used": 1285304800,
      "theoretical_loss": 3.5702247536610976,
      "tokens_seen": 1264844800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038781681482882537,
      "loss": 2.955,
      "theoretical_loss": 3.5702247536610976,
      "tokens_seen": 1264844800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003878002648004539,
      "loss": 2.8803,
      "theoretical_loss": 3.570207905962016,
      "tokens_seen": 1264910336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038778371477208245,
      "loss": 2.4632,
      "theoretical_loss": 3.570191059380199,
      "tokens_seen": 1264975872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038776716474371093,
      "loss": 2.8397,
      "theoretical_loss": 3.570174213915516,
      "tokens_seen": 1265041408
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038775061471533947,
      "loss": 2.7781,
      "theoretical_loss": 3.5701573695678346,
      "tokens_seen": 1265106944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000387734064686968,
      "loss": 2.7407,
      "theoretical_loss": 3.5701405263370223,
      "tokens_seen": 1265172480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038771751465859654,
      "loss": 2.472,
      "theoretical_loss": 3.5701236842229482,
      "tokens_seen": 1265238016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000387700964630225,
      "loss": 2.6761,
      "theoretical_loss": 3.5701068432254797,
      "tokens_seen": 1265303552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038768441460185356,
      "loss": 2.4918,
      "theoretical_loss": 3.570090003344485,
      "tokens_seen": 1265369088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003876678645734821,
      "loss": 2.9448,
      "theoretical_loss": 3.5700731645798323,
      "tokens_seen": 1265434624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038765131454511064,
      "loss": 2.6172,
      "theoretical_loss": 3.57005632693139,
      "tokens_seen": 1265500160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003876347645167391,
      "loss": 2.7517,
      "theoretical_loss": 3.570039490399026,
      "tokens_seen": 1265565696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038761821448836766,
      "loss": 2.7295,
      "theoretical_loss": 3.5700226549826093,
      "tokens_seen": 1265631232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003876016644599962,
      "loss": 2.7596,
      "theoretical_loss": 3.570005820682007,
      "tokens_seen": 1265696768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038758511443162474,
      "loss": 2.6017,
      "theoretical_loss": 3.569988987497088,
      "tokens_seen": 1265762304
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003875685644032532,
      "loss": 3.0771,
      "theoretical_loss": 3.569972155427721,
      "tokens_seen": 1265827840
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038755201437488176,
      "loss": 2.6795,
      "theoretical_loss": 3.5699553244737734,
      "tokens_seen": 1265893376
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003875354643465103,
      "loss": 2.7331,
      "theoretical_loss": 3.5699384946351147,
      "tokens_seen": 1265958912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038751891431813883,
      "loss": 2.4569,
      "theoretical_loss": 3.569921665911612,
      "tokens_seen": 1266024448
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003875023642897673,
      "loss": 2.5998,
      "theoretical_loss": 3.569904838303134,
      "tokens_seen": 1266089984
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038748581426139585,
      "loss": 2.7509,
      "theoretical_loss": 3.56988801180955,
      "tokens_seen": 1266155520
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003874692642330244,
      "loss": 2.5698,
      "theoretical_loss": 3.5698711864307278,
      "tokens_seen": 1266221056
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038745271420465293,
      "loss": 2.3908,
      "theoretical_loss": 3.5698543621665353,
      "tokens_seen": 1266286592
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003874361641762814,
      "loss": 2.7095,
      "theoretical_loss": 3.5698375390168424,
      "tokens_seen": 1266352128
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038741961414790995,
      "loss": 2.704,
      "theoretical_loss": 3.569820716981516,
      "tokens_seen": 1266417664
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003874030641195385,
      "loss": 2.578,
      "theoretical_loss": 3.5698038960604257,
      "tokens_seen": 1266483200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000387386514091167,
      "loss": 2.9218,
      "theoretical_loss": 3.56978707625344,
      "tokens_seen": 1266548736
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003873699640627955,
      "loss": 2.5865,
      "theoretical_loss": 3.5697702575604264,
      "tokens_seen": 1266614272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038735341403442404,
      "loss": 2.5193,
      "theoretical_loss": 3.5697534399812545,
      "tokens_seen": 1266679808
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003873368640060526,
      "loss": 2.6533,
      "theoretical_loss": 3.5697366235157935,
      "tokens_seen": 1266745344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003873203139776811,
      "loss": 2.8533,
      "theoretical_loss": 3.56971980816391,
      "tokens_seen": 1266810880
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003873037639493096,
      "loss": 2.2318,
      "theoretical_loss": 3.5697029939254747,
      "tokens_seen": 1266876416
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038728721392093814,
      "loss": 2.716,
      "theoretical_loss": 3.5696861808003555,
      "tokens_seen": 1266941952
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003872706638925667,
      "loss": 2.6434,
      "theoretical_loss": 3.5696693687884204,
      "tokens_seen": 1267007488
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003872541138641952,
      "loss": 2.9088,
      "theoretical_loss": 3.5696525578895395,
      "tokens_seen": 1267073024
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003872375638358237,
      "loss": 2.6132,
      "theoretical_loss": 3.5696357481035808,
      "tokens_seen": 1267138560
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038722101380745224,
      "loss": 2.7497,
      "theoretical_loss": 3.5696189394304128,
      "tokens_seen": 1267204096
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003872044637790808,
      "loss": 2.5734,
      "theoretical_loss": 3.5696021318699045,
      "tokens_seen": 1267269632
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003871879137507093,
      "loss": 2.7434,
      "theoretical_loss": 3.569585325421925,
      "tokens_seen": 1267335168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003871713637223378,
      "loss": 2.7574,
      "theoretical_loss": 3.5695685200863427,
      "tokens_seen": 1267400704
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038715481369396633,
      "loss": 2.5055,
      "theoretical_loss": 3.5695517158630268,
      "tokens_seen": 1267466240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038713826366559487,
      "loss": 2.6954,
      "theoretical_loss": 3.569534912751846,
      "tokens_seen": 1267531776
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003871217136372234,
      "loss": 2.8808,
      "theoretical_loss": 3.5695181107526697,
      "tokens_seen": 1267597312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003871051636088519,
      "loss": 2.8426,
      "theoretical_loss": 3.569501309865366,
      "tokens_seen": 1267662848
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038708861358048043,
      "loss": 2.7302,
      "theoretical_loss": 3.5694845100898043,
      "tokens_seen": 1267728384
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038707206355210897,
      "loss": 2.5789,
      "theoretical_loss": 3.5694677114258533,
      "tokens_seen": 1267793920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003870555135237375,
      "loss": 2.7513,
      "theoretical_loss": 3.569450913873382,
      "tokens_seen": 1267859456
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000387038963495366,
      "loss": 2.5486,
      "theoretical_loss": 3.5694341174322606,
      "tokens_seen": 1267924992
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003870224134669945,
      "loss": 2.7649,
      "theoretical_loss": 3.5694173221023564,
      "tokens_seen": 1267990528
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038700586343862306,
      "loss": 2.4795,
      "theoretical_loss": 3.5694005278835395,
      "tokens_seen": 1268056064
    },
    {
      "epoch": 0.45,
      "objective/train/avg_token_score": 0.012880174443125725,
      "objective/train/avg_weight": 0.9848463535308838,
      "objective/train/docs_used": 717630,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8774025440216064,
      "objective/train/std_weight": 0.07367522269487381,
      "objective/train/theoretical_loss": 3.5693837347756783,
      "objective/train/tokens_used": 1288581600,
      "theoretical_loss": 3.5693837347756783,
      "tokens_seen": 1268121600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003869893134102516,
      "loss": 2.7391,
      "theoretical_loss": 3.5693837347756783,
      "tokens_seen": 1268121600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038697276338188014,
      "loss": 2.5844,
      "theoretical_loss": 3.5693669427786423,
      "tokens_seen": 1268187136
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003869562133535086,
      "loss": 2.6999,
      "theoretical_loss": 3.5693501518923005,
      "tokens_seen": 1268252672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038693966332513716,
      "loss": 2.7422,
      "theoretical_loss": 3.569333362116523,
      "tokens_seen": 1268318208
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003869231132967657,
      "loss": 2.6974,
      "theoretical_loss": 3.5693165734511774,
      "tokens_seen": 1268383744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038690656326839423,
      "loss": 2.7581,
      "theoretical_loss": 3.569299785896134,
      "tokens_seen": 1268449280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003868900132400227,
      "loss": 2.8669,
      "theoretical_loss": 3.5692829994512607,
      "tokens_seen": 1268514816
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038687346321165125,
      "loss": 2.6381,
      "theoretical_loss": 3.5692662141164284,
      "tokens_seen": 1268580352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003868569131832797,
      "loss": 2.781,
      "theoretical_loss": 3.5692494298915056,
      "tokens_seen": 1268645888
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003868403631549082,
      "loss": 2.4732,
      "theoretical_loss": 3.5692326467763618,
      "tokens_seen": 1268711424
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038682381312653676,
      "loss": 2.8297,
      "theoretical_loss": 3.5692158647708654,
      "tokens_seen": 1268776960
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003868072630981653,
      "loss": 2.7768,
      "theoretical_loss": 3.569199083874887,
      "tokens_seen": 1268842496
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003867907130697938,
      "loss": 2.4492,
      "theoretical_loss": 3.5691823040882946,
      "tokens_seen": 1268908032
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003867741630414223,
      "loss": 2.6352,
      "theoretical_loss": 3.569165525410959,
      "tokens_seen": 1268973568
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038675761301305085,
      "loss": 2.8298,
      "theoretical_loss": 3.569148747842749,
      "tokens_seen": 1269039104
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003867410629846794,
      "loss": 2.8108,
      "theoretical_loss": 3.569131971383533,
      "tokens_seen": 1269104640
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003867245129563079,
      "loss": 2.4421,
      "theoretical_loss": 3.5691151960331826,
      "tokens_seen": 1269170176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003867079629279364,
      "loss": 2.5373,
      "theoretical_loss": 3.5690984217915656,
      "tokens_seen": 1269235712
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038669141289956495,
      "loss": 2.621,
      "theoretical_loss": 3.5690816486585515,
      "tokens_seen": 1269301248
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003866748628711935,
      "loss": 2.6727,
      "theoretical_loss": 3.56906487663401,
      "tokens_seen": 1269366784
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038665831284282197,
      "loss": 2.6668,
      "theoretical_loss": 3.5690481057178114,
      "tokens_seen": 1269432320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003866417628144505,
      "loss": 2.5802,
      "theoretical_loss": 3.5690313359098242,
      "tokens_seen": 1269497856
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038662521278607905,
      "loss": 2.8314,
      "theoretical_loss": 3.5690145672099187,
      "tokens_seen": 1269563392
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003866086627577076,
      "loss": 2.7984,
      "theoretical_loss": 3.5689977996179643,
      "tokens_seen": 1269628928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038659211272933607,
      "loss": 2.5766,
      "theoretical_loss": 3.56898103313383,
      "tokens_seen": 1269694464
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003865755627009646,
      "loss": 2.4571,
      "theoretical_loss": 3.5689642677573863,
      "tokens_seen": 1269760000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038655901267259314,
      "loss": 2.6031,
      "theoretical_loss": 3.568947503488503,
      "tokens_seen": 1269825536
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003865424626442217,
      "loss": 2.6076,
      "theoretical_loss": 3.5689307403270485,
      "tokens_seen": 1269891072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038652591261585016,
      "loss": 2.6061,
      "theoretical_loss": 3.568913978272893,
      "tokens_seen": 1269956608
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003865093625874787,
      "loss": 2.7132,
      "theoretical_loss": 3.5688972173259073,
      "tokens_seen": 1270022144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038649281255910724,
      "loss": 2.4752,
      "theoretical_loss": 3.5688804574859603,
      "tokens_seen": 1270087680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003864762625307358,
      "loss": 2.5622,
      "theoretical_loss": 3.5688636987529216,
      "tokens_seen": 1270153216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038645971250236426,
      "loss": 2.628,
      "theoretical_loss": 3.568846941126661,
      "tokens_seen": 1270218752
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003864431624739928,
      "loss": 2.6199,
      "theoretical_loss": 3.5688301846070485,
      "tokens_seen": 1270284288
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038642661244562133,
      "loss": 2.6718,
      "theoretical_loss": 3.568813429193954,
      "tokens_seen": 1270349824
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038641006241724987,
      "loss": 2.7182,
      "theoretical_loss": 3.5687966748872473,
      "tokens_seen": 1270415360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038639351238887835,
      "loss": 2.6318,
      "theoretical_loss": 3.568779921686798,
      "tokens_seen": 1270480896
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003863769623605069,
      "loss": 2.6354,
      "theoretical_loss": 3.5687631695924766,
      "tokens_seen": 1270546432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038636041233213543,
      "loss": 2.9827,
      "theoretical_loss": 3.568746418604153,
      "tokens_seen": 1270611968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038634386230376397,
      "loss": 2.0222,
      "theoretical_loss": 3.5687296687216965,
      "tokens_seen": 1270677504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038632731227539245,
      "loss": 2.7642,
      "theoretical_loss": 3.568712919944977,
      "tokens_seen": 1270743040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000386310762247021,
      "loss": 2.65,
      "theoretical_loss": 3.5686961722738655,
      "tokens_seen": 1270808576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003862942122186495,
      "loss": 2.7372,
      "theoretical_loss": 3.568679425708231,
      "tokens_seen": 1270874112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038627766219027806,
      "loss": 2.4595,
      "theoretical_loss": 3.568662680247944,
      "tokens_seen": 1270939648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038626111216190655,
      "loss": 2.943,
      "theoretical_loss": 3.5686459358928744,
      "tokens_seen": 1271005184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003862445621335351,
      "loss": 2.5997,
      "theoretical_loss": 3.568629192642892,
      "tokens_seen": 1271070720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003862280121051636,
      "loss": 2.8975,
      "theoretical_loss": 3.5686124504978673,
      "tokens_seen": 1271136256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038621146207679216,
      "loss": 2.5048,
      "theoretical_loss": 3.5685957094576706,
      "tokens_seen": 1271201792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038619491204842064,
      "loss": 2.7385,
      "theoretical_loss": 3.5685789695221715,
      "tokens_seen": 1271267328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003861783620200492,
      "loss": 2.5737,
      "theoretical_loss": 3.5685622306912403,
      "tokens_seen": 1271332864
    },
    {
      "epoch": 0.45,
      "objective/train/avg_token_score": 0.00908350758254528,
      "objective/train/avg_weight": 0.9893068075180054,
      "objective/train/docs_used": 719315,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.890028953552246,
      "objective/train/std_weight": 0.052003201097249985,
      "objective/train/theoretical_loss": 3.5685454929647475,
      "objective/train/tokens_used": 1291858400,
      "theoretical_loss": 3.5685454929647475,
      "tokens_seen": 1271398400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003861618119916777,
      "loss": 2.6927,
      "theoretical_loss": 3.5685454929647475,
      "tokens_seen": 1271398400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038614526196330626,
      "loss": 2.7056,
      "theoretical_loss": 3.5685287563425634,
      "tokens_seen": 1271463936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038612871193493474,
      "loss": 2.8466,
      "theoretical_loss": 3.5685120208245578,
      "tokens_seen": 1271529472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003861121619065633,
      "loss": 2.4161,
      "theoretical_loss": 3.568495286410601,
      "tokens_seen": 1271595008
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003860956118781918,
      "loss": 2.6204,
      "theoretical_loss": 3.5684785531005634,
      "tokens_seen": 1271660544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038607906184982035,
      "loss": 2.6868,
      "theoretical_loss": 3.5684618208943153,
      "tokens_seen": 1271726080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038606251182144883,
      "loss": 2.4679,
      "theoretical_loss": 3.5684450897917266,
      "tokens_seen": 1271791616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038604596179307737,
      "loss": 2.5975,
      "theoretical_loss": 3.5684283597926685,
      "tokens_seen": 1271857152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003860294117647059,
      "loss": 2.7622,
      "theoretical_loss": 3.5684116308970104,
      "tokens_seen": 1271922688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038601286173633445,
      "loss": 2.6662,
      "theoretical_loss": 3.5683949031046236,
      "tokens_seen": 1271988224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038599631170796293,
      "loss": 2.5692,
      "theoretical_loss": 3.5683781764153775,
      "tokens_seen": 1272053760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038597976167959147,
      "loss": 2.9349,
      "theoretical_loss": 3.5683614508291432,
      "tokens_seen": 1272119296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038596321165122,
      "loss": 2.6142,
      "theoretical_loss": 3.5683447263457913,
      "tokens_seen": 1272184832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038594666162284854,
      "loss": 2.638,
      "theoretical_loss": 3.568328002965192,
      "tokens_seen": 1272250368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000385930111594477,
      "loss": 2.5585,
      "theoretical_loss": 3.5683112806872157,
      "tokens_seen": 1272315904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038591356156610556,
      "loss": 2.3178,
      "theoretical_loss": 3.568294559511733,
      "tokens_seen": 1272381440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003858970115377341,
      "loss": 2.8023,
      "theoretical_loss": 3.568277839438614,
      "tokens_seen": 1272446976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038588046150936253,
      "loss": 2.6694,
      "theoretical_loss": 3.5682611204677297,
      "tokens_seen": 1272512512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038586391148099107,
      "loss": 2.6638,
      "theoretical_loss": 3.568244402598951,
      "tokens_seen": 1272578048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003858473614526196,
      "loss": 2.953,
      "theoretical_loss": 3.5682276858321478,
      "tokens_seen": 1272643584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038583081142424814,
      "loss": 2.5528,
      "theoretical_loss": 3.5682109701671916,
      "tokens_seen": 1272709120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003858142613958766,
      "loss": 2.5427,
      "theoretical_loss": 3.568194255603952,
      "tokens_seen": 1272774656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038579771136750516,
      "loss": 2.5568,
      "theoretical_loss": 3.5681775421423003,
      "tokens_seen": 1272840192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003857811613391337,
      "loss": 2.727,
      "theoretical_loss": 3.5681608297821072,
      "tokens_seen": 1272905728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038576461131076224,
      "loss": 2.568,
      "theoretical_loss": 3.568144118523243,
      "tokens_seen": 1272971264
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003857480612823907,
      "loss": 2.8379,
      "theoretical_loss": 3.5681274083655787,
      "tokens_seen": 1273036800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038573151125401926,
      "loss": 2.7523,
      "theoretical_loss": 3.5681106993089853,
      "tokens_seen": 1273102336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003857149612256478,
      "loss": 2.6613,
      "theoretical_loss": 3.568093991353333,
      "tokens_seen": 1273167872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038569841119727634,
      "loss": 2.5996,
      "theoretical_loss": 3.568077284498493,
      "tokens_seen": 1273233408
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003856818611689048,
      "loss": 2.8226,
      "theoretical_loss": 3.568060578744336,
      "tokens_seen": 1273298944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038566531114053336,
      "loss": 2.4,
      "theoretical_loss": 3.5680438740907334,
      "tokens_seen": 1273364480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003856487611121619,
      "loss": 2.4081,
      "theoretical_loss": 3.5680271705375546,
      "tokens_seen": 1273430016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038563221108379043,
      "loss": 2.7946,
      "theoretical_loss": 3.5680104680846725,
      "tokens_seen": 1273495552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003856156610554189,
      "loss": 2.5629,
      "theoretical_loss": 3.5679937667319557,
      "tokens_seen": 1273561088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038559911102704745,
      "loss": 2.7081,
      "theoretical_loss": 3.567977066479277,
      "tokens_seen": 1273626624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000385582560998676,
      "loss": 2.6872,
      "theoretical_loss": 3.567960367326507,
      "tokens_seen": 1273692160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003855660109703045,
      "loss": 2.7183,
      "theoretical_loss": 3.5679436692735154,
      "tokens_seen": 1273757696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000385549460941933,
      "loss": 2.6814,
      "theoretical_loss": 3.567926972320175,
      "tokens_seen": 1273823232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00038553291091356155,
      "loss": 2.683,
      "theoretical_loss": 3.5679102764663555,
      "tokens_seen": 1273888768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0003855163608851901,
      "loss": 2.376,
      "theoretical_loss": 3.567893581711929,
      "tokens_seen": 1273954304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003854998108568186,
      "loss": 2.741,
      "theoretical_loss": 3.567876888056765,
      "tokens_seen": 1274019840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003854832608284471,
      "loss": 2.6577,
      "theoretical_loss": 3.567860195500736,
      "tokens_seen": 1274085376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038546671080007564,
      "loss": 2.6299,
      "theoretical_loss": 3.5678435040437124,
      "tokens_seen": 1274150912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003854501607717042,
      "loss": 2.751,
      "theoretical_loss": 3.567826813685566,
      "tokens_seen": 1274216448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003854336107433327,
      "loss": 2.4921,
      "theoretical_loss": 3.5678101244261677,
      "tokens_seen": 1274281984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003854170607149612,
      "loss": 2.8143,
      "theoretical_loss": 3.5677934362653874,
      "tokens_seen": 1274347520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038540051068658974,
      "loss": 2.7832,
      "theoretical_loss": 3.567776749203098,
      "tokens_seen": 1274413056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003853839606582183,
      "loss": 2.7831,
      "theoretical_loss": 3.56776006323917,
      "tokens_seen": 1274478592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003853674106298468,
      "loss": 2.7491,
      "theoretical_loss": 3.567743378373475,
      "tokens_seen": 1274544128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003853508606014753,
      "loss": 2.5566,
      "theoretical_loss": 3.5677266946058834,
      "tokens_seen": 1274609664
    },
    {
      "epoch": 0.46,
      "objective/train/avg_token_score": 0.01178013626486063,
      "objective/train/avg_weight": 0.9861443638801575,
      "objective/train/docs_used": 720761,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.788282871246338,
      "objective/train/std_weight": 0.06058068946003914,
      "objective/train/theoretical_loss": 3.5677100119362675,
      "objective/train/tokens_used": 1295135200,
      "theoretical_loss": 3.5677100119362675,
      "tokens_seen": 1274675200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038533431057310384,
      "loss": 2.7573,
      "theoretical_loss": 3.5677100119362675,
      "tokens_seen": 1274675200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003853177605447324,
      "loss": 2.611,
      "theoretical_loss": 3.5676933303644978,
      "tokens_seen": 1274740736
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003853012105163609,
      "loss": 2.5035,
      "theoretical_loss": 3.567676649890446,
      "tokens_seen": 1274806272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003852846604879894,
      "loss": 2.5597,
      "theoretical_loss": 3.5676599705139833,
      "tokens_seen": 1274871808
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038526811045961793,
      "loss": 2.9157,
      "theoretical_loss": 3.5676432922349814,
      "tokens_seen": 1274937344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038525156043124647,
      "loss": 2.3044,
      "theoretical_loss": 3.567626615053311,
      "tokens_seen": 1275002880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000385235010402875,
      "loss": 2.7247,
      "theoretical_loss": 3.567609938968844,
      "tokens_seen": 1275068416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003852184603745035,
      "loss": 2.761,
      "theoretical_loss": 3.567593263981452,
      "tokens_seen": 1275133952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038520191034613203,
      "loss": 2.8401,
      "theoretical_loss": 3.5675765900910057,
      "tokens_seen": 1275199488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038518536031776057,
      "loss": 2.757,
      "theoretical_loss": 3.5675599172973778,
      "tokens_seen": 1275265024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003851688102893891,
      "loss": 2.5026,
      "theoretical_loss": 3.5675432456004383,
      "tokens_seen": 1275330560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003851522602610176,
      "loss": 2.4618,
      "theoretical_loss": 3.5675265750000595,
      "tokens_seen": 1275396096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003851357102326461,
      "loss": 2.8242,
      "theoretical_loss": 3.567509905496113,
      "tokens_seen": 1275461632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038511916020427466,
      "loss": 2.766,
      "theoretical_loss": 3.5674932370884704,
      "tokens_seen": 1275527168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003851026101759032,
      "loss": 2.4447,
      "theoretical_loss": 3.5674765697770034,
      "tokens_seen": 1275592704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003850860601475317,
      "loss": 2.6634,
      "theoretical_loss": 3.5674599035615824,
      "tokens_seen": 1275658240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003850695101191602,
      "loss": 2.7259,
      "theoretical_loss": 3.5674432384420807,
      "tokens_seen": 1275723776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038505296009078876,
      "loss": 2.7345,
      "theoretical_loss": 3.567426574418369,
      "tokens_seen": 1275789312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003850364100624173,
      "loss": 2.3973,
      "theoretical_loss": 3.5674099114903193,
      "tokens_seen": 1275854848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003850198600340458,
      "loss": 2.412,
      "theoretical_loss": 3.567393249657803,
      "tokens_seen": 1275920384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003850033100056743,
      "loss": 2.524,
      "theoretical_loss": 3.567376588920692,
      "tokens_seen": 1275985920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038498675997730285,
      "loss": 2.7725,
      "theoretical_loss": 3.567359929278858,
      "tokens_seen": 1276051456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003849702099489314,
      "loss": 2.8091,
      "theoretical_loss": 3.567343270732172,
      "tokens_seen": 1276116992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003849536599205599,
      "loss": 2.6102,
      "theoretical_loss": 3.5673266132805073,
      "tokens_seen": 1276182528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003849371098921884,
      "loss": 2.6899,
      "theoretical_loss": 3.5673099569237348,
      "tokens_seen": 1276248064
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038492055986381695,
      "loss": 2.8993,
      "theoretical_loss": 3.567293301661726,
      "tokens_seen": 1276313600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003849040098354455,
      "loss": 2.7514,
      "theoretical_loss": 3.567276647494354,
      "tokens_seen": 1276379136
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003848874598070739,
      "loss": 2.2878,
      "theoretical_loss": 3.567259994421489,
      "tokens_seen": 1276444672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038487090977870245,
      "loss": 2.666,
      "theoretical_loss": 3.5672433424430037,
      "tokens_seen": 1276510208
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038485435975033094,
      "loss": 2.4614,
      "theoretical_loss": 3.56722669155877,
      "tokens_seen": 1276575744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003848378097219595,
      "loss": 2.5472,
      "theoretical_loss": 3.56721004176866,
      "tokens_seen": 1276641280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000384821259693588,
      "loss": 2.7079,
      "theoretical_loss": 3.5671933930725457,
      "tokens_seen": 1276706816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038480470966521655,
      "loss": 2.7538,
      "theoretical_loss": 3.5671767454702983,
      "tokens_seen": 1276772352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038478815963684503,
      "loss": 2.8102,
      "theoretical_loss": 3.56716009896179,
      "tokens_seen": 1276837888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038477160960847357,
      "loss": 2.7133,
      "theoretical_loss": 3.567143453546894,
      "tokens_seen": 1276903424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003847550595801021,
      "loss": 2.6123,
      "theoretical_loss": 3.567126809225481,
      "tokens_seen": 1276968960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038473850955173065,
      "loss": 2.9725,
      "theoretical_loss": 3.5671101659974234,
      "tokens_seen": 1277034496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038472195952335913,
      "loss": 2.7392,
      "theoretical_loss": 3.567093523862593,
      "tokens_seen": 1277100032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038470540949498767,
      "loss": 2.677,
      "theoretical_loss": 3.567076882820863,
      "tokens_seen": 1277165568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003846888594666162,
      "loss": 2.8202,
      "theoretical_loss": 3.5670602428721043,
      "tokens_seen": 1277231104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038467230943824474,
      "loss": 2.6658,
      "theoretical_loss": 3.56704360401619,
      "tokens_seen": 1277296640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003846557594098732,
      "loss": 2.9781,
      "theoretical_loss": 3.5670269662529916,
      "tokens_seen": 1277362176
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038463920938150176,
      "loss": 2.8327,
      "theoretical_loss": 3.5670103295823816,
      "tokens_seen": 1277427712
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003846226593531303,
      "loss": 2.6187,
      "theoretical_loss": 3.5669936940042315,
      "tokens_seen": 1277493248
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038460610932475884,
      "loss": 2.628,
      "theoretical_loss": 3.5669770595184143,
      "tokens_seen": 1277558784
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003845895592963873,
      "loss": 2.4497,
      "theoretical_loss": 3.566960426124802,
      "tokens_seen": 1277624320
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038457300926801586,
      "loss": 2.5327,
      "theoretical_loss": 3.566943793823267,
      "tokens_seen": 1277689856
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003845564592396444,
      "loss": 2.753,
      "theoretical_loss": 3.5669271626136814,
      "tokens_seen": 1277755392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038453990921127293,
      "loss": 2.4806,
      "theoretical_loss": 3.5669105324959176,
      "tokens_seen": 1277820928
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003845233591829014,
      "loss": 2.49,
      "theoretical_loss": 3.566893903469848,
      "tokens_seen": 1277886464
    },
    {
      "epoch": 0.46,
      "objective/train/avg_token_score": 0.027104580774903297,
      "objective/train/avg_weight": 0.968107283115387,
      "objective/train/docs_used": 722731,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.568668842315674,
      "objective/train/std_weight": 0.12682220339775085,
      "objective/train/theoretical_loss": 3.566877275535345,
      "objective/train/tokens_used": 1298412000,
      "theoretical_loss": 3.566877275535345,
      "tokens_seen": 1277952000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038450680915452995,
      "loss": 2.5855,
      "theoretical_loss": 3.566877275535345,
      "tokens_seen": 1277952000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003844902591261585,
      "loss": 2.711,
      "theoretical_loss": 3.5668606486922805,
      "tokens_seen": 1278017536
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038447370909778703,
      "loss": 2.3961,
      "theoretical_loss": 3.566844022940528,
      "tokens_seen": 1278083072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038445715906941557,
      "loss": 2.5239,
      "theoretical_loss": 3.5668273982799583,
      "tokens_seen": 1278148608
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038444060904104405,
      "loss": 2.7068,
      "theoretical_loss": 3.5668107747104454,
      "tokens_seen": 1278214144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003844240590126726,
      "loss": 2.3713,
      "theoretical_loss": 3.5667941522318607,
      "tokens_seen": 1278279680
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003844075089843011,
      "loss": 2.5592,
      "theoretical_loss": 3.566777530844077,
      "tokens_seen": 1278345216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038439095895592966,
      "loss": 2.7396,
      "theoretical_loss": 3.5667609105469666,
      "tokens_seen": 1278410752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038437440892755815,
      "loss": 2.7707,
      "theoretical_loss": 3.5667442913404024,
      "tokens_seen": 1278476288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003843578588991867,
      "loss": 2.7259,
      "theoretical_loss": 3.5667276732242574,
      "tokens_seen": 1278541824
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003843413088708152,
      "loss": 2.6128,
      "theoretical_loss": 3.566711056198403,
      "tokens_seen": 1278607360
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038432475884244376,
      "loss": 2.7626,
      "theoretical_loss": 3.5666944402627125,
      "tokens_seen": 1278672896
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038430820881407224,
      "loss": 2.8688,
      "theoretical_loss": 3.5666778254170586,
      "tokens_seen": 1278738432
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003842916587857008,
      "loss": 2.6823,
      "theoretical_loss": 3.5666612116613137,
      "tokens_seen": 1278803968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003842751087573293,
      "loss": 2.8156,
      "theoretical_loss": 3.5666445989953504,
      "tokens_seen": 1278869504
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038425855872895785,
      "loss": 2.6622,
      "theoretical_loss": 3.566627987419041,
      "tokens_seen": 1278935040
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038424200870058634,
      "loss": 2.6942,
      "theoretical_loss": 3.5666113769322596,
      "tokens_seen": 1279000576
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003842254586722149,
      "loss": 2.76,
      "theoretical_loss": 3.5665947675348777,
      "tokens_seen": 1279066112
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003842089086438434,
      "loss": 2.8416,
      "theoretical_loss": 3.5665781592267676,
      "tokens_seen": 1279131648
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038419235861547195,
      "loss": 2.7513,
      "theoretical_loss": 3.5665615520078036,
      "tokens_seen": 1279197184
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038417580858710043,
      "loss": 2.6427,
      "theoretical_loss": 3.5665449458778573,
      "tokens_seen": 1279262720
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038415925855872897,
      "loss": 2.2217,
      "theoretical_loss": 3.566528340836802,
      "tokens_seen": 1279328256
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003841427085303575,
      "loss": 2.6367,
      "theoretical_loss": 3.56651173688451,
      "tokens_seen": 1279393792
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038412615850198605,
      "loss": 2.6501,
      "theoretical_loss": 3.5664951340208546,
      "tokens_seen": 1279459328
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038410960847361453,
      "loss": 2.366,
      "theoretical_loss": 3.566478532245709,
      "tokens_seen": 1279524864
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038409305844524307,
      "loss": 2.7268,
      "theoretical_loss": 3.566461931558945,
      "tokens_seen": 1279590400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003840765084168716,
      "loss": 2.6116,
      "theoretical_loss": 3.566445331960437,
      "tokens_seen": 1279655936
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038405995838850014,
      "loss": 2.6097,
      "theoretical_loss": 3.5664287334500564,
      "tokens_seen": 1279721472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003840434083601286,
      "loss": 2.7524,
      "theoretical_loss": 3.5664121360276777,
      "tokens_seen": 1279787008
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038402685833175716,
      "loss": 2.504,
      "theoretical_loss": 3.566395539693172,
      "tokens_seen": 1279852544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003840103083033857,
      "loss": 2.6091,
      "theoretical_loss": 3.566378944446414,
      "tokens_seen": 1279918080
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038399375827501424,
      "loss": 2.6467,
      "theoretical_loss": 3.5663623502872763,
      "tokens_seen": 1279983616
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003839772082466427,
      "loss": 2.6497,
      "theoretical_loss": 3.5663457572156307,
      "tokens_seen": 1280049152
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038396065821827126,
      "loss": 2.4844,
      "theoretical_loss": 3.5663291652313522,
      "tokens_seen": 1280114688
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003839441081898998,
      "loss": 2.8253,
      "theoretical_loss": 3.5663125743343125,
      "tokens_seen": 1280180224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038392755816152833,
      "loss": 2.5681,
      "theoretical_loss": 3.566295984524386,
      "tokens_seen": 1280245760
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038391100813315676,
      "loss": 2.6585,
      "theoretical_loss": 3.566279395801444,
      "tokens_seen": 1280311296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003838944581047853,
      "loss": 2.7369,
      "theoretical_loss": 3.5662628081653605,
      "tokens_seen": 1280376832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003838779080764138,
      "loss": 2.5221,
      "theoretical_loss": 3.566246221616009,
      "tokens_seen": 1280442368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003838613580480423,
      "loss": 2.6116,
      "theoretical_loss": 3.566229636153263,
      "tokens_seen": 1280507904
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038384480801967086,
      "loss": 2.7374,
      "theoretical_loss": 3.5662130517769945,
      "tokens_seen": 1280573440
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003838282579912994,
      "loss": 2.821,
      "theoretical_loss": 3.5661964684870777,
      "tokens_seen": 1280638976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003838117079629279,
      "loss": 2.2645,
      "theoretical_loss": 3.5661798862833853,
      "tokens_seen": 1280704512
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003837951579345564,
      "loss": 2.5308,
      "theoretical_loss": 3.5661633051657917,
      "tokens_seen": 1280770048
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038377860790618496,
      "loss": 2.5457,
      "theoretical_loss": 3.5661467251341685,
      "tokens_seen": 1280835584
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003837620578778135,
      "loss": 2.7608,
      "theoretical_loss": 3.5661301461883896,
      "tokens_seen": 1280901120
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000383745507849442,
      "loss": 2.6196,
      "theoretical_loss": 3.566113568328329,
      "tokens_seen": 1280966656
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003837289578210705,
      "loss": 2.6483,
      "theoretical_loss": 3.5660969915538594,
      "tokens_seen": 1281032192
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038371240779269905,
      "loss": 2.952,
      "theoretical_loss": 3.566080415864855,
      "tokens_seen": 1281097728
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003836958577643276,
      "loss": 2.7891,
      "theoretical_loss": 3.566063841261188,
      "tokens_seen": 1281163264
    },
    {
      "epoch": 0.46,
      "objective/train/avg_token_score": 0.008442874066531658,
      "objective/train/avg_weight": 0.9900636076927185,
      "objective/train/docs_used": 724789,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.568305015563965,
      "objective/train/std_weight": 0.04019840434193611,
      "objective/train/theoretical_loss": 3.566047267742733,
      "objective/train/tokens_used": 1301688800,
      "theoretical_loss": 3.566047267742733,
      "tokens_seen": 1281228800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038367930773595607,
      "loss": 2.6466,
      "theoretical_loss": 3.566047267742733,
      "tokens_seen": 1281228800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003836627577075846,
      "loss": 2.7077,
      "theoretical_loss": 3.5660306953093626,
      "tokens_seen": 1281294336
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038364620767921315,
      "loss": 2.8475,
      "theoretical_loss": 3.56601412396095,
      "tokens_seen": 1281359872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003836296576508417,
      "loss": 2.6915,
      "theoretical_loss": 3.56599755369737,
      "tokens_seen": 1281425408
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038361310762247017,
      "loss": 2.4916,
      "theoretical_loss": 3.565980984518495,
      "tokens_seen": 1281490944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003835965575940987,
      "loss": 2.5969,
      "theoretical_loss": 3.5659644164241984,
      "tokens_seen": 1281556480
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038358000756572724,
      "loss": 2.7713,
      "theoretical_loss": 3.565947849414355,
      "tokens_seen": 1281622016
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003835634575373558,
      "loss": 2.5086,
      "theoretical_loss": 3.565931283488837,
      "tokens_seen": 1281687552
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038354690750898426,
      "loss": 2.6806,
      "theoretical_loss": 3.5659147186475186,
      "tokens_seen": 1281753088
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003835303574806128,
      "loss": 2.4777,
      "theoretical_loss": 3.5658981548902737,
      "tokens_seen": 1281818624
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038351380745224134,
      "loss": 2.7715,
      "theoretical_loss": 3.565881592216975,
      "tokens_seen": 1281884160
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003834972574238699,
      "loss": 2.8597,
      "theoretical_loss": 3.5658650306274975,
      "tokens_seen": 1281949696
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038348070739549836,
      "loss": 2.5303,
      "theoretical_loss": 3.565848470121714,
      "tokens_seen": 1282015232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003834641573671269,
      "loss": 2.6907,
      "theoretical_loss": 3.5658319106994982,
      "tokens_seen": 1282080768
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038344760733875544,
      "loss": 2.6556,
      "theoretical_loss": 3.5658153523607234,
      "tokens_seen": 1282146304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000383431057310384,
      "loss": 2.5498,
      "theoretical_loss": 3.5657987951052648,
      "tokens_seen": 1282211840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038341450728201246,
      "loss": 2.72,
      "theoretical_loss": 3.5657822389329947,
      "tokens_seen": 1282277376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000383397957253641,
      "loss": 2.7338,
      "theoretical_loss": 3.5657656838437877,
      "tokens_seen": 1282342912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038338140722526953,
      "loss": 2.4104,
      "theoretical_loss": 3.565749129837517,
      "tokens_seen": 1282408448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038336485719689807,
      "loss": 2.5627,
      "theoretical_loss": 3.565732576914057,
      "tokens_seen": 1282473984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038334830716852655,
      "loss": 2.8816,
      "theoretical_loss": 3.5657160250732813,
      "tokens_seen": 1282539520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003833317571401551,
      "loss": 2.7935,
      "theoretical_loss": 3.5656994743150636,
      "tokens_seen": 1282605056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038331520711178363,
      "loss": 2.7442,
      "theoretical_loss": 3.565682924639278,
      "tokens_seen": 1282670592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038329865708341216,
      "loss": 2.5684,
      "theoretical_loss": 3.5656663760457987,
      "tokens_seen": 1282736128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038328210705504065,
      "loss": 2.5929,
      "theoretical_loss": 3.565649828534499,
      "tokens_seen": 1282801664
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003832655570266692,
      "loss": 2.5208,
      "theoretical_loss": 3.565633282105253,
      "tokens_seen": 1282867200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003832490069982977,
      "loss": 2.6905,
      "theoretical_loss": 3.565616736757935,
      "tokens_seen": 1282932736
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038323245696992626,
      "loss": 2.5046,
      "theoretical_loss": 3.565600192492419,
      "tokens_seen": 1282998272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038321590694155474,
      "loss": 2.7168,
      "theoretical_loss": 3.5655836493085786,
      "tokens_seen": 1283063808
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003831993569131833,
      "loss": 2.7554,
      "theoretical_loss": 3.565567107206288,
      "tokens_seen": 1283129344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003831828068848118,
      "loss": 2.5146,
      "theoretical_loss": 3.5655505661854217,
      "tokens_seen": 1283194880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038316625685644036,
      "loss": 2.4582,
      "theoretical_loss": 3.565534026245853,
      "tokens_seen": 1283260416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038314970682806884,
      "loss": 2.4721,
      "theoretical_loss": 3.5655174873874564,
      "tokens_seen": 1283325952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003831331567996974,
      "loss": 2.6842,
      "theoretical_loss": 3.565500949610106,
      "tokens_seen": 1283391488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003831166067713259,
      "loss": 2.488,
      "theoretical_loss": 3.5654844129136762,
      "tokens_seen": 1283457024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038310005674295445,
      "loss": 2.7227,
      "theoretical_loss": 3.565467877298041,
      "tokens_seen": 1283522560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038308350671458294,
      "loss": 2.799,
      "theoretical_loss": 3.5654513427630743,
      "tokens_seen": 1283588096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003830669566862115,
      "loss": 2.2997,
      "theoretical_loss": 3.56543480930865,
      "tokens_seen": 1283653632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038305040665784,
      "loss": 2.7474,
      "theoretical_loss": 3.5654182769346434,
      "tokens_seen": 1283719168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038303385662946855,
      "loss": 2.6071,
      "theoretical_loss": 3.565401745640928,
      "tokens_seen": 1283784704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038301730660109703,
      "loss": 2.7975,
      "theoretical_loss": 3.565385215427378,
      "tokens_seen": 1283850240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038300075657272557,
      "loss": 2.6146,
      "theoretical_loss": 3.5653686862938687,
      "tokens_seen": 1283915776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003829842065443541,
      "loss": 2.6327,
      "theoretical_loss": 3.565352158240273,
      "tokens_seen": 1283981312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038296765651598264,
      "loss": 2.6311,
      "theoretical_loss": 3.565335631266466,
      "tokens_seen": 1284046848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003829511064876112,
      "loss": 2.894,
      "theoretical_loss": 3.565319105372322,
      "tokens_seen": 1284112384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003829345564592396,
      "loss": 2.4769,
      "theoretical_loss": 3.565302580557715,
      "tokens_seen": 1284177920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038291800643086815,
      "loss": 2.6305,
      "theoretical_loss": 3.5652860568225204,
      "tokens_seen": 1284243456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038290145640249663,
      "loss": 2.4982,
      "theoretical_loss": 3.565269534166611,
      "tokens_seen": 1284308992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038288490637412517,
      "loss": 2.7245,
      "theoretical_loss": 3.5652530125898627,
      "tokens_seen": 1284374528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003828683563457537,
      "loss": 2.6591,
      "theoretical_loss": 3.565236492092149,
      "tokens_seen": 1284440064
    },
    {
      "epoch": 0.46,
      "objective/train/avg_token_score": 0.00923580676317215,
      "objective/train/avg_weight": 0.9891255497932434,
      "objective/train/docs_used": 727075,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.929180145263672,
      "objective/train/std_weight": 0.060251541435718536,
      "objective/train/theoretical_loss": 3.5652199726733453,
      "objective/train/tokens_used": 1304965600,
      "theoretical_loss": 3.5652199726733453,
      "tokens_seen": 1284505600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038285180631738224,
      "loss": 2.8083,
      "theoretical_loss": 3.5652199726733453,
      "tokens_seen": 1284505600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038283525628901073,
      "loss": 2.6676,
      "theoretical_loss": 3.565203454333325,
      "tokens_seen": 1284571136
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038281870626063927,
      "loss": 2.7207,
      "theoretical_loss": 3.5651869370719633,
      "tokens_seen": 1284636672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003828021562322678,
      "loss": 2.525,
      "theoretical_loss": 3.565170420889135,
      "tokens_seen": 1284702208
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038278560620389634,
      "loss": 2.6873,
      "theoretical_loss": 3.565153905784714,
      "tokens_seen": 1284767744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003827690561755248,
      "loss": 2.8061,
      "theoretical_loss": 3.565137391758575,
      "tokens_seen": 1284833280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038275250614715336,
      "loss": 2.6572,
      "theoretical_loss": 3.565120878810593,
      "tokens_seen": 1284898816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003827359561187819,
      "loss": 2.3641,
      "theoretical_loss": 3.565104366940642,
      "tokens_seen": 1284964352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038271940609041044,
      "loss": 2.69,
      "theoretical_loss": 3.565087856148598,
      "tokens_seen": 1285029888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003827028560620389,
      "loss": 2.6745,
      "theoretical_loss": 3.5650713464343338,
      "tokens_seen": 1285095424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038268630603366746,
      "loss": 2.9982,
      "theoretical_loss": 3.5650548377977254,
      "tokens_seen": 1285160960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000382669756005296,
      "loss": 2.65,
      "theoretical_loss": 3.5650383302386466,
      "tokens_seen": 1285226496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038265320597692453,
      "loss": 2.3034,
      "theoretical_loss": 3.5650218237569735,
      "tokens_seen": 1285292032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000382636655948553,
      "loss": 2.6466,
      "theoretical_loss": 3.565005318352579,
      "tokens_seen": 1285357568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038262010592018155,
      "loss": 2.8677,
      "theoretical_loss": 3.5649888140253396,
      "tokens_seen": 1285423104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003826035558918101,
      "loss": 2.6118,
      "theoretical_loss": 3.564972310775129,
      "tokens_seen": 1285488640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038258700586343863,
      "loss": 2.5919,
      "theoretical_loss": 3.5649558086018223,
      "tokens_seen": 1285554176
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003825704558350671,
      "loss": 2.8139,
      "theoretical_loss": 3.564939307505295,
      "tokens_seen": 1285619712
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038255390580669565,
      "loss": 2.865,
      "theoretical_loss": 3.564922807485421,
      "tokens_seen": 1285685248
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003825373557783242,
      "loss": 2.6486,
      "theoretical_loss": 3.564906308542075,
      "tokens_seen": 1285750784
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003825208057499527,
      "loss": 2.7092,
      "theoretical_loss": 3.5648898106751328,
      "tokens_seen": 1285816320
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003825042557215812,
      "loss": 2.7966,
      "theoretical_loss": 3.564873313884469,
      "tokens_seen": 1285881856
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038248770569320975,
      "loss": 2.4427,
      "theoretical_loss": 3.5648568181699583,
      "tokens_seen": 1285947392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003824711556648383,
      "loss": 2.4066,
      "theoretical_loss": 3.5648403235314756,
      "tokens_seen": 1286012928
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003824546056364668,
      "loss": 2.4494,
      "theoretical_loss": 3.564823829968897,
      "tokens_seen": 1286078464
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003824380556080953,
      "loss": 2.638,
      "theoretical_loss": 3.5648073374820957,
      "tokens_seen": 1286144000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038242150557972384,
      "loss": 2.5786,
      "theoretical_loss": 3.564790846070948,
      "tokens_seen": 1286209536
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003824049555513524,
      "loss": 2.762,
      "theoretical_loss": 3.5647743557353286,
      "tokens_seen": 1286275072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003823884055229809,
      "loss": 2.452,
      "theoretical_loss": 3.5647578664751123,
      "tokens_seen": 1286340608
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003823718554946094,
      "loss": 2.7046,
      "theoretical_loss": 3.564741378290175,
      "tokens_seen": 1286406144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038235530546623794,
      "loss": 2.6631,
      "theoretical_loss": 3.564724891180391,
      "tokens_seen": 1286471680
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003823387554378665,
      "loss": 2.6217,
      "theoretical_loss": 3.564708405145635,
      "tokens_seen": 1286537216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000382322205409495,
      "loss": 2.8377,
      "theoretical_loss": 3.5646919201857834,
      "tokens_seen": 1286602752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003823056553811235,
      "loss": 2.1416,
      "theoretical_loss": 3.5646754363007105,
      "tokens_seen": 1286668288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038228910535275203,
      "loss": 2.4081,
      "theoretical_loss": 3.564658953490292,
      "tokens_seen": 1286733824
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038227255532438057,
      "loss": 2.5929,
      "theoretical_loss": 3.5646424717544027,
      "tokens_seen": 1286799360
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003822560052960091,
      "loss": 2.4177,
      "theoretical_loss": 3.5646259910929183,
      "tokens_seen": 1286864896
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003822394552676376,
      "loss": 2.616,
      "theoretical_loss": 3.5646095115057137,
      "tokens_seen": 1286930432
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038222290523926613,
      "loss": 2.7014,
      "theoretical_loss": 3.564593032992664,
      "tokens_seen": 1286995968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038220635521089467,
      "loss": 2.7099,
      "theoretical_loss": 3.5645765555536446,
      "tokens_seen": 1287061504
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003821898051825232,
      "loss": 2.4427,
      "theoretical_loss": 3.5645600791885315,
      "tokens_seen": 1287127040
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003821732551541517,
      "loss": 2.6227,
      "theoretical_loss": 3.564543603897199,
      "tokens_seen": 1287192576
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003821567051257802,
      "loss": 2.6022,
      "theoretical_loss": 3.564527129679523,
      "tokens_seen": 1287258112
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038214015509740876,
      "loss": 2.4632,
      "theoretical_loss": 3.564510656535379,
      "tokens_seen": 1287323648
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003821236050690373,
      "loss": 2.3198,
      "theoretical_loss": 3.5644941844646416,
      "tokens_seen": 1287389184
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003821070550406658,
      "loss": 2.6795,
      "theoretical_loss": 3.5644777134671877,
      "tokens_seen": 1287454720
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003820905050122943,
      "loss": 2.4936,
      "theoretical_loss": 3.5644612435428913,
      "tokens_seen": 1287520256
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038207395498392286,
      "loss": 2.8436,
      "theoretical_loss": 3.564444774691628,
      "tokens_seen": 1287585792
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003820574049555514,
      "loss": 2.724,
      "theoretical_loss": 3.5644283069132743,
      "tokens_seen": 1287651328
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003820408549271799,
      "loss": 2.6777,
      "theoretical_loss": 3.564411840207705,
      "tokens_seen": 1287716864
    },
    {
      "epoch": 0.46,
      "objective/train/avg_token_score": 0.008972886949777603,
      "objective/train/avg_weight": 0.9894444942474365,
      "objective/train/docs_used": 728554,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.79038143157959,
      "objective/train/std_weight": 0.0443020835518837,
      "objective/train/theoretical_loss": 3.564395374574796,
      "objective/train/tokens_used": 1308242400,
      "theoretical_loss": 3.564395374574796,
      "tokens_seen": 1287782400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003820243048988084,
      "loss": 2.6727,
      "theoretical_loss": 3.564395374574796,
      "tokens_seen": 1287782400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038200775487043695,
      "loss": 2.4778,
      "theoretical_loss": 3.564378910014422,
      "tokens_seen": 1287847936
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003819912048420655,
      "loss": 2.8842,
      "theoretical_loss": 3.5643624465264594,
      "tokens_seen": 1287913472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000381974654813694,
      "loss": 2.6438,
      "theoretical_loss": 3.5643459841107834,
      "tokens_seen": 1287979008
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003819581047853225,
      "loss": 2.653,
      "theoretical_loss": 3.5643295227672698,
      "tokens_seen": 1288044544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000381941554756951,
      "loss": 2.6164,
      "theoretical_loss": 3.564313062495794,
      "tokens_seen": 1288110080
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003819250047285795,
      "loss": 2.8255,
      "theoretical_loss": 3.564296603296232,
      "tokens_seen": 1288175616
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000381908454700208,
      "loss": 2.7501,
      "theoretical_loss": 3.564280145168459,
      "tokens_seen": 1288241152
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038189190467183656,
      "loss": 2.4759,
      "theoretical_loss": 3.5642636881123515,
      "tokens_seen": 1288306688
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003818753546434651,
      "loss": 2.7942,
      "theoretical_loss": 3.564247232127784,
      "tokens_seen": 1288372224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003818588046150936,
      "loss": 2.4757,
      "theoretical_loss": 3.5642307772146333,
      "tokens_seen": 1288437760
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003818422545867221,
      "loss": 2.8076,
      "theoretical_loss": 3.5642143233727746,
      "tokens_seen": 1288503296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038182570455835065,
      "loss": 2.6929,
      "theoretical_loss": 3.5641978706020834,
      "tokens_seen": 1288568832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003818091545299792,
      "loss": 2.6751,
      "theoretical_loss": 3.5641814189024363,
      "tokens_seen": 1288634368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038179260450160767,
      "loss": 2.4928,
      "theoretical_loss": 3.5641649682737087,
      "tokens_seen": 1288699904
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003817760544732362,
      "loss": 2.4503,
      "theoretical_loss": 3.5641485187157764,
      "tokens_seen": 1288765440
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038175950444486475,
      "loss": 2.8434,
      "theoretical_loss": 3.5641320702285153,
      "tokens_seen": 1288830976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003817429544164933,
      "loss": 2.6421,
      "theoretical_loss": 3.5641156228118014,
      "tokens_seen": 1288896512
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038172640438812177,
      "loss": 2.5732,
      "theoretical_loss": 3.5640991764655103,
      "tokens_seen": 1288962048
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003817098543597503,
      "loss": 2.7875,
      "theoretical_loss": 3.564082731189518,
      "tokens_seen": 1289027584
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038169330433137884,
      "loss": 2.8042,
      "theoretical_loss": 3.5640662869837003,
      "tokens_seen": 1289093120
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003816767543030074,
      "loss": 2.6711,
      "theoretical_loss": 3.5640498438479336,
      "tokens_seen": 1289158656
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038166020427463586,
      "loss": 2.8434,
      "theoretical_loss": 3.564033401782094,
      "tokens_seen": 1289224192
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003816436542462644,
      "loss": 2.8352,
      "theoretical_loss": 3.5640169607860566,
      "tokens_seen": 1289289728
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038162710421789294,
      "loss": 2.737,
      "theoretical_loss": 3.564000520859698,
      "tokens_seen": 1289355264
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003816105541895215,
      "loss": 2.6832,
      "theoretical_loss": 3.5639840820028947,
      "tokens_seen": 1289420800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038159400416114996,
      "loss": 2.7744,
      "theoretical_loss": 3.5639676442155217,
      "tokens_seen": 1289486336
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003815774541327785,
      "loss": 2.7217,
      "theoretical_loss": 3.563951207497456,
      "tokens_seen": 1289551872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038156090410440703,
      "loss": 2.7212,
      "theoretical_loss": 3.563934771848573,
      "tokens_seen": 1289617408
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038154435407603557,
      "loss": 2.8433,
      "theoretical_loss": 3.5639183372687495,
      "tokens_seen": 1289682944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038152780404766406,
      "loss": 2.4778,
      "theoretical_loss": 3.5639019037578614,
      "tokens_seen": 1289748480
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003815112540192926,
      "loss": 2.4089,
      "theoretical_loss": 3.5638854713157846,
      "tokens_seen": 1289814016
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038149470399092113,
      "loss": 2.6796,
      "theoretical_loss": 3.5638690399423956,
      "tokens_seen": 1289879552
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038147815396254967,
      "loss": 2.4868,
      "theoretical_loss": 3.5638526096375704,
      "tokens_seen": 1289945088
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038146160393417815,
      "loss": 2.961,
      "theoretical_loss": 3.563836180401185,
      "tokens_seen": 1290010624
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003814450539058067,
      "loss": 2.7789,
      "theoretical_loss": 3.5638197522331163,
      "tokens_seen": 1290076160
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003814285038774352,
      "loss": 2.7893,
      "theoretical_loss": 3.56380332513324,
      "tokens_seen": 1290141696
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038141195384906376,
      "loss": 2.6545,
      "theoretical_loss": 3.5637868991014328,
      "tokens_seen": 1290207232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038139540382069225,
      "loss": 2.5141,
      "theoretical_loss": 3.563770474137571,
      "tokens_seen": 1290272768
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003813788537923208,
      "loss": 2.6457,
      "theoretical_loss": 3.56375405024153,
      "tokens_seen": 1290338304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003813623037639493,
      "loss": 2.4978,
      "theoretical_loss": 3.5637376274131873,
      "tokens_seen": 1290403840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038134575373557786,
      "loss": 2.662,
      "theoretical_loss": 3.563721205652419,
      "tokens_seen": 1290469376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038132920370720634,
      "loss": 2.5958,
      "theoretical_loss": 3.563704784959101,
      "tokens_seen": 1290534912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003813126536788349,
      "loss": 2.8206,
      "theoretical_loss": 3.5636883653331104,
      "tokens_seen": 1290600448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003812961036504634,
      "loss": 2.749,
      "theoretical_loss": 3.563671946774323,
      "tokens_seen": 1290665984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038127955362209196,
      "loss": 2.5344,
      "theoretical_loss": 3.5636555292826158,
      "tokens_seen": 1290731520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038126300359372044,
      "loss": 2.6116,
      "theoretical_loss": 3.563639112857865,
      "tokens_seen": 1290797056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000381246453565349,
      "loss": 2.9862,
      "theoretical_loss": 3.563622697499947,
      "tokens_seen": 1290862592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003812299035369775,
      "loss": 2.7735,
      "theoretical_loss": 3.5636062832087383,
      "tokens_seen": 1290928128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038121335350860605,
      "loss": 2.6662,
      "theoretical_loss": 3.5635898699841158,
      "tokens_seen": 1290993664
    },
    {
      "epoch": 0.46,
      "objective/train/avg_token_score": 0.007539168931543827,
      "objective/train/avg_weight": 0.9911264777183533,
      "objective/train/docs_used": 730680,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.753121852874756,
      "objective/train/std_weight": 0.04433155804872513,
      "objective/train/theoretical_loss": 3.5635734578259557,
      "objective/train/tokens_used": 1311519200,
      "theoretical_loss": 3.5635734578259557,
      "tokens_seen": 1291059200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038119680348023454,
      "loss": 2.5714,
      "theoretical_loss": 3.5635734578259557,
      "tokens_seen": 1291059200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003811802534518631,
      "loss": 2.6537,
      "theoretical_loss": 3.5635570467341346,
      "tokens_seen": 1291124736
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003811637034234916,
      "loss": 2.5671,
      "theoretical_loss": 3.563540636708529,
      "tokens_seen": 1291190272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038114715339512015,
      "loss": 2.6682,
      "theoretical_loss": 3.5635242277490162,
      "tokens_seen": 1291255808
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038113060336674863,
      "loss": 2.8096,
      "theoretical_loss": 3.563507819855472,
      "tokens_seen": 1291321344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038111405333837717,
      "loss": 2.6034,
      "theoretical_loss": 3.5634914130277737,
      "tokens_seen": 1291386880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003810975033100057,
      "loss": 2.845,
      "theoretical_loss": 3.5634750072657972,
      "tokens_seen": 1291452416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038108095328163424,
      "loss": 2.8744,
      "theoretical_loss": 3.56345860256942,
      "tokens_seen": 1291517952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038106440325326273,
      "loss": 2.7801,
      "theoretical_loss": 3.5634421989385183,
      "tokens_seen": 1291583488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038104785322489127,
      "loss": 2.631,
      "theoretical_loss": 3.5634257963729694,
      "tokens_seen": 1291649024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003810313031965198,
      "loss": 2.5861,
      "theoretical_loss": 3.5634093948726493,
      "tokens_seen": 1291714560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038101475316814834,
      "loss": 2.4861,
      "theoretical_loss": 3.5633929944374354,
      "tokens_seen": 1291780096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003809982031397768,
      "loss": 2.5569,
      "theoretical_loss": 3.5633765950672043,
      "tokens_seen": 1291845632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038098165311140536,
      "loss": 2.7696,
      "theoretical_loss": 3.563360196761833,
      "tokens_seen": 1291911168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003809651030830338,
      "loss": 2.766,
      "theoretical_loss": 3.5633437995211974,
      "tokens_seen": 1291976704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038094855305466233,
      "loss": 2.5175,
      "theoretical_loss": 3.5633274033451756,
      "tokens_seen": 1292042240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038093200302629087,
      "loss": 2.5108,
      "theoretical_loss": 3.563311008233644,
      "tokens_seen": 1292107776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003809154529979194,
      "loss": 2.6041,
      "theoretical_loss": 3.5632946141864794,
      "tokens_seen": 1292173312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003808989029695479,
      "loss": 2.8054,
      "theoretical_loss": 3.563278221203559,
      "tokens_seen": 1292238848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003808823529411764,
      "loss": 2.7726,
      "theoretical_loss": 3.5632618292847593,
      "tokens_seen": 1292304384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038086580291280496,
      "loss": 2.6658,
      "theoretical_loss": 3.5632454384299574,
      "tokens_seen": 1292369920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003808492528844335,
      "loss": 2.8427,
      "theoretical_loss": 3.563229048639031,
      "tokens_seen": 1292435456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000380832702856062,
      "loss": 2.5905,
      "theoretical_loss": 3.563212659911856,
      "tokens_seen": 1292500992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003808161528276905,
      "loss": 2.6372,
      "theoretical_loss": 3.5631962722483097,
      "tokens_seen": 1292566528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038079960279931906,
      "loss": 2.7932,
      "theoretical_loss": 3.56317988564827,
      "tokens_seen": 1292632064
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003807830527709476,
      "loss": 2.81,
      "theoretical_loss": 3.5631635001116138,
      "tokens_seen": 1292697600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003807665027425761,
      "loss": 2.638,
      "theoretical_loss": 3.5631471156382166,
      "tokens_seen": 1292763136
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003807499527142046,
      "loss": 2.8387,
      "theoretical_loss": 3.5631307322279575,
      "tokens_seen": 1292828672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038073340268583315,
      "loss": 2.8151,
      "theoretical_loss": 3.5631143498807125,
      "tokens_seen": 1292894208
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003807168526574617,
      "loss": 2.8067,
      "theoretical_loss": 3.5630979685963586,
      "tokens_seen": 1292959744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003807003026290902,
      "loss": 2.7866,
      "theoretical_loss": 3.563081588374774,
      "tokens_seen": 1293025280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003806837526007187,
      "loss": 2.6786,
      "theoretical_loss": 3.563065209215835,
      "tokens_seen": 1293090816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038066720257234725,
      "loss": 2.2893,
      "theoretical_loss": 3.5630488311194193,
      "tokens_seen": 1293156352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003806506525439758,
      "loss": 2.7158,
      "theoretical_loss": 3.5630324540854037,
      "tokens_seen": 1293221888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038063410251560427,
      "loss": 2.9317,
      "theoretical_loss": 3.563016078113666,
      "tokens_seen": 1293287424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003806175524872328,
      "loss": 2.5642,
      "theoretical_loss": 3.562999703204083,
      "tokens_seen": 1293352960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038060100245886135,
      "loss": 2.6308,
      "theoretical_loss": 3.5629833293565323,
      "tokens_seen": 1293418496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003805844524304899,
      "loss": 2.3983,
      "theoretical_loss": 3.5629669565708904,
      "tokens_seen": 1293484032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038056790240211837,
      "loss": 2.5964,
      "theoretical_loss": 3.562950584847036,
      "tokens_seen": 1293549568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003805513523737469,
      "loss": 2.8664,
      "theoretical_loss": 3.5629342141848452,
      "tokens_seen": 1293615104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038053480234537544,
      "loss": 2.6034,
      "theoretical_loss": 3.562917844584197,
      "tokens_seen": 1293680640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000380518252317004,
      "loss": 2.4326,
      "theoretical_loss": 3.5629014760449667,
      "tokens_seen": 1293746176
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038050170228863246,
      "loss": 2.6157,
      "theoretical_loss": 3.562885108567033,
      "tokens_seen": 1293811712
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000380485152260261,
      "loss": 2.7849,
      "theoretical_loss": 3.562868742150273,
      "tokens_seen": 1293877248
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038046860223188954,
      "loss": 2.8077,
      "theoretical_loss": 3.5628523767945643,
      "tokens_seen": 1293942784
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003804520522035181,
      "loss": 2.7747,
      "theoretical_loss": 3.5628360124997847,
      "tokens_seen": 1294008320
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003804355021751466,
      "loss": 2.5871,
      "theoretical_loss": 3.562819649265811,
      "tokens_seen": 1294073856
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003804189521467751,
      "loss": 2.6405,
      "theoretical_loss": 3.562803287092521,
      "tokens_seen": 1294139392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038040240211840363,
      "loss": 2.7665,
      "theoretical_loss": 3.5627869259797924,
      "tokens_seen": 1294204928
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038038585209003217,
      "loss": 2.7599,
      "theoretical_loss": 3.562770565927502,
      "tokens_seen": 1294270464
    },
    {
      "epoch": 0.46,
      "objective/train/avg_token_score": 0.017827458679676056,
      "objective/train/avg_weight": 0.9790127277374268,
      "objective/train/docs_used": 732486,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.709364891052246,
      "objective/train/std_weight": 0.11015570163726807,
      "objective/train/theoretical_loss": 3.5627542069355282,
      "objective/train/tokens_used": 1314796000,
      "theoretical_loss": 3.5627542069355282,
      "tokens_seen": 1294336000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003803693020616607,
      "loss": 2.8745,
      "theoretical_loss": 3.5627542069355282,
      "tokens_seen": 1294336000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003803527520332892,
      "loss": 2.4959,
      "theoretical_loss": 3.562737849003749,
      "tokens_seen": 1294401536
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038033620200491773,
      "loss": 2.605,
      "theoretical_loss": 3.5627214921320407,
      "tokens_seen": 1294467072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038031965197654627,
      "loss": 2.5384,
      "theoretical_loss": 3.5627051363202824,
      "tokens_seen": 1294532608
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003803031019481748,
      "loss": 2.5024,
      "theoretical_loss": 3.56268878156835,
      "tokens_seen": 1294598144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003802865519198033,
      "loss": 2.7672,
      "theoretical_loss": 3.562672427876123,
      "tokens_seen": 1294663680
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003802700018914318,
      "loss": 2.7932,
      "theoretical_loss": 3.5626560752434777,
      "tokens_seen": 1294729216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038025345186306036,
      "loss": 2.7585,
      "theoretical_loss": 3.5626397236702934,
      "tokens_seen": 1294794752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003802369018346889,
      "loss": 2.6709,
      "theoretical_loss": 3.5626233731564456,
      "tokens_seen": 1294860288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003802203518063174,
      "loss": 2.8572,
      "theoretical_loss": 3.562607023701814,
      "tokens_seen": 1294925824
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003802038017779459,
      "loss": 2.7613,
      "theoretical_loss": 3.562590675306276,
      "tokens_seen": 1294991360
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038018725174957446,
      "loss": 2.5489,
      "theoretical_loss": 3.562574327969708,
      "tokens_seen": 1295056896
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000380170701721203,
      "loss": 2.7027,
      "theoretical_loss": 3.56255798169199,
      "tokens_seen": 1295122432
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003801541516928315,
      "loss": 2.5276,
      "theoretical_loss": 3.5625416364729983,
      "tokens_seen": 1295187968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038013760166446,
      "loss": 2.2797,
      "theoretical_loss": 3.562525292312611,
      "tokens_seen": 1295253504
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038012105163608855,
      "loss": 2.3777,
      "theoretical_loss": 3.5625089492107063,
      "tokens_seen": 1295319040
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003801045016077171,
      "loss": 2.7242,
      "theoretical_loss": 3.562492607167162,
      "tokens_seen": 1295384576
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003800879515793456,
      "loss": 2.9716,
      "theoretical_loss": 3.5624762661818563,
      "tokens_seen": 1295450112
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003800714015509741,
      "loss": 2.7358,
      "theoretical_loss": 3.5624599262546663,
      "tokens_seen": 1295515648
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038005485152260265,
      "loss": 2.6034,
      "theoretical_loss": 3.562443587385471,
      "tokens_seen": 1295581184
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003800383014942312,
      "loss": 2.6427,
      "theoretical_loss": 3.5624272495741476,
      "tokens_seen": 1295646720
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00038002175146585967,
      "loss": 2.7182,
      "theoretical_loss": 3.5624109128205745,
      "tokens_seen": 1295712256
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003800052014374882,
      "loss": 2.6838,
      "theoretical_loss": 3.5623945771246297,
      "tokens_seen": 1295777792
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037998865140911675,
      "loss": 2.7461,
      "theoretical_loss": 3.562378242486191,
      "tokens_seen": 1295843328
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003799721013807452,
      "loss": 2.485,
      "theoretical_loss": 3.562361908905137,
      "tokens_seen": 1295908864
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003799555513523737,
      "loss": 2.7885,
      "theoretical_loss": 3.562345576381345,
      "tokens_seen": 1295974400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037993900132400225,
      "loss": 2.7572,
      "theoretical_loss": 3.562329244914694,
      "tokens_seen": 1296039936
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037992245129563073,
      "loss": 2.6435,
      "theoretical_loss": 3.5623129145050614,
      "tokens_seen": 1296105472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037990590126725927,
      "loss": 2.4213,
      "theoretical_loss": 3.5622965851523256,
      "tokens_seen": 1296171008
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003798893512388878,
      "loss": 2.6312,
      "theoretical_loss": 3.5622802568563645,
      "tokens_seen": 1296236544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037987280121051635,
      "loss": 2.761,
      "theoretical_loss": 3.562263929617057,
      "tokens_seen": 1296302080
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037985625118214483,
      "loss": 2.305,
      "theoretical_loss": 3.5622476034342805,
      "tokens_seen": 1296367616
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037983970115377337,
      "loss": 2.6239,
      "theoretical_loss": 3.5622312783079133,
      "tokens_seen": 1296433152
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003798231511254019,
      "loss": 2.6946,
      "theoretical_loss": 3.5622149542378345,
      "tokens_seen": 1296498688
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037980660109703044,
      "loss": 2.5516,
      "theoretical_loss": 3.5621986312239216,
      "tokens_seen": 1296564224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003797900510686589,
      "loss": 2.8164,
      "theoretical_loss": 3.562182309266053,
      "tokens_seen": 1296629760
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037977350104028746,
      "loss": 2.3757,
      "theoretical_loss": 3.5621659883641072,
      "tokens_seen": 1296695296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000379756951011916,
      "loss": 2.7243,
      "theoretical_loss": 3.5621496685179626,
      "tokens_seen": 1296760832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037974040098354454,
      "loss": 2.6942,
      "theoretical_loss": 3.562133349727497,
      "tokens_seen": 1296826368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000379723850955173,
      "loss": 2.5438,
      "theoretical_loss": 3.5621170319925897,
      "tokens_seen": 1296891904
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037970730092680156,
      "loss": 2.4807,
      "theoretical_loss": 3.5621007153131172,
      "tokens_seen": 1296957440
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003796907508984301,
      "loss": 2.8728,
      "theoretical_loss": 3.5620843996889606,
      "tokens_seen": 1297022976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037967420087005863,
      "loss": 2.692,
      "theoretical_loss": 3.5620680851199964,
      "tokens_seen": 1297088512
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003796576508416871,
      "loss": 2.5576,
      "theoretical_loss": 3.5620517716061038,
      "tokens_seen": 1297154048
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037964110081331566,
      "loss": 2.4339,
      "theoretical_loss": 3.5620354591471606,
      "tokens_seen": 1297219584
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003796245507849442,
      "loss": 2.4257,
      "theoretical_loss": 3.5620191477430465,
      "tokens_seen": 1297285120
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037960800075657273,
      "loss": 2.7632,
      "theoretical_loss": 3.5620028373936385,
      "tokens_seen": 1297350656
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003795914507282012,
      "loss": 2.4673,
      "theoretical_loss": 3.5619865280988163,
      "tokens_seen": 1297416192
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037957490069982975,
      "loss": 2.7613,
      "theoretical_loss": 3.5619702198584577,
      "tokens_seen": 1297481728
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003795583506714583,
      "loss": 2.5741,
      "theoretical_loss": 3.561953912672442,
      "tokens_seen": 1297547264
    },
    {
      "epoch": 0.46,
      "objective/train/avg_token_score": 0.013868081383407116,
      "objective/train/avg_weight": 0.9836872220039368,
      "objective/train/docs_used": 734540,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.875692844390869,
      "objective/train/std_weight": 0.06703171133995056,
      "objective/train/theoretical_loss": 3.5619376065406474,
      "objective/train/tokens_used": 1318072800,
      "theoretical_loss": 3.5619376065406474,
      "tokens_seen": 1297612800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003795418006430868,
      "loss": 2.7476,
      "theoretical_loss": 3.5619376065406474,
      "tokens_seen": 1297612800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003795252506147153,
      "loss": 2.4435,
      "theoretical_loss": 3.5619213014629523,
      "tokens_seen": 1297678336
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037950870058634385,
      "loss": 2.5984,
      "theoretical_loss": 3.5619049974392354,
      "tokens_seen": 1297743872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003794921505579724,
      "loss": 2.6479,
      "theoretical_loss": 3.5618886944693755,
      "tokens_seen": 1297809408
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003794756005296009,
      "loss": 2.4243,
      "theoretical_loss": 3.5618723925532514,
      "tokens_seen": 1297874944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003794590505012294,
      "loss": 2.7628,
      "theoretical_loss": 3.561856091690742,
      "tokens_seen": 1297940480
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037944250047285794,
      "loss": 2.7607,
      "theoretical_loss": 3.5618397918817246,
      "tokens_seen": 1298006016
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003794259504444865,
      "loss": 2.7415,
      "theoretical_loss": 3.56182349312608,
      "tokens_seen": 1298071552
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000379409400416115,
      "loss": 2.6157,
      "theoretical_loss": 3.561807195423685,
      "tokens_seen": 1298137088
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003793928503877435,
      "loss": 2.8611,
      "theoretical_loss": 3.5617908987744205,
      "tokens_seen": 1298202624
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037937630035937204,
      "loss": 2.7779,
      "theoretical_loss": 3.561774603178163,
      "tokens_seen": 1298268160
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003793597503310006,
      "loss": 2.6482,
      "theoretical_loss": 3.561758308634793,
      "tokens_seen": 1298333696
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003793432003026291,
      "loss": 2.6221,
      "theoretical_loss": 3.5617420151441888,
      "tokens_seen": 1298399232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003793266502742576,
      "loss": 2.5977,
      "theoretical_loss": 3.5617257227062287,
      "tokens_seen": 1298464768
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037931010024588614,
      "loss": 2.6142,
      "theoretical_loss": 3.561709431320792,
      "tokens_seen": 1298530304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037929355021751467,
      "loss": 2.3865,
      "theoretical_loss": 3.5616931409877584,
      "tokens_seen": 1298595840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003792770001891432,
      "loss": 2.6085,
      "theoretical_loss": 3.5616768517070057,
      "tokens_seen": 1298661376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003792604501607717,
      "loss": 2.8153,
      "theoretical_loss": 3.5616605634784126,
      "tokens_seen": 1298726912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037924390013240023,
      "loss": 2.7127,
      "theoretical_loss": 3.561644276301859,
      "tokens_seen": 1298792448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037922735010402877,
      "loss": 2.7632,
      "theoretical_loss": 3.5616279901772234,
      "tokens_seen": 1298857984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003792108000756573,
      "loss": 2.6655,
      "theoretical_loss": 3.5616117051043856,
      "tokens_seen": 1298923520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003791942500472858,
      "loss": 2.5775,
      "theoretical_loss": 3.561595421083223,
      "tokens_seen": 1298989056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003791777000189143,
      "loss": 2.8312,
      "theoretical_loss": 3.5615791381136157,
      "tokens_seen": 1299054592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037916114999054286,
      "loss": 2.5711,
      "theoretical_loss": 3.561562856195443,
      "tokens_seen": 1299120128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003791445999621714,
      "loss": 2.5213,
      "theoretical_loss": 3.5615465753285833,
      "tokens_seen": 1299185664
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003791280499337999,
      "loss": 2.6044,
      "theoretical_loss": 3.561530295512916,
      "tokens_seen": 1299251200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003791114999054284,
      "loss": 2.514,
      "theoretical_loss": 3.56151401674832,
      "tokens_seen": 1299316736
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037909494987705696,
      "loss": 2.546,
      "theoretical_loss": 3.5614977390346745,
      "tokens_seen": 1299382272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003790783998486855,
      "loss": 2.7372,
      "theoretical_loss": 3.5614814623718587,
      "tokens_seen": 1299447808
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000379061849820314,
      "loss": 2.6954,
      "theoretical_loss": 3.5614651867597518,
      "tokens_seen": 1299513344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003790452997919425,
      "loss": 2.5851,
      "theoretical_loss": 3.5614489121982333,
      "tokens_seen": 1299578880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037902874976357106,
      "loss": 2.8524,
      "theoretical_loss": 3.5614326386871817,
      "tokens_seen": 1299644416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003790121997351996,
      "loss": 2.7076,
      "theoretical_loss": 3.561416366226477,
      "tokens_seen": 1299709952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000378995649706828,
      "loss": 2.6609,
      "theoretical_loss": 3.5614000948159976,
      "tokens_seen": 1299775488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037897909967845656,
      "loss": 2.73,
      "theoretical_loss": 3.561383824455624,
      "tokens_seen": 1299841024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003789625496500851,
      "loss": 2.5864,
      "theoretical_loss": 3.5613675551452344,
      "tokens_seen": 1299906560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003789459996217136,
      "loss": 2.8747,
      "theoretical_loss": 3.561351286884708,
      "tokens_seen": 1299972096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003789294495933421,
      "loss": 2.6917,
      "theoretical_loss": 3.5613350196739253,
      "tokens_seen": 1300037632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037891289956497066,
      "loss": 2.5893,
      "theoretical_loss": 3.561318753512764,
      "tokens_seen": 1300103168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003788963495365992,
      "loss": 2.9438,
      "theoretical_loss": 3.561302488401105,
      "tokens_seen": 1300168704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003788797995082277,
      "loss": 2.4316,
      "theoretical_loss": 3.5612862243388275,
      "tokens_seen": 1300234240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003788632494798562,
      "loss": 2.8676,
      "theoretical_loss": 3.5612699613258094,
      "tokens_seen": 1300299776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037884669945148475,
      "loss": 2.5436,
      "theoretical_loss": 3.561253699361932,
      "tokens_seen": 1300365312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003788301494231133,
      "loss": 2.7805,
      "theoretical_loss": 3.561237438447074,
      "tokens_seen": 1300430848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003788135993947418,
      "loss": 2.5567,
      "theoretical_loss": 3.561221178581114,
      "tokens_seen": 1300496384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003787970493663703,
      "loss": 2.8252,
      "theoretical_loss": 3.5612049197639335,
      "tokens_seen": 1300561920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037878049933799885,
      "loss": 2.5103,
      "theoretical_loss": 3.56118866199541,
      "tokens_seen": 1300627456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003787639493096274,
      "loss": 2.7432,
      "theoretical_loss": 3.5611724052754243,
      "tokens_seen": 1300692992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037874739928125587,
      "loss": 2.4621,
      "theoretical_loss": 3.5611561496038555,
      "tokens_seen": 1300758528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003787308492528844,
      "loss": 2.8163,
      "theoretical_loss": 3.5611398949805833,
      "tokens_seen": 1300824064
    },
    {
      "epoch": 0.46,
      "objective/train/avg_token_score": 0.0175258107483387,
      "objective/train/avg_weight": 0.9793941378593445,
      "objective/train/docs_used": 736491,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.443445682525635,
      "objective/train/std_weight": 0.11709152162075043,
      "objective/train/theoretical_loss": 3.5611236414054868,
      "objective/train/tokens_used": 1321349600,
      "theoretical_loss": 3.5611236414054868,
      "tokens_seen": 1300889600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037871429922451294,
      "loss": 2.7855,
      "theoretical_loss": 3.5611236414054868,
      "tokens_seen": 1300889600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003786977491961415,
      "loss": 2.682,
      "theoretical_loss": 3.561107388878446,
      "tokens_seen": 1300955136
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037868119916776997,
      "loss": 2.7943,
      "theoretical_loss": 3.561091137399341,
      "tokens_seen": 1301020672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003786646491393985,
      "loss": 2.633,
      "theoretical_loss": 3.5610748869680506,
      "tokens_seen": 1301086208
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037864809911102704,
      "loss": 2.6194,
      "theoretical_loss": 3.5610586375844546,
      "tokens_seen": 1301151744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003786315490826556,
      "loss": 2.4395,
      "theoretical_loss": 3.561042389248433,
      "tokens_seen": 1301217280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037861499905428406,
      "loss": 2.4377,
      "theoretical_loss": 3.561026141959866,
      "tokens_seen": 1301282816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003785984490259126,
      "loss": 2.7823,
      "theoretical_loss": 3.5610098957186325,
      "tokens_seen": 1301348352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037858189899754114,
      "loss": 2.7907,
      "theoretical_loss": 3.5609936505246123,
      "tokens_seen": 1301413888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003785653489691697,
      "loss": 2.7266,
      "theoretical_loss": 3.5609774063776856,
      "tokens_seen": 1301479424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037854879894079816,
      "loss": 2.5809,
      "theoretical_loss": 3.5609611632777316,
      "tokens_seen": 1301544960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003785322489124267,
      "loss": 2.6059,
      "theoretical_loss": 3.560944921224631,
      "tokens_seen": 1301610496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037851569888405523,
      "loss": 2.8127,
      "theoretical_loss": 3.5609286802182627,
      "tokens_seen": 1301676032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037849914885568377,
      "loss": 2.5968,
      "theoretical_loss": 3.560912440258507,
      "tokens_seen": 1301741568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037848259882731225,
      "loss": 2.6772,
      "theoretical_loss": 3.560896201345243,
      "tokens_seen": 1301807104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0003784660487989408,
      "loss": 2.4709,
      "theoretical_loss": 3.5608799634783526,
      "tokens_seen": 1301872640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00037844949877056933,
      "loss": 2.6371,
      "theoretical_loss": 3.560863726657714,
      "tokens_seen": 1301938176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037843294874219787,
      "loss": 2.7151,
      "theoretical_loss": 3.560847490883207,
      "tokens_seen": 1302003712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037841639871382635,
      "loss": 2.6012,
      "theoretical_loss": 3.560831256154713,
      "tokens_seen": 1302069248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003783998486854549,
      "loss": 2.317,
      "theoretical_loss": 3.5608150224721102,
      "tokens_seen": 1302134784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003783832986570834,
      "loss": 2.7522,
      "theoretical_loss": 3.5607987898352795,
      "tokens_seen": 1302200320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037836674862871196,
      "loss": 2.7211,
      "theoretical_loss": 3.5607825582441013,
      "tokens_seen": 1302265856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037835019860034045,
      "loss": 2.7548,
      "theoretical_loss": 3.5607663276984547,
      "tokens_seen": 1302331392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000378333648571969,
      "loss": 2.6459,
      "theoretical_loss": 3.5607500981982207,
      "tokens_seen": 1302396928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003783170985435975,
      "loss": 2.6036,
      "theoretical_loss": 3.560733869743278,
      "tokens_seen": 1302462464
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037830054851522606,
      "loss": 2.2849,
      "theoretical_loss": 3.5607176423335085,
      "tokens_seen": 1302528000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037828399848685454,
      "loss": 2.732,
      "theoretical_loss": 3.560701415968791,
      "tokens_seen": 1302593536
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003782674484584831,
      "loss": 2.9873,
      "theoretical_loss": 3.560685190649006,
      "tokens_seen": 1302659072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003782508984301116,
      "loss": 2.3929,
      "theoretical_loss": 3.560668966374034,
      "tokens_seen": 1302724608
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037823434840174015,
      "loss": 2.6887,
      "theoretical_loss": 3.560652743143754,
      "tokens_seen": 1302790144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037821779837336864,
      "loss": 2.5366,
      "theoretical_loss": 3.560636520958047,
      "tokens_seen": 1302855680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003782012483449972,
      "loss": 2.2947,
      "theoretical_loss": 3.5606202998167937,
      "tokens_seen": 1302921216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003781846983166257,
      "loss": 2.8556,
      "theoretical_loss": 3.560604079719874,
      "tokens_seen": 1302986752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037816814828825425,
      "loss": 2.7081,
      "theoretical_loss": 3.5605878606671677,
      "tokens_seen": 1303052288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037815159825988273,
      "loss": 2.6553,
      "theoretical_loss": 3.560571642658555,
      "tokens_seen": 1303117824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037813504823151127,
      "loss": 2.5763,
      "theoretical_loss": 3.5605554256939165,
      "tokens_seen": 1303183360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003781184982031398,
      "loss": 2.5582,
      "theoretical_loss": 3.560539209773133,
      "tokens_seen": 1303248896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037810194817476835,
      "loss": 2.6267,
      "theoretical_loss": 3.560522994896084,
      "tokens_seen": 1303314432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037808539814639683,
      "loss": 2.9322,
      "theoretical_loss": 3.5605067810626503,
      "tokens_seen": 1303379968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037806884811802537,
      "loss": 2.9186,
      "theoretical_loss": 3.560490568272712,
      "tokens_seen": 1303445504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003780522980896539,
      "loss": 2.8536,
      "theoretical_loss": 3.5604743565261496,
      "tokens_seen": 1303511040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037803574806128244,
      "loss": 2.6695,
      "theoretical_loss": 3.5604581458228437,
      "tokens_seen": 1303576576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037801919803291087,
      "loss": 2.7201,
      "theoretical_loss": 3.5604419361626745,
      "tokens_seen": 1303642112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003780026480045394,
      "loss": 2.8033,
      "theoretical_loss": 3.560425727545522,
      "tokens_seen": 1303707648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037798609797616795,
      "loss": 2.8238,
      "theoretical_loss": 3.5604095199712678,
      "tokens_seen": 1303773184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037796954794779643,
      "loss": 2.6642,
      "theoretical_loss": 3.560393313439791,
      "tokens_seen": 1303838720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037795299791942497,
      "loss": 2.7001,
      "theoretical_loss": 3.5603771079509734,
      "tokens_seen": 1303904256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003779364478910535,
      "loss": 2.6919,
      "theoretical_loss": 3.5603609035046944,
      "tokens_seen": 1303969792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037791989786268204,
      "loss": 2.742,
      "theoretical_loss": 3.5603447001008357,
      "tokens_seen": 1304035328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003779033478343105,
      "loss": 2.3561,
      "theoretical_loss": 3.5603284977392766,
      "tokens_seen": 1304100864
    },
    {
      "epoch": 0.47,
      "objective/train/avg_token_score": 0.004544904455542564,
      "objective/train/avg_weight": 0.9946537613868713,
      "objective/train/docs_used": 738318,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0811948776245117,
      "objective/train/std_weight": 0.020697496831417084,
      "objective/train/theoretical_loss": 3.560312296419899,
      "objective/train/tokens_used": 1324626400,
      "theoretical_loss": 3.560312296419899,
      "tokens_seen": 1304166400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037788679780593906,
      "loss": 2.9228,
      "theoretical_loss": 3.560312296419899,
      "tokens_seen": 1304166400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003778702477775676,
      "loss": 2.6599,
      "theoretical_loss": 3.560296096142582,
      "tokens_seen": 1304231936
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037785369774919614,
      "loss": 2.8793,
      "theoretical_loss": 3.5602798969072076,
      "tokens_seen": 1304297472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003778371477208246,
      "loss": 2.8325,
      "theoretical_loss": 3.5602636987136558,
      "tokens_seen": 1304363008
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037782059769245316,
      "loss": 2.7201,
      "theoretical_loss": 3.5602475015618067,
      "tokens_seen": 1304428544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003778040476640817,
      "loss": 2.5471,
      "theoretical_loss": 3.560231305451542,
      "tokens_seen": 1304494080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037778749763571023,
      "loss": 2.6949,
      "theoretical_loss": 3.5602151103827424,
      "tokens_seen": 1304559616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003777709476073387,
      "loss": 2.8838,
      "theoretical_loss": 3.5601989163552883,
      "tokens_seen": 1304625152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037775439757896725,
      "loss": 2.6188,
      "theoretical_loss": 3.56018272336906,
      "tokens_seen": 1304690688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003777378475505958,
      "loss": 2.7237,
      "theoretical_loss": 3.5601665314239384,
      "tokens_seen": 1304756224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037772129752222433,
      "loss": 2.7144,
      "theoretical_loss": 3.5601503405198045,
      "tokens_seen": 1304821760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003777047474938528,
      "loss": 2.7164,
      "theoretical_loss": 3.560134150656539,
      "tokens_seen": 1304887296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037768819746548135,
      "loss": 2.6741,
      "theoretical_loss": 3.560117961834023,
      "tokens_seen": 1304952832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003776716474371099,
      "loss": 2.7489,
      "theoretical_loss": 3.5601017740521366,
      "tokens_seen": 1305018368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003776550974087384,
      "loss": 2.7318,
      "theoretical_loss": 3.5600855873107617,
      "tokens_seen": 1305083904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003776385473803669,
      "loss": 2.8594,
      "theoretical_loss": 3.5600694016097787,
      "tokens_seen": 1305149440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037762199735199545,
      "loss": 2.7016,
      "theoretical_loss": 3.5600532169490684,
      "tokens_seen": 1305214976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000377605447323624,
      "loss": 2.9611,
      "theoretical_loss": 3.560037033328512,
      "tokens_seen": 1305280512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003775888972952525,
      "loss": 2.8285,
      "theoretical_loss": 3.560020850747989,
      "tokens_seen": 1305346048
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000377572347266881,
      "loss": 2.5502,
      "theoretical_loss": 3.560004669207383,
      "tokens_seen": 1305411584
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037755579723850954,
      "loss": 2.835,
      "theoretical_loss": 3.5599884887065727,
      "tokens_seen": 1305477120
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003775392472101381,
      "loss": 2.4904,
      "theoretical_loss": 3.55997230924544,
      "tokens_seen": 1305542656
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003775226971817666,
      "loss": 2.34,
      "theoretical_loss": 3.5599561308238656,
      "tokens_seen": 1305608192
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003775061471533951,
      "loss": 2.6829,
      "theoretical_loss": 3.559939953441731,
      "tokens_seen": 1305673728
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037748959712502364,
      "loss": 2.5116,
      "theoretical_loss": 3.559923777098917,
      "tokens_seen": 1305739264
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003774730470966522,
      "loss": 2.7721,
      "theoretical_loss": 3.5599076017953046,
      "tokens_seen": 1305804800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003774564970682807,
      "loss": 2.871,
      "theoretical_loss": 3.5598914275307747,
      "tokens_seen": 1305870336
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003774399470399092,
      "loss": 2.7743,
      "theoretical_loss": 3.5598752543052092,
      "tokens_seen": 1305935872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037742339701153773,
      "loss": 2.6906,
      "theoretical_loss": 3.5598590821184883,
      "tokens_seen": 1306001408
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037740684698316627,
      "loss": 2.6342,
      "theoretical_loss": 3.5598429109704934,
      "tokens_seen": 1306066944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003773902969547948,
      "loss": 2.6818,
      "theoretical_loss": 3.559826740861106,
      "tokens_seen": 1306132480
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003773737469264233,
      "loss": 2.688,
      "theoretical_loss": 3.5598105717902073,
      "tokens_seen": 1306198016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037735719689805183,
      "loss": 2.5434,
      "theoretical_loss": 3.559794403757678,
      "tokens_seen": 1306263552
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037734064686968037,
      "loss": 2.8907,
      "theoretical_loss": 3.5597782367633997,
      "tokens_seen": 1306329088
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003773240968413089,
      "loss": 2.6861,
      "theoretical_loss": 3.5597620708072535,
      "tokens_seen": 1306394624
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003773075468129374,
      "loss": 2.7121,
      "theoretical_loss": 3.5597459058891205,
      "tokens_seen": 1306460160
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003772909967845659,
      "loss": 2.1798,
      "theoretical_loss": 3.5597297420088827,
      "tokens_seen": 1306525696
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037727444675619446,
      "loss": 2.8391,
      "theoretical_loss": 3.559713579166421,
      "tokens_seen": 1306591232
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000377257896727823,
      "loss": 2.5372,
      "theoretical_loss": 3.5596974173616163,
      "tokens_seen": 1306656768
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003772413466994515,
      "loss": 2.5375,
      "theoretical_loss": 3.55968125659435,
      "tokens_seen": 1306722304
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037722479667108,
      "loss": 2.5939,
      "theoretical_loss": 3.559665096864504,
      "tokens_seen": 1306787840
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037720824664270856,
      "loss": 2.6734,
      "theoretical_loss": 3.5596489381719594,
      "tokens_seen": 1306853376
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003771916966143371,
      "loss": 2.6966,
      "theoretical_loss": 3.5596327805165977,
      "tokens_seen": 1306918912
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003771751465859656,
      "loss": 2.4326,
      "theoretical_loss": 3.5596166238983002,
      "tokens_seen": 1306984448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003771585965575941,
      "loss": 2.1036,
      "theoretical_loss": 3.5596004683169484,
      "tokens_seen": 1307049984
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037714204652922266,
      "loss": 2.474,
      "theoretical_loss": 3.5595843137724237,
      "tokens_seen": 1307115520
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003771254965008512,
      "loss": 2.8863,
      "theoretical_loss": 3.5595681602646074,
      "tokens_seen": 1307181056
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003771089464724797,
      "loss": 2.7212,
      "theoretical_loss": 3.559552007793382,
      "tokens_seen": 1307246592
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003770923964441082,
      "loss": 2.8818,
      "theoretical_loss": 3.5595358563586275,
      "tokens_seen": 1307312128
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037707584641573675,
      "loss": 2.5882,
      "theoretical_loss": 3.559519705960226,
      "tokens_seen": 1307377664
    },
    {
      "epoch": 0.47,
      "objective/train/avg_token_score": 0.013016336597502232,
      "objective/train/avg_weight": 0.9846771359443665,
      "objective/train/docs_used": 740377,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1253433227539062,
      "objective/train/std_weight": 0.09036663919687271,
      "objective/train/theoretical_loss": 3.55950355659806,
      "objective/train/tokens_used": 1327903200,
      "theoretical_loss": 3.55950355659806,
      "tokens_seen": 1307443200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003770592963873653,
      "loss": 2.7917,
      "theoretical_loss": 3.55950355659806,
      "tokens_seen": 1307443200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037704274635899377,
      "loss": 2.6301,
      "theoretical_loss": 3.55948740827201,
      "tokens_seen": 1307508736
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037702619633062226,
      "loss": 2.7133,
      "theoretical_loss": 3.5594712609819577,
      "tokens_seen": 1307574272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037700964630225074,
      "loss": 2.9027,
      "theoretical_loss": 3.5594551147277853,
      "tokens_seen": 1307639808
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003769930962738793,
      "loss": 2.68,
      "theoretical_loss": 3.5594389695093738,
      "tokens_seen": 1307705344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003769765462455078,
      "loss": 2.6819,
      "theoretical_loss": 3.559422825326605,
      "tokens_seen": 1307770880
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037695999621713635,
      "loss": 2.8035,
      "theoretical_loss": 3.559406682179361,
      "tokens_seen": 1307836416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037694344618876484,
      "loss": 2.6689,
      "theoretical_loss": 3.5593905400675228,
      "tokens_seen": 1307901952
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003769268961603934,
      "loss": 2.6566,
      "theoretical_loss": 3.5593743989909727,
      "tokens_seen": 1307967488
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003769103461320219,
      "loss": 2.448,
      "theoretical_loss": 3.559358258949592,
      "tokens_seen": 1308033024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037689379610365045,
      "loss": 2.8424,
      "theoretical_loss": 3.5593421199432633,
      "tokens_seen": 1308098560
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037687724607527893,
      "loss": 2.5349,
      "theoretical_loss": 3.5593259819718677,
      "tokens_seen": 1308164096
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037686069604690747,
      "loss": 2.7031,
      "theoretical_loss": 3.5593098450352865,
      "tokens_seen": 1308229632
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000376844146018536,
      "loss": 2.7251,
      "theoretical_loss": 3.559293709133402,
      "tokens_seen": 1308295168
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037682759599016454,
      "loss": 2.3664,
      "theoretical_loss": 3.5592775742660967,
      "tokens_seen": 1308360704
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037681104596179303,
      "loss": 2.9072,
      "theoretical_loss": 3.5592614404332514,
      "tokens_seen": 1308426240
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037679449593342156,
      "loss": 2.6012,
      "theoretical_loss": 3.5592453076347486,
      "tokens_seen": 1308491776
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003767779459050501,
      "loss": 2.5953,
      "theoretical_loss": 3.55922917587047,
      "tokens_seen": 1308557312
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037676139587667864,
      "loss": 2.5173,
      "theoretical_loss": 3.5592130451402975,
      "tokens_seen": 1308622848
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003767448458483071,
      "loss": 2.765,
      "theoretical_loss": 3.5591969154441125,
      "tokens_seen": 1308688384
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037672829581993566,
      "loss": 2.5468,
      "theoretical_loss": 3.5591807867817984,
      "tokens_seen": 1308753920
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003767117457915642,
      "loss": 2.6441,
      "theoretical_loss": 3.559164659153235,
      "tokens_seen": 1308819456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037669519576319274,
      "loss": 2.5575,
      "theoretical_loss": 3.5591485325583068,
      "tokens_seen": 1308884992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003766786457348212,
      "loss": 2.5955,
      "theoretical_loss": 3.559132406996894,
      "tokens_seen": 1308950528
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037666209570644976,
      "loss": 2.6706,
      "theoretical_loss": 3.559116282468879,
      "tokens_seen": 1309016064
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003766455456780783,
      "loss": 2.5899,
      "theoretical_loss": 3.559100158974144,
      "tokens_seen": 1309081600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037662899564970683,
      "loss": 2.7966,
      "theoretical_loss": 3.559084036512571,
      "tokens_seen": 1309147136
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003766124456213353,
      "loss": 2.7472,
      "theoretical_loss": 3.5590679150840425,
      "tokens_seen": 1309212672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037659589559296385,
      "loss": 2.5973,
      "theoretical_loss": 3.5590517946884406,
      "tokens_seen": 1309278208
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003765793455645924,
      "loss": 2.7217,
      "theoretical_loss": 3.559035675325646,
      "tokens_seen": 1309343744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037656279553622093,
      "loss": 2.5462,
      "theoretical_loss": 3.5590195569955423,
      "tokens_seen": 1309409280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003765462455078494,
      "loss": 2.3569,
      "theoretical_loss": 3.5590034396980115,
      "tokens_seen": 1309474816
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037652969547947795,
      "loss": 2.7066,
      "theoretical_loss": 3.5589873234329357,
      "tokens_seen": 1309540352
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003765131454511065,
      "loss": 2.6157,
      "theoretical_loss": 3.5589712082001963,
      "tokens_seen": 1309605888
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000376496595422735,
      "loss": 2.3039,
      "theoretical_loss": 3.5589550939996766,
      "tokens_seen": 1309671424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037648004539436356,
      "loss": 2.7845,
      "theoretical_loss": 3.5589389808312584,
      "tokens_seen": 1309736960
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037646349536599204,
      "loss": 2.5749,
      "theoretical_loss": 3.5589228686948235,
      "tokens_seen": 1309802496
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003764469453376206,
      "loss": 2.7873,
      "theoretical_loss": 3.5589067575902553,
      "tokens_seen": 1309868032
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003764303953092491,
      "loss": 2.3998,
      "theoretical_loss": 3.558890647517435,
      "tokens_seen": 1309933568
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037641384528087766,
      "loss": 2.9234,
      "theoretical_loss": 3.5588745384762452,
      "tokens_seen": 1309999104
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037639729525250614,
      "loss": 2.4631,
      "theoretical_loss": 3.558858430466569,
      "tokens_seen": 1310064640
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003763807452241347,
      "loss": 2.745,
      "theoretical_loss": 3.558842323488287,
      "tokens_seen": 1310130176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003763641951957632,
      "loss": 2.7598,
      "theoretical_loss": 3.5588262175412835,
      "tokens_seen": 1310195712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037634764516739175,
      "loss": 2.6648,
      "theoretical_loss": 3.55881011262544,
      "tokens_seen": 1310261248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037633109513902024,
      "loss": 2.7247,
      "theoretical_loss": 3.558794008740639,
      "tokens_seen": 1310326784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003763145451106488,
      "loss": 2.685,
      "theoretical_loss": 3.558777905886763,
      "tokens_seen": 1310392320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003762979950822773,
      "loss": 2.7165,
      "theoretical_loss": 3.5587618040636944,
      "tokens_seen": 1310457856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037628144505390585,
      "loss": 2.7147,
      "theoretical_loss": 3.558745703271315,
      "tokens_seen": 1310523392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037626489502553433,
      "loss": 2.6342,
      "theoretical_loss": 3.5587296035095086,
      "tokens_seen": 1310588928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037624834499716287,
      "loss": 2.7389,
      "theoretical_loss": 3.558713504778156,
      "tokens_seen": 1310654464
    },
    {
      "debugging/Self-BLEU-5": 0.49020908264157476,
      "debugging/distinct-1-grams": 0.768901113497886,
      "debugging/distinct-2-grams": 0.9428782333551957,
      "debugging/entropy-1-grams": 6.085999550681761,
      "debugging/entropy-2-grams": 7.0033060167714964,
      "debugging/length": 490.2352941176471,
      "debugging/num_segments": 17,
      "epoch": 0.47,
      "objective/train/avg_token_score": 0.02056093141436577,
      "objective/train/avg_weight": 0.9757911562919617,
      "objective/train/docs_used": 741674,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 6.077814102172852,
      "objective/train/std_weight": 0.12924915552139282,
      "objective/train/theoretical_loss": 3.558697407077142,
      "objective/train/tokens_used": 1331180000,
      "theoretical_loss": 3.558697407077142,
      "tokens_seen": 1310720000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003762317949687914,
      "loss": 2.4445,
      "theoretical_loss": 3.558697407077142,
      "tokens_seen": 1310720000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037621524494041995,
      "loss": 2.6912,
      "theoretical_loss": 3.5586813104063477,
      "tokens_seen": 1310785536
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037619869491204843,
      "loss": 2.5335,
      "theoretical_loss": 3.5586652147656554,
      "tokens_seen": 1310851072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037618214488367697,
      "loss": 2.8432,
      "theoretical_loss": 3.558649120154948,
      "tokens_seen": 1310916608
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003761655948553055,
      "loss": 2.5599,
      "theoretical_loss": 3.5586330265741086,
      "tokens_seen": 1310982144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037614904482693404,
      "loss": 2.6265,
      "theoretical_loss": 3.5586169340230196,
      "tokens_seen": 1311047680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003761324947985625,
      "loss": 2.5731,
      "theoretical_loss": 3.5586008425015634,
      "tokens_seen": 1311113216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037611594477019106,
      "loss": 2.743,
      "theoretical_loss": 3.5585847520096228,
      "tokens_seen": 1311178752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003760993947418196,
      "loss": 2.8605,
      "theoretical_loss": 3.5585686625470805,
      "tokens_seen": 1311244288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037608284471344814,
      "loss": 2.8228,
      "theoretical_loss": 3.558552574113819,
      "tokens_seen": 1311309824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003760662946850766,
      "loss": 2.7314,
      "theoretical_loss": 3.558536486709721,
      "tokens_seen": 1311375360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003760497446567051,
      "loss": 2.607,
      "theoretical_loss": 3.55852040033467,
      "tokens_seen": 1311440896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003760331946283336,
      "loss": 2.6388,
      "theoretical_loss": 3.5585043149885474,
      "tokens_seen": 1311506432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003760166445999621,
      "loss": 2.6125,
      "theoretical_loss": 3.5584882306712373,
      "tokens_seen": 1311571968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037600009457159066,
      "loss": 2.2913,
      "theoretical_loss": 3.5584721473826217,
      "tokens_seen": 1311637504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003759835445432192,
      "loss": 2.8782,
      "theoretical_loss": 3.5584560651225834,
      "tokens_seen": 1311703040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003759669945148477,
      "loss": 2.6541,
      "theoretical_loss": 3.558439983891006,
      "tokens_seen": 1311768576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003759504444864762,
      "loss": 2.4995,
      "theoretical_loss": 3.558423903687772,
      "tokens_seen": 1311834112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037593389445810476,
      "loss": 2.6958,
      "theoretical_loss": 3.5584078245127637,
      "tokens_seen": 1311899648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003759173444297333,
      "loss": 2.7005,
      "theoretical_loss": 3.5583917463658645,
      "tokens_seen": 1311965184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003759007944013618,
      "loss": 2.5826,
      "theoretical_loss": 3.558375669246957,
      "tokens_seen": 1312030720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003758842443729903,
      "loss": 2.4873,
      "theoretical_loss": 3.5583595931559246,
      "tokens_seen": 1312096256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037586769434461885,
      "loss": 2.8585,
      "theoretical_loss": 3.5583435180926495,
      "tokens_seen": 1312161792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003758511443162474,
      "loss": 2.5499,
      "theoretical_loss": 3.558327444057016,
      "tokens_seen": 1312227328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003758345942878759,
      "loss": 2.8049,
      "theoretical_loss": 3.558311371048906,
      "tokens_seen": 1312292864
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003758180442595044,
      "loss": 2.5224,
      "theoretical_loss": 3.5582952990682024,
      "tokens_seen": 1312358400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037580149423113295,
      "loss": 2.8764,
      "theoretical_loss": 3.5582792281147886,
      "tokens_seen": 1312423936
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003757849442027615,
      "loss": 2.69,
      "theoretical_loss": 3.5582631581885478,
      "tokens_seen": 1312489472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037576839417438997,
      "loss": 2.8688,
      "theoretical_loss": 3.5582470892893627,
      "tokens_seen": 1312555008
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003757518441460185,
      "loss": 2.8044,
      "theoretical_loss": 3.558231021417117,
      "tokens_seen": 1312620544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037573529411764705,
      "loss": 2.5092,
      "theoretical_loss": 3.5582149545716932,
      "tokens_seen": 1312686080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003757187440892756,
      "loss": 2.5197,
      "theoretical_loss": 3.5581988887529743,
      "tokens_seen": 1312751616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037570219406090407,
      "loss": 2.8618,
      "theoretical_loss": 3.558182823960844,
      "tokens_seen": 1312817152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003756856440325326,
      "loss": 2.5115,
      "theoretical_loss": 3.558166760195185,
      "tokens_seen": 1312882688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037566909400416114,
      "loss": 2.393,
      "theoretical_loss": 3.558150697455881,
      "tokens_seen": 1312948224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003756525439757897,
      "loss": 2.9381,
      "theoretical_loss": 3.5581346357428147,
      "tokens_seen": 1313013760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037563599394741816,
      "loss": 2.5936,
      "theoretical_loss": 3.5581185750558695,
      "tokens_seen": 1313079296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003756194439190467,
      "loss": 2.9079,
      "theoretical_loss": 3.5581025153949284,
      "tokens_seen": 1313144832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037560289389067524,
      "loss": 2.686,
      "theoretical_loss": 3.558086456759875,
      "tokens_seen": 1313210368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003755863438623038,
      "loss": 2.4745,
      "theoretical_loss": 3.5580703991505924,
      "tokens_seen": 1313275904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037556979383393226,
      "loss": 2.6501,
      "theoretical_loss": 3.5580543425669644,
      "tokens_seen": 1313341440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003755532438055608,
      "loss": 2.6822,
      "theoretical_loss": 3.5580382870088734,
      "tokens_seen": 1313406976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037553669377718933,
      "loss": 2.742,
      "theoretical_loss": 3.5580222324762034,
      "tokens_seen": 1313472512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037552014374881787,
      "loss": 2.8074,
      "theoretical_loss": 3.5580061789688378,
      "tokens_seen": 1313538048
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037550359372044636,
      "loss": 2.7911,
      "theoretical_loss": 3.557990126486659,
      "tokens_seen": 1313603584
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003754870436920749,
      "loss": 2.9143,
      "theoretical_loss": 3.5579740750295517,
      "tokens_seen": 1313669120
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037547049366370343,
      "loss": 2.6075,
      "theoretical_loss": 3.5579580245973985,
      "tokens_seen": 1313734656
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037545394363533197,
      "loss": 2.7885,
      "theoretical_loss": 3.557941975190083,
      "tokens_seen": 1313800192
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037543739360696045,
      "loss": 2.8225,
      "theoretical_loss": 3.5579259268074885,
      "tokens_seen": 1313865728
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000375420843578589,
      "loss": 2.5783,
      "theoretical_loss": 3.557909879449499,
      "tokens_seen": 1313931264
    },
    {
      "epoch": 0.47,
      "objective/train/avg_token_score": 0.016822705045342445,
      "objective/train/avg_weight": 0.980201244354248,
      "objective/train/docs_used": 743814,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6687989234924316,
      "objective/train/std_weight": 0.07436075061559677,
      "objective/train/theoretical_loss": 3.5578938331159975,
      "objective/train/tokens_used": 1334456800,
      "theoretical_loss": 3.5578938331159975,
      "tokens_seen": 1313996800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003754042935502175,
      "loss": 2.5709,
      "theoretical_loss": 3.5578938331159975,
      "tokens_seen": 1313996800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037538774352184606,
      "loss": 2.8106,
      "theoretical_loss": 3.5578777878068677,
      "tokens_seen": 1314062336
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037537119349347455,
      "loss": 2.877,
      "theoretical_loss": 3.5578617435219932,
      "tokens_seen": 1314127872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003753546434651031,
      "loss": 2.6393,
      "theoretical_loss": 3.557845700261257,
      "tokens_seen": 1314193408
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003753380934367316,
      "loss": 2.3588,
      "theoretical_loss": 3.557829658024543,
      "tokens_seen": 1314258944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037532154340836016,
      "loss": 2.527,
      "theoretical_loss": 3.5578136168117354,
      "tokens_seen": 1314324480
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037530499337998864,
      "loss": 2.5042,
      "theoretical_loss": 3.5577975766227166,
      "tokens_seen": 1314390016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003752884433516172,
      "loss": 2.3891,
      "theoretical_loss": 3.5577815374573714,
      "tokens_seen": 1314455552
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003752718933232457,
      "loss": 2.734,
      "theoretical_loss": 3.5577654993155825,
      "tokens_seen": 1314521088
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037525534329487426,
      "loss": 2.5466,
      "theoretical_loss": 3.5577494621972345,
      "tokens_seen": 1314586624
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037523879326650274,
      "loss": 2.521,
      "theoretical_loss": 3.55773342610221,
      "tokens_seen": 1314652160
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003752222432381313,
      "loss": 2.5138,
      "theoretical_loss": 3.557717391030393,
      "tokens_seen": 1314717696
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003752056932097598,
      "loss": 2.6174,
      "theoretical_loss": 3.557701356981668,
      "tokens_seen": 1314783232
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037518914318138835,
      "loss": 2.529,
      "theoretical_loss": 3.5576853239559183,
      "tokens_seen": 1314848768
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037517259315301683,
      "loss": 2.7565,
      "theoretical_loss": 3.557669291953027,
      "tokens_seen": 1314914304
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037515604312464537,
      "loss": 2.3547,
      "theoretical_loss": 3.5576532609728786,
      "tokens_seen": 1314979840
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003751394930962739,
      "loss": 2.6177,
      "theoretical_loss": 3.557637231015357,
      "tokens_seen": 1315045376
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037512294306790245,
      "loss": 2.6849,
      "theoretical_loss": 3.5576212020803446,
      "tokens_seen": 1315110912
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037510639303953093,
      "loss": 2.6872,
      "theoretical_loss": 3.557605174167727,
      "tokens_seen": 1315176448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037508984301115947,
      "loss": 2.8415,
      "theoretical_loss": 3.5575891472773877,
      "tokens_seen": 1315241984
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000375073292982788,
      "loss": 2.5697,
      "theoretical_loss": 3.5575731214092094,
      "tokens_seen": 1315307520
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037505674295441644,
      "loss": 2.5744,
      "theoretical_loss": 3.5575570965630776,
      "tokens_seen": 1315373056
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037504019292604497,
      "loss": 2.687,
      "theoretical_loss": 3.557541072738875,
      "tokens_seen": 1315438592
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003750236428976735,
      "loss": 2.6544,
      "theoretical_loss": 3.557525049936486,
      "tokens_seen": 1315504128
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037500709286930205,
      "loss": 2.6438,
      "theoretical_loss": 3.5575090281557946,
      "tokens_seen": 1315569664
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037499054284093053,
      "loss": 2.388,
      "theoretical_loss": 3.557493007396684,
      "tokens_seen": 1315635200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037497399281255907,
      "loss": 2.6945,
      "theoretical_loss": 3.5574769876590397,
      "tokens_seen": 1315700736
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003749574427841876,
      "loss": 2.5665,
      "theoretical_loss": 3.5574609689427446,
      "tokens_seen": 1315766272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037494089275581614,
      "loss": 2.7581,
      "theoretical_loss": 3.5574449512476822,
      "tokens_seen": 1315831808
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003749243427274446,
      "loss": 2.5655,
      "theoretical_loss": 3.5574289345737378,
      "tokens_seen": 1315897344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037490779269907316,
      "loss": 2.6319,
      "theoretical_loss": 3.557412918920795,
      "tokens_seen": 1315962880
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003748912426707017,
      "loss": 2.5464,
      "theoretical_loss": 3.5573969042887374,
      "tokens_seen": 1316028416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037487469264233024,
      "loss": 2.5837,
      "theoretical_loss": 3.55738089067745,
      "tokens_seen": 1316093952
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003748581426139587,
      "loss": 2.7773,
      "theoretical_loss": 3.557364878086816,
      "tokens_seen": 1316159488
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037484159258558726,
      "loss": 2.7777,
      "theoretical_loss": 3.55734886651672,
      "tokens_seen": 1316225024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003748250425572158,
      "loss": 2.7944,
      "theoretical_loss": 3.5573328559670463,
      "tokens_seen": 1316290560
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037480849252884434,
      "loss": 2.5832,
      "theoretical_loss": 3.5573168464376783,
      "tokens_seen": 1316356096
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003747919425004728,
      "loss": 2.7585,
      "theoretical_loss": 3.557300837928501,
      "tokens_seen": 1316421632
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037477539247210136,
      "loss": 2.5239,
      "theoretical_loss": 3.5572848304393982,
      "tokens_seen": 1316487168
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003747588424437299,
      "loss": 2.6604,
      "theoretical_loss": 3.5572688239702543,
      "tokens_seen": 1316552704
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037474229241535843,
      "loss": 2.7001,
      "theoretical_loss": 3.5572528185209538,
      "tokens_seen": 1316618240
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003747257423869869,
      "loss": 2.5773,
      "theoretical_loss": 3.55723681409138,
      "tokens_seen": 1316683776
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037470919235861545,
      "loss": 2.8887,
      "theoretical_loss": 3.5572208106814185,
      "tokens_seen": 1316749312
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000374692642330244,
      "loss": 2.6411,
      "theoretical_loss": 3.557204808290953,
      "tokens_seen": 1316814848
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037467609230187253,
      "loss": 2.7908,
      "theoretical_loss": 3.557188806919867,
      "tokens_seen": 1316880384
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000374659542273501,
      "loss": 2.6635,
      "theoretical_loss": 3.5571728065680457,
      "tokens_seen": 1316945920
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037464299224512955,
      "loss": 2.7413,
      "theoretical_loss": 3.5571568072353736,
      "tokens_seen": 1317011456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003746264422167581,
      "loss": 2.6042,
      "theoretical_loss": 3.5571408089217353,
      "tokens_seen": 1317076992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003746098921883866,
      "loss": 2.8567,
      "theoretical_loss": 3.557124811627014,
      "tokens_seen": 1317142528
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003745933421600151,
      "loss": 2.582,
      "theoretical_loss": 3.5571088153510955,
      "tokens_seen": 1317208064
    },
    {
      "epoch": 0.47,
      "objective/train/avg_token_score": 0.009160229004919529,
      "objective/train/avg_weight": 0.9892166256904602,
      "objective/train/docs_used": 745506,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0793232917785645,
      "objective/train/std_weight": 0.06002482771873474,
      "objective/train/theoretical_loss": 3.557092820093863,
      "objective/train/tokens_used": 1337733600,
      "theoretical_loss": 3.557092820093863,
      "tokens_seen": 1317273600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037457679213164364,
      "loss": 3.0478,
      "theoretical_loss": 3.557092820093863,
      "tokens_seen": 1317273600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003745602421032722,
      "loss": 2.5195,
      "theoretical_loss": 3.5570768258552015,
      "tokens_seen": 1317339136
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003745436920749007,
      "loss": 2.8322,
      "theoretical_loss": 3.5570608326349955,
      "tokens_seen": 1317404672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003745271420465292,
      "loss": 2.7297,
      "theoretical_loss": 3.5570448404331296,
      "tokens_seen": 1317470208
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037451059201815774,
      "loss": 2.6823,
      "theoretical_loss": 3.5570288492494884,
      "tokens_seen": 1317535744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003744940419897863,
      "loss": 2.7273,
      "theoretical_loss": 3.5570128590839563,
      "tokens_seen": 1317601280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003744774919614148,
      "loss": 2.6549,
      "theoretical_loss": 3.5569968699364174,
      "tokens_seen": 1317666816
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003744609419330433,
      "loss": 2.7464,
      "theoretical_loss": 3.5569808818067568,
      "tokens_seen": 1317732352
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037444439190467184,
      "loss": 2.4749,
      "theoretical_loss": 3.5569648946948584,
      "tokens_seen": 1317797888
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003744278418763004,
      "loss": 2.6551,
      "theoretical_loss": 3.556948908600608,
      "tokens_seen": 1317863424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003744112918479289,
      "loss": 2.2883,
      "theoretical_loss": 3.556932923523889,
      "tokens_seen": 1317928960
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003743947418195574,
      "loss": 2.4396,
      "theoretical_loss": 3.556916939464587,
      "tokens_seen": 1317994496
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037437819179118593,
      "loss": 2.7529,
      "theoretical_loss": 3.556900956422586,
      "tokens_seen": 1318060032
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037436164176281447,
      "loss": 2.4902,
      "theoretical_loss": 3.5568849743977706,
      "tokens_seen": 1318125568
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000374345091734443,
      "loss": 2.4869,
      "theoretical_loss": 3.5568689933900264,
      "tokens_seen": 1318191104
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003743285417060715,
      "loss": 2.6151,
      "theoretical_loss": 3.5568530133992375,
      "tokens_seen": 1318256640
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037431199167770003,
      "loss": 2.8421,
      "theoretical_loss": 3.556837034425288,
      "tokens_seen": 1318322176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037429544164932857,
      "loss": 2.8377,
      "theoretical_loss": 3.556821056468064,
      "tokens_seen": 1318387712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003742788916209571,
      "loss": 2.7818,
      "theoretical_loss": 3.5568050795274493,
      "tokens_seen": 1318453248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003742623415925856,
      "loss": 2.6489,
      "theoretical_loss": 3.556789103603329,
      "tokens_seen": 1318518784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003742457915642141,
      "loss": 2.7978,
      "theoretical_loss": 3.5567731286955873,
      "tokens_seen": 1318584320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037422924153584266,
      "loss": 2.5365,
      "theoretical_loss": 3.55675715480411,
      "tokens_seen": 1318649856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003742126915074712,
      "loss": 2.7757,
      "theoretical_loss": 3.5567411819287815,
      "tokens_seen": 1318715392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003741961414790997,
      "loss": 2.6878,
      "theoretical_loss": 3.556725210069487,
      "tokens_seen": 1318780928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003741795914507282,
      "loss": 2.8285,
      "theoretical_loss": 3.5567092392261106,
      "tokens_seen": 1318846464
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037416304142235676,
      "loss": 2.4713,
      "theoretical_loss": 3.5566932693985382,
      "tokens_seen": 1318912000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003741464913939853,
      "loss": 2.5194,
      "theoretical_loss": 3.556677300586654,
      "tokens_seen": 1318977536
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003741299413656138,
      "loss": 2.5132,
      "theoretical_loss": 3.556661332790343,
      "tokens_seen": 1319043072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003741133913372423,
      "loss": 2.4705,
      "theoretical_loss": 3.55664536600949,
      "tokens_seen": 1319108608
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037409684130887085,
      "loss": 2.785,
      "theoretical_loss": 3.556629400243981,
      "tokens_seen": 1319174144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003740802912804993,
      "loss": 2.7238,
      "theoretical_loss": 3.5566134354937,
      "tokens_seen": 1319239680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003740637412521278,
      "loss": 2.7746,
      "theoretical_loss": 3.5565974717585322,
      "tokens_seen": 1319305216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037404719122375636,
      "loss": 2.4608,
      "theoretical_loss": 3.556581509038363,
      "tokens_seen": 1319370752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037403064119538484,
      "loss": 2.7937,
      "theoretical_loss": 3.5565655473330766,
      "tokens_seen": 1319436288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003740140911670134,
      "loss": 2.6672,
      "theoretical_loss": 3.556549586642559,
      "tokens_seen": 1319501824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003739975411386419,
      "loss": 2.609,
      "theoretical_loss": 3.5565336269666954,
      "tokens_seen": 1319567360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037398099111027045,
      "loss": 2.63,
      "theoretical_loss": 3.5565176683053696,
      "tokens_seen": 1319632896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000373964441081899,
      "loss": 2.9616,
      "theoretical_loss": 3.5565017106584684,
      "tokens_seen": 1319698432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003739478910535275,
      "loss": 2.9349,
      "theoretical_loss": 3.556485754025876,
      "tokens_seen": 1319763968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000373931341025156,
      "loss": 2.7101,
      "theoretical_loss": 3.556469798407477,
      "tokens_seen": 1319829504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037391479099678455,
      "loss": 2.464,
      "theoretical_loss": 3.556453843803158,
      "tokens_seen": 1319895040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003738982409684131,
      "loss": 2.5907,
      "theoretical_loss": 3.556437890212803,
      "tokens_seen": 1319960576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037388169094004157,
      "loss": 2.9638,
      "theoretical_loss": 3.5564219376362978,
      "tokens_seen": 1320026112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003738651409116701,
      "loss": 2.5163,
      "theoretical_loss": 3.5564059860735275,
      "tokens_seen": 1320091648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037384859088329865,
      "loss": 2.6783,
      "theoretical_loss": 3.5563900355243776,
      "tokens_seen": 1320157184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003738320408549272,
      "loss": 2.4608,
      "theoretical_loss": 3.5563740859887334,
      "tokens_seen": 1320222720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037381549082655567,
      "loss": 2.7425,
      "theoretical_loss": 3.5563581374664794,
      "tokens_seen": 1320288256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003737989407981842,
      "loss": 2.528,
      "theoretical_loss": 3.556342189957502,
      "tokens_seen": 1320353792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037378239076981274,
      "loss": 2.8113,
      "theoretical_loss": 3.556326243461686,
      "tokens_seen": 1320419328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003737658407414413,
      "loss": 2.9104,
      "theoretical_loss": 3.556310297978916,
      "tokens_seen": 1320484864
    },
    {
      "epoch": 0.47,
      "objective/train/avg_token_score": 0.020005233585834503,
      "objective/train/avg_weight": 0.9764440059661865,
      "objective/train/docs_used": 747613,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.145772933959961,
      "objective/train/std_weight": 0.1086675152182579,
      "objective/train/theoretical_loss": 3.556294353509079,
      "objective/train/tokens_used": 1341010400,
      "theoretical_loss": 3.556294353509079,
      "tokens_seen": 1320550400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037374929071306976,
      "loss": 2.8916,
      "theoretical_loss": 3.556294353509079,
      "tokens_seen": 1320550400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003737327406846983,
      "loss": 2.6159,
      "theoretical_loss": 3.5562784100520592,
      "tokens_seen": 1320615936
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037371619065632684,
      "loss": 2.6093,
      "theoretical_loss": 3.556262467607742,
      "tokens_seen": 1320681472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003736996406279554,
      "loss": 2.454,
      "theoretical_loss": 3.5562465261760137,
      "tokens_seen": 1320747008
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037368309059958386,
      "loss": 2.7199,
      "theoretical_loss": 3.556230585756759,
      "tokens_seen": 1320812544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003736665405712124,
      "loss": 2.7626,
      "theoretical_loss": 3.556214646349864,
      "tokens_seen": 1320878080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037364999054284093,
      "loss": 2.5577,
      "theoretical_loss": 3.5561987079552133,
      "tokens_seen": 1320943616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037363344051446947,
      "loss": 2.6942,
      "theoretical_loss": 3.5561827705726925,
      "tokens_seen": 1321009152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037361689048609795,
      "loss": 2.9197,
      "theoretical_loss": 3.5561668342021884,
      "tokens_seen": 1321074688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003736003404577265,
      "loss": 2.4893,
      "theoretical_loss": 3.556150898843585,
      "tokens_seen": 1321140224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037358379042935503,
      "loss": 2.4193,
      "theoretical_loss": 3.556134964496769,
      "tokens_seen": 1321205760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037356724040098357,
      "loss": 2.4694,
      "theoretical_loss": 3.556119031161624,
      "tokens_seen": 1321271296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037355069037261205,
      "loss": 2.7721,
      "theoretical_loss": 3.5561030988380384,
      "tokens_seen": 1321336832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003735341403442406,
      "loss": 2.7288,
      "theoretical_loss": 3.556087167525896,
      "tokens_seen": 1321402368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003735175903158691,
      "loss": 2.6738,
      "theoretical_loss": 3.5560712372250833,
      "tokens_seen": 1321467904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037350104028749766,
      "loss": 2.7257,
      "theoretical_loss": 3.556055307935485,
      "tokens_seen": 1321533440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037348449025912615,
      "loss": 2.8855,
      "theoretical_loss": 3.556039379656988,
      "tokens_seen": 1321598976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003734679402307547,
      "loss": 2.7698,
      "theoretical_loss": 3.5560234523894763,
      "tokens_seen": 1321664512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003734513902023832,
      "loss": 2.637,
      "theoretical_loss": 3.556007526132837,
      "tokens_seen": 1321730048
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037343484017401176,
      "loss": 2.8063,
      "theoretical_loss": 3.555991600886956,
      "tokens_seen": 1321795584
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037341829014564024,
      "loss": 2.6051,
      "theoretical_loss": 3.555975676651718,
      "tokens_seen": 1321861120
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003734017401172688,
      "loss": 2.9044,
      "theoretical_loss": 3.555959753427009,
      "tokens_seen": 1321926656
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003733851900888973,
      "loss": 2.4265,
      "theoretical_loss": 3.5559438312127147,
      "tokens_seen": 1321992192
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037336864006052586,
      "loss": 2.6868,
      "theoretical_loss": 3.555927910008722,
      "tokens_seen": 1322057728
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037335209003215434,
      "loss": 2.4062,
      "theoretical_loss": 3.5559119898149154,
      "tokens_seen": 1322123264
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003733355400037829,
      "loss": 2.7023,
      "theoretical_loss": 3.5558960706311815,
      "tokens_seen": 1322188800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003733189899754114,
      "loss": 2.48,
      "theoretical_loss": 3.555880152457406,
      "tokens_seen": 1322254336
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037330243994703995,
      "loss": 2.4942,
      "theoretical_loss": 3.555864235293474,
      "tokens_seen": 1322319872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037328588991866843,
      "loss": 2.5973,
      "theoretical_loss": 3.5558483191392734,
      "tokens_seen": 1322385408
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037326933989029697,
      "loss": 2.5557,
      "theoretical_loss": 3.5558324039946876,
      "tokens_seen": 1322450944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003732527898619255,
      "loss": 2.7889,
      "theoretical_loss": 3.555816489859604,
      "tokens_seen": 1322516480
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037323623983355405,
      "loss": 3.0967,
      "theoretical_loss": 3.5558005767339083,
      "tokens_seen": 1322582016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037321968980518253,
      "loss": 2.7118,
      "theoretical_loss": 3.5557846646174864,
      "tokens_seen": 1322647552
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037320313977681107,
      "loss": 2.6585,
      "theoretical_loss": 3.5557687535102245,
      "tokens_seen": 1322713088
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003731865897484396,
      "loss": 2.522,
      "theoretical_loss": 3.5557528434120083,
      "tokens_seen": 1322778624
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037317003972006814,
      "loss": 2.5919,
      "theoretical_loss": 3.5557369343227236,
      "tokens_seen": 1322844160
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003731534896916966,
      "loss": 2.7177,
      "theoretical_loss": 3.5557210262422574,
      "tokens_seen": 1322909696
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037313693966332516,
      "loss": 2.6946,
      "theoretical_loss": 3.5557051191704945,
      "tokens_seen": 1322975232
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003731203896349537,
      "loss": 2.5985,
      "theoretical_loss": 3.555689213107322,
      "tokens_seen": 1323040768
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037310383960658213,
      "loss": 2.8507,
      "theoretical_loss": 3.5556733080526257,
      "tokens_seen": 1323106304
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037308728957821067,
      "loss": 2.8943,
      "theoretical_loss": 3.5556574040062916,
      "tokens_seen": 1323171840
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003730707395498392,
      "loss": 2.7282,
      "theoretical_loss": 3.555641500968205,
      "tokens_seen": 1323237376
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003730541895214677,
      "loss": 2.6902,
      "theoretical_loss": 3.5556255989382537,
      "tokens_seen": 1323302912
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003730376394930962,
      "loss": 2.7831,
      "theoretical_loss": 3.5556096979163234,
      "tokens_seen": 1323368448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037302108946472476,
      "loss": 3.0043,
      "theoretical_loss": 3.555593797902299,
      "tokens_seen": 1323433984
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003730045394363533,
      "loss": 2.8455,
      "theoretical_loss": 3.5555778988960682,
      "tokens_seen": 1323499520
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003729879894079818,
      "loss": 2.7801,
      "theoretical_loss": 3.555562000897517,
      "tokens_seen": 1323565056
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003729714393796103,
      "loss": 2.5999,
      "theoretical_loss": 3.5555461039065306,
      "tokens_seen": 1323630592
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037295488935123886,
      "loss": 2.7015,
      "theoretical_loss": 3.5555302079229962,
      "tokens_seen": 1323696128
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003729383393228674,
      "loss": 2.5296,
      "theoretical_loss": 3.5555143129467996,
      "tokens_seen": 1323761664
    },
    {
      "epoch": 0.47,
      "objective/train/avg_token_score": 0.01610785536468029,
      "objective/train/avg_weight": 0.9810377359390259,
      "objective/train/docs_used": 749700,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.9360086917877197,
      "objective/train/std_weight": 0.09517740458250046,
      "objective/train/theoretical_loss": 3.555498418977828,
      "objective/train/tokens_used": 1344287200,
      "theoretical_loss": 3.555498418977828,
      "tokens_seen": 1323827200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003729217892944959,
      "loss": 2.4227,
      "theoretical_loss": 3.555498418977828,
      "tokens_seen": 1323827200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003729052392661244,
      "loss": 2.6842,
      "theoretical_loss": 3.555482526015967,
      "tokens_seen": 1323892736
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037288868923775296,
      "loss": 2.7013,
      "theoretical_loss": 3.5554666340611023,
      "tokens_seen": 1323958272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003728721392093815,
      "loss": 2.472,
      "theoretical_loss": 3.5554507431131217,
      "tokens_seen": 1324023808
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037285558918101,
      "loss": 2.6618,
      "theoretical_loss": 3.555434853171911,
      "tokens_seen": 1324089344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003728390391526385,
      "loss": 2.818,
      "theoretical_loss": 3.555418964237356,
      "tokens_seen": 1324154880
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037282248912426705,
      "loss": 2.7979,
      "theoretical_loss": 3.5554030763093434,
      "tokens_seen": 1324220416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003728059390958956,
      "loss": 2.3802,
      "theoretical_loss": 3.55538718938776,
      "tokens_seen": 1324285952
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037278938906752407,
      "loss": 2.6592,
      "theoretical_loss": 3.555371303472492,
      "tokens_seen": 1324351488
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003727728390391526,
      "loss": 2.58,
      "theoretical_loss": 3.555355418563426,
      "tokens_seen": 1324417024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037275628901078115,
      "loss": 2.6581,
      "theoretical_loss": 3.5553395346604484,
      "tokens_seen": 1324482560
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003727397389824097,
      "loss": 2.5905,
      "theoretical_loss": 3.5553236517634454,
      "tokens_seen": 1324548096
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037272318895403817,
      "loss": 2.6142,
      "theoretical_loss": 3.5553077698723037,
      "tokens_seen": 1324613632
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003727066389256667,
      "loss": 2.7173,
      "theoretical_loss": 3.5552918889869103,
      "tokens_seen": 1324679168
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037269008889729524,
      "loss": 2.5359,
      "theoretical_loss": 3.5552760091071507,
      "tokens_seen": 1324744704
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003726735388689238,
      "loss": 2.6099,
      "theoretical_loss": 3.555260130232913,
      "tokens_seen": 1324810240
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037265698884055226,
      "loss": 2.8153,
      "theoretical_loss": 3.5552442523640826,
      "tokens_seen": 1324875776
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003726404388121808,
      "loss": 2.6626,
      "theoretical_loss": 3.5552283755005463,
      "tokens_seen": 1324941312
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037262388878380934,
      "loss": 2.6412,
      "theoretical_loss": 3.5552124996421908,
      "tokens_seen": 1325006848
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003726073387554379,
      "loss": 2.4843,
      "theoretical_loss": 3.5551966247889033,
      "tokens_seen": 1325072384
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037259078872706636,
      "loss": 2.6552,
      "theoretical_loss": 3.5551807509405697,
      "tokens_seen": 1325137920
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003725742386986949,
      "loss": 2.4898,
      "theoretical_loss": 3.5551648780970773,
      "tokens_seen": 1325203456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037255768867032344,
      "loss": 2.9828,
      "theoretical_loss": 3.555149006258312,
      "tokens_seen": 1325268992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000372541138641952,
      "loss": 2.8075,
      "theoretical_loss": 3.5551331354241613,
      "tokens_seen": 1325334528
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037252458861358046,
      "loss": 2.6209,
      "theoretical_loss": 3.5551172655945114,
      "tokens_seen": 1325400064
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000372508038585209,
      "loss": 2.6312,
      "theoretical_loss": 3.55510139676925,
      "tokens_seen": 1325465600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037249148855683753,
      "loss": 2.75,
      "theoretical_loss": 3.555085528948262,
      "tokens_seen": 1325531136
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037247493852846607,
      "loss": 2.9705,
      "theoretical_loss": 3.5550696621314364,
      "tokens_seen": 1325596672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003724583885000946,
      "loss": 2.463,
      "theoretical_loss": 3.555053796318659,
      "tokens_seen": 1325662208
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003724418384717231,
      "loss": 2.5754,
      "theoretical_loss": 3.5550379315098164,
      "tokens_seen": 1325727744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037242528844335163,
      "loss": 2.6092,
      "theoretical_loss": 3.5550220677047952,
      "tokens_seen": 1325793280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037240873841498017,
      "loss": 2.4414,
      "theoretical_loss": 3.555006204903483,
      "tokens_seen": 1325858816
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003723921883866087,
      "loss": 2.7817,
      "theoretical_loss": 3.554990343105767,
      "tokens_seen": 1325924352
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003723756383582372,
      "loss": 2.8457,
      "theoretical_loss": 3.554974482311533,
      "tokens_seen": 1325989888
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003723590883298657,
      "loss": 2.8709,
      "theoretical_loss": 3.554958622520669,
      "tokens_seen": 1326055424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037234253830149426,
      "loss": 2.7238,
      "theoretical_loss": 3.5549427637330604,
      "tokens_seen": 1326120960
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003723259882731228,
      "loss": 2.5806,
      "theoretical_loss": 3.554926905948596,
      "tokens_seen": 1326186496
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003723094382447513,
      "loss": 2.591,
      "theoretical_loss": 3.5549110491671616,
      "tokens_seen": 1326252032
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003722928882163798,
      "loss": 2.6474,
      "theoretical_loss": 3.554895193388645,
      "tokens_seen": 1326317568
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037227633818800836,
      "loss": 2.8058,
      "theoretical_loss": 3.5548793386129325,
      "tokens_seen": 1326383104
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003722597881596369,
      "loss": 2.7291,
      "theoretical_loss": 3.5548634848399114,
      "tokens_seen": 1326448640
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003722432381312654,
      "loss": 2.3269,
      "theoretical_loss": 3.5548476320694684,
      "tokens_seen": 1326514176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003722266881028939,
      "loss": 2.4561,
      "theoretical_loss": 3.554831780301491,
      "tokens_seen": 1326579712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037221013807452245,
      "loss": 2.8425,
      "theoretical_loss": 3.554815929535867,
      "tokens_seen": 1326645248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000372193588046151,
      "loss": 2.7456,
      "theoretical_loss": 3.554800079772482,
      "tokens_seen": 1326710784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003721770380177795,
      "loss": 2.7168,
      "theoretical_loss": 3.554784231011224,
      "tokens_seen": 1326776320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000372160487989408,
      "loss": 2.7669,
      "theoretical_loss": 3.55476838325198,
      "tokens_seen": 1326841856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037214393796103655,
      "loss": 2.5804,
      "theoretical_loss": 3.5547525364946373,
      "tokens_seen": 1326907392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003721273879326651,
      "loss": 2.6227,
      "theoretical_loss": 3.554736690739083,
      "tokens_seen": 1326972928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003721108379042935,
      "loss": 2.7831,
      "theoretical_loss": 3.5547208459852033,
      "tokens_seen": 1327038464
    },
    {
      "epoch": 0.47,
      "objective/train/avg_token_score": 0.041607752442359924,
      "objective/train/avg_weight": 0.9510558247566223,
      "objective/train/docs_used": 752163,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.683650016784668,
      "objective/train/std_weight": 0.20603998005390167,
      "objective/train/theoretical_loss": 3.5547050022328874,
      "objective/train/tokens_used": 1347564000,
      "theoretical_loss": 3.5547050022328874,
      "tokens_seen": 1327104000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037209428787592205,
      "loss": 2.6482,
      "theoretical_loss": 3.5547050022328874,
      "tokens_seen": 1327104000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037207773784755054,
      "loss": 2.7358,
      "theoretical_loss": 3.554689159482021,
      "tokens_seen": 1327169536
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003720611878191791,
      "loss": 2.4192,
      "theoretical_loss": 3.554673317732492,
      "tokens_seen": 1327235072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003720446377908076,
      "loss": 2.5873,
      "theoretical_loss": 3.5546574769841874,
      "tokens_seen": 1327300608
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037202808776243615,
      "loss": 2.7249,
      "theoretical_loss": 3.554641637236995,
      "tokens_seen": 1327366144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037201153773406463,
      "loss": 2.6124,
      "theoretical_loss": 3.554625798490801,
      "tokens_seen": 1327431680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037199498770569317,
      "loss": 2.633,
      "theoretical_loss": 3.554609960745494,
      "tokens_seen": 1327497216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003719784376773217,
      "loss": 2.8281,
      "theoretical_loss": 3.5545941240009604,
      "tokens_seen": 1327562752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037196188764895025,
      "loss": 2.5583,
      "theoretical_loss": 3.5545782882570878,
      "tokens_seen": 1327628288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037194533762057873,
      "loss": 2.9761,
      "theoretical_loss": 3.554562453513764,
      "tokens_seen": 1327693824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037192878759220727,
      "loss": 2.8882,
      "theoretical_loss": 3.554546619770876,
      "tokens_seen": 1327759360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003719122375638358,
      "loss": 2.5707,
      "theoretical_loss": 3.5545307870283116,
      "tokens_seen": 1327824896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037189568753546434,
      "loss": 2.6658,
      "theoretical_loss": 3.554514955285958,
      "tokens_seen": 1327890432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003718791375070928,
      "loss": 2.554,
      "theoretical_loss": 3.5544991245437023,
      "tokens_seen": 1327955968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037186258747872136,
      "loss": 2.7173,
      "theoretical_loss": 3.5544832948014324,
      "tokens_seen": 1328021504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003718460374503499,
      "loss": 2.8465,
      "theoretical_loss": 3.5544674660590356,
      "tokens_seen": 1328087040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037182948742197844,
      "loss": 2.4255,
      "theoretical_loss": 3.5544516383164,
      "tokens_seen": 1328152576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003718129373936069,
      "loss": 2.7073,
      "theoretical_loss": 3.554435811573412,
      "tokens_seen": 1328218112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037179638736523546,
      "loss": 2.8043,
      "theoretical_loss": 3.55441998582996,
      "tokens_seen": 1328283648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000371779837336864,
      "loss": 2.7477,
      "theoretical_loss": 3.554404161085931,
      "tokens_seen": 1328349184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037176328730849253,
      "loss": 2.7124,
      "theoretical_loss": 3.554388337341213,
      "tokens_seen": 1328414720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000371746737280121,
      "loss": 2.8154,
      "theoretical_loss": 3.554372514595694,
      "tokens_seen": 1328480256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037173018725174955,
      "loss": 2.7379,
      "theoretical_loss": 3.554356692849261,
      "tokens_seen": 1328545792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003717136372233781,
      "loss": 2.5522,
      "theoretical_loss": 3.5543408721018013,
      "tokens_seen": 1328611328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037169708719500663,
      "loss": 2.7417,
      "theoretical_loss": 3.5543250523532035,
      "tokens_seen": 1328676864
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003716805371666351,
      "loss": 2.4561,
      "theoretical_loss": 3.5543092336033544,
      "tokens_seen": 1328742400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037166398713826365,
      "loss": 2.6608,
      "theoretical_loss": 3.5542934158521424,
      "tokens_seen": 1328807936
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003716474371098922,
      "loss": 2.5473,
      "theoretical_loss": 3.554277599099455,
      "tokens_seen": 1328873472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003716308870815207,
      "loss": 2.7674,
      "theoretical_loss": 3.5542617833451793,
      "tokens_seen": 1328939008
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003716143370531492,
      "loss": 2.6015,
      "theoretical_loss": 3.554245968589204,
      "tokens_seen": 1329004544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037159778702477775,
      "loss": 2.5938,
      "theoretical_loss": 3.554230154831416,
      "tokens_seen": 1329070080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003715812369964063,
      "loss": 2.2698,
      "theoretical_loss": 3.5542143420717043,
      "tokens_seen": 1329135616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003715646869680348,
      "loss": 2.9458,
      "theoretical_loss": 3.554198530309955,
      "tokens_seen": 1329201152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003715481369396633,
      "loss": 2.7023,
      "theoretical_loss": 3.5541827195460574,
      "tokens_seen": 1329266688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037153158691129184,
      "loss": 2.5967,
      "theoretical_loss": 3.554166909779898,
      "tokens_seen": 1329332224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003715150368829204,
      "loss": 2.3778,
      "theoretical_loss": 3.5541511010113664,
      "tokens_seen": 1329397760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003714984868545489,
      "loss": 2.439,
      "theoretical_loss": 3.554135293240349,
      "tokens_seen": 1329463296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003714819368261774,
      "loss": 2.5128,
      "theoretical_loss": 3.554119486466734,
      "tokens_seen": 1329528832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037146538679780594,
      "loss": 2.4672,
      "theoretical_loss": 3.55410368069041,
      "tokens_seen": 1329594368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003714488367694345,
      "loss": 2.5175,
      "theoretical_loss": 3.554087875911264,
      "tokens_seen": 1329659904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000371432286741063,
      "loss": 2.6058,
      "theoretical_loss": 3.554072072129184,
      "tokens_seen": 1329725440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0003714157367126915,
      "loss": 2.8449,
      "theoretical_loss": 3.554056269344059,
      "tokens_seen": 1329790976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037139918668432003,
      "loss": 2.7629,
      "theoretical_loss": 3.554040467555776,
      "tokens_seen": 1329856512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00037138263665594857,
      "loss": 2.7536,
      "theoretical_loss": 3.5540246667642235,
      "tokens_seen": 1329922048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003713660866275771,
      "loss": 2.5584,
      "theoretical_loss": 3.554008866969289,
      "tokens_seen": 1329987584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003713495365992056,
      "loss": 2.6764,
      "theoretical_loss": 3.553993068170861,
      "tokens_seen": 1330053120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037133298657083413,
      "loss": 2.7345,
      "theoretical_loss": 3.5539772703688275,
      "tokens_seen": 1330118656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037131643654246267,
      "loss": 2.3567,
      "theoretical_loss": 3.553961473563076,
      "tokens_seen": 1330184192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003712998865140912,
      "loss": 2.8087,
      "theoretical_loss": 3.5539456777534957,
      "tokens_seen": 1330249728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003712833364857197,
      "loss": 2.6215,
      "theoretical_loss": 3.5539298829399737,
      "tokens_seen": 1330315264
    },
    {
      "epoch": 0.48,
      "objective/train/avg_token_score": 0.02648327872157097,
      "objective/train/avg_weight": 0.9688311815261841,
      "objective/train/docs_used": 753856,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0848546028137207,
      "objective/train/std_weight": 0.1427544504404068,
      "objective/train/theoretical_loss": 3.553914089122399,
      "objective/train/tokens_used": 1350840800,
      "theoretical_loss": 3.553914089122399,
      "tokens_seen": 1330380800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003712667864573482,
      "loss": 2.6164,
      "theoretical_loss": 3.553914089122399,
      "tokens_seen": 1330380800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037125023642897676,
      "loss": 2.6339,
      "theoretical_loss": 3.5538982963006585,
      "tokens_seen": 1330446336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003712336864006053,
      "loss": 2.6986,
      "theoretical_loss": 3.5538825044746414,
      "tokens_seen": 1330511872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003712171363722338,
      "loss": 2.967,
      "theoretical_loss": 3.5538667136442355,
      "tokens_seen": 1330577408
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003712005863438623,
      "loss": 2.6195,
      "theoretical_loss": 3.553850923809329,
      "tokens_seen": 1330642944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037118403631549086,
      "loss": 2.4324,
      "theoretical_loss": 3.5538351349698107,
      "tokens_seen": 1330708480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003711674862871194,
      "loss": 2.8023,
      "theoretical_loss": 3.553819347125568,
      "tokens_seen": 1330774016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003711509362587479,
      "loss": 2.477,
      "theoretical_loss": 3.5538035602764895,
      "tokens_seen": 1330839552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037113438623037636,
      "loss": 2.6505,
      "theoretical_loss": 3.5537877744224633,
      "tokens_seen": 1330905088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003711178362020049,
      "loss": 2.8605,
      "theoretical_loss": 3.553771989563378,
      "tokens_seen": 1330970624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003711012861736334,
      "loss": 3.0968,
      "theoretical_loss": 3.5537562056991217,
      "tokens_seen": 1331036160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003710847361452619,
      "loss": 2.3095,
      "theoretical_loss": 3.5537404228295824,
      "tokens_seen": 1331101696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037106818611689046,
      "loss": 2.6889,
      "theoretical_loss": 3.5537246409546492,
      "tokens_seen": 1331167232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000371051636088519,
      "loss": 2.5236,
      "theoretical_loss": 3.5537088600742104,
      "tokens_seen": 1331232768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003710350860601475,
      "loss": 2.7045,
      "theoretical_loss": 3.5536930801881534,
      "tokens_seen": 1331298304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000371018536031776,
      "loss": 2.6561,
      "theoretical_loss": 3.5536773012963674,
      "tokens_seen": 1331363840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037100198600340456,
      "loss": 2.7366,
      "theoretical_loss": 3.553661523398741,
      "tokens_seen": 1331429376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003709854359750331,
      "loss": 2.8057,
      "theoretical_loss": 3.553645746495162,
      "tokens_seen": 1331494912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003709688859466616,
      "loss": 2.755,
      "theoretical_loss": 3.553629970585519,
      "tokens_seen": 1331560448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003709523359182901,
      "loss": 2.4983,
      "theoretical_loss": 3.5536141956697005,
      "tokens_seen": 1331625984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037093578588991865,
      "loss": 2.864,
      "theoretical_loss": 3.5535984217475955,
      "tokens_seen": 1331691520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003709192358615472,
      "loss": 2.8389,
      "theoretical_loss": 3.5535826488190922,
      "tokens_seen": 1331757056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037090268583317567,
      "loss": 2.6898,
      "theoretical_loss": 3.5535668768840782,
      "tokens_seen": 1331822592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003708861358048042,
      "loss": 2.4077,
      "theoretical_loss": 3.553551105942444,
      "tokens_seen": 1331888128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037086958577643275,
      "loss": 2.5779,
      "theoretical_loss": 3.5535353359940762,
      "tokens_seen": 1331953664
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003708530357480613,
      "loss": 2.9029,
      "theoretical_loss": 3.5535195670388644,
      "tokens_seen": 1332019200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037083648571968977,
      "loss": 2.7808,
      "theoretical_loss": 3.553503799076697,
      "tokens_seen": 1332084736
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003708199356913183,
      "loss": 2.7245,
      "theoretical_loss": 3.5534880321074627,
      "tokens_seen": 1332150272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037080338566294684,
      "loss": 2.817,
      "theoretical_loss": 3.55347226613105,
      "tokens_seen": 1332215808
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003707868356345754,
      "loss": 2.5439,
      "theoretical_loss": 3.5534565011473473,
      "tokens_seen": 1332281344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037077028560620386,
      "loss": 2.5838,
      "theoretical_loss": 3.5534407371562438,
      "tokens_seen": 1332346880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003707537355778324,
      "loss": 2.4648,
      "theoretical_loss": 3.553424974157628,
      "tokens_seen": 1332412416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037073718554946094,
      "loss": 2.3839,
      "theoretical_loss": 3.5534092121513883,
      "tokens_seen": 1332477952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003707206355210895,
      "loss": 2.5017,
      "theoretical_loss": 3.5533934511374135,
      "tokens_seen": 1332543488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037070408549271796,
      "loss": 2.9561,
      "theoretical_loss": 3.553377691115593,
      "tokens_seen": 1332609024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003706875354643465,
      "loss": 2.8239,
      "theoretical_loss": 3.5533619320858145,
      "tokens_seen": 1332674560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037067098543597504,
      "loss": 2.8099,
      "theoretical_loss": 3.5533461740479675,
      "tokens_seen": 1332740096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037065443540760357,
      "loss": 2.6712,
      "theoretical_loss": 3.553330417001941,
      "tokens_seen": 1332805632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037063788537923206,
      "loss": 2.5861,
      "theoretical_loss": 3.553314660947623,
      "tokens_seen": 1332871168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003706213353508606,
      "loss": 2.3301,
      "theoretical_loss": 3.553298905884903,
      "tokens_seen": 1332936704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037060478532248913,
      "loss": 2.7789,
      "theoretical_loss": 3.5532831518136696,
      "tokens_seen": 1333002240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037058823529411767,
      "loss": 2.8891,
      "theoretical_loss": 3.5532673987338113,
      "tokens_seen": 1333067776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037057168526574615,
      "loss": 2.4861,
      "theoretical_loss": 3.5532516466452173,
      "tokens_seen": 1333133312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003705551352373747,
      "loss": 2.4747,
      "theoretical_loss": 3.553235895547777,
      "tokens_seen": 1333198848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037053858520900323,
      "loss": 2.5908,
      "theoretical_loss": 3.5532201454413785,
      "tokens_seen": 1333264384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037052203518063176,
      "loss": 2.5838,
      "theoretical_loss": 3.5532043963259112,
      "tokens_seen": 1333329920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037050548515226025,
      "loss": 2.5306,
      "theoretical_loss": 3.553188648201264,
      "tokens_seen": 1333395456
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003704889351238888,
      "loss": 2.0043,
      "theoretical_loss": 3.5531729010673256,
      "tokens_seen": 1333460992
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003704723850955173,
      "loss": 2.5228,
      "theoretical_loss": 3.5531571549239853,
      "tokens_seen": 1333526528
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037045583506714586,
      "loss": 2.8771,
      "theoretical_loss": 3.5531414097711322,
      "tokens_seen": 1333592064
    },
    {
      "epoch": 0.48,
      "objective/train/avg_token_score": 0.02437640354037285,
      "objective/train/avg_weight": 0.971311092376709,
      "objective/train/docs_used": 755628,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.695122480392456,
      "objective/train/std_weight": 0.10065309703350067,
      "objective/train/theoretical_loss": 3.553125665608655,
      "objective/train/tokens_used": 1354117600,
      "theoretical_loss": 3.553125665608655,
      "tokens_seen": 1333657600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037043928503877434,
      "loss": 2.5409,
      "theoretical_loss": 3.553125665608655,
      "tokens_seen": 1333657600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003704227350104029,
      "loss": 2.6912,
      "theoretical_loss": 3.553109922436443,
      "tokens_seen": 1333723136
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003704061849820314,
      "loss": 2.6627,
      "theoretical_loss": 3.553094180254385,
      "tokens_seen": 1333788672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037038963495365996,
      "loss": 2.607,
      "theoretical_loss": 3.5530784390623698,
      "tokens_seen": 1333854208
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037037308492528844,
      "loss": 2.5608,
      "theoretical_loss": 3.5530626988602876,
      "tokens_seen": 1333919744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000370356534896917,
      "loss": 2.513,
      "theoretical_loss": 3.5530469596480265,
      "tokens_seen": 1333985280
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003703399848685455,
      "loss": 2.6228,
      "theoretical_loss": 3.553031221425476,
      "tokens_seen": 1334050816
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037032343484017405,
      "loss": 2.8664,
      "theoretical_loss": 3.553015484192525,
      "tokens_seen": 1334116352
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037030688481180254,
      "loss": 2.864,
      "theoretical_loss": 3.5529997479490634,
      "tokens_seen": 1334181888
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003702903347834311,
      "loss": 2.4822,
      "theoretical_loss": 3.552984012694979,
      "tokens_seen": 1334247424
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003702737847550596,
      "loss": 2.4228,
      "theoretical_loss": 3.5529682784301624,
      "tokens_seen": 1334312960
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037025723472668815,
      "loss": 2.5881,
      "theoretical_loss": 3.5529525451545023,
      "tokens_seen": 1334378496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037024068469831663,
      "loss": 2.7846,
      "theoretical_loss": 3.552936812867888,
      "tokens_seen": 1334444032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037022413466994517,
      "loss": 2.6735,
      "theoretical_loss": 3.552921081570209,
      "tokens_seen": 1334509568
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003702075846415737,
      "loss": 2.5303,
      "theoretical_loss": 3.5529053512613533,
      "tokens_seen": 1334575104
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037019103461320224,
      "loss": 2.9289,
      "theoretical_loss": 3.552889621941212,
      "tokens_seen": 1334640640
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037017448458483073,
      "loss": 2.6028,
      "theoretical_loss": 3.5528738936096724,
      "tokens_seen": 1334706176
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037015793455645927,
      "loss": 2.6109,
      "theoretical_loss": 3.552858166266626,
      "tokens_seen": 1334771712
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003701413845280877,
      "loss": 2.5239,
      "theoretical_loss": 3.552842439911961,
      "tokens_seen": 1334837248
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037012483449971623,
      "loss": 2.7234,
      "theoretical_loss": 3.5528267145455668,
      "tokens_seen": 1334902784
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037010828447134477,
      "loss": 2.673,
      "theoretical_loss": 3.552810990167333,
      "tokens_seen": 1334968320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003700917344429733,
      "loss": 2.7898,
      "theoretical_loss": 3.5527952667771485,
      "tokens_seen": 1335033856
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003700751844146018,
      "loss": 2.7652,
      "theoretical_loss": 3.5527795443749035,
      "tokens_seen": 1335099392
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037005863438623033,
      "loss": 2.7038,
      "theoretical_loss": 3.5527638229604865,
      "tokens_seen": 1335164928
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037004208435785887,
      "loss": 2.7043,
      "theoretical_loss": 3.552748102533788,
      "tokens_seen": 1335230464
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003700255343294874,
      "loss": 2.8532,
      "theoretical_loss": 3.552732383094696,
      "tokens_seen": 1335296000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00037000898430111594,
      "loss": 2.6327,
      "theoretical_loss": 3.5527166646431025,
      "tokens_seen": 1335361536
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003699924342727444,
      "loss": 2.7578,
      "theoretical_loss": 3.552700947178894,
      "tokens_seen": 1335427072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036997588424437296,
      "loss": 2.8516,
      "theoretical_loss": 3.5526852307019623,
      "tokens_seen": 1335492608
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003699593342160015,
      "loss": 2.6781,
      "theoretical_loss": 3.5526695152121963,
      "tokens_seen": 1335558144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036994278418763004,
      "loss": 2.6717,
      "theoretical_loss": 3.5526538007094848,
      "tokens_seen": 1335623680
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003699262341592585,
      "loss": 2.4713,
      "theoretical_loss": 3.552638087193718,
      "tokens_seen": 1335689216
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036990968413088706,
      "loss": 2.5476,
      "theoretical_loss": 3.552622374664786,
      "tokens_seen": 1335754752
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003698931341025156,
      "loss": 2.7575,
      "theoretical_loss": 3.5526066631225772,
      "tokens_seen": 1335820288
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036987658407414413,
      "loss": 2.7704,
      "theoretical_loss": 3.5525909525669817,
      "tokens_seen": 1335885824
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003698600340457726,
      "loss": 2.7005,
      "theoretical_loss": 3.5525752429978894,
      "tokens_seen": 1335951360
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036984348401740115,
      "loss": 2.4265,
      "theoretical_loss": 3.552559534415191,
      "tokens_seen": 1336016896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003698269339890297,
      "loss": 2.8808,
      "theoretical_loss": 3.552543826818774,
      "tokens_seen": 1336082432
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036981038396065823,
      "loss": 2.7031,
      "theoretical_loss": 3.552528120208529,
      "tokens_seen": 1336147968
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003697938339322867,
      "loss": 2.7066,
      "theoretical_loss": 3.5525124145843465,
      "tokens_seen": 1336213504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036977728390391525,
      "loss": 2.3554,
      "theoretical_loss": 3.5524967099461158,
      "tokens_seen": 1336279040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003697607338755438,
      "loss": 2.6975,
      "theoretical_loss": 3.5524810062937258,
      "tokens_seen": 1336344576
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003697441838471723,
      "loss": 2.7898,
      "theoretical_loss": 3.5524653036270673,
      "tokens_seen": 1336410112
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003697276338188008,
      "loss": 2.5734,
      "theoretical_loss": 3.55244960194603,
      "tokens_seen": 1336475648
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036971108379042935,
      "loss": 2.6243,
      "theoretical_loss": 3.5524339012505033,
      "tokens_seen": 1336541184
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003696945337620579,
      "loss": 2.6565,
      "theoretical_loss": 3.552418201540377,
      "tokens_seen": 1336606720
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003696779837336864,
      "loss": 2.7419,
      "theoretical_loss": 3.5524025028155415,
      "tokens_seen": 1336672256
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003696614337053149,
      "loss": 2.5631,
      "theoretical_loss": 3.5523868050758862,
      "tokens_seen": 1336737792
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036964488367694344,
      "loss": 2.8627,
      "theoretical_loss": 3.552371108321301,
      "tokens_seen": 1336803328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000369628333648572,
      "loss": 2.4323,
      "theoretical_loss": 3.552355412551676,
      "tokens_seen": 1336868864
    },
    {
      "epoch": 0.48,
      "objective/train/avg_token_score": 0.017261125147342682,
      "objective/train/avg_weight": 0.9796912670135498,
      "objective/train/docs_used": 757781,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9634814262390137,
      "objective/train/std_weight": 0.07162501662969589,
      "objective/train/theoretical_loss": 3.5523397177669005,
      "objective/train/tokens_used": 1357394400,
      "theoretical_loss": 3.5523397177669005,
      "tokens_seen": 1336934400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003696117836202005,
      "loss": 2.7161,
      "theoretical_loss": 3.5523397177669005,
      "tokens_seen": 1336934400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000369595233591829,
      "loss": 2.5154,
      "theoretical_loss": 3.5523240239668654,
      "tokens_seen": 1336999936
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036957868356345754,
      "loss": 2.7406,
      "theoretical_loss": 3.55230833115146,
      "tokens_seen": 1337065472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003695621335350861,
      "loss": 2.748,
      "theoretical_loss": 3.5522926393205743,
      "tokens_seen": 1337131008
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003695455835067146,
      "loss": 2.6719,
      "theoretical_loss": 3.5522769484740984,
      "tokens_seen": 1337196544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003695290334783431,
      "loss": 2.693,
      "theoretical_loss": 3.5522612586119227,
      "tokens_seen": 1337262080
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036951248344997163,
      "loss": 2.9702,
      "theoretical_loss": 3.5522455697339366,
      "tokens_seen": 1337327616
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036949593342160017,
      "loss": 2.6395,
      "theoretical_loss": 3.55222988184003,
      "tokens_seen": 1337393152
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003694793833932287,
      "loss": 2.5577,
      "theoretical_loss": 3.552214194930094,
      "tokens_seen": 1337458688
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003694628333648572,
      "loss": 2.6716,
      "theoretical_loss": 3.552198509004018,
      "tokens_seen": 1337524224
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036944628333648573,
      "loss": 2.7485,
      "theoretical_loss": 3.5521828240616915,
      "tokens_seen": 1337589760
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036942973330811427,
      "loss": 2.5408,
      "theoretical_loss": 3.5521671401030055,
      "tokens_seen": 1337655296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003694131832797428,
      "loss": 2.7853,
      "theoretical_loss": 3.5521514571278496,
      "tokens_seen": 1337720832
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003693966332513713,
      "loss": 2.8427,
      "theoretical_loss": 3.552135775136115,
      "tokens_seen": 1337786368
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003693800832229998,
      "loss": 2.76,
      "theoretical_loss": 3.5521200941276905,
      "tokens_seen": 1337851904
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036936353319462836,
      "loss": 2.7453,
      "theoretical_loss": 3.552104414102467,
      "tokens_seen": 1337917440
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003693469831662569,
      "loss": 2.8926,
      "theoretical_loss": 3.552088735060334,
      "tokens_seen": 1337982976
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003693304331378854,
      "loss": 2.2731,
      "theoretical_loss": 3.5520730570011825,
      "tokens_seen": 1338048512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003693138831095139,
      "loss": 2.6389,
      "theoretical_loss": 3.5520573799249027,
      "tokens_seen": 1338114048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036929733308114246,
      "loss": 2.8116,
      "theoretical_loss": 3.5520417038313843,
      "tokens_seen": 1338179584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000369280783052771,
      "loss": 2.8332,
      "theoretical_loss": 3.552026028720518,
      "tokens_seen": 1338245120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003692642330243995,
      "loss": 2.6596,
      "theoretical_loss": 3.552010354592194,
      "tokens_seen": 1338310656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000369247682996028,
      "loss": 2.7964,
      "theoretical_loss": 3.5519946814463026,
      "tokens_seen": 1338376192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036923113296765655,
      "loss": 2.6531,
      "theoretical_loss": 3.551979009282734,
      "tokens_seen": 1338441728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003692145829392851,
      "loss": 2.5311,
      "theoretical_loss": 3.5519633381013787,
      "tokens_seen": 1338507264
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003691980329109136,
      "loss": 2.6628,
      "theoretical_loss": 3.5519476679021267,
      "tokens_seen": 1338572800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003691814828825421,
      "loss": 2.7198,
      "theoretical_loss": 3.551931998684869,
      "tokens_seen": 1338638336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036916493285417054,
      "loss": 2.6011,
      "theoretical_loss": 3.5519163304494956,
      "tokens_seen": 1338703872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003691483828257991,
      "loss": 2.8762,
      "theoretical_loss": 3.551900663195897,
      "tokens_seen": 1338769408
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003691318327974276,
      "loss": 2.9198,
      "theoretical_loss": 3.5518849969239636,
      "tokens_seen": 1338834944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036911528276905615,
      "loss": 2.6161,
      "theoretical_loss": 3.551869331633586,
      "tokens_seen": 1338900480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036909873274068464,
      "loss": 2.5854,
      "theoretical_loss": 3.5518536673246537,
      "tokens_seen": 1338966016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003690821827123132,
      "loss": 2.7012,
      "theoretical_loss": 3.5518380039970587,
      "tokens_seen": 1339031552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003690656326839417,
      "loss": 2.4819,
      "theoretical_loss": 3.5518223416506904,
      "tokens_seen": 1339097088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036904908265557025,
      "loss": 2.5282,
      "theoretical_loss": 3.55180668028544,
      "tokens_seen": 1339162624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036903253262719873,
      "loss": 2.6039,
      "theoretical_loss": 3.5517910199011977,
      "tokens_seen": 1339228160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036901598259882727,
      "loss": 2.7905,
      "theoretical_loss": 3.551775360497854,
      "tokens_seen": 1339293696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003689994325704558,
      "loss": 2.6815,
      "theoretical_loss": 3.551759702075299,
      "tokens_seen": 1339359232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036898288254208435,
      "loss": 2.5963,
      "theoretical_loss": 3.5517440446334243,
      "tokens_seen": 1339424768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036896633251371283,
      "loss": 2.5545,
      "theoretical_loss": 3.55172838817212,
      "tokens_seen": 1339490304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036894978248534137,
      "loss": 2.7585,
      "theoretical_loss": 3.5517127326912767,
      "tokens_seen": 1339555840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003689332324569699,
      "loss": 2.6569,
      "theoretical_loss": 3.5516970781907853,
      "tokens_seen": 1339621376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036891668242859844,
      "loss": 2.7418,
      "theoretical_loss": 3.5516814246705355,
      "tokens_seen": 1339686912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003689001324002269,
      "loss": 2.6246,
      "theoretical_loss": 3.5516657721304195,
      "tokens_seen": 1339752448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036888358237185546,
      "loss": 2.5649,
      "theoretical_loss": 3.5516501205703266,
      "tokens_seen": 1339817984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000368867032343484,
      "loss": 2.1093,
      "theoretical_loss": 3.551634469990148,
      "tokens_seen": 1339883520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036885048231511254,
      "loss": 2.6456,
      "theoretical_loss": 3.551618820389775,
      "tokens_seen": 1339949056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000368833932286741,
      "loss": 2.6975,
      "theoretical_loss": 3.5516031717690977,
      "tokens_seen": 1340014592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036881738225836956,
      "loss": 2.805,
      "theoretical_loss": 3.5515875241280073,
      "tokens_seen": 1340080128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003688008322299981,
      "loss": 2.6424,
      "theoretical_loss": 3.551571877466394,
      "tokens_seen": 1340145664
    },
    {
      "epoch": 0.48,
      "objective/train/avg_token_score": 0.0058165728114545345,
      "objective/train/avg_weight": 0.9931523203849792,
      "objective/train/docs_used": 759683,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.862051486968994,
      "objective/train/std_weight": 0.05214110016822815,
      "objective/train/theoretical_loss": 3.551556231784149,
      "objective/train/tokens_used": 1360671200,
      "theoretical_loss": 3.551556231784149,
      "tokens_seen": 1340211200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036878428220162663,
      "loss": 2.7163,
      "theoretical_loss": 3.551556231784149,
      "tokens_seen": 1340211200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003687677321732551,
      "loss": 2.3335,
      "theoretical_loss": 3.5515405870811625,
      "tokens_seen": 1340276736
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036875118214488366,
      "loss": 2.6322,
      "theoretical_loss": 3.551524943357326,
      "tokens_seen": 1340342272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003687346321165122,
      "loss": 2.6434,
      "theoretical_loss": 3.5515093006125307,
      "tokens_seen": 1340407808
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036871808208814073,
      "loss": 2.5638,
      "theoretical_loss": 3.551493658846667,
      "tokens_seen": 1340473344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003687015320597692,
      "loss": 2.7053,
      "theoretical_loss": 3.5514780180596253,
      "tokens_seen": 1340538880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036868498203139775,
      "loss": 2.6819,
      "theoretical_loss": 3.551462378251297,
      "tokens_seen": 1340604416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003686684320030263,
      "loss": 2.3514,
      "theoretical_loss": 3.5514467394215723,
      "tokens_seen": 1340669952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003686518819746548,
      "loss": 2.1785,
      "theoretical_loss": 3.551431101570344,
      "tokens_seen": 1340735488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003686353319462833,
      "loss": 2.7339,
      "theoretical_loss": 3.551415464697501,
      "tokens_seen": 1340801024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036861878191791185,
      "loss": 2.7101,
      "theoretical_loss": 3.5513998288029356,
      "tokens_seen": 1340866560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003686022318895404,
      "loss": 2.8457,
      "theoretical_loss": 3.5513841938865376,
      "tokens_seen": 1340932096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003685856818611689,
      "loss": 2.8378,
      "theoretical_loss": 3.551368559948199,
      "tokens_seen": 1340997632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003685691318327974,
      "loss": 2.5981,
      "theoretical_loss": 3.5513529269878106,
      "tokens_seen": 1341063168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036855258180442594,
      "loss": 2.4642,
      "theoretical_loss": 3.551337295005264,
      "tokens_seen": 1341128704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003685360317760545,
      "loss": 2.5466,
      "theoretical_loss": 3.5513216640004486,
      "tokens_seen": 1341194240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000368519481747683,
      "loss": 2.5393,
      "theoretical_loss": 3.551306033973257,
      "tokens_seen": 1341259776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003685029317193115,
      "loss": 2.4488,
      "theoretical_loss": 3.5512904049235794,
      "tokens_seen": 1341325312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036848638169094004,
      "loss": 2.3392,
      "theoretical_loss": 3.551274776851308,
      "tokens_seen": 1341390848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003684698316625686,
      "loss": 2.7902,
      "theoretical_loss": 3.5512591497563326,
      "tokens_seen": 1341456384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003684532816341971,
      "loss": 2.8392,
      "theoretical_loss": 3.551243523638545,
      "tokens_seen": 1341521920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036843673160582565,
      "loss": 2.495,
      "theoretical_loss": 3.551227898497836,
      "tokens_seen": 1341587456
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036842018157745414,
      "loss": 2.6114,
      "theoretical_loss": 3.5512122743340977,
      "tokens_seen": 1341652992
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003684036315490827,
      "loss": 2.6308,
      "theoretical_loss": 3.5511966511472206,
      "tokens_seen": 1341718528
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003683870815207112,
      "loss": 2.8464,
      "theoretical_loss": 3.551181028937096,
      "tokens_seen": 1341784064
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036837053149233975,
      "loss": 2.3605,
      "theoretical_loss": 3.551165407703615,
      "tokens_seen": 1341849600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036835398146396823,
      "loss": 2.582,
      "theoretical_loss": 3.551149787446669,
      "tokens_seen": 1341915136
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036833743143559677,
      "loss": 2.7321,
      "theoretical_loss": 3.5511341681661492,
      "tokens_seen": 1341980672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003683208814072253,
      "loss": 2.7378,
      "theoretical_loss": 3.551118549861947,
      "tokens_seen": 1342046208
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036830433137885384,
      "loss": 2.5816,
      "theoretical_loss": 3.5511029325339534,
      "tokens_seen": 1342111744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036828778135048233,
      "loss": 2.4761,
      "theoretical_loss": 3.5510873161820604,
      "tokens_seen": 1342177280
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036827123132211086,
      "loss": 2.6452,
      "theoretical_loss": 3.5510717008061587,
      "tokens_seen": 1342242816
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003682546812937394,
      "loss": 2.7212,
      "theoretical_loss": 3.55105608640614,
      "tokens_seen": 1342308352
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036823813126536794,
      "loss": 2.6361,
      "theoretical_loss": 3.5510404729818954,
      "tokens_seen": 1342373888
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003682215812369964,
      "loss": 2.6232,
      "theoretical_loss": 3.5510248605333166,
      "tokens_seen": 1342439424
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036820503120862496,
      "loss": 2.7721,
      "theoretical_loss": 3.5510092490602942,
      "tokens_seen": 1342504960
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003681884811802535,
      "loss": 2.8018,
      "theoretical_loss": 3.550993638562721,
      "tokens_seen": 1342570496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036817193115188193,
      "loss": 2.7502,
      "theoretical_loss": 3.550978029040487,
      "tokens_seen": 1342636032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036815538112351047,
      "loss": 2.8667,
      "theoretical_loss": 3.5509624204934847,
      "tokens_seen": 1342701568
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000368138831095139,
      "loss": 2.6597,
      "theoretical_loss": 3.5509468129216053,
      "tokens_seen": 1342767104
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003681222810667675,
      "loss": 2.9028,
      "theoretical_loss": 3.55093120632474,
      "tokens_seen": 1342832640
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000368105731038396,
      "loss": 2.53,
      "theoretical_loss": 3.5509156007027807,
      "tokens_seen": 1342898176
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036808918101002456,
      "loss": 2.7224,
      "theoretical_loss": 3.5508999960556182,
      "tokens_seen": 1342963712
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003680726309816531,
      "loss": 2.4666,
      "theoretical_loss": 3.550884392383145,
      "tokens_seen": 1343029248
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003680560809532816,
      "loss": 2.7222,
      "theoretical_loss": 3.550868789685252,
      "tokens_seen": 1343094784
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003680395309249101,
      "loss": 2.5682,
      "theoretical_loss": 3.5508531879618315,
      "tokens_seen": 1343160320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036802298089653866,
      "loss": 2.6142,
      "theoretical_loss": 3.550837587212774,
      "tokens_seen": 1343225856
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003680064308681672,
      "loss": 2.3747,
      "theoretical_loss": 3.5508219874379723,
      "tokens_seen": 1343291392
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003679898808397957,
      "loss": 2.5299,
      "theoretical_loss": 3.5508063886373176,
      "tokens_seen": 1343356928
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003679733308114242,
      "loss": 2.6225,
      "theoretical_loss": 3.5507907908107006,
      "tokens_seen": 1343422464
    },
    {
      "epoch": 0.48,
      "objective/train/avg_token_score": 0.015723444521427155,
      "objective/train/avg_weight": 0.9814993739128113,
      "objective/train/docs_used": 761574,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.639876365661621,
      "objective/train/std_weight": 0.07378087937831879,
      "objective/train/theoretical_loss": 3.5507751939580148,
      "objective/train/tokens_used": 1363948000,
      "theoretical_loss": 3.5507751939580148,
      "tokens_seen": 1343488000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036795678078305275,
      "loss": 2.5056,
      "theoretical_loss": 3.5507751939580148,
      "tokens_seen": 1343488000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003679402307546813,
      "loss": 2.6367,
      "theoretical_loss": 3.55075959807915,
      "tokens_seen": 1343553536
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003679236807263098,
      "loss": 2.7338,
      "theoretical_loss": 3.550744003173999,
      "tokens_seen": 1343619072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003679071306979383,
      "loss": 2.5226,
      "theoretical_loss": 3.5507284092424536,
      "tokens_seen": 1343684608
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036789058066956685,
      "loss": 2.5817,
      "theoretical_loss": 3.550712816284405,
      "tokens_seen": 1343750144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003678740306411954,
      "loss": 2.6333,
      "theoretical_loss": 3.550697224299745,
      "tokens_seen": 1343815680
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036785748061282387,
      "loss": 2.7179,
      "theoretical_loss": 3.5506816332883657,
      "tokens_seen": 1343881216
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003678409305844524,
      "loss": 2.621,
      "theoretical_loss": 3.5506660432501587,
      "tokens_seen": 1343946752
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036782438055608095,
      "loss": 2.9,
      "theoretical_loss": 3.550650454185016,
      "tokens_seen": 1344012288
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003678078305277095,
      "loss": 2.8022,
      "theoretical_loss": 3.550634866092829,
      "tokens_seen": 1344077824
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036779128049933797,
      "loss": 2.5675,
      "theoretical_loss": 3.55061927897349,
      "tokens_seen": 1344143360
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003677747304709665,
      "loss": 2.7269,
      "theoretical_loss": 3.5506036928268907,
      "tokens_seen": 1344208896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036775818044259504,
      "loss": 2.6523,
      "theoretical_loss": 3.5505881076529224,
      "tokens_seen": 1344274432
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003677416304142236,
      "loss": 2.5635,
      "theoretical_loss": 3.550572523451478,
      "tokens_seen": 1344339968
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036772508038585206,
      "loss": 2.3931,
      "theoretical_loss": 3.5505569402224486,
      "tokens_seen": 1344405504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003677085303574806,
      "loss": 2.556,
      "theoretical_loss": 3.550541357965727,
      "tokens_seen": 1344471040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036769198032910914,
      "loss": 2.5701,
      "theoretical_loss": 3.550525776681204,
      "tokens_seen": 1344536576
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003676754303007377,
      "loss": 2.408,
      "theoretical_loss": 3.5505101963687724,
      "tokens_seen": 1344602112
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036765888027236616,
      "loss": 3.0155,
      "theoretical_loss": 3.550494617028324,
      "tokens_seen": 1344667648
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003676423302439947,
      "loss": 2.582,
      "theoretical_loss": 3.5504790386597507,
      "tokens_seen": 1344733184
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036762578021562323,
      "loss": 2.5796,
      "theoretical_loss": 3.5504634612629444,
      "tokens_seen": 1344798720
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036760923018725177,
      "loss": 2.6473,
      "theoretical_loss": 3.5504478848377974,
      "tokens_seen": 1344864256
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036759268015888025,
      "loss": 2.7557,
      "theoretical_loss": 3.5504323093842016,
      "tokens_seen": 1344929792
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003675761301305088,
      "loss": 2.478,
      "theoretical_loss": 3.550416734902049,
      "tokens_seen": 1344995328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036755958010213733,
      "loss": 2.5838,
      "theoretical_loss": 3.550401161391232,
      "tokens_seen": 1345060864
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036754303007376587,
      "loss": 2.7273,
      "theoretical_loss": 3.550385588851642,
      "tokens_seen": 1345126400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036752648004539435,
      "loss": 2.7958,
      "theoretical_loss": 3.550370017283172,
      "tokens_seen": 1345191936
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003675099300170229,
      "loss": 2.6109,
      "theoretical_loss": 3.5503544466857133,
      "tokens_seen": 1345257472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003674933799886514,
      "loss": 2.6899,
      "theoretical_loss": 3.5503388770591586,
      "tokens_seen": 1345323008
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036747682996027996,
      "loss": 2.6712,
      "theoretical_loss": 3.5503233084034003,
      "tokens_seen": 1345388544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036746027993190845,
      "loss": 2.6238,
      "theoretical_loss": 3.5503077407183294,
      "tokens_seen": 1345454080
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000367443729903537,
      "loss": 2.7274,
      "theoretical_loss": 3.5502921740038396,
      "tokens_seen": 1345519616
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003674271798751655,
      "loss": 2.6233,
      "theoretical_loss": 3.5502766082598223,
      "tokens_seen": 1345585152
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036741062984679406,
      "loss": 2.6878,
      "theoretical_loss": 3.55026104348617,
      "tokens_seen": 1345650688
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036739407981842254,
      "loss": 2.3522,
      "theoretical_loss": 3.5502454796827743,
      "tokens_seen": 1345716224
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003673775297900511,
      "loss": 2.6317,
      "theoretical_loss": 3.5502299168495286,
      "tokens_seen": 1345781760
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003673609797616796,
      "loss": 2.1197,
      "theoretical_loss": 3.550214354986324,
      "tokens_seen": 1345847296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036734442973330815,
      "loss": 2.6951,
      "theoretical_loss": 3.5501987940930535,
      "tokens_seen": 1345912832
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036732787970493664,
      "loss": 2.5494,
      "theoretical_loss": 3.5501832341696096,
      "tokens_seen": 1345978368
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003673113296765652,
      "loss": 2.7921,
      "theoretical_loss": 3.550167675215884,
      "tokens_seen": 1346043904
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003672947796481937,
      "loss": 2.4449,
      "theoretical_loss": 3.55015211723177,
      "tokens_seen": 1346109440
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036727822961982225,
      "loss": 2.7882,
      "theoretical_loss": 3.5501365602171586,
      "tokens_seen": 1346174976
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036726167959145073,
      "loss": 2.785,
      "theoretical_loss": 3.5501210041719435,
      "tokens_seen": 1346240512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036724512956307927,
      "loss": 2.9472,
      "theoretical_loss": 3.550105449096016,
      "tokens_seen": 1346306048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003672285795347078,
      "loss": 2.5733,
      "theoretical_loss": 3.5500898949892696,
      "tokens_seen": 1346371584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036721202950633635,
      "loss": 2.7034,
      "theoretical_loss": 3.5500743418515963,
      "tokens_seen": 1346437120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003671954794779648,
      "loss": 2.5919,
      "theoretical_loss": 3.550058789682888,
      "tokens_seen": 1346502656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003671789294495933,
      "loss": 2.88,
      "theoretical_loss": 3.5500432384830383,
      "tokens_seen": 1346568192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036716237942122185,
      "loss": 3.1895,
      "theoretical_loss": 3.550027688251938,
      "tokens_seen": 1346633728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036714582939285033,
      "loss": 2.8276,
      "theoretical_loss": 3.550012138989482,
      "tokens_seen": 1346699264
    },
    {
      "epoch": 0.48,
      "objective/train/avg_token_score": 0.0235238429158926,
      "objective/train/avg_weight": 0.9723167419433594,
      "objective/train/docs_used": 763625,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.780709743499756,
      "objective/train/std_weight": 0.08779799193143845,
      "objective/train/theoretical_loss": 3.5499965906955606,
      "objective/train/tokens_used": 1367224800,
      "theoretical_loss": 3.5499965906955606,
      "tokens_seen": 1346764800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036712927936447887,
      "loss": 2.7977,
      "theoretical_loss": 3.5499965906955606,
      "tokens_seen": 1346764800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003671127293361074,
      "loss": 2.7767,
      "theoretical_loss": 3.549981043370068,
      "tokens_seen": 1346830336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036709617930773595,
      "loss": 2.7431,
      "theoretical_loss": 3.549965497012895,
      "tokens_seen": 1346895872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036707962927936443,
      "loss": 2.4468,
      "theoretical_loss": 3.549949951623936,
      "tokens_seen": 1346961408
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036706307925099297,
      "loss": 2.6259,
      "theoretical_loss": 3.5499344072030823,
      "tokens_seen": 1347026944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003670465292226215,
      "loss": 2.8369,
      "theoretical_loss": 3.5499188637502277,
      "tokens_seen": 1347092480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036702997919425004,
      "loss": 2.8164,
      "theoretical_loss": 3.5499033212652638,
      "tokens_seen": 1347158016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003670134291658785,
      "loss": 2.7869,
      "theoretical_loss": 3.549887779748083,
      "tokens_seen": 1347223552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036699687913750706,
      "loss": 2.7714,
      "theoretical_loss": 3.549872239198579,
      "tokens_seen": 1347289088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003669803291091356,
      "loss": 2.6872,
      "theoretical_loss": 3.5498566996166443,
      "tokens_seen": 1347354624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036696377908076414,
      "loss": 2.6734,
      "theoretical_loss": 3.549841161002171,
      "tokens_seen": 1347420160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003669472290523926,
      "loss": 2.6641,
      "theoretical_loss": 3.5498256233550523,
      "tokens_seen": 1347485696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036693067902402116,
      "loss": 2.7225,
      "theoretical_loss": 3.5498100866751807,
      "tokens_seen": 1347551232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003669141289956497,
      "loss": 2.6478,
      "theoretical_loss": 3.549794550962449,
      "tokens_seen": 1347616768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036689757896727823,
      "loss": 2.7229,
      "theoretical_loss": 3.5497790162167497,
      "tokens_seen": 1347682304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003668810289389067,
      "loss": 2.6556,
      "theoretical_loss": 3.5497634824379762,
      "tokens_seen": 1347747840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036686447891053526,
      "loss": 2.7814,
      "theoretical_loss": 3.5497479496260214,
      "tokens_seen": 1347813376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003668479288821638,
      "loss": 2.75,
      "theoretical_loss": 3.549732417780777,
      "tokens_seen": 1347878912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036683137885379233,
      "loss": 2.6383,
      "theoretical_loss": 3.5497168869021367,
      "tokens_seen": 1347944448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003668148288254208,
      "loss": 2.5357,
      "theoretical_loss": 3.549701356989994,
      "tokens_seen": 1348009984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036679827879704935,
      "loss": 2.547,
      "theoretical_loss": 3.54968582804424,
      "tokens_seen": 1348075520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003667817287686779,
      "loss": 2.5674,
      "theoretical_loss": 3.549670300064769,
      "tokens_seen": 1348141056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003667651787403064,
      "loss": 2.6517,
      "theoretical_loss": 3.549654773051473,
      "tokens_seen": 1348206592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003667486287119349,
      "loss": 2.6582,
      "theoretical_loss": 3.549639247004246,
      "tokens_seen": 1348272128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036673207868356345,
      "loss": 2.5233,
      "theoretical_loss": 3.54962372192298,
      "tokens_seen": 1348337664
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000366715528655192,
      "loss": 2.8008,
      "theoretical_loss": 3.549608197807568,
      "tokens_seen": 1348403200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003666989786268205,
      "loss": 2.7669,
      "theoretical_loss": 3.5495926746579034,
      "tokens_seen": 1348468736
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000366682428598449,
      "loss": 2.4284,
      "theoretical_loss": 3.5495771524738795,
      "tokens_seen": 1348534272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036666587857007754,
      "loss": 2.7123,
      "theoretical_loss": 3.5495616312553886,
      "tokens_seen": 1348599808
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003666493285417061,
      "loss": 2.611,
      "theoretical_loss": 3.549546111002324,
      "tokens_seen": 1348665344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003666327785133346,
      "loss": 2.7238,
      "theoretical_loss": 3.5495305917145785,
      "tokens_seen": 1348730880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003666162284849631,
      "loss": 2.8535,
      "theoretical_loss": 3.5495150733920457,
      "tokens_seen": 1348796416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036659967845659164,
      "loss": 2.5847,
      "theoretical_loss": 3.549499556034618,
      "tokens_seen": 1348861952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003665831284282202,
      "loss": 2.6113,
      "theoretical_loss": 3.5494840396421887,
      "tokens_seen": 1348927488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003665665783998487,
      "loss": 2.6856,
      "theoretical_loss": 3.5494685242146513,
      "tokens_seen": 1348993024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003665500283714772,
      "loss": 2.8321,
      "theoretical_loss": 3.549453009751899,
      "tokens_seen": 1349058560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036653347834310574,
      "loss": 2.6893,
      "theoretical_loss": 3.549437496253824,
      "tokens_seen": 1349124096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036651692831473427,
      "loss": 2.6666,
      "theoretical_loss": 3.549421983720321,
      "tokens_seen": 1349189632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003665003782863628,
      "loss": 2.5436,
      "theoretical_loss": 3.549406472151281,
      "tokens_seen": 1349255168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003664838282579913,
      "loss": 2.6838,
      "theoretical_loss": 3.5493909615465995,
      "tokens_seen": 1349320704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036646727822961983,
      "loss": 2.8752,
      "theoretical_loss": 3.549375451906168,
      "tokens_seen": 1349386240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036645072820124837,
      "loss": 2.7488,
      "theoretical_loss": 3.549359943229881,
      "tokens_seen": 1349451776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003664341781728769,
      "loss": 2.8098,
      "theoretical_loss": 3.5493444355176305,
      "tokens_seen": 1349517312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003664176281445054,
      "loss": 2.8577,
      "theoretical_loss": 3.549328928769311,
      "tokens_seen": 1349582848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003664010781161339,
      "loss": 2.8893,
      "theoretical_loss": 3.549313422984815,
      "tokens_seen": 1349648384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036638452808776246,
      "loss": 2.6215,
      "theoretical_loss": 3.5492979181640356,
      "tokens_seen": 1349713920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000366367978059391,
      "loss": 2.4901,
      "theoretical_loss": 3.549282414306867,
      "tokens_seen": 1349779456
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003663514280310195,
      "loss": 2.6171,
      "theoretical_loss": 3.549266911413201,
      "tokens_seen": 1349844992
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000366334878002648,
      "loss": 2.6922,
      "theoretical_loss": 3.549251409482933,
      "tokens_seen": 1349910528
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036631832797427656,
      "loss": 2.7454,
      "theoretical_loss": 3.5492359085159553,
      "tokens_seen": 1349976064
    },
    {
      "epoch": 0.48,
      "objective/train/avg_token_score": 0.01844686083495617,
      "objective/train/avg_weight": 0.9782994985580444,
      "objective/train/docs_used": 764744,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5705833435058594,
      "objective/train/std_weight": 0.07013782113790512,
      "objective/train/theoretical_loss": 3.549220408512161,
      "objective/train/tokens_used": 1370501600,
      "theoretical_loss": 3.549220408512161,
      "tokens_seen": 1350041600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003663017779459051,
      "loss": 2.5305,
      "theoretical_loss": 3.549220408512161,
      "tokens_seen": 1350041600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003662852279175336,
      "loss": 2.6398,
      "theoretical_loss": 3.549204909471444,
      "tokens_seen": 1350107136
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003662686778891621,
      "loss": 2.8733,
      "theoretical_loss": 3.549189411393697,
      "tokens_seen": 1350172672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036625212786079066,
      "loss": 2.5071,
      "theoretical_loss": 3.5491739142788146,
      "tokens_seen": 1350238208
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003662355778324192,
      "loss": 2.7741,
      "theoretical_loss": 3.549158418126689,
      "tokens_seen": 1350303744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003662190278040476,
      "loss": 2.7559,
      "theoretical_loss": 3.5491429229372153,
      "tokens_seen": 1350369280
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036620247777567616,
      "loss": 2.8016,
      "theoretical_loss": 3.549127428710285,
      "tokens_seen": 1350434816
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036618592774730464,
      "loss": 2.4882,
      "theoretical_loss": 3.5491119354457927,
      "tokens_seen": 1350500352
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003661693777189332,
      "loss": 2.6406,
      "theoretical_loss": 3.549096443143632,
      "tokens_seen": 1350565888
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003661528276905617,
      "loss": 2.6535,
      "theoretical_loss": 3.5490809518036963,
      "tokens_seen": 1350631424
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036613627766219026,
      "loss": 2.6445,
      "theoretical_loss": 3.5490654614258785,
      "tokens_seen": 1350696960
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036611972763381874,
      "loss": 2.4834,
      "theoretical_loss": 3.5490499720100734,
      "tokens_seen": 1350762496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003661031776054473,
      "loss": 2.8136,
      "theoretical_loss": 3.5490344835561736,
      "tokens_seen": 1350828032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003660866275770758,
      "loss": 2.7011,
      "theoretical_loss": 3.5490189960640732,
      "tokens_seen": 1350893568
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036607007754870435,
      "loss": 2.4238,
      "theoretical_loss": 3.5490035095336654,
      "tokens_seen": 1350959104
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036605352752033284,
      "loss": 2.4501,
      "theoretical_loss": 3.5489880239648444,
      "tokens_seen": 1351024640
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003660369774919614,
      "loss": 2.5157,
      "theoretical_loss": 3.5489725393575036,
      "tokens_seen": 1351090176
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003660204274635899,
      "loss": 2.719,
      "theoretical_loss": 3.548957055711536,
      "tokens_seen": 1351155712
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036600387743521845,
      "loss": 2.3552,
      "theoretical_loss": 3.548941573026836,
      "tokens_seen": 1351221248
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000365987327406847,
      "loss": 2.6818,
      "theoretical_loss": 3.548926091303297,
      "tokens_seen": 1351286784
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036597077737847547,
      "loss": 2.6765,
      "theoretical_loss": 3.5489106105408137,
      "tokens_seen": 1351352320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000365954227350104,
      "loss": 2.6151,
      "theoretical_loss": 3.548895130739279,
      "tokens_seen": 1351417856
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036593767732173254,
      "loss": 2.6054,
      "theoretical_loss": 3.548879651898586,
      "tokens_seen": 1351483392
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003659211272933611,
      "loss": 2.3057,
      "theoretical_loss": 3.5488641740186293,
      "tokens_seen": 1351548928
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036590457726498957,
      "loss": 2.8451,
      "theoretical_loss": 3.5488486970993027,
      "tokens_seen": 1351614464
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003658880272366181,
      "loss": 2.7507,
      "theoretical_loss": 3.5488332211405,
      "tokens_seen": 1351680000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036587147720824664,
      "loss": 2.6713,
      "theoretical_loss": 3.548817746142115,
      "tokens_seen": 1351745536
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003658549271798752,
      "loss": 2.8019,
      "theoretical_loss": 3.548802272104041,
      "tokens_seen": 1351811072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036583837715150366,
      "loss": 2.6216,
      "theoretical_loss": 3.548786799026172,
      "tokens_seen": 1351876608
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003658218271231322,
      "loss": 2.5386,
      "theoretical_loss": 3.548771326908403,
      "tokens_seen": 1351942144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036580527709476074,
      "loss": 2.8935,
      "theoretical_loss": 3.5487558557506267,
      "tokens_seen": 1352007680
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003657887270663893,
      "loss": 2.5493,
      "theoretical_loss": 3.548740385552737,
      "tokens_seen": 1352073216
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036577217703801776,
      "loss": 2.8507,
      "theoretical_loss": 3.5487249163146286,
      "tokens_seen": 1352138752
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003657556270096463,
      "loss": 2.6562,
      "theoretical_loss": 3.5487094480361945,
      "tokens_seen": 1352204288
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036573907698127483,
      "loss": 2.7335,
      "theoretical_loss": 3.5486939807173297,
      "tokens_seen": 1352269824
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036572252695290337,
      "loss": 2.7462,
      "theoretical_loss": 3.5486785143579276,
      "tokens_seen": 1352335360
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036570597692453185,
      "loss": 2.614,
      "theoretical_loss": 3.548663048957882,
      "tokens_seen": 1352400896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003656894268961604,
      "loss": 2.6885,
      "theoretical_loss": 3.548647584517087,
      "tokens_seen": 1352466432
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036567287686778893,
      "loss": 2.7807,
      "theoretical_loss": 3.5486321210354372,
      "tokens_seen": 1352531968
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036565632683941747,
      "loss": 2.8362,
      "theoretical_loss": 3.548616658512826,
      "tokens_seen": 1352597504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036563977681104595,
      "loss": 2.4422,
      "theoretical_loss": 3.5486011969491478,
      "tokens_seen": 1352663040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003656232267826745,
      "loss": 2.8822,
      "theoretical_loss": 3.548585736344296,
      "tokens_seen": 1352728576
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000365606676754303,
      "loss": 2.7601,
      "theoretical_loss": 3.5485702766981655,
      "tokens_seen": 1352794112
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036559012672593156,
      "loss": 2.6951,
      "theoretical_loss": 3.5485548180106505,
      "tokens_seen": 1352859648
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036557357669756005,
      "loss": 2.5606,
      "theoretical_loss": 3.5485393602816444,
      "tokens_seen": 1352925184
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003655570266691886,
      "loss": 2.4966,
      "theoretical_loss": 3.548523903511042,
      "tokens_seen": 1352990720
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003655404766408171,
      "loss": 2.6066,
      "theoretical_loss": 3.5485084476987367,
      "tokens_seen": 1353056256
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036552392661244566,
      "loss": 2.6143,
      "theoretical_loss": 3.548492992844623,
      "tokens_seen": 1353121792
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036550737658407414,
      "loss": 2.7247,
      "theoretical_loss": 3.548477538948596,
      "tokens_seen": 1353187328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003654908265557027,
      "loss": 2.6954,
      "theoretical_loss": 3.5484620860105487,
      "tokens_seen": 1353252864
    },
    {
      "epoch": 0.48,
      "objective/train/avg_token_score": 0.010542012751102448,
      "objective/train/avg_weight": 0.9875929355621338,
      "objective/train/docs_used": 766515,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7977170944213867,
      "objective/train/std_weight": 0.07056954503059387,
      "objective/train/theoretical_loss": 3.5484466340303755,
      "objective/train/tokens_used": 1373778400,
      "theoretical_loss": 3.5484466340303755,
      "tokens_seen": 1353318400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003654742765273312,
      "loss": 2.7142,
      "theoretical_loss": 3.5484466340303755,
      "tokens_seen": 1353318400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036545772649895975,
      "loss": 2.6175,
      "theoretical_loss": 3.5484311830079713,
      "tokens_seen": 1353383936
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036544117647058824,
      "loss": 2.4992,
      "theoretical_loss": 3.54841573294323,
      "tokens_seen": 1353449472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003654246264422168,
      "loss": 2.7963,
      "theoretical_loss": 3.5484002838360453,
      "tokens_seen": 1353515008
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003654080764138453,
      "loss": 2.5214,
      "theoretical_loss": 3.5483848356863126,
      "tokens_seen": 1353580544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036539152638547385,
      "loss": 2.672,
      "theoretical_loss": 3.5483693884939256,
      "tokens_seen": 1353646080
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036537497635710233,
      "loss": 2.8124,
      "theoretical_loss": 3.548353942258778,
      "tokens_seen": 1353711616
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036535842632873087,
      "loss": 2.6733,
      "theoretical_loss": 3.5483384969807656,
      "tokens_seen": 1353777152
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003653418763003594,
      "loss": 2.4578,
      "theoretical_loss": 3.548323052659782,
      "tokens_seen": 1353842688
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036532532627198795,
      "loss": 2.6855,
      "theoretical_loss": 3.548307609295721,
      "tokens_seen": 1353908224
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036530877624361643,
      "loss": 2.6698,
      "theoretical_loss": 3.548292166888478,
      "tokens_seen": 1353973760
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036529222621524497,
      "loss": 2.6806,
      "theoretical_loss": 3.5482767254379466,
      "tokens_seen": 1354039296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003652756761868735,
      "loss": 2.482,
      "theoretical_loss": 3.5482612849440223,
      "tokens_seen": 1354104832
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036525912615850204,
      "loss": 2.4361,
      "theoretical_loss": 3.5482458454065977,
      "tokens_seen": 1354170368
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003652425761301305,
      "loss": 2.7789,
      "theoretical_loss": 3.548230406825569,
      "tokens_seen": 1354235904
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000365226026101759,
      "loss": 2.6698,
      "theoretical_loss": 3.54821496920083,
      "tokens_seen": 1354301440
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003652094760733875,
      "loss": 2.5674,
      "theoretical_loss": 3.5481995325322755,
      "tokens_seen": 1354366976
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036519292604501603,
      "loss": 2.7846,
      "theoretical_loss": 3.5481840968197993,
      "tokens_seen": 1354432512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036517637601664457,
      "loss": 2.6708,
      "theoretical_loss": 3.5481686620632966,
      "tokens_seen": 1354498048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003651598259882731,
      "loss": 2.4929,
      "theoretical_loss": 3.548153228262662,
      "tokens_seen": 1354563584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003651432759599016,
      "loss": 2.5966,
      "theoretical_loss": 3.5481377954177895,
      "tokens_seen": 1354629120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003651267259315301,
      "loss": 2.5704,
      "theoretical_loss": 3.548122363528574,
      "tokens_seen": 1354694656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036511017590315866,
      "loss": 2.717,
      "theoretical_loss": 3.5481069325949104,
      "tokens_seen": 1354760192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003650936258747872,
      "loss": 2.691,
      "theoretical_loss": 3.5480915026166926,
      "tokens_seen": 1354825728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003650770758464157,
      "loss": 2.6159,
      "theoretical_loss": 3.5480760735938155,
      "tokens_seen": 1354891264
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003650605258180442,
      "loss": 2.652,
      "theoretical_loss": 3.5480606455261743,
      "tokens_seen": 1354956800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036504397578967276,
      "loss": 2.8236,
      "theoretical_loss": 3.5480452184136633,
      "tokens_seen": 1355022336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003650274257613013,
      "loss": 2.8429,
      "theoretical_loss": 3.5480297922561768,
      "tokens_seen": 1355087872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003650108757329298,
      "loss": 2.5424,
      "theoretical_loss": 3.54801436705361,
      "tokens_seen": 1355153408
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003649943257045583,
      "loss": 2.6473,
      "theoretical_loss": 3.547998942805857,
      "tokens_seen": 1355218944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036497777567618685,
      "loss": 2.8224,
      "theoretical_loss": 3.547983519512813,
      "tokens_seen": 1355284480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003649612256478154,
      "loss": 2.589,
      "theoretical_loss": 3.547968097174373,
      "tokens_seen": 1355350016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003649446756194439,
      "loss": 2.3951,
      "theoretical_loss": 3.547952675790431,
      "tokens_seen": 1355415552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003649281255910724,
      "loss": 2.8065,
      "theoretical_loss": 3.547937255360883,
      "tokens_seen": 1355481088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036491157556270095,
      "loss": 2.7104,
      "theoretical_loss": 3.5479218358856226,
      "tokens_seen": 1355546624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003648950255343295,
      "loss": 2.7182,
      "theoretical_loss": 3.547906417364545,
      "tokens_seen": 1355612160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036487847550595797,
      "loss": 2.8326,
      "theoretical_loss": 3.547890999797545,
      "tokens_seen": 1355677696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003648619254775865,
      "loss": 2.5056,
      "theoretical_loss": 3.5478755831845175,
      "tokens_seen": 1355743232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036484537544921505,
      "loss": 2.5174,
      "theoretical_loss": 3.547860167525357,
      "tokens_seen": 1355808768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003648288254208436,
      "loss": 2.8469,
      "theoretical_loss": 3.5478447528199597,
      "tokens_seen": 1355874304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036481227539247207,
      "loss": 2.6558,
      "theoretical_loss": 3.547829339068219,
      "tokens_seen": 1355939840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003647957253641006,
      "loss": 2.622,
      "theoretical_loss": 3.54781392627003,
      "tokens_seen": 1356005376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036477917533572914,
      "loss": 2.7631,
      "theoretical_loss": 3.5477985144252884,
      "tokens_seen": 1356070912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003647626253073577,
      "loss": 2.637,
      "theoretical_loss": 3.547783103533889,
      "tokens_seen": 1356136448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036474607527898616,
      "loss": 2.7724,
      "theoretical_loss": 3.547767693595726,
      "tokens_seen": 1356201984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003647295252506147,
      "loss": 2.9207,
      "theoretical_loss": 3.547752284610695,
      "tokens_seen": 1356267520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036471297522224324,
      "loss": 2.8172,
      "theoretical_loss": 3.547736876578691,
      "tokens_seen": 1356333056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003646964251938718,
      "loss": 2.6728,
      "theoretical_loss": 3.547721469499608,
      "tokens_seen": 1356398592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036467987516550026,
      "loss": 2.7111,
      "theoretical_loss": 3.547706063373343,
      "tokens_seen": 1356464128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003646633251371288,
      "loss": 2.4973,
      "theoretical_loss": 3.54769065819979,
      "tokens_seen": 1356529664
    },
    {
      "epoch": 0.48,
      "objective/train/avg_token_score": 0.008803254924714565,
      "objective/train/avg_weight": 0.9896420240402222,
      "objective/train/docs_used": 768508,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.580826997756958,
      "objective/train/std_weight": 0.05009785667061806,
      "objective/train/theoretical_loss": 3.547675253978843,
      "objective/train/tokens_used": 1377055200,
      "theoretical_loss": 3.547675253978843,
      "tokens_seen": 1356595200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036464677510875733,
      "loss": 2.5916,
      "theoretical_loss": 3.547675253978843,
      "tokens_seen": 1356595200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036463022508038587,
      "loss": 2.8118,
      "theoretical_loss": 3.5476598507103994,
      "tokens_seen": 1356660736
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036461367505201436,
      "loss": 2.6255,
      "theoretical_loss": 3.547644448394352,
      "tokens_seen": 1356726272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003645971250236429,
      "loss": 2.5424,
      "theoretical_loss": 3.547629047030597,
      "tokens_seen": 1356791808
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036458057499527143,
      "loss": 2.7022,
      "theoretical_loss": 3.5476136466190296,
      "tokens_seen": 1356857344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036456402496689997,
      "loss": 2.805,
      "theoretical_loss": 3.547598247159545,
      "tokens_seen": 1356922880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036454747493852845,
      "loss": 2.7702,
      "theoretical_loss": 3.5475828486520378,
      "tokens_seen": 1356988416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000364530924910157,
      "loss": 2.6378,
      "theoretical_loss": 3.5475674510964037,
      "tokens_seen": 1357053952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003645143748817855,
      "loss": 2.7186,
      "theoretical_loss": 3.547552054492537,
      "tokens_seen": 1357119488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036449782485341406,
      "loss": 2.2041,
      "theoretical_loss": 3.5475366588403343,
      "tokens_seen": 1357185024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003644812748250426,
      "loss": 2.672,
      "theoretical_loss": 3.5475212641396903,
      "tokens_seen": 1357250560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003644647247966711,
      "loss": 2.4685,
      "theoretical_loss": 3.5475058703905,
      "tokens_seen": 1357316096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003644481747682996,
      "loss": 2.7346,
      "theoretical_loss": 3.5474904775926586,
      "tokens_seen": 1357381632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036443162473992816,
      "loss": 2.6704,
      "theoretical_loss": 3.5474750857460613,
      "tokens_seen": 1357447168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003644150747115567,
      "loss": 2.6859,
      "theoretical_loss": 3.547459694850604,
      "tokens_seen": 1357512704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003643985246831852,
      "loss": 2.8542,
      "theoretical_loss": 3.5474443049061812,
      "tokens_seen": 1357578240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003643819746548137,
      "loss": 2.6951,
      "theoretical_loss": 3.5474289159126893,
      "tokens_seen": 1357643776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036436542462644226,
      "loss": 2.6854,
      "theoretical_loss": 3.5474135278700225,
      "tokens_seen": 1357709312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003643488745980708,
      "loss": 2.4387,
      "theoretical_loss": 3.5473981407780766,
      "tokens_seen": 1357774848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003643323245696993,
      "loss": 2.6024,
      "theoretical_loss": 3.5473827546367476,
      "tokens_seen": 1357840384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0003643157745413278,
      "loss": 2.6012,
      "theoretical_loss": 3.54736736944593,
      "tokens_seen": 1357905920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00036429922451295635,
      "loss": 2.6723,
      "theoretical_loss": 3.5473519852055193,
      "tokens_seen": 1357971456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003642826744845849,
      "loss": 2.4123,
      "theoretical_loss": 3.5473366019154113,
      "tokens_seen": 1358036992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036426612445621337,
      "loss": 2.7556,
      "theoretical_loss": 3.5473212195755015,
      "tokens_seen": 1358102528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036424957442784186,
      "loss": 2.5999,
      "theoretical_loss": 3.547305838185685,
      "tokens_seen": 1358168064
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036423302439947034,
      "loss": 2.7681,
      "theoretical_loss": 3.5472904577458575,
      "tokens_seen": 1358233600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003642164743710989,
      "loss": 2.8018,
      "theoretical_loss": 3.5472750782559146,
      "tokens_seen": 1358299136
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003641999243427274,
      "loss": 2.441,
      "theoretical_loss": 3.5472596997157515,
      "tokens_seen": 1358364672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036418337431435595,
      "loss": 2.8764,
      "theoretical_loss": 3.5472443221252634,
      "tokens_seen": 1358430208
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036416682428598444,
      "loss": 2.7306,
      "theoretical_loss": 3.547228945484347,
      "tokens_seen": 1358495744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000364150274257613,
      "loss": 2.6956,
      "theoretical_loss": 3.5472135697928966,
      "tokens_seen": 1358561280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003641337242292415,
      "loss": 2.4523,
      "theoretical_loss": 3.547198195050809,
      "tokens_seen": 1358626816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036411717420087005,
      "loss": 2.7266,
      "theoretical_loss": 3.5471828212579792,
      "tokens_seen": 1358692352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036410062417249853,
      "loss": 2.2899,
      "theoretical_loss": 3.547167448414302,
      "tokens_seen": 1358757888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036408407414412707,
      "loss": 2.7852,
      "theoretical_loss": 3.5471520765196747,
      "tokens_seen": 1358823424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003640675241157556,
      "loss": 2.7195,
      "theoretical_loss": 3.5471367055739913,
      "tokens_seen": 1358888960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036405097408738414,
      "loss": 2.607,
      "theoretical_loss": 3.547121335577148,
      "tokens_seen": 1358954496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036403442405901263,
      "loss": 2.4645,
      "theoretical_loss": 3.5471059665290414,
      "tokens_seen": 1359020032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036401787403064116,
      "loss": 2.8502,
      "theoretical_loss": 3.547090598429566,
      "tokens_seen": 1359085568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003640013240022697,
      "loss": 2.7327,
      "theoretical_loss": 3.547075231278618,
      "tokens_seen": 1359151104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036398477397389824,
      "loss": 2.9298,
      "theoretical_loss": 3.5470598650760934,
      "tokens_seen": 1359216640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003639682239455267,
      "loss": 2.8444,
      "theoretical_loss": 3.5470444998218875,
      "tokens_seen": 1359282176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036395167391715526,
      "loss": 2.7227,
      "theoretical_loss": 3.5470291355158956,
      "tokens_seen": 1359347712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003639351238887838,
      "loss": 2.8664,
      "theoretical_loss": 3.5470137721580146,
      "tokens_seen": 1359413248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036391857386041234,
      "loss": 2.609,
      "theoretical_loss": 3.5469984097481397,
      "tokens_seen": 1359478784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003639020238320408,
      "loss": 2.6673,
      "theoretical_loss": 3.546983048286167,
      "tokens_seen": 1359544320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036388547380366936,
      "loss": 2.6219,
      "theoretical_loss": 3.5469676877719913,
      "tokens_seen": 1359609856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003638689237752979,
      "loss": 2.5898,
      "theoretical_loss": 3.54695232820551,
      "tokens_seen": 1359675392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036385237374692643,
      "loss": 2.8898,
      "theoretical_loss": 3.5469369695866177,
      "tokens_seen": 1359740928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003638358237185549,
      "loss": 2.7038,
      "theoretical_loss": 3.546921611915211,
      "tokens_seen": 1359806464
    },
    {
      "epoch": 0.49,
      "objective/train/avg_token_score": 0.005285386927425861,
      "objective/train/avg_weight": 0.9937793016433716,
      "objective/train/docs_used": 770636,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7089648246765137,
      "objective/train/std_weight": 0.03277360275387764,
      "objective/train/theoretical_loss": 3.5469062551911854,
      "objective/train/tokens_used": 1380332000,
      "theoretical_loss": 3.5469062551911854,
      "tokens_seen": 1359872000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036381927369018345,
      "loss": 2.6772,
      "theoretical_loss": 3.5469062551911854,
      "tokens_seen": 1359872000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000363802723661812,
      "loss": 2.4801,
      "theoretical_loss": 3.5468908994144366,
      "tokens_seen": 1359937536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036378617363344053,
      "loss": 2.6154,
      "theoretical_loss": 3.546875544584861,
      "tokens_seen": 1360003072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000363769623605069,
      "loss": 2.6444,
      "theoretical_loss": 3.546860190702355,
      "tokens_seen": 1360068608
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036375307357669755,
      "loss": 2.741,
      "theoretical_loss": 3.5468448377668134,
      "tokens_seen": 1360134144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003637365235483261,
      "loss": 2.5474,
      "theoretical_loss": 3.5468294857781326,
      "tokens_seen": 1360199680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003637199735199546,
      "loss": 2.6036,
      "theoretical_loss": 3.546814134736209,
      "tokens_seen": 1360265216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003637034234915831,
      "loss": 2.5185,
      "theoretical_loss": 3.546798784640939,
      "tokens_seen": 1360330752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036368687346321164,
      "loss": 2.5556,
      "theoretical_loss": 3.5467834354922165,
      "tokens_seen": 1360396288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003636703234348402,
      "loss": 2.5847,
      "theoretical_loss": 3.54676808728994,
      "tokens_seen": 1360461824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003636537734064687,
      "loss": 2.5046,
      "theoretical_loss": 3.546752740034004,
      "tokens_seen": 1360527360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003636372233780972,
      "loss": 2.8553,
      "theoretical_loss": 3.546737393724305,
      "tokens_seen": 1360592896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036362067334972574,
      "loss": 2.8098,
      "theoretical_loss": 3.5467220483607393,
      "tokens_seen": 1360658432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003636041233213543,
      "loss": 2.5485,
      "theoretical_loss": 3.546706703943203,
      "tokens_seen": 1360723968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003635875732929828,
      "loss": 2.7704,
      "theoretical_loss": 3.546691360471592,
      "tokens_seen": 1360789504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003635710232646113,
      "loss": 2.7873,
      "theoretical_loss": 3.546676017945803,
      "tokens_seen": 1360855040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036355447323623984,
      "loss": 2.8359,
      "theoretical_loss": 3.5466606763657316,
      "tokens_seen": 1360920576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003635379232078684,
      "loss": 2.7442,
      "theoretical_loss": 3.5466453357312737,
      "tokens_seen": 1360986112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003635213731794969,
      "loss": 2.5953,
      "theoretical_loss": 3.5466299960423258,
      "tokens_seen": 1361051648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003635048231511254,
      "loss": 2.6553,
      "theoretical_loss": 3.5466146572987842,
      "tokens_seen": 1361117184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036348827312275393,
      "loss": 2.7731,
      "theoretical_loss": 3.5465993195005456,
      "tokens_seen": 1361182720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036347172309438247,
      "loss": 2.463,
      "theoretical_loss": 3.5465839826475047,
      "tokens_seen": 1361248256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000363455173066011,
      "loss": 2.7911,
      "theoretical_loss": 3.5465686467395594,
      "tokens_seen": 1361313792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003634386230376395,
      "loss": 2.6655,
      "theoretical_loss": 3.5465533117766053,
      "tokens_seen": 1361379328
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036342207300926803,
      "loss": 2.8292,
      "theoretical_loss": 3.5465379777585384,
      "tokens_seen": 1361444864
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036340552298089657,
      "loss": 2.8346,
      "theoretical_loss": 3.546522644685256,
      "tokens_seen": 1361510400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003633889729525251,
      "loss": 2.5641,
      "theoretical_loss": 3.5465073125566526,
      "tokens_seen": 1361575936
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003633724229241536,
      "loss": 2.6496,
      "theoretical_loss": 3.5464919813726263,
      "tokens_seen": 1361641472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003633558728957821,
      "loss": 2.8168,
      "theoretical_loss": 3.546476651133073,
      "tokens_seen": 1361707008
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036333932286741066,
      "loss": 2.8355,
      "theoretical_loss": 3.5464613218378886,
      "tokens_seen": 1361772544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003633227728390392,
      "loss": 2.7046,
      "theoretical_loss": 3.5464459934869694,
      "tokens_seen": 1361838080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003633062228106677,
      "loss": 2.6623,
      "theoretical_loss": 3.5464306660802123,
      "tokens_seen": 1361903616
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003632896727822962,
      "loss": 2.8341,
      "theoretical_loss": 3.5464153396175138,
      "tokens_seen": 1361969152
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036327312275392476,
      "loss": 2.8893,
      "theoretical_loss": 3.54640001409877,
      "tokens_seen": 1362034688
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003632565727255532,
      "loss": 2.6854,
      "theoretical_loss": 3.5463846895238773,
      "tokens_seen": 1362100224
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003632400226971817,
      "loss": 2.6969,
      "theoretical_loss": 3.5463693658927324,
      "tokens_seen": 1362165760
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036322347266881026,
      "loss": 2.6364,
      "theoretical_loss": 3.5463540432052314,
      "tokens_seen": 1362231296
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003632069226404388,
      "loss": 2.88,
      "theoretical_loss": 3.546338721461271,
      "tokens_seen": 1362296832
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003631903726120673,
      "loss": 2.6145,
      "theoretical_loss": 3.5463234006607482,
      "tokens_seen": 1362362368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003631738225836958,
      "loss": 2.4832,
      "theoretical_loss": 3.546308080803559,
      "tokens_seen": 1362427904
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036315727255532436,
      "loss": 2.8516,
      "theoretical_loss": 3.5462927618895996,
      "tokens_seen": 1362493440
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003631407225269529,
      "loss": 3.0372,
      "theoretical_loss": 3.546277443918768,
      "tokens_seen": 1362558976
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003631241724985814,
      "loss": 2.7938,
      "theoretical_loss": 3.5462621268909587,
      "tokens_seen": 1362624512
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003631076224702099,
      "loss": 2.5611,
      "theoretical_loss": 3.5462468108060703,
      "tokens_seen": 1362690048
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036309107244183845,
      "loss": 2.9011,
      "theoretical_loss": 3.5462314956639975,
      "tokens_seen": 1362755584
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000363074522413467,
      "loss": 2.6809,
      "theoretical_loss": 3.5462161814646387,
      "tokens_seen": 1362821120
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003630579723850955,
      "loss": 2.4828,
      "theoretical_loss": 3.5462008682078894,
      "tokens_seen": 1362886656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000363041422356724,
      "loss": 2.7571,
      "theoretical_loss": 3.546185555893647,
      "tokens_seen": 1362952192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036302487232835255,
      "loss": 2.2289,
      "theoretical_loss": 3.5461702445218073,
      "tokens_seen": 1363017728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003630083222999811,
      "loss": 2.7076,
      "theoretical_loss": 3.5461549340922676,
      "tokens_seen": 1363083264
    },
    {
      "epoch": 0.49,
      "objective/train/avg_token_score": 0.025066375732421875,
      "objective/train/avg_weight": 0.9705035090446472,
      "objective/train/docs_used": 772517,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.015625476837158,
      "objective/train/std_weight": 0.09583219885826111,
      "objective/train/theoretical_loss": 3.5461396246049244,
      "objective/train/tokens_used": 1383608800,
      "theoretical_loss": 3.5461396246049244,
      "tokens_seen": 1363148800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036299177227160957,
      "loss": 2.4821,
      "theoretical_loss": 3.5461396246049244,
      "tokens_seen": 1363148800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003629752222432381,
      "loss": 2.6428,
      "theoretical_loss": 3.546124316059675,
      "tokens_seen": 1363214336
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036295867221486665,
      "loss": 2.5063,
      "theoretical_loss": 3.546109008456415,
      "tokens_seen": 1363279872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003629421221864952,
      "loss": 2.5206,
      "theoretical_loss": 3.5460937017950425,
      "tokens_seen": 1363345408
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036292557215812367,
      "loss": 2.3559,
      "theoretical_loss": 3.546078396075453,
      "tokens_seen": 1363410944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003629090221297522,
      "loss": 2.8094,
      "theoretical_loss": 3.546063091297544,
      "tokens_seen": 1363476480
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036289247210138074,
      "loss": 2.5552,
      "theoretical_loss": 3.5460477874612124,
      "tokens_seen": 1363542016
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003628759220730093,
      "loss": 2.8484,
      "theoretical_loss": 3.5460324845663544,
      "tokens_seen": 1363607552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036285937204463776,
      "loss": 2.6879,
      "theoretical_loss": 3.5460171826128675,
      "tokens_seen": 1363673088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003628428220162663,
      "loss": 2.513,
      "theoretical_loss": 3.5460018816006484,
      "tokens_seen": 1363738624
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036282627198789484,
      "loss": 2.489,
      "theoretical_loss": 3.5459865815295935,
      "tokens_seen": 1363804160
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003628097219595234,
      "loss": 2.5767,
      "theoretical_loss": 3.5459712823996004,
      "tokens_seen": 1363869696
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036279317193115186,
      "loss": 2.4519,
      "theoretical_loss": 3.5459559842105657,
      "tokens_seen": 1363935232
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003627766219027804,
      "loss": 2.8003,
      "theoretical_loss": 3.5459406869623855,
      "tokens_seen": 1364000768
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036276007187440893,
      "loss": 2.8867,
      "theoretical_loss": 3.5459253906549586,
      "tokens_seen": 1364066304
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036274352184603747,
      "loss": 2.465,
      "theoretical_loss": 3.54591009528818,
      "tokens_seen": 1364131840
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036272697181766595,
      "loss": 2.7076,
      "theoretical_loss": 3.545894800861948,
      "tokens_seen": 1364197376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003627104217892945,
      "loss": 2.72,
      "theoretical_loss": 3.545879507376159,
      "tokens_seen": 1364262912
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036269387176092303,
      "loss": 2.791,
      "theoretical_loss": 3.54586421483071,
      "tokens_seen": 1364328448
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036267732173255157,
      "loss": 2.5989,
      "theoretical_loss": 3.545848923225498,
      "tokens_seen": 1364393984
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036266077170418005,
      "loss": 2.8472,
      "theoretical_loss": 3.5458336325604205,
      "tokens_seen": 1364459520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003626442216758086,
      "loss": 2.5893,
      "theoretical_loss": 3.545818342835374,
      "tokens_seen": 1364525056
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003626276716474371,
      "loss": 2.6061,
      "theoretical_loss": 3.545803054050256,
      "tokens_seen": 1364590592
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036261112161906566,
      "loss": 2.803,
      "theoretical_loss": 3.5457877662049633,
      "tokens_seen": 1364656128
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036259457159069415,
      "loss": 2.3203,
      "theoretical_loss": 3.545772479299393,
      "tokens_seen": 1364721664
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003625780215623227,
      "loss": 2.557,
      "theoretical_loss": 3.5457571933334417,
      "tokens_seen": 1364787200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003625614715339512,
      "loss": 2.8129,
      "theoretical_loss": 3.545741908307008,
      "tokens_seen": 1364852736
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036254492150557976,
      "loss": 2.8223,
      "theoretical_loss": 3.545726624219988,
      "tokens_seen": 1364918272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036252837147720824,
      "loss": 2.7234,
      "theoretical_loss": 3.5457113410722787,
      "tokens_seen": 1364983808
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003625118214488368,
      "loss": 2.6809,
      "theoretical_loss": 3.5456960588637774,
      "tokens_seen": 1365049344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003624952714204653,
      "loss": 2.8217,
      "theoretical_loss": 3.5456807775943817,
      "tokens_seen": 1365114880
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036247872139209386,
      "loss": 2.6003,
      "theoretical_loss": 3.5456654972639883,
      "tokens_seen": 1365180416
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036246217136372234,
      "loss": 2.8437,
      "theoretical_loss": 3.5456502178724953,
      "tokens_seen": 1365245952
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003624456213353509,
      "loss": 2.5412,
      "theoretical_loss": 3.5456349394197986,
      "tokens_seen": 1365311488
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003624290713069794,
      "loss": 2.7445,
      "theoretical_loss": 3.545619661905797,
      "tokens_seen": 1365377024
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036241252127860795,
      "loss": 2.5717,
      "theoretical_loss": 3.5456043853303862,
      "tokens_seen": 1365442560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036239597125023643,
      "loss": 2.9699,
      "theoretical_loss": 3.545589109693465,
      "tokens_seen": 1365508096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036237942122186497,
      "loss": 2.3503,
      "theoretical_loss": 3.54557383499493,
      "tokens_seen": 1365573632
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003623628711934935,
      "loss": 2.642,
      "theoretical_loss": 3.545558561234678,
      "tokens_seen": 1365639168
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036234632116512205,
      "loss": 2.8237,
      "theoretical_loss": 3.545543288412607,
      "tokens_seen": 1365704704
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036232977113675053,
      "loss": 2.3341,
      "theoretical_loss": 3.5455280165286145,
      "tokens_seen": 1365770240
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036231322110837907,
      "loss": 2.5979,
      "theoretical_loss": 3.545512745582597,
      "tokens_seen": 1365835776
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003622966710800076,
      "loss": 2.7419,
      "theoretical_loss": 3.545497475574453,
      "tokens_seen": 1365901312
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036228012105163603,
      "loss": 2.7659,
      "theoretical_loss": 3.5454822065040794,
      "tokens_seen": 1365966848
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036226357102326457,
      "loss": 2.5523,
      "theoretical_loss": 3.5454669383713737,
      "tokens_seen": 1366032384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003622470209948931,
      "loss": 2.4238,
      "theoretical_loss": 3.5454516711762327,
      "tokens_seen": 1366097920
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003622304709665216,
      "loss": 2.5622,
      "theoretical_loss": 3.5454364049185547,
      "tokens_seen": 1366163456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036221392093815013,
      "loss": 2.8506,
      "theoretical_loss": 3.545421139598237,
      "tokens_seen": 1366228992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036219737090977867,
      "loss": 2.9117,
      "theoretical_loss": 3.5454058752151765,
      "tokens_seen": 1366294528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003621808208814072,
      "loss": 2.6463,
      "theoretical_loss": 3.5453906117692715,
      "tokens_seen": 1366360064
    },
    {
      "epoch": 0.49,
      "objective/train/avg_token_score": 0.007199227809906006,
      "objective/train/avg_weight": 0.9915366768836975,
      "objective/train/docs_used": 774661,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.045898914337158,
      "objective/train/std_weight": 0.05738738551735878,
      "objective/train/theoretical_loss": 3.545375349260419,
      "objective/train/tokens_used": 1386885600,
      "theoretical_loss": 3.545375349260419,
      "tokens_seen": 1366425600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003621642708530357,
      "loss": 2.7276,
      "theoretical_loss": 3.545375349260419,
      "tokens_seen": 1366425600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003621477208246642,
      "loss": 2.6449,
      "theoretical_loss": 3.545360087688517,
      "tokens_seen": 1366491136
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036213117079629276,
      "loss": 2.9476,
      "theoretical_loss": 3.545344827053462,
      "tokens_seen": 1366556672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003621146207679213,
      "loss": 2.7794,
      "theoretical_loss": 3.5453295673551537,
      "tokens_seen": 1366622208
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003620980707395498,
      "loss": 2.7547,
      "theoretical_loss": 3.545314308593487,
      "tokens_seen": 1366687744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003620815207111783,
      "loss": 2.7879,
      "theoretical_loss": 3.5452990507683615,
      "tokens_seen": 1366753280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036206497068280686,
      "loss": 2.8971,
      "theoretical_loss": 3.545283793879674,
      "tokens_seen": 1366818816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003620484206544354,
      "loss": 2.7253,
      "theoretical_loss": 3.545268537927322,
      "tokens_seen": 1366884352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003620318706260639,
      "loss": 2.7849,
      "theoretical_loss": 3.5452532829112036,
      "tokens_seen": 1366949888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003620153205976924,
      "loss": 2.6831,
      "theoretical_loss": 3.545238028831216,
      "tokens_seen": 1367015424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036199877056932096,
      "loss": 2.614,
      "theoretical_loss": 3.5452227756872574,
      "tokens_seen": 1367080960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003619822205409495,
      "loss": 2.4576,
      "theoretical_loss": 3.545207523479225,
      "tokens_seen": 1367146496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036196567051257803,
      "loss": 2.7242,
      "theoretical_loss": 3.545192272207017,
      "tokens_seen": 1367212032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003619491204842065,
      "loss": 2.9942,
      "theoretical_loss": 3.545177021870531,
      "tokens_seen": 1367277568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036193257045583505,
      "loss": 2.6425,
      "theoretical_loss": 3.5451617724696645,
      "tokens_seen": 1367343104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003619160204274636,
      "loss": 2.9093,
      "theoretical_loss": 3.5451465240043145,
      "tokens_seen": 1367408640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036189947039909213,
      "loss": 2.7233,
      "theoretical_loss": 3.5451312764743808,
      "tokens_seen": 1367474176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003618829203707206,
      "loss": 2.5458,
      "theoretical_loss": 3.5451160298797593,
      "tokens_seen": 1367539712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036186637034234915,
      "loss": 2.839,
      "theoretical_loss": 3.545100784220349,
      "tokens_seen": 1367605248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003618498203139777,
      "loss": 2.7678,
      "theoretical_loss": 3.545085539496047,
      "tokens_seen": 1367670784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003618332702856062,
      "loss": 2.6571,
      "theoretical_loss": 3.5450702957067515,
      "tokens_seen": 1367736320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003618167202572347,
      "loss": 2.4961,
      "theoretical_loss": 3.5450550528523603,
      "tokens_seen": 1367801856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036180017022886324,
      "loss": 2.7217,
      "theoretical_loss": 3.545039810932771,
      "tokens_seen": 1367867392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003617836202004918,
      "loss": 2.704,
      "theoretical_loss": 3.5450245699478824,
      "tokens_seen": 1367932928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003617670701721203,
      "loss": 2.4586,
      "theoretical_loss": 3.5450093298975913,
      "tokens_seen": 1367998464
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003617505201437488,
      "loss": 2.9027,
      "theoretical_loss": 3.544994090781796,
      "tokens_seen": 1368064000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036173397011537734,
      "loss": 2.3664,
      "theoretical_loss": 3.5449788526003942,
      "tokens_seen": 1368129536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003617174200870059,
      "loss": 2.6001,
      "theoretical_loss": 3.5449636153532844,
      "tokens_seen": 1368195072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003617008700586344,
      "loss": 2.5439,
      "theoretical_loss": 3.544948379040364,
      "tokens_seen": 1368260608
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003616843200302629,
      "loss": 2.7908,
      "theoretical_loss": 3.544933143661532,
      "tokens_seen": 1368326144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036166777000189144,
      "loss": 2.6706,
      "theoretical_loss": 3.5449179092166854,
      "tokens_seen": 1368391680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036165121997352,
      "loss": 2.8731,
      "theoretical_loss": 3.5449026757057225,
      "tokens_seen": 1368457216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003616346699451485,
      "loss": 2.7294,
      "theoretical_loss": 3.544887443128541,
      "tokens_seen": 1368522752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000361618119916777,
      "loss": 2.9094,
      "theoretical_loss": 3.5448722114850395,
      "tokens_seen": 1368588288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036160156988840553,
      "loss": 2.6615,
      "theoretical_loss": 3.544856980775116,
      "tokens_seen": 1368653824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036158501986003407,
      "loss": 2.6185,
      "theoretical_loss": 3.5448417509986685,
      "tokens_seen": 1368719360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003615684698316626,
      "loss": 2.7744,
      "theoretical_loss": 3.544826522155595,
      "tokens_seen": 1368784896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003615519198032911,
      "loss": 2.5556,
      "theoretical_loss": 3.544811294245793,
      "tokens_seen": 1368850432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036153536977491963,
      "loss": 2.6534,
      "theoretical_loss": 3.5447960672691616,
      "tokens_seen": 1368915968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036151881974654817,
      "loss": 2.6889,
      "theoretical_loss": 3.5447808412255988,
      "tokens_seen": 1368981504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003615022697181767,
      "loss": 2.4367,
      "theoretical_loss": 3.5447656161150025,
      "tokens_seen": 1369047040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003614857196898052,
      "loss": 2.5842,
      "theoretical_loss": 3.544750391937271,
      "tokens_seen": 1369112576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003614691696614337,
      "loss": 2.7075,
      "theoretical_loss": 3.5447351686923025,
      "tokens_seen": 1369178112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036145261963306226,
      "loss": 2.7738,
      "theoretical_loss": 3.544719946379995,
      "tokens_seen": 1369243648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003614360696046908,
      "loss": 2.6772,
      "theoretical_loss": 3.544704725000247,
      "tokens_seen": 1369309184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003614195195763193,
      "loss": 2.9124,
      "theoretical_loss": 3.5446895045529567,
      "tokens_seen": 1369374720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003614029695479478,
      "loss": 2.6629,
      "theoretical_loss": 3.5446742850380217,
      "tokens_seen": 1369440256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036138641951957636,
      "loss": 2.3681,
      "theoretical_loss": 3.544659066455341,
      "tokens_seen": 1369505792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003613698694912049,
      "loss": 2.8675,
      "theoretical_loss": 3.5446438488048133,
      "tokens_seen": 1369571328
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003613533194628334,
      "loss": 2.7889,
      "theoretical_loss": 3.5446286320863356,
      "tokens_seen": 1369636864
    },
    {
      "epoch": 0.49,
      "objective/train/avg_token_score": 0.010788600891828537,
      "objective/train/avg_weight": 0.9872981309890747,
      "objective/train/docs_used": 776810,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.66348934173584,
      "objective/train/std_weight": 0.05296250432729721,
      "objective/train/theoretical_loss": 3.544613416299808,
      "objective/train/tokens_used": 1390162400,
      "theoretical_loss": 3.544613416299808,
      "tokens_seen": 1369702400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003613367694344619,
      "loss": 2.6663,
      "theoretical_loss": 3.544613416299808,
      "tokens_seen": 1369702400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036132021940609045,
      "loss": 2.8279,
      "theoretical_loss": 3.544598201445127,
      "tokens_seen": 1369767936
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003613036693777189,
      "loss": 2.6209,
      "theoretical_loss": 3.5445829875221917,
      "tokens_seen": 1369833472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003612871193493474,
      "loss": 2.7098,
      "theoretical_loss": 3.5445677745309006,
      "tokens_seen": 1369899008
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036127056932097596,
      "loss": 2.6481,
      "theoretical_loss": 3.544552562471152,
      "tokens_seen": 1369964544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036125401929260444,
      "loss": 2.6,
      "theoretical_loss": 3.544537351342844,
      "tokens_seen": 1370030080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000361237469264233,
      "loss": 2.7093,
      "theoretical_loss": 3.544522141145876,
      "tokens_seen": 1370095616
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003612209192358615,
      "loss": 2.7892,
      "theoretical_loss": 3.5445069318801457,
      "tokens_seen": 1370161152
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036120436920749005,
      "loss": 2.8262,
      "theoretical_loss": 3.544491723545551,
      "tokens_seen": 1370226688
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036118781917911854,
      "loss": 2.4097,
      "theoretical_loss": 3.5444765161419913,
      "tokens_seen": 1370292224
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003611712691507471,
      "loss": 2.6467,
      "theoretical_loss": 3.5444613096693645,
      "tokens_seen": 1370357760
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003611547191223756,
      "loss": 2.6886,
      "theoretical_loss": 3.5444461041275694,
      "tokens_seen": 1370423296
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036113816909400415,
      "loss": 2.765,
      "theoretical_loss": 3.5444308995165046,
      "tokens_seen": 1370488832
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036112161906563263,
      "loss": 2.5502,
      "theoretical_loss": 3.544415695836068,
      "tokens_seen": 1370554368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036110506903726117,
      "loss": 2.6103,
      "theoretical_loss": 3.544400493086159,
      "tokens_seen": 1370619904
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003610885190088897,
      "loss": 2.8766,
      "theoretical_loss": 3.544385291266676,
      "tokens_seen": 1370685440
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036107196898051825,
      "loss": 2.6628,
      "theoretical_loss": 3.544370090377517,
      "tokens_seen": 1370750976
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036105541895214673,
      "loss": 2.6354,
      "theoretical_loss": 3.544354890418581,
      "tokens_seen": 1370816512
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036103886892377527,
      "loss": 2.405,
      "theoretical_loss": 3.5443396913897667,
      "tokens_seen": 1370882048
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003610223188954038,
      "loss": 2.7222,
      "theoretical_loss": 3.544324493290972,
      "tokens_seen": 1370947584
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036100576886703234,
      "loss": 2.6448,
      "theoretical_loss": 3.5443092961220968,
      "tokens_seen": 1371013120
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003609892188386608,
      "loss": 2.7145,
      "theoretical_loss": 3.5442940998830386,
      "tokens_seen": 1371078656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036097266881028936,
      "loss": 2.6558,
      "theoretical_loss": 3.544278904573697,
      "tokens_seen": 1371144192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003609561187819179,
      "loss": 2.7794,
      "theoretical_loss": 3.54426371019397,
      "tokens_seen": 1371209728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036093956875354644,
      "loss": 2.7255,
      "theoretical_loss": 3.544248516743756,
      "tokens_seen": 1371275264
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003609230187251749,
      "loss": 2.6139,
      "theoretical_loss": 3.5442333242229545,
      "tokens_seen": 1371340800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036090646869680346,
      "loss": 2.6866,
      "theoretical_loss": 3.544218132631464,
      "tokens_seen": 1371406336
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000360889918668432,
      "loss": 2.6276,
      "theoretical_loss": 3.5442029419691834,
      "tokens_seen": 1371471872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036087336864006053,
      "loss": 2.5969,
      "theoretical_loss": 3.5441877522360112,
      "tokens_seen": 1371537408
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000360856818611689,
      "loss": 2.582,
      "theoretical_loss": 3.544172563431846,
      "tokens_seen": 1371602944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036084026858331755,
      "loss": 2.8347,
      "theoretical_loss": 3.544157375556587,
      "tokens_seen": 1371668480
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003608237185549461,
      "loss": 2.6797,
      "theoretical_loss": 3.5441421886101336,
      "tokens_seen": 1371734016
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036080716852657463,
      "loss": 2.6863,
      "theoretical_loss": 3.544127002592383,
      "tokens_seen": 1371799552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003607906184982031,
      "loss": 2.7014,
      "theoretical_loss": 3.544111817503235,
      "tokens_seen": 1371865088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036077406846983165,
      "loss": 2.5948,
      "theoretical_loss": 3.5440966333425887,
      "tokens_seen": 1371930624
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003607575184414602,
      "loss": 2.7861,
      "theoretical_loss": 3.544081450110342,
      "tokens_seen": 1371996160
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003607409684130887,
      "loss": 2.7939,
      "theoretical_loss": 3.5440662678063957,
      "tokens_seen": 1372061696
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003607244183847172,
      "loss": 2.5788,
      "theoretical_loss": 3.5440510864306467,
      "tokens_seen": 1372127232
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036070786835634575,
      "loss": 2.7254,
      "theoretical_loss": 3.544035905982995,
      "tokens_seen": 1372192768
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003606913183279743,
      "loss": 2.5921,
      "theoretical_loss": 3.544020726463339,
      "tokens_seen": 1372258304
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003606747682996028,
      "loss": 2.6899,
      "theoretical_loss": 3.5440055478715777,
      "tokens_seen": 1372323840
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003606582182712313,
      "loss": 2.5207,
      "theoretical_loss": 3.5439903702076108,
      "tokens_seen": 1372389376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036064166824285984,
      "loss": 2.7677,
      "theoretical_loss": 3.5439751934713364,
      "tokens_seen": 1372454912
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003606251182144884,
      "loss": 2.5852,
      "theoretical_loss": 3.543960017662654,
      "tokens_seen": 1372520448
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003606085681861169,
      "loss": 2.6374,
      "theoretical_loss": 3.543944842781462,
      "tokens_seen": 1372585984
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003605920181577454,
      "loss": 2.7574,
      "theoretical_loss": 3.5439296688276607,
      "tokens_seen": 1372651520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036057546812937394,
      "loss": 2.5711,
      "theoretical_loss": 3.5439144958011477,
      "tokens_seen": 1372717056
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003605589181010025,
      "loss": 2.766,
      "theoretical_loss": 3.5438993237018233,
      "tokens_seen": 1372782592
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000360542368072631,
      "loss": 2.6738,
      "theoretical_loss": 3.5438841525295857,
      "tokens_seen": 1372848128
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003605258180442595,
      "loss": 2.6849,
      "theoretical_loss": 3.5438689822843346,
      "tokens_seen": 1372913664
    },
    {
      "epoch": 0.49,
      "objective/train/avg_token_score": 0.024327285587787628,
      "objective/train/avg_weight": 0.971400797367096,
      "objective/train/docs_used": 778576,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.205732583999634,
      "objective/train/std_weight": 0.10012760013341904,
      "objective/train/theoretical_loss": 3.5438538129659687,
      "objective/train/tokens_used": 1393439200,
      "theoretical_loss": 3.5438538129659687,
      "tokens_seen": 1372979200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036050926801588803,
      "loss": 2.7522,
      "theoretical_loss": 3.5438538129659687,
      "tokens_seen": 1372979200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036049271798751657,
      "loss": 2.6485,
      "theoretical_loss": 3.543838644574387,
      "tokens_seen": 1373044736
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003604761679591451,
      "loss": 2.6241,
      "theoretical_loss": 3.543823477109489,
      "tokens_seen": 1373110272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036045961793077365,
      "loss": 2.8674,
      "theoretical_loss": 3.543808310571174,
      "tokens_seen": 1373175808
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036044306790240213,
      "loss": 2.8041,
      "theoretical_loss": 3.5437931449593405,
      "tokens_seen": 1373241344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036042651787403067,
      "loss": 2.6277,
      "theoretical_loss": 3.5437779802738887,
      "tokens_seen": 1373306880
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003604099678456592,
      "loss": 2.4642,
      "theoretical_loss": 3.543762816514717,
      "tokens_seen": 1373372416
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036039341781728774,
      "loss": 2.9267,
      "theoretical_loss": 3.5437476536817245,
      "tokens_seen": 1373437952
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003603768677889162,
      "loss": 2.8,
      "theoretical_loss": 3.5437324917748114,
      "tokens_seen": 1373503488
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036036031776054476,
      "loss": 2.8115,
      "theoretical_loss": 3.543717330793876,
      "tokens_seen": 1373569024
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003603437677321733,
      "loss": 2.6837,
      "theoretical_loss": 3.543702170738818,
      "tokens_seen": 1373634560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036032721770380184,
      "loss": 2.5948,
      "theoretical_loss": 3.5436870116095367,
      "tokens_seen": 1373700096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036031066767543027,
      "loss": 2.7361,
      "theoretical_loss": 3.543671853405931,
      "tokens_seen": 1373765632
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003602941176470588,
      "loss": 2.3914,
      "theoretical_loss": 3.543656696127901,
      "tokens_seen": 1373831168
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003602775676186873,
      "loss": 2.7093,
      "theoretical_loss": 3.5436415397753454,
      "tokens_seen": 1373896704
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003602610175903158,
      "loss": 2.6767,
      "theoretical_loss": 3.5436263843481637,
      "tokens_seen": 1373962240
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036024446756194436,
      "loss": 2.8837,
      "theoretical_loss": 3.543611229846255,
      "tokens_seen": 1374027776
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003602279175335729,
      "loss": 2.4942,
      "theoretical_loss": 3.5435960762695196,
      "tokens_seen": 1374093312
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003602113675052014,
      "loss": 2.7282,
      "theoretical_loss": 3.543580923617856,
      "tokens_seen": 1374158848
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003601948174768299,
      "loss": 2.6097,
      "theoretical_loss": 3.5435657718911635,
      "tokens_seen": 1374224384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036017826744845846,
      "loss": 2.6547,
      "theoretical_loss": 3.5435506210893424,
      "tokens_seen": 1374289920
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000360161717420087,
      "loss": 2.4313,
      "theoretical_loss": 3.5435354712122917,
      "tokens_seen": 1374355456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003601451673917155,
      "loss": 2.9995,
      "theoretical_loss": 3.5435203222599103,
      "tokens_seen": 1374420992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000360128617363344,
      "loss": 2.7245,
      "theoretical_loss": 3.5435051742320987,
      "tokens_seen": 1374486528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036011206733497256,
      "loss": 2.9622,
      "theoretical_loss": 3.5434900271287555,
      "tokens_seen": 1374552064
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003600955173066011,
      "loss": 2.7367,
      "theoretical_loss": 3.5434748809497814,
      "tokens_seen": 1374617600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003600789672782296,
      "loss": 2.7636,
      "theoretical_loss": 3.543459735695074,
      "tokens_seen": 1374683136
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003600624172498581,
      "loss": 2.6642,
      "theoretical_loss": 3.543444591364535,
      "tokens_seen": 1374748672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036004586722148665,
      "loss": 2.6932,
      "theoretical_loss": 3.5434294479580624,
      "tokens_seen": 1374814208
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003600293171931152,
      "loss": 2.7006,
      "theoretical_loss": 3.5434143054755562,
      "tokens_seen": 1374879744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00036001276716474367,
      "loss": 2.7633,
      "theoretical_loss": 3.5433991639169165,
      "tokens_seen": 1374945280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003599962171363722,
      "loss": 2.509,
      "theoretical_loss": 3.5433840232820417,
      "tokens_seen": 1375010816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035997966710800075,
      "loss": 2.6815,
      "theoretical_loss": 3.5433688835708326,
      "tokens_seen": 1375076352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003599631170796293,
      "loss": 2.6307,
      "theoretical_loss": 3.5433537447831887,
      "tokens_seen": 1375141888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035994656705125777,
      "loss": 2.4795,
      "theoretical_loss": 3.543338606919009,
      "tokens_seen": 1375207424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003599300170228863,
      "loss": 2.4391,
      "theoretical_loss": 3.543323469978194,
      "tokens_seen": 1375272960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035991346699451484,
      "loss": 2.7279,
      "theoretical_loss": 3.5433083339606424,
      "tokens_seen": 1375338496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003598969169661434,
      "loss": 2.7959,
      "theoretical_loss": 3.543293198866255,
      "tokens_seen": 1375404032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035988036693777186,
      "loss": 2.8758,
      "theoretical_loss": 3.5432780646949302,
      "tokens_seen": 1375469568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003598638169094004,
      "loss": 2.6181,
      "theoretical_loss": 3.543262931446569,
      "tokens_seen": 1375535104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035984726688102894,
      "loss": 3.0675,
      "theoretical_loss": 3.5432477991210702,
      "tokens_seen": 1375600640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003598307168526575,
      "loss": 2.6443,
      "theoretical_loss": 3.5432326677183346,
      "tokens_seen": 1375666176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035981416682428596,
      "loss": 2.5577,
      "theoretical_loss": 3.5432175372382604,
      "tokens_seen": 1375731712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003597976167959145,
      "loss": 2.8577,
      "theoretical_loss": 3.543202407680749,
      "tokens_seen": 1375797248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035978106676754304,
      "loss": 2.7803,
      "theoretical_loss": 3.543187279045699,
      "tokens_seen": 1375862784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003597645167391716,
      "loss": 2.7854,
      "theoretical_loss": 3.543172151333011,
      "tokens_seen": 1375928320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035974796671080006,
      "loss": 2.4325,
      "theoretical_loss": 3.543157024542585,
      "tokens_seen": 1375993856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003597314166824286,
      "loss": 2.571,
      "theoretical_loss": 3.54314189867432,
      "tokens_seen": 1376059392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035971486665405713,
      "loss": 2.522,
      "theoretical_loss": 3.5431267737281167,
      "tokens_seen": 1376124928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035969831662568567,
      "loss": 2.5229,
      "theoretical_loss": 3.5431116497038744,
      "tokens_seen": 1376190464
    },
    {
      "epoch": 0.49,
      "objective/train/avg_token_score": 0.023317284882068634,
      "objective/train/avg_weight": 0.9725581407546997,
      "objective/train/docs_used": 780651,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.002244234085083,
      "objective/train/std_weight": 0.1072935163974762,
      "objective/train/theoretical_loss": 3.5430965266014933,
      "objective/train/tokens_used": 1396716000,
      "theoretical_loss": 3.5430965266014933,
      "tokens_seen": 1376256000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035968176659731415,
      "loss": 2.781,
      "theoretical_loss": 3.5430965266014933,
      "tokens_seen": 1376256000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003596652165689427,
      "loss": 2.9667,
      "theoretical_loss": 3.5430814044208727,
      "tokens_seen": 1376321536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035964866654057123,
      "loss": 2.6716,
      "theoretical_loss": 3.5430662831619135,
      "tokens_seen": 1376387072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035963211651219977,
      "loss": 2.7844,
      "theoretical_loss": 3.543051162824515,
      "tokens_seen": 1376452608
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035961556648382825,
      "loss": 2.5085,
      "theoretical_loss": 3.5430360434085775,
      "tokens_seen": 1376518144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003595990164554568,
      "loss": 2.6647,
      "theoretical_loss": 3.543020924914001,
      "tokens_seen": 1376583680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003595824664270853,
      "loss": 2.6384,
      "theoretical_loss": 3.543005807340685,
      "tokens_seen": 1376649216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035956591639871386,
      "loss": 2.8362,
      "theoretical_loss": 3.542990690688529,
      "tokens_seen": 1376714752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035954936637034234,
      "loss": 2.5646,
      "theoretical_loss": 3.542975574957435,
      "tokens_seen": 1376780288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003595328163419709,
      "loss": 2.809,
      "theoretical_loss": 3.5429604601473015,
      "tokens_seen": 1376845824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003595162663135994,
      "loss": 2.7662,
      "theoretical_loss": 3.5429453462580294,
      "tokens_seen": 1376911360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035949971628522796,
      "loss": 2.6438,
      "theoretical_loss": 3.542930233289518,
      "tokens_seen": 1376976896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035948316625685644,
      "loss": 2.7944,
      "theoretical_loss": 3.5429151212416676,
      "tokens_seen": 1377042432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000359466616228485,
      "loss": 2.8681,
      "theoretical_loss": 3.542900010114378,
      "tokens_seen": 1377107968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003594500662001135,
      "loss": 2.739,
      "theoretical_loss": 3.5428848999075506,
      "tokens_seen": 1377173504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035943351617174205,
      "loss": 2.6997,
      "theoretical_loss": 3.542869790621084,
      "tokens_seen": 1377239040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035941696614337054,
      "loss": 2.6142,
      "theoretical_loss": 3.5428546822548794,
      "tokens_seen": 1377304576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003594004161149991,
      "loss": 3.0687,
      "theoretical_loss": 3.542839574808837,
      "tokens_seen": 1377370112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003593838660866276,
      "loss": 2.6853,
      "theoretical_loss": 3.5428244682828556,
      "tokens_seen": 1377435648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035936731605825615,
      "loss": 2.682,
      "theoretical_loss": 3.5428093626768367,
      "tokens_seen": 1377501184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035935076602988463,
      "loss": 2.5225,
      "theoretical_loss": 3.54279425799068,
      "tokens_seen": 1377566720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003593342160015131,
      "loss": 2.5475,
      "theoretical_loss": 3.542779154224286,
      "tokens_seen": 1377632256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035931766597314165,
      "loss": 2.8058,
      "theoretical_loss": 3.5427640513775556,
      "tokens_seen": 1377697792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035930111594477014,
      "loss": 2.5955,
      "theoretical_loss": 3.542748949450387,
      "tokens_seen": 1377763328
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003592845659163987,
      "loss": 2.6411,
      "theoretical_loss": 3.5427338484426825,
      "tokens_seen": 1377828864
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003592680158880272,
      "loss": 2.58,
      "theoretical_loss": 3.542718748354342,
      "tokens_seen": 1377894400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035925146585965575,
      "loss": 2.8851,
      "theoretical_loss": 3.542703649185265,
      "tokens_seen": 1377959936
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035923491583128423,
      "loss": 2.7435,
      "theoretical_loss": 3.542688550935352,
      "tokens_seen": 1378025472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035921836580291277,
      "loss": 2.5249,
      "theoretical_loss": 3.542673453604504,
      "tokens_seen": 1378091008
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003592018157745413,
      "loss": 2.8542,
      "theoretical_loss": 3.5426583571926207,
      "tokens_seen": 1378156544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035918526574616985,
      "loss": 2.7752,
      "theoretical_loss": 3.5426432616996033,
      "tokens_seen": 1378222080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035916871571779833,
      "loss": 2.748,
      "theoretical_loss": 3.5426281671253506,
      "tokens_seen": 1378287616
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035915216568942687,
      "loss": 2.5662,
      "theoretical_loss": 3.542613073469765,
      "tokens_seen": 1378353152
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003591356156610554,
      "loss": 2.9715,
      "theoretical_loss": 3.5425979807327455,
      "tokens_seen": 1378418688
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035911906563268394,
      "loss": 2.9641,
      "theoretical_loss": 3.5425828889141933,
      "tokens_seen": 1378484224
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003591025156043124,
      "loss": 2.7533,
      "theoretical_loss": 3.542567798014008,
      "tokens_seen": 1378549760
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035908596557594096,
      "loss": 2.6777,
      "theoretical_loss": 3.542552708032091,
      "tokens_seen": 1378615296
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003590694155475695,
      "loss": 2.7725,
      "theoretical_loss": 3.5425376189683417,
      "tokens_seen": 1378680832
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035905286551919804,
      "loss": 2.7623,
      "theoretical_loss": 3.542522530822662,
      "tokens_seen": 1378746368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003590363154908265,
      "loss": 2.814,
      "theoretical_loss": 3.5425074435949515,
      "tokens_seen": 1378811904
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035901976546245506,
      "loss": 2.5601,
      "theoretical_loss": 3.54249235728511,
      "tokens_seen": 1378877440
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003590032154340836,
      "loss": 2.6689,
      "theoretical_loss": 3.54247727189304,
      "tokens_seen": 1378942976
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035898666540571213,
      "loss": 2.4835,
      "theoretical_loss": 3.5424621874186406,
      "tokens_seen": 1379008512
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003589701153773406,
      "loss": 2.7622,
      "theoretical_loss": 3.5424471038618126,
      "tokens_seen": 1379074048
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035895356534896915,
      "loss": 2.4618,
      "theoretical_loss": 3.5424320212224565,
      "tokens_seen": 1379139584
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003589370153205977,
      "loss": 2.643,
      "theoretical_loss": 3.5424169395004736,
      "tokens_seen": 1379205120
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035892046529222623,
      "loss": 2.7591,
      "theoretical_loss": 3.542401858695764,
      "tokens_seen": 1379270656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003589039152638547,
      "loss": 2.4786,
      "theoretical_loss": 3.542386778808228,
      "tokens_seen": 1379336192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035888736523548325,
      "loss": 2.3205,
      "theoretical_loss": 3.5423716998377666,
      "tokens_seen": 1379401728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003588708152071118,
      "loss": 2.5974,
      "theoretical_loss": 3.542356621784281,
      "tokens_seen": 1379467264
    },
    {
      "epoch": 0.49,
      "objective/train/avg_token_score": 0.009273147210478783,
      "objective/train/avg_weight": 0.9890860915184021,
      "objective/train/docs_used": 781840,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7279105186462402,
      "objective/train/std_weight": 0.040280334651470184,
      "objective/train/theoretical_loss": 3.5423415446476705,
      "objective/train/tokens_used": 1399992800,
      "theoretical_loss": 3.5423415446476705,
      "tokens_seen": 1379532800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003588542651787403,
      "loss": 2.6086,
      "theoretical_loss": 3.5423415446476705,
      "tokens_seen": 1379532800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003588377151503688,
      "loss": 2.9955,
      "theoretical_loss": 3.542326468427837,
      "tokens_seen": 1379598336
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035882116512199735,
      "loss": 2.7797,
      "theoretical_loss": 3.542311393124681,
      "tokens_seen": 1379663872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003588046150936259,
      "loss": 2.4759,
      "theoretical_loss": 3.5422963187381034,
      "tokens_seen": 1379729408
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003587880650652544,
      "loss": 2.7519,
      "theoretical_loss": 3.542281245268004,
      "tokens_seen": 1379794944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003587715150368829,
      "loss": 2.6062,
      "theoretical_loss": 3.5422661727142843,
      "tokens_seen": 1379860480
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035875496500851144,
      "loss": 2.5368,
      "theoretical_loss": 3.542251101076845,
      "tokens_seen": 1379926016
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035873841498014,
      "loss": 2.5674,
      "theoretical_loss": 3.542236030355587,
      "tokens_seen": 1379991552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003587218649517685,
      "loss": 2.497,
      "theoretical_loss": 3.5422209605504107,
      "tokens_seen": 1380057088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000358705314923397,
      "loss": 2.8067,
      "theoretical_loss": 3.542205891661217,
      "tokens_seen": 1380122624
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035868876489502554,
      "loss": 2.7289,
      "theoretical_loss": 3.542190823687907,
      "tokens_seen": 1380188160
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003586722148666541,
      "loss": 2.5025,
      "theoretical_loss": 3.5421757566303818,
      "tokens_seen": 1380253696
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003586556648382826,
      "loss": 2.4692,
      "theoretical_loss": 3.542160690488541,
      "tokens_seen": 1380319232
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003586391148099111,
      "loss": 2.6327,
      "theoretical_loss": 3.542145625262287,
      "tokens_seen": 1380384768
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035862256478153963,
      "loss": 2.7476,
      "theoretical_loss": 3.54213056095152,
      "tokens_seen": 1380450304
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035860601475316817,
      "loss": 2.6463,
      "theoretical_loss": 3.542115497556141,
      "tokens_seen": 1380515840
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003585894647247967,
      "loss": 2.7044,
      "theoretical_loss": 3.5421004350760508,
      "tokens_seen": 1380581376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003585729146964252,
      "loss": 2.7753,
      "theoretical_loss": 3.54208537351115,
      "tokens_seen": 1380646912
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035855636466805373,
      "loss": 2.6787,
      "theoretical_loss": 3.54207031286134,
      "tokens_seen": 1380712448
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035853981463968227,
      "loss": 2.5074,
      "theoretical_loss": 3.542055253126523,
      "tokens_seen": 1380777984
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003585232646113108,
      "loss": 2.73,
      "theoretical_loss": 3.5420401943065976,
      "tokens_seen": 1380843520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003585067145829393,
      "loss": 2.4637,
      "theoretical_loss": 3.542025136401466,
      "tokens_seen": 1380909056
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003584901645545678,
      "loss": 2.7422,
      "theoretical_loss": 3.542010079411029,
      "tokens_seen": 1380974592
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035847361452619636,
      "loss": 2.6793,
      "theoretical_loss": 3.5419950233351885,
      "tokens_seen": 1381040128
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003584570644978249,
      "loss": 2.6027,
      "theoretical_loss": 3.5419799681738446,
      "tokens_seen": 1381105664
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003584405144694534,
      "loss": 2.5976,
      "theoretical_loss": 3.5419649139268987,
      "tokens_seen": 1381171200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003584239644410819,
      "loss": 2.6906,
      "theoretical_loss": 3.541949860594251,
      "tokens_seen": 1381236736
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035840741441271046,
      "loss": 2.7515,
      "theoretical_loss": 3.541934808175804,
      "tokens_seen": 1381302272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000358390864384339,
      "loss": 2.7025,
      "theoretical_loss": 3.541919756671458,
      "tokens_seen": 1381367808
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003583743143559675,
      "loss": 2.8126,
      "theoretical_loss": 3.541904706081114,
      "tokens_seen": 1381433344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000358357764327596,
      "loss": 2.7345,
      "theoretical_loss": 3.541889656404674,
      "tokens_seen": 1381498880
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035834121429922445,
      "loss": 2.5626,
      "theoretical_loss": 3.541874607642039,
      "tokens_seen": 1381564416
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000358324664270853,
      "loss": 2.5244,
      "theoretical_loss": 3.541859559793109,
      "tokens_seen": 1381629952
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003583081142424815,
      "loss": 2.4358,
      "theoretical_loss": 3.5418445128577862,
      "tokens_seen": 1381695488
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035829156421411006,
      "loss": 2.6444,
      "theoretical_loss": 3.5418294668359716,
      "tokens_seen": 1381761024
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035827501418573854,
      "loss": 2.6332,
      "theoretical_loss": 3.541814421727566,
      "tokens_seen": 1381826560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003582584641573671,
      "loss": 2.6208,
      "theoretical_loss": 3.5417993775324716,
      "tokens_seen": 1381892096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003582419141289956,
      "loss": 2.4403,
      "theoretical_loss": 3.541784334250589,
      "tokens_seen": 1381957632
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035822536410062416,
      "loss": 2.5445,
      "theoretical_loss": 3.5417692918818187,
      "tokens_seen": 1382023168
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035820881407225264,
      "loss": 2.481,
      "theoretical_loss": 3.541754250426064,
      "tokens_seen": 1382088704
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003581922640438812,
      "loss": 2.6904,
      "theoretical_loss": 3.5417392098832243,
      "tokens_seen": 1382154240
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003581757140155097,
      "loss": 2.7319,
      "theoretical_loss": 3.541724170253201,
      "tokens_seen": 1382219776
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035815916398713825,
      "loss": 2.5054,
      "theoretical_loss": 3.5417091315358973,
      "tokens_seen": 1382285312
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035814261395876673,
      "loss": 2.7532,
      "theoretical_loss": 3.5416940937312127,
      "tokens_seen": 1382350848
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035812606393039527,
      "loss": 2.6143,
      "theoretical_loss": 3.5416790568390493,
      "tokens_seen": 1382416384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003581095139020238,
      "loss": 2.6698,
      "theoretical_loss": 3.541664020859308,
      "tokens_seen": 1382481920
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035809296387365235,
      "loss": 2.5896,
      "theoretical_loss": 3.5416489857918902,
      "tokens_seen": 1382547456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035807641384528083,
      "loss": 2.5633,
      "theoretical_loss": 3.5416339516366984,
      "tokens_seen": 1382612992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035805986381690937,
      "loss": 2.7843,
      "theoretical_loss": 3.541618918393633,
      "tokens_seen": 1382678528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003580433137885379,
      "loss": 2.5118,
      "theoretical_loss": 3.541603886062595,
      "tokens_seen": 1382744064
    },
    {
      "epoch": 0.49,
      "objective/train/avg_token_score": 0.03043193370103836,
      "objective/train/avg_weight": 0.9641802906990051,
      "objective/train/docs_used": 784187,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.76712703704834,
      "objective/train/std_weight": 0.12112551927566528,
      "objective/train/theoretical_loss": 3.541588854643487,
      "objective/train/tokens_used": 1403269600,
      "theoretical_loss": 3.541588854643487,
      "tokens_seen": 1382809600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035802676376016644,
      "loss": 2.4172,
      "theoretical_loss": 3.541588854643487,
      "tokens_seen": 1382809600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003580102137317949,
      "loss": 2.6154,
      "theoretical_loss": 3.54157382413621,
      "tokens_seen": 1382875136
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035799366370342346,
      "loss": 2.3885,
      "theoretical_loss": 3.541558794540665,
      "tokens_seen": 1382940672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000357977113675052,
      "loss": 2.601,
      "theoretical_loss": 3.5415437658567543,
      "tokens_seen": 1383006208
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035796056364668054,
      "loss": 2.7114,
      "theoretical_loss": 3.541528738084379,
      "tokens_seen": 1383071744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003579440136183091,
      "loss": 2.5767,
      "theoretical_loss": 3.5415137112234403,
      "tokens_seen": 1383137280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035792746358993756,
      "loss": 2.7387,
      "theoretical_loss": 3.54149868527384,
      "tokens_seen": 1383202816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003579109135615661,
      "loss": 2.6715,
      "theoretical_loss": 3.541483660235481,
      "tokens_seen": 1383268352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035789436353319464,
      "loss": 2.7658,
      "theoretical_loss": 3.541468636108262,
      "tokens_seen": 1383333888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035787781350482317,
      "loss": 2.7941,
      "theoretical_loss": 3.541453612892087,
      "tokens_seen": 1383399424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035786126347645166,
      "loss": 2.5699,
      "theoretical_loss": 3.541438590586857,
      "tokens_seen": 1383464960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003578447134480802,
      "loss": 2.5858,
      "theoretical_loss": 3.5414235691924727,
      "tokens_seen": 1383530496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035782816341970873,
      "loss": 2.775,
      "theoretical_loss": 3.541408548708837,
      "tokens_seen": 1383596032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035781161339133727,
      "loss": 2.7181,
      "theoretical_loss": 3.541393529135851,
      "tokens_seen": 1383661568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035779506336296575,
      "loss": 2.7479,
      "theoretical_loss": 3.5413785104734163,
      "tokens_seen": 1383727104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003577785133345943,
      "loss": 2.4956,
      "theoretical_loss": 3.5413634927214344,
      "tokens_seen": 1383792640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035776196330622283,
      "loss": 2.6993,
      "theoretical_loss": 3.5413484758798073,
      "tokens_seen": 1383858176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035774541327785136,
      "loss": 2.8421,
      "theoretical_loss": 3.5413334599484365,
      "tokens_seen": 1383923712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035772886324947985,
      "loss": 2.7783,
      "theoretical_loss": 3.5413184449272244,
      "tokens_seen": 1383989248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003577123132211084,
      "loss": 2.6111,
      "theoretical_loss": 3.541303430816072,
      "tokens_seen": 1384054784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003576957631927369,
      "loss": 2.7551,
      "theoretical_loss": 3.5412884176148807,
      "tokens_seen": 1384120320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035767921316436546,
      "loss": 2.7006,
      "theoretical_loss": 3.541273405323553,
      "tokens_seen": 1384185856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035766266313599394,
      "loss": 2.9574,
      "theoretical_loss": 3.5412583939419906,
      "tokens_seen": 1384251392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003576461131076225,
      "loss": 2.6071,
      "theoretical_loss": 3.5412433834700954,
      "tokens_seen": 1384316928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000357629563079251,
      "loss": 2.4505,
      "theoretical_loss": 3.541228373907769,
      "tokens_seen": 1384382464
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035761301305087956,
      "loss": 2.4001,
      "theoretical_loss": 3.541213365254913,
      "tokens_seen": 1384448000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035759646302250804,
      "loss": 2.6797,
      "theoretical_loss": 3.5411983575114294,
      "tokens_seen": 1384513536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003575799129941366,
      "loss": 2.4887,
      "theoretical_loss": 3.5411833506772203,
      "tokens_seen": 1384579072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003575633629657651,
      "loss": 2.6903,
      "theoretical_loss": 3.5411683447521876,
      "tokens_seen": 1384644608
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035754681293739365,
      "loss": 2.686,
      "theoretical_loss": 3.5411533397362325,
      "tokens_seen": 1384710144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035753026290902214,
      "loss": 2.7667,
      "theoretical_loss": 3.541138335629258,
      "tokens_seen": 1384775680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003575137128806507,
      "loss": 2.5112,
      "theoretical_loss": 3.541123332431165,
      "tokens_seen": 1384841216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003574971628522792,
      "loss": 2.4301,
      "theoretical_loss": 3.541108330141856,
      "tokens_seen": 1384906752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035748061282390775,
      "loss": 2.6206,
      "theoretical_loss": 3.5410933287612325,
      "tokens_seen": 1384972288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035746406279553623,
      "loss": 2.5368,
      "theoretical_loss": 3.5410783282891973,
      "tokens_seen": 1385037824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035744751276716477,
      "loss": 2.7505,
      "theoretical_loss": 3.541063328725652,
      "tokens_seen": 1385103360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003574309627387933,
      "loss": 2.9041,
      "theoretical_loss": 3.541048330070498,
      "tokens_seen": 1385168896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035741441271042184,
      "loss": 2.8184,
      "theoretical_loss": 3.541033332323638,
      "tokens_seen": 1385234432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035739786268205033,
      "loss": 2.7466,
      "theoretical_loss": 3.541018335484974,
      "tokens_seen": 1385299968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035738131265367887,
      "loss": 2.7014,
      "theoretical_loss": 3.5410033395544076,
      "tokens_seen": 1385365504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003573647626253073,
      "loss": 2.7417,
      "theoretical_loss": 3.540988344531841,
      "tokens_seen": 1385431040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035734821259693583,
      "loss": 2.8685,
      "theoretical_loss": 3.5409733504171763,
      "tokens_seen": 1385496576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035733166256856437,
      "loss": 2.3201,
      "theoretical_loss": 3.540958357210316,
      "tokens_seen": 1385562112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003573151125401929,
      "loss": 2.8323,
      "theoretical_loss": 3.5409433649111617,
      "tokens_seen": 1385627648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003572985625118214,
      "loss": 2.8408,
      "theoretical_loss": 3.5409283735196153,
      "tokens_seen": 1385693184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035728201248344993,
      "loss": 2.6312,
      "theoretical_loss": 3.54091338303558,
      "tokens_seen": 1385758720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00035726546245507847,
      "loss": 2.6861,
      "theoretical_loss": 3.5408983934589573,
      "tokens_seen": 1385824256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000357248912426707,
      "loss": 2.5381,
      "theoretical_loss": 3.540883404789649,
      "tokens_seen": 1385889792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0003572323623983355,
      "loss": 2.6745,
      "theoretical_loss": 3.5408684170275575,
      "tokens_seen": 1385955328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000357215812369964,
      "loss": 2.7805,
      "theoretical_loss": 3.5408534301725854,
      "tokens_seen": 1386020864
    },
    {
      "epoch": 0.5,
      "objective/train/avg_token_score": 0.020880233496427536,
      "objective/train/avg_weight": 0.9754285216331482,
      "objective/train/docs_used": 786072,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.592090129852295,
      "objective/train/std_weight": 0.09247780591249466,
      "objective/train/theoretical_loss": 3.5408384442246343,
      "objective/train/tokens_used": 1406546400,
      "theoretical_loss": 3.5408384442246343,
      "tokens_seen": 1386086400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035719926234159256,
      "loss": 2.6392,
      "theoretical_loss": 3.5408384442246343,
      "tokens_seen": 1386086400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003571827123132211,
      "loss": 2.6547,
      "theoretical_loss": 3.540823459183607,
      "tokens_seen": 1386151936
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003571661622848496,
      "loss": 2.5135,
      "theoretical_loss": 3.5408084750494053,
      "tokens_seen": 1386217472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003571496122564781,
      "loss": 2.5112,
      "theoretical_loss": 3.540793491821932,
      "tokens_seen": 1386283008
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035713306222810666,
      "loss": 2.612,
      "theoretical_loss": 3.540778509501089,
      "tokens_seen": 1386348544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003571165121997352,
      "loss": 2.7698,
      "theoretical_loss": 3.540763528086778,
      "tokens_seen": 1386414080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003570999621713637,
      "loss": 2.5245,
      "theoretical_loss": 3.540748547578902,
      "tokens_seen": 1386479616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003570834121429922,
      "loss": 2.1143,
      "theoretical_loss": 3.540733567977364,
      "tokens_seen": 1386545152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035706686211462075,
      "loss": 2.5055,
      "theoretical_loss": 3.540718589282065,
      "tokens_seen": 1386610688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003570503120862493,
      "loss": 2.6197,
      "theoretical_loss": 3.5407036114929076,
      "tokens_seen": 1386676224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003570337620578778,
      "loss": 2.4764,
      "theoretical_loss": 3.540688634609795,
      "tokens_seen": 1386741760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003570172120295063,
      "loss": 2.6407,
      "theoretical_loss": 3.5406736586326284,
      "tokens_seen": 1386807296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035700066200113485,
      "loss": 2.8468,
      "theoretical_loss": 3.540658683561311,
      "tokens_seen": 1386872832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003569841119727634,
      "loss": 2.6008,
      "theoretical_loss": 3.5406437093957455,
      "tokens_seen": 1386938368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035696756194439187,
      "loss": 2.7202,
      "theoretical_loss": 3.540628736135834,
      "tokens_seen": 1387003904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003569510119160204,
      "loss": 2.559,
      "theoretical_loss": 3.540613763781478,
      "tokens_seen": 1387069440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035693446188764895,
      "loss": 2.5895,
      "theoretical_loss": 3.540598792332581,
      "tokens_seen": 1387134976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003569179118592775,
      "loss": 2.5901,
      "theoretical_loss": 3.5405838217890455,
      "tokens_seen": 1387200512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035690136183090597,
      "loss": 2.5671,
      "theoretical_loss": 3.5405688521507734,
      "tokens_seen": 1387266048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003568848118025345,
      "loss": 2.5679,
      "theoretical_loss": 3.5405538834176675,
      "tokens_seen": 1387331584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035686826177416304,
      "loss": 2.7598,
      "theoretical_loss": 3.5405389155896305,
      "tokens_seen": 1387397120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003568517117457916,
      "loss": 2.7343,
      "theoretical_loss": 3.5405239486665643,
      "tokens_seen": 1387462656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035683516171742006,
      "loss": 2.5914,
      "theoretical_loss": 3.540508982648372,
      "tokens_seen": 1387528192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003568186116890486,
      "loss": 2.6774,
      "theoretical_loss": 3.5404940175349564,
      "tokens_seen": 1387593728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035680206166067714,
      "loss": 2.7044,
      "theoretical_loss": 3.5404790533262194,
      "tokens_seen": 1387659264
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003567855116323057,
      "loss": 2.5338,
      "theoretical_loss": 3.5404640900220636,
      "tokens_seen": 1387724800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035676896160393416,
      "loss": 2.6355,
      "theoretical_loss": 3.5404491276223924,
      "tokens_seen": 1387790336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003567524115755627,
      "loss": 2.5135,
      "theoretical_loss": 3.540434166127108,
      "tokens_seen": 1387855872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035673586154719123,
      "loss": 2.672,
      "theoretical_loss": 3.5404192055361126,
      "tokens_seen": 1387921408
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035671931151881977,
      "loss": 2.8354,
      "theoretical_loss": 3.5404042458493086,
      "tokens_seen": 1387986944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035670276149044825,
      "loss": 2.6802,
      "theoretical_loss": 3.5403892870666,
      "tokens_seen": 1388052480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003566862114620768,
      "loss": 2.6436,
      "theoretical_loss": 3.5403743291878884,
      "tokens_seen": 1388118016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035666966143370533,
      "loss": 2.5969,
      "theoretical_loss": 3.5403593722130773,
      "tokens_seen": 1388183552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035665311140533387,
      "loss": 2.5595,
      "theoretical_loss": 3.5403444161420685,
      "tokens_seen": 1388249088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035663656137696235,
      "loss": 2.787,
      "theoretical_loss": 3.5403294609747653,
      "tokens_seen": 1388314624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003566200113485909,
      "loss": 2.6995,
      "theoretical_loss": 3.54031450671107,
      "tokens_seen": 1388380160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003566034613202194,
      "loss": 2.6306,
      "theoretical_loss": 3.540299553350886,
      "tokens_seen": 1388445696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035658691129184796,
      "loss": 2.3657,
      "theoretical_loss": 3.5402846008941156,
      "tokens_seen": 1388511232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035657036126347645,
      "loss": 2.4759,
      "theoretical_loss": 3.5402696493406616,
      "tokens_seen": 1388576768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000356553811235105,
      "loss": 2.5788,
      "theoretical_loss": 3.5402546986904273,
      "tokens_seen": 1388642304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003565372612067335,
      "loss": 2.7347,
      "theoretical_loss": 3.5402397489433146,
      "tokens_seen": 1388707840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035652071117836206,
      "loss": 2.742,
      "theoretical_loss": 3.540224800099227,
      "tokens_seen": 1388773376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035650416114999054,
      "loss": 2.6704,
      "theoretical_loss": 3.5402098521580676,
      "tokens_seen": 1388838912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003564876111216191,
      "loss": 2.7118,
      "theoretical_loss": 3.540194905119738,
      "tokens_seen": 1388904448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003564710610932476,
      "loss": 2.6557,
      "theoretical_loss": 3.5401799589841425,
      "tokens_seen": 1388969984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035645451106487615,
      "loss": 2.5991,
      "theoretical_loss": 3.5401650137511833,
      "tokens_seen": 1389035520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003564379610365047,
      "loss": 2.8375,
      "theoretical_loss": 3.5401500694207635,
      "tokens_seen": 1389101056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003564214110081332,
      "loss": 2.6462,
      "theoretical_loss": 3.5401351259927853,
      "tokens_seen": 1389166592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003564048609797617,
      "loss": 2.5666,
      "theoretical_loss": 3.540120183467153,
      "tokens_seen": 1389232128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035638831095139014,
      "loss": 2.7343,
      "theoretical_loss": 3.5401052418437686,
      "tokens_seen": 1389297664
    },
    {
      "epoch": 0.5,
      "objective/train/avg_token_score": 0.01979769766330719,
      "objective/train/avg_weight": 0.976690411567688,
      "objective/train/docs_used": 787611,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0776898860931396,
      "objective/train/std_weight": 0.0931376963853836,
      "objective/train/theoretical_loss": 3.540090301122535,
      "objective/train/tokens_used": 1409823200,
      "theoretical_loss": 3.540090301122535,
      "tokens_seen": 1389363200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003563717609230187,
      "loss": 2.8402,
      "theoretical_loss": 3.540090301122535,
      "tokens_seen": 1389363200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003563552108946472,
      "loss": 2.7972,
      "theoretical_loss": 3.540075361303356,
      "tokens_seen": 1389428736
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035633866086627575,
      "loss": 2.7111,
      "theoretical_loss": 3.5400604223861336,
      "tokens_seen": 1389494272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035632211083790424,
      "loss": 2.5539,
      "theoretical_loss": 3.540045484370771,
      "tokens_seen": 1389559808
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003563055608095328,
      "loss": 2.754,
      "theoretical_loss": 3.5400305472571723,
      "tokens_seen": 1389625344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003562890107811613,
      "loss": 2.7244,
      "theoretical_loss": 3.540015611045239,
      "tokens_seen": 1389690880
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035627246075278985,
      "loss": 2.5131,
      "theoretical_loss": 3.540000675734875,
      "tokens_seen": 1389756416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035625591072441833,
      "loss": 2.4799,
      "theoretical_loss": 3.5399857413259834,
      "tokens_seen": 1389821952
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035623936069604687,
      "loss": 2.8104,
      "theoretical_loss": 3.5399708078184675,
      "tokens_seen": 1389887488
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003562228106676754,
      "loss": 2.6009,
      "theoretical_loss": 3.5399558752122293,
      "tokens_seen": 1389953024
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035620626063930395,
      "loss": 2.5126,
      "theoretical_loss": 3.5399409435071734,
      "tokens_seen": 1390018560
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035618971061093243,
      "loss": 2.5358,
      "theoretical_loss": 3.5399260127032015,
      "tokens_seen": 1390084096
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035617316058256097,
      "loss": 2.6488,
      "theoretical_loss": 3.5399110828002174,
      "tokens_seen": 1390149632
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003561566105541895,
      "loss": 2.7831,
      "theoretical_loss": 3.5398961537981246,
      "tokens_seen": 1390215168
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035614006052581804,
      "loss": 2.7914,
      "theoretical_loss": 3.539881225696826,
      "tokens_seen": 1390280704
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003561235104974465,
      "loss": 2.6053,
      "theoretical_loss": 3.539866298496224,
      "tokens_seen": 1390346240
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035610696046907506,
      "loss": 2.3864,
      "theoretical_loss": 3.5398513721962237,
      "tokens_seen": 1390411776
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003560904104407036,
      "loss": 2.4915,
      "theoretical_loss": 3.5398364467967265,
      "tokens_seen": 1390477312
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035607386041233214,
      "loss": 2.8795,
      "theoretical_loss": 3.539821522297636,
      "tokens_seen": 1390542848
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003560573103839606,
      "loss": 2.7057,
      "theoretical_loss": 3.5398065986988563,
      "tokens_seen": 1390608384
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035604076035558916,
      "loss": 2.932,
      "theoretical_loss": 3.53979167600029,
      "tokens_seen": 1390673920
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003560242103272177,
      "loss": 2.6714,
      "theoretical_loss": 3.5397767542018403,
      "tokens_seen": 1390739456
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035600766029884623,
      "loss": 2.8323,
      "theoretical_loss": 3.5397618333034107,
      "tokens_seen": 1390804992
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003559911102704747,
      "loss": 2.3928,
      "theoretical_loss": 3.5397469133049047,
      "tokens_seen": 1390870528
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035597456024210326,
      "loss": 2.5092,
      "theoretical_loss": 3.5397319942062255,
      "tokens_seen": 1390936064
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003559580102137318,
      "loss": 2.7873,
      "theoretical_loss": 3.539717076007276,
      "tokens_seen": 1391001600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035594146018536033,
      "loss": 2.696,
      "theoretical_loss": 3.53970215870796,
      "tokens_seen": 1391067136
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003559249101569888,
      "loss": 2.5119,
      "theoretical_loss": 3.539687242308181,
      "tokens_seen": 1391132672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035590836012861735,
      "loss": 2.6405,
      "theoretical_loss": 3.5396723268078416,
      "tokens_seen": 1391198208
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003558918101002459,
      "loss": 2.6299,
      "theoretical_loss": 3.539657412206846,
      "tokens_seen": 1391263744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003558752600718744,
      "loss": 2.8246,
      "theoretical_loss": 3.5396424985050974,
      "tokens_seen": 1391329280
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003558587100435029,
      "loss": 2.6795,
      "theoretical_loss": 3.5396275857024992,
      "tokens_seen": 1391394816
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035584216001513145,
      "loss": 2.5649,
      "theoretical_loss": 3.5396126737989553,
      "tokens_seen": 1391460352
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035582560998676,
      "loss": 2.4999,
      "theoretical_loss": 3.5395977627943678,
      "tokens_seen": 1391525888
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003558090599583885,
      "loss": 2.4002,
      "theoretical_loss": 3.5395828526886417,
      "tokens_seen": 1391591424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000355792509930017,
      "loss": 2.6396,
      "theoretical_loss": 3.5395679434816794,
      "tokens_seen": 1391656960
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035577595990164554,
      "loss": 2.5996,
      "theoretical_loss": 3.5395530351733853,
      "tokens_seen": 1391722496
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003557594098732741,
      "loss": 2.6886,
      "theoretical_loss": 3.539538127763662,
      "tokens_seen": 1391788032
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003557428598449026,
      "loss": 2.6073,
      "theoretical_loss": 3.539523221252414,
      "tokens_seen": 1391853568
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003557263098165311,
      "loss": 2.7698,
      "theoretical_loss": 3.5395083156395444,
      "tokens_seen": 1391919104
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035570975978815964,
      "loss": 2.667,
      "theoretical_loss": 3.5394934109249565,
      "tokens_seen": 1391984640
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003556932097597882,
      "loss": 2.5787,
      "theoretical_loss": 3.5394785071085537,
      "tokens_seen": 1392050176
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003556766597314167,
      "loss": 2.6668,
      "theoretical_loss": 3.5394636041902405,
      "tokens_seen": 1392115712
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003556601097030452,
      "loss": 2.5629,
      "theoretical_loss": 3.5394487021699197,
      "tokens_seen": 1392181248
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035564355967467374,
      "loss": 2.6509,
      "theoretical_loss": 3.5394338010474957,
      "tokens_seen": 1392246784
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003556270096463023,
      "loss": 2.5201,
      "theoretical_loss": 3.539418900822871,
      "tokens_seen": 1392312320
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003556104596179308,
      "loss": 2.6251,
      "theoretical_loss": 3.5394040014959502,
      "tokens_seen": 1392377856
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003555939095895593,
      "loss": 2.5756,
      "theoretical_loss": 3.539389103066637,
      "tokens_seen": 1392443392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035557735956118783,
      "loss": 2.6367,
      "theoretical_loss": 3.539374205534834,
      "tokens_seen": 1392508928
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035556080953281637,
      "loss": 2.7284,
      "theoretical_loss": 3.539359308900446,
      "tokens_seen": 1392574464
    },
    {
      "epoch": 0.5,
      "objective/train/avg_token_score": 0.008933682925999165,
      "objective/train/avg_weight": 0.9895046353340149,
      "objective/train/docs_used": 789336,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.807175636291504,
      "objective/train/std_weight": 0.06561707705259323,
      "objective/train/theoretical_loss": 3.5393444131633762,
      "objective/train/tokens_used": 1413100000,
      "theoretical_loss": 3.5393444131633762,
      "tokens_seen": 1392640000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003555442595044449,
      "loss": 2.7654,
      "theoretical_loss": 3.5393444131633762,
      "tokens_seen": 1392640000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003555277094760734,
      "loss": 2.7587,
      "theoretical_loss": 3.5393295183235294,
      "tokens_seen": 1392705536
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035551115944770193,
      "loss": 2.7887,
      "theoretical_loss": 3.5393146243808076,
      "tokens_seen": 1392771072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035549460941933046,
      "loss": 2.4702,
      "theoretical_loss": 3.539299731335115,
      "tokens_seen": 1392836608
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000355478059390959,
      "loss": 2.7035,
      "theoretical_loss": 3.5392848391863563,
      "tokens_seen": 1392902144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003554615093625875,
      "loss": 2.6416,
      "theoretical_loss": 3.539269947934435,
      "tokens_seen": 1392967680
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000355444959334216,
      "loss": 2.9163,
      "theoretical_loss": 3.539255057579254,
      "tokens_seen": 1393033216
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035542840930584456,
      "loss": 2.9817,
      "theoretical_loss": 3.539240168120718,
      "tokens_seen": 1393098752
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003554118592774731,
      "loss": 2.7281,
      "theoretical_loss": 3.539225279558731,
      "tokens_seen": 1393164288
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035539530924910153,
      "loss": 2.8539,
      "theoretical_loss": 3.539210391893196,
      "tokens_seen": 1393229824
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035537875922073007,
      "loss": 2.6429,
      "theoretical_loss": 3.539195505124017,
      "tokens_seen": 1393295360
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003553622091923586,
      "loss": 2.8031,
      "theoretical_loss": 3.5391806192510984,
      "tokens_seen": 1393360896
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003553456591639871,
      "loss": 2.6676,
      "theoretical_loss": 3.539165734274344,
      "tokens_seen": 1393426432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003553291091356156,
      "loss": 2.6016,
      "theoretical_loss": 3.539150850193658,
      "tokens_seen": 1393491968
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035531255910724416,
      "loss": 2.7559,
      "theoretical_loss": 3.539135967008943,
      "tokens_seen": 1393557504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003552960090788727,
      "loss": 2.4244,
      "theoretical_loss": 3.5391210847201044,
      "tokens_seen": 1393623040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003552794590505012,
      "loss": 2.7331,
      "theoretical_loss": 3.5391062033270453,
      "tokens_seen": 1393688576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003552629090221297,
      "loss": 2.5163,
      "theoretical_loss": 3.5390913228296697,
      "tokens_seen": 1393754112
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035524635899375826,
      "loss": 2.5724,
      "theoretical_loss": 3.539076443227882,
      "tokens_seen": 1393819648
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003552298089653868,
      "loss": 2.4891,
      "theoretical_loss": 3.539061564521586,
      "tokens_seen": 1393885184
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003552132589370153,
      "loss": 2.528,
      "theoretical_loss": 3.5390466867106856,
      "tokens_seen": 1393950720
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003551967089086438,
      "loss": 2.7829,
      "theoretical_loss": 3.539031809795085,
      "tokens_seen": 1394016256
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035518015888027235,
      "loss": 2.7533,
      "theoretical_loss": 3.5390169337746884,
      "tokens_seen": 1394081792
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003551636088519009,
      "loss": 2.8647,
      "theoretical_loss": 3.539002058649399,
      "tokens_seen": 1394147328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003551470588235294,
      "loss": 2.656,
      "theoretical_loss": 3.5389871844191223,
      "tokens_seen": 1394212864
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003551305087951579,
      "loss": 2.5585,
      "theoretical_loss": 3.5389723110837608,
      "tokens_seen": 1394278400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035511395876678645,
      "loss": 2.585,
      "theoretical_loss": 3.53895743864322,
      "tokens_seen": 1394343936
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000355097408738415,
      "loss": 2.7346,
      "theoretical_loss": 3.538942567097403,
      "tokens_seen": 1394409472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035508085871004347,
      "loss": 2.68,
      "theoretical_loss": 3.538927696446214,
      "tokens_seen": 1394475008
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000355064308681672,
      "loss": 3.142,
      "theoretical_loss": 3.538912826689558,
      "tokens_seen": 1394540544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035504775865330054,
      "loss": 2.8176,
      "theoretical_loss": 3.5388979578273383,
      "tokens_seen": 1394606080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003550312086249291,
      "loss": 2.7891,
      "theoretical_loss": 3.538883089859459,
      "tokens_seen": 1394671616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035501465859655757,
      "loss": 2.6281,
      "theoretical_loss": 3.5388682227858252,
      "tokens_seen": 1394737152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003549981085681861,
      "loss": 2.6694,
      "theoretical_loss": 3.5388533566063405,
      "tokens_seen": 1394802688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035498155853981464,
      "loss": 2.6253,
      "theoretical_loss": 3.538838491320909,
      "tokens_seen": 1394868224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003549650085114432,
      "loss": 2.8324,
      "theoretical_loss": 3.5388236269294353,
      "tokens_seen": 1394933760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035494845848307166,
      "loss": 2.8149,
      "theoretical_loss": 3.538808763431823,
      "tokens_seen": 1394999296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003549319084547002,
      "loss": 2.7872,
      "theoretical_loss": 3.5387939008279767,
      "tokens_seen": 1395064832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035491535842632874,
      "loss": 2.3083,
      "theoretical_loss": 3.5387790391178013,
      "tokens_seen": 1395130368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003548988083979573,
      "loss": 2.5659,
      "theoretical_loss": 3.5387641783012,
      "tokens_seen": 1395195904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035488225836958576,
      "loss": 2.8122,
      "theoretical_loss": 3.5387493183780783,
      "tokens_seen": 1395261440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003548657083412143,
      "loss": 2.4716,
      "theoretical_loss": 3.5387344593483396,
      "tokens_seen": 1395326976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035484915831284283,
      "loss": 2.5691,
      "theoretical_loss": 3.5387196012118887,
      "tokens_seen": 1395392512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035483260828447137,
      "loss": 2.8145,
      "theoretical_loss": 3.538704743968629,
      "tokens_seen": 1395458048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035481605825609985,
      "loss": 2.713,
      "theoretical_loss": 3.5386898876184665,
      "tokens_seen": 1395523584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003547995082277284,
      "loss": 2.6417,
      "theoretical_loss": 3.5386750321613043,
      "tokens_seen": 1395589120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035478295819935693,
      "loss": 2.6196,
      "theoretical_loss": 3.538660177597047,
      "tokens_seen": 1395654656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035476640817098547,
      "loss": 2.7679,
      "theoretical_loss": 3.5386453239255995,
      "tokens_seen": 1395720192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035474985814261395,
      "loss": 2.6256,
      "theoretical_loss": 3.538630471146866,
      "tokens_seen": 1395785728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003547333081142425,
      "loss": 2.7181,
      "theoretical_loss": 3.5386156192607503,
      "tokens_seen": 1395851264
    },
    {
      "epoch": 0.5,
      "objective/train/avg_token_score": 0.033521734178066254,
      "objective/train/avg_weight": 0.9605625867843628,
      "objective/train/docs_used": 791369,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1706080436706543,
      "objective/train/std_weight": 0.11861860007047653,
      "objective/train/theoretical_loss": 3.5386007682671576,
      "objective/train/tokens_used": 1416376800,
      "theoretical_loss": 3.5386007682671576,
      "tokens_seen": 1395916800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000354716758085871,
      "loss": 2.9845,
      "theoretical_loss": 3.5386007682671576,
      "tokens_seen": 1395916800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035470020805749956,
      "loss": 2.8706,
      "theoretical_loss": 3.5385859181659924,
      "tokens_seen": 1395982336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035468365802912805,
      "loss": 2.63,
      "theoretical_loss": 3.538571068957159,
      "tokens_seen": 1396047872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003546671080007566,
      "loss": 2.5218,
      "theoretical_loss": 3.5385562206405616,
      "tokens_seen": 1396113408
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003546505579723851,
      "loss": 2.6513,
      "theoretical_loss": 3.5385413732161046,
      "tokens_seen": 1396178944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035463400794401366,
      "loss": 2.6101,
      "theoretical_loss": 3.5385265266836936,
      "tokens_seen": 1396244480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035461745791564214,
      "loss": 2.7099,
      "theoretical_loss": 3.538511681043232,
      "tokens_seen": 1396310016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003546009078872707,
      "loss": 2.6113,
      "theoretical_loss": 3.538496836294625,
      "tokens_seen": 1396375552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003545843578588992,
      "loss": 2.8722,
      "theoretical_loss": 3.5384819924377764,
      "tokens_seen": 1396441088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035456780783052775,
      "loss": 2.6229,
      "theoretical_loss": 3.538467149472592,
      "tokens_seen": 1396506624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035455125780215624,
      "loss": 2.4567,
      "theoretical_loss": 3.538452307398975,
      "tokens_seen": 1396572160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003545347077737848,
      "loss": 2.7188,
      "theoretical_loss": 3.5384374662168314,
      "tokens_seen": 1396637696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003545181577454133,
      "loss": 2.851,
      "theoretical_loss": 3.5384226259260645,
      "tokens_seen": 1396703232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035450160771704185,
      "loss": 2.7783,
      "theoretical_loss": 3.53840778652658,
      "tokens_seen": 1396768768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035448505768867033,
      "loss": 2.4811,
      "theoretical_loss": 3.5383929480182825,
      "tokens_seen": 1396834304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035446850766029887,
      "loss": 2.5444,
      "theoretical_loss": 3.5383781104010756,
      "tokens_seen": 1396899840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003544519576319274,
      "loss": 2.7097,
      "theoretical_loss": 3.5383632736748654,
      "tokens_seen": 1396965376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035443540760355595,
      "loss": 2.7352,
      "theoretical_loss": 3.5383484378395553,
      "tokens_seen": 1397030912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003544188575751844,
      "loss": 2.7002,
      "theoretical_loss": 3.538333602895051,
      "tokens_seen": 1397096448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003544023075468129,
      "loss": 2.6113,
      "theoretical_loss": 3.5383187688412567,
      "tokens_seen": 1397161984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003543857575184414,
      "loss": 2.642,
      "theoretical_loss": 3.538303935678077,
      "tokens_seen": 1397227520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035436920749006993,
      "loss": 2.8093,
      "theoretical_loss": 3.5382891034054174,
      "tokens_seen": 1397293056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035435265746169847,
      "loss": 2.4013,
      "theoretical_loss": 3.538274272023182,
      "tokens_seen": 1397358592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000354336107433327,
      "loss": 2.6557,
      "theoretical_loss": 3.538259441531276,
      "tokens_seen": 1397424128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003543195574049555,
      "loss": 2.6915,
      "theoretical_loss": 3.5382446119296036,
      "tokens_seen": 1397489664
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035430300737658403,
      "loss": 2.7453,
      "theoretical_loss": 3.5382297832180702,
      "tokens_seen": 1397555200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035428645734821257,
      "loss": 2.8061,
      "theoretical_loss": 3.5382149553965805,
      "tokens_seen": 1397620736
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003542699073198411,
      "loss": 2.9387,
      "theoretical_loss": 3.5382001284650397,
      "tokens_seen": 1397686272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003542533572914696,
      "loss": 2.8155,
      "theoretical_loss": 3.538185302423352,
      "tokens_seen": 1397751808
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003542368072630981,
      "loss": 2.8129,
      "theoretical_loss": 3.5381704772714224,
      "tokens_seen": 1397817344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035422025723472666,
      "loss": 2.577,
      "theoretical_loss": 3.5381556530091562,
      "tokens_seen": 1397882880
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003542037072063552,
      "loss": 2.736,
      "theoretical_loss": 3.5381408296364576,
      "tokens_seen": 1397948416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003541871571779837,
      "loss": 2.4826,
      "theoretical_loss": 3.5381260071532328,
      "tokens_seen": 1398013952
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003541706071496122,
      "loss": 2.8075,
      "theoretical_loss": 3.538111185559385,
      "tokens_seen": 1398079488
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035415405712124076,
      "loss": 2.6048,
      "theoretical_loss": 3.5380963648548205,
      "tokens_seen": 1398145024
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003541375070928693,
      "loss": 2.4595,
      "theoretical_loss": 3.538081545039444,
      "tokens_seen": 1398210560
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003541209570644978,
      "loss": 2.588,
      "theoretical_loss": 3.5380667261131595,
      "tokens_seen": 1398276096
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003541044070361263,
      "loss": 2.6704,
      "theoretical_loss": 3.5380519080758734,
      "tokens_seen": 1398341632
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035408785700775486,
      "loss": 2.7035,
      "theoretical_loss": 3.5380370909274896,
      "tokens_seen": 1398407168
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003540713069793834,
      "loss": 2.877,
      "theoretical_loss": 3.538022274667914,
      "tokens_seen": 1398472704
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003540547569510119,
      "loss": 2.7714,
      "theoretical_loss": 3.5380074592970514,
      "tokens_seen": 1398538240
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003540382069226404,
      "loss": 2.7385,
      "theoretical_loss": 3.537992644814807,
      "tokens_seen": 1398603776
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035402165689426895,
      "loss": 2.7033,
      "theoretical_loss": 3.5379778312210846,
      "tokens_seen": 1398669312
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003540051068658975,
      "loss": 2.7243,
      "theoretical_loss": 3.537963018515791,
      "tokens_seen": 1398734848
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000353988556837526,
      "loss": 2.6293,
      "theoretical_loss": 3.53794820669883,
      "tokens_seen": 1398800384
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003539720068091545,
      "loss": 2.6994,
      "theoretical_loss": 3.537933395770108,
      "tokens_seen": 1398865920
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035395545678078305,
      "loss": 2.9383,
      "theoretical_loss": 3.5379185857295283,
      "tokens_seen": 1398931456
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003539389067524116,
      "loss": 2.5805,
      "theoretical_loss": 3.5379037765769983,
      "tokens_seen": 1398996992
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003539223567240401,
      "loss": 2.7889,
      "theoretical_loss": 3.537888968312421,
      "tokens_seen": 1399062528
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003539058066956686,
      "loss": 2.9203,
      "theoretical_loss": 3.5378741609357034,
      "tokens_seen": 1399128064
    },
    {
      "epoch": 0.5,
      "objective/train/avg_token_score": 0.010972633957862854,
      "objective/train/avg_weight": 0.9870882630348206,
      "objective/train/docs_used": 793176,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7209882736206055,
      "objective/train/std_weight": 0.04937069118022919,
      "objective/train/theoretical_loss": 3.5378593544467494,
      "objective/train/tokens_used": 1419653600,
      "theoretical_loss": 3.5378593544467494,
      "tokens_seen": 1399193600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035388925666729714,
      "loss": 2.7401,
      "theoretical_loss": 3.5378593544467494,
      "tokens_seen": 1399193600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003538727066389257,
      "loss": 2.3824,
      "theoretical_loss": 3.537844548845464,
      "tokens_seen": 1399259136
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003538561566105542,
      "loss": 2.747,
      "theoretical_loss": 3.537829744131754,
      "tokens_seen": 1399324672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003538396065821827,
      "loss": 2.5659,
      "theoretical_loss": 3.5378149403055232,
      "tokens_seen": 1399390208
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035382305655381124,
      "loss": 2.9168,
      "theoretical_loss": 3.537800137366677,
      "tokens_seen": 1399455744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003538065065254398,
      "loss": 2.7097,
      "theoretical_loss": 3.5377853353151214,
      "tokens_seen": 1399521280
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003537899564970683,
      "loss": 2.4897,
      "theoretical_loss": 3.537770534150761,
      "tokens_seen": 1399586816
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003537734064686968,
      "loss": 2.6639,
      "theoretical_loss": 3.5377557338735017,
      "tokens_seen": 1399652352
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035375685644032533,
      "loss": 2.8896,
      "theoretical_loss": 3.537740934483248,
      "tokens_seen": 1399717888
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035374030641195387,
      "loss": 2.5224,
      "theoretical_loss": 3.5377261359799057,
      "tokens_seen": 1399783424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003537237563835824,
      "loss": 2.6145,
      "theoretical_loss": 3.5377113383633803,
      "tokens_seen": 1399848960
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003537072063552109,
      "loss": 2.4196,
      "theoretical_loss": 3.5376965416335766,
      "tokens_seen": 1399914496
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035369065632683943,
      "loss": 2.7932,
      "theoretical_loss": 3.5376817457904,
      "tokens_seen": 1399980032
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035367410629846797,
      "loss": 2.6867,
      "theoretical_loss": 3.5376669508337564,
      "tokens_seen": 1400045568
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003536575562700965,
      "loss": 2.4853,
      "theoretical_loss": 3.537652156763551,
      "tokens_seen": 1400111104
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000353641006241725,
      "loss": 2.7665,
      "theoretical_loss": 3.537637363579689,
      "tokens_seen": 1400176640
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003536244562133535,
      "loss": 2.5705,
      "theoretical_loss": 3.537622571282076,
      "tokens_seen": 1400242176
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035360790618498206,
      "loss": 2.5854,
      "theoretical_loss": 3.5376077798706174,
      "tokens_seen": 1400307712
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003535913561566106,
      "loss": 2.7649,
      "theoretical_loss": 3.5375929893452187,
      "tokens_seen": 1400373248
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003535748061282391,
      "loss": 2.4937,
      "theoretical_loss": 3.537578199705785,
      "tokens_seen": 1400438784
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003535582560998676,
      "loss": 2.587,
      "theoretical_loss": 3.537563410952222,
      "tokens_seen": 1400504320
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035354170607149616,
      "loss": 2.7803,
      "theoretical_loss": 3.5375486230844353,
      "tokens_seen": 1400569856
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003535251560431247,
      "loss": 2.748,
      "theoretical_loss": 3.5375338361023303,
      "tokens_seen": 1400635392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003535086060147532,
      "loss": 2.7282,
      "theoretical_loss": 3.5375190500058125,
      "tokens_seen": 1400700928
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003534920559863817,
      "loss": 2.6893,
      "theoretical_loss": 3.5375042647947876,
      "tokens_seen": 1400766464
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035347550595801026,
      "loss": 2.4931,
      "theoretical_loss": 3.537489480469161,
      "tokens_seen": 1400832000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003534589559296388,
      "loss": 2.7335,
      "theoretical_loss": 3.5374746970288378,
      "tokens_seen": 1400897536
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003534424059012673,
      "loss": 2.6011,
      "theoretical_loss": 3.5374599144737244,
      "tokens_seen": 1400963072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035342585587289576,
      "loss": 2.671,
      "theoretical_loss": 3.537445132803726,
      "tokens_seen": 1401028608
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035340930584452424,
      "loss": 2.7865,
      "theoretical_loss": 3.537430352018749,
      "tokens_seen": 1401094144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003533927558161528,
      "loss": 2.6323,
      "theoretical_loss": 3.537415572118697,
      "tokens_seen": 1401159680
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003533762057877813,
      "loss": 2.7893,
      "theoretical_loss": 3.5374007931034774,
      "tokens_seen": 1401225216
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035335965575940986,
      "loss": 2.9919,
      "theoretical_loss": 3.5373860149729954,
      "tokens_seen": 1401290752
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035334310573103834,
      "loss": 2.4829,
      "theoretical_loss": 3.5373712377271564,
      "tokens_seen": 1401356288
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003533265557026669,
      "loss": 2.8341,
      "theoretical_loss": 3.537356461365867,
      "tokens_seen": 1401421824
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003533100056742954,
      "loss": 2.4504,
      "theoretical_loss": 3.5373416858890314,
      "tokens_seen": 1401487360
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035329345564592395,
      "loss": 2.608,
      "theoretical_loss": 3.5373269112965566,
      "tokens_seen": 1401552896
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035327690561755244,
      "loss": 2.8332,
      "theoretical_loss": 3.537312137588347,
      "tokens_seen": 1401618432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000353260355589181,
      "loss": 2.6786,
      "theoretical_loss": 3.53729736476431,
      "tokens_seen": 1401683968
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003532438055608095,
      "loss": 2.5227,
      "theoretical_loss": 3.53728259282435,
      "tokens_seen": 1401749504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035322725553243805,
      "loss": 2.7893,
      "theoretical_loss": 3.5372678217683733,
      "tokens_seen": 1401815040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035321070550406653,
      "loss": 2.6742,
      "theoretical_loss": 3.5372530515962852,
      "tokens_seen": 1401880576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035319415547569507,
      "loss": 2.6897,
      "theoretical_loss": 3.5372382823079924,
      "tokens_seen": 1401946112
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003531776054473236,
      "loss": 2.6668,
      "theoretical_loss": 3.5372235139034,
      "tokens_seen": 1402011648
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035316105541895214,
      "loss": 2.6124,
      "theoretical_loss": 3.5372087463824142,
      "tokens_seen": 1402077184
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035314450539058063,
      "loss": 2.7136,
      "theoretical_loss": 3.5371939797449405,
      "tokens_seen": 1402142720
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035312795536220917,
      "loss": 2.4817,
      "theoretical_loss": 3.5371792139908855,
      "tokens_seen": 1402208256
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003531114053338377,
      "loss": 2.6667,
      "theoretical_loss": 3.5371644491201537,
      "tokens_seen": 1402273792
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035309485530546624,
      "loss": 2.8508,
      "theoretical_loss": 3.5371496851326523,
      "tokens_seen": 1402339328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003530783052770947,
      "loss": 2.7731,
      "theoretical_loss": 3.5371349220282866,
      "tokens_seen": 1402404864
    },
    {
      "epoch": 0.5,
      "objective/train/avg_token_score": 0.01858796551823616,
      "objective/train/avg_weight": 0.9781519174575806,
      "objective/train/docs_used": 795165,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.555966854095459,
      "objective/train/std_weight": 0.09123383462429047,
      "objective/train/theoretical_loss": 3.537120159806962,
      "objective/train/tokens_used": 1422930400,
      "theoretical_loss": 3.537120159806962,
      "tokens_seen": 1402470400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035306175524872326,
      "loss": 2.4729,
      "theoretical_loss": 3.537120159806962,
      "tokens_seen": 1402470400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003530452052203518,
      "loss": 2.5956,
      "theoretical_loss": 3.5371053984685856,
      "tokens_seen": 1402535936
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035302865519198034,
      "loss": 2.7305,
      "theoretical_loss": 3.5370906380130624,
      "tokens_seen": 1402601472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003530121051636088,
      "loss": 2.6979,
      "theoretical_loss": 3.537075878440299,
      "tokens_seen": 1402667008
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035299555513523736,
      "loss": 2.5361,
      "theoretical_loss": 3.5370611197502004,
      "tokens_seen": 1402732544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003529790051068659,
      "loss": 2.7211,
      "theoretical_loss": 3.5370463619426733,
      "tokens_seen": 1402798080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035296245507849443,
      "loss": 2.9144,
      "theoretical_loss": 3.5370316050176243,
      "tokens_seen": 1402863616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003529459050501229,
      "loss": 2.7277,
      "theoretical_loss": 3.5370168489749583,
      "tokens_seen": 1402929152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035292935502175145,
      "loss": 2.7425,
      "theoretical_loss": 3.5370020938145816,
      "tokens_seen": 1402994688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035291280499338,
      "loss": 2.8481,
      "theoretical_loss": 3.5369873395364007,
      "tokens_seen": 1403060224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035289625496500853,
      "loss": 2.6816,
      "theoretical_loss": 3.5369725861403207,
      "tokens_seen": 1403125760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000352879704936637,
      "loss": 2.6116,
      "theoretical_loss": 3.536957833626249,
      "tokens_seen": 1403191296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035286315490826555,
      "loss": 2.6602,
      "theoretical_loss": 3.5369430819940906,
      "tokens_seen": 1403256832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003528466048798941,
      "loss": 2.6959,
      "theoretical_loss": 3.536928331243752,
      "tokens_seen": 1403322368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003528300548515226,
      "loss": 2.7782,
      "theoretical_loss": 3.5369135813751393,
      "tokens_seen": 1403387904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003528135048231511,
      "loss": 2.5526,
      "theoretical_loss": 3.536898832388159,
      "tokens_seen": 1403453440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035279695479477965,
      "loss": 2.6721,
      "theoretical_loss": 3.5368840842827165,
      "tokens_seen": 1403518976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003527804047664082,
      "loss": 2.7072,
      "theoretical_loss": 3.536869337058718,
      "tokens_seen": 1403584512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003527638547380367,
      "loss": 2.4882,
      "theoretical_loss": 3.5368545907160707,
      "tokens_seen": 1403650048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003527473047096652,
      "loss": 2.7719,
      "theoretical_loss": 3.536839845254679,
      "tokens_seen": 1403715584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035273075468129374,
      "loss": 2.6946,
      "theoretical_loss": 3.536825100674451,
      "tokens_seen": 1403781120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003527142046529223,
      "loss": 2.4644,
      "theoretical_loss": 3.5368103569752916,
      "tokens_seen": 1403846656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003526976546245508,
      "loss": 2.6806,
      "theoretical_loss": 3.5367956141571075,
      "tokens_seen": 1403912192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003526811045961793,
      "loss": 2.6619,
      "theoretical_loss": 3.536780872219805,
      "tokens_seen": 1403977728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035266455456780784,
      "loss": 2.7653,
      "theoretical_loss": 3.53676613116329,
      "tokens_seen": 1404043264
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003526480045394364,
      "loss": 2.7218,
      "theoretical_loss": 3.5367513909874697,
      "tokens_seen": 1404108800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003526314545110649,
      "loss": 2.6597,
      "theoretical_loss": 3.536736651692249,
      "tokens_seen": 1404174336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003526149044826934,
      "loss": 2.5281,
      "theoretical_loss": 3.5367219132775354,
      "tokens_seen": 1404239872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035259835445432193,
      "loss": 2.8489,
      "theoretical_loss": 3.5367071757432345,
      "tokens_seen": 1404305408
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035258180442595047,
      "loss": 2.6966,
      "theoretical_loss": 3.5366924390892525,
      "tokens_seen": 1404370944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000352565254397579,
      "loss": 2.6978,
      "theoretical_loss": 3.5366777033154966,
      "tokens_seen": 1404436480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003525487043692075,
      "loss": 2.7076,
      "theoretical_loss": 3.536662968421872,
      "tokens_seen": 1404502016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035253215434083603,
      "loss": 2.4798,
      "theoretical_loss": 3.536648234408286,
      "tokens_seen": 1404567552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035251560431246457,
      "loss": 2.6186,
      "theoretical_loss": 3.536633501274645,
      "tokens_seen": 1404633088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003524990542840931,
      "loss": 2.4351,
      "theoretical_loss": 3.536618769020855,
      "tokens_seen": 1404698624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035248250425572164,
      "loss": 2.8712,
      "theoretical_loss": 3.536604037646822,
      "tokens_seen": 1404764160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003524659542273501,
      "loss": 2.5357,
      "theoretical_loss": 3.5365893071524535,
      "tokens_seen": 1404829696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003524494041989786,
      "loss": 2.5825,
      "theoretical_loss": 3.536574577537655,
      "tokens_seen": 1404895232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003524328541706071,
      "loss": 2.7139,
      "theoretical_loss": 3.5365598488023333,
      "tokens_seen": 1404960768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035241630414223563,
      "loss": 2.6356,
      "theoretical_loss": 3.536545120946395,
      "tokens_seen": 1405026304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035239975411386417,
      "loss": 2.7921,
      "theoretical_loss": 3.5365303939697466,
      "tokens_seen": 1405091840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003523832040854927,
      "loss": 2.7295,
      "theoretical_loss": 3.536515667872294,
      "tokens_seen": 1405157376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003523666540571212,
      "loss": 2.7123,
      "theoretical_loss": 3.5365009426539444,
      "tokens_seen": 1405222912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003523501040287497,
      "loss": 2.6504,
      "theoretical_loss": 3.5364862183146037,
      "tokens_seen": 1405288448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035233355400037826,
      "loss": 2.5041,
      "theoretical_loss": 3.5364714948541796,
      "tokens_seen": 1405353984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003523170039720068,
      "loss": 2.5308,
      "theoretical_loss": 3.536456772272577,
      "tokens_seen": 1405419520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003523004539436353,
      "loss": 2.7482,
      "theoretical_loss": 3.536442050569704,
      "tokens_seen": 1405485056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003522839039152638,
      "loss": 2.5855,
      "theoretical_loss": 3.5364273297454663,
      "tokens_seen": 1405550592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035226735388689236,
      "loss": 2.5773,
      "theoretical_loss": 3.5364126097997706,
      "tokens_seen": 1405616128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003522508038585209,
      "loss": 2.7109,
      "theoretical_loss": 3.5363978907325233,
      "tokens_seen": 1405681664
    },
    {
      "epoch": 0.5,
      "objective/train/avg_token_score": 0.01821300759911537,
      "objective/train/avg_weight": 0.9785669445991516,
      "objective/train/docs_used": 797297,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.253760576248169,
      "objective/train/std_weight": 0.10831209272146225,
      "objective/train/theoretical_loss": 3.5363831725436317,
      "objective/train/tokens_used": 1426207200,
      "theoretical_loss": 3.5363831725436317,
      "tokens_seen": 1405747200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003522342538301494,
      "loss": 2.7956,
      "theoretical_loss": 3.5363831725436317,
      "tokens_seen": 1405747200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003522177038017779,
      "loss": 2.666,
      "theoretical_loss": 3.536368455233003,
      "tokens_seen": 1405812736
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035220115377340645,
      "loss": 2.668,
      "theoretical_loss": 3.5363537388005417,
      "tokens_seen": 1405878272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000352184603745035,
      "loss": 2.8502,
      "theoretical_loss": 3.536339023246156,
      "tokens_seen": 1405943808
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003521680537166635,
      "loss": 2.6371,
      "theoretical_loss": 3.5363243085697524,
      "tokens_seen": 1406009344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000352151503688292,
      "loss": 2.8749,
      "theoretical_loss": 3.5363095947712377,
      "tokens_seen": 1406074880
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035213495365992055,
      "loss": 2.7817,
      "theoretical_loss": 3.5362948818505178,
      "tokens_seen": 1406140416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003521184036315491,
      "loss": 2.4698,
      "theoretical_loss": 3.5362801698075006,
      "tokens_seen": 1406205952
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035210185360317757,
      "loss": 2.7483,
      "theoretical_loss": 3.536265458642092,
      "tokens_seen": 1406271488
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003520853035748061,
      "loss": 2.5835,
      "theoretical_loss": 3.536250748354199,
      "tokens_seen": 1406337024
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035206875354643465,
      "loss": 2.4791,
      "theoretical_loss": 3.536236038943729,
      "tokens_seen": 1406402560
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003520522035180632,
      "loss": 2.5036,
      "theoretical_loss": 3.536221330410587,
      "tokens_seen": 1406468096
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035203565348969167,
      "loss": 2.7932,
      "theoretical_loss": 3.5362066227546816,
      "tokens_seen": 1406533632
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003520191034613202,
      "loss": 2.6039,
      "theoretical_loss": 3.5361919159759196,
      "tokens_seen": 1406599168
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035200255343294874,
      "loss": 2.5436,
      "theoretical_loss": 3.5361772100742064,
      "tokens_seen": 1406664704
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003519860034045773,
      "loss": 2.6164,
      "theoretical_loss": 3.5361625050494494,
      "tokens_seen": 1406730240
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035196945337620576,
      "loss": 2.6843,
      "theoretical_loss": 3.5361478009015563,
      "tokens_seen": 1406795776
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003519529033478343,
      "loss": 2.6559,
      "theoretical_loss": 3.536133097630433,
      "tokens_seen": 1406861312
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035193635331946284,
      "loss": 2.479,
      "theoretical_loss": 3.536118395235987,
      "tokens_seen": 1406926848
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003519198032910914,
      "loss": 2.5654,
      "theoretical_loss": 3.5361036937181254,
      "tokens_seen": 1406992384
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035190325326271986,
      "loss": 2.5443,
      "theoretical_loss": 3.5360889930767536,
      "tokens_seen": 1407057920
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003518867032343484,
      "loss": 2.5729,
      "theoretical_loss": 3.5360742933117804,
      "tokens_seen": 1407123456
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035187015320597693,
      "loss": 2.6312,
      "theoretical_loss": 3.5360595944231115,
      "tokens_seen": 1407188992
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035185360317760547,
      "loss": 2.6584,
      "theoretical_loss": 3.536044896410654,
      "tokens_seen": 1407254528
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035183705314923396,
      "loss": 2.6843,
      "theoretical_loss": 3.5360301992743155,
      "tokens_seen": 1407320064
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003518205031208625,
      "loss": 2.7318,
      "theoretical_loss": 3.536015503014003,
      "tokens_seen": 1407385600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035180395309249103,
      "loss": 2.7368,
      "theoretical_loss": 3.5360008076296223,
      "tokens_seen": 1407451136
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035178740306411957,
      "loss": 2.7625,
      "theoretical_loss": 3.5359861131210817,
      "tokens_seen": 1407516672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035177085303574805,
      "loss": 2.5785,
      "theoretical_loss": 3.5359714194882876,
      "tokens_seen": 1407582208
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003517543030073766,
      "loss": 2.5108,
      "theoretical_loss": 3.535956726731147,
      "tokens_seen": 1407647744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003517377529790051,
      "loss": 2.7752,
      "theoretical_loss": 3.535942034849567,
      "tokens_seen": 1407713280
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035172120295063366,
      "loss": 2.8221,
      "theoretical_loss": 3.535927343843455,
      "tokens_seen": 1407778816
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035170465292226215,
      "loss": 2.5912,
      "theoretical_loss": 3.5359126537127183,
      "tokens_seen": 1407844352
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003516881028938907,
      "loss": 2.4872,
      "theoretical_loss": 3.535897964457263,
      "tokens_seen": 1407909888
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003516715528655192,
      "loss": 2.629,
      "theoretical_loss": 3.535883276076997,
      "tokens_seen": 1407975424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035165500283714776,
      "loss": 2.61,
      "theoretical_loss": 3.5358685885718275,
      "tokens_seen": 1408040960
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035163845280877624,
      "loss": 2.4585,
      "theoretical_loss": 3.535853901941661,
      "tokens_seen": 1408106496
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003516219027804048,
      "loss": 2.8761,
      "theoretical_loss": 3.535839216186405,
      "tokens_seen": 1408172032
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003516053527520333,
      "loss": 2.7153,
      "theoretical_loss": 3.535824531305966,
      "tokens_seen": 1408237568
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035158880272366186,
      "loss": 2.644,
      "theoretical_loss": 3.5358098473002526,
      "tokens_seen": 1408303104
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035157225269529034,
      "loss": 2.4983,
      "theoretical_loss": 3.5357951641691714,
      "tokens_seen": 1408368640
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003515557026669189,
      "loss": 2.4775,
      "theoretical_loss": 3.5357804819126293,
      "tokens_seen": 1408434176
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003515391526385474,
      "loss": 2.5717,
      "theoretical_loss": 3.5357658005305335,
      "tokens_seen": 1408499712
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035152260261017595,
      "loss": 2.7431,
      "theoretical_loss": 3.535751120022791,
      "tokens_seen": 1408565248
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035150605258180444,
      "loss": 2.6593,
      "theoretical_loss": 3.53573644038931,
      "tokens_seen": 1408630784
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035148950255343297,
      "loss": 2.7376,
      "theoretical_loss": 3.535721761629997,
      "tokens_seen": 1408696320
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035147295252506146,
      "loss": 2.5725,
      "theoretical_loss": 3.53570708374476,
      "tokens_seen": 1408761856
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035145640249668994,
      "loss": 2.6333,
      "theoretical_loss": 3.5356924067335047,
      "tokens_seen": 1408827392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003514398524683185,
      "loss": 2.534,
      "theoretical_loss": 3.53567773059614,
      "tokens_seen": 1408892928
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000351423302439947,
      "loss": 2.5816,
      "theoretical_loss": 3.5356630553325727,
      "tokens_seen": 1408958464
    },
    {
      "epoch": 0.5,
      "objective/train/avg_token_score": 0.01757778227329254,
      "objective/train/avg_weight": 0.9793093800544739,
      "objective/train/docs_used": 799421,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6493234634399414,
      "objective/train/std_weight": 0.0872621089220047,
      "objective/train/theoretical_loss": 3.5356483809427104,
      "objective/train/tokens_used": 1429484000,
      "theoretical_loss": 3.5356483809427104,
      "tokens_seen": 1409024000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035140675241157555,
      "loss": 2.6797,
      "theoretical_loss": 3.5356483809427104,
      "tokens_seen": 1409024000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035139020238320404,
      "loss": 2.502,
      "theoretical_loss": 3.5356337074264603,
      "tokens_seen": 1409089536
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003513736523548326,
      "loss": 2.5875,
      "theoretical_loss": 3.5356190347837293,
      "tokens_seen": 1409155072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003513571023264611,
      "loss": 2.5555,
      "theoretical_loss": 3.535604363014425,
      "tokens_seen": 1409220608
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035134055229808965,
      "loss": 2.7634,
      "theoretical_loss": 3.535589692118455,
      "tokens_seen": 1409286144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035132400226971813,
      "loss": 2.7404,
      "theoretical_loss": 3.5355750220957267,
      "tokens_seen": 1409351680
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035130745224134667,
      "loss": 2.765,
      "theoretical_loss": 3.5355603529461477,
      "tokens_seen": 1409417216
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003512909022129752,
      "loss": 2.7059,
      "theoretical_loss": 3.535545684669625,
      "tokens_seen": 1409482752
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035127435218460374,
      "loss": 2.5049,
      "theoretical_loss": 3.535531017266066,
      "tokens_seen": 1409548288
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035125780215623223,
      "loss": 2.5585,
      "theoretical_loss": 3.535516350735379,
      "tokens_seen": 1409613824
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035124125212786076,
      "loss": 2.6554,
      "theoretical_loss": 3.5355016850774703,
      "tokens_seen": 1409679360
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003512247020994893,
      "loss": 2.6821,
      "theoretical_loss": 3.535487020292248,
      "tokens_seen": 1409744896
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035120815207111784,
      "loss": 2.7201,
      "theoretical_loss": 3.5354723563796195,
      "tokens_seen": 1409810432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003511916020427463,
      "loss": 2.7116,
      "theoretical_loss": 3.5354576933394926,
      "tokens_seen": 1409875968
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035117505201437486,
      "loss": 2.6428,
      "theoretical_loss": 3.5354430311717744,
      "tokens_seen": 1409941504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003511585019860034,
      "loss": 2.6667,
      "theoretical_loss": 3.535428369876373,
      "tokens_seen": 1410007040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035114195195763194,
      "loss": 2.4197,
      "theoretical_loss": 3.535413709453196,
      "tokens_seen": 1410072576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003511254019292604,
      "loss": 2.6582,
      "theoretical_loss": 3.53539904990215,
      "tokens_seen": 1410138112
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035110885190088896,
      "loss": 2.5674,
      "theoretical_loss": 3.5353843912231433,
      "tokens_seen": 1410203648
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003510923018725175,
      "loss": 2.4837,
      "theoretical_loss": 3.535369733416083,
      "tokens_seen": 1410269184
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035107575184414603,
      "loss": 2.5487,
      "theoretical_loss": 3.5353550764808777,
      "tokens_seen": 1410334720
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003510592018157745,
      "loss": 2.6887,
      "theoretical_loss": 3.535340420417434,
      "tokens_seen": 1410400256
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035104265178740305,
      "loss": 2.469,
      "theoretical_loss": 3.53532576522566,
      "tokens_seen": 1410465792
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003510261017590316,
      "loss": 2.593,
      "theoretical_loss": 3.5353111109054636,
      "tokens_seen": 1410531328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035100955173066013,
      "loss": 2.4376,
      "theoretical_loss": 3.5352964574567523,
      "tokens_seen": 1410596864
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003509930017022886,
      "loss": 2.7653,
      "theoretical_loss": 3.5352818048794337,
      "tokens_seen": 1410662400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035097645167391715,
      "loss": 2.5297,
      "theoretical_loss": 3.5352671531734154,
      "tokens_seen": 1410727936
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003509599016455457,
      "loss": 2.6614,
      "theoretical_loss": 3.535252502338605,
      "tokens_seen": 1410793472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003509433516171742,
      "loss": 2.6218,
      "theoretical_loss": 3.5352378523749106,
      "tokens_seen": 1410859008
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003509268015888027,
      "loss": 2.4431,
      "theoretical_loss": 3.53522320328224,
      "tokens_seen": 1410924544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035091025156043124,
      "loss": 2.5007,
      "theoretical_loss": 3.5352085550605006,
      "tokens_seen": 1410990080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003508937015320598,
      "loss": 2.5845,
      "theoretical_loss": 3.5351939077096004,
      "tokens_seen": 1411055616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003508771515036883,
      "loss": 2.631,
      "theoretical_loss": 3.5351792612294473,
      "tokens_seen": 1411121152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003508606014753168,
      "loss": 2.6616,
      "theoretical_loss": 3.5351646156199488,
      "tokens_seen": 1411186688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035084405144694534,
      "loss": 2.5752,
      "theoretical_loss": 3.5351499708810126,
      "tokens_seen": 1411252224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003508275014185739,
      "loss": 2.5883,
      "theoretical_loss": 3.5351353270125463,
      "tokens_seen": 1411317760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003508109513902024,
      "loss": 2.5761,
      "theoretical_loss": 3.5351206840144593,
      "tokens_seen": 1411383296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003507944013618309,
      "loss": 2.6975,
      "theoretical_loss": 3.5351060418866576,
      "tokens_seen": 1411448832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035077785133345944,
      "loss": 2.6049,
      "theoretical_loss": 3.5350914006290504,
      "tokens_seen": 1411514368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000350761301305088,
      "loss": 2.5538,
      "theoretical_loss": 3.5350767602415445,
      "tokens_seen": 1411579904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003507447512767165,
      "loss": 2.643,
      "theoretical_loss": 3.5350621207240485,
      "tokens_seen": 1411645440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000350728201248345,
      "loss": 2.4949,
      "theoretical_loss": 3.53504748207647,
      "tokens_seen": 1411710976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035071165121997353,
      "loss": 2.9268,
      "theoretical_loss": 3.5350328442987173,
      "tokens_seen": 1411776512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035069510119160207,
      "loss": 2.7522,
      "theoretical_loss": 3.535018207390698,
      "tokens_seen": 1411842048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003506785511632306,
      "loss": 2.7567,
      "theoretical_loss": 3.53500357135232,
      "tokens_seen": 1411907584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003506620011348591,
      "loss": 2.7949,
      "theoretical_loss": 3.534988936183492,
      "tokens_seen": 1411973120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035064545110648763,
      "loss": 2.7021,
      "theoretical_loss": 3.5349743018841204,
      "tokens_seen": 1412038656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035062890107811617,
      "loss": 2.4594,
      "theoretical_loss": 3.534959668454115,
      "tokens_seen": 1412104192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003506123510497447,
      "loss": 2.6774,
      "theoretical_loss": 3.5349450358933825,
      "tokens_seen": 1412169728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003505958010213732,
      "loss": 2.6146,
      "theoretical_loss": 3.5349304042018317,
      "tokens_seen": 1412235264
    },
    {
      "epoch": 0.5,
      "objective/train/avg_token_score": 0.019894368946552277,
      "objective/train/avg_weight": 0.9765930771827698,
      "objective/train/docs_used": 800610,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.716184139251709,
      "objective/train/std_weight": 0.07435636222362518,
      "objective/train/theoretical_loss": 3.53491577337937,
      "objective/train/tokens_used": 1432760800,
      "theoretical_loss": 3.53491577337937,
      "tokens_seen": 1412300800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003505792509930017,
      "loss": 2.7618,
      "theoretical_loss": 3.53491577337937,
      "tokens_seen": 1412300800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035056270096463026,
      "loss": 2.6,
      "theoretical_loss": 3.5349011434259063,
      "tokens_seen": 1412366336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003505461509362588,
      "loss": 2.732,
      "theoretical_loss": 3.5348865143413484,
      "tokens_seen": 1412431872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003505296009078873,
      "loss": 2.5452,
      "theoretical_loss": 3.534871886125604,
      "tokens_seen": 1412497408
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003505130508795158,
      "loss": 2.6788,
      "theoretical_loss": 3.534857258778581,
      "tokens_seen": 1412562944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035049650085114436,
      "loss": 2.8787,
      "theoretical_loss": 3.5348426323001876,
      "tokens_seen": 1412628480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003504799508227728,
      "loss": 2.4087,
      "theoretical_loss": 3.534828006690333,
      "tokens_seen": 1412694016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003504634007944013,
      "loss": 2.4044,
      "theoretical_loss": 3.534813381948924,
      "tokens_seen": 1412759552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035044685076602986,
      "loss": 2.6422,
      "theoretical_loss": 3.53479875807587,
      "tokens_seen": 1412825088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035043030073765835,
      "loss": 2.7836,
      "theoretical_loss": 3.534784135071077,
      "tokens_seen": 1412890624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003504137507092869,
      "loss": 2.4588,
      "theoretical_loss": 3.5347695129344556,
      "tokens_seen": 1412956160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003503972006809154,
      "loss": 2.5843,
      "theoretical_loss": 3.534754891665913,
      "tokens_seen": 1413021696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035038065065254396,
      "loss": 2.5924,
      "theoretical_loss": 3.534740271265358,
      "tokens_seen": 1413087232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035036410062417244,
      "loss": 2.5816,
      "theoretical_loss": 3.5347256517326975,
      "tokens_seen": 1413152768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000350347550595801,
      "loss": 2.7249,
      "theoretical_loss": 3.5347110330678406,
      "tokens_seen": 1413218304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003503310005674295,
      "loss": 2.5487,
      "theoretical_loss": 3.5346964152706954,
      "tokens_seen": 1413283840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035031445053905805,
      "loss": 2.5652,
      "theoretical_loss": 3.53468179834117,
      "tokens_seen": 1413349376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035029790051068654,
      "loss": 2.7744,
      "theoretical_loss": 3.534667182279173,
      "tokens_seen": 1413414912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003502813504823151,
      "loss": 2.5882,
      "theoretical_loss": 3.5346525670846125,
      "tokens_seen": 1413480448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003502648004539436,
      "loss": 2.5586,
      "theoretical_loss": 3.5346379527573974,
      "tokens_seen": 1413545984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035024825042557215,
      "loss": 2.7002,
      "theoretical_loss": 3.534623339297435,
      "tokens_seen": 1413611520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035023170039720063,
      "loss": 2.8446,
      "theoretical_loss": 3.534608726704634,
      "tokens_seen": 1413677056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035021515036882917,
      "loss": 2.6437,
      "theoretical_loss": 3.534594114978903,
      "tokens_seen": 1413742592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0003501986003404577,
      "loss": 2.2656,
      "theoretical_loss": 3.5345795041201504,
      "tokens_seen": 1413808128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035018205031208625,
      "loss": 2.4697,
      "theoretical_loss": 3.5345648941282843,
      "tokens_seen": 1413873664
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00035016550028371473,
      "loss": 2.5949,
      "theoretical_loss": 3.5345502850032133,
      "tokens_seen": 1413939200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00035014895025534327,
      "loss": 2.602,
      "theoretical_loss": 3.534535676744845,
      "tokens_seen": 1414004736
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003501324002269718,
      "loss": 2.7079,
      "theoretical_loss": 3.5345210693530893,
      "tokens_seen": 1414070272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00035011585019860034,
      "loss": 2.4747,
      "theoretical_loss": 3.5345064628278537,
      "tokens_seen": 1414135808
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003500993001702288,
      "loss": 2.8383,
      "theoretical_loss": 3.5344918571690465,
      "tokens_seen": 1414201344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00035008275014185736,
      "loss": 2.6404,
      "theoretical_loss": 3.5344772523765764,
      "tokens_seen": 1414266880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003500662001134859,
      "loss": 2.6098,
      "theoretical_loss": 3.534462648450352,
      "tokens_seen": 1414332416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00035004965008511444,
      "loss": 2.4762,
      "theoretical_loss": 3.534448045390282,
      "tokens_seen": 1414397952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003500331000567429,
      "loss": 2.5187,
      "theoretical_loss": 3.5344334431962743,
      "tokens_seen": 1414463488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00035001655002837146,
      "loss": 2.7399,
      "theoretical_loss": 3.5344188418682374,
      "tokens_seen": 1414529024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00035,
      "loss": 2.7701,
      "theoretical_loss": 3.5344042414060803,
      "tokens_seen": 1414594560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034998344997162853,
      "loss": 2.4395,
      "theoretical_loss": 3.5343896418097116,
      "tokens_seen": 1414660096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034996689994325707,
      "loss": 2.7567,
      "theoretical_loss": 3.5343750430790397,
      "tokens_seen": 1414725632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034995034991488555,
      "loss": 2.8995,
      "theoretical_loss": 3.5343604452139727,
      "tokens_seen": 1414791168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003499337998865141,
      "loss": 2.5637,
      "theoretical_loss": 3.5343458482144197,
      "tokens_seen": 1414856704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034991724985814263,
      "loss": 2.502,
      "theoretical_loss": 3.534331252080289,
      "tokens_seen": 1414922240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034990069982977117,
      "loss": 3.0244,
      "theoretical_loss": 3.5343166568114897,
      "tokens_seen": 1414987776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034988414980139965,
      "loss": 2.6971,
      "theoretical_loss": 3.5343020624079298,
      "tokens_seen": 1415053312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003498675997730282,
      "loss": 2.6295,
      "theoretical_loss": 3.5342874688695183,
      "tokens_seen": 1415118848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003498510497446567,
      "loss": 2.6219,
      "theoretical_loss": 3.534272876196164,
      "tokens_seen": 1415184384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034983449971628526,
      "loss": 2.8693,
      "theoretical_loss": 3.534258284387775,
      "tokens_seen": 1415249920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034981794968791375,
      "loss": 2.6883,
      "theoretical_loss": 3.5342436934442607,
      "tokens_seen": 1415315456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003498013996595423,
      "loss": 2.6729,
      "theoretical_loss": 3.5342291033655293,
      "tokens_seen": 1415380992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003497848496311708,
      "loss": 2.9062,
      "theoretical_loss": 3.53421451415149,
      "tokens_seen": 1415446528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034976829960279936,
      "loss": 2.6161,
      "theoretical_loss": 3.5341999258020502,
      "tokens_seen": 1415512064
    },
    {
      "epoch": 0.51,
      "objective/train/avg_token_score": 0.011515786871314049,
      "objective/train/avg_weight": 0.986451268196106,
      "objective/train/docs_used": 802603,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.083113193511963,
      "objective/train/std_weight": 0.06448455154895782,
      "objective/train/theoretical_loss": 3.5341853383171205,
      "objective/train/tokens_used": 1436037600,
      "theoretical_loss": 3.5341853383171205,
      "tokens_seen": 1415577600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034975174957442784,
      "loss": 2.6144,
      "theoretical_loss": 3.5341853383171205,
      "tokens_seen": 1415577600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003497351995460563,
      "loss": 2.6582,
      "theoretical_loss": 3.5341707516966085,
      "tokens_seen": 1415643136
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034971864951768486,
      "loss": 2.5528,
      "theoretical_loss": 3.534156165940423,
      "tokens_seen": 1415708672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003497020994893134,
      "loss": 2.8339,
      "theoretical_loss": 3.5341415810484733,
      "tokens_seen": 1415774208
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034968554946094194,
      "loss": 2.4472,
      "theoretical_loss": 3.5341269970206675,
      "tokens_seen": 1415839744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003496689994325704,
      "loss": 2.6269,
      "theoretical_loss": 3.534112413856915,
      "tokens_seen": 1415905280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034965244940419896,
      "loss": 2.8355,
      "theoretical_loss": 3.5340978315571245,
      "tokens_seen": 1415970816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003496358993758275,
      "loss": 2.4667,
      "theoretical_loss": 3.534083250121204,
      "tokens_seen": 1416036352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034961934934745603,
      "loss": 2.7033,
      "theoretical_loss": 3.5340686695490637,
      "tokens_seen": 1416101888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003496027993190845,
      "loss": 2.7003,
      "theoretical_loss": 3.5340540898406116,
      "tokens_seen": 1416167424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034958624929071306,
      "loss": 2.5925,
      "theoretical_loss": 3.534039510995757,
      "tokens_seen": 1416232960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003495696992623416,
      "loss": 2.6961,
      "theoretical_loss": 3.5340249330144085,
      "tokens_seen": 1416298496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034955314923397013,
      "loss": 2.6474,
      "theoretical_loss": 3.5340103558964753,
      "tokens_seen": 1416364032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003495365992055986,
      "loss": 2.6175,
      "theoretical_loss": 3.5339957796418657,
      "tokens_seen": 1416429568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034952004917722715,
      "loss": 2.773,
      "theoretical_loss": 3.5339812042504897,
      "tokens_seen": 1416495104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003495034991488557,
      "loss": 2.6993,
      "theoretical_loss": 3.533966629722255,
      "tokens_seen": 1416560640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003494869491204842,
      "loss": 2.4624,
      "theoretical_loss": 3.533952056057071,
      "tokens_seen": 1416626176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003494703990921127,
      "loss": 2.7423,
      "theoretical_loss": 3.5339374832548476,
      "tokens_seen": 1416691712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034945384906374125,
      "loss": 2.6457,
      "theoretical_loss": 3.5339229113154924,
      "tokens_seen": 1416757248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003494372990353698,
      "loss": 2.6725,
      "theoretical_loss": 3.5339083402389146,
      "tokens_seen": 1416822784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003494207490069983,
      "loss": 2.5093,
      "theoretical_loss": 3.5338937700250246,
      "tokens_seen": 1416888320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003494041989786268,
      "loss": 2.6044,
      "theoretical_loss": 3.53387920067373,
      "tokens_seen": 1416953856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034938764895025534,
      "loss": 2.4955,
      "theoretical_loss": 3.53386463218494,
      "tokens_seen": 1417019392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003493710989218839,
      "loss": 2.6941,
      "theoretical_loss": 3.533850064558564,
      "tokens_seen": 1417084928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003493545488935124,
      "loss": 2.6335,
      "theoretical_loss": 3.5338354977945112,
      "tokens_seen": 1417150464
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003493379988651409,
      "loss": 2.5666,
      "theoretical_loss": 3.5338209318926905,
      "tokens_seen": 1417216000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034932144883676944,
      "loss": 2.7021,
      "theoretical_loss": 3.5338063668530104,
      "tokens_seen": 1417281536
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000349304898808398,
      "loss": 2.8191,
      "theoretical_loss": 3.533791802675381,
      "tokens_seen": 1417347072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003492883487800265,
      "loss": 2.4603,
      "theoretical_loss": 3.533777239359711,
      "tokens_seen": 1417412608
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000349271798751655,
      "loss": 2.7014,
      "theoretical_loss": 3.5337626769059094,
      "tokens_seen": 1417478144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003492552487232835,
      "loss": 2.9276,
      "theoretical_loss": 3.5337481153138857,
      "tokens_seen": 1417543680
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000349238698694912,
      "loss": 2.626,
      "theoretical_loss": 3.5337335545835487,
      "tokens_seen": 1417609216
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034922214866654056,
      "loss": 2.5419,
      "theoretical_loss": 3.5337189947148078,
      "tokens_seen": 1417674752
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003492055986381691,
      "loss": 2.6352,
      "theoretical_loss": 3.5337044357075715,
      "tokens_seen": 1417740288
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003491890486097976,
      "loss": 2.5955,
      "theoretical_loss": 3.5336898775617502,
      "tokens_seen": 1417805824
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003491724985814261,
      "loss": 2.7421,
      "theoretical_loss": 3.5336753202772524,
      "tokens_seen": 1417871360
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034915594855305465,
      "loss": 2.424,
      "theoretical_loss": 3.5336607638539874,
      "tokens_seen": 1417936896
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003491393985246832,
      "loss": 3.0435,
      "theoretical_loss": 3.5336462082918647,
      "tokens_seen": 1418002432
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003491228484963117,
      "loss": 2.6866,
      "theoretical_loss": 3.5336316535907937,
      "tokens_seen": 1418067968
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003491062984679402,
      "loss": 2.7878,
      "theoretical_loss": 3.533617099750683,
      "tokens_seen": 1418133504
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034908974843956875,
      "loss": 2.8737,
      "theoretical_loss": 3.5336025467714416,
      "tokens_seen": 1418199040
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003490731984111973,
      "loss": 2.9179,
      "theoretical_loss": 3.5335879946529802,
      "tokens_seen": 1418264576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034905664838282577,
      "loss": 2.5784,
      "theoretical_loss": 3.5335734433952073,
      "tokens_seen": 1418330112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003490400983544543,
      "loss": 2.5996,
      "theoretical_loss": 3.5335588929980317,
      "tokens_seen": 1418395648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034902354832608284,
      "loss": 2.7155,
      "theoretical_loss": 3.5335443434613643,
      "tokens_seen": 1418461184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003490069982977114,
      "loss": 2.731,
      "theoretical_loss": 3.5335297947851125,
      "tokens_seen": 1418526720
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034899044826933987,
      "loss": 2.8034,
      "theoretical_loss": 3.533515246969187,
      "tokens_seen": 1418592256
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003489738982409684,
      "loss": 2.608,
      "theoretical_loss": 3.533500700013497,
      "tokens_seen": 1418657792
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034895734821259694,
      "loss": 2.7258,
      "theoretical_loss": 3.533486153917951,
      "tokens_seen": 1418723328
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003489407981842255,
      "loss": 2.6596,
      "theoretical_loss": 3.5334716086824605,
      "tokens_seen": 1418788864
    },
    {
      "epoch": 0.51,
      "objective/train/avg_token_score": 0.012168517336249352,
      "objective/train/avg_weight": 0.9856815338134766,
      "objective/train/docs_used": 804391,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.745335578918457,
      "objective/train/std_weight": 0.07229627668857574,
      "objective/train/theoretical_loss": 3.5334570643069325,
      "objective/train/tokens_used": 1439314400,
      "theoretical_loss": 3.5334570643069325,
      "tokens_seen": 1418854400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034892424815585396,
      "loss": 2.4046,
      "theoretical_loss": 3.5334570643069325,
      "tokens_seen": 1418854400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003489076981274825,
      "loss": 2.6102,
      "theoretical_loss": 3.5334425207912776,
      "tokens_seen": 1418919936
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034889114809911104,
      "loss": 2.7054,
      "theoretical_loss": 3.533427978135405,
      "tokens_seen": 1418985472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003488745980707396,
      "loss": 2.6307,
      "theoretical_loss": 3.5334134363392247,
      "tokens_seen": 1419051008
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034885804804236806,
      "loss": 2.7053,
      "theoretical_loss": 3.533398895402646,
      "tokens_seen": 1419116544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003488414980139966,
      "loss": 2.6924,
      "theoretical_loss": 3.5333843553255777,
      "tokens_seen": 1419182080
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034882494798562513,
      "loss": 2.5625,
      "theoretical_loss": 3.53336981610793,
      "tokens_seen": 1419247616
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034880839795725367,
      "loss": 2.6372,
      "theoretical_loss": 3.5333552777496124,
      "tokens_seen": 1419313152
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034879184792888215,
      "loss": 2.7506,
      "theoretical_loss": 3.5333407402505337,
      "tokens_seen": 1419378688
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003487752979005107,
      "loss": 2.7841,
      "theoretical_loss": 3.5333262036106046,
      "tokens_seen": 1419444224
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003487587478721392,
      "loss": 2.6646,
      "theoretical_loss": 3.533311667829734,
      "tokens_seen": 1419509760
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003487421978437677,
      "loss": 2.6816,
      "theoretical_loss": 3.5332971329078315,
      "tokens_seen": 1419575296
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034872564781539625,
      "loss": 2.5726,
      "theoretical_loss": 3.5332825988448064,
      "tokens_seen": 1419640832
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003487090977870248,
      "loss": 2.6667,
      "theoretical_loss": 3.533268065640569,
      "tokens_seen": 1419706368
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034869254775865327,
      "loss": 2.5942,
      "theoretical_loss": 3.533253533295028,
      "tokens_seen": 1419771904
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003486759977302818,
      "loss": 2.9231,
      "theoretical_loss": 3.533239001808094,
      "tokens_seen": 1419837440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034865944770191034,
      "loss": 2.5431,
      "theoretical_loss": 3.5332244711796763,
      "tokens_seen": 1419902976
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003486428976735389,
      "loss": 2.8673,
      "theoretical_loss": 3.533209941409684,
      "tokens_seen": 1419968512
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034862634764516737,
      "loss": 2.7016,
      "theoretical_loss": 3.533195412498028,
      "tokens_seen": 1420034048
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003486097976167959,
      "loss": 2.7448,
      "theoretical_loss": 3.5331808844446164,
      "tokens_seen": 1420099584
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034859324758842444,
      "loss": 2.5669,
      "theoretical_loss": 3.5331663572493603,
      "tokens_seen": 1420165120
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000348576697560053,
      "loss": 2.5799,
      "theoretical_loss": 3.5331518309121686,
      "tokens_seen": 1420230656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034856014753168146,
      "loss": 2.6964,
      "theoretical_loss": 3.533137305432951,
      "tokens_seen": 1420296192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034854359750331,
      "loss": 2.875,
      "theoretical_loss": 3.5331227808116177,
      "tokens_seen": 1420361728
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034852704747493854,
      "loss": 2.7152,
      "theoretical_loss": 3.5331082570480787,
      "tokens_seen": 1420427264
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003485104974465671,
      "loss": 2.4792,
      "theoretical_loss": 3.533093734142243,
      "tokens_seen": 1420492800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034849394741819556,
      "loss": 2.8123,
      "theoretical_loss": 3.5330792120940204,
      "tokens_seen": 1420558336
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003484773973898241,
      "loss": 2.6262,
      "theoretical_loss": 3.5330646909033216,
      "tokens_seen": 1420623872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034846084736145263,
      "loss": 2.5193,
      "theoretical_loss": 3.533050170570055,
      "tokens_seen": 1420689408
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034844429733308117,
      "loss": 2.8584,
      "theoretical_loss": 3.5330356510941323,
      "tokens_seen": 1420754944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034842774730470965,
      "loss": 2.7696,
      "theoretical_loss": 3.5330211324754615,
      "tokens_seen": 1420820480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003484111972763382,
      "loss": 2.5542,
      "theoretical_loss": 3.533006614713954,
      "tokens_seen": 1420886016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034839464724796673,
      "loss": 2.5045,
      "theoretical_loss": 3.532992097809518,
      "tokens_seen": 1420951552
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034837809721959527,
      "loss": 2.5347,
      "theoretical_loss": 3.5329775817620646,
      "tokens_seen": 1421017088
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034836154719122375,
      "loss": 2.5526,
      "theoretical_loss": 3.532963066571503,
      "tokens_seen": 1421082624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003483449971628523,
      "loss": 2.8255,
      "theoretical_loss": 3.5329485522377437,
      "tokens_seen": 1421148160
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003483284471344808,
      "loss": 2.6057,
      "theoretical_loss": 3.532934038760697,
      "tokens_seen": 1421213696
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034831189710610936,
      "loss": 2.7662,
      "theoretical_loss": 3.5329195261402715,
      "tokens_seen": 1421279232
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034829534707773785,
      "loss": 2.4505,
      "theoretical_loss": 3.532905014376378,
      "tokens_seen": 1421344768
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034827879704936633,
      "loss": 2.4873,
      "theoretical_loss": 3.532890503468926,
      "tokens_seen": 1421410304
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034826224702099487,
      "loss": 2.5351,
      "theoretical_loss": 3.532875993417826,
      "tokens_seen": 1421475840
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003482456969926234,
      "loss": 2.5735,
      "theoretical_loss": 3.5328614842229875,
      "tokens_seen": 1421541376
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034822914696425194,
      "loss": 2.4502,
      "theoretical_loss": 3.532846975884321,
      "tokens_seen": 1421606912
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003482125969358804,
      "loss": 2.7162,
      "theoretical_loss": 3.5328324684017365,
      "tokens_seen": 1421672448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034819604690750896,
      "loss": 2.6218,
      "theoretical_loss": 3.5328179617751436,
      "tokens_seen": 1421737984
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003481794968791375,
      "loss": 2.8991,
      "theoretical_loss": 3.5328034560044523,
      "tokens_seen": 1421803520
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034816294685076604,
      "loss": 2.5796,
      "theoretical_loss": 3.5327889510895734,
      "tokens_seen": 1421869056
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003481463968223945,
      "loss": 2.8261,
      "theoretical_loss": 3.532774447030416,
      "tokens_seen": 1421934592
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034812984679402306,
      "loss": 2.665,
      "theoretical_loss": 3.532759943826891,
      "tokens_seen": 1422000128
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003481132967656516,
      "loss": 2.5196,
      "theoretical_loss": 3.532745441478908,
      "tokens_seen": 1422065664
    },
    {
      "epoch": 0.51,
      "objective/train/avg_token_score": 0.0252030361443758,
      "objective/train/avg_weight": 0.9703326225280762,
      "objective/train/docs_used": 806541,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9002938270568848,
      "objective/train/std_weight": 0.10464319586753845,
      "objective/train/theoretical_loss": 3.5327309399863767,
      "objective/train/tokens_used": 1442591200,
      "theoretical_loss": 3.5327309399863767,
      "tokens_seen": 1422131200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034809674673728013,
      "loss": 2.7681,
      "theoretical_loss": 3.5327309399863767,
      "tokens_seen": 1422131200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003480801967089086,
      "loss": 2.6614,
      "theoretical_loss": 3.5327164393492083,
      "tokens_seen": 1422196736
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034806364668053715,
      "loss": 2.8354,
      "theoretical_loss": 3.5327019395673123,
      "tokens_seen": 1422262272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003480470966521657,
      "loss": 2.8057,
      "theoretical_loss": 3.532687440640599,
      "tokens_seen": 1422327808
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034803054662379423,
      "loss": 3.0085,
      "theoretical_loss": 3.5326729425689782,
      "tokens_seen": 1422393344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003480139965954227,
      "loss": 2.598,
      "theoretical_loss": 3.5326584453523604,
      "tokens_seen": 1422458880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034799744656705125,
      "loss": 2.5903,
      "theoretical_loss": 3.532643948990656,
      "tokens_seen": 1422524416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003479808965386798,
      "loss": 2.637,
      "theoretical_loss": 3.532629453483775,
      "tokens_seen": 1422589952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003479643465103083,
      "loss": 2.1973,
      "theoretical_loss": 3.532614958831627,
      "tokens_seen": 1422655488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003479477964819368,
      "loss": 2.7479,
      "theoretical_loss": 3.5326004650341236,
      "tokens_seen": 1422721024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034793124645356535,
      "loss": 2.3438,
      "theoretical_loss": 3.532585972091174,
      "tokens_seen": 1422786560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003479146964251939,
      "loss": 2.6764,
      "theoretical_loss": 3.532571480002688,
      "tokens_seen": 1422852096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003478981463968224,
      "loss": 2.6146,
      "theoretical_loss": 3.5325569887685777,
      "tokens_seen": 1422917632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003478815963684509,
      "loss": 2.8283,
      "theoretical_loss": 3.532542498388752,
      "tokens_seen": 1422983168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034786504634007944,
      "loss": 2.6242,
      "theoretical_loss": 3.532528008863121,
      "tokens_seen": 1423048704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000347848496311708,
      "loss": 2.7717,
      "theoretical_loss": 3.5325135201915954,
      "tokens_seen": 1423114240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003478319462833365,
      "loss": 2.9393,
      "theoretical_loss": 3.532499032374086,
      "tokens_seen": 1423179776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000347815396254965,
      "loss": 2.7559,
      "theoretical_loss": 3.5324845454105027,
      "tokens_seen": 1423245312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034779884622659354,
      "loss": 2.5628,
      "theoretical_loss": 3.5324700593007554,
      "tokens_seen": 1423310848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000347782296198222,
      "loss": 2.7811,
      "theoretical_loss": 3.5324555740447554,
      "tokens_seen": 1423376384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034776574616985056,
      "loss": 2.6471,
      "theoretical_loss": 3.532441089642413,
      "tokens_seen": 1423441920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003477491961414791,
      "loss": 2.6647,
      "theoretical_loss": 3.5324266060936376,
      "tokens_seen": 1423507456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003477326461131076,
      "loss": 2.4523,
      "theoretical_loss": 3.5324121233983403,
      "tokens_seen": 1423572992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003477160960847361,
      "loss": 2.7017,
      "theoretical_loss": 3.5323976415564315,
      "tokens_seen": 1423638528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034769954605636466,
      "loss": 2.3533,
      "theoretical_loss": 3.5323831605678215,
      "tokens_seen": 1423704064
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003476829960279932,
      "loss": 2.9888,
      "theoretical_loss": 3.5323686804324215,
      "tokens_seen": 1423769600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003476664459996217,
      "loss": 2.5682,
      "theoretical_loss": 3.5323542011501408,
      "tokens_seen": 1423835136
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003476498959712502,
      "loss": 2.6431,
      "theoretical_loss": 3.53233972272089,
      "tokens_seen": 1423900672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034763334594287875,
      "loss": 2.7018,
      "theoretical_loss": 3.5323252451445804,
      "tokens_seen": 1423966208
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003476167959145073,
      "loss": 2.7069,
      "theoretical_loss": 3.5323107684211217,
      "tokens_seen": 1424031744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034760024588613577,
      "loss": 2.3623,
      "theoretical_loss": 3.5322962925504253,
      "tokens_seen": 1424097280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003475836958577643,
      "loss": 2.6723,
      "theoretical_loss": 3.5322818175324007,
      "tokens_seen": 1424162816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034756714582939285,
      "loss": 2.5949,
      "theoretical_loss": 3.532267343366959,
      "tokens_seen": 1424228352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003475505958010214,
      "loss": 2.9061,
      "theoretical_loss": 3.532252870054011,
      "tokens_seen": 1424293888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034753404577264987,
      "loss": 2.8359,
      "theoretical_loss": 3.532238397593466,
      "tokens_seen": 1424359424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003475174957442784,
      "loss": 2.6387,
      "theoretical_loss": 3.532223925985237,
      "tokens_seen": 1424424960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034750094571590694,
      "loss": 2.5633,
      "theoretical_loss": 3.532209455229232,
      "tokens_seen": 1424490496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003474843956875355,
      "loss": 2.8442,
      "theoretical_loss": 3.532194985325363,
      "tokens_seen": 1424556032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034746784565916396,
      "loss": 2.5355,
      "theoretical_loss": 3.5321805162735407,
      "tokens_seen": 1424621568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003474512956307925,
      "loss": 2.8482,
      "theoretical_loss": 3.532166048073675,
      "tokens_seen": 1424687104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034743474560242104,
      "loss": 2.8901,
      "theoretical_loss": 3.532151580725677,
      "tokens_seen": 1424752640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003474181955740496,
      "loss": 2.2919,
      "theoretical_loss": 3.5321371142294575,
      "tokens_seen": 1424818176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034740164554567806,
      "loss": 2.6428,
      "theoretical_loss": 3.532122648584927,
      "tokens_seen": 1424883712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003473850955173066,
      "loss": 2.7142,
      "theoretical_loss": 3.532108183791996,
      "tokens_seen": 1424949248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034736854548893513,
      "loss": 2.8113,
      "theoretical_loss": 3.5320937198505753,
      "tokens_seen": 1425014784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034735199546056367,
      "loss": 2.6221,
      "theoretical_loss": 3.532079256760576,
      "tokens_seen": 1425080320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034733544543219216,
      "loss": 2.6926,
      "theoretical_loss": 3.532064794521908,
      "tokens_seen": 1425145856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003473188954038207,
      "loss": 2.8299,
      "theoretical_loss": 3.5320503331344826,
      "tokens_seen": 1425211392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034730234537544923,
      "loss": 2.6808,
      "theoretical_loss": 3.532035872598211,
      "tokens_seen": 1425276928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003472857953470777,
      "loss": 2.7069,
      "theoretical_loss": 3.5320214129130036,
      "tokens_seen": 1425342464
    },
    {
      "epoch": 0.51,
      "objective/train/avg_token_score": 0.008538750931620598,
      "objective/train/avg_weight": 0.9899464845657349,
      "objective/train/docs_used": 808636,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8855602741241455,
      "objective/train/std_weight": 0.05488242954015732,
      "objective/train/theoretical_loss": 3.5320069540787706,
      "objective/train/tokens_used": 1445868000,
      "theoretical_loss": 3.5320069540787706,
      "tokens_seen": 1425408000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034726924531870625,
      "loss": 2.6912,
      "theoretical_loss": 3.5320069540787706,
      "tokens_seen": 1425408000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003472526952903348,
      "loss": 2.6086,
      "theoretical_loss": 3.531992496095423,
      "tokens_seen": 1425473536
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034723614526196327,
      "loss": 2.6647,
      "theoretical_loss": 3.5319780389628725,
      "tokens_seen": 1425539072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003472195952335918,
      "loss": 2.5707,
      "theoretical_loss": 3.531963582681029,
      "tokens_seen": 1425604608
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034720304520522035,
      "loss": 2.7882,
      "theoretical_loss": 3.531949127249804,
      "tokens_seen": 1425670144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003471864951768489,
      "loss": 2.5906,
      "theoretical_loss": 3.5319346726691077,
      "tokens_seen": 1425735680
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034716994514847737,
      "loss": 2.8015,
      "theoretical_loss": 3.5319202189388514,
      "tokens_seen": 1425801216
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003471533951201059,
      "loss": 2.6885,
      "theoretical_loss": 3.5319057660589452,
      "tokens_seen": 1425866752
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034713684509173444,
      "loss": 2.4652,
      "theoretical_loss": 3.531891314029301,
      "tokens_seen": 1425932288
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000347120295063363,
      "loss": 2.7042,
      "theoretical_loss": 3.53187686284983,
      "tokens_seen": 1425997824
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034710374503499146,
      "loss": 2.3521,
      "theoretical_loss": 3.531862412520442,
      "tokens_seen": 1426063360
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034708719500662,
      "loss": 2.519,
      "theoretical_loss": 3.531847963041048,
      "tokens_seen": 1426128896
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034707064497824854,
      "loss": 2.6562,
      "theoretical_loss": 3.53183351441156,
      "tokens_seen": 1426194432
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003470540949498771,
      "loss": 2.542,
      "theoretical_loss": 3.531819066631887,
      "tokens_seen": 1426259968
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034703754492150556,
      "loss": 2.6349,
      "theoretical_loss": 3.5318046197019424,
      "tokens_seen": 1426325504
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003470209948931341,
      "loss": 2.7311,
      "theoretical_loss": 3.531790173621636,
      "tokens_seen": 1426391040
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034700444486476264,
      "loss": 2.6035,
      "theoretical_loss": 3.531775728390879,
      "tokens_seen": 1426456576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003469878948363912,
      "loss": 2.7176,
      "theoretical_loss": 3.5317612840095816,
      "tokens_seen": 1426522112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034697134480801966,
      "loss": 2.6532,
      "theoretical_loss": 3.5317468404776555,
      "tokens_seen": 1426587648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003469547947796482,
      "loss": 2.6208,
      "theoretical_loss": 3.531732397795012,
      "tokens_seen": 1426653184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034693824475127673,
      "loss": 2.5175,
      "theoretical_loss": 3.531717955961562,
      "tokens_seen": 1426718720
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034692169472290527,
      "loss": 2.4852,
      "theoretical_loss": 3.531703514977216,
      "tokens_seen": 1426784256
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034690514469453375,
      "loss": 1.9283,
      "theoretical_loss": 3.5316890748418857,
      "tokens_seen": 1426849792
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003468885946661623,
      "loss": 2.5876,
      "theoretical_loss": 3.531674635555482,
      "tokens_seen": 1426915328
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034687204463779083,
      "loss": 2.6449,
      "theoretical_loss": 3.5316601971179162,
      "tokens_seen": 1426980864
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034685549460941937,
      "loss": 2.7179,
      "theoretical_loss": 3.531645759529099,
      "tokens_seen": 1427046400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034683894458104785,
      "loss": 2.5341,
      "theoretical_loss": 3.5316313227889413,
      "tokens_seen": 1427111936
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003468223945526764,
      "loss": 2.7745,
      "theoretical_loss": 3.5316168868973556,
      "tokens_seen": 1427177472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003468058445243049,
      "loss": 2.8242,
      "theoretical_loss": 3.5316024518542517,
      "tokens_seen": 1427243008
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003467892944959334,
      "loss": 2.6712,
      "theoretical_loss": 3.5315880176595416,
      "tokens_seen": 1427308544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034677274446756194,
      "loss": 2.8015,
      "theoretical_loss": 3.5315735843131355,
      "tokens_seen": 1427374080
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034675619443919043,
      "loss": 2.5341,
      "theoretical_loss": 3.5315591518149456,
      "tokens_seen": 1427439616
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034673964441081897,
      "loss": 2.8123,
      "theoretical_loss": 3.5315447201648826,
      "tokens_seen": 1427505152
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003467230943824475,
      "loss": 2.7272,
      "theoretical_loss": 3.531530289362858,
      "tokens_seen": 1427570688
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034670654435407604,
      "loss": 2.5099,
      "theoretical_loss": 3.531515859408783,
      "tokens_seen": 1427636224
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003466899943257045,
      "loss": 2.6292,
      "theoretical_loss": 3.5315014303025682,
      "tokens_seen": 1427701760
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034667344429733306,
      "loss": 2.4539,
      "theoretical_loss": 3.531487002044126,
      "tokens_seen": 1427767296
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003466568942689616,
      "loss": 2.6465,
      "theoretical_loss": 3.5314725746333666,
      "tokens_seen": 1427832832
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034664034424059014,
      "loss": 2.677,
      "theoretical_loss": 3.531458148070202,
      "tokens_seen": 1427898368
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003466237942122186,
      "loss": 2.7966,
      "theoretical_loss": 3.5314437223545436,
      "tokens_seen": 1427963904
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034660724418384716,
      "loss": 2.7497,
      "theoretical_loss": 3.5314292974863024,
      "tokens_seen": 1428029440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003465906941554757,
      "loss": 2.5776,
      "theoretical_loss": 3.5314148734653896,
      "tokens_seen": 1428094976
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034657414412710423,
      "loss": 2.7269,
      "theoretical_loss": 3.5314004502917165,
      "tokens_seen": 1428160512
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003465575940987327,
      "loss": 2.5482,
      "theoretical_loss": 3.531386027965195,
      "tokens_seen": 1428226048
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034654104407036125,
      "loss": 2.8473,
      "theoretical_loss": 3.5313716064857354,
      "tokens_seen": 1428291584
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003465244940419898,
      "loss": 2.7796,
      "theoretical_loss": 3.531357185853251,
      "tokens_seen": 1428357120
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034650794401361833,
      "loss": 2.6287,
      "theoretical_loss": 3.5313427660676506,
      "tokens_seen": 1428422656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003464913939852468,
      "loss": 2.7447,
      "theoretical_loss": 3.5313283471288477,
      "tokens_seen": 1428488192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034647484395687535,
      "loss": 2.6892,
      "theoretical_loss": 3.5313139290367532,
      "tokens_seen": 1428553728
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003464582939285039,
      "loss": 2.8758,
      "theoretical_loss": 3.5312995117912784,
      "tokens_seen": 1428619264
    },
    {
      "epoch": 0.51,
      "objective/train/avg_token_score": 0.011311979033052921,
      "objective/train/avg_weight": 0.9866854548454285,
      "objective/train/docs_used": 810397,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.655975341796875,
      "objective/train/std_weight": 0.05427398160099983,
      "objective/train/theoretical_loss": 3.531285095392335,
      "objective/train/tokens_used": 1449144800,
      "theoretical_loss": 3.531285095392335,
      "tokens_seen": 1428684800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003464417439001324,
      "loss": 2.554,
      "theoretical_loss": 3.531285095392335,
      "tokens_seen": 1428684800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003464251938717609,
      "loss": 2.5632,
      "theoretical_loss": 3.531270679839834,
      "tokens_seen": 1428750336
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034640864384338945,
      "loss": 2.8991,
      "theoretical_loss": 3.5312562651336865,
      "tokens_seen": 1428815872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000346392093815018,
      "loss": 2.5728,
      "theoretical_loss": 3.5312418512738053,
      "tokens_seen": 1428881408
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003463755437866465,
      "loss": 2.6938,
      "theoretical_loss": 3.5312274382601005,
      "tokens_seen": 1428946944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000346358993758275,
      "loss": 2.6903,
      "theoretical_loss": 3.5312130260924848,
      "tokens_seen": 1429012480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034634244372990354,
      "loss": 2.6796,
      "theoretical_loss": 3.5311986147708696,
      "tokens_seen": 1429078016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003463258937015321,
      "loss": 2.789,
      "theoretical_loss": 3.5311842042951653,
      "tokens_seen": 1429143552
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034630934367316056,
      "loss": 2.492,
      "theoretical_loss": 3.531169794665285,
      "tokens_seen": 1429209088
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003462927936447891,
      "loss": 2.7874,
      "theoretical_loss": 3.531155385881139,
      "tokens_seen": 1429274624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003462762436164176,
      "loss": 2.5801,
      "theoretical_loss": 3.5311409779426395,
      "tokens_seen": 1429340160
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003462596935880461,
      "loss": 2.3638,
      "theoretical_loss": 3.531126570849698,
      "tokens_seen": 1429405696
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034624314355967466,
      "loss": 2.6021,
      "theoretical_loss": 3.5311121646022263,
      "tokens_seen": 1429471232
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003462265935313032,
      "loss": 2.4529,
      "theoretical_loss": 3.5310977592001356,
      "tokens_seen": 1429536768
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003462100435029317,
      "loss": 2.7721,
      "theoretical_loss": 3.531083354643338,
      "tokens_seen": 1429602304
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003461934934745602,
      "loss": 2.5812,
      "theoretical_loss": 3.5310689509317448,
      "tokens_seen": 1429667840
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034617694344618875,
      "loss": 2.7717,
      "theoretical_loss": 3.531054548065268,
      "tokens_seen": 1429733376
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003461603934178173,
      "loss": 2.649,
      "theoretical_loss": 3.5310401460438188,
      "tokens_seen": 1429798912
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003461438433894458,
      "loss": 2.5232,
      "theoretical_loss": 3.5310257448673092,
      "tokens_seen": 1429864448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003461272933610743,
      "loss": 2.781,
      "theoretical_loss": 3.531011344535651,
      "tokens_seen": 1429929984
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034611074333270285,
      "loss": 2.7065,
      "theoretical_loss": 3.5309969450487557,
      "tokens_seen": 1429995520
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003460941933043314,
      "loss": 2.8618,
      "theoretical_loss": 3.5309825464065354,
      "tokens_seen": 1430061056
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034607764327595987,
      "loss": 2.609,
      "theoretical_loss": 3.5309681486089013,
      "tokens_seen": 1430126592
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003460610932475884,
      "loss": 2.5727,
      "theoretical_loss": 3.5309537516557654,
      "tokens_seen": 1430192128
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034604454321921695,
      "loss": 2.7469,
      "theoretical_loss": 3.53093935554704,
      "tokens_seen": 1430257664
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003460279931908455,
      "loss": 2.8557,
      "theoretical_loss": 3.530924960282636,
      "tokens_seen": 1430323200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034601144316247397,
      "loss": 2.7151,
      "theoretical_loss": 3.5309105658624658,
      "tokens_seen": 1430388736
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003459948931341025,
      "loss": 2.6284,
      "theoretical_loss": 3.5308961722864405,
      "tokens_seen": 1430454272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034597834310573104,
      "loss": 2.6445,
      "theoretical_loss": 3.530881779554473,
      "tokens_seen": 1430519808
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003459617930773596,
      "loss": 2.7971,
      "theoretical_loss": 3.5308673876664742,
      "tokens_seen": 1430585344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003459452430489881,
      "loss": 2.6411,
      "theoretical_loss": 3.5308529966223565,
      "tokens_seen": 1430650880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003459286930206166,
      "loss": 2.3606,
      "theoretical_loss": 3.5308386064220314,
      "tokens_seen": 1430716416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034591214299224514,
      "loss": 2.665,
      "theoretical_loss": 3.530824217065411,
      "tokens_seen": 1430781952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003458955929638737,
      "loss": 2.5981,
      "theoretical_loss": 3.530809828552407,
      "tokens_seen": 1430847488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003458790429355022,
      "loss": 2.9597,
      "theoretical_loss": 3.530795440882932,
      "tokens_seen": 1430913024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003458624929071307,
      "loss": 2.695,
      "theoretical_loss": 3.530781054056897,
      "tokens_seen": 1430978560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034584594287875923,
      "loss": 2.6324,
      "theoretical_loss": 3.5307666680742145,
      "tokens_seen": 1431044096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034582939285038777,
      "loss": 2.5768,
      "theoretical_loss": 3.530752282934796,
      "tokens_seen": 1431109632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034581284282201625,
      "loss": 2.7178,
      "theoretical_loss": 3.530737898638554,
      "tokens_seen": 1431175168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003457962927936448,
      "loss": 2.5933,
      "theoretical_loss": 3.5307235151854,
      "tokens_seen": 1431240704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003457797427652733,
      "loss": 2.6701,
      "theoretical_loss": 3.5307091325752458,
      "tokens_seen": 1431306240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003457631927369018,
      "loss": 2.5593,
      "theoretical_loss": 3.5306947508080047,
      "tokens_seen": 1431371776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034574664270853035,
      "loss": 2.4606,
      "theoretical_loss": 3.530680369883587,
      "tokens_seen": 1431437312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003457300926801589,
      "loss": 2.5561,
      "theoretical_loss": 3.5306659898019053,
      "tokens_seen": 1431502848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034571354265178737,
      "loss": 2.5217,
      "theoretical_loss": 3.530651610562872,
      "tokens_seen": 1431568384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003456969926234159,
      "loss": 2.7732,
      "theoretical_loss": 3.5306372321663995,
      "tokens_seen": 1431633920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034568044259504445,
      "loss": 2.6018,
      "theoretical_loss": 3.530622854612399,
      "tokens_seen": 1431699456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000345663892566673,
      "loss": 2.8116,
      "theoretical_loss": 3.530608477900783,
      "tokens_seen": 1431764992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034564734253830147,
      "loss": 2.5518,
      "theoretical_loss": 3.5305941020314635,
      "tokens_seen": 1431830528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034563079250993,
      "loss": 2.6999,
      "theoretical_loss": 3.5305797270043526,
      "tokens_seen": 1431896064
    },
    {
      "epoch": 0.51,
      "objective/train/avg_token_score": 0.016850709915161133,
      "objective/train/avg_weight": 0.9801775813102722,
      "objective/train/docs_used": 812373,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.462169647216797,
      "objective/train/std_weight": 0.0835515707731247,
      "objective/train/theoretical_loss": 3.5305653528193623,
      "objective/train/tokens_used": 1452421600,
      "theoretical_loss": 3.5305653528193623,
      "tokens_seen": 1431961600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034561424248155854,
      "loss": 2.9482,
      "theoretical_loss": 3.5305653528193623,
      "tokens_seen": 1431961600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003455976924531871,
      "loss": 2.7396,
      "theoretical_loss": 3.530550979476405,
      "tokens_seen": 1432027136
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034558114242481556,
      "loss": 2.6985,
      "theoretical_loss": 3.5305366069753927,
      "tokens_seen": 1432092672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003455645923964441,
      "loss": 2.6021,
      "theoretical_loss": 3.530522235316237,
      "tokens_seen": 1432158208
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034554804236807264,
      "loss": 2.7708,
      "theoretical_loss": 3.5305078644988512,
      "tokens_seen": 1432223744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003455314923397012,
      "loss": 2.8574,
      "theoretical_loss": 3.530493494523147,
      "tokens_seen": 1432289280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034551494231132966,
      "loss": 2.8214,
      "theoretical_loss": 3.5304791253890366,
      "tokens_seen": 1432354816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003454983922829582,
      "loss": 2.7137,
      "theoretical_loss": 3.530464757096432,
      "tokens_seen": 1432420352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034548184225458673,
      "loss": 2.64,
      "theoretical_loss": 3.530450389645245,
      "tokens_seen": 1432485888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034546529222621527,
      "loss": 2.445,
      "theoretical_loss": 3.5304360230353886,
      "tokens_seen": 1432551424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034544874219784376,
      "loss": 2.5923,
      "theoretical_loss": 3.530421657266775,
      "tokens_seen": 1432616960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003454321921694723,
      "loss": 2.7399,
      "theoretical_loss": 3.5304072923393166,
      "tokens_seen": 1432682496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034541564214110083,
      "loss": 2.6669,
      "theoretical_loss": 3.530392928252925,
      "tokens_seen": 1432748032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034539909211272937,
      "loss": 2.8102,
      "theoretical_loss": 3.5303785650075126,
      "tokens_seen": 1432813568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034538254208435785,
      "loss": 2.5105,
      "theoretical_loss": 3.5303642026029918,
      "tokens_seen": 1432879104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003453659920559864,
      "loss": 2.7736,
      "theoretical_loss": 3.5303498410392757,
      "tokens_seen": 1432944640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003453494420276149,
      "loss": 2.4177,
      "theoretical_loss": 3.5303354803162756,
      "tokens_seen": 1433010176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034533289199924346,
      "loss": 2.4275,
      "theoretical_loss": 3.5303211204339044,
      "tokens_seen": 1433075712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034531634197087195,
      "loss": 2.5557,
      "theoretical_loss": 3.530306761392074,
      "tokens_seen": 1433141248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034529979194250043,
      "loss": 2.5196,
      "theoretical_loss": 3.530292403190697,
      "tokens_seen": 1433206784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034528324191412897,
      "loss": 2.6658,
      "theoretical_loss": 3.530278045829686,
      "tokens_seen": 1433272320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003452666918857575,
      "loss": 2.6896,
      "theoretical_loss": 3.530263689308953,
      "tokens_seen": 1433337856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034525014185738604,
      "loss": 2.8632,
      "theoretical_loss": 3.5302493336284106,
      "tokens_seen": 1433403392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003452335918290145,
      "loss": 2.6147,
      "theoretical_loss": 3.5302349787879717,
      "tokens_seen": 1433468928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034521704180064306,
      "loss": 2.5481,
      "theoretical_loss": 3.530220624787548,
      "tokens_seen": 1433534464
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003452004917722716,
      "loss": 2.5809,
      "theoretical_loss": 3.530206271627052,
      "tokens_seen": 1433600000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034518394174390014,
      "loss": 2.594,
      "theoretical_loss": 3.5301919193063966,
      "tokens_seen": 1433665536
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003451673917155286,
      "loss": 2.6238,
      "theoretical_loss": 3.5301775678254943,
      "tokens_seen": 1433731072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034515084168715716,
      "loss": 2.5689,
      "theoretical_loss": 3.5301632171842567,
      "tokens_seen": 1433796608
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003451342916587857,
      "loss": 2.5389,
      "theoretical_loss": 3.5301488673825974,
      "tokens_seen": 1433862144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034511774163041424,
      "loss": 2.6336,
      "theoretical_loss": 3.5301345184204282,
      "tokens_seen": 1433927680
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003451011916020427,
      "loss": 2.67,
      "theoretical_loss": 3.5301201702976615,
      "tokens_seen": 1433993216
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034508464157367126,
      "loss": 2.3722,
      "theoretical_loss": 3.530105823014211,
      "tokens_seen": 1434058752
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003450680915452998,
      "loss": 2.581,
      "theoretical_loss": 3.5300914765699876,
      "tokens_seen": 1434124288
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034505154151692833,
      "loss": 2.7555,
      "theoretical_loss": 3.5300771309649055,
      "tokens_seen": 1434189824
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003450349914885568,
      "loss": 2.5465,
      "theoretical_loss": 3.530062786198876,
      "tokens_seen": 1434255360
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034501844146018535,
      "loss": 2.3543,
      "theoretical_loss": 3.530048442271812,
      "tokens_seen": 1434320896
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003450018914318139,
      "loss": 2.607,
      "theoretical_loss": 3.5300340991836268,
      "tokens_seen": 1434386432
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003449853414034424,
      "loss": 2.2153,
      "theoretical_loss": 3.530019756934232,
      "tokens_seen": 1434451968
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003449687913750709,
      "loss": 2.7181,
      "theoretical_loss": 3.530005415523541,
      "tokens_seen": 1434517504
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034495224134669945,
      "loss": 2.5271,
      "theoretical_loss": 3.5299910749514662,
      "tokens_seen": 1434583040
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000344935691318328,
      "loss": 2.1028,
      "theoretical_loss": 3.52997673521792,
      "tokens_seen": 1434648576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003449191412899565,
      "loss": 2.8509,
      "theoretical_loss": 3.529962396322815,
      "tokens_seen": 1434714112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000344902591261585,
      "loss": 2.6747,
      "theoretical_loss": 3.529948058266065,
      "tokens_seen": 1434779648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034488604123321354,
      "loss": 2.7598,
      "theoretical_loss": 3.529933721047581,
      "tokens_seen": 1434845184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003448694912048421,
      "loss": 2.5892,
      "theoretical_loss": 3.529919384667277,
      "tokens_seen": 1434910720
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003448529411764706,
      "loss": 2.6765,
      "theoretical_loss": 3.529905049125065,
      "tokens_seen": 1434976256
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003448363911480991,
      "loss": 2.4783,
      "theoretical_loss": 3.5298907144208584,
      "tokens_seen": 1435041792
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003448198411197276,
      "loss": 2.8547,
      "theoretical_loss": 3.529876380554569,
      "tokens_seen": 1435107328
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003448032910913561,
      "loss": 2.787,
      "theoretical_loss": 3.5298620475261107,
      "tokens_seen": 1435172864
    },
    {
      "epoch": 0.51,
      "objective/train/avg_token_score": 0.008029872551560402,
      "objective/train/avg_weight": 0.9905571341514587,
      "objective/train/docs_used": 813976,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.621913433074951,
      "objective/train/std_weight": 0.03252668306231499,
      "objective/train/theoretical_loss": 3.5298477153353955,
      "objective/train/tokens_used": 1455698400,
      "theoretical_loss": 3.5298477153353955,
      "tokens_seen": 1435238400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034478674106298466,
      "loss": 2.5298,
      "theoretical_loss": 3.5298477153353955,
      "tokens_seen": 1435238400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003447701910346132,
      "loss": 2.8865,
      "theoretical_loss": 3.5298333839823366,
      "tokens_seen": 1435303936
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034475364100624174,
      "loss": 2.4507,
      "theoretical_loss": 3.529819053466846,
      "tokens_seen": 1435369472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003447370909778702,
      "loss": 2.5612,
      "theoretical_loss": 3.5298047237888373,
      "tokens_seen": 1435435008
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034472054094949876,
      "loss": 2.8217,
      "theoretical_loss": 3.529790394948223,
      "tokens_seen": 1435500544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003447039909211273,
      "loss": 2.732,
      "theoretical_loss": 3.5297760669449163,
      "tokens_seen": 1435566080
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034468744089275583,
      "loss": 2.4404,
      "theoretical_loss": 3.52976173977883,
      "tokens_seen": 1435631616
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003446708908643843,
      "loss": 2.5153,
      "theoretical_loss": 3.5297474134498765,
      "tokens_seen": 1435697152
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034465434083601285,
      "loss": 2.5795,
      "theoretical_loss": 3.529733087957969,
      "tokens_seen": 1435762688
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003446377908076414,
      "loss": 2.1094,
      "theoretical_loss": 3.5297187633030203,
      "tokens_seen": 1435828224
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034462124077926993,
      "loss": 2.4229,
      "theoretical_loss": 3.5297044394849433,
      "tokens_seen": 1435893760
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003446046907508984,
      "loss": 2.5155,
      "theoretical_loss": 3.5296901165036507,
      "tokens_seen": 1435959296
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034458814072252695,
      "loss": 2.4933,
      "theoretical_loss": 3.529675794359056,
      "tokens_seen": 1436024832
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003445715906941555,
      "loss": 2.4879,
      "theoretical_loss": 3.529661473051072,
      "tokens_seen": 1436090368
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000344555040665784,
      "loss": 2.6613,
      "theoretical_loss": 3.529647152579611,
      "tokens_seen": 1436155904
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003445384906374125,
      "loss": 2.3736,
      "theoretical_loss": 3.5296328329445865,
      "tokens_seen": 1436221440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034452194060904104,
      "loss": 2.6163,
      "theoretical_loss": 3.5296185141459118,
      "tokens_seen": 1436286976
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003445053905806696,
      "loss": 2.3434,
      "theoretical_loss": 3.529604196183499,
      "tokens_seen": 1436352512
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003444888405522981,
      "loss": 2.6435,
      "theoretical_loss": 3.529589879057262,
      "tokens_seen": 1436418048
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003444722905239266,
      "loss": 2.6089,
      "theoretical_loss": 3.5295755627671133,
      "tokens_seen": 1436483584
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034445574049555514,
      "loss": 2.5089,
      "theoretical_loss": 3.5295612473129663,
      "tokens_seen": 1436549120
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003444391904671837,
      "loss": 2.6669,
      "theoretical_loss": 3.5295469326947337,
      "tokens_seen": 1436614656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003444226404388122,
      "loss": 2.8442,
      "theoretical_loss": 3.529532618912328,
      "tokens_seen": 1436680192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003444060904104407,
      "loss": 2.764,
      "theoretical_loss": 3.529518305965664,
      "tokens_seen": 1436745728
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034438954038206924,
      "loss": 2.706,
      "theoretical_loss": 3.5295039938546537,
      "tokens_seen": 1436811264
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003443729903536978,
      "loss": 2.5996,
      "theoretical_loss": 3.52948968257921,
      "tokens_seen": 1436876800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003443564403253263,
      "loss": 2.6382,
      "theoretical_loss": 3.5294753721392453,
      "tokens_seen": 1436942336
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003443398902969548,
      "loss": 2.6801,
      "theoretical_loss": 3.5294610625346747,
      "tokens_seen": 1437007872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003443233402685833,
      "loss": 2.6777,
      "theoretical_loss": 3.52944675376541,
      "tokens_seen": 1437073408
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003443067902402118,
      "loss": 2.5939,
      "theoretical_loss": 3.5294324458313646,
      "tokens_seen": 1437138944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034429024021184035,
      "loss": 2.4317,
      "theoretical_loss": 3.529418138732452,
      "tokens_seen": 1437204480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003442736901834689,
      "loss": 2.5974,
      "theoretical_loss": 3.529403832468585,
      "tokens_seen": 1437270016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003442571401550974,
      "loss": 2.7076,
      "theoretical_loss": 3.529389527039677,
      "tokens_seen": 1437335552
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003442405901267259,
      "loss": 2.4601,
      "theoretical_loss": 3.529375222445641,
      "tokens_seen": 1437401088
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034422404009835445,
      "loss": 2.6844,
      "theoretical_loss": 3.5293609186863906,
      "tokens_seen": 1437466624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000344207490069983,
      "loss": 2.6754,
      "theoretical_loss": 3.5293466157618383,
      "tokens_seen": 1437532160
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034419094004161147,
      "loss": 2.3306,
      "theoretical_loss": 3.5293323136718975,
      "tokens_seen": 1437597696
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034417439001324,
      "loss": 2.468,
      "theoretical_loss": 3.529318012416482,
      "tokens_seen": 1437663232
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034415783998486855,
      "loss": 2.3711,
      "theoretical_loss": 3.529303711995505,
      "tokens_seen": 1437728768
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003441412899564971,
      "loss": 2.5906,
      "theoretical_loss": 3.52928941240888,
      "tokens_seen": 1437794304
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034412473992812557,
      "loss": 2.5913,
      "theoretical_loss": 3.529275113656519,
      "tokens_seen": 1437859840
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003441081898997541,
      "loss": 2.6514,
      "theoretical_loss": 3.5292608157383363,
      "tokens_seen": 1437925376
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034409163987138264,
      "loss": 2.7778,
      "theoretical_loss": 3.5292465186542454,
      "tokens_seen": 1437990912
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003440750898430112,
      "loss": 2.7924,
      "theoretical_loss": 3.529232222404159,
      "tokens_seen": 1438056448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034405853981463966,
      "loss": 2.6047,
      "theoretical_loss": 3.5292179269879913,
      "tokens_seen": 1438121984
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003440419897862682,
      "loss": 2.7216,
      "theoretical_loss": 3.529203632405655,
      "tokens_seen": 1438187520
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034402543975789674,
      "loss": 2.4956,
      "theoretical_loss": 3.529189338657063,
      "tokens_seen": 1438253056
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003440088897295253,
      "loss": 2.545,
      "theoretical_loss": 3.5291750457421296,
      "tokens_seen": 1438318592
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034399233970115376,
      "loss": 2.8741,
      "theoretical_loss": 3.5291607536607685,
      "tokens_seen": 1438384128
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003439757896727823,
      "loss": 2.537,
      "theoretical_loss": 3.529146462412892,
      "tokens_seen": 1438449664
    },
    {
      "epoch": 0.51,
      "objective/train/avg_token_score": 0.014263682067394257,
      "objective/train/avg_weight": 0.9832130074501038,
      "objective/train/docs_used": 815908,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6340155601501465,
      "objective/train/std_weight": 0.07831459492444992,
      "objective/train/theoretical_loss": 3.5291321719984134,
      "objective/train/tokens_used": 1458975200,
      "theoretical_loss": 3.5291321719984134,
      "tokens_seen": 1438515200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034395923964441083,
      "loss": 2.4247,
      "theoretical_loss": 3.5291321719984134,
      "tokens_seen": 1438515200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034394268961603937,
      "loss": 2.5709,
      "theoretical_loss": 3.5291178824172476,
      "tokens_seen": 1438580736
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034392613958766785,
      "loss": 2.8391,
      "theoretical_loss": 3.5291035936693067,
      "tokens_seen": 1438646272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003439095895592964,
      "loss": 2.6488,
      "theoretical_loss": 3.529089305754505,
      "tokens_seen": 1438711808
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034389303953092493,
      "loss": 2.5292,
      "theoretical_loss": 3.5290750186727555,
      "tokens_seen": 1438777344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034387648950255347,
      "loss": 2.4925,
      "theoretical_loss": 3.529060732423972,
      "tokens_seen": 1438842880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034385993947418195,
      "loss": 2.5061,
      "theoretical_loss": 3.5290464470080676,
      "tokens_seen": 1438908416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034384338944581043,
      "loss": 2.5854,
      "theoretical_loss": 3.529032162424956,
      "tokens_seen": 1438973952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034382683941743897,
      "loss": 2.6148,
      "theoretical_loss": 3.529017878674551,
      "tokens_seen": 1439039488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003438102893890675,
      "loss": 2.502,
      "theoretical_loss": 3.5290035957567656,
      "tokens_seen": 1439105024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034379373936069605,
      "loss": 2.8147,
      "theoretical_loss": 3.528989313671514,
      "tokens_seen": 1439170560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034377718933232453,
      "loss": 2.8305,
      "theoretical_loss": 3.5289750324187086,
      "tokens_seen": 1439236096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034376063930395307,
      "loss": 2.5183,
      "theoretical_loss": 3.5289607519982646,
      "tokens_seen": 1439301632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003437440892755816,
      "loss": 2.537,
      "theoretical_loss": 3.5289464724100945,
      "tokens_seen": 1439367168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034372753924721014,
      "loss": 2.7903,
      "theoretical_loss": 3.5289321936541125,
      "tokens_seen": 1439432704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003437109892188386,
      "loss": 2.5938,
      "theoretical_loss": 3.5289179157302315,
      "tokens_seen": 1439498240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034369443919046716,
      "loss": 2.2778,
      "theoretical_loss": 3.528903638638366,
      "tokens_seen": 1439563776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003436778891620957,
      "loss": 2.6819,
      "theoretical_loss": 3.5288893623784285,
      "tokens_seen": 1439629312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034366133913372424,
      "loss": 2.4988,
      "theoretical_loss": 3.5288750869503334,
      "tokens_seen": 1439694848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003436447891053527,
      "loss": 2.7521,
      "theoretical_loss": 3.5288608123539946,
      "tokens_seen": 1439760384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034362823907698126,
      "loss": 2.4108,
      "theoretical_loss": 3.5288465385893257,
      "tokens_seen": 1439825920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003436116890486098,
      "loss": 2.5783,
      "theoretical_loss": 3.5288322656562396,
      "tokens_seen": 1439891456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034359513902023833,
      "loss": 2.4607,
      "theoretical_loss": 3.528817993554651,
      "tokens_seen": 1439956992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003435785889918668,
      "loss": 2.1414,
      "theoretical_loss": 3.528803722284473,
      "tokens_seen": 1440022528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034356203896349535,
      "loss": 2.5439,
      "theoretical_loss": 3.5287894518456193,
      "tokens_seen": 1440088064
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003435454889351239,
      "loss": 2.6827,
      "theoretical_loss": 3.5287751822380047,
      "tokens_seen": 1440153600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034352893890675243,
      "loss": 2.6122,
      "theoretical_loss": 3.528760913461541,
      "tokens_seen": 1440219136
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003435123888783809,
      "loss": 2.7808,
      "theoretical_loss": 3.528746645516144,
      "tokens_seen": 1440284672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034349583885000945,
      "loss": 2.4865,
      "theoretical_loss": 3.5287323784017266,
      "tokens_seen": 1440350208
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000343479288821638,
      "loss": 2.5586,
      "theoretical_loss": 3.528718112118202,
      "tokens_seen": 1440415744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003434627387932665,
      "loss": 2.7319,
      "theoretical_loss": 3.5287038466654845,
      "tokens_seen": 1440481280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000343446188764895,
      "loss": 2.8231,
      "theoretical_loss": 3.5286895820434885,
      "tokens_seen": 1440546816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034342963873652355,
      "loss": 2.9414,
      "theoretical_loss": 3.5286753182521275,
      "tokens_seen": 1440612352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003434130887081521,
      "loss": 2.5616,
      "theoretical_loss": 3.5286610552913147,
      "tokens_seen": 1440677888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003433965386797806,
      "loss": 2.4806,
      "theoretical_loss": 3.5286467931609646,
      "tokens_seen": 1440743424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003433799886514091,
      "loss": 2.6879,
      "theoretical_loss": 3.528632531860991,
      "tokens_seen": 1440808960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034336343862303764,
      "loss": 2.505,
      "theoretical_loss": 3.5286182713913075,
      "tokens_seen": 1440874496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003433468885946661,
      "loss": 2.423,
      "theoretical_loss": 3.5286040117518285,
      "tokens_seen": 1440940032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034333033856629466,
      "loss": 2.635,
      "theoretical_loss": 3.5285897529424677,
      "tokens_seen": 1441005568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003433137885379232,
      "loss": 2.3522,
      "theoretical_loss": 3.5285754949631385,
      "tokens_seen": 1441071104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034329723850955174,
      "loss": 2.4622,
      "theoretical_loss": 3.5285612378137556,
      "tokens_seen": 1441136640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003432806884811802,
      "loss": 2.638,
      "theoretical_loss": 3.528546981494233,
      "tokens_seen": 1441202176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034326413845280876,
      "loss": 2.7874,
      "theoretical_loss": 3.5285327260044834,
      "tokens_seen": 1441267712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003432475884244373,
      "loss": 2.7526,
      "theoretical_loss": 3.528518471344422,
      "tokens_seen": 1441333248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034323103839606583,
      "loss": 2.7335,
      "theoretical_loss": 3.528504217513963,
      "tokens_seen": 1441398784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003432144883676943,
      "loss": 2.8534,
      "theoretical_loss": 3.528489964513019,
      "tokens_seen": 1441464320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034319793833932286,
      "loss": 2.4761,
      "theoretical_loss": 3.528475712341506,
      "tokens_seen": 1441529856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003431813883109514,
      "loss": 2.6085,
      "theoretical_loss": 3.528461460999336,
      "tokens_seen": 1441595392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034316483828257993,
      "loss": 2.6684,
      "theoretical_loss": 3.5284472104864246,
      "tokens_seen": 1441660928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003431482882542084,
      "loss": 2.5604,
      "theoretical_loss": 3.528432960802685,
      "tokens_seen": 1441726464
    },
    {
      "epoch": 0.51,
      "objective/train/avg_token_score": 0.02621074579656124,
      "objective/train/avg_weight": 0.9691607356071472,
      "objective/train/docs_used": 817781,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7484827041625977,
      "objective/train/std_weight": 0.11313387751579285,
      "objective/train/theoretical_loss": 3.5284187119480315,
      "objective/train/tokens_used": 1462252000,
      "theoretical_loss": 3.5284187119480315,
      "tokens_seen": 1441792000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00034313173822583695,
      "loss": 2.8554,
      "theoretical_loss": 3.5284187119480315,
      "tokens_seen": 1441792000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0003431151881974655,
      "loss": 2.5568,
      "theoretical_loss": 3.528404463922378,
      "tokens_seen": 1441857536
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000343098638169094,
      "loss": 2.7574,
      "theoretical_loss": 3.5283902167256382,
      "tokens_seen": 1441923072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003430820881407225,
      "loss": 2.5127,
      "theoretical_loss": 3.528375970357728,
      "tokens_seen": 1441988608
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034306553811235105,
      "loss": 2.4061,
      "theoretical_loss": 3.5283617248185593,
      "tokens_seen": 1442054144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003430489880839796,
      "loss": 2.5924,
      "theoretical_loss": 3.5283474801080477,
      "tokens_seen": 1442119680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003430324380556081,
      "loss": 2.5455,
      "theoretical_loss": 3.528333236226106,
      "tokens_seen": 1442185216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003430158880272366,
      "loss": 2.4766,
      "theoretical_loss": 3.5283189931726504,
      "tokens_seen": 1442250752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034299933799886514,
      "loss": 2.6214,
      "theoretical_loss": 3.5283047509475933,
      "tokens_seen": 1442316288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003429827879704937,
      "loss": 2.626,
      "theoretical_loss": 3.528290509550849,
      "tokens_seen": 1442381824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003429662379421222,
      "loss": 2.535,
      "theoretical_loss": 3.528276268982333,
      "tokens_seen": 1442447360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003429496879137507,
      "loss": 2.4618,
      "theoretical_loss": 3.528262029241958,
      "tokens_seen": 1442512896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034293313788537924,
      "loss": 2.5978,
      "theoretical_loss": 3.5282477903296394,
      "tokens_seen": 1442578432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003429165878570078,
      "loss": 2.5995,
      "theoretical_loss": 3.5282335522452906,
      "tokens_seen": 1442643968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003429000378286363,
      "loss": 2.4272,
      "theoretical_loss": 3.528219314988826,
      "tokens_seen": 1442709504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003428834878002648,
      "loss": 2.5064,
      "theoretical_loss": 3.52820507856016,
      "tokens_seen": 1442775040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003428669377718933,
      "loss": 2.7406,
      "theoretical_loss": 3.528190842959207,
      "tokens_seen": 1442840576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003428503877435218,
      "loss": 2.6219,
      "theoretical_loss": 3.5281766081858814,
      "tokens_seen": 1442906112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034283383771515036,
      "loss": 2.5546,
      "theoretical_loss": 3.5281623742400967,
      "tokens_seen": 1442971648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003428172876867789,
      "loss": 2.7258,
      "theoretical_loss": 3.5281481411217683,
      "tokens_seen": 1443037184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003428007376584074,
      "loss": 2.6458,
      "theoretical_loss": 3.52813390883081,
      "tokens_seen": 1443102720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003427841876300359,
      "loss": 2.501,
      "theoretical_loss": 3.528119677367136,
      "tokens_seen": 1443168256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034276763760166445,
      "loss": 2.6298,
      "theoretical_loss": 3.5281054467306605,
      "tokens_seen": 1443233792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000342751087573293,
      "loss": 2.4374,
      "theoretical_loss": 3.5280912169212986,
      "tokens_seen": 1443299328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003427345375449215,
      "loss": 2.5998,
      "theoretical_loss": 3.528076987938964,
      "tokens_seen": 1443364864
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034271798751655,
      "loss": 2.5017,
      "theoretical_loss": 3.5280627597835705,
      "tokens_seen": 1443430400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034270143748817855,
      "loss": 2.4613,
      "theoretical_loss": 3.528048532455034,
      "tokens_seen": 1443495936
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003426848874598071,
      "loss": 2.8247,
      "theoretical_loss": 3.5280343059532684,
      "tokens_seen": 1443561472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034266833743143557,
      "loss": 2.9367,
      "theoretical_loss": 3.528020080278188,
      "tokens_seen": 1443627008
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003426517874030641,
      "loss": 2.5009,
      "theoretical_loss": 3.528005855429707,
      "tokens_seen": 1443692544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034263523737469264,
      "loss": 2.726,
      "theoretical_loss": 3.5279916314077395,
      "tokens_seen": 1443758080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003426186873463212,
      "loss": 2.7046,
      "theoretical_loss": 3.527977408212201,
      "tokens_seen": 1443823616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034260213731794967,
      "loss": 2.482,
      "theoretical_loss": 3.527963185843005,
      "tokens_seen": 1443889152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003425855872895782,
      "loss": 2.5595,
      "theoretical_loss": 3.5279489643000668,
      "tokens_seen": 1443954688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034256903726120674,
      "loss": 2.5555,
      "theoretical_loss": 3.527934743583301,
      "tokens_seen": 1444020224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003425524872328353,
      "loss": 2.4764,
      "theoretical_loss": 3.5279205236926208,
      "tokens_seen": 1444085760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034253593720446376,
      "loss": 2.3455,
      "theoretical_loss": 3.5279063046279417,
      "tokens_seen": 1444151296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003425193871760923,
      "loss": 2.8721,
      "theoretical_loss": 3.5278920863891785,
      "tokens_seen": 1444216832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034250283714772084,
      "loss": 2.6423,
      "theoretical_loss": 3.527877868976245,
      "tokens_seen": 1444282368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003424862871193494,
      "loss": 2.579,
      "theoretical_loss": 3.5278636523890565,
      "tokens_seen": 1444347904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034246973709097786,
      "loss": 2.5788,
      "theoretical_loss": 3.527849436627527,
      "tokens_seen": 1444413440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003424531870626064,
      "loss": 2.0592,
      "theoretical_loss": 3.5278352216915714,
      "tokens_seen": 1444478976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034243663703423493,
      "loss": 2.7237,
      "theoretical_loss": 3.527821007581104,
      "tokens_seen": 1444544512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034242008700586347,
      "loss": 2.6057,
      "theoretical_loss": 3.5278067942960396,
      "tokens_seen": 1444610048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034240353697749195,
      "loss": 2.5094,
      "theoretical_loss": 3.5277925818362927,
      "tokens_seen": 1444675584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003423869869491205,
      "loss": 2.5638,
      "theoretical_loss": 3.527778370201778,
      "tokens_seen": 1444741120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000342370436920749,
      "loss": 2.7023,
      "theoretical_loss": 3.527764159392411,
      "tokens_seen": 1444806656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003423538868923775,
      "loss": 2.4312,
      "theoretical_loss": 3.527749949408105,
      "tokens_seen": 1444872192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034233733686400605,
      "loss": 2.7485,
      "theoretical_loss": 3.5277357402487755,
      "tokens_seen": 1444937728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034232078683563453,
      "loss": 2.808,
      "theoretical_loss": 3.527721531914337,
      "tokens_seen": 1445003264
    },
    {
      "epoch": 0.52,
      "objective/train/avg_token_score": 0.025895804166793823,
      "objective/train/avg_weight": 0.9695385098457336,
      "objective/train/docs_used": 819121,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.017639636993408,
      "objective/train/std_weight": 0.14582692086696625,
      "objective/train/theoretical_loss": 3.527707324404704,
      "objective/train/tokens_used": 1465528800,
      "theoretical_loss": 3.527707324404704,
      "tokens_seen": 1445068800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034230423680726307,
      "loss": 2.4918,
      "theoretical_loss": 3.527707324404704,
      "tokens_seen": 1445068800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003422876867788916,
      "loss": 2.4817,
      "theoretical_loss": 3.5276931177197914,
      "tokens_seen": 1445134336
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034227113675052014,
      "loss": 2.5989,
      "theoretical_loss": 3.527678911859514,
      "tokens_seen": 1445199872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034225458672214863,
      "loss": 2.4989,
      "theoretical_loss": 3.5276647068237867,
      "tokens_seen": 1445265408
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034223803669377717,
      "loss": 2.4583,
      "theoretical_loss": 3.5276505026125236,
      "tokens_seen": 1445330944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003422214866654057,
      "loss": 2.9231,
      "theoretical_loss": 3.52763629922564,
      "tokens_seen": 1445396480
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034220493663703424,
      "loss": 2.611,
      "theoretical_loss": 3.5276220966630505,
      "tokens_seen": 1445462016
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003421883866086627,
      "loss": 2.8883,
      "theoretical_loss": 3.52760789492467,
      "tokens_seen": 1445527552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034217183658029126,
      "loss": 2.561,
      "theoretical_loss": 3.527593694010413,
      "tokens_seen": 1445593088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003421552865519198,
      "loss": 2.5156,
      "theoretical_loss": 3.5275794939201948,
      "tokens_seen": 1445658624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034213873652354834,
      "loss": 2.9038,
      "theoretical_loss": 3.52756529465393,
      "tokens_seen": 1445724160
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003421221864951768,
      "loss": 2.3956,
      "theoretical_loss": 3.5275510962115337,
      "tokens_seen": 1445789696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034210563646680536,
      "loss": 2.3747,
      "theoretical_loss": 3.5275368985929205,
      "tokens_seen": 1445855232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003420890864384339,
      "loss": 2.6338,
      "theoretical_loss": 3.527522701798005,
      "tokens_seen": 1445920768
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034207253641006243,
      "loss": 2.5293,
      "theoretical_loss": 3.5275085058267024,
      "tokens_seen": 1445986304
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003420559863816909,
      "loss": 2.6739,
      "theoretical_loss": 3.5274943106789274,
      "tokens_seen": 1446051840
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034203943635331945,
      "loss": 2.7076,
      "theoretical_loss": 3.527480116354595,
      "tokens_seen": 1446117376
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000342022886324948,
      "loss": 2.6482,
      "theoretical_loss": 3.5274659228536205,
      "tokens_seen": 1446182912
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034200633629657653,
      "loss": 2.5626,
      "theoretical_loss": 3.5274517301759185,
      "tokens_seen": 1446248448
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034198978626820507,
      "loss": 2.5821,
      "theoretical_loss": 3.527437538321404,
      "tokens_seen": 1446313984
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034197323623983355,
      "loss": 2.7425,
      "theoretical_loss": 3.5274233472899916,
      "tokens_seen": 1446379520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003419566862114621,
      "loss": 2.6685,
      "theoretical_loss": 3.527409157081597,
      "tokens_seen": 1446445056
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003419401361830906,
      "loss": 2.7061,
      "theoretical_loss": 3.527394967696134,
      "tokens_seen": 1446510592
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034192358615471916,
      "loss": 2.4722,
      "theoretical_loss": 3.5273807791335194,
      "tokens_seen": 1446576128
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034190703612634765,
      "loss": 2.4986,
      "theoretical_loss": 3.5273665913936663,
      "tokens_seen": 1446641664
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003418904860979762,
      "loss": 2.6488,
      "theoretical_loss": 3.527352404476491,
      "tokens_seen": 1446707200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034187393606960467,
      "loss": 2.66,
      "theoretical_loss": 3.527338218381908,
      "tokens_seen": 1446772736
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003418573860412332,
      "loss": 2.6926,
      "theoretical_loss": 3.5273240331098323,
      "tokens_seen": 1446838272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034184083601286174,
      "loss": 2.5077,
      "theoretical_loss": 3.5273098486601793,
      "tokens_seen": 1446903808
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003418242859844902,
      "loss": 2.582,
      "theoretical_loss": 3.527295665032864,
      "tokens_seen": 1446969344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034180773595611876,
      "loss": 2.7,
      "theoretical_loss": 3.527281482227801,
      "tokens_seen": 1447034880
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003417911859277473,
      "loss": 2.5807,
      "theoretical_loss": 3.527267300244906,
      "tokens_seen": 1447100416
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034177463589937584,
      "loss": 2.5562,
      "theoretical_loss": 3.5272531190840937,
      "tokens_seen": 1447165952
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003417580858710043,
      "loss": 2.3658,
      "theoretical_loss": 3.52723893874528,
      "tokens_seen": 1447231488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034174153584263286,
      "loss": 2.6124,
      "theoretical_loss": 3.5272247592283787,
      "tokens_seen": 1447297024
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003417249858142614,
      "loss": 2.4372,
      "theoretical_loss": 3.527210580533306,
      "tokens_seen": 1447362560
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034170843578588993,
      "loss": 2.7367,
      "theoretical_loss": 3.527196402659977,
      "tokens_seen": 1447428096
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003416918857575184,
      "loss": 2.6368,
      "theoretical_loss": 3.527182225608306,
      "tokens_seen": 1447493632
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034167533572914695,
      "loss": 2.4124,
      "theoretical_loss": 3.527168049378209,
      "tokens_seen": 1447559168
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003416587857007755,
      "loss": 2.6839,
      "theoretical_loss": 3.5271538739696013,
      "tokens_seen": 1447624704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034164223567240403,
      "loss": 2.6866,
      "theoretical_loss": 3.5271396993823974,
      "tokens_seen": 1447690240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003416256856440325,
      "loss": 2.9106,
      "theoretical_loss": 3.5271255256165133,
      "tokens_seen": 1447755776
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034160913561566105,
      "loss": 2.8567,
      "theoretical_loss": 3.527111352671864,
      "tokens_seen": 1447821312
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003415925855872896,
      "loss": 2.6761,
      "theoretical_loss": 3.527097180548364,
      "tokens_seen": 1447886848
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003415760355589181,
      "loss": 2.8401,
      "theoretical_loss": 3.5270830092459295,
      "tokens_seen": 1447952384
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003415594855305466,
      "loss": 2.5567,
      "theoretical_loss": 3.527068838764475,
      "tokens_seen": 1448017920
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034154293550217515,
      "loss": 2.7945,
      "theoretical_loss": 3.527054669103917,
      "tokens_seen": 1448083456
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003415263854738037,
      "loss": 2.4728,
      "theoretical_loss": 3.527040500264169,
      "tokens_seen": 1448148992
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003415098354454322,
      "loss": 2.5033,
      "theoretical_loss": 3.527026332245148,
      "tokens_seen": 1448214528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003414932854170607,
      "loss": 2.6844,
      "theoretical_loss": 3.527012165046769,
      "tokens_seen": 1448280064
    },
    {
      "epoch": 0.52,
      "objective/train/avg_token_score": 0.023511910811066628,
      "objective/train/avg_weight": 0.9723325371742249,
      "objective/train/docs_used": 821111,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.5331430435180664,
      "objective/train/std_weight": 0.10877357423305511,
      "objective/train/theoretical_loss": 3.5269979986689464,
      "objective/train/tokens_used": 1468805600,
      "theoretical_loss": 3.5269979986689464,
      "tokens_seen": 1448345600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034147673538868924,
      "loss": 2.6481,
      "theoretical_loss": 3.5269979986689464,
      "tokens_seen": 1448345600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003414601853603178,
      "loss": 2.9563,
      "theoretical_loss": 3.526983833111596,
      "tokens_seen": 1448411136
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003414436353319463,
      "loss": 2.4271,
      "theoretical_loss": 3.526969668374634,
      "tokens_seen": 1448476672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003414270853035748,
      "loss": 2.3893,
      "theoretical_loss": 3.5269555044579746,
      "tokens_seen": 1448542208
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034141053527520334,
      "loss": 2.7616,
      "theoretical_loss": 3.5269413413615336,
      "tokens_seen": 1448607744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003413939852468318,
      "loss": 2.4701,
      "theoretical_loss": 3.5269271790852263,
      "tokens_seen": 1448673280
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034137743521846036,
      "loss": 2.5374,
      "theoretical_loss": 3.526913017628969,
      "tokens_seen": 1448738816
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003413608851900889,
      "loss": 2.783,
      "theoretical_loss": 3.5268988569926756,
      "tokens_seen": 1448804352
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003413443351617174,
      "loss": 2.5695,
      "theoretical_loss": 3.5268846971762624,
      "tokens_seen": 1448869888
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003413277851333459,
      "loss": 2.6006,
      "theoretical_loss": 3.5268705381796455,
      "tokens_seen": 1448935424
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034131123510497446,
      "loss": 2.8038,
      "theoretical_loss": 3.526856380002739,
      "tokens_seen": 1449000960
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000341294685076603,
      "loss": 2.7691,
      "theoretical_loss": 3.5268422226454588,
      "tokens_seen": 1449066496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003412781350482315,
      "loss": 2.6045,
      "theoretical_loss": 3.526828066107721,
      "tokens_seen": 1449132032
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034126158501986,
      "loss": 2.56,
      "theoretical_loss": 3.5268139103894405,
      "tokens_seen": 1449197568
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034124503499148855,
      "loss": 2.4682,
      "theoretical_loss": 3.526799755490533,
      "tokens_seen": 1449263104
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003412284849631171,
      "loss": 2.4836,
      "theoretical_loss": 3.5267856014109142,
      "tokens_seen": 1449328640
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034121193493474557,
      "loss": 2.7725,
      "theoretical_loss": 3.526771448150499,
      "tokens_seen": 1449394176
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003411953849063741,
      "loss": 2.8859,
      "theoretical_loss": 3.5267572957092037,
      "tokens_seen": 1449459712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034117883487800265,
      "loss": 2.6483,
      "theoretical_loss": 3.526743144086944,
      "tokens_seen": 1449525248
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003411622848496312,
      "loss": 2.5639,
      "theoretical_loss": 3.5267289932836343,
      "tokens_seen": 1449590784
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034114573482125967,
      "loss": 2.639,
      "theoretical_loss": 3.5267148432991915,
      "tokens_seen": 1449656320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003411291847928882,
      "loss": 2.5507,
      "theoretical_loss": 3.52670069413353,
      "tokens_seen": 1449721856
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034111263476451674,
      "loss": 2.6032,
      "theoretical_loss": 3.526686545786567,
      "tokens_seen": 1449787392
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003410960847361453,
      "loss": 2.5833,
      "theoretical_loss": 3.526672398258216,
      "tokens_seen": 1449852928
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034107953470777376,
      "loss": 2.612,
      "theoretical_loss": 3.5266582515483944,
      "tokens_seen": 1449918464
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003410629846794023,
      "loss": 2.5022,
      "theoretical_loss": 3.5266441056570175,
      "tokens_seen": 1449984000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034104643465103084,
      "loss": 2.8176,
      "theoretical_loss": 3.526629960584,
      "tokens_seen": 1450049536
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003410298846226594,
      "loss": 2.767,
      "theoretical_loss": 3.526615816329259,
      "tokens_seen": 1450115072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034101333459428786,
      "loss": 2.5245,
      "theoretical_loss": 3.526601672892709,
      "tokens_seen": 1450180608
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003409967845659164,
      "loss": 2.5204,
      "theoretical_loss": 3.526587530274266,
      "tokens_seen": 1450246144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034098023453754493,
      "loss": 2.9598,
      "theoretical_loss": 3.5265733884738464,
      "tokens_seen": 1450311680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034096368450917347,
      "loss": 2.611,
      "theoretical_loss": 3.526559247491365,
      "tokens_seen": 1450377216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034094713448080196,
      "loss": 2.6554,
      "theoretical_loss": 3.526545107326738,
      "tokens_seen": 1450442752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003409305844524305,
      "loss": 2.6858,
      "theoretical_loss": 3.5265309679798813,
      "tokens_seen": 1450508288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034091403442405903,
      "loss": 2.5145,
      "theoretical_loss": 3.52651682945071,
      "tokens_seen": 1450573824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003408974843956875,
      "loss": 2.5831,
      "theoretical_loss": 3.526502691739141,
      "tokens_seen": 1450639360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034088093436731605,
      "loss": 2.6674,
      "theoretical_loss": 3.5264885548450886,
      "tokens_seen": 1450704896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034086438433894454,
      "loss": 2.7934,
      "theoretical_loss": 3.52647441876847,
      "tokens_seen": 1450770432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034084783431057307,
      "loss": 2.671,
      "theoretical_loss": 3.5264602835092003,
      "tokens_seen": 1450835968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003408312842822016,
      "loss": 2.4352,
      "theoretical_loss": 3.526446149067195,
      "tokens_seen": 1450901504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034081473425383015,
      "loss": 2.7849,
      "theoretical_loss": 3.5264320154423707,
      "tokens_seen": 1450967040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034079818422545863,
      "loss": 2.527,
      "theoretical_loss": 3.526417882634643,
      "tokens_seen": 1451032576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034078163419708717,
      "loss": 2.4419,
      "theoretical_loss": 3.5264037506439276,
      "tokens_seen": 1451098112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003407650841687157,
      "loss": 2.4452,
      "theoretical_loss": 3.52638961947014,
      "tokens_seen": 1451163648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034074853414034424,
      "loss": 2.6681,
      "theoretical_loss": 3.5263754891131973,
      "tokens_seen": 1451229184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003407319841119728,
      "loss": 2.7363,
      "theoretical_loss": 3.5263613595730146,
      "tokens_seen": 1451294720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034071543408360126,
      "loss": 2.8492,
      "theoretical_loss": 3.526347230849507,
      "tokens_seen": 1451360256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003406988840552298,
      "loss": 2.6201,
      "theoretical_loss": 3.526333102942592,
      "tokens_seen": 1451425792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034068233402685834,
      "loss": 2.64,
      "theoretical_loss": 3.5263189758521847,
      "tokens_seen": 1451491328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003406657839984869,
      "loss": 2.6448,
      "theoretical_loss": 3.5263048495782012,
      "tokens_seen": 1451556864
    },
    {
      "epoch": 0.52,
      "objective/train/avg_token_score": 0.015410682186484337,
      "objective/train/avg_weight": 0.9818550944328308,
      "objective/train/docs_used": 823345,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8926329612731934,
      "objective/train/std_weight": 0.08770640194416046,
      "objective/train/theoretical_loss": 3.5262907241205568,
      "objective/train/tokens_used": 1472082400,
      "theoretical_loss": 3.5262907241205568,
      "tokens_seen": 1451622400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034064923397011536,
      "loss": 2.6018,
      "theoretical_loss": 3.5262907241205568,
      "tokens_seen": 1451622400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003406326839417439,
      "loss": 2.7379,
      "theoretical_loss": 3.5262765994791687,
      "tokens_seen": 1451687936
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034061613391337244,
      "loss": 2.8533,
      "theoretical_loss": 3.5262624756539522,
      "tokens_seen": 1451753472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000340599583885001,
      "loss": 2.64,
      "theoretical_loss": 3.5262483526448234,
      "tokens_seen": 1451819008
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034058303385662946,
      "loss": 2.5174,
      "theoretical_loss": 3.526234230451698,
      "tokens_seen": 1451884544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000340566483828258,
      "loss": 2.4035,
      "theoretical_loss": 3.5262201090744925,
      "tokens_seen": 1451950080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034054993379988653,
      "loss": 2.7558,
      "theoretical_loss": 3.526205988513123,
      "tokens_seen": 1452015616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034053338377151507,
      "loss": 2.6368,
      "theoretical_loss": 3.5261918687675045,
      "tokens_seen": 1452081152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034051683374314355,
      "loss": 2.3757,
      "theoretical_loss": 3.5261777498375544,
      "tokens_seen": 1452146688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003405002837147721,
      "loss": 2.4762,
      "theoretical_loss": 3.526163631723188,
      "tokens_seen": 1452212224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034048373368640063,
      "loss": 2.717,
      "theoretical_loss": 3.526149514424322,
      "tokens_seen": 1452277760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034046718365802917,
      "loss": 2.3811,
      "theoretical_loss": 3.5261353979408714,
      "tokens_seen": 1452343296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034045063362965765,
      "loss": 2.4901,
      "theoretical_loss": 3.5261212822727535,
      "tokens_seen": 1452408832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003404340836012862,
      "loss": 2.6185,
      "theoretical_loss": 3.526107167419884,
      "tokens_seen": 1452474368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003404175335729147,
      "loss": 2.6449,
      "theoretical_loss": 3.526093053382179,
      "tokens_seen": 1452539904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003404009835445432,
      "loss": 2.6949,
      "theoretical_loss": 3.5260789401595543,
      "tokens_seen": 1452605440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034038443351617174,
      "loss": 2.526,
      "theoretical_loss": 3.5260648277519264,
      "tokens_seen": 1452670976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034036788348780023,
      "loss": 2.6359,
      "theoretical_loss": 3.5260507161592116,
      "tokens_seen": 1452736512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034035133345942877,
      "loss": 2.5815,
      "theoretical_loss": 3.526036605381326,
      "tokens_seen": 1452802048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003403347834310573,
      "loss": 2.7423,
      "theoretical_loss": 3.5260224954181854,
      "tokens_seen": 1452867584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034031823340268584,
      "loss": 2.7789,
      "theoretical_loss": 3.5260083862697065,
      "tokens_seen": 1452933120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003403016833743143,
      "loss": 2.5345,
      "theoretical_loss": 3.5259942779358053,
      "tokens_seen": 1452998656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034028513334594286,
      "loss": 2.7482,
      "theoretical_loss": 3.525980170416398,
      "tokens_seen": 1453064192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003402685833175714,
      "loss": 2.5661,
      "theoretical_loss": 3.525966063711401,
      "tokens_seen": 1453129728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034025203328919994,
      "loss": 2.6477,
      "theoretical_loss": 3.525951957820731,
      "tokens_seen": 1453195264
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003402354832608284,
      "loss": 2.5962,
      "theoretical_loss": 3.5259378527443035,
      "tokens_seen": 1453260800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034021893323245696,
      "loss": 2.6918,
      "theoretical_loss": 3.5259237484820347,
      "tokens_seen": 1453326336
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003402023832040855,
      "loss": 2.5743,
      "theoretical_loss": 3.525909645033841,
      "tokens_seen": 1453391872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034018583317571403,
      "loss": 2.5923,
      "theoretical_loss": 3.5258955423996396,
      "tokens_seen": 1453457408
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003401692831473425,
      "loss": 2.3433,
      "theoretical_loss": 3.525881440579346,
      "tokens_seen": 1453522944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034015273311897105,
      "loss": 2.5122,
      "theoretical_loss": 3.5258673395728763,
      "tokens_seen": 1453588480
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003401361830905996,
      "loss": 2.5185,
      "theoretical_loss": 3.5258532393801474,
      "tokens_seen": 1453654016
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034011963306222813,
      "loss": 2.6403,
      "theoretical_loss": 3.525839140001076,
      "tokens_seen": 1453719552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003401030830338566,
      "loss": 2.7173,
      "theoretical_loss": 3.525825041435578,
      "tokens_seen": 1453785088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034008653300548515,
      "loss": 2.5898,
      "theoretical_loss": 3.525810943683569,
      "tokens_seen": 1453850624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003400699829771137,
      "loss": 2.7602,
      "theoretical_loss": 3.5257968467449663,
      "tokens_seen": 1453916160
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003400534329487422,
      "loss": 2.402,
      "theoretical_loss": 3.525782750619687,
      "tokens_seen": 1453981696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003400368829203707,
      "loss": 2.7446,
      "theoretical_loss": 3.5257686553076457,
      "tokens_seen": 1454047232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00034002033289199925,
      "loss": 2.7046,
      "theoretical_loss": 3.5257545608087604,
      "tokens_seen": 1454112768
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003400037828636278,
      "loss": 2.6151,
      "theoretical_loss": 3.5257404671229464,
      "tokens_seen": 1454178304
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003399872328352563,
      "loss": 2.5308,
      "theoretical_loss": 3.5257263742501213,
      "tokens_seen": 1454243840
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003399706828068848,
      "loss": 2.7406,
      "theoretical_loss": 3.5257122821902005,
      "tokens_seen": 1454309376
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033995413277851334,
      "loss": 2.5926,
      "theoretical_loss": 3.525698190943101,
      "tokens_seen": 1454374912
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003399375827501419,
      "loss": 2.6759,
      "theoretical_loss": 3.5256841005087396,
      "tokens_seen": 1454440448
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033992103272177036,
      "loss": 2.7662,
      "theoretical_loss": 3.5256700108870325,
      "tokens_seen": 1454505984
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003399044826933989,
      "loss": 2.4746,
      "theoretical_loss": 3.5256559220778954,
      "tokens_seen": 1454571520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003398879326650274,
      "loss": 2.5568,
      "theoretical_loss": 3.525641834081246,
      "tokens_seen": 1454637056
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003398713826366559,
      "loss": 2.7678,
      "theoretical_loss": 3.525627746897001,
      "tokens_seen": 1454702592
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033985483260828446,
      "loss": 2.6736,
      "theoretical_loss": 3.5256136605250763,
      "tokens_seen": 1454768128
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000339838282579913,
      "loss": 2.3987,
      "theoretical_loss": 3.525599574965388,
      "tokens_seen": 1454833664
    },
    {
      "epoch": 0.52,
      "objective/train/avg_token_score": 0.020806394517421722,
      "objective/train/avg_weight": 0.9755421876907349,
      "objective/train/docs_used": 825270,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6069695949554443,
      "objective/train/std_weight": 0.07848712801933289,
      "objective/train/theoretical_loss": 3.5255854902178534,
      "objective/train/tokens_used": 1475359200,
      "theoretical_loss": 3.5255854902178534,
      "tokens_seen": 1454899200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003398217325515415,
      "loss": 2.4368,
      "theoretical_loss": 3.5255854902178534,
      "tokens_seen": 1454899200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033980518252317,
      "loss": 2.3845,
      "theoretical_loss": 3.5255714062823893,
      "tokens_seen": 1454964736
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033978863249479855,
      "loss": 2.839,
      "theoretical_loss": 3.5255573231589117,
      "tokens_seen": 1455030272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003397720824664271,
      "loss": 2.5391,
      "theoretical_loss": 3.5255432408473375,
      "tokens_seen": 1455095808
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003397555324380556,
      "loss": 2.4496,
      "theoretical_loss": 3.525529159347583,
      "tokens_seen": 1455161344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003397389824096841,
      "loss": 2.6857,
      "theoretical_loss": 3.5255150786595655,
      "tokens_seen": 1455226880
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033972243238131265,
      "loss": 2.5492,
      "theoretical_loss": 3.525500998783201,
      "tokens_seen": 1455292416
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003397058823529412,
      "loss": 2.8329,
      "theoretical_loss": 3.5254869197184067,
      "tokens_seen": 1455357952
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033968933232456967,
      "loss": 2.7128,
      "theoretical_loss": 3.525472841465099,
      "tokens_seen": 1455423488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003396727822961982,
      "loss": 2.5102,
      "theoretical_loss": 3.5254587640231945,
      "tokens_seen": 1455489024
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033965623226782675,
      "loss": 2.5224,
      "theoretical_loss": 3.52544468739261,
      "tokens_seen": 1455554560
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003396396822394553,
      "loss": 2.4091,
      "theoretical_loss": 3.5254306115732623,
      "tokens_seen": 1455620096
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033962313221108377,
      "loss": 2.6627,
      "theoretical_loss": 3.525416536565068,
      "tokens_seen": 1455685632
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003396065821827123,
      "loss": 2.4719,
      "theoretical_loss": 3.525402462367944,
      "tokens_seen": 1455751168
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033959003215434084,
      "loss": 2.6115,
      "theoretical_loss": 3.525388388981807,
      "tokens_seen": 1455816704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003395734821259694,
      "loss": 2.7107,
      "theoretical_loss": 3.525374316406574,
      "tokens_seen": 1455882240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033955693209759786,
      "loss": 2.6137,
      "theoretical_loss": 3.525360244642161,
      "tokens_seen": 1455947776
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003395403820692264,
      "loss": 2.7905,
      "theoretical_loss": 3.5253461736884852,
      "tokens_seen": 1456013312
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033952383204085494,
      "loss": 2.4789,
      "theoretical_loss": 3.525332103545464,
      "tokens_seen": 1456078848
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003395072820124835,
      "loss": 2.5709,
      "theoretical_loss": 3.5253180342130133,
      "tokens_seen": 1456144384
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033949073198411196,
      "loss": 2.7356,
      "theoretical_loss": 3.5253039656910508,
      "tokens_seen": 1456209920
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003394741819557405,
      "loss": 2.6368,
      "theoretical_loss": 3.5252898979794924,
      "tokens_seen": 1456275456
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033945763192736903,
      "loss": 2.5215,
      "theoretical_loss": 3.5252758310782557,
      "tokens_seen": 1456340992
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033944108189899757,
      "loss": 2.5907,
      "theoretical_loss": 3.525261764987257,
      "tokens_seen": 1456406528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033942453187062605,
      "loss": 2.9011,
      "theoretical_loss": 3.5252476997064135,
      "tokens_seen": 1456472064
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003394079818422546,
      "loss": 2.5768,
      "theoretical_loss": 3.525233635235642,
      "tokens_seen": 1456537600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003393914318138831,
      "loss": 2.601,
      "theoretical_loss": 3.5252195715748598,
      "tokens_seen": 1456603136
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003393748817855116,
      "loss": 2.387,
      "theoretical_loss": 3.5252055087239835,
      "tokens_seen": 1456668672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033935833175714015,
      "loss": 2.5269,
      "theoretical_loss": 3.5251914466829293,
      "tokens_seen": 1456734208
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003393417817287687,
      "loss": 2.5687,
      "theoretical_loss": 3.5251773854516153,
      "tokens_seen": 1456799744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033932523170039717,
      "loss": 2.5328,
      "theoretical_loss": 3.5251633250299577,
      "tokens_seen": 1456865280
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003393086816720257,
      "loss": 2.5902,
      "theoretical_loss": 3.525149265417874,
      "tokens_seen": 1456930816
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033929213164365425,
      "loss": 2.6183,
      "theoretical_loss": 3.525135206615281,
      "tokens_seen": 1456996352
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003392755816152828,
      "loss": 2.5476,
      "theoretical_loss": 3.5251211486220955,
      "tokens_seen": 1457061888
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033925903158691127,
      "loss": 2.4703,
      "theoretical_loss": 3.525107091438234,
      "tokens_seen": 1457127424
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003392424815585398,
      "loss": 2.4562,
      "theoretical_loss": 3.5250930350636147,
      "tokens_seen": 1457192960
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033922593153016834,
      "loss": 2.6769,
      "theoretical_loss": 3.525078979498154,
      "tokens_seen": 1457258496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003392093815017969,
      "loss": 2.8232,
      "theoretical_loss": 3.525064924741769,
      "tokens_seen": 1457324032
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033919283147342536,
      "loss": 2.5012,
      "theoretical_loss": 3.5250508707943764,
      "tokens_seen": 1457389568
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003391762814450539,
      "loss": 2.7066,
      "theoretical_loss": 3.5250368176558933,
      "tokens_seen": 1457455104
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033915973141668244,
      "loss": 2.5889,
      "theoretical_loss": 3.5250227653262374,
      "tokens_seen": 1457520640
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000339143181388311,
      "loss": 2.1347,
      "theoretical_loss": 3.5250087138053257,
      "tokens_seen": 1457586176
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033912663135993946,
      "loss": 2.4841,
      "theoretical_loss": 3.524994663093074,
      "tokens_seen": 1457651712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000339110081331568,
      "loss": 2.4762,
      "theoretical_loss": 3.5249806131894013,
      "tokens_seen": 1457717248
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033909353130319653,
      "loss": 2.6461,
      "theoretical_loss": 3.524966564094224,
      "tokens_seen": 1457782784
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033907698127482507,
      "loss": 2.6398,
      "theoretical_loss": 3.524952515807458,
      "tokens_seen": 1457848320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033906043124645356,
      "loss": 2.6209,
      "theoretical_loss": 3.5249384683290224,
      "tokens_seen": 1457913856
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003390438812180821,
      "loss": 2.5854,
      "theoretical_loss": 3.524924421658833,
      "tokens_seen": 1457979392
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033902733118971063,
      "loss": 2.8096,
      "theoretical_loss": 3.5249103757968077,
      "tokens_seen": 1458044928
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033901078116133917,
      "loss": 2.5183,
      "theoretical_loss": 3.5248963307428633,
      "tokens_seen": 1458110464
    },
    {
      "epoch": 0.52,
      "objective/train/avg_token_score": 0.02130289375782013,
      "objective/train/avg_weight": 0.9749473333358765,
      "objective/train/docs_used": 827227,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6375763416290283,
      "objective/train/std_weight": 0.08238950371742249,
      "objective/train/theoretical_loss": 3.524882286496917,
      "objective/train/tokens_used": 1478636000,
      "theoretical_loss": 3.524882286496917,
      "tokens_seen": 1458176000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033899423113296765,
      "loss": 2.5009,
      "theoretical_loss": 3.524882286496917,
      "tokens_seen": 1458176000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003389776811045962,
      "loss": 2.6453,
      "theoretical_loss": 3.5248682430588865,
      "tokens_seen": 1458241536
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003389611310762247,
      "loss": 2.5434,
      "theoretical_loss": 3.5248542004286882,
      "tokens_seen": 1458307072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033894458104785326,
      "loss": 2.6643,
      "theoretical_loss": 3.52484015860624,
      "tokens_seen": 1458372608
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033892803101948175,
      "loss": 2.6189,
      "theoretical_loss": 3.5248261175914584,
      "tokens_seen": 1458438144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033891148099111023,
      "loss": 2.8339,
      "theoretical_loss": 3.5248120773842615,
      "tokens_seen": 1458503680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033889493096273877,
      "loss": 2.5637,
      "theoretical_loss": 3.5247980379845663,
      "tokens_seen": 1458569216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003388783809343673,
      "loss": 2.386,
      "theoretical_loss": 3.52478399939229,
      "tokens_seen": 1458634752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033886183090599584,
      "loss": 2.6303,
      "theoretical_loss": 3.5247699616073502,
      "tokens_seen": 1458700288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003388452808776243,
      "loss": 2.7788,
      "theoretical_loss": 3.5247559246296634,
      "tokens_seen": 1458765824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033882873084925286,
      "loss": 2.5437,
      "theoretical_loss": 3.5247418884591473,
      "tokens_seen": 1458831360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003388121808208814,
      "loss": 2.6931,
      "theoretical_loss": 3.52472785309572,
      "tokens_seen": 1458896896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033879563079250994,
      "loss": 2.4756,
      "theoretical_loss": 3.524713818539297,
      "tokens_seen": 1458962432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003387790807641384,
      "loss": 2.4569,
      "theoretical_loss": 3.5246997847897976,
      "tokens_seen": 1459027968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033876253073576696,
      "loss": 2.5083,
      "theoretical_loss": 3.5246857518471386,
      "tokens_seen": 1459093504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003387459807073955,
      "loss": 2.4532,
      "theoretical_loss": 3.524671719711237,
      "tokens_seen": 1459159040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033872943067902404,
      "loss": 2.5751,
      "theoretical_loss": 3.52465768838201,
      "tokens_seen": 1459224576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003387128806506525,
      "loss": 2.566,
      "theoretical_loss": 3.5246436578593756,
      "tokens_seen": 1459290112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033869633062228106,
      "loss": 2.6744,
      "theoretical_loss": 3.5246296281432508,
      "tokens_seen": 1459355648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003386797805939096,
      "loss": 2.7325,
      "theoretical_loss": 3.5246155992335533,
      "tokens_seen": 1459421184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033866323056553813,
      "loss": 2.534,
      "theoretical_loss": 3.5246015711302006,
      "tokens_seen": 1459486720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003386466805371666,
      "loss": 2.7244,
      "theoretical_loss": 3.5245875438331096,
      "tokens_seen": 1459552256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033863013050879515,
      "loss": 2.6323,
      "theoretical_loss": 3.5245735173421986,
      "tokens_seen": 1459617792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003386135804804237,
      "loss": 2.5638,
      "theoretical_loss": 3.524559491657384,
      "tokens_seen": 1459683328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003385970304520522,
      "loss": 2.547,
      "theoretical_loss": 3.5245454667785845,
      "tokens_seen": 1459748864
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003385804804236807,
      "loss": 2.7538,
      "theoretical_loss": 3.524531442705716,
      "tokens_seen": 1459814400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033856393039530925,
      "loss": 2.6665,
      "theoretical_loss": 3.5245174194386983,
      "tokens_seen": 1459879936
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003385473803669378,
      "loss": 2.4722,
      "theoretical_loss": 3.5245033969774466,
      "tokens_seen": 1459945472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003385308303385663,
      "loss": 2.5282,
      "theoretical_loss": 3.5244893753218802,
      "tokens_seen": 1460011008
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003385142803101948,
      "loss": 2.772,
      "theoretical_loss": 3.5244753544719156,
      "tokens_seen": 1460076544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033849773028182334,
      "loss": 2.5794,
      "theoretical_loss": 3.5244613344274707,
      "tokens_seen": 1460142080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003384811802534519,
      "loss": 2.9388,
      "theoretical_loss": 3.5244473151884628,
      "tokens_seen": 1460207616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003384646302250804,
      "loss": 2.6572,
      "theoretical_loss": 3.52443329675481,
      "tokens_seen": 1460273152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003384480801967089,
      "loss": 2.6258,
      "theoretical_loss": 3.5244192791264295,
      "tokens_seen": 1460338688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003384315301683374,
      "loss": 2.7597,
      "theoretical_loss": 3.524405262303239,
      "tokens_seen": 1460404224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003384149801399659,
      "loss": 2.5895,
      "theoretical_loss": 3.524391246285156,
      "tokens_seen": 1460469760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033839843011159446,
      "loss": 2.2681,
      "theoretical_loss": 3.524377231072098,
      "tokens_seen": 1460535296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000338381880083223,
      "loss": 2.978,
      "theoretical_loss": 3.5243632166639838,
      "tokens_seen": 1460600832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003383653300548515,
      "loss": 2.5488,
      "theoretical_loss": 3.5243492030607295,
      "tokens_seen": 1460666368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033834878002648,
      "loss": 2.4994,
      "theoretical_loss": 3.5243351902622537,
      "tokens_seen": 1460731904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033833222999810856,
      "loss": 2.4977,
      "theoretical_loss": 3.5243211782684734,
      "tokens_seen": 1460797440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003383156799697371,
      "loss": 2.7428,
      "theoretical_loss": 3.524307167079307,
      "tokens_seen": 1460862976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003382991299413656,
      "loss": 2.4492,
      "theoretical_loss": 3.5242931566946716,
      "tokens_seen": 1460928512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003382825799129941,
      "loss": 2.559,
      "theoretical_loss": 3.524279147114486,
      "tokens_seen": 1460994048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033826602988462265,
      "loss": 2.5836,
      "theoretical_loss": 3.5242651383386665,
      "tokens_seen": 1461059584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003382494798562512,
      "loss": 2.643,
      "theoretical_loss": 3.5242511303671313,
      "tokens_seen": 1461125120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003382329298278797,
      "loss": 2.6607,
      "theoretical_loss": 3.5242371231997986,
      "tokens_seen": 1461190656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003382163797995082,
      "loss": 2.5307,
      "theoretical_loss": 3.524223116836586,
      "tokens_seen": 1461256192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033819982977113675,
      "loss": 2.6589,
      "theoretical_loss": 3.524209111277411,
      "tokens_seen": 1461321728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003381832797427653,
      "loss": 2.6415,
      "theoretical_loss": 3.5241951065221917,
      "tokens_seen": 1461387264
    },
    {
      "epoch": 0.52,
      "objective/train/avg_token_score": 0.007561920210719109,
      "objective/train/avg_weight": 0.9911026358604431,
      "objective/train/docs_used": 829193,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8184239864349365,
      "objective/train/std_weight": 0.03605998679995537,
      "objective/train/theoretical_loss": 3.524181102570846,
      "objective/train/tokens_used": 1481912800,
      "theoretical_loss": 3.524181102570846,
      "tokens_seen": 1461452800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033816672971439377,
      "loss": 2.2034,
      "theoretical_loss": 3.524181102570846,
      "tokens_seen": 1461452800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003381501796860223,
      "loss": 2.2347,
      "theoretical_loss": 3.524167099423291,
      "tokens_seen": 1461518336
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033813362965765084,
      "loss": 2.5465,
      "theoretical_loss": 3.5241530970794455,
      "tokens_seen": 1461583872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003381170796292794,
      "loss": 2.3973,
      "theoretical_loss": 3.524139095539226,
      "tokens_seen": 1461649408
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033810052960090787,
      "loss": 2.6043,
      "theoretical_loss": 3.524125094802552,
      "tokens_seen": 1461714944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003380839795725364,
      "loss": 2.3879,
      "theoretical_loss": 3.5241110948693404,
      "tokens_seen": 1461780480
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033806742954416494,
      "loss": 2.655,
      "theoretical_loss": 3.524097095739509,
      "tokens_seen": 1461846016
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003380508795157935,
      "loss": 2.5497,
      "theoretical_loss": 3.5240830974129764,
      "tokens_seen": 1461911552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033803432948742196,
      "loss": 2.5693,
      "theoretical_loss": 3.5240690998896596,
      "tokens_seen": 1461977088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003380177794590505,
      "loss": 2.7072,
      "theoretical_loss": 3.524055103169477,
      "tokens_seen": 1462042624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033800122943067904,
      "loss": 2.5811,
      "theoretical_loss": 3.524041107252346,
      "tokens_seen": 1462108160
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003379846794023076,
      "loss": 2.6265,
      "theoretical_loss": 3.524027112138186,
      "tokens_seen": 1462173696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003379681293739361,
      "loss": 2.613,
      "theoretical_loss": 3.524013117826913,
      "tokens_seen": 1462239232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003379515793455646,
      "loss": 2.7428,
      "theoretical_loss": 3.523999124318446,
      "tokens_seen": 1462304768
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003379350293171931,
      "loss": 2.6336,
      "theoretical_loss": 3.5239851316127035,
      "tokens_seen": 1462370304
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003379184792888216,
      "loss": 2.7507,
      "theoretical_loss": 3.5239711397096025,
      "tokens_seen": 1462435840
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033790192926045015,
      "loss": 2.4765,
      "theoretical_loss": 3.523957148609061,
      "tokens_seen": 1462501376
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003378853792320787,
      "loss": 2.4724,
      "theoretical_loss": 3.5239431583109972,
      "tokens_seen": 1462566912
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003378688292037072,
      "loss": 2.601,
      "theoretical_loss": 3.5239291688153296,
      "tokens_seen": 1462632448
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003378522791753357,
      "loss": 2.8243,
      "theoretical_loss": 3.5239151801219757,
      "tokens_seen": 1462697984
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033783572914696425,
      "loss": 2.8046,
      "theoretical_loss": 3.523901192230854,
      "tokens_seen": 1462763520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003378191791185928,
      "loss": 2.795,
      "theoretical_loss": 3.523887205141882,
      "tokens_seen": 1462829056
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033780262909022127,
      "loss": 2.5259,
      "theoretical_loss": 3.5238732188549777,
      "tokens_seen": 1462894592
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003377860790618498,
      "loss": 2.5843,
      "theoretical_loss": 3.52385923337006,
      "tokens_seen": 1462960128
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033776952903347835,
      "loss": 2.8159,
      "theoretical_loss": 3.5238452486870466,
      "tokens_seen": 1463025664
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003377529790051069,
      "loss": 2.6512,
      "theoretical_loss": 3.5238312648058554,
      "tokens_seen": 1463091200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033773642897673537,
      "loss": 2.6785,
      "theoretical_loss": 3.5238172817264046,
      "tokens_seen": 1463156736
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003377198789483639,
      "loss": 2.3481,
      "theoretical_loss": 3.523803299448612,
      "tokens_seen": 1463222272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033770332891999244,
      "loss": 2.539,
      "theoretical_loss": 3.5237893179723967,
      "tokens_seen": 1463287808
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000337686778891621,
      "loss": 2.6085,
      "theoretical_loss": 3.5237753372976757,
      "tokens_seen": 1463353344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033767022886324946,
      "loss": 2.6682,
      "theoretical_loss": 3.523761357424368,
      "tokens_seen": 1463418880
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000337653678834878,
      "loss": 2.6488,
      "theoretical_loss": 3.5237473783523914,
      "tokens_seen": 1463484416
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033763712880650654,
      "loss": 2.7867,
      "theoretical_loss": 3.523733400081664,
      "tokens_seen": 1463549952
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003376205787781351,
      "loss": 2.7522,
      "theoretical_loss": 3.5237194226121042,
      "tokens_seen": 1463615488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033760402874976356,
      "loss": 2.6989,
      "theoretical_loss": 3.52370544594363,
      "tokens_seen": 1463681024
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003375874787213921,
      "loss": 2.5422,
      "theoretical_loss": 3.5236914700761606,
      "tokens_seen": 1463746560
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033757092869302063,
      "loss": 2.8871,
      "theoretical_loss": 3.5236774950096126,
      "tokens_seen": 1463812096
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033755437866464917,
      "loss": 2.8305,
      "theoretical_loss": 3.523663520743905,
      "tokens_seen": 1463877632
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033753782863627765,
      "loss": 2.4201,
      "theoretical_loss": 3.5236495472789566,
      "tokens_seen": 1463943168
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003375212786079062,
      "loss": 2.5416,
      "theoretical_loss": 3.5236355746146852,
      "tokens_seen": 1464008704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033750472857953473,
      "loss": 2.6042,
      "theoretical_loss": 3.5236216027510086,
      "tokens_seen": 1464074240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033748817855116327,
      "loss": 2.8542,
      "theoretical_loss": 3.523607631687846,
      "tokens_seen": 1464139776
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033747162852279175,
      "loss": 2.4972,
      "theoretical_loss": 3.523593661425115,
      "tokens_seen": 1464205312
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033745507849442023,
      "loss": 2.3127,
      "theoretical_loss": 3.5235796919627345,
      "tokens_seen": 1464270848
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033743852846604877,
      "loss": 2.798,
      "theoretical_loss": 3.5235657233006226,
      "tokens_seen": 1464336384
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003374219784376773,
      "loss": 2.6864,
      "theoretical_loss": 3.523551755438697,
      "tokens_seen": 1464401920
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033740542840930585,
      "loss": 2.5314,
      "theoretical_loss": 3.523537788376877,
      "tokens_seen": 1464467456
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033738887838093433,
      "loss": 2.5954,
      "theoretical_loss": 3.523523822115081,
      "tokens_seen": 1464532992
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033737232835256287,
      "loss": 2.6494,
      "theoretical_loss": 3.5235098566532264,
      "tokens_seen": 1464598528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003373557783241914,
      "loss": 2.4985,
      "theoretical_loss": 3.5234958919912316,
      "tokens_seen": 1464664064
    },
    {
      "epoch": 0.52,
      "objective/train/avg_token_score": 0.042604994028806686,
      "objective/train/avg_weight": 0.9498737454414368,
      "objective/train/docs_used": 831131,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.566676139831543,
      "objective/train/std_weight": 0.170751690864563,
      "objective/train/theoretical_loss": 3.5234819281290166,
      "objective/train/tokens_used": 1485189600,
      "theoretical_loss": 3.5234819281290166,
      "tokens_seen": 1464729600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033733922829581994,
      "loss": 2.5436,
      "theoretical_loss": 3.5234819281290166,
      "tokens_seen": 1464729600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003373226782674484,
      "loss": 2.6881,
      "theoretical_loss": 3.5234679650664984,
      "tokens_seen": 1464795136
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033730612823907696,
      "loss": 2.5975,
      "theoretical_loss": 3.5234540028035957,
      "tokens_seen": 1464860672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003372895782107055,
      "loss": 2.6217,
      "theoretical_loss": 3.523440041340227,
      "tokens_seen": 1464926208
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033727302818233404,
      "loss": 2.7731,
      "theoretical_loss": 3.523426080676311,
      "tokens_seen": 1464991744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003372564781539625,
      "loss": 2.5327,
      "theoretical_loss": 3.523412120811766,
      "tokens_seen": 1465057280
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033723992812559106,
      "loss": 2.8467,
      "theoretical_loss": 3.5233981617465107,
      "tokens_seen": 1465122816
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003372233780972196,
      "loss": 2.7168,
      "theoretical_loss": 3.5233842034804628,
      "tokens_seen": 1465188352
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033720682806884813,
      "loss": 2.7238,
      "theoretical_loss": 3.523370246013541,
      "tokens_seen": 1465253888
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003371902780404766,
      "loss": 2.7226,
      "theoretical_loss": 3.5233562893456645,
      "tokens_seen": 1465319424
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033717372801210515,
      "loss": 2.5196,
      "theoretical_loss": 3.5233423334767515,
      "tokens_seen": 1465384960
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003371571779837337,
      "loss": 2.6805,
      "theoretical_loss": 3.5233283784067204,
      "tokens_seen": 1465450496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033714062795536223,
      "loss": 2.6371,
      "theoretical_loss": 3.5233144241354903,
      "tokens_seen": 1465516032
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003371240779269907,
      "loss": 2.6478,
      "theoretical_loss": 3.5233004706629787,
      "tokens_seen": 1465581568
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033710752789861925,
      "loss": 2.4626,
      "theoretical_loss": 3.5232865179891046,
      "tokens_seen": 1465647104
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003370909778702478,
      "loss": 2.677,
      "theoretical_loss": 3.5232725661137874,
      "tokens_seen": 1465712640
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003370744278418763,
      "loss": 2.6335,
      "theoretical_loss": 3.5232586150369443,
      "tokens_seen": 1465778176
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003370578778135048,
      "loss": 2.8694,
      "theoretical_loss": 3.5232446647584954,
      "tokens_seen": 1465843712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033704132778513335,
      "loss": 2.8613,
      "theoretical_loss": 3.5232307152783577,
      "tokens_seen": 1465909248
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003370247777567619,
      "loss": 2.5336,
      "theoretical_loss": 3.523216766596451,
      "tokens_seen": 1465974784
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003370082277283904,
      "loss": 2.7703,
      "theoretical_loss": 3.523202818712694,
      "tokens_seen": 1466040320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003369916777000189,
      "loss": 2.5415,
      "theoretical_loss": 3.5231888716270046,
      "tokens_seen": 1466105856
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033697512767164744,
      "loss": 2.7771,
      "theoretical_loss": 3.523174925339302,
      "tokens_seen": 1466171392
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003369585776432759,
      "loss": 2.5324,
      "theoretical_loss": 3.5231609798495045,
      "tokens_seen": 1466236928
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033694202761490446,
      "loss": 2.3904,
      "theoretical_loss": 3.523147035157531,
      "tokens_seen": 1466302464
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000336925477586533,
      "loss": 2.6394,
      "theoretical_loss": 3.5231330912633005,
      "tokens_seen": 1466368000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003369089275581615,
      "loss": 2.6982,
      "theoretical_loss": 3.523119148166731,
      "tokens_seen": 1466433536
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033689237752979,
      "loss": 2.6407,
      "theoretical_loss": 3.5231052058677417,
      "tokens_seen": 1466499072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033687582750141856,
      "loss": 2.5602,
      "theoretical_loss": 3.523091264366252,
      "tokens_seen": 1466564608
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003368592774730471,
      "loss": 2.6682,
      "theoretical_loss": 3.523077323662179,
      "tokens_seen": 1466630144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003368427274446756,
      "loss": 2.6504,
      "theoretical_loss": 3.523063383755443,
      "tokens_seen": 1466695680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003368261774163041,
      "loss": 2.6246,
      "theoretical_loss": 3.523049444645962,
      "tokens_seen": 1466761216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033680962738793266,
      "loss": 2.6837,
      "theoretical_loss": 3.5230355063336547,
      "tokens_seen": 1466826752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003367930773595612,
      "loss": 2.4212,
      "theoretical_loss": 3.5230215688184403,
      "tokens_seen": 1466892288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003367765273311897,
      "loss": 2.5566,
      "theoretical_loss": 3.5230076321002373,
      "tokens_seen": 1466957824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003367599773028182,
      "loss": 2.5546,
      "theoretical_loss": 3.5229936961789647,
      "tokens_seen": 1467023360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033674342727444675,
      "loss": 2.9175,
      "theoretical_loss": 3.5229797610545415,
      "tokens_seen": 1467088896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003367268772460753,
      "loss": 2.5046,
      "theoretical_loss": 3.522965826726886,
      "tokens_seen": 1467154432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003367103272177038,
      "loss": 2.5139,
      "theoretical_loss": 3.522951893195918,
      "tokens_seen": 1467219968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003366937771893323,
      "loss": 2.4834,
      "theoretical_loss": 3.5229379604615554,
      "tokens_seen": 1467285504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033667722716096085,
      "loss": 2.6812,
      "theoretical_loss": 3.5229240285237173,
      "tokens_seen": 1467351040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003366606771325894,
      "loss": 2.6438,
      "theoretical_loss": 3.522910097382323,
      "tokens_seen": 1467416576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003366441271042179,
      "loss": 2.4279,
      "theoretical_loss": 3.522896167037291,
      "tokens_seen": 1467482112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003366275770758464,
      "loss": 2.5606,
      "theoretical_loss": 3.52288223748854,
      "tokens_seen": 1467547648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033661102704747494,
      "loss": 2.6549,
      "theoretical_loss": 3.5228683087359896,
      "tokens_seen": 1467613184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003365944770191035,
      "loss": 2.5611,
      "theoretical_loss": 3.5228543807795587,
      "tokens_seen": 1467678720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000336577926990732,
      "loss": 2.5501,
      "theoretical_loss": 3.5228404536191658,
      "tokens_seen": 1467744256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003365613769623605,
      "loss": 2.7996,
      "theoretical_loss": 3.52282652725473,
      "tokens_seen": 1467809792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033654482693398904,
      "loss": 2.8017,
      "theoretical_loss": 3.5228126016861703,
      "tokens_seen": 1467875328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003365282769056176,
      "loss": 2.7261,
      "theoretical_loss": 3.522798676913405,
      "tokens_seen": 1467940864
    },
    {
      "epoch": 0.52,
      "objective/train/avg_token_score": 0.028022421523928642,
      "objective/train/avg_weight": 0.9670459032058716,
      "objective/train/docs_used": 832511,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.122345924377441,
      "objective/train/std_weight": 0.15765903890132904,
      "objective/train/theoretical_loss": 3.5227847529363547,
      "objective/train/tokens_used": 1488466400,
      "theoretical_loss": 3.5227847529363547,
      "tokens_seen": 1468006400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003365117268772461,
      "loss": 2.607,
      "theoretical_loss": 3.5227847529363547,
      "tokens_seen": 1468006400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003364951768488746,
      "loss": 2.589,
      "theoretical_loss": 3.522770829754937,
      "tokens_seen": 1468071936
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003364786268205031,
      "loss": 2.617,
      "theoretical_loss": 3.5227569073690717,
      "tokens_seen": 1468137472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003364620767921316,
      "loss": 2.7042,
      "theoretical_loss": 3.522742985778678,
      "tokens_seen": 1468203008
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033644552676376016,
      "loss": 2.7523,
      "theoretical_loss": 3.5227290649836736,
      "tokens_seen": 1468268544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003364289767353887,
      "loss": 2.7002,
      "theoretical_loss": 3.5227151449839793,
      "tokens_seen": 1468334080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003364124267070172,
      "loss": 2.9525,
      "theoretical_loss": 3.5227012257795125,
      "tokens_seen": 1468399616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003363958766786457,
      "loss": 2.6567,
      "theoretical_loss": 3.5226873073701936,
      "tokens_seen": 1468465152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033637932665027425,
      "loss": 2.5768,
      "theoretical_loss": 3.522673389755941,
      "tokens_seen": 1468530688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003363627766219028,
      "loss": 2.3108,
      "theoretical_loss": 3.522659472936674,
      "tokens_seen": 1468596224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003363462265935313,
      "loss": 2.569,
      "theoretical_loss": 3.522645556912312,
      "tokens_seen": 1468661760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003363296765651598,
      "loss": 2.6059,
      "theoretical_loss": 3.5226316416827737,
      "tokens_seen": 1468727296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033631312653678835,
      "loss": 2.5531,
      "theoretical_loss": 3.5226177272479786,
      "tokens_seen": 1468792832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003362965765084169,
      "loss": 2.7154,
      "theoretical_loss": 3.5226038136078452,
      "tokens_seen": 1468858368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033628002648004537,
      "loss": 2.8349,
      "theoretical_loss": 3.522589900762293,
      "tokens_seen": 1468923904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003362634764516739,
      "loss": 2.7195,
      "theoretical_loss": 3.522575988711242,
      "tokens_seen": 1468989440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033624692642330244,
      "loss": 2.6783,
      "theoretical_loss": 3.5225620774546105,
      "tokens_seen": 1469054976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000336230376394931,
      "loss": 2.5852,
      "theoretical_loss": 3.5225481669923173,
      "tokens_seen": 1469120512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033621382636655946,
      "loss": 2.803,
      "theoretical_loss": 3.522534257324282,
      "tokens_seen": 1469186048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000336197276338188,
      "loss": 2.5589,
      "theoretical_loss": 3.522520348450425,
      "tokens_seen": 1469251584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033618072630981654,
      "loss": 2.4707,
      "theoretical_loss": 3.5225064403706643,
      "tokens_seen": 1469317120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003361641762814451,
      "loss": 2.7021,
      "theoretical_loss": 3.522492533084919,
      "tokens_seen": 1469382656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033614762625307356,
      "loss": 2.3089,
      "theoretical_loss": 3.5224786265931085,
      "tokens_seen": 1469448192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003361310762247021,
      "loss": 2.6702,
      "theoretical_loss": 3.5224647208951527,
      "tokens_seen": 1469513728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033611452619633064,
      "loss": 2.5245,
      "theoretical_loss": 3.522450815990971,
      "tokens_seen": 1469579264
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003360979761679592,
      "loss": 2.621,
      "theoretical_loss": 3.5224369118804817,
      "tokens_seen": 1469644800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033608142613958766,
      "loss": 2.6757,
      "theoretical_loss": 3.5224230085636044,
      "tokens_seen": 1469710336
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0003360648761112162,
      "loss": 2.8931,
      "theoretical_loss": 3.522409106040259,
      "tokens_seen": 1469775872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033604832608284473,
      "loss": 2.6647,
      "theoretical_loss": 3.522395204310364,
      "tokens_seen": 1469841408
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033603177605447327,
      "loss": 2.6836,
      "theoretical_loss": 3.5223813033738396,
      "tokens_seen": 1469906944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00033601522602610175,
      "loss": 2.7125,
      "theoretical_loss": 3.5223674032306045,
      "tokens_seen": 1469972480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003359986759977303,
      "loss": 2.5247,
      "theoretical_loss": 3.5223535038805784,
      "tokens_seen": 1470038016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003359821259693588,
      "loss": 2.666,
      "theoretical_loss": 3.5223396053236806,
      "tokens_seen": 1470103552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003359655759409873,
      "loss": 2.6951,
      "theoretical_loss": 3.5223257075598307,
      "tokens_seen": 1470169088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033594902591261585,
      "loss": 2.4095,
      "theoretical_loss": 3.5223118105889473,
      "tokens_seen": 1470234624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033593247588424433,
      "loss": 2.5424,
      "theoretical_loss": 3.5222979144109505,
      "tokens_seen": 1470300160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033591592585587287,
      "loss": 2.4881,
      "theoretical_loss": 3.52228401902576,
      "tokens_seen": 1470365696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003358993758275014,
      "loss": 2.4116,
      "theoretical_loss": 3.5222701244332946,
      "tokens_seen": 1470431232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033588282579912994,
      "loss": 2.435,
      "theoretical_loss": 3.5222562306334737,
      "tokens_seen": 1470496768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033586627577075843,
      "loss": 2.7127,
      "theoretical_loss": 3.522242337626218,
      "tokens_seen": 1470562304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033584972574238697,
      "loss": 2.7356,
      "theoretical_loss": 3.5222284454114448,
      "tokens_seen": 1470627840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003358331757140155,
      "loss": 2.711,
      "theoretical_loss": 3.5222145539890755,
      "tokens_seen": 1470693376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033581662568564404,
      "loss": 2.7419,
      "theoretical_loss": 3.522200663359029,
      "tokens_seen": 1470758912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003358000756572725,
      "loss": 2.5383,
      "theoretical_loss": 3.5221867735212244,
      "tokens_seen": 1470824448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033578352562890106,
      "loss": 2.7246,
      "theoretical_loss": 3.5221728844755815,
      "tokens_seen": 1470889984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003357669756005296,
      "loss": 2.617,
      "theoretical_loss": 3.5221589962220197,
      "tokens_seen": 1470955520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033575042557215814,
      "loss": 2.5646,
      "theoretical_loss": 3.5221451087604594,
      "tokens_seen": 1471021056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003357338755437866,
      "loss": 2.6757,
      "theoretical_loss": 3.5221312220908185,
      "tokens_seen": 1471086592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033571732551541516,
      "loss": 2.7682,
      "theoretical_loss": 3.522117336213018,
      "tokens_seen": 1471152128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003357007754870437,
      "loss": 2.7903,
      "theoretical_loss": 3.522103451126977,
      "tokens_seen": 1471217664
    },
    {
      "epoch": 0.53,
      "objective/train/avg_token_score": 0.007857661694288254,
      "objective/train/avg_weight": 0.9907535314559937,
      "objective/train/docs_used": 834109,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4770326614379883,
      "objective/train/std_weight": 0.03588344529271126,
      "objective/train/theoretical_loss": 3.522089566832615,
      "objective/train/tokens_used": 1491743200,
      "theoretical_loss": 3.522089566832615,
      "tokens_seen": 1471283200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033568422545867223,
      "loss": 2.6408,
      "theoretical_loss": 3.522089566832615,
      "tokens_seen": 1471283200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003356676754303007,
      "loss": 2.6083,
      "theoretical_loss": 3.5220756833298514,
      "tokens_seen": 1471348736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033565112540192925,
      "loss": 2.6074,
      "theoretical_loss": 3.522061800618607,
      "tokens_seen": 1471414272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003356345753735578,
      "loss": 2.6781,
      "theoretical_loss": 3.5220479186987994,
      "tokens_seen": 1471479808
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033561802534518633,
      "loss": 2.953,
      "theoretical_loss": 3.52203403757035,
      "tokens_seen": 1471545344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003356014753168148,
      "loss": 2.5779,
      "theoretical_loss": 3.5220201572331775,
      "tokens_seen": 1471610880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033558492528844335,
      "loss": 2.4952,
      "theoretical_loss": 3.522006277687202,
      "tokens_seen": 1471676416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003355683752600719,
      "loss": 2.6065,
      "theoretical_loss": 3.521992398932343,
      "tokens_seen": 1471741952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003355518252317004,
      "loss": 2.6653,
      "theoretical_loss": 3.52197852096852,
      "tokens_seen": 1471807488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003355352752033289,
      "loss": 2.5606,
      "theoretical_loss": 3.521964643795653,
      "tokens_seen": 1471873024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033551872517495745,
      "loss": 2.8984,
      "theoretical_loss": 3.521950767413662,
      "tokens_seen": 1471938560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000335502175146586,
      "loss": 2.7822,
      "theoretical_loss": 3.521936891822466,
      "tokens_seen": 1472004096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033548562511821447,
      "loss": 2.7077,
      "theoretical_loss": 3.521923017021985,
      "tokens_seen": 1472069632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000335469075089843,
      "loss": 2.4602,
      "theoretical_loss": 3.521909143012139,
      "tokens_seen": 1472135168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033545252506147154,
      "loss": 2.6707,
      "theoretical_loss": 3.5218952697928474,
      "tokens_seen": 1472200704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003354359750331,
      "loss": 2.6608,
      "theoretical_loss": 3.5218813973640297,
      "tokens_seen": 1472266240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033541942500472856,
      "loss": 2.5906,
      "theoretical_loss": 3.5218675257256065,
      "tokens_seen": 1472331776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003354028749763571,
      "loss": 2.7808,
      "theoretical_loss": 3.5218536548774972,
      "tokens_seen": 1472397312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033538632494798564,
      "loss": 2.3975,
      "theoretical_loss": 3.5218397848196217,
      "tokens_seen": 1472462848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003353697749196141,
      "loss": 2.6709,
      "theoretical_loss": 3.521825915551899,
      "tokens_seen": 1472528384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033535322489124266,
      "loss": 2.8191,
      "theoretical_loss": 3.52181204707425,
      "tokens_seen": 1472593920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003353366748628712,
      "loss": 2.4747,
      "theoretical_loss": 3.5217981793865945,
      "tokens_seen": 1472659456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033532012483449973,
      "loss": 2.6053,
      "theoretical_loss": 3.5217843124888515,
      "tokens_seen": 1472724992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003353035748061282,
      "loss": 2.6155,
      "theoretical_loss": 3.5217704463809416,
      "tokens_seen": 1472790528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033528702477775675,
      "loss": 2.6991,
      "theoretical_loss": 3.5217565810627844,
      "tokens_seen": 1472856064
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003352704747493853,
      "loss": 2.5607,
      "theoretical_loss": 3.5217427165343,
      "tokens_seen": 1472921600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033525392472101383,
      "loss": 2.5711,
      "theoretical_loss": 3.521728852795407,
      "tokens_seen": 1472987136
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003352373746926423,
      "loss": 2.7239,
      "theoretical_loss": 3.5217149898460276,
      "tokens_seen": 1473052672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033522082466427085,
      "loss": 2.7481,
      "theoretical_loss": 3.52170112768608,
      "tokens_seen": 1473118208
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003352042746358994,
      "loss": 2.5856,
      "theoretical_loss": 3.5216872663154852,
      "tokens_seen": 1473183744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003351877246075279,
      "loss": 2.6184,
      "theoretical_loss": 3.5216734057341617,
      "tokens_seen": 1473249280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003351711745791564,
      "loss": 2.5652,
      "theoretical_loss": 3.521659545942031,
      "tokens_seen": 1473314816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033515462455078495,
      "loss": 2.7527,
      "theoretical_loss": 3.5216456869390123,
      "tokens_seen": 1473380352
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003351380745224135,
      "loss": 2.4816,
      "theoretical_loss": 3.5216318287250257,
      "tokens_seen": 1473445888
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000335121524494042,
      "loss": 2.5669,
      "theoretical_loss": 3.5216179712999907,
      "tokens_seen": 1473511424
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003351049744656705,
      "loss": 2.5982,
      "theoretical_loss": 3.5216041146638277,
      "tokens_seen": 1473576960
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033508842443729904,
      "loss": 2.3596,
      "theoretical_loss": 3.521590258816457,
      "tokens_seen": 1473642496
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003350718744089276,
      "loss": 2.8761,
      "theoretical_loss": 3.5215764037577983,
      "tokens_seen": 1473708032
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003350553243805561,
      "loss": 2.5887,
      "theoretical_loss": 3.521562549487772,
      "tokens_seen": 1473773568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003350387743521846,
      "loss": 2.5558,
      "theoretical_loss": 3.521548696006298,
      "tokens_seen": 1473839104
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033502222432381314,
      "loss": 2.7911,
      "theoretical_loss": 3.521534843313296,
      "tokens_seen": 1473904640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003350056742954416,
      "loss": 2.4912,
      "theoretical_loss": 3.5215209914086865,
      "tokens_seen": 1473970176
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033498912426707016,
      "loss": 2.6718,
      "theoretical_loss": 3.521507140292389,
      "tokens_seen": 1474035712
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003349725742386987,
      "loss": 2.7323,
      "theoretical_loss": 3.521493289964324,
      "tokens_seen": 1474101248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003349560242103272,
      "loss": 2.5946,
      "theoretical_loss": 3.521479440424412,
      "tokens_seen": 1474166784
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003349394741819557,
      "loss": 2.727,
      "theoretical_loss": 3.5214655916725714,
      "tokens_seen": 1474232320
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033492292415358426,
      "loss": 2.5315,
      "theoretical_loss": 3.5214517437087247,
      "tokens_seen": 1474297856
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003349063741252128,
      "loss": 2.7733,
      "theoretical_loss": 3.521437896532791,
      "tokens_seen": 1474363392
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003348898240968413,
      "loss": 2.8907,
      "theoretical_loss": 3.52142405014469,
      "tokens_seen": 1474428928
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003348732740684698,
      "loss": 2.6864,
      "theoretical_loss": 3.5214102045443423,
      "tokens_seen": 1474494464
    },
    {
      "epoch": 0.53,
      "objective/train/avg_token_score": 0.03047107346355915,
      "objective/train/avg_weight": 0.9641354084014893,
      "objective/train/docs_used": 835915,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0526270866394043,
      "objective/train/std_weight": 0.1337636113166809,
      "objective/train/theoretical_loss": 3.521396359731668,
      "objective/train/tokens_used": 1495020000,
      "theoretical_loss": 3.521396359731668,
      "tokens_seen": 1474560000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033485672404009835,
      "loss": 2.5428,
      "theoretical_loss": 3.521396359731668,
      "tokens_seen": 1474560000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003348401740117269,
      "loss": 2.8129,
      "theoretical_loss": 3.5213825157065877,
      "tokens_seen": 1474625536
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033482362398335537,
      "loss": 2.6694,
      "theoretical_loss": 3.5213686724690207,
      "tokens_seen": 1474691072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003348070739549839,
      "loss": 2.8596,
      "theoretical_loss": 3.521354830018888,
      "tokens_seen": 1474756608
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033479052392661245,
      "loss": 2.5627,
      "theoretical_loss": 3.5213409883561093,
      "tokens_seen": 1474822144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000334773973898241,
      "loss": 2.3477,
      "theoretical_loss": 3.521327147480605,
      "tokens_seen": 1474887680
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033475742386986947,
      "loss": 2.582,
      "theoretical_loss": 3.5213133073922958,
      "tokens_seen": 1474953216
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000334740873841498,
      "loss": 2.672,
      "theoretical_loss": 3.5212994680911014,
      "tokens_seen": 1475018752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033472432381312654,
      "loss": 2.8728,
      "theoretical_loss": 3.5212856295769424,
      "tokens_seen": 1475084288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003347077737847551,
      "loss": 2.56,
      "theoretical_loss": 3.5212717918497383,
      "tokens_seen": 1475149824
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033469122375638356,
      "loss": 2.7078,
      "theoretical_loss": 3.5212579549094105,
      "tokens_seen": 1475215360
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003346746737280121,
      "loss": 2.9806,
      "theoretical_loss": 3.5212441187558787,
      "tokens_seen": 1475280896
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033465812369964064,
      "loss": 2.7729,
      "theoretical_loss": 3.5212302833890634,
      "tokens_seen": 1475346432
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003346415736712692,
      "loss": 2.7857,
      "theoretical_loss": 3.5212164488088846,
      "tokens_seen": 1475411968
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033462502364289766,
      "loss": 2.7176,
      "theoretical_loss": 3.5212026150152633,
      "tokens_seen": 1475477504
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003346084736145262,
      "loss": 2.1155,
      "theoretical_loss": 3.521188782008119,
      "tokens_seen": 1475543040
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033459192358615473,
      "loss": 2.9462,
      "theoretical_loss": 3.5211749497873726,
      "tokens_seen": 1475608576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033457537355778327,
      "loss": 2.596,
      "theoretical_loss": 3.521161118352944,
      "tokens_seen": 1475674112
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033455882352941176,
      "loss": 2.951,
      "theoretical_loss": 3.5211472877047543,
      "tokens_seen": 1475739648
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003345422735010403,
      "loss": 2.7983,
      "theoretical_loss": 3.5211334578427236,
      "tokens_seen": 1475805184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033452572347266883,
      "loss": 2.6279,
      "theoretical_loss": 3.5211196287667716,
      "tokens_seen": 1475870720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003345091734442973,
      "loss": 2.5408,
      "theoretical_loss": 3.5211058004768203,
      "tokens_seen": 1475936256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033449262341592585,
      "loss": 2.8102,
      "theoretical_loss": 3.5210919729727888,
      "tokens_seen": 1476001792
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033447607338755434,
      "loss": 2.2621,
      "theoretical_loss": 3.521078146254598,
      "tokens_seen": 1476067328
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033445952335918287,
      "loss": 2.6831,
      "theoretical_loss": 3.521064320322168,
      "tokens_seen": 1476132864
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003344429733308114,
      "loss": 2.4857,
      "theoretical_loss": 3.5210504951754196,
      "tokens_seen": 1476198400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033442642330243995,
      "loss": 2.6859,
      "theoretical_loss": 3.521036670814273,
      "tokens_seen": 1476263936
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033440987327406843,
      "loss": 2.5457,
      "theoretical_loss": 3.5210228472386493,
      "tokens_seen": 1476329472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033439332324569697,
      "loss": 2.5446,
      "theoretical_loss": 3.521009024448468,
      "tokens_seen": 1476395008
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003343767732173255,
      "loss": 2.804,
      "theoretical_loss": 3.5209952024436504,
      "tokens_seen": 1476460544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033436022318895404,
      "loss": 2.8618,
      "theoretical_loss": 3.520981381224117,
      "tokens_seen": 1476526080
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003343436731605825,
      "loss": 2.7283,
      "theoretical_loss": 3.520967560789788,
      "tokens_seen": 1476591616
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033432712313221106,
      "loss": 2.6703,
      "theoretical_loss": 3.5209537411405845,
      "tokens_seen": 1476657152
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003343105731038396,
      "loss": 2.6306,
      "theoretical_loss": 3.520939922276426,
      "tokens_seen": 1476722688
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033429402307546814,
      "loss": 2.5328,
      "theoretical_loss": 3.520926104197234,
      "tokens_seen": 1476788224
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003342774730470966,
      "loss": 2.5801,
      "theoretical_loss": 3.520912286902929,
      "tokens_seen": 1476853760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033426092301872516,
      "loss": 2.5668,
      "theoretical_loss": 3.5208984703934316,
      "tokens_seen": 1476919296
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003342443729903537,
      "loss": 2.4466,
      "theoretical_loss": 3.520884654668662,
      "tokens_seen": 1476984832
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033422782296198224,
      "loss": 2.7676,
      "theoretical_loss": 3.5208708397285404,
      "tokens_seen": 1477050368
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003342112729336107,
      "loss": 2.7601,
      "theoretical_loss": 3.5208570255729885,
      "tokens_seen": 1477115904
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033419472290523926,
      "loss": 2.4992,
      "theoretical_loss": 3.5208432122019264,
      "tokens_seen": 1477181440
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003341781728768678,
      "loss": 2.6261,
      "theoretical_loss": 3.520829399615275,
      "tokens_seen": 1477246976
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033416162284849633,
      "loss": 2.594,
      "theoretical_loss": 3.5208155878129546,
      "tokens_seen": 1477312512
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003341450728201248,
      "loss": 2.7739,
      "theoretical_loss": 3.5208017767948863,
      "tokens_seen": 1477378048
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033412852279175335,
      "loss": 2.5697,
      "theoretical_loss": 3.5207879665609902,
      "tokens_seen": 1477443584
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003341119727633819,
      "loss": 2.8131,
      "theoretical_loss": 3.5207741571111875,
      "tokens_seen": 1477509120
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033409542273501043,
      "loss": 2.5449,
      "theoretical_loss": 3.520760348445399,
      "tokens_seen": 1477574656
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003340788727066389,
      "loss": 2.5191,
      "theoretical_loss": 3.520746540563544,
      "tokens_seen": 1477640192
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033406232267826745,
      "loss": 2.4775,
      "theoretical_loss": 3.520732733465546,
      "tokens_seen": 1477705728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000334045772649896,
      "loss": 2.6182,
      "theoretical_loss": 3.520718927151323,
      "tokens_seen": 1477771264
    },
    {
      "epoch": 0.53,
      "objective/train/avg_token_score": 0.00587688572704792,
      "objective/train/avg_weight": 0.9931036233901978,
      "objective/train/docs_used": 837471,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7184839248657227,
      "objective/train/std_weight": 0.04493831843137741,
      "objective/train/theoretical_loss": 3.520705121620797,
      "objective/train/tokens_used": 1498296800,
      "theoretical_loss": 3.520705121620797,
      "tokens_seen": 1477836800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003340292226215245,
      "loss": 2.6823,
      "theoretical_loss": 3.520705121620797,
      "tokens_seen": 1477836800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000334012672593153,
      "loss": 2.5797,
      "theoretical_loss": 3.5206913168738887,
      "tokens_seen": 1477902336
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033399612256478154,
      "loss": 2.4077,
      "theoretical_loss": 3.520677512910519,
      "tokens_seen": 1477967872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033397957253641003,
      "loss": 2.5627,
      "theoretical_loss": 3.5206637097306084,
      "tokens_seen": 1478033408
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033396302250803857,
      "loss": 2.7227,
      "theoretical_loss": 3.520649907334078,
      "tokens_seen": 1478098944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003339464724796671,
      "loss": 2.4875,
      "theoretical_loss": 3.5206361057208477,
      "tokens_seen": 1478164480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033392992245129564,
      "loss": 2.5624,
      "theoretical_loss": 3.5206223048908396,
      "tokens_seen": 1478230016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003339133724229241,
      "loss": 2.4642,
      "theoretical_loss": 3.520608504843974,
      "tokens_seen": 1478295552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033389682239455266,
      "loss": 2.32,
      "theoretical_loss": 3.520594705580171,
      "tokens_seen": 1478361088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003338802723661812,
      "loss": 2.5677,
      "theoretical_loss": 3.520580907099353,
      "tokens_seen": 1478426624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033386372233780974,
      "loss": 2.7108,
      "theoretical_loss": 3.5205671094014397,
      "tokens_seen": 1478492160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003338471723094382,
      "loss": 2.6411,
      "theoretical_loss": 3.5205533124863524,
      "tokens_seen": 1478557696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033383062228106676,
      "loss": 2.5005,
      "theoretical_loss": 3.520539516354012,
      "tokens_seen": 1478623232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003338140722526953,
      "loss": 2.6045,
      "theoretical_loss": 3.520525721004339,
      "tokens_seen": 1478688768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033379752222432383,
      "loss": 2.8233,
      "theoretical_loss": 3.520511926437255,
      "tokens_seen": 1478754304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003337809721959523,
      "loss": 2.5532,
      "theoretical_loss": 3.52049813265268,
      "tokens_seen": 1478819840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033376442216758085,
      "loss": 2.7772,
      "theoretical_loss": 3.520484339650536,
      "tokens_seen": 1478885376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003337478721392094,
      "loss": 2.6047,
      "theoretical_loss": 3.520470547430743,
      "tokens_seen": 1478950912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033373132211083793,
      "loss": 2.6325,
      "theoretical_loss": 3.520456755993222,
      "tokens_seen": 1479016448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003337147720824664,
      "loss": 2.7953,
      "theoretical_loss": 3.520442965337895,
      "tokens_seen": 1479081984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033369822205409495,
      "loss": 2.7173,
      "theoretical_loss": 3.520429175464682,
      "tokens_seen": 1479147520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003336816720257235,
      "loss": 2.6119,
      "theoretical_loss": 3.5204153863735046,
      "tokens_seen": 1479213056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000333665121997352,
      "loss": 2.5612,
      "theoretical_loss": 3.5204015980642835,
      "tokens_seen": 1479278592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003336485719689805,
      "loss": 2.478,
      "theoretical_loss": 3.5203878105369397,
      "tokens_seen": 1479344128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033363202194060905,
      "loss": 2.4551,
      "theoretical_loss": 3.520374023791394,
      "tokens_seen": 1479409664
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003336154719122376,
      "loss": 2.6271,
      "theoretical_loss": 3.520360237827568,
      "tokens_seen": 1479475200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003335989218838661,
      "loss": 2.6558,
      "theoretical_loss": 3.5203464526453825,
      "tokens_seen": 1479540736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003335823718554946,
      "loss": 2.6427,
      "theoretical_loss": 3.5203326682447584,
      "tokens_seen": 1479606272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033356582182712314,
      "loss": 2.7578,
      "theoretical_loss": 3.520318884625617,
      "tokens_seen": 1479671808
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003335492717987517,
      "loss": 2.4452,
      "theoretical_loss": 3.520305101787879,
      "tokens_seen": 1479737344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003335327217703802,
      "loss": 2.7698,
      "theoretical_loss": 3.5202913197314665,
      "tokens_seen": 1479802880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003335161717420087,
      "loss": 2.7584,
      "theoretical_loss": 3.520277538456299,
      "tokens_seen": 1479868416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003334996217136372,
      "loss": 2.6988,
      "theoretical_loss": 3.520263757962299,
      "tokens_seen": 1479933952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003334830716852657,
      "loss": 2.6151,
      "theoretical_loss": 3.520249978249387,
      "tokens_seen": 1479999488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033346652165689426,
      "loss": 2.4812,
      "theoretical_loss": 3.5202361993174844,
      "tokens_seen": 1480065024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003334499716285228,
      "loss": 2.6732,
      "theoretical_loss": 3.5202224211665123,
      "tokens_seen": 1480130560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003334334216001513,
      "loss": 2.6991,
      "theoretical_loss": 3.5202086437963915,
      "tokens_seen": 1480196096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003334168715717798,
      "loss": 2.6411,
      "theoretical_loss": 3.520194867207044,
      "tokens_seen": 1480261632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033340032154340835,
      "loss": 2.3675,
      "theoretical_loss": 3.52018109139839,
      "tokens_seen": 1480327168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003333837715150369,
      "loss": 2.6571,
      "theoretical_loss": 3.520167316370351,
      "tokens_seen": 1480392704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003333672214866654,
      "loss": 2.6004,
      "theoretical_loss": 3.520153542122849,
      "tokens_seen": 1480458240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003333506714582939,
      "loss": 2.6909,
      "theoretical_loss": 3.5201397686558042,
      "tokens_seen": 1480523776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033333412142992245,
      "loss": 2.719,
      "theoretical_loss": 3.5201259959691384,
      "tokens_seen": 1480589312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000333317571401551,
      "loss": 2.814,
      "theoretical_loss": 3.5201122240627725,
      "tokens_seen": 1480654848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033330102137317947,
      "loss": 2.472,
      "theoretical_loss": 3.520098452936628,
      "tokens_seen": 1480720384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000333284471344808,
      "loss": 2.5375,
      "theoretical_loss": 3.5200846825906265,
      "tokens_seen": 1480785920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033326792131643655,
      "loss": 2.6677,
      "theoretical_loss": 3.5200709130246883,
      "tokens_seen": 1480851456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003332513712880651,
      "loss": 2.6769,
      "theoretical_loss": 3.5200571442387356,
      "tokens_seen": 1480916992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033323482125969357,
      "loss": 2.7587,
      "theoretical_loss": 3.5200433762326897,
      "tokens_seen": 1480982528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003332182712313221,
      "loss": 2.8292,
      "theoretical_loss": 3.520029609006471,
      "tokens_seen": 1481048064
    },
    {
      "epoch": 0.53,
      "objective/train/avg_token_score": 0.023678729310631752,
      "objective/train/avg_weight": 0.9721649289131165,
      "objective/train/docs_used": 839384,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.652683734893799,
      "objective/train/std_weight": 0.09892963618040085,
      "objective/train/theoretical_loss": 3.520015842560001,
      "objective/train/tokens_used": 1501573600,
      "theoretical_loss": 3.520015842560001,
      "tokens_seen": 1481113600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033320172120295064,
      "loss": 2.7598,
      "theoretical_loss": 3.520015842560001,
      "tokens_seen": 1481113600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003331851711745792,
      "loss": 2.712,
      "theoretical_loss": 3.5200020768932028,
      "tokens_seen": 1481179136
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033316862114620766,
      "loss": 2.6626,
      "theoretical_loss": 3.519988312005996,
      "tokens_seen": 1481244672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003331520711178362,
      "loss": 2.5957,
      "theoretical_loss": 3.5199745478983013,
      "tokens_seen": 1481310208
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033313552108946474,
      "loss": 2.6163,
      "theoretical_loss": 3.519960784570042,
      "tokens_seen": 1481375744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003331189710610933,
      "loss": 2.7319,
      "theoretical_loss": 3.519947022021139,
      "tokens_seen": 1481441280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033310242103272176,
      "loss": 2.4825,
      "theoretical_loss": 3.5199332602515128,
      "tokens_seen": 1481506816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003330858710043503,
      "loss": 2.7396,
      "theoretical_loss": 3.5199194992610856,
      "tokens_seen": 1481572352
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033306932097597883,
      "loss": 2.5611,
      "theoretical_loss": 3.5199057390497783,
      "tokens_seen": 1481637888
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033305277094760737,
      "loss": 2.4457,
      "theoretical_loss": 3.5198919796175128,
      "tokens_seen": 1481703424
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033303622091923585,
      "loss": 2.7142,
      "theoretical_loss": 3.51987822096421,
      "tokens_seen": 1481768960
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033301967089086434,
      "loss": 2.5455,
      "theoretical_loss": 3.5198644630897915,
      "tokens_seen": 1481834496
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003330031208624929,
      "loss": 2.5933,
      "theoretical_loss": 3.5198507059941795,
      "tokens_seen": 1481900032
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003329865708341214,
      "loss": 2.4651,
      "theoretical_loss": 3.5198369496772948,
      "tokens_seen": 1481965568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033297002080574995,
      "loss": 2.655,
      "theoretical_loss": 3.5198231941390588,
      "tokens_seen": 1482031104
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033295347077737843,
      "loss": 2.6648,
      "theoretical_loss": 3.5198094393793933,
      "tokens_seen": 1482096640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033293692074900697,
      "loss": 2.664,
      "theoretical_loss": 3.5197956853982193,
      "tokens_seen": 1482162176
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003329203707206355,
      "loss": 2.2043,
      "theoretical_loss": 3.5197819321954595,
      "tokens_seen": 1482227712
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033290382069226405,
      "loss": 2.7677,
      "theoretical_loss": 3.5197681797710336,
      "tokens_seen": 1482293248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033288727066389253,
      "loss": 2.5744,
      "theoretical_loss": 3.519754428124865,
      "tokens_seen": 1482358784
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033287072063552107,
      "loss": 2.7358,
      "theoretical_loss": 3.5197406772568742,
      "tokens_seen": 1482424320
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003328541706071496,
      "loss": 2.6623,
      "theoretical_loss": 3.519726927166983,
      "tokens_seen": 1482489856
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033283762057877814,
      "loss": 2.8184,
      "theoretical_loss": 3.519713177855113,
      "tokens_seen": 1482555392
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003328210705504066,
      "loss": 2.6703,
      "theoretical_loss": 3.519699429321186,
      "tokens_seen": 1482620928
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033280452052203516,
      "loss": 2.6429,
      "theoretical_loss": 3.519685681565123,
      "tokens_seen": 1482686464
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003327879704936637,
      "loss": 2.6792,
      "theoretical_loss": 3.519671934586846,
      "tokens_seen": 1482752000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033277142046529224,
      "loss": 2.675,
      "theoretical_loss": 3.519658188386277,
      "tokens_seen": 1482817536
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003327548704369207,
      "loss": 2.6987,
      "theoretical_loss": 3.5196444429633367,
      "tokens_seen": 1482883072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033273832040854926,
      "loss": 2.7778,
      "theoretical_loss": 3.5196306983179477,
      "tokens_seen": 1482948608
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003327217703801778,
      "loss": 2.4899,
      "theoretical_loss": 3.519616954450031,
      "tokens_seen": 1483014144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033270522035180633,
      "loss": 2.7211,
      "theoretical_loss": 3.5196032113595086,
      "tokens_seen": 1483079680
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033268867032343487,
      "loss": 2.802,
      "theoretical_loss": 3.519589469046302,
      "tokens_seen": 1483145216
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033267212029506336,
      "loss": 2.6945,
      "theoretical_loss": 3.519575727510333,
      "tokens_seen": 1483210752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003326555702666919,
      "loss": 2.6988,
      "theoretical_loss": 3.5195619867515235,
      "tokens_seen": 1483276288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033263902023832043,
      "loss": 2.5441,
      "theoretical_loss": 3.5195482467697943,
      "tokens_seen": 1483341824
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033262247020994897,
      "loss": 2.5959,
      "theoretical_loss": 3.5195345075650684,
      "tokens_seen": 1483407360
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033260592018157745,
      "loss": 2.5826,
      "theoretical_loss": 3.5195207691372667,
      "tokens_seen": 1483472896
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000332589370153206,
      "loss": 2.807,
      "theoretical_loss": 3.5195070314863113,
      "tokens_seen": 1483538432
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003325728201248345,
      "loss": 2.6533,
      "theoretical_loss": 3.519493294612124,
      "tokens_seen": 1483603968
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033255627009646306,
      "loss": 2.9143,
      "theoretical_loss": 3.519479558514626,
      "tokens_seen": 1483669504
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033253972006809155,
      "loss": 2.6218,
      "theoretical_loss": 3.51946582319374,
      "tokens_seen": 1483735040
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033252317003972003,
      "loss": 2.6116,
      "theoretical_loss": 3.519452088649387,
      "tokens_seen": 1483800576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033250662001134857,
      "loss": 2.6231,
      "theoretical_loss": 3.5194383548814896,
      "tokens_seen": 1483866112
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003324900699829771,
      "loss": 2.6375,
      "theoretical_loss": 3.5194246218899683,
      "tokens_seen": 1483931648
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033247351995460564,
      "loss": 2.7192,
      "theoretical_loss": 3.519410889674746,
      "tokens_seen": 1483997184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003324569699262341,
      "loss": 2.7072,
      "theoretical_loss": 3.519397158235745,
      "tokens_seen": 1484062720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033244041989786266,
      "loss": 2.8422,
      "theoretical_loss": 3.5193834275728855,
      "tokens_seen": 1484128256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003324238698694912,
      "loss": 2.4361,
      "theoretical_loss": 3.5193696976860904,
      "tokens_seen": 1484193792
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033240731984111974,
      "loss": 2.5573,
      "theoretical_loss": 3.519355968575282,
      "tokens_seen": 1484259328
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003323907698127482,
      "loss": 2.8271,
      "theoretical_loss": 3.5193422402403813,
      "tokens_seen": 1484324864
    },
    {
      "epoch": 0.53,
      "objective/train/avg_token_score": 0.017632870003581047,
      "objective/train/avg_weight": 0.9792490601539612,
      "objective/train/docs_used": 841281,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.784003496170044,
      "objective/train/std_weight": 0.08513686060905457,
      "objective/train/theoretical_loss": 3.5193285126813105,
      "objective/train/tokens_used": 1504850400,
      "theoretical_loss": 3.5193285126813105,
      "tokens_seen": 1484390400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033237421978437676,
      "loss": 2.638,
      "theoretical_loss": 3.5193285126813105,
      "tokens_seen": 1484390400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003323576697560053,
      "loss": 2.7283,
      "theoretical_loss": 3.5193147858979916,
      "tokens_seen": 1484455936
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033234111972763384,
      "loss": 2.8584,
      "theoretical_loss": 3.5193010598903465,
      "tokens_seen": 1484521472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003323245696992623,
      "loss": 2.7086,
      "theoretical_loss": 3.519287334658297,
      "tokens_seen": 1484587008
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033230801967089086,
      "loss": 2.5777,
      "theoretical_loss": 3.5192736102017648,
      "tokens_seen": 1484652544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003322914696425194,
      "loss": 2.8938,
      "theoretical_loss": 3.519259886520673,
      "tokens_seen": 1484718080
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033227491961414793,
      "loss": 2.9036,
      "theoretical_loss": 3.519246163614942,
      "tokens_seen": 1484783616
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003322583695857764,
      "loss": 2.759,
      "theoretical_loss": 3.519232441484495,
      "tokens_seen": 1484849152
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033224181955740495,
      "loss": 2.7156,
      "theoretical_loss": 3.519218720129253,
      "tokens_seen": 1484914688
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003322252695290335,
      "loss": 2.7138,
      "theoretical_loss": 3.519204999549139,
      "tokens_seen": 1484980224
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000332208719500662,
      "loss": 2.5358,
      "theoretical_loss": 3.519191279744074,
      "tokens_seen": 1485045760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003321921694722905,
      "loss": 2.8685,
      "theoretical_loss": 3.519177560713981,
      "tokens_seen": 1485111296
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033217561944391905,
      "loss": 2.6296,
      "theoretical_loss": 3.5191638424587808,
      "tokens_seen": 1485176832
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003321590694155476,
      "loss": 2.6895,
      "theoretical_loss": 3.519150124978397,
      "tokens_seen": 1485242368
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003321425193871761,
      "loss": 2.6842,
      "theoretical_loss": 3.5191364082727503,
      "tokens_seen": 1485307904
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003321259693588046,
      "loss": 2.6126,
      "theoretical_loss": 3.519122692341764,
      "tokens_seen": 1485373440
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033210941933043314,
      "loss": 2.6328,
      "theoretical_loss": 3.5191089771853585,
      "tokens_seen": 1485438976
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003320928693020617,
      "loss": 2.6077,
      "theoretical_loss": 3.5190952628034573,
      "tokens_seen": 1485504512
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003320763192736902,
      "loss": 2.709,
      "theoretical_loss": 3.519081549195982,
      "tokens_seen": 1485570048
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003320597692453187,
      "loss": 2.5102,
      "theoretical_loss": 3.519067836362855,
      "tokens_seen": 1485635584
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003320432192169472,
      "loss": 2.6622,
      "theoretical_loss": 3.519054124303998,
      "tokens_seen": 1485701120
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003320266691885757,
      "loss": 2.6783,
      "theoretical_loss": 3.519040413019334,
      "tokens_seen": 1485766656
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033201011916020426,
      "loss": 2.7813,
      "theoretical_loss": 3.519026702508783,
      "tokens_seen": 1485832192
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003319935691318328,
      "loss": 2.5539,
      "theoretical_loss": 3.5190129927722698,
      "tokens_seen": 1485897728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003319770191034613,
      "loss": 2.7192,
      "theoretical_loss": 3.518999283809715,
      "tokens_seen": 1485963264
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003319604690750898,
      "loss": 2.4754,
      "theoretical_loss": 3.5189855756210413,
      "tokens_seen": 1486028800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033194391904671836,
      "loss": 2.5619,
      "theoretical_loss": 3.5189718682061706,
      "tokens_seen": 1486094336
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003319273690183469,
      "loss": 2.59,
      "theoretical_loss": 3.518958161565025,
      "tokens_seen": 1486159872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003319108189899754,
      "loss": 2.6465,
      "theoretical_loss": 3.518944455697527,
      "tokens_seen": 1486225408
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003318942689616039,
      "loss": 2.6195,
      "theoretical_loss": 3.5189307506035994,
      "tokens_seen": 1486290944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033187771893323245,
      "loss": 2.3864,
      "theoretical_loss": 3.518917046283163,
      "tokens_seen": 1486356480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000331861168904861,
      "loss": 2.8222,
      "theoretical_loss": 3.5189033427361416,
      "tokens_seen": 1486422016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003318446188764895,
      "loss": 2.5876,
      "theoretical_loss": 3.518889639962456,
      "tokens_seen": 1486487552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000331828068848118,
      "loss": 2.438,
      "theoretical_loss": 3.51887593796203,
      "tokens_seen": 1486553088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033181151881974655,
      "loss": 2.8388,
      "theoretical_loss": 3.5188622367347846,
      "tokens_seen": 1486618624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003317949687913751,
      "loss": 2.8147,
      "theoretical_loss": 3.5188485362806423,
      "tokens_seen": 1486684160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033177841876300357,
      "loss": 2.7184,
      "theoretical_loss": 3.5188348365995257,
      "tokens_seen": 1486749696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003317618687346321,
      "loss": 2.4879,
      "theoretical_loss": 3.5188211376913574,
      "tokens_seen": 1486815232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033174531870626064,
      "loss": 2.6471,
      "theoretical_loss": 3.518807439556059,
      "tokens_seen": 1486880768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003317287686778892,
      "loss": 2.7967,
      "theoretical_loss": 3.5187937421935533,
      "tokens_seen": 1486946304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033171221864951767,
      "loss": 2.6869,
      "theoretical_loss": 3.518780045603763,
      "tokens_seen": 1487011840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003316956686211462,
      "loss": 2.5078,
      "theoretical_loss": 3.5187663497866097,
      "tokens_seen": 1487077376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033167911859277474,
      "loss": 2.9607,
      "theoretical_loss": 3.518752654742016,
      "tokens_seen": 1487142912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003316625685644033,
      "loss": 2.3859,
      "theoretical_loss": 3.5187389604699044,
      "tokens_seen": 1487208448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033164601853603176,
      "loss": 2.433,
      "theoretical_loss": 3.518725266970197,
      "tokens_seen": 1487273984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003316294685076603,
      "loss": 2.6763,
      "theoretical_loss": 3.5187115742428166,
      "tokens_seen": 1487339520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033161291847928884,
      "loss": 2.7793,
      "theoretical_loss": 3.5186978822876855,
      "tokens_seen": 1487405056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003315963684509174,
      "loss": 2.7799,
      "theoretical_loss": 3.518684191104726,
      "tokens_seen": 1487470592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033157981842254586,
      "loss": 2.8523,
      "theoretical_loss": 3.518670500693861,
      "tokens_seen": 1487536128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033156326839417434,
      "loss": 2.5318,
      "theoretical_loss": 3.518656811055012,
      "tokens_seen": 1487601664
    },
    {
      "epoch": 0.53,
      "objective/train/avg_token_score": 0.025652356445789337,
      "objective/train/avg_weight": 0.9698153138160706,
      "objective/train/docs_used": 843256,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3580434322357178,
      "objective/train/std_weight": 0.11254582554101944,
      "objective/train/theoretical_loss": 3.5186431221881023,
      "objective/train/tokens_used": 1508127200,
      "theoretical_loss": 3.5186431221881023,
      "tokens_seen": 1487667200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003315467183658029,
      "loss": 2.7273,
      "theoretical_loss": 3.5186431221881023,
      "tokens_seen": 1487667200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003315301683374314,
      "loss": 2.6846,
      "theoretical_loss": 3.5186294340930537,
      "tokens_seen": 1487732736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033151361830905995,
      "loss": 2.5575,
      "theoretical_loss": 3.5186157467697896,
      "tokens_seen": 1487798272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003314970682806885,
      "loss": 2.6906,
      "theoretical_loss": 3.5186020602182317,
      "tokens_seen": 1487863808
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000331480518252317,
      "loss": 2.7018,
      "theoretical_loss": 3.518588374438303,
      "tokens_seen": 1487929344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003314639682239455,
      "loss": 2.5931,
      "theoretical_loss": 3.5185746894299257,
      "tokens_seen": 1487994880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033144741819557405,
      "loss": 2.7838,
      "theoretical_loss": 3.518561005193022,
      "tokens_seen": 1488060416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003314308681672026,
      "loss": 2.5399,
      "theoretical_loss": 3.5185473217275147,
      "tokens_seen": 1488125952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033141431813883107,
      "loss": 2.8557,
      "theoretical_loss": 3.5185336390333273,
      "tokens_seen": 1488191488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003313977681104596,
      "loss": 2.6877,
      "theoretical_loss": 3.5185199571103807,
      "tokens_seen": 1488257024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033138121808208815,
      "loss": 2.6659,
      "theoretical_loss": 3.518506275958599,
      "tokens_seen": 1488322560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003313646680537167,
      "loss": 2.5677,
      "theoretical_loss": 3.5184925955779036,
      "tokens_seen": 1488388096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033134811802534517,
      "loss": 2.6522,
      "theoretical_loss": 3.5184789159682177,
      "tokens_seen": 1488453632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003313315679969737,
      "loss": 2.8748,
      "theoretical_loss": 3.518465237129464,
      "tokens_seen": 1488519168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033131501796860224,
      "loss": 2.6516,
      "theoretical_loss": 3.518451559061565,
      "tokens_seen": 1488584704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003312984679402308,
      "loss": 2.7223,
      "theoretical_loss": 3.5184378817644424,
      "tokens_seen": 1488650240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033128191791185926,
      "loss": 2.498,
      "theoretical_loss": 3.5184242052380204,
      "tokens_seen": 1488715776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003312653678834878,
      "loss": 2.6806,
      "theoretical_loss": 3.5184105294822206,
      "tokens_seen": 1488781312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033124881785511634,
      "loss": 2.4932,
      "theoretical_loss": 3.5183968544969657,
      "tokens_seen": 1488846848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003312322678267449,
      "loss": 2.7303,
      "theoretical_loss": 3.5183831802821794,
      "tokens_seen": 1488912384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033121571779837336,
      "loss": 2.7687,
      "theoretical_loss": 3.518369506837783,
      "tokens_seen": 1488977920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003311991677700019,
      "loss": 2.6538,
      "theoretical_loss": 3.5183558341637,
      "tokens_seen": 1489043456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033118261774163043,
      "loss": 2.4757,
      "theoretical_loss": 3.5183421622598527,
      "tokens_seen": 1489108992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033116606771325897,
      "loss": 2.6834,
      "theoretical_loss": 3.518328491126164,
      "tokens_seen": 1489174528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033114951768488745,
      "loss": 2.5602,
      "theoretical_loss": 3.5183148207625567,
      "tokens_seen": 1489240064
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000331132967656516,
      "loss": 2.3939,
      "theoretical_loss": 3.518301151168954,
      "tokens_seen": 1489305600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033111641762814453,
      "loss": 2.5014,
      "theoretical_loss": 3.5182874823452774,
      "tokens_seen": 1489371136
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033109986759977307,
      "loss": 2.3685,
      "theoretical_loss": 3.518273814291451,
      "tokens_seen": 1489436672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033108331757140155,
      "loss": 2.7988,
      "theoretical_loss": 3.518260147007396,
      "tokens_seen": 1489502208
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033106676754303003,
      "loss": 2.6705,
      "theoretical_loss": 3.5182464804930373,
      "tokens_seen": 1489567744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033105021751465857,
      "loss": 2.5935,
      "theoretical_loss": 3.5182328147482957,
      "tokens_seen": 1489633280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003310336674862871,
      "loss": 2.8207,
      "theoretical_loss": 3.5182191497730955,
      "tokens_seen": 1489698816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033101711745791565,
      "loss": 2.6215,
      "theoretical_loss": 3.5182054855673583,
      "tokens_seen": 1489764352
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033100056742954413,
      "loss": 2.5536,
      "theoretical_loss": 3.5181918221310076,
      "tokens_seen": 1489829888
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033098401740117267,
      "loss": 2.7094,
      "theoretical_loss": 3.5181781594639663,
      "tokens_seen": 1489895424
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003309674673728012,
      "loss": 2.5645,
      "theoretical_loss": 3.518164497566157,
      "tokens_seen": 1489960960
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033095091734442974,
      "loss": 2.6425,
      "theoretical_loss": 3.5181508364375023,
      "tokens_seen": 1490026496
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003309343673160582,
      "loss": 2.6294,
      "theoretical_loss": 3.518137176077926,
      "tokens_seen": 1490092032
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033091781728768676,
      "loss": 2.908,
      "theoretical_loss": 3.51812351648735,
      "tokens_seen": 1490157568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003309012672593153,
      "loss": 2.7354,
      "theoretical_loss": 3.518109857665697,
      "tokens_seen": 1490223104
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033088471723094384,
      "loss": 2.7142,
      "theoretical_loss": 3.518096199612891,
      "tokens_seen": 1490288640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003308681672025723,
      "loss": 2.8487,
      "theoretical_loss": 3.5180825423288544,
      "tokens_seen": 1490354176
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033085161717420086,
      "loss": 2.9853,
      "theoretical_loss": 3.51806888581351,
      "tokens_seen": 1490419712
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003308350671458294,
      "loss": 2.9333,
      "theoretical_loss": 3.518055230066781,
      "tokens_seen": 1490485248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033081851711745793,
      "loss": 2.7043,
      "theoretical_loss": 3.51804157508859,
      "tokens_seen": 1490550784
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003308019670890864,
      "loss": 2.7,
      "theoretical_loss": 3.5180279208788603,
      "tokens_seen": 1490616320
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033078541706071495,
      "loss": 2.8138,
      "theoretical_loss": 3.5180142674375148,
      "tokens_seen": 1490681856
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003307688670323435,
      "loss": 2.9283,
      "theoretical_loss": 3.5180006147644756,
      "tokens_seen": 1490747392
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033075231700397203,
      "loss": 2.6083,
      "theoretical_loss": 3.517986962859667,
      "tokens_seen": 1490812928
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003307357669756005,
      "loss": 2.6849,
      "theoretical_loss": 3.5179733117230123,
      "tokens_seen": 1490878464
    },
    {
      "epoch": 0.53,
      "objective/train/avg_token_score": 0.019977184012532234,
      "objective/train/avg_weight": 0.9765017628669739,
      "objective/train/docs_used": 844947,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8573007583618164,
      "objective/train/std_weight": 0.08870706707239151,
      "objective/train/theoretical_loss": 3.517959661354433,
      "objective/train/tokens_used": 1511404000,
      "theoretical_loss": 3.517959661354433,
      "tokens_seen": 1490944000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033071921694722905,
      "loss": 2.8026,
      "theoretical_loss": 3.517959661354433,
      "tokens_seen": 1490944000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003307026669188576,
      "loss": 2.4616,
      "theoretical_loss": 3.5179460117538524,
      "tokens_seen": 1491009536
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003306861168904861,
      "loss": 2.6703,
      "theoretical_loss": 3.5179323629211945,
      "tokens_seen": 1491075072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003306695668621146,
      "loss": 2.5941,
      "theoretical_loss": 3.517918714856382,
      "tokens_seen": 1491140608
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033065301683374315,
      "loss": 2.5008,
      "theoretical_loss": 3.5179050675593375,
      "tokens_seen": 1491206144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003306364668053717,
      "loss": 2.8066,
      "theoretical_loss": 3.517891421029984,
      "tokens_seen": 1491271680
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003306199167770002,
      "loss": 2.6663,
      "theoretical_loss": 3.5178777752682455,
      "tokens_seen": 1491337216
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003306033667486287,
      "loss": 2.7731,
      "theoretical_loss": 3.5178641302740443,
      "tokens_seen": 1491402752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033058681672025724,
      "loss": 2.6419,
      "theoretical_loss": 3.517850486047304,
      "tokens_seen": 1491468288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003305702666918857,
      "loss": 2.6663,
      "theoretical_loss": 3.517836842587947,
      "tokens_seen": 1491533824
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033055371666351426,
      "loss": 2.6829,
      "theoretical_loss": 3.517823199895897,
      "tokens_seen": 1491599360
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003305371666351428,
      "loss": 2.7286,
      "theoretical_loss": 3.517809557971077,
      "tokens_seen": 1491664896
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003305206166067713,
      "loss": 2.5746,
      "theoretical_loss": 3.51779591681341,
      "tokens_seen": 1491730432
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003305040665783998,
      "loss": 2.4684,
      "theoretical_loss": 3.5177822764228193,
      "tokens_seen": 1491795968
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033048751655002836,
      "loss": 2.8027,
      "theoretical_loss": 3.5177686367992287,
      "tokens_seen": 1491861504
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003304709665216569,
      "loss": 2.715,
      "theoretical_loss": 3.51775499794256,
      "tokens_seen": 1491927040
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003304544164932854,
      "loss": 2.8947,
      "theoretical_loss": 3.517741359852738,
      "tokens_seen": 1491992576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003304378664649139,
      "loss": 2.4574,
      "theoretical_loss": 3.517727722529684,
      "tokens_seen": 1492058112
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033042131643654246,
      "loss": 2.8376,
      "theoretical_loss": 3.517714085973323,
      "tokens_seen": 1492123648
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000330404766408171,
      "loss": 2.8034,
      "theoretical_loss": 3.517700450183577,
      "tokens_seen": 1492189184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003303882163797995,
      "loss": 2.843,
      "theoretical_loss": 3.5176868151603697,
      "tokens_seen": 1492254720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000330371666351428,
      "loss": 2.5974,
      "theoretical_loss": 3.5176731809036252,
      "tokens_seen": 1492320256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033035511632305655,
      "loss": 2.8639,
      "theoretical_loss": 3.517659547413265,
      "tokens_seen": 1492385792
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003303385662946851,
      "loss": 2.8726,
      "theoretical_loss": 3.517645914689214,
      "tokens_seen": 1492451328
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033032201626631357,
      "loss": 2.7435,
      "theoretical_loss": 3.5176322827313946,
      "tokens_seen": 1492516864
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003303054662379421,
      "loss": 2.6979,
      "theoretical_loss": 3.5176186515397303,
      "tokens_seen": 1492582400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033028891620957065,
      "loss": 2.6546,
      "theoretical_loss": 3.517605021114144,
      "tokens_seen": 1492647936
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003302723661811992,
      "loss": 2.7964,
      "theoretical_loss": 3.51759139145456,
      "tokens_seen": 1492713472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033025581615282767,
      "loss": 2.4827,
      "theoretical_loss": 3.5175777625609004,
      "tokens_seen": 1492779008
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003302392661244562,
      "loss": 2.6607,
      "theoretical_loss": 3.5175641344330892,
      "tokens_seen": 1492844544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033022271609608474,
      "loss": 2.7557,
      "theoretical_loss": 3.51755050707105,
      "tokens_seen": 1492910080
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033022271609608474,
      "loss": 3.048,
      "theoretical_loss": 3.517536880474706,
      "tokens_seen": 1492975616
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003302061660677133,
      "loss": 2.4301,
      "theoretical_loss": 3.51752325464398,
      "tokens_seen": 1493041152
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003302061660677133,
      "loss": 2.8585,
      "theoretical_loss": 3.5175096295787958,
      "tokens_seen": 1493106688
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033018961603934176,
      "loss": 2.6447,
      "theoretical_loss": 3.517496005279077,
      "tokens_seen": 1493172224
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003301730660109703,
      "loss": 2.796,
      "theoretical_loss": 3.5174823817447467,
      "tokens_seen": 1493237760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033015651598259884,
      "loss": 3.9041,
      "theoretical_loss": 3.5174687589757285,
      "tokens_seen": 1493303296
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033015651598259884,
      "loss": 10.2201,
      "theoretical_loss": 3.517455136971946,
      "tokens_seen": 1493368832
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003301399659542274,
      "loss": 2.7738,
      "theoretical_loss": 3.5174415157333216,
      "tokens_seen": 1493434368
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033012341592585586,
      "loss": 2.8677,
      "theoretical_loss": 3.5174278952597806,
      "tokens_seen": 1493499904
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003301068658974844,
      "loss": 2.4692,
      "theoretical_loss": 3.5174142755512445,
      "tokens_seen": 1493565440
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003300903158691129,
      "loss": 2.8932,
      "theoretical_loss": 3.517400656607638,
      "tokens_seen": 1493630976
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003300737658407414,
      "loss": 2.8126,
      "theoretical_loss": 3.5173870384288843,
      "tokens_seen": 1493696512
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00033005721581236996,
      "loss": 2.5438,
      "theoretical_loss": 3.5173734210149066,
      "tokens_seen": 1493762048
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003300406657839985,
      "loss": 2.8639,
      "theoretical_loss": 3.5173598043656282,
      "tokens_seen": 1493827584
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000330024115755627,
      "loss": 2.4373,
      "theoretical_loss": 3.517346188480974,
      "tokens_seen": 1493893120
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003300075657272555,
      "loss": 2.8584,
      "theoretical_loss": 3.517332573360866,
      "tokens_seen": 1493958656
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032999101569888405,
      "loss": 2.4722,
      "theoretical_loss": 3.517318959005228,
      "tokens_seen": 1494024192
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003299744656705126,
      "loss": 2.6289,
      "theoretical_loss": 3.517305345413984,
      "tokens_seen": 1494089728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003299579156421411,
      "loss": 2.8013,
      "theoretical_loss": 3.517291732587058,
      "tokens_seen": 1494155264
    },
    {
      "epoch": 0.53,
      "objective/train/avg_token_score": 0.019158657640218735,
      "objective/train/avg_weight": 0.9774720072746277,
      "objective/train/docs_used": 846677,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.796180248260498,
      "objective/train/std_weight": 0.10704844444990158,
      "objective/train/theoretical_loss": 3.517278120524372,
      "objective/train/tokens_used": 1514680800,
      "theoretical_loss": 3.517278120524372,
      "tokens_seen": 1494220800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003299413656137696,
      "loss": 2.7142,
      "theoretical_loss": 3.517278120524372,
      "tokens_seen": 1494220800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032992481558539815,
      "loss": 2.7436,
      "theoretical_loss": 3.517264509225851,
      "tokens_seen": 1494286336
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003299082655570267,
      "loss": 2.7602,
      "theoretical_loss": 3.517250898691418,
      "tokens_seen": 1494351872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032989171552865517,
      "loss": 2.6911,
      "theoretical_loss": 3.517237288920997,
      "tokens_seen": 1494417408
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003298751655002837,
      "loss": 2.9584,
      "theoretical_loss": 3.517223679914511,
      "tokens_seen": 1494482944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032985861547191224,
      "loss": 2.8904,
      "theoretical_loss": 3.5172100716718844,
      "tokens_seen": 1494548480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003298420654435408,
      "loss": 2.7211,
      "theoretical_loss": 3.5171964641930398,
      "tokens_seen": 1494614016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032982551541516926,
      "loss": 2.6332,
      "theoretical_loss": 3.517182857477902,
      "tokens_seen": 1494679552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003298089653867978,
      "loss": 2.7408,
      "theoretical_loss": 3.5171692515263935,
      "tokens_seen": 1494745088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032979241535842634,
      "loss": 2.7395,
      "theoretical_loss": 3.5171556463384386,
      "tokens_seen": 1494810624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003297758653300549,
      "loss": 2.6006,
      "theoretical_loss": 3.5171420419139614,
      "tokens_seen": 1494876160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032975931530168336,
      "loss": 2.6798,
      "theoretical_loss": 3.5171284382528847,
      "tokens_seen": 1494941696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003297427652733119,
      "loss": 2.6076,
      "theoretical_loss": 3.517114835355133,
      "tokens_seen": 1495007232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032972621524494044,
      "loss": 2.6244,
      "theoretical_loss": 3.5171012332206293,
      "tokens_seen": 1495072768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000329709665216569,
      "loss": 2.7893,
      "theoretical_loss": 3.517087631849298,
      "tokens_seen": 1495138304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032969311518819746,
      "loss": 2.5737,
      "theoretical_loss": 3.517074031241062,
      "tokens_seen": 1495203840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000329676565159826,
      "loss": 2.6221,
      "theoretical_loss": 3.5170604313958456,
      "tokens_seen": 1495269376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032966001513145453,
      "loss": 2.6959,
      "theoretical_loss": 3.5170468323135724,
      "tokens_seen": 1495334912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032964346510308307,
      "loss": 2.622,
      "theoretical_loss": 3.517033233994167,
      "tokens_seen": 1495400448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032962691507471155,
      "loss": 2.9009,
      "theoretical_loss": 3.5170196364375514,
      "tokens_seen": 1495465984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003296103650463401,
      "loss": 2.6171,
      "theoretical_loss": 3.5170060396436504,
      "tokens_seen": 1495531520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003295938150179686,
      "loss": 2.9548,
      "theoretical_loss": 3.5169924436123887,
      "tokens_seen": 1495597056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003295772649895971,
      "loss": 2.6323,
      "theoretical_loss": 3.5169788483436886,
      "tokens_seen": 1495662592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032956071496122565,
      "loss": 2.6217,
      "theoretical_loss": 3.5169652538374745,
      "tokens_seen": 1495728128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032954416493285413,
      "loss": 2.6547,
      "theoretical_loss": 3.51695166009367,
      "tokens_seen": 1495793664
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032952761490448267,
      "loss": 2.6596,
      "theoretical_loss": 3.5169380671122,
      "tokens_seen": 1495859200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003295110648761112,
      "loss": 2.6675,
      "theoretical_loss": 3.516924474892987,
      "tokens_seen": 1495924736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032949451484773974,
      "loss": 2.6231,
      "theoretical_loss": 3.5169108834359557,
      "tokens_seen": 1495990272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032947796481936823,
      "loss": 2.7313,
      "theoretical_loss": 3.5168972927410294,
      "tokens_seen": 1496055808
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032946141479099677,
      "loss": 2.6356,
      "theoretical_loss": 3.5168837028081326,
      "tokens_seen": 1496121344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003294448647626253,
      "loss": 2.7293,
      "theoretical_loss": 3.5168701136371885,
      "tokens_seen": 1496186880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032942831473425384,
      "loss": 2.8296,
      "theoretical_loss": 3.516856525228121,
      "tokens_seen": 1496252416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003294117647058823,
      "loss": 2.5822,
      "theoretical_loss": 3.516842937580855,
      "tokens_seen": 1496317952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032939521467751086,
      "loss": 2.8135,
      "theoretical_loss": 3.516829350695314,
      "tokens_seen": 1496383488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003293786646491394,
      "loss": 2.6925,
      "theoretical_loss": 3.5168157645714215,
      "tokens_seen": 1496449024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032936211462076794,
      "loss": 2.6112,
      "theoretical_loss": 3.516802179209101,
      "tokens_seen": 1496514560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003293455645923964,
      "loss": 2.6421,
      "theoretical_loss": 3.516788594608278,
      "tokens_seen": 1496580096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032932901456402496,
      "loss": 2.6021,
      "theoretical_loss": 3.5167750107688756,
      "tokens_seen": 1496645632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003293124645356535,
      "loss": 2.7857,
      "theoretical_loss": 3.5167614276908177,
      "tokens_seen": 1496711168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032929591450728203,
      "loss": 2.4718,
      "theoretical_loss": 3.5167478453740286,
      "tokens_seen": 1496776704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003292793644789105,
      "loss": 2.813,
      "theoretical_loss": 3.5167342638184316,
      "tokens_seen": 1496842240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032926281445053905,
      "loss": 2.8988,
      "theoretical_loss": 3.5167206830239515,
      "tokens_seen": 1496907776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003292462644221676,
      "loss": 2.6005,
      "theoretical_loss": 3.5167071029905124,
      "tokens_seen": 1496973312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032922971439379613,
      "loss": 2.7277,
      "theoretical_loss": 3.516693523718038,
      "tokens_seen": 1497038848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003292131643654246,
      "loss": 2.4664,
      "theoretical_loss": 3.5166799452064517,
      "tokens_seen": 1497104384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032919661433705315,
      "loss": 2.5741,
      "theoretical_loss": 3.5166663674556786,
      "tokens_seen": 1497169920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003291800643086817,
      "loss": 2.5411,
      "theoretical_loss": 3.5166527904656424,
      "tokens_seen": 1497235456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003291635142803102,
      "loss": 2.6367,
      "theoretical_loss": 3.5166392142362675,
      "tokens_seen": 1497300992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003291469642519387,
      "loss": 2.5778,
      "theoretical_loss": 3.516625638767477,
      "tokens_seen": 1497366528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032913041422356725,
      "loss": 2.8805,
      "theoretical_loss": 3.516612064059196,
      "tokens_seen": 1497432064
    },
    {
      "epoch": 0.53,
      "objective/train/avg_token_score": 0.018181461840867996,
      "objective/train/avg_weight": 0.978614330291748,
      "objective/train/docs_used": 848503,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.871346950531006,
      "objective/train/std_weight": 0.1126461774110794,
      "objective/train/theoretical_loss": 3.516598490111348,
      "objective/train/tokens_used": 1517957600,
      "theoretical_loss": 3.516598490111348,
      "tokens_seen": 1497497600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003291138641951958,
      "loss": 2.8797,
      "theoretical_loss": 3.516598490111348,
      "tokens_seen": 1497497600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032909731416682427,
      "loss": 2.7655,
      "theoretical_loss": 3.516584916923858,
      "tokens_seen": 1497563136
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003290807641384528,
      "loss": 2.9128,
      "theoretical_loss": 3.516571344496649,
      "tokens_seen": 1497628672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003290642141100813,
      "loss": 2.8996,
      "theoretical_loss": 3.516557772829646,
      "tokens_seen": 1497694208
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003290476640817098,
      "loss": 2.8907,
      "theoretical_loss": 3.516544201922773,
      "tokens_seen": 1497759744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00032903111405333836,
      "loss": 2.6322,
      "theoretical_loss": 3.5165306317759537,
      "tokens_seen": 1497825280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003290145640249669,
      "loss": 2.8807,
      "theoretical_loss": 3.516517062389112,
      "tokens_seen": 1497890816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0003289980139965954,
      "loss": 2.5019,
      "theoretical_loss": 3.5165034937621735,
      "tokens_seen": 1497956352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003289814639682239,
      "loss": 2.6309,
      "theoretical_loss": 3.5164899258950615,
      "tokens_seen": 1498021888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032896491393985246,
      "loss": 2.6558,
      "theoretical_loss": 3.5164763587877,
      "tokens_seen": 1498087424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000328948363911481,
      "loss": 2.6535,
      "theoretical_loss": 3.516462792440014,
      "tokens_seen": 1498152960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003289318138831095,
      "loss": 2.724,
      "theoretical_loss": 3.5164492268519267,
      "tokens_seen": 1498218496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000328915263854738,
      "loss": 2.5589,
      "theoretical_loss": 3.516435662023363,
      "tokens_seen": 1498284032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032889871382636655,
      "loss": 2.7256,
      "theoretical_loss": 3.5164220979542478,
      "tokens_seen": 1498349568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003288821637979951,
      "loss": 2.7241,
      "theoretical_loss": 3.5164085346445035,
      "tokens_seen": 1498415104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003288656137696236,
      "loss": 3.0153,
      "theoretical_loss": 3.516394972094056,
      "tokens_seen": 1498480640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003288490637412521,
      "loss": 2.7471,
      "theoretical_loss": 3.516381410302829,
      "tokens_seen": 1498546176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032883251371288065,
      "loss": 2.6384,
      "theoretical_loss": 3.5163678492707473,
      "tokens_seen": 1498611712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003288159636845092,
      "loss": 2.7917,
      "theoretical_loss": 3.5163542889977344,
      "tokens_seen": 1498677248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032879941365613767,
      "loss": 2.5999,
      "theoretical_loss": 3.516340729483715,
      "tokens_seen": 1498742784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003287828636277662,
      "loss": 2.8234,
      "theoretical_loss": 3.516327170728614,
      "tokens_seen": 1498808320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032876631359939475,
      "loss": 2.8461,
      "theoretical_loss": 3.5163136127323544,
      "tokens_seen": 1498873856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003287497635710233,
      "loss": 2.804,
      "theoretical_loss": 3.516300055494862,
      "tokens_seen": 1498939392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003287332135426518,
      "loss": 2.5601,
      "theoretical_loss": 3.5162864990160596,
      "tokens_seen": 1499004928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003287166635142803,
      "loss": 2.7712,
      "theoretical_loss": 3.5162729432958733,
      "tokens_seen": 1499070464
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032870011348590884,
      "loss": 2.5792,
      "theoretical_loss": 3.516259388334227,
      "tokens_seen": 1499136000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003286835634575374,
      "loss": 2.4207,
      "theoretical_loss": 3.516245834131044,
      "tokens_seen": 1499201536
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003286670134291659,
      "loss": 2.6206,
      "theoretical_loss": 3.5162322806862494,
      "tokens_seen": 1499267072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003286504634007944,
      "loss": 2.7503,
      "theoretical_loss": 3.516218727999768,
      "tokens_seen": 1499332608
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032863391337242294,
      "loss": 2.8235,
      "theoretical_loss": 3.516205176071524,
      "tokens_seen": 1499398144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003286173633440515,
      "loss": 2.7986,
      "theoretical_loss": 3.5161916249014413,
      "tokens_seen": 1499463680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032860081331567996,
      "loss": 2.5309,
      "theoretical_loss": 3.5161780744894453,
      "tokens_seen": 1499529216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003285842632873085,
      "loss": 2.7821,
      "theoretical_loss": 3.51616452483546,
      "tokens_seen": 1499594752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000328567713258937,
      "loss": 2.7705,
      "theoretical_loss": 3.5161509759394094,
      "tokens_seen": 1499660288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003285511632305655,
      "loss": 2.7574,
      "theoretical_loss": 3.516137427801219,
      "tokens_seen": 1499725824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032853461320219405,
      "loss": 2.6841,
      "theoretical_loss": 3.516123880420812,
      "tokens_seen": 1499791360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003285180631738226,
      "loss": 2.2552,
      "theoretical_loss": 3.5161103337981143,
      "tokens_seen": 1499856896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003285015131454511,
      "loss": 2.755,
      "theoretical_loss": 3.5160967879330496,
      "tokens_seen": 1499922432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003284849631170796,
      "loss": 2.6541,
      "theoretical_loss": 3.516083242825542,
      "tokens_seen": 1499987968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032846841308870815,
      "loss": 2.8476,
      "theoretical_loss": 3.5160696984755173,
      "tokens_seen": 1500053504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003284518630603367,
      "loss": 2.9023,
      "theoretical_loss": 3.5160561548828992,
      "tokens_seen": 1500119040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032843531303196517,
      "loss": 2.8809,
      "theoretical_loss": 3.5160426120476123,
      "tokens_seen": 1500184576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003284187630035937,
      "loss": 2.557,
      "theoretical_loss": 3.5160290699695813,
      "tokens_seen": 1500250112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032840221297522225,
      "loss": 2.5095,
      "theoretical_loss": 3.5160155286487313,
      "tokens_seen": 1500315648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003283856629468508,
      "loss": 2.649,
      "theoretical_loss": 3.5160019880849855,
      "tokens_seen": 1500381184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032836911291847927,
      "loss": 2.6968,
      "theoretical_loss": 3.5159884482782697,
      "tokens_seen": 1500446720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003283525628901078,
      "loss": 2.7599,
      "theoretical_loss": 3.5159749092285084,
      "tokens_seen": 1500512256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032833601286173634,
      "loss": 2.357,
      "theoretical_loss": 3.5159613709356257,
      "tokens_seen": 1500577792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003283194628333649,
      "loss": 2.7179,
      "theoretical_loss": 3.5159478333995464,
      "tokens_seen": 1500643328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032830291280499336,
      "loss": 2.7759,
      "theoretical_loss": 3.5159342966201956,
      "tokens_seen": 1500708864
    },
    {
      "epoch": 0.54,
      "objective/train/avg_token_score": 0.02170385979115963,
      "objective/train/avg_weight": 0.9744846820831299,
      "objective/train/docs_used": 850745,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1596755981445312,
      "objective/train/std_weight": 0.10622738301753998,
      "objective/train/theoretical_loss": 3.515920760597498,
      "objective/train/tokens_used": 1521234400,
      "theoretical_loss": 3.515920760597498,
      "tokens_seen": 1500774400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003282863627766219,
      "loss": 2.8476,
      "theoretical_loss": 3.515920760597498,
      "tokens_seen": 1500774400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032826981274825044,
      "loss": 2.7559,
      "theoretical_loss": 3.515907225331377,
      "tokens_seen": 1500839936
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000328253262719879,
      "loss": 2.8576,
      "theoretical_loss": 3.515893690821759,
      "tokens_seen": 1500905472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032823671269150746,
      "loss": 2.7797,
      "theoretical_loss": 3.515880157068567,
      "tokens_seen": 1500971008
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000328220162663136,
      "loss": 2.5412,
      "theoretical_loss": 3.5158666240717276,
      "tokens_seen": 1501036544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032820361263476453,
      "loss": 2.2872,
      "theoretical_loss": 3.515853091831164,
      "tokens_seen": 1501102080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032818706260639307,
      "loss": 2.5208,
      "theoretical_loss": 3.5158395603468016,
      "tokens_seen": 1501167616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032817051257802156,
      "loss": 2.71,
      "theoretical_loss": 3.515826029618564,
      "tokens_seen": 1501233152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003281539625496501,
      "loss": 2.7127,
      "theoretical_loss": 3.515812499646378,
      "tokens_seen": 1501298688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032813741252127863,
      "loss": 2.011,
      "theoretical_loss": 3.5157989704301666,
      "tokens_seen": 1501364224
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003281208624929071,
      "loss": 2.6274,
      "theoretical_loss": 3.5157854419698555,
      "tokens_seen": 1501429760
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032810431246453565,
      "loss": 2.8926,
      "theoretical_loss": 3.515771914265369,
      "tokens_seen": 1501495296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032808776243616413,
      "loss": 2.7543,
      "theoretical_loss": 3.515758387316632,
      "tokens_seen": 1501560832
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032807121240779267,
      "loss": 2.5981,
      "theoretical_loss": 3.51574486112357,
      "tokens_seen": 1501626368
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003280546623794212,
      "loss": 2.2997,
      "theoretical_loss": 3.5157313356861066,
      "tokens_seen": 1501691904
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032803811235104975,
      "loss": 2.4134,
      "theoretical_loss": 3.5157178110041674,
      "tokens_seen": 1501757440
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032802156232267823,
      "loss": 2.6808,
      "theoretical_loss": 3.5157042870776767,
      "tokens_seen": 1501822976
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032800501229430677,
      "loss": 2.6364,
      "theoretical_loss": 3.5156907639065595,
      "tokens_seen": 1501888512
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003279884622659353,
      "loss": 2.5872,
      "theoretical_loss": 3.5156772414907413,
      "tokens_seen": 1501954048
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032797191223756384,
      "loss": 2.6347,
      "theoretical_loss": 3.515663719830146,
      "tokens_seen": 1502019584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003279553622091923,
      "loss": 2.8161,
      "theoretical_loss": 3.5156501989246993,
      "tokens_seen": 1502085120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032793881218082086,
      "loss": 2.5713,
      "theoretical_loss": 3.5156366787743254,
      "tokens_seen": 1502150656
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003279222621524494,
      "loss": 2.819,
      "theoretical_loss": 3.5156231593789498,
      "tokens_seen": 1502216192
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032790571212407794,
      "loss": 2.0897,
      "theoretical_loss": 3.5156096407384974,
      "tokens_seen": 1502281728
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003278891620957064,
      "loss": 2.55,
      "theoretical_loss": 3.515596122852892,
      "tokens_seen": 1502347264
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032787261206733496,
      "loss": 2.6582,
      "theoretical_loss": 3.51558260572206,
      "tokens_seen": 1502412800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003278560620389635,
      "loss": 2.7982,
      "theoretical_loss": 3.5155690893459255,
      "tokens_seen": 1502478336
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032783951201059204,
      "loss": 2.3524,
      "theoretical_loss": 3.5155555737244133,
      "tokens_seen": 1502543872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003278229619822205,
      "loss": 2.6307,
      "theoretical_loss": 3.515542058857449,
      "tokens_seen": 1502609408
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032780641195384906,
      "loss": 2.8817,
      "theoretical_loss": 3.5155285447449574,
      "tokens_seen": 1502674944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003277898619254776,
      "loss": 2.6221,
      "theoretical_loss": 3.515515031386863,
      "tokens_seen": 1502740480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032777331189710613,
      "loss": 2.4353,
      "theoretical_loss": 3.5155015187830916,
      "tokens_seen": 1502806016
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003277567618687346,
      "loss": 2.7502,
      "theoretical_loss": 3.515488006933567,
      "tokens_seen": 1502871552
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032774021184036315,
      "loss": 2.4682,
      "theoretical_loss": 3.515474495838216,
      "tokens_seen": 1502937088
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003277236618119917,
      "loss": 2.8097,
      "theoretical_loss": 3.5154609854969614,
      "tokens_seen": 1503002624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032770711178362023,
      "loss": 2.658,
      "theoretical_loss": 3.5154474759097303,
      "tokens_seen": 1503068160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003276905617552487,
      "loss": 2.6939,
      "theoretical_loss": 3.5154339670764463,
      "tokens_seen": 1503133696
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032767401172687725,
      "loss": 2.7132,
      "theoretical_loss": 3.5154204589970353,
      "tokens_seen": 1503199232
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003276574616985058,
      "loss": 2.4919,
      "theoretical_loss": 3.515406951671422,
      "tokens_seen": 1503264768
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003276409116701343,
      "loss": 2.7496,
      "theoretical_loss": 3.5153934450995314,
      "tokens_seen": 1503330304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003276243616417628,
      "loss": 2.6921,
      "theoretical_loss": 3.5153799392812886,
      "tokens_seen": 1503395840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003276078116133913,
      "loss": 2.6303,
      "theoretical_loss": 3.5153664342166193,
      "tokens_seen": 1503461376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032759126158501983,
      "loss": 2.7944,
      "theoretical_loss": 3.515352929905448,
      "tokens_seen": 1503526912
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032757471155664837,
      "loss": 2.6142,
      "theoretical_loss": 3.5153394263476994,
      "tokens_seen": 1503592448
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003275581615282769,
      "loss": 2.8459,
      "theoretical_loss": 3.5153259235432994,
      "tokens_seen": 1503657984
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003275416114999054,
      "loss": 2.6233,
      "theoretical_loss": 3.5153124214921734,
      "tokens_seen": 1503723520
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003275250614715339,
      "loss": 2.4561,
      "theoretical_loss": 3.5152989201942457,
      "tokens_seen": 1503789056
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032750851144316246,
      "loss": 2.8716,
      "theoretical_loss": 3.515285419649442,
      "tokens_seen": 1503854592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000327491961414791,
      "loss": 2.7341,
      "theoretical_loss": 3.515271919857687,
      "tokens_seen": 1503920128
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032747541138641954,
      "loss": 2.7688,
      "theoretical_loss": 3.515258420818906,
      "tokens_seen": 1503985664
    },
    {
      "epoch": 0.54,
      "objective/train/avg_token_score": 0.016090868040919304,
      "objective/train/avg_weight": 0.9810699224472046,
      "objective/train/docs_used": 852130,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.733710289001465,
      "objective/train/std_weight": 0.09029201418161392,
      "objective/train/theoretical_loss": 3.515244922533025,
      "objective/train/tokens_used": 1524511200,
      "theoretical_loss": 3.515244922533025,
      "tokens_seen": 1504051200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000327458861358048,
      "loss": 2.7359,
      "theoretical_loss": 3.515244922533025,
      "tokens_seen": 1504051200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032744231132967656,
      "loss": 2.5857,
      "theoretical_loss": 3.5152314249999685,
      "tokens_seen": 1504116736
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003274257613013051,
      "loss": 2.5923,
      "theoretical_loss": 3.5152179282196614,
      "tokens_seen": 1504182272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032740921127293363,
      "loss": 2.6152,
      "theoretical_loss": 3.5152044321920295,
      "tokens_seen": 1504247808
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003273926612445621,
      "loss": 2.4049,
      "theoretical_loss": 3.5151909369169982,
      "tokens_seen": 1504313344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032737611121619065,
      "loss": 2.6544,
      "theoretical_loss": 3.515177442394492,
      "tokens_seen": 1504378880
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003273595611878192,
      "loss": 2.6926,
      "theoretical_loss": 3.5151639486244366,
      "tokens_seen": 1504444416
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032734301115944773,
      "loss": 2.7465,
      "theoretical_loss": 3.515150455606757,
      "tokens_seen": 1504509952
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003273264611310762,
      "loss": 2.4097,
      "theoretical_loss": 3.515136963341379,
      "tokens_seen": 1504575488
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032730991110270475,
      "loss": 2.7261,
      "theoretical_loss": 3.515123471828228,
      "tokens_seen": 1504641024
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003272933610743333,
      "loss": 2.4595,
      "theoretical_loss": 3.515109981067228,
      "tokens_seen": 1504706560
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003272768110459618,
      "loss": 2.1301,
      "theoretical_loss": 3.5150964910583062,
      "tokens_seen": 1504772096
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003272602610175903,
      "loss": 2.6752,
      "theoretical_loss": 3.5150830018013863,
      "tokens_seen": 1504837632
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032724371098921885,
      "loss": 2.5882,
      "theoretical_loss": 3.5150695132963943,
      "tokens_seen": 1504903168
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003272271609608474,
      "loss": 2.4197,
      "theoretical_loss": 3.515056025543256,
      "tokens_seen": 1504968704
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003272106109324759,
      "loss": 2.7081,
      "theoretical_loss": 3.5150425385418957,
      "tokens_seen": 1505034240
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003271940609041044,
      "loss": 2.45,
      "theoretical_loss": 3.515029052292239,
      "tokens_seen": 1505099776
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032717751087573294,
      "loss": 2.5511,
      "theoretical_loss": 3.5150155667942125,
      "tokens_seen": 1505165312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003271609608473615,
      "loss": 2.2356,
      "theoretical_loss": 3.51500208204774,
      "tokens_seen": 1505230848
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032714441081899,
      "loss": 2.578,
      "theoretical_loss": 3.514988598052748,
      "tokens_seen": 1505296384
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003271278607906185,
      "loss": 2.6756,
      "theoretical_loss": 3.5149751148091615,
      "tokens_seen": 1505361920
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000327111310762247,
      "loss": 2.6263,
      "theoretical_loss": 3.514961632316906,
      "tokens_seen": 1505427456
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003270947607338755,
      "loss": 2.3299,
      "theoretical_loss": 3.5149481505759064,
      "tokens_seen": 1505492992
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032707821070550406,
      "loss": 2.652,
      "theoretical_loss": 3.514934669586089,
      "tokens_seen": 1505558528
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003270616606771326,
      "loss": 2.7064,
      "theoretical_loss": 3.5149211893473784,
      "tokens_seen": 1505624064
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003270451106487611,
      "loss": 2.732,
      "theoretical_loss": 3.514907709859701,
      "tokens_seen": 1505689600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003270285606203896,
      "loss": 2.9076,
      "theoretical_loss": 3.514894231122981,
      "tokens_seen": 1505755136
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032701201059201815,
      "loss": 2.7485,
      "theoretical_loss": 3.5148807531371453,
      "tokens_seen": 1505820672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003269954605636467,
      "loss": 2.62,
      "theoretical_loss": 3.5148672759021182,
      "tokens_seen": 1505886208
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003269789105352752,
      "loss": 2.6142,
      "theoretical_loss": 3.514853799417826,
      "tokens_seen": 1505951744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003269623605069037,
      "loss": 2.6403,
      "theoretical_loss": 3.514840323684194,
      "tokens_seen": 1506017280
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032694581047853225,
      "loss": 2.6437,
      "theoretical_loss": 3.514826848701148,
      "tokens_seen": 1506082816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003269292604501608,
      "loss": 2.636,
      "theoretical_loss": 3.5148133744686127,
      "tokens_seen": 1506148352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032691271042178927,
      "loss": 2.6537,
      "theoretical_loss": 3.5147999009865143,
      "tokens_seen": 1506213888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003268961603934178,
      "loss": 2.547,
      "theoretical_loss": 3.5147864282547783,
      "tokens_seen": 1506279424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032687961036504635,
      "loss": 2.6267,
      "theoretical_loss": 3.51477295627333,
      "tokens_seen": 1506344960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003268630603366749,
      "loss": 2.7054,
      "theoretical_loss": 3.514759485042095,
      "tokens_seen": 1506410496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032684651030830337,
      "loss": 2.7741,
      "theoretical_loss": 3.514746014560999,
      "tokens_seen": 1506476032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003268299602799319,
      "loss": 2.6486,
      "theoretical_loss": 3.514732544829968,
      "tokens_seen": 1506541568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032681341025156044,
      "loss": 2.7084,
      "theoretical_loss": 3.5147190758489275,
      "tokens_seen": 1506607104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000326796860223189,
      "loss": 2.3398,
      "theoretical_loss": 3.5147056076178025,
      "tokens_seen": 1506672640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032678031019481746,
      "loss": 2.6461,
      "theoretical_loss": 3.514692140136519,
      "tokens_seen": 1506738176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000326763760166446,
      "loss": 2.6637,
      "theoretical_loss": 3.514678673405003,
      "tokens_seen": 1506803712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032674721013807454,
      "loss": 2.4328,
      "theoretical_loss": 3.514665207423179,
      "tokens_seen": 1506869248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003267306601097031,
      "loss": 2.863,
      "theoretical_loss": 3.514651742190974,
      "tokens_seen": 1506934784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032671411008133156,
      "loss": 2.6156,
      "theoretical_loss": 3.514638277708313,
      "tokens_seen": 1507000320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003266975600529601,
      "loss": 2.7121,
      "theoretical_loss": 3.5146248139751215,
      "tokens_seen": 1507065856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032668101002458863,
      "loss": 2.6168,
      "theoretical_loss": 3.514611350991326,
      "tokens_seen": 1507131392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032666445999621717,
      "loss": 2.9384,
      "theoretical_loss": 3.5145978887568514,
      "tokens_seen": 1507196928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032664790996784565,
      "loss": 2.7806,
      "theoretical_loss": 3.5145844272716236,
      "tokens_seen": 1507262464
    },
    {
      "epoch": 0.54,
      "objective/train/avg_token_score": 0.02060885541141033,
      "objective/train/avg_weight": 0.9757462739944458,
      "objective/train/docs_used": 853500,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.914759635925293,
      "objective/train/std_weight": 0.10687510669231415,
      "objective/train/theoretical_loss": 3.514570966535569,
      "objective/train/tokens_used": 1527788000,
      "theoretical_loss": 3.514570966535569,
      "tokens_seen": 1507328000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032663135993947414,
      "loss": 2.7778,
      "theoretical_loss": 3.514570966535569,
      "tokens_seen": 1507328000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003266148099111027,
      "loss": 2.6541,
      "theoretical_loss": 3.5145575065486123,
      "tokens_seen": 1507393536
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003265982598827312,
      "loss": 2.6853,
      "theoretical_loss": 3.51454404731068,
      "tokens_seen": 1507459072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032658170985435975,
      "loss": 2.8002,
      "theoretical_loss": 3.5145305888216973,
      "tokens_seen": 1507524608
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032656515982598823,
      "loss": 2.7187,
      "theoretical_loss": 3.5145171310815906,
      "tokens_seen": 1507590144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032654860979761677,
      "loss": 2.6413,
      "theoretical_loss": 3.5145036740902853,
      "tokens_seen": 1507655680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003265320597692453,
      "loss": 2.5732,
      "theoretical_loss": 3.514490217847707,
      "tokens_seen": 1507721216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032651550974087385,
      "loss": 2.792,
      "theoretical_loss": 3.5144767623537816,
      "tokens_seen": 1507786752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032649895971250233,
      "loss": 2.4146,
      "theoretical_loss": 3.514463307608435,
      "tokens_seen": 1507852288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032648240968413087,
      "loss": 2.7577,
      "theoretical_loss": 3.514449853611594,
      "tokens_seen": 1507917824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003264658596557594,
      "loss": 2.8442,
      "theoretical_loss": 3.5144364003631825,
      "tokens_seen": 1507983360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032644930962738794,
      "loss": 2.75,
      "theoretical_loss": 3.5144229478631273,
      "tokens_seen": 1508048896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003264327595990164,
      "loss": 2.8487,
      "theoretical_loss": 3.5144094961113543,
      "tokens_seen": 1508114432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032641620957064496,
      "loss": 2.9558,
      "theoretical_loss": 3.51439604510779,
      "tokens_seen": 1508179968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003263996595422735,
      "loss": 2.8183,
      "theoretical_loss": 3.514382594852359,
      "tokens_seen": 1508245504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032638310951390204,
      "loss": 2.5214,
      "theoretical_loss": 3.514369145344988,
      "tokens_seen": 1508311040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003263665594855305,
      "loss": 2.861,
      "theoretical_loss": 3.5143556965856027,
      "tokens_seen": 1508376576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032635000945715906,
      "loss": 2.779,
      "theoretical_loss": 3.514342248574129,
      "tokens_seen": 1508442112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003263334594287876,
      "loss": 2.6661,
      "theoretical_loss": 3.5143288013104925,
      "tokens_seen": 1508507648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032631690940041613,
      "loss": 2.555,
      "theoretical_loss": 3.51431535479462,
      "tokens_seen": 1508573184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003263003593720446,
      "loss": 2.7534,
      "theoretical_loss": 3.514301909026437,
      "tokens_seen": 1508638720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032628380934367316,
      "loss": 2.8499,
      "theoretical_loss": 3.5142884640058685,
      "tokens_seen": 1508704256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003262672593153017,
      "loss": 2.4652,
      "theoretical_loss": 3.5142750197328416,
      "tokens_seen": 1508769792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032625070928693023,
      "loss": 2.4154,
      "theoretical_loss": 3.514261576207282,
      "tokens_seen": 1508835328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003262341592585587,
      "loss": 2.5956,
      "theoretical_loss": 3.514248133429116,
      "tokens_seen": 1508900864
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032621760923018725,
      "loss": 2.8088,
      "theoretical_loss": 3.514234691398269,
      "tokens_seen": 1508966400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003262010592018158,
      "loss": 2.5078,
      "theoretical_loss": 3.514221250114667,
      "tokens_seen": 1509031936
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003261845091734443,
      "loss": 2.712,
      "theoretical_loss": 3.5142078095782363,
      "tokens_seen": 1509097472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003261679591450728,
      "loss": 2.8917,
      "theoretical_loss": 3.514194369788903,
      "tokens_seen": 1509163008
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032615140911670135,
      "loss": 2.707,
      "theoretical_loss": 3.514180930746593,
      "tokens_seen": 1509228544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032613485908832983,
      "loss": 2.7357,
      "theoretical_loss": 3.5141674924512323,
      "tokens_seen": 1509294080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032611830905995837,
      "loss": 2.8565,
      "theoretical_loss": 3.514154054902747,
      "tokens_seen": 1509359616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003261017590315869,
      "loss": 2.7098,
      "theoretical_loss": 3.514140618101063,
      "tokens_seen": 1509425152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032608520900321544,
      "loss": 2.7218,
      "theoretical_loss": 3.5141271820461064,
      "tokens_seen": 1509490688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003260686589748439,
      "loss": 2.8534,
      "theoretical_loss": 3.514113746737804,
      "tokens_seen": 1509556224
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032605210894647246,
      "loss": 2.5793,
      "theoretical_loss": 3.5141003121760805,
      "tokens_seen": 1509621760
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000326035558918101,
      "loss": 2.487,
      "theoretical_loss": 3.514086878360863,
      "tokens_seen": 1509687296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032601900888972954,
      "loss": 2.6883,
      "theoretical_loss": 3.5140734452920777,
      "tokens_seen": 1509752832
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000326002458861358,
      "loss": 2.8454,
      "theoretical_loss": 3.514060012969651,
      "tokens_seen": 1509818368
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032598590883298656,
      "loss": 2.5679,
      "theoretical_loss": 3.5140465813935076,
      "tokens_seen": 1509883904
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003259693588046151,
      "loss": 2.6606,
      "theoretical_loss": 3.5140331505635745,
      "tokens_seen": 1509949440
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032595280877624364,
      "loss": 2.6031,
      "theoretical_loss": 3.5140197204797783,
      "tokens_seen": 1510014976
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003259362587478721,
      "loss": 2.6971,
      "theoretical_loss": 3.514006291142045,
      "tokens_seen": 1510080512
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032591970871950066,
      "loss": 2.6099,
      "theoretical_loss": 3.5139928625503,
      "tokens_seen": 1510146048
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003259031586911292,
      "loss": 2.8224,
      "theoretical_loss": 3.5139794347044706,
      "tokens_seen": 1510211584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032588660866275773,
      "loss": 2.773,
      "theoretical_loss": 3.513966007604482,
      "tokens_seen": 1510277120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003258700586343862,
      "loss": 2.6308,
      "theoretical_loss": 3.5139525812502606,
      "tokens_seen": 1510342656
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032585350860601475,
      "loss": 2.7196,
      "theoretical_loss": 3.513939155641733,
      "tokens_seen": 1510408192
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003258369585776433,
      "loss": 2.7323,
      "theoretical_loss": 3.5139257307788254,
      "tokens_seen": 1510473728
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003258204085492718,
      "loss": 2.7656,
      "theoretical_loss": 3.5139123066614646,
      "tokens_seen": 1510539264
    },
    {
      "epoch": 0.54,
      "objective/train/avg_token_score": 0.013696974143385887,
      "objective/train/avg_weight": 0.9838927984237671,
      "objective/train/docs_used": 854981,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.499075174331665,
      "objective/train/std_weight": 0.05619911476969719,
      "objective/train/theoretical_loss": 3.513898883289575,
      "objective/train/tokens_used": 1531064800,
      "theoretical_loss": 3.513898883289575,
      "tokens_seen": 1510604800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003258038585209003,
      "loss": 2.5431,
      "theoretical_loss": 3.513898883289575,
      "tokens_seen": 1510604800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032578730849252885,
      "loss": 2.639,
      "theoretical_loss": 3.5138854606630847,
      "tokens_seen": 1510670336
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003257707584641574,
      "loss": 2.5543,
      "theoretical_loss": 3.5138720387819196,
      "tokens_seen": 1510735872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003257542084357859,
      "loss": 2.7439,
      "theoretical_loss": 3.5138586176460054,
      "tokens_seen": 1510801408
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003257376584074144,
      "loss": 2.5417,
      "theoretical_loss": 3.513845197255269,
      "tokens_seen": 1510866944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032572110837904294,
      "loss": 2.7599,
      "theoretical_loss": 3.513831777609636,
      "tokens_seen": 1510932480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003257045583506715,
      "loss": 2.6678,
      "theoretical_loss": 3.513818358709033,
      "tokens_seen": 1510998016
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003256880083223,
      "loss": 2.8937,
      "theoretical_loss": 3.5138049405533867,
      "tokens_seen": 1511063552
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003256714582939285,
      "loss": 2.819,
      "theoretical_loss": 3.5137915231426233,
      "tokens_seen": 1511129088
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000325654908265557,
      "loss": 2.5435,
      "theoretical_loss": 3.5137781064766687,
      "tokens_seen": 1511194624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003256383582371855,
      "loss": 2.4242,
      "theoretical_loss": 3.51376469055545,
      "tokens_seen": 1511260160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032562180820881406,
      "loss": 2.6855,
      "theoretical_loss": 3.513751275378893,
      "tokens_seen": 1511325696
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003256052581804426,
      "loss": 2.5975,
      "theoretical_loss": 3.5137378609469243,
      "tokens_seen": 1511391232
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003255887081520711,
      "loss": 2.6762,
      "theoretical_loss": 3.51372444725947,
      "tokens_seen": 1511456768
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003255721581236996,
      "loss": 2.8053,
      "theoretical_loss": 3.513711034316457,
      "tokens_seen": 1511522304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032555560809532816,
      "loss": 2.4584,
      "theoretical_loss": 3.513697622117811,
      "tokens_seen": 1511587840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003255390580669567,
      "loss": 2.6118,
      "theoretical_loss": 3.5136842106634596,
      "tokens_seen": 1511653376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003255225080385852,
      "loss": 2.5994,
      "theoretical_loss": 3.513670799953328,
      "tokens_seen": 1511718912
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003255059580102137,
      "loss": 2.3276,
      "theoretical_loss": 3.513657389987343,
      "tokens_seen": 1511784448
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032548940798184225,
      "loss": 2.5797,
      "theoretical_loss": 3.5136439807654316,
      "tokens_seen": 1511849984
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003254728579534708,
      "loss": 2.7445,
      "theoretical_loss": 3.5136305722875196,
      "tokens_seen": 1511915520
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003254563079250993,
      "loss": 2.8062,
      "theoretical_loss": 3.5136171645535335,
      "tokens_seen": 1511981056
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003254397578967278,
      "loss": 2.8521,
      "theoretical_loss": 3.5136037575634003,
      "tokens_seen": 1512046592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032542320786835635,
      "loss": 2.6084,
      "theoretical_loss": 3.5135903513170463,
      "tokens_seen": 1512112128
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003254066578399849,
      "loss": 2.8207,
      "theoretical_loss": 3.513576945814398,
      "tokens_seen": 1512177664
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032539010781161337,
      "loss": 2.7078,
      "theoretical_loss": 3.5135635410553814,
      "tokens_seen": 1512243200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003253735577832419,
      "loss": 2.7824,
      "theoretical_loss": 3.513550137039924,
      "tokens_seen": 1512308736
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032535700775487044,
      "loss": 2.7224,
      "theoretical_loss": 3.513536733767951,
      "tokens_seen": 1512374272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000325340457726499,
      "loss": 2.5914,
      "theoretical_loss": 3.5135233312393903,
      "tokens_seen": 1512439808
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032532390769812747,
      "loss": 2.7175,
      "theoretical_loss": 3.513509929454168,
      "tokens_seen": 1512505344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000325307357669756,
      "loss": 2.7111,
      "theoretical_loss": 3.51349652841221,
      "tokens_seen": 1512570880
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032529080764138454,
      "loss": 2.9388,
      "theoretical_loss": 3.5134831281134433,
      "tokens_seen": 1512636416
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003252742576130131,
      "loss": 2.4823,
      "theoretical_loss": 3.513469728557795,
      "tokens_seen": 1512701952
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032525770758464156,
      "loss": 2.6544,
      "theoretical_loss": 3.5134563297451917,
      "tokens_seen": 1512767488
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003252411575562701,
      "loss": 2.5595,
      "theoretical_loss": 3.5134429316755593,
      "tokens_seen": 1512833024
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032522460752789864,
      "loss": 2.4676,
      "theoretical_loss": 3.513429534348824,
      "tokens_seen": 1512898560
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003252080574995272,
      "loss": 2.9341,
      "theoretical_loss": 3.5134161377649145,
      "tokens_seen": 1512964096
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032519150747115566,
      "loss": 2.5819,
      "theoretical_loss": 3.5134027419237555,
      "tokens_seen": 1513029632
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032517495744278414,
      "loss": 2.5424,
      "theoretical_loss": 3.513389346825274,
      "tokens_seen": 1513095168
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003251584074144127,
      "loss": 2.7514,
      "theoretical_loss": 3.513375952469397,
      "tokens_seen": 1513160704
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003251418573860412,
      "loss": 2.7085,
      "theoretical_loss": 3.513362558856051,
      "tokens_seen": 1513226240
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032512530735766975,
      "loss": 2.7259,
      "theoretical_loss": 3.5133491659851632,
      "tokens_seen": 1513291776
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032510875732929824,
      "loss": 2.6238,
      "theoretical_loss": 3.5133357738566597,
      "tokens_seen": 1513357312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003250922073009268,
      "loss": 2.722,
      "theoretical_loss": 3.5133223824704674,
      "tokens_seen": 1513422848
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003250756572725553,
      "loss": 2.6304,
      "theoretical_loss": 3.513308991826513,
      "tokens_seen": 1513488384
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032505910724418385,
      "loss": 2.6535,
      "theoretical_loss": 3.513295601924723,
      "tokens_seen": 1513553920
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032504255721581233,
      "loss": 2.8377,
      "theoretical_loss": 3.513282212765024,
      "tokens_seen": 1513619456
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032502600718744087,
      "loss": 2.6635,
      "theoretical_loss": 3.5132688243473438,
      "tokens_seen": 1513684992
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003250094571590694,
      "loss": 2.6803,
      "theoretical_loss": 3.5132554366716082,
      "tokens_seen": 1513750528
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032499290713069795,
      "loss": 2.7756,
      "theoretical_loss": 3.5132420497377437,
      "tokens_seen": 1513816064
    },
    {
      "epoch": 0.54,
      "objective/train/avg_token_score": 0.019163120537996292,
      "objective/train/avg_weight": 0.9774447083473206,
      "objective/train/docs_used": 857043,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7570624351501465,
      "objective/train/std_weight": 0.09026959538459778,
      "objective/train/theoretical_loss": 3.513228663545678,
      "objective/train/tokens_used": 1534341600,
      "theoretical_loss": 3.513228663545678,
      "tokens_seen": 1513881600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032497635710232643,
      "loss": 2.6644,
      "theoretical_loss": 3.513228663545678,
      "tokens_seen": 1513881600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032495980707395497,
      "loss": 2.8897,
      "theoretical_loss": 3.5132152780953376,
      "tokens_seen": 1513947136
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003249432570455835,
      "loss": 2.6429,
      "theoretical_loss": 3.513201893386649,
      "tokens_seen": 1514012672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032492670701721204,
      "loss": 2.6616,
      "theoretical_loss": 3.513188509419539,
      "tokens_seen": 1514078208
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003249101569888405,
      "loss": 2.8267,
      "theoretical_loss": 3.513175126193935,
      "tokens_seen": 1514143744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032489360696046906,
      "loss": 2.8214,
      "theoretical_loss": 3.5131617437097633,
      "tokens_seen": 1514209280
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003248770569320976,
      "loss": 2.9045,
      "theoretical_loss": 3.5131483619669504,
      "tokens_seen": 1514274816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032486050690372614,
      "loss": 2.5735,
      "theoretical_loss": 3.513134980965424,
      "tokens_seen": 1514340352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003248439568753546,
      "loss": 2.6735,
      "theoretical_loss": 3.5131216007051105,
      "tokens_seen": 1514405888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032482740684698316,
      "loss": 2.161,
      "theoretical_loss": 3.513108221185937,
      "tokens_seen": 1514471424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003248108568186117,
      "loss": 2.6896,
      "theoretical_loss": 3.5130948424078294,
      "tokens_seen": 1514536960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032479430679024023,
      "loss": 2.6432,
      "theoretical_loss": 3.5130814643707162,
      "tokens_seen": 1514602496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003247777567618687,
      "loss": 2.7652,
      "theoretical_loss": 3.5130680870745232,
      "tokens_seen": 1514668032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032476120673349725,
      "loss": 2.61,
      "theoretical_loss": 3.5130547105191776,
      "tokens_seen": 1514733568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003247446567051258,
      "loss": 2.7069,
      "theoretical_loss": 3.5130413347046066,
      "tokens_seen": 1514799104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032472810667675433,
      "loss": 2.6475,
      "theoretical_loss": 3.5130279596307363,
      "tokens_seen": 1514864640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032471155664838287,
      "loss": 2.6489,
      "theoretical_loss": 3.513014585297494,
      "tokens_seen": 1514930176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032469500662001135,
      "loss": 2.5045,
      "theoretical_loss": 3.5130012117048075,
      "tokens_seen": 1514995712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032467845659163983,
      "loss": 2.7073,
      "theoretical_loss": 3.5129878388526032,
      "tokens_seen": 1515061248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032466190656326837,
      "loss": 2.3051,
      "theoretical_loss": 3.5129744667408076,
      "tokens_seen": 1515126784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003246453565348969,
      "loss": 2.513,
      "theoretical_loss": 3.5129610953693478,
      "tokens_seen": 1515192320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032462880650652545,
      "loss": 2.7447,
      "theoretical_loss": 3.5129477247381518,
      "tokens_seen": 1515257856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032461225647815393,
      "loss": 2.8801,
      "theoretical_loss": 3.512934354847145,
      "tokens_seen": 1515323392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032459570644978247,
      "loss": 2.7526,
      "theoretical_loss": 3.5129209856962555,
      "tokens_seen": 1515388928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000324579156421411,
      "loss": 2.7364,
      "theoretical_loss": 3.5129076172854106,
      "tokens_seen": 1515454464
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032456260639303954,
      "loss": 2.7924,
      "theoretical_loss": 3.5128942496145363,
      "tokens_seen": 1515520000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000324546056364668,
      "loss": 2.8751,
      "theoretical_loss": 3.5128808826835605,
      "tokens_seen": 1515585536
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032452950633629656,
      "loss": 2.9183,
      "theoretical_loss": 3.5128675164924097,
      "tokens_seen": 1515651072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003245129563079251,
      "loss": 2.8156,
      "theoretical_loss": 3.5128541510410116,
      "tokens_seen": 1515716608
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032449640627955364,
      "loss": 2.6231,
      "theoretical_loss": 3.5128407863292925,
      "tokens_seen": 1515782144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003244798562511821,
      "loss": 2.7671,
      "theoretical_loss": 3.51282742235718,
      "tokens_seen": 1515847680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032446330622281066,
      "loss": 2.561,
      "theoretical_loss": 3.512814059124601,
      "tokens_seen": 1515913216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003244467561944392,
      "loss": 2.7239,
      "theoretical_loss": 3.5128006966314826,
      "tokens_seen": 1515978752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032443020616606773,
      "loss": 2.7117,
      "theoretical_loss": 3.5127873348777516,
      "tokens_seen": 1516044288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003244136561376962,
      "loss": 2.6319,
      "theoretical_loss": 3.512773973863336,
      "tokens_seen": 1516109824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032439710610932475,
      "loss": 2.5703,
      "theoretical_loss": 3.5127606135881626,
      "tokens_seen": 1516175360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003243805560809533,
      "loss": 1.9206,
      "theoretical_loss": 3.512747254052158,
      "tokens_seen": 1516240896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032436400605258183,
      "loss": 2.8034,
      "theoretical_loss": 3.51273389525525,
      "tokens_seen": 1516306432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003243474560242103,
      "loss": 2.1101,
      "theoretical_loss": 3.512720537197365,
      "tokens_seen": 1516371968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032433090599583885,
      "loss": 2.7793,
      "theoretical_loss": 3.512707179878431,
      "tokens_seen": 1516437504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003243143559674674,
      "loss": 2.7009,
      "theoretical_loss": 3.512693823298375,
      "tokens_seen": 1516503040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003242978059390959,
      "loss": 2.5955,
      "theoretical_loss": 3.512680467457124,
      "tokens_seen": 1516568576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003242812559107244,
      "loss": 2.6996,
      "theoretical_loss": 3.5126671123546056,
      "tokens_seen": 1516634112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032426470588235295,
      "loss": 2.6875,
      "theoretical_loss": 3.5126537579907464,
      "tokens_seen": 1516699648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003242481558539815,
      "loss": 2.6035,
      "theoretical_loss": 3.5126404043654738,
      "tokens_seen": 1516765184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032423160582561,
      "loss": 2.3568,
      "theoretical_loss": 3.5126270514787157,
      "tokens_seen": 1516830720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003242150557972385,
      "loss": 2.6234,
      "theoretical_loss": 3.5126136993303985,
      "tokens_seen": 1516896256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032419850576886704,
      "loss": 2.8058,
      "theoretical_loss": 3.5126003479204497,
      "tokens_seen": 1516961792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003241819557404955,
      "loss": 2.8317,
      "theoretical_loss": 3.512586997248797,
      "tokens_seen": 1517027328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032416540571212406,
      "loss": 2.6539,
      "theoretical_loss": 3.5125736473153673,
      "tokens_seen": 1517092864
    },
    {
      "epoch": 0.54,
      "objective/train/avg_token_score": 0.01424703374505043,
      "objective/train/avg_weight": 0.9832313060760498,
      "objective/train/docs_used": 859161,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7267496585845947,
      "objective/train/std_weight": 0.0848308652639389,
      "objective/train/theoretical_loss": 3.5125602981200874,
      "objective/train/tokens_used": 1537618400,
      "theoretical_loss": 3.5125602981200874,
      "tokens_seen": 1517158400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003241488556837526,
      "loss": 2.7531,
      "theoretical_loss": 3.5125602981200874,
      "tokens_seen": 1517158400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003241323056553811,
      "loss": 2.776,
      "theoretical_loss": 3.5125469496628856,
      "tokens_seen": 1517223936
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003241157556270096,
      "loss": 2.6266,
      "theoretical_loss": 3.512533601943689,
      "tokens_seen": 1517289472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032409920559863816,
      "loss": 2.5955,
      "theoretical_loss": 3.512520254962424,
      "tokens_seen": 1517355008
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003240826555702667,
      "loss": 2.445,
      "theoretical_loss": 3.5125069087190193,
      "tokens_seen": 1517420544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003240661055418952,
      "loss": 2.6159,
      "theoretical_loss": 3.5124935632134013,
      "tokens_seen": 1517486080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003240495555135237,
      "loss": 2.7617,
      "theoretical_loss": 3.5124802184454973,
      "tokens_seen": 1517551616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032403300548515226,
      "loss": 2.7601,
      "theoretical_loss": 3.512466874415235,
      "tokens_seen": 1517617152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003240164554567808,
      "loss": 2.9151,
      "theoretical_loss": 3.512453531122542,
      "tokens_seen": 1517682688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003239999054284093,
      "loss": 2.6003,
      "theoretical_loss": 3.512440188567346,
      "tokens_seen": 1517748224
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003239833554000378,
      "loss": 2.7014,
      "theoretical_loss": 3.512426846749573,
      "tokens_seen": 1517813760
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032396680537166635,
      "loss": 2.6373,
      "theoretical_loss": 3.512413505669152,
      "tokens_seen": 1517879296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003239502553432949,
      "loss": 2.6259,
      "theoretical_loss": 3.5124001653260093,
      "tokens_seen": 1517944832
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032393370531492337,
      "loss": 2.7437,
      "theoretical_loss": 3.512386825720072,
      "tokens_seen": 1518010368
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003239171552865519,
      "loss": 2.6119,
      "theoretical_loss": 3.5123734868512693,
      "tokens_seen": 1518075904
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032390060525818045,
      "loss": 2.6717,
      "theoretical_loss": 3.512360148719527,
      "tokens_seen": 1518141440
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000323884055229809,
      "loss": 2.8064,
      "theoretical_loss": 3.5123468113247736,
      "tokens_seen": 1518206976
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032386750520143747,
      "loss": 2.7671,
      "theoretical_loss": 3.5123334746669355,
      "tokens_seen": 1518272512
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000323850955173066,
      "loss": 2.482,
      "theoretical_loss": 3.512320138745941,
      "tokens_seen": 1518338048
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032383440514469454,
      "loss": 2.686,
      "theoretical_loss": 3.5123068035617173,
      "tokens_seen": 1518403584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003238178551163231,
      "loss": 2.6189,
      "theoretical_loss": 3.512293469114192,
      "tokens_seen": 1518469120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032380130508795156,
      "loss": 2.5457,
      "theoretical_loss": 3.5122801354032926,
      "tokens_seen": 1518534656
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003237847550595801,
      "loss": 2.6592,
      "theoretical_loss": 3.5122668024289467,
      "tokens_seen": 1518600192
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032376820503120864,
      "loss": 2.5906,
      "theoretical_loss": 3.5122534701910815,
      "tokens_seen": 1518665728
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003237516550028372,
      "loss": 2.3352,
      "theoretical_loss": 3.512240138689625,
      "tokens_seen": 1518731264
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032373510497446566,
      "loss": 2.8627,
      "theoretical_loss": 3.5122268079245043,
      "tokens_seen": 1518796800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003237185549460942,
      "loss": 2.7878,
      "theoretical_loss": 3.5122134778956475,
      "tokens_seen": 1518862336
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032370200491772274,
      "loss": 2.6707,
      "theoretical_loss": 3.5122001486029815,
      "tokens_seen": 1518927872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003236854548893512,
      "loss": 2.5073,
      "theoretical_loss": 3.512186820046434,
      "tokens_seen": 1518993408
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032366890486097976,
      "loss": 2.7095,
      "theoretical_loss": 3.512173492225933,
      "tokens_seen": 1519058944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032365235483260824,
      "loss": 2.6573,
      "theoretical_loss": 3.512160165141406,
      "tokens_seen": 1519124480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003236358048042368,
      "loss": 2.6833,
      "theoretical_loss": 3.5121468387927806,
      "tokens_seen": 1519190016
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003236192547758653,
      "loss": 2.5151,
      "theoretical_loss": 3.5121335131799847,
      "tokens_seen": 1519255552
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032360270474749385,
      "loss": 2.7039,
      "theoretical_loss": 3.512120188302945,
      "tokens_seen": 1519321088
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032358615471912234,
      "loss": 2.5067,
      "theoretical_loss": 3.5121068641615896,
      "tokens_seen": 1519386624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003235696046907509,
      "loss": 2.3756,
      "theoretical_loss": 3.512093540755847,
      "tokens_seen": 1519452160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003235530546623794,
      "loss": 2.8402,
      "theoretical_loss": 3.5120802180856434,
      "tokens_seen": 1519517696
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032353650463400795,
      "loss": 2.3631,
      "theoretical_loss": 3.512066896150907,
      "tokens_seen": 1519583232
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032351995460563643,
      "loss": 2.7758,
      "theoretical_loss": 3.5120535749515667,
      "tokens_seen": 1519648768
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032350340457726497,
      "loss": 2.6986,
      "theoretical_loss": 3.5120402544875486,
      "tokens_seen": 1519714304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003234868545488935,
      "loss": 2.6778,
      "theoretical_loss": 3.512026934758781,
      "tokens_seen": 1519779840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032347030452052204,
      "loss": 2.5652,
      "theoretical_loss": 3.5120136157651913,
      "tokens_seen": 1519845376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003234537544921506,
      "loss": 2.5754,
      "theoretical_loss": 3.5120002975067077,
      "tokens_seen": 1519910912
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032343720446377906,
      "loss": 2.6508,
      "theoretical_loss": 3.5119869799832584,
      "tokens_seen": 1519976448
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003234206544354076,
      "loss": 2.7906,
      "theoretical_loss": 3.5119736631947696,
      "tokens_seen": 1520041984
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032340410440703614,
      "loss": 2.5694,
      "theoretical_loss": 3.5119603471411702,
      "tokens_seen": 1520107520
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003233875543786647,
      "loss": 2.4703,
      "theoretical_loss": 3.511947031822388,
      "tokens_seen": 1520173056
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032337100435029316,
      "loss": 2.7318,
      "theoretical_loss": 3.5119337172383505,
      "tokens_seen": 1520238592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003233544543219217,
      "loss": 2.5681,
      "theoretical_loss": 3.5119204033889853,
      "tokens_seen": 1520304128
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032333790429355024,
      "loss": 2.6002,
      "theoretical_loss": 3.51190709027422,
      "tokens_seen": 1520369664
    },
    {
      "epoch": 0.54,
      "objective/train/avg_token_score": 0.007394648622721434,
      "objective/train/avg_weight": 0.991329550743103,
      "objective/train/docs_used": 861286,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.829930543899536,
      "objective/train/std_weight": 0.046335890889167786,
      "objective/train/theoretical_loss": 3.5118937778939836,
      "objective/train/tokens_used": 1540895200,
      "theoretical_loss": 3.5118937778939836,
      "tokens_seen": 1520435200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003233213542651788,
      "loss": 2.5563,
      "theoretical_loss": 3.5118937778939836,
      "tokens_seen": 1520435200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032330480423680726,
      "loss": 2.6855,
      "theoretical_loss": 3.5118804662482024,
      "tokens_seen": 1520500736
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003232882542084358,
      "loss": 2.8461,
      "theoretical_loss": 3.5118671553368053,
      "tokens_seen": 1520566272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032327170418006433,
      "loss": 2.6912,
      "theoretical_loss": 3.5118538451597194,
      "tokens_seen": 1520631808
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032325515415169287,
      "loss": 2.6268,
      "theoretical_loss": 3.511840535716873,
      "tokens_seen": 1520697344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032323860412332135,
      "loss": 2.8026,
      "theoretical_loss": 3.511827227008194,
      "tokens_seen": 1520762880
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003232220540949499,
      "loss": 2.414,
      "theoretical_loss": 3.51181391903361,
      "tokens_seen": 1520828416
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003232055040665784,
      "loss": 2.7364,
      "theoretical_loss": 3.511800611793049,
      "tokens_seen": 1520893952
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003231889540382069,
      "loss": 2.5245,
      "theoretical_loss": 3.5117873052864392,
      "tokens_seen": 1520959488
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032317240400983545,
      "loss": 2.528,
      "theoretical_loss": 3.511773999513708,
      "tokens_seen": 1521025024
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032315585398146393,
      "loss": 2.4883,
      "theoretical_loss": 3.5117606944747832,
      "tokens_seen": 1521090560
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032313930395309247,
      "loss": 2.7302,
      "theoretical_loss": 3.5117473901695933,
      "tokens_seen": 1521156096
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000323122753924721,
      "loss": 2.6691,
      "theoretical_loss": 3.511734086598066,
      "tokens_seen": 1521221632
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032310620389634954,
      "loss": 2.7201,
      "theoretical_loss": 3.511720783760129,
      "tokens_seen": 1521287168
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032308965386797803,
      "loss": 2.6137,
      "theoretical_loss": 3.5117074816557103,
      "tokens_seen": 1521352704
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032307310383960657,
      "loss": 2.456,
      "theoretical_loss": 3.511694180284738,
      "tokens_seen": 1521418240
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003230565538112351,
      "loss": 2.6687,
      "theoretical_loss": 3.51168087964714,
      "tokens_seen": 1521483776
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032304000378286364,
      "loss": 2.5242,
      "theoretical_loss": 3.511667579742845,
      "tokens_seen": 1521549312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003230234537544921,
      "loss": 2.5865,
      "theoretical_loss": 3.5116542805717796,
      "tokens_seen": 1521614848
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032300690372612066,
      "loss": 2.693,
      "theoretical_loss": 3.5116409821338728,
      "tokens_seen": 1521680384
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003229903536977492,
      "loss": 2.834,
      "theoretical_loss": 3.5116276844290524,
      "tokens_seen": 1521745920
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032297380366937774,
      "loss": 2.6459,
      "theoretical_loss": 3.511614387457246,
      "tokens_seen": 1521811456
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003229572536410062,
      "loss": 2.6292,
      "theoretical_loss": 3.511601091218382,
      "tokens_seen": 1521876992
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032294070361263476,
      "loss": 2.655,
      "theoretical_loss": 3.511587795712389,
      "tokens_seen": 1521942528
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003229241535842633,
      "loss": 2.6915,
      "theoretical_loss": 3.5115745009391937,
      "tokens_seen": 1522008064
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032290760355589183,
      "loss": 2.7388,
      "theoretical_loss": 3.5115612068987256,
      "tokens_seen": 1522073600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003228910535275203,
      "loss": 2.62,
      "theoretical_loss": 3.5115479135909116,
      "tokens_seen": 1522139136
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032287450349914885,
      "loss": 2.5615,
      "theoretical_loss": 3.5115346210156804,
      "tokens_seen": 1522204672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003228579534707774,
      "loss": 2.2891,
      "theoretical_loss": 3.51152132917296,
      "tokens_seen": 1522270208
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032284140344240593,
      "loss": 2.4351,
      "theoretical_loss": 3.5115080380626784,
      "tokens_seen": 1522335744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003228248534140344,
      "loss": 2.5117,
      "theoretical_loss": 3.5114947476847638,
      "tokens_seen": 1522401280
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032280830338566295,
      "loss": 2.8295,
      "theoretical_loss": 3.5114814580391442,
      "tokens_seen": 1522466816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003227917533572915,
      "loss": 2.5785,
      "theoretical_loss": 3.5114681691257483,
      "tokens_seen": 1522532352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032277520332892,
      "loss": 2.7181,
      "theoretical_loss": 3.5114548809445036,
      "tokens_seen": 1522597888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003227586533005485,
      "loss": 2.6663,
      "theoretical_loss": 3.511441593495338,
      "tokens_seen": 1522663424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032274210327217705,
      "loss": 2.4608,
      "theoretical_loss": 3.51142830677818,
      "tokens_seen": 1522728960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003227255532438056,
      "loss": 2.6135,
      "theoretical_loss": 3.511415020792959,
      "tokens_seen": 1522794496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032270900321543407,
      "loss": 2.8006,
      "theoretical_loss": 3.511401735539601,
      "tokens_seen": 1522860032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003226924531870626,
      "loss": 2.7159,
      "theoretical_loss": 3.5113884510180355,
      "tokens_seen": 1522925568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003226759031586911,
      "loss": 2.5841,
      "theoretical_loss": 3.51137516722819,
      "tokens_seen": 1522991104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003226593531303196,
      "loss": 2.6094,
      "theoretical_loss": 3.5113618841699936,
      "tokens_seen": 1523056640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032264280310194816,
      "loss": 2.561,
      "theoretical_loss": 3.511348601843374,
      "tokens_seen": 1523122176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003226262530735767,
      "loss": 2.7257,
      "theoretical_loss": 3.51133532024826,
      "tokens_seen": 1523187712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003226097030452052,
      "loss": 2.6083,
      "theoretical_loss": 3.511322039384578,
      "tokens_seen": 1523253248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003225931530168337,
      "loss": 2.8039,
      "theoretical_loss": 3.511308759252259,
      "tokens_seen": 1523318784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032257660298846226,
      "loss": 2.7052,
      "theoretical_loss": 3.5112954798512286,
      "tokens_seen": 1523384320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003225600529600908,
      "loss": 2.5099,
      "theoretical_loss": 3.5112822011814173,
      "tokens_seen": 1523449856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003225435029317193,
      "loss": 2.1081,
      "theoretical_loss": 3.511268923242752,
      "tokens_seen": 1523515392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003225269529033478,
      "loss": 2.5845,
      "theoretical_loss": 3.5112556460351616,
      "tokens_seen": 1523580928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032251040287497635,
      "loss": 2.94,
      "theoretical_loss": 3.511242369558574,
      "tokens_seen": 1523646464
    },
    {
      "epoch": 0.54,
      "objective/train/avg_token_score": 0.010019734501838684,
      "objective/train/avg_weight": 0.9882251024246216,
      "objective/train/docs_used": 863379,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.3968822956085205,
      "objective/train/std_weight": 0.0631350725889206,
      "objective/train/theoretical_loss": 3.5112290938129176,
      "objective/train/tokens_used": 1544172000,
      "theoretical_loss": 3.5112290938129176,
      "tokens_seen": 1523712000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003224938528466049,
      "loss": 2.58,
      "theoretical_loss": 3.5112290938129176,
      "tokens_seen": 1523712000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003224773028182334,
      "loss": 2.4573,
      "theoretical_loss": 3.5112158187981213,
      "tokens_seen": 1523777536
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003224607527898619,
      "loss": 2.8212,
      "theoretical_loss": 3.511202544514113,
      "tokens_seen": 1523843072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032244420276149045,
      "loss": 2.6548,
      "theoretical_loss": 3.5111892709608203,
      "tokens_seen": 1523908608
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000322427652733119,
      "loss": 2.6506,
      "theoretical_loss": 3.5111759981381727,
      "tokens_seen": 1523974144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032241110270474747,
      "loss": 2.7405,
      "theoretical_loss": 3.5111627260460985,
      "tokens_seen": 1524039680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000322394552676376,
      "loss": 2.7542,
      "theoretical_loss": 3.511149454684525,
      "tokens_seen": 1524105216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032237800264800455,
      "loss": 2.6225,
      "theoretical_loss": 3.511136184053382,
      "tokens_seen": 1524170752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003223614526196331,
      "loss": 2.682,
      "theoretical_loss": 3.5111229141525966,
      "tokens_seen": 1524236288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032234490259126157,
      "loss": 2.5837,
      "theoretical_loss": 3.511109644982098,
      "tokens_seen": 1524301824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003223283525628901,
      "loss": 2.5242,
      "theoretical_loss": 3.511096376541815,
      "tokens_seen": 1524367360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032231180253451864,
      "loss": 2.5925,
      "theoretical_loss": 3.5110831088316745,
      "tokens_seen": 1524432896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003222952525061472,
      "loss": 2.8748,
      "theoretical_loss": 3.511069841851606,
      "tokens_seen": 1524498432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032227870247777566,
      "loss": 2.7022,
      "theoretical_loss": 3.511056575601539,
      "tokens_seen": 1524563968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003222621524494042,
      "loss": 2.5979,
      "theoretical_loss": 3.5110433100813996,
      "tokens_seen": 1524629504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032224560242103274,
      "loss": 2.8038,
      "theoretical_loss": 3.5110300452911174,
      "tokens_seen": 1524695040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003222290523926613,
      "loss": 2.6846,
      "theoretical_loss": 3.511016781230621,
      "tokens_seen": 1524760576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032221250236428976,
      "loss": 2.6807,
      "theoretical_loss": 3.511003517899839,
      "tokens_seen": 1524826112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003221959523359183,
      "loss": 2.5901,
      "theoretical_loss": 3.5109902552987,
      "tokens_seen": 1524891648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003221794023075468,
      "loss": 2.5934,
      "theoretical_loss": 3.510976993427132,
      "tokens_seen": 1524957184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003221628522791753,
      "loss": 2.764,
      "theoretical_loss": 3.510963732285063,
      "tokens_seen": 1525022720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032214630225080385,
      "loss": 2.5556,
      "theoretical_loss": 3.510950471872423,
      "tokens_seen": 1525088256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003221297522224324,
      "loss": 2.6912,
      "theoretical_loss": 3.51093721218914,
      "tokens_seen": 1525153792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003221132021940609,
      "loss": 2.878,
      "theoretical_loss": 3.5109239532351415,
      "tokens_seen": 1525219328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003220966521656894,
      "loss": 2.7189,
      "theoretical_loss": 3.5109106950103577,
      "tokens_seen": 1525284864
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032208010213731795,
      "loss": 2.5755,
      "theoretical_loss": 3.5108974375147155,
      "tokens_seen": 1525350400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003220635521089465,
      "loss": 2.6244,
      "theoretical_loss": 3.510884180748145,
      "tokens_seen": 1525415936
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032204700208057497,
      "loss": 2.5908,
      "theoretical_loss": 3.5108709247105736,
      "tokens_seen": 1525481472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003220304520522035,
      "loss": 2.9498,
      "theoretical_loss": 3.510857669401931,
      "tokens_seen": 1525547008
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032201390202383205,
      "loss": 2.6103,
      "theoretical_loss": 3.5108444148221443,
      "tokens_seen": 1525612544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003219973519954606,
      "loss": 2.3938,
      "theoretical_loss": 3.5108311609711436,
      "tokens_seen": 1525678080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032198080196708907,
      "loss": 2.5183,
      "theoretical_loss": 3.510817907848857,
      "tokens_seen": 1525743616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003219642519387176,
      "loss": 2.6177,
      "theoretical_loss": 3.510804655455213,
      "tokens_seen": 1525809152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00032194770191034614,
      "loss": 2.664,
      "theoretical_loss": 3.51079140379014,
      "tokens_seen": 1525874688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0003219311518819747,
      "loss": 2.5279,
      "theoretical_loss": 3.5107781528535673,
      "tokens_seen": 1525940224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032191460185360316,
      "loss": 2.8152,
      "theoretical_loss": 3.5107649026454233,
      "tokens_seen": 1526005760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003218980518252317,
      "loss": 2.7799,
      "theoretical_loss": 3.510751653165636,
      "tokens_seen": 1526071296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032188150179686024,
      "loss": 2.6034,
      "theoretical_loss": 3.510738404414136,
      "tokens_seen": 1526136832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003218649517684888,
      "loss": 2.8471,
      "theoretical_loss": 3.5107251563908495,
      "tokens_seen": 1526202368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032184840174011726,
      "loss": 1.3805,
      "theoretical_loss": 3.510711909095707,
      "tokens_seen": 1526267904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003218318517117458,
      "loss": 2.5132,
      "theoretical_loss": 3.5106986625286365,
      "tokens_seen": 1526333440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032181530168337433,
      "loss": 2.7654,
      "theoretical_loss": 3.510685416689567,
      "tokens_seen": 1526398976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032179875165500287,
      "loss": 2.6897,
      "theoretical_loss": 3.5106721715784266,
      "tokens_seen": 1526464512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032178220162663136,
      "loss": 2.4321,
      "theoretical_loss": 3.510658927195145,
      "tokens_seen": 1526530048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003217656515982599,
      "loss": 2.5962,
      "theoretical_loss": 3.5106456835396505,
      "tokens_seen": 1526595584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032174910156988843,
      "loss": 2.7636,
      "theoretical_loss": 3.5106324406118716,
      "tokens_seen": 1526661120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003217325515415169,
      "loss": 2.4326,
      "theoretical_loss": 3.510619198411738,
      "tokens_seen": 1526726656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032171600151314545,
      "loss": 2.7639,
      "theoretical_loss": 3.5106059569391768,
      "tokens_seen": 1526792192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032169945148477393,
      "loss": 2.6803,
      "theoretical_loss": 3.510592716194118,
      "tokens_seen": 1526857728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032168290145640247,
      "loss": 2.3202,
      "theoretical_loss": 3.510579476176491,
      "tokens_seen": 1526923264
    },
    {
      "epoch": 0.55,
      "objective/train/avg_token_score": 0.018338985741138458,
      "objective/train/avg_weight": 0.9784134030342102,
      "objective/train/docs_used": 865312,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.677236795425415,
      "objective/train/std_weight": 0.08125273138284683,
      "objective/train/theoretical_loss": 3.5105662368862234,
      "objective/train/tokens_used": 1547448800,
      "theoretical_loss": 3.5105662368862234,
      "tokens_seen": 1526988800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000321666351428031,
      "loss": 2.7103,
      "theoretical_loss": 3.5105662368862234,
      "tokens_seen": 1526988800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032164980139965955,
      "loss": 2.5552,
      "theoretical_loss": 3.510552998323244,
      "tokens_seen": 1527054336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032163325137128803,
      "loss": 2.7971,
      "theoretical_loss": 3.5105397604874833,
      "tokens_seen": 1527119872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032161670134291657,
      "loss": 2.5655,
      "theoretical_loss": 3.510526523378868,
      "tokens_seen": 1527185408
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003216001513145451,
      "loss": 2.6683,
      "theoretical_loss": 3.510513286997328,
      "tokens_seen": 1527250944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032158360128617364,
      "loss": 2.6927,
      "theoretical_loss": 3.5105000513427926,
      "tokens_seen": 1527316480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003215670512578021,
      "loss": 2.8594,
      "theoretical_loss": 3.5104868164151894,
      "tokens_seen": 1527382016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032155050122943066,
      "loss": 2.5334,
      "theoretical_loss": 3.5104735822144484,
      "tokens_seen": 1527447552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003215339512010592,
      "loss": 2.6671,
      "theoretical_loss": 3.510460348740498,
      "tokens_seen": 1527513088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032151740117268774,
      "loss": 2.2486,
      "theoretical_loss": 3.5104471159932675,
      "tokens_seen": 1527578624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003215008511443162,
      "loss": 2.4507,
      "theoretical_loss": 3.510433883972686,
      "tokens_seen": 1527644160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032148430111594476,
      "loss": 2.585,
      "theoretical_loss": 3.510420652678681,
      "tokens_seen": 1527709696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003214677510875733,
      "loss": 2.6364,
      "theoretical_loss": 3.510407422111183,
      "tokens_seen": 1527775232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032145120105920184,
      "loss": 3.0315,
      "theoretical_loss": 3.5103941922701205,
      "tokens_seen": 1527840768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003214346510308303,
      "loss": 2.5818,
      "theoretical_loss": 3.5103809631554217,
      "tokens_seen": 1527906304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032141810100245886,
      "loss": 2.5589,
      "theoretical_loss": 3.5103677347670166,
      "tokens_seen": 1527971840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003214015509740874,
      "loss": 2.5392,
      "theoretical_loss": 3.510354507104834,
      "tokens_seen": 1528037376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032138500094571593,
      "loss": 2.7305,
      "theoretical_loss": 3.5103412801688023,
      "tokens_seen": 1528102912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003213684509173444,
      "loss": 2.6269,
      "theoretical_loss": 3.510328053958851,
      "tokens_seen": 1528168448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032135190088897295,
      "loss": 2.8304,
      "theoretical_loss": 3.5103148284749093,
      "tokens_seen": 1528233984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003213353508606015,
      "loss": 2.587,
      "theoretical_loss": 3.5103016037169055,
      "tokens_seen": 1528299520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032131880083223003,
      "loss": 2.5991,
      "theoretical_loss": 3.510288379684769,
      "tokens_seen": 1528365056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003213022508038585,
      "loss": 2.5879,
      "theoretical_loss": 3.510275156378429,
      "tokens_seen": 1528430592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032128570077548705,
      "loss": 2.9323,
      "theoretical_loss": 3.510261933797814,
      "tokens_seen": 1528496128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003212691507471156,
      "loss": 2.4965,
      "theoretical_loss": 3.510248711942854,
      "tokens_seen": 1528561664
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003212526007187441,
      "loss": 2.6111,
      "theoretical_loss": 3.5102354908134776,
      "tokens_seen": 1528627200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003212360506903726,
      "loss": 2.524,
      "theoretical_loss": 3.510222270409613,
      "tokens_seen": 1528692736
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003212195006620011,
      "loss": 2.7143,
      "theoretical_loss": 3.5102090507311905,
      "tokens_seen": 1528758272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032120295063362963,
      "loss": 2.5735,
      "theoretical_loss": 3.510195831778139,
      "tokens_seen": 1528823808
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032118640060525817,
      "loss": 2.7668,
      "theoretical_loss": 3.5101826135503873,
      "tokens_seen": 1528889344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003211698505768867,
      "loss": 2.324,
      "theoretical_loss": 3.5101693960478646,
      "tokens_seen": 1528954880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003211533005485152,
      "loss": 2.7116,
      "theoretical_loss": 3.5101561792705,
      "tokens_seen": 1529020416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003211367505201437,
      "loss": 2.5938,
      "theoretical_loss": 3.5101429632182226,
      "tokens_seen": 1529085952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032112020049177226,
      "loss": 2.6494,
      "theoretical_loss": 3.5101297478909617,
      "tokens_seen": 1529151488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003211036504634008,
      "loss": 2.7625,
      "theoretical_loss": 3.510116533288646,
      "tokens_seen": 1529217024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003210871004350293,
      "loss": 2.5202,
      "theoretical_loss": 3.5101033194112055,
      "tokens_seen": 1529282560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003210705504066578,
      "loss": 2.6954,
      "theoretical_loss": 3.510090106258569,
      "tokens_seen": 1529348096
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032105400037828636,
      "loss": 2.5658,
      "theoretical_loss": 3.5100768938306652,
      "tokens_seen": 1529413632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003210374503499149,
      "loss": 2.3749,
      "theoretical_loss": 3.5100636821274236,
      "tokens_seen": 1529479168
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003210209003215434,
      "loss": 2.699,
      "theoretical_loss": 3.510050471148774,
      "tokens_seen": 1529544704
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003210043502931719,
      "loss": 2.5816,
      "theoretical_loss": 3.510037260894645,
      "tokens_seen": 1529610240
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032098780026480045,
      "loss": 2.6848,
      "theoretical_loss": 3.5100240513649656,
      "tokens_seen": 1529675776
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000320971250236429,
      "loss": 2.6993,
      "theoretical_loss": 3.510010842559666,
      "tokens_seen": 1529741312
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003209547002080575,
      "loss": 2.3573,
      "theoretical_loss": 3.5099976344786743,
      "tokens_seen": 1529806848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000320938150179686,
      "loss": 2.4916,
      "theoretical_loss": 3.5099844271219203,
      "tokens_seen": 1529872384
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032092160015131455,
      "loss": 2.6405,
      "theoretical_loss": 3.5099712204893327,
      "tokens_seen": 1529937920
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003209050501229431,
      "loss": 2.6027,
      "theoretical_loss": 3.5099580145808424,
      "tokens_seen": 1530003456
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032088850009457157,
      "loss": 2.5943,
      "theoretical_loss": 3.509944809396377,
      "tokens_seen": 1530068992
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003208719500662001,
      "loss": 2.6648,
      "theoretical_loss": 3.5099316049358666,
      "tokens_seen": 1530134528
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032085540003782864,
      "loss": 2.5723,
      "theoretical_loss": 3.50991840119924,
      "tokens_seen": 1530200064
    },
    {
      "epoch": 0.55,
      "objective/train/avg_token_score": 0.005584646016359329,
      "objective/train/avg_weight": 0.9934273362159729,
      "objective/train/docs_used": 867418,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6337409019470215,
      "objective/train/std_weight": 0.04247665032744408,
      "objective/train/theoretical_loss": 3.5099051981864275,
      "objective/train/tokens_used": 1550725600,
      "theoretical_loss": 3.5099051981864275,
      "tokens_seen": 1530265600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003208388500094572,
      "loss": 2.6408,
      "theoretical_loss": 3.5099051981864275,
      "tokens_seen": 1530265600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032082229998108567,
      "loss": 2.6592,
      "theoretical_loss": 3.509891995897357,
      "tokens_seen": 1530331136
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003208057499527142,
      "loss": 2.5071,
      "theoretical_loss": 3.509878794331959,
      "tokens_seen": 1530396672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032078919992434274,
      "loss": 2.4507,
      "theoretical_loss": 3.5098655934901624,
      "tokens_seen": 1530462208
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003207726498959713,
      "loss": 2.6411,
      "theoretical_loss": 3.5098523933718964,
      "tokens_seen": 1530527744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032075609986759976,
      "loss": 2.7833,
      "theoretical_loss": 3.5098391939770908,
      "tokens_seen": 1530593280
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003207395498392283,
      "loss": 2.6131,
      "theoretical_loss": 3.5098259953056745,
      "tokens_seen": 1530658816
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003207229998108568,
      "loss": 2.5315,
      "theoretical_loss": 3.5098127973575775,
      "tokens_seen": 1530724352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003207064497824853,
      "loss": 2.7413,
      "theoretical_loss": 3.5097996001327285,
      "tokens_seen": 1530789888
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032068989975411386,
      "loss": 2.5507,
      "theoretical_loss": 3.509786403631057,
      "tokens_seen": 1530855424
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003206733497257424,
      "loss": 2.6874,
      "theoretical_loss": 3.509773207852493,
      "tokens_seen": 1530920960
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003206567996973709,
      "loss": 1.7841,
      "theoretical_loss": 3.5097600127969657,
      "tokens_seen": 1530986496
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003206402496689994,
      "loss": 2.7435,
      "theoretical_loss": 3.509746818464404,
      "tokens_seen": 1531052032
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032062369964062795,
      "loss": 2.8098,
      "theoretical_loss": 3.5097336248547375,
      "tokens_seen": 1531117568
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003206071496122565,
      "loss": 2.7406,
      "theoretical_loss": 3.5097204319678967,
      "tokens_seen": 1531183104
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000320590599583885,
      "loss": 2.4858,
      "theoretical_loss": 3.5097072398038094,
      "tokens_seen": 1531248640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003205740495555135,
      "loss": 2.4199,
      "theoretical_loss": 3.509694048362406,
      "tokens_seen": 1531314176
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032055749952714205,
      "loss": 2.5239,
      "theoretical_loss": 3.509680857643617,
      "tokens_seen": 1531379712
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003205409494987706,
      "loss": 2.5488,
      "theoretical_loss": 3.5096676676473697,
      "tokens_seen": 1531445248
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032052439947039907,
      "loss": 2.678,
      "theoretical_loss": 3.5096544783735952,
      "tokens_seen": 1531510784
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003205078494420276,
      "loss": 2.5668,
      "theoretical_loss": 3.509641289822222,
      "tokens_seen": 1531576320
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032049129941365615,
      "loss": 2.7704,
      "theoretical_loss": 3.5096281019931803,
      "tokens_seen": 1531641856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003204747493852847,
      "loss": 2.7088,
      "theoretical_loss": 3.5096149148864,
      "tokens_seen": 1531707392
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032045819935691317,
      "loss": 2.4081,
      "theoretical_loss": 3.5096017285018095,
      "tokens_seen": 1531772928
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003204416493285417,
      "loss": 2.6697,
      "theoretical_loss": 3.5095885428393396,
      "tokens_seen": 1531838464
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032042509930017024,
      "loss": 2.7449,
      "theoretical_loss": 3.5095753578989184,
      "tokens_seen": 1531904000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003204085492717988,
      "loss": 2.3081,
      "theoretical_loss": 3.509562173680477,
      "tokens_seen": 1531969536
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032039199924342726,
      "loss": 2.6199,
      "theoretical_loss": 3.5095489901839443,
      "tokens_seen": 1532035072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003203754492150558,
      "loss": 2.6726,
      "theoretical_loss": 3.5095358074092493,
      "tokens_seen": 1532100608
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032035889918668434,
      "loss": 2.5465,
      "theoretical_loss": 3.5095226253563228,
      "tokens_seen": 1532166144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003203423491583129,
      "loss": 2.7628,
      "theoretical_loss": 3.5095094440250936,
      "tokens_seen": 1532231680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032032579912994136,
      "loss": 2.6082,
      "theoretical_loss": 3.5094962634154916,
      "tokens_seen": 1532297216
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003203092491015699,
      "loss": 2.6789,
      "theoretical_loss": 3.5094830835274458,
      "tokens_seen": 1532362752
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032029269907319843,
      "loss": 2.7635,
      "theoretical_loss": 3.509469904360887,
      "tokens_seen": 1532428288
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032027614904482697,
      "loss": 2.7255,
      "theoretical_loss": 3.509456725915744,
      "tokens_seen": 1532493824
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032025959901645545,
      "loss": 2.7151,
      "theoretical_loss": 3.509443548191947,
      "tokens_seen": 1532559360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032024304898808394,
      "loss": 2.7156,
      "theoretical_loss": 3.509430371189425,
      "tokens_seen": 1532624896
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003202264989597125,
      "loss": 2.5311,
      "theoretical_loss": 3.5094171949081083,
      "tokens_seen": 1532690432
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000320209948931341,
      "loss": 2.4038,
      "theoretical_loss": 3.509404019347926,
      "tokens_seen": 1532755968
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032019339890296955,
      "loss": 2.2475,
      "theoretical_loss": 3.5093908445088084,
      "tokens_seen": 1532821504
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032017684887459803,
      "loss": 2.6191,
      "theoretical_loss": 3.509377670390685,
      "tokens_seen": 1532887040
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032016029884622657,
      "loss": 2.8206,
      "theoretical_loss": 3.5093644969934856,
      "tokens_seen": 1532952576
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003201437488178551,
      "loss": 2.454,
      "theoretical_loss": 3.50935132431714,
      "tokens_seen": 1533018112
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032012719878948365,
      "loss": 2.6028,
      "theoretical_loss": 3.509338152361577,
      "tokens_seen": 1533083648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032011064876111213,
      "loss": 2.5532,
      "theoretical_loss": 3.5093249811267277,
      "tokens_seen": 1533149184
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032009409873274067,
      "loss": 2.7173,
      "theoretical_loss": 3.5093118106125214,
      "tokens_seen": 1533214720
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003200775487043692,
      "loss": 2.9835,
      "theoretical_loss": 3.509298640818888,
      "tokens_seen": 1533280256
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032006099867599774,
      "loss": 2.6227,
      "theoretical_loss": 3.5092854717457564,
      "tokens_seen": 1533345792
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003200444486476262,
      "loss": 2.5949,
      "theoretical_loss": 3.5092723033930566,
      "tokens_seen": 1533411328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00032002789861925476,
      "loss": 2.693,
      "theoretical_loss": 3.5092591357607197,
      "tokens_seen": 1533476864
    },
    {
      "epoch": 0.55,
      "objective/train/avg_token_score": 0.013790605589747429,
      "objective/train/avg_weight": 0.9837678670883179,
      "objective/train/docs_used": 869264,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.724614381790161,
      "objective/train/std_weight": 0.07405279576778412,
      "objective/train/theoretical_loss": 3.509245968848674,
      "objective/train/tokens_used": 1554002400,
      "theoretical_loss": 3.509245968848674,
      "tokens_seen": 1533542400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003200113485908833,
      "loss": 2.5925,
      "theoretical_loss": 3.509245968848674,
      "tokens_seen": 1533542400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031999479856251184,
      "loss": 2.8238,
      "theoretical_loss": 3.5092328026568507,
      "tokens_seen": 1533607936
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003199782485341403,
      "loss": 2.4658,
      "theoretical_loss": 3.5092196371851783,
      "tokens_seen": 1533673472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031996169850576886,
      "loss": 2.4307,
      "theoretical_loss": 3.509206472433587,
      "tokens_seen": 1533739008
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003199451484773974,
      "loss": 2.7432,
      "theoretical_loss": 3.5091933084020077,
      "tokens_seen": 1533804544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031992859844902593,
      "loss": 2.4837,
      "theoretical_loss": 3.509180145090369,
      "tokens_seen": 1533870080
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003199120484206544,
      "loss": 2.6955,
      "theoretical_loss": 3.509166982498601,
      "tokens_seen": 1533935616
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031989549839228296,
      "loss": 2.648,
      "theoretical_loss": 3.5091538206266337,
      "tokens_seen": 1534001152
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003198789483639115,
      "loss": 2.6448,
      "theoretical_loss": 3.5091406594743972,
      "tokens_seen": 1534066688
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031986239833554003,
      "loss": 2.5919,
      "theoretical_loss": 3.5091274990418215,
      "tokens_seen": 1534132224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003198458483071685,
      "loss": 2.3271,
      "theoretical_loss": 3.5091143393288355,
      "tokens_seen": 1534197760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031982929827879705,
      "loss": 2.6098,
      "theoretical_loss": 3.509101180335371,
      "tokens_seen": 1534263296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003198127482504256,
      "loss": 2.535,
      "theoretical_loss": 3.509088022061356,
      "tokens_seen": 1534328832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003197961982220541,
      "loss": 2.492,
      "theoretical_loss": 3.5090748645067213,
      "tokens_seen": 1534394368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003197796481936826,
      "loss": 2.6304,
      "theoretical_loss": 3.509061707671397,
      "tokens_seen": 1534459904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003197630981653111,
      "loss": 2.4259,
      "theoretical_loss": 3.5090485515553134,
      "tokens_seen": 1534525440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031974654813693963,
      "loss": 2.5802,
      "theoretical_loss": 3.5090353961583993,
      "tokens_seen": 1534590976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031972999810856817,
      "loss": 2.7599,
      "theoretical_loss": 3.509022241480585,
      "tokens_seen": 1534656512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003197134480801967,
      "loss": 2.7321,
      "theoretical_loss": 3.509009087521801,
      "tokens_seen": 1534722048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003196968980518252,
      "loss": 2.8044,
      "theoretical_loss": 3.5089959342819776,
      "tokens_seen": 1534787584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003196803480234537,
      "loss": 2.5495,
      "theoretical_loss": 3.508982781761044,
      "tokens_seen": 1534853120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031966379799508226,
      "loss": 2.6748,
      "theoretical_loss": 3.50896962995893,
      "tokens_seen": 1534918656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003196472479667108,
      "loss": 2.3002,
      "theoretical_loss": 3.5089564788755667,
      "tokens_seen": 1534984192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003196306979383393,
      "loss": 2.6447,
      "theoretical_loss": 3.5089433285108838,
      "tokens_seen": 1535049728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003196141479099678,
      "loss": 2.7413,
      "theoretical_loss": 3.5089301788648104,
      "tokens_seen": 1535115264
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031959759788159636,
      "loss": 2.4008,
      "theoretical_loss": 3.5089170299372774,
      "tokens_seen": 1535180800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003195810478532249,
      "loss": 2.8274,
      "theoretical_loss": 3.508903881728215,
      "tokens_seen": 1535246336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003195644978248534,
      "loss": 2.3928,
      "theoretical_loss": 3.5088907342375526,
      "tokens_seen": 1535311872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003195479477964819,
      "loss": 2.7357,
      "theoretical_loss": 3.5088775874652214,
      "tokens_seen": 1535377408
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031953139776811046,
      "loss": 2.6013,
      "theoretical_loss": 3.5088644414111503,
      "tokens_seen": 1535442944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000319514847739739,
      "loss": 2.8756,
      "theoretical_loss": 3.50885129607527,
      "tokens_seen": 1535508480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031949829771136753,
      "loss": 2.6264,
      "theoretical_loss": 3.5088381514575104,
      "tokens_seen": 1535574016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000319481747682996,
      "loss": 2.693,
      "theoretical_loss": 3.5088250075578014,
      "tokens_seen": 1535639552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031946519765462455,
      "loss": 2.8856,
      "theoretical_loss": 3.508811864376074,
      "tokens_seen": 1535705088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003194486476262531,
      "loss": 2.8651,
      "theoretical_loss": 3.5087987219122576,
      "tokens_seen": 1535770624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003194320975978816,
      "loss": 2.5614,
      "theoretical_loss": 3.5087855801662826,
      "tokens_seen": 1535836160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003194155475695101,
      "loss": 2.849,
      "theoretical_loss": 3.5087724391380792,
      "tokens_seen": 1535901696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031939899754113865,
      "loss": 2.7482,
      "theoretical_loss": 3.5087592988275773,
      "tokens_seen": 1535967232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003193824475127672,
      "loss": 2.5412,
      "theoretical_loss": 3.5087461592347076,
      "tokens_seen": 1536032768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003193658974843957,
      "loss": 2.4195,
      "theoretical_loss": 3.5087330203594,
      "tokens_seen": 1536098304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003193493474560242,
      "loss": 2.7993,
      "theoretical_loss": 3.508719882201584,
      "tokens_seen": 1536163840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031933279742765274,
      "loss": 2.702,
      "theoretical_loss": 3.5087067447611906,
      "tokens_seen": 1536229376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003193162473992813,
      "loss": 2.5739,
      "theoretical_loss": 3.5086936080381506,
      "tokens_seen": 1536294912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003192996973709098,
      "loss": 2.7877,
      "theoretical_loss": 3.5086804720323927,
      "tokens_seen": 1536360448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003192831473425383,
      "loss": 2.4311,
      "theoretical_loss": 3.5086673367438483,
      "tokens_seen": 1536425984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003192665973141668,
      "loss": 2.4995,
      "theoretical_loss": 3.5086542021724476,
      "tokens_seen": 1536491520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003192500472857953,
      "loss": 2.6371,
      "theoretical_loss": 3.50864106831812,
      "tokens_seen": 1536557056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031923349725742386,
      "loss": 2.6554,
      "theoretical_loss": 3.508627935180797,
      "tokens_seen": 1536622592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003192169472290524,
      "loss": 2.63,
      "theoretical_loss": 3.508614802760408,
      "tokens_seen": 1536688128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003192003972006809,
      "loss": 2.565,
      "theoretical_loss": 3.5086016710568835,
      "tokens_seen": 1536753664
    },
    {
      "epoch": 0.55,
      "objective/train/avg_token_score": 0.014612684026360512,
      "objective/train/avg_weight": 0.982804536819458,
      "objective/train/docs_used": 871030,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5946381092071533,
      "objective/train/std_weight": 0.07685697823762894,
      "objective/train/theoretical_loss": 3.5085885400701535,
      "objective/train/tokens_used": 1557279200,
      "theoretical_loss": 3.5085885400701535,
      "tokens_seen": 1536819200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003191838471723094,
      "loss": 2.6281,
      "theoretical_loss": 3.5085885400701535,
      "tokens_seen": 1536819200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031916729714393796,
      "loss": 2.5804,
      "theoretical_loss": 3.508575409800149,
      "tokens_seen": 1536884736
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003191507471155665,
      "loss": 2.7728,
      "theoretical_loss": 3.5085622802467995,
      "tokens_seen": 1536950272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000319134197087195,
      "loss": 2.7363,
      "theoretical_loss": 3.508549151410037,
      "tokens_seen": 1537015808
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003191176470588235,
      "loss": 3.0819,
      "theoretical_loss": 3.508536023289789,
      "tokens_seen": 1537081344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031910109703045205,
      "loss": 2.8748,
      "theoretical_loss": 3.5085228958859886,
      "tokens_seen": 1537146880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003190845470020806,
      "loss": 2.609,
      "theoretical_loss": 3.5085097691985645,
      "tokens_seen": 1537212416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003190679969737091,
      "loss": 2.6891,
      "theoretical_loss": 3.508496643227448,
      "tokens_seen": 1537277952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003190514469453376,
      "loss": 2.6093,
      "theoretical_loss": 3.508483517972569,
      "tokens_seen": 1537343488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031903489691696615,
      "loss": 2.4391,
      "theoretical_loss": 3.5084703934338575,
      "tokens_seen": 1537409024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003190183468885947,
      "loss": 2.8092,
      "theoretical_loss": 3.5084572696112444,
      "tokens_seen": 1537474560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031900179686022317,
      "loss": 2.4808,
      "theoretical_loss": 3.5084441465046607,
      "tokens_seen": 1537540096
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003189852468318517,
      "loss": 2.6808,
      "theoretical_loss": 3.508431024114035,
      "tokens_seen": 1537605632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031896869680348024,
      "loss": 2.4929,
      "theoretical_loss": 3.5084179024393,
      "tokens_seen": 1537671168
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003189521467751088,
      "loss": 2.3555,
      "theoretical_loss": 3.508404781480385,
      "tokens_seen": 1537736704
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031893559674673727,
      "loss": 2.6403,
      "theoretical_loss": 3.50839166123722,
      "tokens_seen": 1537802240
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003189190467183658,
      "loss": 2.7098,
      "theoretical_loss": 3.508378541709736,
      "tokens_seen": 1537867776
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031890249668999434,
      "loss": 2.4843,
      "theoretical_loss": 3.508365422897864,
      "tokens_seen": 1537933312
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003188859466616229,
      "loss": 2.7628,
      "theoretical_loss": 3.508352304801533,
      "tokens_seen": 1537998848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031886939663325136,
      "loss": 2.6577,
      "theoretical_loss": 3.508339187420675,
      "tokens_seen": 1538064384
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003188528466048799,
      "loss": 2.3899,
      "theoretical_loss": 3.5083260707552193,
      "tokens_seen": 1538129920
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031883629657650844,
      "loss": 2.4444,
      "theoretical_loss": 3.5083129548050973,
      "tokens_seen": 1538195456
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000318819746548137,
      "loss": 2.3552,
      "theoretical_loss": 3.508299839570239,
      "tokens_seen": 1538260992
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031880319651976546,
      "loss": 2.5138,
      "theoretical_loss": 3.508286725050575,
      "tokens_seen": 1538326528
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000318786646491394,
      "loss": 2.5233,
      "theoretical_loss": 3.5082736112460364,
      "tokens_seen": 1538392064
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003187700964630225,
      "loss": 2.6813,
      "theoretical_loss": 3.5082604981565533,
      "tokens_seen": 1538457600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000318753546434651,
      "loss": 2.7522,
      "theoretical_loss": 3.508247385782056,
      "tokens_seen": 1538523136
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031873699640627955,
      "loss": 2.7238,
      "theoretical_loss": 3.508234274122475,
      "tokens_seen": 1538588672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031872044637790804,
      "loss": 2.5781,
      "theoretical_loss": 3.5082211631777414,
      "tokens_seen": 1538654208
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003187038963495366,
      "loss": 2.7251,
      "theoretical_loss": 3.5082080529477855,
      "tokens_seen": 1538719744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003186873463211651,
      "loss": 2.5585,
      "theoretical_loss": 3.5081949434325383,
      "tokens_seen": 1538785280
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031867079629279365,
      "loss": 2.5857,
      "theoretical_loss": 3.5081818346319293,
      "tokens_seen": 1538850816
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031865424626442213,
      "loss": 2.52,
      "theoretical_loss": 3.5081687265458905,
      "tokens_seen": 1538916352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031863769623605067,
      "loss": 2.7343,
      "theoretical_loss": 3.5081556191743517,
      "tokens_seen": 1538981888
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003186211462076792,
      "loss": 2.3952,
      "theoretical_loss": 3.5081425125172436,
      "tokens_seen": 1539047424
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031860459617930775,
      "loss": 2.4565,
      "theoretical_loss": 3.508129406574497,
      "tokens_seen": 1539112960
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031858804615093623,
      "loss": 2.5256,
      "theoretical_loss": 3.5081163013460426,
      "tokens_seen": 1539178496
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031857149612256477,
      "loss": 2.1973,
      "theoretical_loss": 3.5081031968318106,
      "tokens_seen": 1539244032
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003185549460941933,
      "loss": 2.439,
      "theoretical_loss": 3.5080900930317327,
      "tokens_seen": 1539309568
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031853839606582184,
      "loss": 2.7895,
      "theoretical_loss": 3.508076989945738,
      "tokens_seen": 1539375104
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003185218460374503,
      "loss": 2.4155,
      "theoretical_loss": 3.5080638875737584,
      "tokens_seen": 1539440640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031850529600907886,
      "loss": 2.5161,
      "theoretical_loss": 3.5080507859157244,
      "tokens_seen": 1539506176
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003184887459807074,
      "loss": 2.3743,
      "theoretical_loss": 3.5080376849715664,
      "tokens_seen": 1539571712
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031847219595233594,
      "loss": 2.441,
      "theoretical_loss": 3.5080245847412153,
      "tokens_seen": 1539637248
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003184556459239644,
      "loss": 2.4034,
      "theoretical_loss": 3.5080114852246025,
      "tokens_seen": 1539702784
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031843909589559296,
      "loss": 2.7011,
      "theoretical_loss": 3.507998386421657,
      "tokens_seen": 1539768320
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003184225458672215,
      "loss": 2.6086,
      "theoretical_loss": 3.5079852883323115,
      "tokens_seen": 1539833856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031840599583885003,
      "loss": 2.7592,
      "theoretical_loss": 3.5079721909564956,
      "tokens_seen": 1539899392
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003183894458104785,
      "loss": 2.4881,
      "theoretical_loss": 3.50795909429414,
      "tokens_seen": 1539964928
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031837289578210705,
      "loss": 2.732,
      "theoretical_loss": 3.5079459983451757,
      "tokens_seen": 1540030464
    },
    {
      "epoch": 0.55,
      "objective/train/avg_token_score": 0.01922745257616043,
      "objective/train/avg_weight": 0.9773696660995483,
      "objective/train/docs_used": 872615,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.113034725189209,
      "objective/train/std_weight": 0.09107836335897446,
      "objective/train/theoretical_loss": 3.507932903109534,
      "objective/train/tokens_used": 1560556000,
      "theoretical_loss": 3.507932903109534,
      "tokens_seen": 1540096000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003183563457537356,
      "loss": 2.6534,
      "theoretical_loss": 3.507932903109534,
      "tokens_seen": 1540096000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031833979572536413,
      "loss": 2.6835,
      "theoretical_loss": 3.5079198085871455,
      "tokens_seen": 1540161536
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003183232456969926,
      "loss": 2.6716,
      "theoretical_loss": 3.50790671477794,
      "tokens_seen": 1540227072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031830669566862115,
      "loss": 2.4284,
      "theoretical_loss": 3.50789362168185,
      "tokens_seen": 1540292608
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031829014564024963,
      "loss": 2.3897,
      "theoretical_loss": 3.507880529298805,
      "tokens_seen": 1540358144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031827359561187817,
      "loss": 2.5007,
      "theoretical_loss": 3.507867437628736,
      "tokens_seen": 1540423680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003182570455835067,
      "loss": 2.5287,
      "theoretical_loss": 3.507854346671575,
      "tokens_seen": 1540489216
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031824049555513525,
      "loss": 2.5668,
      "theoretical_loss": 3.507841256427251,
      "tokens_seen": 1540554752
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031822394552676373,
      "loss": 2.4788,
      "theoretical_loss": 3.5078281668956963,
      "tokens_seen": 1540620288
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031820739549839227,
      "loss": 2.598,
      "theoretical_loss": 3.507815078076841,
      "tokens_seen": 1540685824
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003181908454700208,
      "loss": 2.5446,
      "theoretical_loss": 3.5078019899706163,
      "tokens_seen": 1540751360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031817429544164934,
      "loss": 2.8456,
      "theoretical_loss": 3.5077889025769533,
      "tokens_seen": 1540816896
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003181577454132778,
      "loss": 2.5069,
      "theoretical_loss": 3.5077758158957826,
      "tokens_seen": 1540882432
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031814119538490636,
      "loss": 2.6215,
      "theoretical_loss": 3.507762729927035,
      "tokens_seen": 1540947968
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003181246453565349,
      "loss": 2.6258,
      "theoretical_loss": 3.5077496446706418,
      "tokens_seen": 1541013504
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031810809532816344,
      "loss": 2.5783,
      "theoretical_loss": 3.5077365601265336,
      "tokens_seen": 1541079040
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003180915452997919,
      "loss": 2.5261,
      "theoretical_loss": 3.5077234762946414,
      "tokens_seen": 1541144576
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031807499527142046,
      "loss": 2.6132,
      "theoretical_loss": 3.5077103931748965,
      "tokens_seen": 1541210112
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000318058445243049,
      "loss": 2.7194,
      "theoretical_loss": 3.507697310767229,
      "tokens_seen": 1541275648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031804189521467753,
      "loss": 2.6652,
      "theoretical_loss": 3.507684229071571,
      "tokens_seen": 1541341184
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000318025345186306,
      "loss": 2.5309,
      "theoretical_loss": 3.5076711480878524,
      "tokens_seen": 1541406720
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031800879515793455,
      "loss": 2.7479,
      "theoretical_loss": 3.5076580678160054,
      "tokens_seen": 1541472256
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003179922451295631,
      "loss": 2.61,
      "theoretical_loss": 3.5076449882559597,
      "tokens_seen": 1541537792
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031797569510119163,
      "loss": 2.491,
      "theoretical_loss": 3.507631909407647,
      "tokens_seen": 1541603328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003179591450728201,
      "loss": 2.7406,
      "theoretical_loss": 3.5076188312709986,
      "tokens_seen": 1541668864
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031794259504444865,
      "loss": 2.6251,
      "theoretical_loss": 3.5076057538459446,
      "tokens_seen": 1541734400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003179260450160772,
      "loss": 2.6409,
      "theoretical_loss": 3.5075926771324166,
      "tokens_seen": 1541799936
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003179094949877057,
      "loss": 2.4882,
      "theoretical_loss": 3.507579601130346,
      "tokens_seen": 1541865472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003178929449593342,
      "loss": 2.7068,
      "theoretical_loss": 3.5075665258396636,
      "tokens_seen": 1541931008
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031787639493096275,
      "loss": 2.5736,
      "theoretical_loss": 3.5075534512603,
      "tokens_seen": 1541996544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003178598449025913,
      "loss": 2.5158,
      "theoretical_loss": 3.5075403773921865,
      "tokens_seen": 1542062080
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003178432948742198,
      "loss": 2.6105,
      "theoretical_loss": 3.5075273042352544,
      "tokens_seen": 1542127616
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003178267448458483,
      "loss": 2.6054,
      "theoretical_loss": 3.507514231789435,
      "tokens_seen": 1542193152
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031781019481747684,
      "loss": 2.8066,
      "theoretical_loss": 3.507501160054659,
      "tokens_seen": 1542258688
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003177936447891053,
      "loss": 2.7031,
      "theoretical_loss": 3.5074880890308577,
      "tokens_seen": 1542324224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031777709476073386,
      "loss": 2.5933,
      "theoretical_loss": 3.5074750187179617,
      "tokens_seen": 1542389760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003177605447323624,
      "loss": 2.6595,
      "theoretical_loss": 3.507461949115903,
      "tokens_seen": 1542455296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003177439947039909,
      "loss": 2.8231,
      "theoretical_loss": 3.5074488802246115,
      "tokens_seen": 1542520832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003177274446756194,
      "loss": 2.6434,
      "theoretical_loss": 3.5074358120440206,
      "tokens_seen": 1542586368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031771089464724796,
      "loss": 2.7552,
      "theoretical_loss": 3.5074227445740584,
      "tokens_seen": 1542651904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003176943446188765,
      "loss": 2.672,
      "theoretical_loss": 3.5074096778146586,
      "tokens_seen": 1542717440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000317677794590505,
      "loss": 2.6891,
      "theoretical_loss": 3.507396611765752,
      "tokens_seen": 1542782976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003176612445621335,
      "loss": 2.5965,
      "theoretical_loss": 3.5073835464272682,
      "tokens_seen": 1542848512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031764469453376206,
      "loss": 2.6045,
      "theoretical_loss": 3.50737048179914,
      "tokens_seen": 1542914048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003176281445053906,
      "loss": 2.7141,
      "theoretical_loss": 3.5073574178812974,
      "tokens_seen": 1542979584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003176115944770191,
      "loss": 2.6321,
      "theoretical_loss": 3.5073443546736733,
      "tokens_seen": 1543045120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003175950444486476,
      "loss": 2.51,
      "theoretical_loss": 3.507331292176197,
      "tokens_seen": 1543110656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031757849442027615,
      "loss": 2.2305,
      "theoretical_loss": 3.507318230388801,
      "tokens_seen": 1543176192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003175619443919047,
      "loss": 2.5913,
      "theoretical_loss": 3.507305169311416,
      "tokens_seen": 1543241728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031754539436353317,
      "loss": 2.6646,
      "theoretical_loss": 3.507292108943974,
      "tokens_seen": 1543307264
    },
    {
      "epoch": 0.55,
      "objective/train/avg_token_score": 0.007569483481347561,
      "objective/train/avg_weight": 0.9910898208618164,
      "objective/train/docs_used": 874504,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7507336139678955,
      "objective/train/std_weight": 0.04726743325591087,
      "objective/train/theoretical_loss": 3.5072790492864057,
      "objective/train/tokens_used": 1563832800,
      "theoretical_loss": 3.5072790492864057,
      "tokens_seen": 1543372800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003175288443351617,
      "loss": 2.7469,
      "theoretical_loss": 3.5072790492864057,
      "tokens_seen": 1543372800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031751229430679025,
      "loss": 2.5841,
      "theoretical_loss": 3.5072659903386416,
      "tokens_seen": 1543438336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003174957442784188,
      "loss": 2.6486,
      "theoretical_loss": 3.5072529321006147,
      "tokens_seen": 1543503872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031747919425004727,
      "loss": 2.5484,
      "theoretical_loss": 3.507239874572255,
      "tokens_seen": 1543569408
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003174626442216758,
      "loss": 2.5069,
      "theoretical_loss": 3.5072268177534935,
      "tokens_seen": 1543634944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031744609419330434,
      "loss": 2.2333,
      "theoretical_loss": 3.5072137616442625,
      "tokens_seen": 1543700480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003174295441649329,
      "loss": 2.5089,
      "theoretical_loss": 3.5072007062444936,
      "tokens_seen": 1543766016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031741299413656136,
      "loss": 2.5366,
      "theoretical_loss": 3.5071876515541174,
      "tokens_seen": 1543831552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003173964441081899,
      "loss": 2.5837,
      "theoretical_loss": 3.507174597573065,
      "tokens_seen": 1543897088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031737989407981844,
      "loss": 2.6149,
      "theoretical_loss": 3.507161544301268,
      "tokens_seen": 1543962624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000317363344051447,
      "loss": 2.6041,
      "theoretical_loss": 3.5071484917386586,
      "tokens_seen": 1544028160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031734679402307546,
      "loss": 2.6937,
      "theoretical_loss": 3.5071354398851673,
      "tokens_seen": 1544093696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000317330243994704,
      "loss": 2.7843,
      "theoretical_loss": 3.507122388740725,
      "tokens_seen": 1544159232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031731369396633254,
      "loss": 2.8073,
      "theoretical_loss": 3.5071093383052645,
      "tokens_seen": 1544224768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000317297143937961,
      "loss": 2.4326,
      "theoretical_loss": 3.5070962885787154,
      "tokens_seen": 1544290304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031728059390958956,
      "loss": 2.657,
      "theoretical_loss": 3.507083239561011,
      "tokens_seen": 1544355840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031726404388121804,
      "loss": 2.4437,
      "theoretical_loss": 3.507070191252082,
      "tokens_seen": 1544421376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003172474938528466,
      "loss": 2.4369,
      "theoretical_loss": 3.507057143651859,
      "tokens_seen": 1544486912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003172309438244751,
      "loss": 2.672,
      "theoretical_loss": 3.507044096760275,
      "tokens_seen": 1544552448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031721439379610365,
      "loss": 2.6387,
      "theoretical_loss": 3.5070310505772597,
      "tokens_seen": 1544617984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031719784376773214,
      "loss": 2.5977,
      "theoretical_loss": 3.5070180051027458,
      "tokens_seen": 1544683520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003171812937393607,
      "loss": 2.5475,
      "theoretical_loss": 3.5070049603366646,
      "tokens_seen": 1544749056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003171647437109892,
      "loss": 2.7929,
      "theoretical_loss": 3.506991916278947,
      "tokens_seen": 1544814592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031714819368261775,
      "loss": 2.9268,
      "theoretical_loss": 3.5069788729295253,
      "tokens_seen": 1544880128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031713164365424623,
      "loss": 2.7204,
      "theoretical_loss": 3.5069658302883298,
      "tokens_seen": 1544945664
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031711509362587477,
      "loss": 2.682,
      "theoretical_loss": 3.5069527883552936,
      "tokens_seen": 1545011200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003170985435975033,
      "loss": 2.4843,
      "theoretical_loss": 3.506939747130347,
      "tokens_seen": 1545076736
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031708199356913184,
      "loss": 2.6238,
      "theoretical_loss": 3.506926706613422,
      "tokens_seen": 1545142272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031706544354076033,
      "loss": 2.7127,
      "theoretical_loss": 3.50691366680445,
      "tokens_seen": 1545207808
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031704889351238886,
      "loss": 2.4793,
      "theoretical_loss": 3.506900627703362,
      "tokens_seen": 1545273344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003170323434840174,
      "loss": 2.7099,
      "theoretical_loss": 3.5068875893100913,
      "tokens_seen": 1545338880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031701579345564594,
      "loss": 2.8135,
      "theoretical_loss": 3.506874551624567,
      "tokens_seen": 1545404416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003169992434272744,
      "loss": 2.7716,
      "theoretical_loss": 3.506861514646723,
      "tokens_seen": 1545469952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031698269339890296,
      "loss": 2.7349,
      "theoretical_loss": 3.5068484783764893,
      "tokens_seen": 1545535488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003169661433705315,
      "loss": 2.5168,
      "theoretical_loss": 3.5068354428137978,
      "tokens_seen": 1545601024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031694959334216004,
      "loss": 2.7707,
      "theoretical_loss": 3.5068224079585812,
      "tokens_seen": 1545666560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003169330433137885,
      "loss": 2.7863,
      "theoretical_loss": 3.5068093738107695,
      "tokens_seen": 1545732096
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031691649328541706,
      "loss": 2.7292,
      "theoretical_loss": 3.506796340370295,
      "tokens_seen": 1545797632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003168999432570456,
      "loss": 2.6456,
      "theoretical_loss": 3.5067833076370896,
      "tokens_seen": 1545863168
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031688339322867413,
      "loss": 2.7577,
      "theoretical_loss": 3.506770275611085,
      "tokens_seen": 1545928704
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003168668432003026,
      "loss": 2.5705,
      "theoretical_loss": 3.5067572442922126,
      "tokens_seen": 1545994240
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031685029317193115,
      "loss": 2.6697,
      "theoretical_loss": 3.5067442136804035,
      "tokens_seen": 1546059776
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003168337431435597,
      "loss": 2.781,
      "theoretical_loss": 3.50673118377559,
      "tokens_seen": 1546125312
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003168171931151882,
      "loss": 2.7122,
      "theoretical_loss": 3.5067181545777037,
      "tokens_seen": 1546190848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003168006430868167,
      "loss": 2.6714,
      "theoretical_loss": 3.5067051260866764,
      "tokens_seen": 1546256384
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031678409305844525,
      "loss": 2.6173,
      "theoretical_loss": 3.50669209830244,
      "tokens_seen": 1546321920
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031676754303007373,
      "loss": 2.7519,
      "theoretical_loss": 3.5066790712249256,
      "tokens_seen": 1546387456
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031675099300170227,
      "loss": 2.5817,
      "theoretical_loss": 3.5066660448540645,
      "tokens_seen": 1546452992
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003167344429733308,
      "loss": 2.417,
      "theoretical_loss": 3.50665301918979,
      "tokens_seen": 1546518528
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031671789294495934,
      "loss": 2.5715,
      "theoretical_loss": 3.5066399942320325,
      "tokens_seen": 1546584064
    },
    {
      "epoch": 0.55,
      "objective/train/avg_token_score": 0.013138110749423504,
      "objective/train/avg_weight": 0.9845436215400696,
      "objective/train/docs_used": 876453,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8855366706848145,
      "objective/train/std_weight": 0.07567085325717926,
      "objective/train/theoretical_loss": 3.506626969980724,
      "objective/train/tokens_used": 1567109600,
      "theoretical_loss": 3.506626969980724,
      "tokens_seen": 1546649600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031670134291658783,
      "loss": 2.6873,
      "theoretical_loss": 3.506626969980724,
      "tokens_seen": 1546649600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031668479288821637,
      "loss": 2.4993,
      "theoretical_loss": 3.506613946435797,
      "tokens_seen": 1546715136
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003166682428598449,
      "loss": 2.6278,
      "theoretical_loss": 3.5066009235971825,
      "tokens_seen": 1546780672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031665169283147344,
      "loss": 2.6197,
      "theoretical_loss": 3.506587901464812,
      "tokens_seen": 1546846208
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003166351428031019,
      "loss": 2.8154,
      "theoretical_loss": 3.5065748800386176,
      "tokens_seen": 1546911744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031661859277473046,
      "loss": 2.5493,
      "theoretical_loss": 3.506561859318532,
      "tokens_seen": 1546977280
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000316602042746359,
      "loss": 2.6202,
      "theoretical_loss": 3.506548839304486,
      "tokens_seen": 1547042816
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031658549271798754,
      "loss": 2.851,
      "theoretical_loss": 3.506535819996411,
      "tokens_seen": 1547108352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000316568942689616,
      "loss": 2.5131,
      "theoretical_loss": 3.50652280139424,
      "tokens_seen": 1547173888
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031655239266124456,
      "loss": 2.7522,
      "theoretical_loss": 3.506509783497904,
      "tokens_seen": 1547239424
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003165358426328731,
      "loss": 2.7932,
      "theoretical_loss": 3.506496766307335,
      "tokens_seen": 1547304960
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031651929260450163,
      "loss": 2.8195,
      "theoretical_loss": 3.5064837498224657,
      "tokens_seen": 1547370496
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003165027425761301,
      "loss": 2.7533,
      "theoretical_loss": 3.5064707340432264,
      "tokens_seen": 1547436032
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031648619254775865,
      "loss": 2.8972,
      "theoretical_loss": 3.5064577189695503,
      "tokens_seen": 1547501568
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003164696425193872,
      "loss": 2.7505,
      "theoretical_loss": 3.5064447046013685,
      "tokens_seen": 1547567104
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031645309249101573,
      "loss": 2.4531,
      "theoretical_loss": 3.506431690938613,
      "tokens_seen": 1547632640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003164365424626442,
      "loss": 2.4649,
      "theoretical_loss": 3.506418677981216,
      "tokens_seen": 1547698176
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031641999243427275,
      "loss": 2.807,
      "theoretical_loss": 3.5064056657291096,
      "tokens_seen": 1547763712
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003164034424059013,
      "loss": 2.5564,
      "theoretical_loss": 3.506392654182225,
      "tokens_seen": 1547829248
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003163868923775298,
      "loss": 2.9971,
      "theoretical_loss": 3.5063796433404946,
      "tokens_seen": 1547894784
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003163703423491583,
      "loss": 2.4963,
      "theoretical_loss": 3.50636663320385,
      "tokens_seen": 1547960320
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031635379232078685,
      "loss": 2.5954,
      "theoretical_loss": 3.506353623772223,
      "tokens_seen": 1548025856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003163372422924154,
      "loss": 2.5965,
      "theoretical_loss": 3.5063406150455467,
      "tokens_seen": 1548091392
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031632069226404387,
      "loss": 2.4199,
      "theoretical_loss": 3.506327607023752,
      "tokens_seen": 1548156928
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003163041422356724,
      "loss": 2.5812,
      "theoretical_loss": 3.50631459970677,
      "tokens_seen": 1548222464
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003162875922073009,
      "loss": 2.6704,
      "theoretical_loss": 3.5063015930945354,
      "tokens_seen": 1548288000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003162710421789294,
      "loss": 2.5244,
      "theoretical_loss": 3.506288587186978,
      "tokens_seen": 1548353536
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031625449215055796,
      "loss": 2.8263,
      "theoretical_loss": 3.50627558198403,
      "tokens_seen": 1548419072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003162379421221865,
      "loss": 2.1502,
      "theoretical_loss": 3.506262577485624,
      "tokens_seen": 1548484608
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000316221392093815,
      "loss": 2.5364,
      "theoretical_loss": 3.506249573691692,
      "tokens_seen": 1548550144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003162048420654435,
      "loss": 2.6641,
      "theoretical_loss": 3.506236570602166,
      "tokens_seen": 1548615680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031618829203707206,
      "loss": 2.4973,
      "theoretical_loss": 3.5062235682169773,
      "tokens_seen": 1548681216
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003161717420087006,
      "loss": 2.813,
      "theoretical_loss": 3.5062105665360592,
      "tokens_seen": 1548746752
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003161551919803291,
      "loss": 3.0352,
      "theoretical_loss": 3.5061975655593423,
      "tokens_seen": 1548812288
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003161386419519576,
      "loss": 2.5536,
      "theoretical_loss": 3.50618456528676,
      "tokens_seen": 1548877824
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031612209192358615,
      "loss": 2.7588,
      "theoretical_loss": 3.5061715657182435,
      "tokens_seen": 1548943360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003161055418952147,
      "loss": 2.5697,
      "theoretical_loss": 3.506158566853725,
      "tokens_seen": 1549008896
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003160889918668432,
      "loss": 2.7616,
      "theoretical_loss": 3.506145568693137,
      "tokens_seen": 1549074432
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003160724418384717,
      "loss": 2.5888,
      "theoretical_loss": 3.506132571236412,
      "tokens_seen": 1549139968
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031605589181010025,
      "loss": 2.7162,
      "theoretical_loss": 3.5061195744834808,
      "tokens_seen": 1549205504
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003160393417817288,
      "loss": 2.5682,
      "theoretical_loss": 3.5061065784342764,
      "tokens_seen": 1549271040
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031602279175335727,
      "loss": 2.5823,
      "theoretical_loss": 3.506093583088731,
      "tokens_seen": 1549336576
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003160062417249858,
      "loss": 2.749,
      "theoretical_loss": 3.5060805884467756,
      "tokens_seen": 1549402112
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031598969169661435,
      "loss": 2.7333,
      "theoretical_loss": 3.5060675945083437,
      "tokens_seen": 1549467648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003159731416682429,
      "loss": 2.6439,
      "theoretical_loss": 3.5060546012733673,
      "tokens_seen": 1549533184
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031595659163987137,
      "loss": 2.7613,
      "theoretical_loss": 3.506041608741778,
      "tokens_seen": 1549598720
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003159400416114999,
      "loss": 2.594,
      "theoretical_loss": 3.5060286169135084,
      "tokens_seen": 1549664256
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031592349158312844,
      "loss": 2.5226,
      "theoretical_loss": 3.5060156257884905,
      "tokens_seen": 1549729792
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000315906941554757,
      "loss": 2.6905,
      "theoretical_loss": 3.5060026353666562,
      "tokens_seen": 1549795328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031589039152638546,
      "loss": 2.6262,
      "theoretical_loss": 3.5059896456479382,
      "tokens_seen": 1549860864
    },
    {
      "epoch": 0.55,
      "objective/train/avg_token_score": 0.012624753639101982,
      "objective/train/avg_weight": 0.9851396083831787,
      "objective/train/docs_used": 878255,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8999595642089844,
      "objective/train/std_weight": 0.0650639608502388,
      "objective/train/theoretical_loss": 3.505976656632269,
      "objective/train/tokens_used": 1570386400,
      "theoretical_loss": 3.505976656632269,
      "tokens_seen": 1549926400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000315873841498014,
      "loss": 2.7243,
      "theoretical_loss": 3.505976656632269,
      "tokens_seen": 1549926400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031585729146964254,
      "loss": 2.8403,
      "theoretical_loss": 3.5059636683195796,
      "tokens_seen": 1549991936
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003158407414412711,
      "loss": 2.6899,
      "theoretical_loss": 3.5059506807098035,
      "tokens_seen": 1550057472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031582419141289956,
      "loss": 2.7795,
      "theoretical_loss": 3.505937693802872,
      "tokens_seen": 1550123008
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031580764138452804,
      "loss": 2.4057,
      "theoretical_loss": 3.505924707598718,
      "tokens_seen": 1550188544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003157910913561566,
      "loss": 2.4752,
      "theoretical_loss": 3.5059117220972738,
      "tokens_seen": 1550254080
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003157745413277851,
      "loss": 2.8085,
      "theoretical_loss": 3.5058987372984713,
      "tokens_seen": 1550319616
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031575799129941365,
      "loss": 2.7617,
      "theoretical_loss": 3.5058857532022425,
      "tokens_seen": 1550385152
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031574144127104214,
      "loss": 2.7277,
      "theoretical_loss": 3.50587276980852,
      "tokens_seen": 1550450688
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003157248912426707,
      "loss": 2.3942,
      "theoretical_loss": 3.5058597871172372,
      "tokens_seen": 1550516224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003157083412142992,
      "loss": 2.6718,
      "theoretical_loss": 3.5058468051283245,
      "tokens_seen": 1550581760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031569179118592775,
      "loss": 2.6179,
      "theoretical_loss": 3.5058338238417157,
      "tokens_seen": 1550647296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031567524115755623,
      "loss": 2.6481,
      "theoretical_loss": 3.505820843257342,
      "tokens_seen": 1550712832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031565869112918477,
      "loss": 2.5388,
      "theoretical_loss": 3.5058078633751366,
      "tokens_seen": 1550778368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003156421411008133,
      "loss": 2.7511,
      "theoretical_loss": 3.505794884195031,
      "tokens_seen": 1550843904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031562559107244185,
      "loss": 2.7231,
      "theoretical_loss": 3.5057819057169586,
      "tokens_seen": 1550909440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031560904104407033,
      "loss": 2.7671,
      "theoretical_loss": 3.505768927940851,
      "tokens_seen": 1550974976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031559249101569887,
      "loss": 2.6427,
      "theoretical_loss": 3.505755950866641,
      "tokens_seen": 1551040512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003155759409873274,
      "loss": 2.5615,
      "theoretical_loss": 3.505742974494261,
      "tokens_seen": 1551106048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031555939095895594,
      "loss": 2.631,
      "theoretical_loss": 3.5057299988236426,
      "tokens_seen": 1551171584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003155428409305844,
      "loss": 2.6593,
      "theoretical_loss": 3.505717023854719,
      "tokens_seen": 1551237120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031552629090221296,
      "loss": 2.606,
      "theoretical_loss": 3.5057040495874228,
      "tokens_seen": 1551302656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003155097408738415,
      "loss": 2.3602,
      "theoretical_loss": 3.5056910760216855,
      "tokens_seen": 1551368192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031549319084547004,
      "loss": 2.7268,
      "theoretical_loss": 3.5056781031574404,
      "tokens_seen": 1551433728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003154766408170986,
      "loss": 2.6046,
      "theoretical_loss": 3.505665130994619,
      "tokens_seen": 1551499264
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031546009078872706,
      "loss": 2.6624,
      "theoretical_loss": 3.505652159533155,
      "tokens_seen": 1551564800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003154435407603556,
      "loss": 2.8428,
      "theoretical_loss": 3.5056391887729803,
      "tokens_seen": 1551630336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031542699073198413,
      "loss": 2.5036,
      "theoretical_loss": 3.505626218714027,
      "tokens_seen": 1551695872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031541044070361267,
      "loss": 2.4798,
      "theoretical_loss": 3.5056132493562275,
      "tokens_seen": 1551761408
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031539389067524116,
      "loss": 2.8456,
      "theoretical_loss": 3.5056002806995155,
      "tokens_seen": 1551826944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003153773406468697,
      "loss": 2.7327,
      "theoretical_loss": 3.5055873127438217,
      "tokens_seen": 1551892480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031536079061849823,
      "loss": 2.611,
      "theoretical_loss": 3.50557434548908,
      "tokens_seen": 1551958016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031534424059012677,
      "loss": 2.7666,
      "theoretical_loss": 3.505561378935223,
      "tokens_seen": 1552023552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031532769056175525,
      "loss": 2.5889,
      "theoretical_loss": 3.5055484130821815,
      "tokens_seen": 1552089088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031531114053338373,
      "loss": 2.7345,
      "theoretical_loss": 3.5055354479298897,
      "tokens_seen": 1552154624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031529459050501227,
      "loss": 2.7727,
      "theoretical_loss": 3.5055224834782797,
      "tokens_seen": 1552220160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003152780404766408,
      "loss": 2.5771,
      "theoretical_loss": 3.5055095197272843,
      "tokens_seen": 1552285696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031526149044826935,
      "loss": 2.5002,
      "theoretical_loss": 3.5054965566768352,
      "tokens_seen": 1552351232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031524494041989783,
      "loss": 2.4576,
      "theoretical_loss": 3.505483594326866,
      "tokens_seen": 1552416768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031522839039152637,
      "loss": 2.2615,
      "theoretical_loss": 3.5054706326773086,
      "tokens_seen": 1552482304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003152118403631549,
      "loss": 2.5496,
      "theoretical_loss": 3.5054576717280956,
      "tokens_seen": 1552547840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031519529033478344,
      "loss": 2.4857,
      "theoretical_loss": 3.50544471147916,
      "tokens_seen": 1552613376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003151787403064119,
      "loss": 2.5112,
      "theoretical_loss": 3.505431751930434,
      "tokens_seen": 1552678912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031516219027804046,
      "loss": 2.7009,
      "theoretical_loss": 3.505418793081851,
      "tokens_seen": 1552744448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000315145640249669,
      "loss": 2.6413,
      "theoretical_loss": 3.5054058349333426,
      "tokens_seen": 1552809984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031512909022129754,
      "loss": 2.8009,
      "theoretical_loss": 3.5053928774848417,
      "tokens_seen": 1552875520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000315112540192926,
      "loss": 2.6457,
      "theoretical_loss": 3.5053799207362815,
      "tokens_seen": 1552941056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031509599016455456,
      "loss": 2.5595,
      "theoretical_loss": 3.5053669646875942,
      "tokens_seen": 1553006592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003150794401361831,
      "loss": 2.5147,
      "theoretical_loss": 3.5053540093387126,
      "tokens_seen": 1553072128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031506289010781164,
      "loss": 2.5286,
      "theoretical_loss": 3.50534105468957,
      "tokens_seen": 1553137664
    },
    {
      "epoch": 0.55,
      "objective/train/avg_token_score": 0.00904814712703228,
      "objective/train/avg_weight": 0.9893512725830078,
      "objective/train/docs_used": 880445,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5095536708831787,
      "objective/train/std_weight": 0.061026766896247864,
      "objective/train/theoretical_loss": 3.505328100740097,
      "objective/train/tokens_used": 1573663200,
      "theoretical_loss": 3.505328100740097,
      "tokens_seen": 1553203200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003150463400794401,
      "loss": 2.4964,
      "theoretical_loss": 3.505328100740097,
      "tokens_seen": 1553203200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031502979005106866,
      "loss": 2.6759,
      "theoretical_loss": 3.5053151474902284,
      "tokens_seen": 1553268736
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003150132400226972,
      "loss": 2.4914,
      "theoretical_loss": 3.505302194939896,
      "tokens_seen": 1553334272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031499668999432573,
      "loss": 2.5867,
      "theoretical_loss": 3.505289243089033,
      "tokens_seen": 1553399808
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003149801399659542,
      "loss": 2.6607,
      "theoretical_loss": 3.5052762919375717,
      "tokens_seen": 1553465344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031496358993758275,
      "loss": 2.7193,
      "theoretical_loss": 3.5052633414854446,
      "tokens_seen": 1553530880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003149470399092113,
      "loss": 2.7461,
      "theoretical_loss": 3.5052503917325852,
      "tokens_seen": 1553596416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031493048988083983,
      "loss": 2.8175,
      "theoretical_loss": 3.505237442678926,
      "tokens_seen": 1553661952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003149139398524683,
      "loss": 2.4949,
      "theoretical_loss": 3.5052244943243993,
      "tokens_seen": 1553727488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00031489738982409685,
      "loss": 2.6987,
      "theoretical_loss": 3.505211546668938,
      "tokens_seen": 1553793024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003148808397957254,
      "loss": 2.6372,
      "theoretical_loss": 3.5051985997124757,
      "tokens_seen": 1553858560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0003148642897673539,
      "loss": 2.5638,
      "theoretical_loss": 3.5051856534549444,
      "tokens_seen": 1553924096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003148477397389824,
      "loss": 2.6674,
      "theoretical_loss": 3.505172707896276,
      "tokens_seen": 1553989632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003148311897106109,
      "loss": 2.4671,
      "theoretical_loss": 3.5051597630364055,
      "tokens_seen": 1554055168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031481463968223943,
      "loss": 2.6647,
      "theoretical_loss": 3.505146818875264,
      "tokens_seen": 1554120704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031479808965386797,
      "loss": 2.5295,
      "theoretical_loss": 3.5051338754127848,
      "tokens_seen": 1554186240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003147815396254965,
      "loss": 2.7243,
      "theoretical_loss": 3.505120932648901,
      "tokens_seen": 1554251776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000314764989597125,
      "loss": 2.6951,
      "theoretical_loss": 3.505107990583545,
      "tokens_seen": 1554317312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003147484395687535,
      "loss": 2.68,
      "theoretical_loss": 3.50509504921665,
      "tokens_seen": 1554382848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031473188954038206,
      "loss": 2.573,
      "theoretical_loss": 3.505082108548149,
      "tokens_seen": 1554448384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003147153395120106,
      "loss": 2.8098,
      "theoretical_loss": 3.5050691685779745,
      "tokens_seen": 1554513920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003146987894836391,
      "loss": 2.8264,
      "theoretical_loss": 3.505056229306059,
      "tokens_seen": 1554579456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003146822394552676,
      "loss": 2.7278,
      "theoretical_loss": 3.5050432907323366,
      "tokens_seen": 1554644992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031466568942689616,
      "loss": 2.6601,
      "theoretical_loss": 3.505030352856739,
      "tokens_seen": 1554710528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003146491393985247,
      "loss": 2.7219,
      "theoretical_loss": 3.5050174156791996,
      "tokens_seen": 1554776064
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003146325893701532,
      "loss": 2.6367,
      "theoretical_loss": 3.505004479199651,
      "tokens_seen": 1554841600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003146160393417817,
      "loss": 2.711,
      "theoretical_loss": 3.504991543418027,
      "tokens_seen": 1554907136
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031459948931341025,
      "loss": 2.4789,
      "theoretical_loss": 3.50497860833426,
      "tokens_seen": 1554972672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003145829392850388,
      "loss": 2.5462,
      "theoretical_loss": 3.5049656739482824,
      "tokens_seen": 1555038208
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003145663892566673,
      "loss": 2.6114,
      "theoretical_loss": 3.504952740260028,
      "tokens_seen": 1555103744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003145498392282958,
      "loss": 2.5295,
      "theoretical_loss": 3.504939807269429,
      "tokens_seen": 1555169280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031453328919992435,
      "loss": 2.4925,
      "theoretical_loss": 3.5049268749764186,
      "tokens_seen": 1555234816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003145167391715529,
      "loss": 2.763,
      "theoretical_loss": 3.504913943380931,
      "tokens_seen": 1555300352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031450018914318137,
      "loss": 2.4615,
      "theoretical_loss": 3.5049010124828968,
      "tokens_seen": 1555365888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003144836391148099,
      "loss": 2.3966,
      "theoretical_loss": 3.5048880822822515,
      "tokens_seen": 1555431424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031446708908643844,
      "loss": 2.3644,
      "theoretical_loss": 3.5048751527789266,
      "tokens_seen": 1555496960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000314450539058067,
      "loss": 2.7199,
      "theoretical_loss": 3.504862223972855,
      "tokens_seen": 1555562496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031443398902969547,
      "loss": 2.734,
      "theoretical_loss": 3.5048492958639708,
      "tokens_seen": 1555628032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000314417439001324,
      "loss": 2.4868,
      "theoretical_loss": 3.5048363684522057,
      "tokens_seen": 1555693568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031440088897295254,
      "loss": 2.6743,
      "theoretical_loss": 3.504823441737494,
      "tokens_seen": 1555759104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003143843389445811,
      "loss": 2.6839,
      "theoretical_loss": 3.504810515719768,
      "tokens_seen": 1555824640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031436778891620956,
      "loss": 2.6856,
      "theoretical_loss": 3.5047975903989617,
      "tokens_seen": 1555890176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031435123888783805,
      "loss": 2.6873,
      "theoretical_loss": 3.5047846657750066,
      "tokens_seen": 1555955712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003143346888594666,
      "loss": 2.6702,
      "theoretical_loss": 3.504771741847837,
      "tokens_seen": 1556021248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003143181388310951,
      "loss": 2.5442,
      "theoretical_loss": 3.5047588186173853,
      "tokens_seen": 1556086784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031430158880272366,
      "loss": 2.9269,
      "theoretical_loss": 3.5047458960835853,
      "tokens_seen": 1556152320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031428503877435214,
      "loss": 2.7397,
      "theoretical_loss": 3.5047329742463695,
      "tokens_seen": 1556217856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003142684887459807,
      "loss": 2.7228,
      "theoretical_loss": 3.5047200531056713,
      "tokens_seen": 1556283392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003142519387176092,
      "loss": 2.8091,
      "theoretical_loss": 3.504707132661424,
      "tokens_seen": 1556348928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031423538868923775,
      "loss": 2.4128,
      "theoretical_loss": 3.5046942129135603,
      "tokens_seen": 1556414464
    },
    {
      "epoch": 0.56,
      "objective/train/avg_token_score": 0.00903913751244545,
      "objective/train/avg_weight": 0.989361584186554,
      "objective/train/docs_used": 881444,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8103365898132324,
      "objective/train/std_weight": 0.05955254286527634,
      "objective/train/theoretical_loss": 3.5046812938620135,
      "objective/train/tokens_used": 1576940000,
      "theoretical_loss": 3.5046812938620135,
      "tokens_seen": 1556480000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003142188386608663,
      "loss": 2.5743,
      "theoretical_loss": 3.5046812938620135,
      "tokens_seen": 1556480000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003142022886324948,
      "loss": 2.5304,
      "theoretical_loss": 3.504668375506717,
      "tokens_seen": 1556545536
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003141857386041233,
      "loss": 2.7176,
      "theoretical_loss": 3.504655457847604,
      "tokens_seen": 1556611072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031416918857575185,
      "loss": 2.5245,
      "theoretical_loss": 3.504642540884607,
      "tokens_seen": 1556676608
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003141526385473804,
      "loss": 2.8523,
      "theoretical_loss": 3.5046296246176594,
      "tokens_seen": 1556742144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031413608851900887,
      "loss": 2.6544,
      "theoretical_loss": 3.5046167090466955,
      "tokens_seen": 1556807680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003141195384906374,
      "loss": 2.8149,
      "theoretical_loss": 3.504603794171647,
      "tokens_seen": 1556873216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031410298846226595,
      "loss": 2.6496,
      "theoretical_loss": 3.504590879992448,
      "tokens_seen": 1556938752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003140864384338945,
      "loss": 2.8015,
      "theoretical_loss": 3.5045779665090313,
      "tokens_seen": 1557004288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031406988840552297,
      "loss": 2.5183,
      "theoretical_loss": 3.5045650537213304,
      "tokens_seen": 1557069824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003140533383771515,
      "loss": 2.5305,
      "theoretical_loss": 3.5045521416292784,
      "tokens_seen": 1557135360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031403678834878004,
      "loss": 2.9869,
      "theoretical_loss": 3.5045392302328087,
      "tokens_seen": 1557200896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003140202383204086,
      "loss": 2.6136,
      "theoretical_loss": 3.5045263195318546,
      "tokens_seen": 1557266432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031400368829203706,
      "loss": 2.709,
      "theoretical_loss": 3.5045134095263486,
      "tokens_seen": 1557331968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003139871382636656,
      "loss": 2.6384,
      "theoretical_loss": 3.504500500216225,
      "tokens_seen": 1557397504
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031397058823529414,
      "loss": 2.9428,
      "theoretical_loss": 3.504487591601417,
      "tokens_seen": 1557463040
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003139540382069227,
      "loss": 2.7617,
      "theoretical_loss": 3.504474683681857,
      "tokens_seen": 1557528576
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031393748817855116,
      "loss": 2.5993,
      "theoretical_loss": 3.504461776457479,
      "tokens_seen": 1557594112
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003139209381501797,
      "loss": 2.766,
      "theoretical_loss": 3.504448869928216,
      "tokens_seen": 1557659648
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031390438812180823,
      "loss": 2.665,
      "theoretical_loss": 3.504435964094002,
      "tokens_seen": 1557725184
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031388783809343677,
      "loss": 2.6398,
      "theoretical_loss": 3.504423058954769,
      "tokens_seen": 1557790720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031387128806506525,
      "loss": 2.6357,
      "theoretical_loss": 3.504410154510452,
      "tokens_seen": 1557856256
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031385473803669374,
      "loss": 2.5699,
      "theoretical_loss": 3.5043972507609835,
      "tokens_seen": 1557921792
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003138381880083223,
      "loss": 2.8436,
      "theoretical_loss": 3.5043843477062966,
      "tokens_seen": 1557987328
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003138216379799508,
      "loss": 2.5222,
      "theoretical_loss": 3.504371445346325,
      "tokens_seen": 1558052864
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031380508795157935,
      "loss": 2.5048,
      "theoretical_loss": 3.5043585436810014,
      "tokens_seen": 1558118400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031378853792320783,
      "loss": 2.3434,
      "theoretical_loss": 3.504345642710261,
      "tokens_seen": 1558183936
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031377198789483637,
      "loss": 2.6688,
      "theoretical_loss": 3.504332742434035,
      "tokens_seen": 1558249472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003137554378664649,
      "loss": 2.6167,
      "theoretical_loss": 3.504319842852258,
      "tokens_seen": 1558315008
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031373888783809345,
      "loss": 2.5759,
      "theoretical_loss": 3.5043069439648633,
      "tokens_seen": 1558380544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031372233780972193,
      "loss": 2.4424,
      "theoretical_loss": 3.5042940457717844,
      "tokens_seen": 1558446080
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031370578778135047,
      "loss": 2.4399,
      "theoretical_loss": 3.5042811482729546,
      "tokens_seen": 1558511616
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000313689237752979,
      "loss": 2.6357,
      "theoretical_loss": 3.5042682514683072,
      "tokens_seen": 1558577152
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031367268772460754,
      "loss": 2.5215,
      "theoretical_loss": 3.5042553553577758,
      "tokens_seen": 1558642688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000313656137696236,
      "loss": 2.5297,
      "theoretical_loss": 3.5042424599412936,
      "tokens_seen": 1558708224
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031363958766786456,
      "loss": 2.6282,
      "theoretical_loss": 3.504229565218794,
      "tokens_seen": 1558773760
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003136230376394931,
      "loss": 2.8409,
      "theoretical_loss": 3.5042166711902114,
      "tokens_seen": 1558839296
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031360648761112164,
      "loss": 2.7451,
      "theoretical_loss": 3.504203777855478,
      "tokens_seen": 1558904832
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003135899375827501,
      "loss": 2.5329,
      "theoretical_loss": 3.5041908852145287,
      "tokens_seen": 1558970368
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031357338755437866,
      "loss": 2.6033,
      "theoretical_loss": 3.5041779932672963,
      "tokens_seen": 1559035904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003135568375260072,
      "loss": 2.8522,
      "theoretical_loss": 3.5041651020137135,
      "tokens_seen": 1559101440
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031354028749763573,
      "loss": 2.6757,
      "theoretical_loss": 3.5041522114537154,
      "tokens_seen": 1559166976
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003135237374692642,
      "loss": 2.558,
      "theoretical_loss": 3.504139321587234,
      "tokens_seen": 1559232512
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031350718744089276,
      "loss": 2.651,
      "theoretical_loss": 3.504126432414204,
      "tokens_seen": 1559298048
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003134906374125213,
      "loss": 2.5919,
      "theoretical_loss": 3.504113543934558,
      "tokens_seen": 1559363584
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031347408738414983,
      "loss": 2.5038,
      "theoretical_loss": 3.5041006561482306,
      "tokens_seen": 1559429120
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003134575373557783,
      "loss": 2.66,
      "theoretical_loss": 3.504087769055155,
      "tokens_seen": 1559494656
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031344098732740685,
      "loss": 2.7315,
      "theoretical_loss": 3.504074882655264,
      "tokens_seen": 1559560192
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003134244372990354,
      "loss": 2.5248,
      "theoretical_loss": 3.5040619969484923,
      "tokens_seen": 1559625728
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003134078872706639,
      "loss": 2.5297,
      "theoretical_loss": 3.5040491119347728,
      "tokens_seen": 1559691264
    },
    {
      "epoch": 0.56,
      "objective/train/avg_token_score": 0.02818768098950386,
      "objective/train/avg_weight": 0.966819167137146,
      "objective/train/docs_used": 883210,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.925438165664673,
      "objective/train/std_weight": 0.12601692974567413,
      "objective/train/theoretical_loss": 3.5040362276140393,
      "objective/train/tokens_used": 1580216800,
      "theoretical_loss": 3.5040362276140393,
      "tokens_seen": 1559756800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003133913372422924,
      "loss": 2.6668,
      "theoretical_loss": 3.5040362276140393,
      "tokens_seen": 1559756800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003133747872139209,
      "loss": 2.3813,
      "theoretical_loss": 3.5040233439862254,
      "tokens_seen": 1559822336
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031335823718554943,
      "loss": 2.5418,
      "theoretical_loss": 3.504010461051265,
      "tokens_seen": 1559887872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031334168715717797,
      "loss": 2.325,
      "theoretical_loss": 3.5039975788090914,
      "tokens_seen": 1559953408
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003133251371288065,
      "loss": 2.5859,
      "theoretical_loss": 3.503984697259638,
      "tokens_seen": 1560018944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000313308587100435,
      "loss": 2.7028,
      "theoretical_loss": 3.503971816402839,
      "tokens_seen": 1560084480
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003132920370720635,
      "loss": 2.4048,
      "theoretical_loss": 3.503958936238628,
      "tokens_seen": 1560150016
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031327548704369206,
      "loss": 2.4612,
      "theoretical_loss": 3.503946056766938,
      "tokens_seen": 1560215552
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003132589370153206,
      "loss": 2.1349,
      "theoretical_loss": 3.5039331779877037,
      "tokens_seen": 1560281088
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003132423869869491,
      "loss": 2.4947,
      "theoretical_loss": 3.503920299900858,
      "tokens_seen": 1560346624
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003132258369585776,
      "loss": 2.5397,
      "theoretical_loss": 3.5039074225063347,
      "tokens_seen": 1560412160
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031320928693020616,
      "loss": 2.7061,
      "theoretical_loss": 3.5038945458040684,
      "tokens_seen": 1560477696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003131927369018347,
      "loss": 2.7006,
      "theoretical_loss": 3.5038816697939916,
      "tokens_seen": 1560543232
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003131761868734632,
      "loss": 2.6603,
      "theoretical_loss": 3.5038687944760385,
      "tokens_seen": 1560608768
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003131596368450917,
      "loss": 2.5138,
      "theoretical_loss": 3.503855919850143,
      "tokens_seen": 1560674304
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031314308681672026,
      "loss": 2.7062,
      "theoretical_loss": 3.5038430459162386,
      "tokens_seen": 1560739840
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003131265367883488,
      "loss": 2.706,
      "theoretical_loss": 3.503830172674259,
      "tokens_seen": 1560805376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003131099867599773,
      "loss": 2.5269,
      "theoretical_loss": 3.503817300124138,
      "tokens_seen": 1560870912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003130934367316058,
      "loss": 2.633,
      "theoretical_loss": 3.50380442826581,
      "tokens_seen": 1560936448
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031307688670323435,
      "loss": 2.7447,
      "theoretical_loss": 3.503791557099208,
      "tokens_seen": 1561001984
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003130603366748629,
      "loss": 2.2961,
      "theoretical_loss": 3.503778686624266,
      "tokens_seen": 1561067520
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031304378664649137,
      "loss": 2.5487,
      "theoretical_loss": 3.503765816840918,
      "tokens_seen": 1561133056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003130272366181199,
      "loss": 2.8924,
      "theoretical_loss": 3.5037529477490974,
      "tokens_seen": 1561198592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031301068658974845,
      "loss": 2.6458,
      "theoretical_loss": 3.5037400793487383,
      "tokens_seen": 1561264128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000312994136561377,
      "loss": 2.4088,
      "theoretical_loss": 3.5037272116397746,
      "tokens_seen": 1561329664
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031297758653300547,
      "loss": 2.5121,
      "theoretical_loss": 3.50371434462214,
      "tokens_seen": 1561395200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000312961036504634,
      "loss": 2.5612,
      "theoretical_loss": 3.5037014782957683,
      "tokens_seen": 1561460736
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031294448647626254,
      "loss": 2.3695,
      "theoretical_loss": 3.5036886126605937,
      "tokens_seen": 1561526272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003129279364478911,
      "loss": 2.6134,
      "theoretical_loss": 3.5036757477165494,
      "tokens_seen": 1561591808
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031291138641951956,
      "loss": 2.3931,
      "theoretical_loss": 3.5036628834635697,
      "tokens_seen": 1561657344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003128948363911481,
      "loss": 2.5919,
      "theoretical_loss": 3.5036500199015883,
      "tokens_seen": 1561722880
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003128782863627766,
      "loss": 2.8418,
      "theoretical_loss": 3.50363715703054,
      "tokens_seen": 1561788416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003128617363344051,
      "loss": 2.751,
      "theoretical_loss": 3.503624294850357,
      "tokens_seen": 1561853952
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031284518630603366,
      "loss": 2.7791,
      "theoretical_loss": 3.5036114333609745,
      "tokens_seen": 1561919488
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003128286362776622,
      "loss": 2.7568,
      "theoretical_loss": 3.5035985725623258,
      "tokens_seen": 1561985024
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003128120862492907,
      "loss": 2.5213,
      "theoretical_loss": 3.5035857124543455,
      "tokens_seen": 1562050560
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003127955362209192,
      "loss": 2.7768,
      "theoretical_loss": 3.503572853036967,
      "tokens_seen": 1562116096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031277898619254776,
      "loss": 2.8832,
      "theoretical_loss": 3.503559994310124,
      "tokens_seen": 1562181632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003127624361641763,
      "loss": 2.507,
      "theoretical_loss": 3.503547136273751,
      "tokens_seen": 1562247168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003127458861358048,
      "loss": 2.6979,
      "theoretical_loss": 3.5035342789277824,
      "tokens_seen": 1562312704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003127293361074333,
      "loss": 2.3534,
      "theoretical_loss": 3.5035214222721507,
      "tokens_seen": 1562378240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031271278607906185,
      "loss": 2.5659,
      "theoretical_loss": 3.503508566306791,
      "tokens_seen": 1562443776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003126962360506904,
      "loss": 2.8772,
      "theoretical_loss": 3.5034957110316367,
      "tokens_seen": 1562509312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003126796860223189,
      "loss": 2.7676,
      "theoretical_loss": 3.503482856446622,
      "tokens_seen": 1562574848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003126631359939474,
      "loss": 2.6776,
      "theoretical_loss": 3.5034700025516816,
      "tokens_seen": 1562640384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031264658596557595,
      "loss": 2.8146,
      "theoretical_loss": 3.5034571493467483,
      "tokens_seen": 1562705920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003126300359372045,
      "loss": 2.7143,
      "theoretical_loss": 3.5034442968317574,
      "tokens_seen": 1562771456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031261348590883297,
      "loss": 2.7703,
      "theoretical_loss": 3.5034314450066417,
      "tokens_seen": 1562836992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003125969358804615,
      "loss": 2.5616,
      "theoretical_loss": 3.503418593871336,
      "tokens_seen": 1562902528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031258038585209004,
      "loss": 2.746,
      "theoretical_loss": 3.503405743425774,
      "tokens_seen": 1562968064
    },
    {
      "epoch": 0.56,
      "objective/train/avg_token_score": 0.011233927682042122,
      "objective/train/avg_weight": 0.9868052005767822,
      "objective/train/docs_used": 884815,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6979126930236816,
      "objective/train/std_weight": 0.05318503454327583,
      "objective/train/theoretical_loss": 3.5033928936698904,
      "objective/train/tokens_used": 1583493600,
      "theoretical_loss": 3.5033928936698904,
      "tokens_seen": 1563033600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003125638358237186,
      "loss": 2.5553,
      "theoretical_loss": 3.5033928936698904,
      "tokens_seen": 1563033600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031254728579534707,
      "loss": 2.8153,
      "theoretical_loss": 3.5033800446036185,
      "tokens_seen": 1563099136
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003125307357669756,
      "loss": 2.6135,
      "theoretical_loss": 3.5033671962268924,
      "tokens_seen": 1563164672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031251418573860414,
      "loss": 2.7003,
      "theoretical_loss": 3.5033543485396472,
      "tokens_seen": 1563230208
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003124976357102327,
      "loss": 2.6976,
      "theoretical_loss": 3.503341501541816,
      "tokens_seen": 1563295744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031248108568186116,
      "loss": 2.712,
      "theoretical_loss": 3.5033286552333327,
      "tokens_seen": 1563361280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003124645356534897,
      "loss": 2.2783,
      "theoretical_loss": 3.503315809614132,
      "tokens_seen": 1563426816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031244798562511824,
      "loss": 2.6215,
      "theoretical_loss": 3.5033029646841483,
      "tokens_seen": 1563492352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003124314355967468,
      "loss": 2.6631,
      "theoretical_loss": 3.5032901204433156,
      "tokens_seen": 1563557888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031241488556837526,
      "loss": 2.6125,
      "theoretical_loss": 3.503277276891567,
      "tokens_seen": 1563623424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003123983355400038,
      "loss": 2.6891,
      "theoretical_loss": 3.5032644340288375,
      "tokens_seen": 1563688960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003123817855116323,
      "loss": 2.4659,
      "theoretical_loss": 3.5032515918550615,
      "tokens_seen": 1563754496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003123652354832608,
      "loss": 2.7121,
      "theoretical_loss": 3.503238750370173,
      "tokens_seen": 1563820032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031234868545488935,
      "loss": 2.5463,
      "theoretical_loss": 3.503225909574106,
      "tokens_seen": 1563885568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031233213542651784,
      "loss": 2.6426,
      "theoretical_loss": 3.5032130694667947,
      "tokens_seen": 1563951104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003123155853981464,
      "loss": 2.5847,
      "theoretical_loss": 3.503200230048173,
      "tokens_seen": 1564016640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003122990353697749,
      "loss": 2.5757,
      "theoretical_loss": 3.5031873913181757,
      "tokens_seen": 1564082176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031228248534140345,
      "loss": 2.511,
      "theoretical_loss": 3.503174553276737,
      "tokens_seen": 1564147712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031226593531303193,
      "loss": 2.594,
      "theoretical_loss": 3.5031617159237904,
      "tokens_seen": 1564213248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031224938528466047,
      "loss": 2.6362,
      "theoretical_loss": 3.503148879259271,
      "tokens_seen": 1564278784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000312232835256289,
      "loss": 2.515,
      "theoretical_loss": 3.5031360432831127,
      "tokens_seen": 1564344320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031221628522791755,
      "loss": 2.4941,
      "theoretical_loss": 3.5031232079952495,
      "tokens_seen": 1564409856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031219973519954603,
      "loss": 2.6827,
      "theoretical_loss": 3.5031103733956157,
      "tokens_seen": 1564475392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031218318517117457,
      "loss": 2.8102,
      "theoretical_loss": 3.5030975394841457,
      "tokens_seen": 1564540928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003121666351428031,
      "loss": 2.5763,
      "theoretical_loss": 3.5030847062607746,
      "tokens_seen": 1564606464
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031215008511443164,
      "loss": 2.3515,
      "theoretical_loss": 3.5030718737254354,
      "tokens_seen": 1564672000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003121335350860601,
      "loss": 2.5615,
      "theoretical_loss": 3.5030590418780623,
      "tokens_seen": 1564737536
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031211698505768866,
      "loss": 2.7623,
      "theoretical_loss": 3.503046210718591,
      "tokens_seen": 1564803072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003121004350293172,
      "loss": 2.8929,
      "theoretical_loss": 3.503033380246955,
      "tokens_seen": 1564868608
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031208388500094574,
      "loss": 2.6464,
      "theoretical_loss": 3.5030205504630882,
      "tokens_seen": 1564934144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003120673349725742,
      "loss": 2.8368,
      "theoretical_loss": 3.5030077213669255,
      "tokens_seen": 1564999680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031205078494420276,
      "loss": 2.8098,
      "theoretical_loss": 3.5029948929584016,
      "tokens_seen": 1565065216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003120342349158313,
      "loss": 2.8425,
      "theoretical_loss": 3.5029820652374495,
      "tokens_seen": 1565130752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031201768488745983,
      "loss": 2.5056,
      "theoretical_loss": 3.502969238204005,
      "tokens_seen": 1565196288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003120011348590883,
      "loss": 2.6708,
      "theoretical_loss": 3.502956411858002,
      "tokens_seen": 1565261824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031198458483071685,
      "loss": 2.3851,
      "theoretical_loss": 3.5029435861993745,
      "tokens_seen": 1565327360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003119680348023454,
      "loss": 2.5736,
      "theoretical_loss": 3.502930761228057,
      "tokens_seen": 1565392896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031195148477397393,
      "loss": 2.5251,
      "theoretical_loss": 3.5029179369439847,
      "tokens_seen": 1565458432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003119349347456024,
      "loss": 2.8021,
      "theoretical_loss": 3.5029051133470905,
      "tokens_seen": 1565523968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031191838471723095,
      "loss": 2.7276,
      "theoretical_loss": 3.5028922904373103,
      "tokens_seen": 1565589504
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031190183468885943,
      "loss": 2.6159,
      "theoretical_loss": 3.5028794682145774,
      "tokens_seen": 1565655040
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031188528466048797,
      "loss": 2.6988,
      "theoretical_loss": 3.5028666466788274,
      "tokens_seen": 1565720576
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003118687346321165,
      "loss": 2.506,
      "theoretical_loss": 3.5028538258299937,
      "tokens_seen": 1565786112
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000311852184603745,
      "loss": 2.5095,
      "theoretical_loss": 3.502841005668011,
      "tokens_seen": 1565851648
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031183563457537353,
      "loss": 2.7432,
      "theoretical_loss": 3.502828186192814,
      "tokens_seen": 1565917184
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031181908454700207,
      "loss": 2.7381,
      "theoretical_loss": 3.502815367404337,
      "tokens_seen": 1565982720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003118025345186306,
      "loss": 2.5147,
      "theoretical_loss": 3.5028025493025146,
      "tokens_seen": 1566048256
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003117859844902591,
      "loss": 2.6029,
      "theoretical_loss": 3.5027897318872814,
      "tokens_seen": 1566113792
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003117694344618876,
      "loss": 2.7227,
      "theoretical_loss": 3.502776915158571,
      "tokens_seen": 1566179328
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031175288443351616,
      "loss": 2.662,
      "theoretical_loss": 3.5027640991163196,
      "tokens_seen": 1566244864
    },
    {
      "epoch": 0.56,
      "objective/train/avg_token_score": 0.012778505682945251,
      "objective/train/avg_weight": 0.9849910736083984,
      "objective/train/docs_used": 886685,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.459707736968994,
      "objective/train/std_weight": 0.057656627148389816,
      "objective/train/theoretical_loss": 3.50275128376046,
      "objective/train/tokens_used": 1586770400,
      "theoretical_loss": 3.50275128376046,
      "tokens_seen": 1566310400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003117363344051447,
      "loss": 2.6592,
      "theoretical_loss": 3.50275128376046,
      "tokens_seen": 1566310400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003117197843767732,
      "loss": 2.5821,
      "theoretical_loss": 3.502738469090928,
      "tokens_seen": 1566375936
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003117032343484017,
      "loss": 2.867,
      "theoretical_loss": 3.502725655107657,
      "tokens_seen": 1566441472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031168668432003026,
      "loss": 2.5747,
      "theoretical_loss": 3.502712841810582,
      "tokens_seen": 1566507008
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003116701342916588,
      "loss": 2.6629,
      "theoretical_loss": 3.502700029199638,
      "tokens_seen": 1566572544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003116535842632873,
      "loss": 2.7467,
      "theoretical_loss": 3.5026872172747594,
      "tokens_seen": 1566638080
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003116370342349158,
      "loss": 2.4819,
      "theoretical_loss": 3.5026744060358803,
      "tokens_seen": 1566703616
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031162048420654435,
      "loss": 2.4098,
      "theoretical_loss": 3.5026615954829357,
      "tokens_seen": 1566769152
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003116039341781729,
      "loss": 2.5865,
      "theoretical_loss": 3.50264878561586,
      "tokens_seen": 1566834688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003115873841498014,
      "loss": 2.5809,
      "theoretical_loss": 3.502635976434588,
      "tokens_seen": 1566900224
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003115708341214299,
      "loss": 2.4178,
      "theoretical_loss": 3.5026231679390536,
      "tokens_seen": 1566965760
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031155428409305845,
      "loss": 2.5863,
      "theoretical_loss": 3.5026103601291925,
      "tokens_seen": 1567031296
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000311537734064687,
      "loss": 2.4916,
      "theoretical_loss": 3.5025975530049385,
      "tokens_seen": 1567096832
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031152118403631547,
      "loss": 2.4806,
      "theoretical_loss": 3.5025847465662268,
      "tokens_seen": 1567162368
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000311504634007944,
      "loss": 2.5362,
      "theoretical_loss": 3.5025719408129916,
      "tokens_seen": 1567227904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031148808397957255,
      "loss": 2.4799,
      "theoretical_loss": 3.5025591357451673,
      "tokens_seen": 1567293440
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003114715339512011,
      "loss": 2.6116,
      "theoretical_loss": 3.5025463313626894,
      "tokens_seen": 1567358976
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003114549839228296,
      "loss": 2.5313,
      "theoretical_loss": 3.502533527665492,
      "tokens_seen": 1567424512
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003114384338944581,
      "loss": 2.6653,
      "theoretical_loss": 3.5025207246535097,
      "tokens_seen": 1567490048
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031142188386608664,
      "loss": 2.3645,
      "theoretical_loss": 3.5025079223266777,
      "tokens_seen": 1567555584
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003114053338377151,
      "loss": 2.6566,
      "theoretical_loss": 3.50249512068493,
      "tokens_seen": 1567621120
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031138878380934366,
      "loss": 2.5099,
      "theoretical_loss": 3.502482319728202,
      "tokens_seen": 1567686656
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003113722337809722,
      "loss": 2.3511,
      "theoretical_loss": 3.5024695194564277,
      "tokens_seen": 1567752192
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003113556837526007,
      "loss": 2.4331,
      "theoretical_loss": 3.502456719869542,
      "tokens_seen": 1567817728
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003113391337242292,
      "loss": 2.521,
      "theoretical_loss": 3.5024439209674805,
      "tokens_seen": 1567883264
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031132258369585776,
      "loss": 2.5165,
      "theoretical_loss": 3.5024311227501768,
      "tokens_seen": 1567948800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003113060336674863,
      "loss": 2.9231,
      "theoretical_loss": 3.5024183252175662,
      "tokens_seen": 1568014336
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003112894836391148,
      "loss": 2.6568,
      "theoretical_loss": 3.502405528369583,
      "tokens_seen": 1568079872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003112729336107433,
      "loss": 2.3805,
      "theoretical_loss": 3.5023927322061637,
      "tokens_seen": 1568145408
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031125638358237186,
      "loss": 2.4981,
      "theoretical_loss": 3.50237993672724,
      "tokens_seen": 1568210944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003112398335540004,
      "loss": 2.7327,
      "theoretical_loss": 3.5023671419327496,
      "tokens_seen": 1568276480
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003112232835256289,
      "loss": 2.2194,
      "theoretical_loss": 3.5023543478226253,
      "tokens_seen": 1568342016
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003112067334972574,
      "loss": 2.4905,
      "theoretical_loss": 3.5023415543968026,
      "tokens_seen": 1568407552
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031119018346888595,
      "loss": 2.8176,
      "theoretical_loss": 3.5023287616552166,
      "tokens_seen": 1568473088
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003111736334405145,
      "loss": 2.7046,
      "theoretical_loss": 3.5023159695978023,
      "tokens_seen": 1568538624
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031115708341214297,
      "loss": 2.7371,
      "theoretical_loss": 3.5023031782244933,
      "tokens_seen": 1568604160
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003111405333837715,
      "loss": 2.4781,
      "theoretical_loss": 3.502290387535226,
      "tokens_seen": 1568669696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031112398335540005,
      "loss": 2.7739,
      "theoretical_loss": 3.5022775975299343,
      "tokens_seen": 1568735232
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003111074333270286,
      "loss": 2.6415,
      "theoretical_loss": 3.502264808208553,
      "tokens_seen": 1568800768
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031109088329865707,
      "loss": 2.4669,
      "theoretical_loss": 3.502252019571017,
      "tokens_seen": 1568866304
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003110743332702856,
      "loss": 2.4726,
      "theoretical_loss": 3.5022392316172617,
      "tokens_seen": 1568931840
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031105778324191414,
      "loss": 2.777,
      "theoretical_loss": 3.502226444347222,
      "tokens_seen": 1568997376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003110412332135427,
      "loss": 2.6712,
      "theoretical_loss": 3.502213657760832,
      "tokens_seen": 1569062912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031102468318517116,
      "loss": 2.8374,
      "theoretical_loss": 3.502200871858027,
      "tokens_seen": 1569128448
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003110081331567997,
      "loss": 2.6581,
      "theoretical_loss": 3.5021880866387423,
      "tokens_seen": 1569193984
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031099158312842824,
      "loss": 2.6113,
      "theoretical_loss": 3.5021753021029123,
      "tokens_seen": 1569259520
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003109750331000568,
      "loss": 2.8149,
      "theoretical_loss": 3.502162518250472,
      "tokens_seen": 1569325056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031095848307168526,
      "loss": 2.5095,
      "theoretical_loss": 3.5021497350813564,
      "tokens_seen": 1569390592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003109419330433138,
      "loss": 2.6213,
      "theoretical_loss": 3.5021369525955004,
      "tokens_seen": 1569456128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031092538301494234,
      "loss": 2.5271,
      "theoretical_loss": 3.5021241707928397,
      "tokens_seen": 1569521664
    },
    {
      "epoch": 0.56,
      "objective/train/avg_token_score": 0.027438264340162277,
      "objective/train/avg_weight": 0.967717707157135,
      "objective/train/docs_used": 888616,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.6523542404174805,
      "objective/train/std_weight": 0.12086282670497894,
      "objective/train/theoretical_loss": 3.5021113896733076,
      "objective/train/tokens_used": 1590047200,
      "theoretical_loss": 3.5021113896733076,
      "tokens_seen": 1569587200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003109088329865708,
      "loss": 2.5661,
      "theoretical_loss": 3.5021113896733076,
      "tokens_seen": 1569587200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031089228295819936,
      "loss": 2.4678,
      "theoretical_loss": 3.5020986092368407,
      "tokens_seen": 1569652736
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031087573292982784,
      "loss": 2.5813,
      "theoretical_loss": 3.5020858294833728,
      "tokens_seen": 1569718272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003108591829014564,
      "loss": 2.2686,
      "theoretical_loss": 3.5020730504128395,
      "tokens_seen": 1569783808
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003108426328730849,
      "loss": 2.5758,
      "theoretical_loss": 3.5020602720251763,
      "tokens_seen": 1569849344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031082608284471345,
      "loss": 2.6115,
      "theoretical_loss": 3.5020474943203173,
      "tokens_seen": 1569914880
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031080953281634194,
      "loss": 2.7063,
      "theoretical_loss": 3.502034717298198,
      "tokens_seen": 1569980416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003107929827879705,
      "loss": 2.429,
      "theoretical_loss": 3.5020219409587527,
      "tokens_seen": 1570045952
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000310776432759599,
      "loss": 2.7982,
      "theoretical_loss": 3.502009165301917,
      "tokens_seen": 1570111488
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031075988273122755,
      "loss": 2.6219,
      "theoretical_loss": 3.5019963903276263,
      "tokens_seen": 1570177024
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031074333270285603,
      "loss": 2.6344,
      "theoretical_loss": 3.5019836160358153,
      "tokens_seen": 1570242560
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031072678267448457,
      "loss": 2.6853,
      "theoretical_loss": 3.5019708424264184,
      "tokens_seen": 1570308096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003107102326461131,
      "loss": 2.6585,
      "theoretical_loss": 3.501958069499372,
      "tokens_seen": 1570373632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031069368261774164,
      "loss": 2.5905,
      "theoretical_loss": 3.5019452972546103,
      "tokens_seen": 1570439168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031067713258937013,
      "loss": 2.3748,
      "theoretical_loss": 3.5019325256920686,
      "tokens_seen": 1570504704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031066058256099866,
      "loss": 2.5519,
      "theoretical_loss": 3.5019197548116816,
      "tokens_seen": 1570570240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003106440325326272,
      "loss": 2.2807,
      "theoretical_loss": 3.501906984613385,
      "tokens_seen": 1570635776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031062748250425574,
      "loss": 2.5162,
      "theoretical_loss": 3.5018942150971135,
      "tokens_seen": 1570701312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003106109324758842,
      "loss": 2.7677,
      "theoretical_loss": 3.5018814462628027,
      "tokens_seen": 1570766848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031059438244751276,
      "loss": 2.3899,
      "theoretical_loss": 3.5018686781103874,
      "tokens_seen": 1570832384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003105778324191413,
      "loss": 2.4823,
      "theoretical_loss": 3.5018559106398026,
      "tokens_seen": 1570897920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031056128239076984,
      "loss": 2.5129,
      "theoretical_loss": 3.501843143850983,
      "tokens_seen": 1570963456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003105447323623983,
      "loss": 2.64,
      "theoretical_loss": 3.5018303777438646,
      "tokens_seen": 1571028992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031052818233402686,
      "loss": 2.3111,
      "theoretical_loss": 3.501817612318383,
      "tokens_seen": 1571094528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003105116323056554,
      "loss": 2.5151,
      "theoretical_loss": 3.5018048475744723,
      "tokens_seen": 1571160064
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031049508227728393,
      "loss": 2.4156,
      "theoretical_loss": 3.501792083512068,
      "tokens_seen": 1571225600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003104785322489124,
      "loss": 2.5095,
      "theoretical_loss": 3.5017793201311056,
      "tokens_seen": 1571291136
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031046198222054095,
      "loss": 2.7767,
      "theoretical_loss": 3.5017665574315195,
      "tokens_seen": 1571356672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003104454321921695,
      "loss": 2.6654,
      "theoretical_loss": 3.5017537954132463,
      "tokens_seen": 1571422208
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031042888216379803,
      "loss": 2.5123,
      "theoretical_loss": 3.5017410340762196,
      "tokens_seen": 1571487744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003104123321354265,
      "loss": 2.5021,
      "theoretical_loss": 3.501728273420376,
      "tokens_seen": 1571553280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000310395782107055,
      "loss": 2.68,
      "theoretical_loss": 3.5017155134456495,
      "tokens_seen": 1571618816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031037923207868353,
      "loss": 2.773,
      "theoretical_loss": 3.5017027541519763,
      "tokens_seen": 1571684352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031036268205031207,
      "loss": 2.7265,
      "theoretical_loss": 3.5016899955392917,
      "tokens_seen": 1571749888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003103461320219406,
      "loss": 2.5137,
      "theoretical_loss": 3.50167723760753,
      "tokens_seen": 1571815424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003103295819935691,
      "loss": 2.7636,
      "theoretical_loss": 3.5016644803566277,
      "tokens_seen": 1571880960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031031303196519763,
      "loss": 2.5597,
      "theoretical_loss": 3.501651723786519,
      "tokens_seen": 1571946496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031029648193682617,
      "loss": 2.4663,
      "theoretical_loss": 3.50163896789714,
      "tokens_seen": 1572012032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003102799319084547,
      "loss": 2.8005,
      "theoretical_loss": 3.5016262126884254,
      "tokens_seen": 1572077568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003102633818800832,
      "loss": 2.6897,
      "theoretical_loss": 3.5016134581603104,
      "tokens_seen": 1572143104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003102468318517117,
      "loss": 2.2954,
      "theoretical_loss": 3.5016007043127315,
      "tokens_seen": 1572208640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031023028182334026,
      "loss": 2.5774,
      "theoretical_loss": 3.5015879511456225,
      "tokens_seen": 1572274176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003102137317949688,
      "loss": 2.5884,
      "theoretical_loss": 3.50157519865892,
      "tokens_seen": 1572339712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031019718176659734,
      "loss": 2.82,
      "theoretical_loss": 3.5015624468525584,
      "tokens_seen": 1572405248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003101806317382258,
      "loss": 2.4745,
      "theoretical_loss": 3.5015496957264736,
      "tokens_seen": 1572470784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031016408170985436,
      "loss": 2.5624,
      "theoretical_loss": 3.5015369452806,
      "tokens_seen": 1572536320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003101475316814829,
      "loss": 2.5626,
      "theoretical_loss": 3.5015241955148744,
      "tokens_seen": 1572601856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031013098165311143,
      "loss": 2.6206,
      "theoretical_loss": 3.501511446429232,
      "tokens_seen": 1572667392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003101144316247399,
      "loss": 2.7473,
      "theoretical_loss": 3.5014986980236067,
      "tokens_seen": 1572732928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031009788159636845,
      "loss": 2.6463,
      "theoretical_loss": 3.5014859502979356,
      "tokens_seen": 1572798464
    },
    {
      "epoch": 0.56,
      "objective/train/avg_token_score": 0.011344833299517632,
      "objective/train/avg_weight": 0.9866426587104797,
      "objective/train/docs_used": 890514,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.153562545776367,
      "objective/train/std_weight": 0.07907364517450333,
      "objective/train/theoretical_loss": 3.501473203252153,
      "objective/train/tokens_used": 1593324000,
      "theoretical_loss": 3.501473203252153,
      "tokens_seen": 1572864000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000310081331567997,
      "loss": 2.3922,
      "theoretical_loss": 3.501473203252153,
      "tokens_seen": 1572864000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031006478153962553,
      "loss": 2.5828,
      "theoretical_loss": 3.501460456886195,
      "tokens_seen": 1572929536
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000310048231511254,
      "loss": 2.8004,
      "theoretical_loss": 3.5014477111999964,
      "tokens_seen": 1572995072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00031003168148288255,
      "loss": 2.5958,
      "theoretical_loss": 3.5014349661934934,
      "tokens_seen": 1573060608
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003100151314545111,
      "loss": 2.6697,
      "theoretical_loss": 3.501422221866621,
      "tokens_seen": 1573126144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003099985814261396,
      "loss": 2.6665,
      "theoretical_loss": 3.5014094782193146,
      "tokens_seen": 1573191680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003099820313977681,
      "loss": 2.7759,
      "theoretical_loss": 3.501396735251509,
      "tokens_seen": 1573257216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030996548136939665,
      "loss": 2.6208,
      "theoretical_loss": 3.501383992963141,
      "tokens_seen": 1573322752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003099489313410252,
      "loss": 2.676,
      "theoretical_loss": 3.5013712513541453,
      "tokens_seen": 1573388288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030993238131265367,
      "loss": 2.5754,
      "theoretical_loss": 3.501358510424458,
      "tokens_seen": 1573453824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003099158312842822,
      "loss": 2.7313,
      "theoretical_loss": 3.501345770174013,
      "tokens_seen": 1573519360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003098992812559107,
      "loss": 2.8325,
      "theoretical_loss": 3.501333030602748,
      "tokens_seen": 1573584896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003098827312275392,
      "loss": 2.614,
      "theoretical_loss": 3.501320291710597,
      "tokens_seen": 1573650432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030986618119916776,
      "loss": 2.6418,
      "theoretical_loss": 3.501307553497496,
      "tokens_seen": 1573715968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003098496311707963,
      "loss": 2.5344,
      "theoretical_loss": 3.501294815963381,
      "tokens_seen": 1573781504
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003098330811424248,
      "loss": 2.6127,
      "theoretical_loss": 3.5012820791081865,
      "tokens_seen": 1573847040
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003098165311140533,
      "loss": 2.9091,
      "theoretical_loss": 3.501269342931848,
      "tokens_seen": 1573912576
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030979998108568186,
      "loss": 2.5758,
      "theoretical_loss": 3.5012566074343026,
      "tokens_seen": 1573978112
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003097834310573104,
      "loss": 2.8612,
      "theoretical_loss": 3.5012438726154844,
      "tokens_seen": 1574043648
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003097668810289389,
      "loss": 2.5161,
      "theoretical_loss": 3.501231138475329,
      "tokens_seen": 1574109184
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003097503310005674,
      "loss": 2.585,
      "theoretical_loss": 3.5012184050137733,
      "tokens_seen": 1574174720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030973378097219595,
      "loss": 2.7337,
      "theoretical_loss": 3.5012056722307516,
      "tokens_seen": 1574240256
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003097172309438245,
      "loss": 2.8708,
      "theoretical_loss": 3.5011929401261996,
      "tokens_seen": 1574305792
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000309700680915453,
      "loss": 2.7971,
      "theoretical_loss": 3.5011802087000534,
      "tokens_seen": 1574371328
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003096841308870815,
      "loss": 2.7003,
      "theoretical_loss": 3.501167477952249,
      "tokens_seen": 1574436864
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030966758085871005,
      "loss": 2.7491,
      "theoretical_loss": 3.5011547478827207,
      "tokens_seen": 1574502400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003096510308303386,
      "loss": 2.757,
      "theoretical_loss": 3.5011420184914055,
      "tokens_seen": 1574567936
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030963448080196707,
      "loss": 2.1975,
      "theoretical_loss": 3.501129289778238,
      "tokens_seen": 1574633472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003096179307735956,
      "loss": 2.3713,
      "theoretical_loss": 3.5011165617431543,
      "tokens_seen": 1574699008
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030960138074522415,
      "loss": 2.7111,
      "theoretical_loss": 3.50110383438609,
      "tokens_seen": 1574764544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003095848307168527,
      "loss": 2.5131,
      "theoretical_loss": 3.501091107706981,
      "tokens_seen": 1574830080
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030956828068848117,
      "loss": 2.2763,
      "theoretical_loss": 3.5010783817057627,
      "tokens_seen": 1574895616
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003095517306601097,
      "loss": 2.8141,
      "theoretical_loss": 3.501065656382371,
      "tokens_seen": 1574961152
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030953518063173824,
      "loss": 2.6586,
      "theoretical_loss": 3.5010529317367416,
      "tokens_seen": 1575026688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003095186306033668,
      "loss": 2.7833,
      "theoretical_loss": 3.5010402077688094,
      "tokens_seen": 1575092224
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030950208057499526,
      "loss": 2.478,
      "theoretical_loss": 3.501027484478511,
      "tokens_seen": 1575157760
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003094855305466238,
      "loss": 2.3812,
      "theoretical_loss": 3.5010147618657825,
      "tokens_seen": 1575223296
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030946898051825234,
      "loss": 2.6293,
      "theoretical_loss": 3.501002039930558,
      "tokens_seen": 1575288832
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003094524304898809,
      "loss": 2.7215,
      "theoretical_loss": 3.500989318672775,
      "tokens_seen": 1575354368
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030943588046150936,
      "loss": 2.5948,
      "theoretical_loss": 3.5009765980923686,
      "tokens_seen": 1575419904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030941933043313784,
      "loss": 2.6139,
      "theoretical_loss": 3.5009638781892742,
      "tokens_seen": 1575485440
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003094027804047664,
      "loss": 2.8133,
      "theoretical_loss": 3.500951158963428,
      "tokens_seen": 1575550976
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003093862303763949,
      "loss": 2.6375,
      "theoretical_loss": 3.500938440414765,
      "tokens_seen": 1575616512
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030936968034802345,
      "loss": 2.7274,
      "theoretical_loss": 3.500925722543222,
      "tokens_seen": 1575682048
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030935313031965194,
      "loss": 2.5614,
      "theoretical_loss": 3.5009130053487345,
      "tokens_seen": 1575747584
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003093365802912805,
      "loss": 2.776,
      "theoretical_loss": 3.500900288831238,
      "tokens_seen": 1575813120
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000309320030262909,
      "loss": 2.4042,
      "theoretical_loss": 3.5008875729906688,
      "tokens_seen": 1575878656
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030930348023453755,
      "loss": 2.6374,
      "theoretical_loss": 3.500874857826962,
      "tokens_seen": 1575944192
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030928693020616603,
      "loss": 2.487,
      "theoretical_loss": 3.500862143340054,
      "tokens_seen": 1576009728
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030927038017779457,
      "loss": 2.5428,
      "theoretical_loss": 3.50084942952988,
      "tokens_seen": 1576075264
    },
    {
      "epoch": 0.56,
      "objective/train/avg_token_score": 0.038241319358348846,
      "objective/train/avg_weight": 0.9549973011016846,
      "objective/train/docs_used": 892478,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.766745090484619,
      "objective/train/std_weight": 0.14063760638237,
      "objective/train/theoretical_loss": 3.500836716396377,
      "objective/train/tokens_used": 1596600800,
      "theoretical_loss": 3.500836716396377,
      "tokens_seen": 1576140800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003092538301494231,
      "loss": 2.8357,
      "theoretical_loss": 3.500836716396377,
      "tokens_seen": 1576140800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030923728012105165,
      "loss": 2.4847,
      "theoretical_loss": 3.5008240039394796,
      "tokens_seen": 1576206336
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030922073009268013,
      "loss": 2.8147,
      "theoretical_loss": 3.5008112921591246,
      "tokens_seen": 1576271872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030920418006430867,
      "loss": 2.5467,
      "theoretical_loss": 3.5007985810552475,
      "tokens_seen": 1576337408
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003091876300359372,
      "loss": 2.7418,
      "theoretical_loss": 3.5007858706277837,
      "tokens_seen": 1576402944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030917108000756574,
      "loss": 2.6698,
      "theoretical_loss": 3.50077316087667,
      "tokens_seen": 1576468480
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003091545299791942,
      "loss": 2.7503,
      "theoretical_loss": 3.500760451801842,
      "tokens_seen": 1576534016
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030913797995082276,
      "loss": 2.7785,
      "theoretical_loss": 3.5007477434032346,
      "tokens_seen": 1576599552
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003091214299224513,
      "loss": 2.8517,
      "theoretical_loss": 3.500735035680785,
      "tokens_seen": 1576665088
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030910487989407984,
      "loss": 2.6202,
      "theoretical_loss": 3.5007223286344287,
      "tokens_seen": 1576730624
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003090883298657083,
      "loss": 2.5797,
      "theoretical_loss": 3.500709622264102,
      "tokens_seen": 1576796160
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030907177983733686,
      "loss": 2.8523,
      "theoretical_loss": 3.50069691656974,
      "tokens_seen": 1576861696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003090552298089654,
      "loss": 2.7669,
      "theoretical_loss": 3.5006842115512793,
      "tokens_seen": 1576927232
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030903867978059393,
      "loss": 2.6323,
      "theoretical_loss": 3.500671507208655,
      "tokens_seen": 1576992768
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003090221297522224,
      "loss": 2.493,
      "theoretical_loss": 3.5006588035418043,
      "tokens_seen": 1577058304
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030900557972385096,
      "loss": 2.4033,
      "theoretical_loss": 3.5006461005506626,
      "tokens_seen": 1577123840
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003089890296954795,
      "loss": 2.4613,
      "theoretical_loss": 3.5006333982351663,
      "tokens_seen": 1577189376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030897247966710803,
      "loss": 2.5705,
      "theoretical_loss": 3.5006206965952504,
      "tokens_seen": 1577254912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003089559296387365,
      "loss": 2.8578,
      "theoretical_loss": 3.500607995630851,
      "tokens_seen": 1577320448
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030893937961036505,
      "loss": 2.549,
      "theoretical_loss": 3.5005952953419053,
      "tokens_seen": 1577385984
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030892282958199353,
      "loss": 2.7927,
      "theoretical_loss": 3.5005825957283485,
      "tokens_seen": 1577451520
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030890627955362207,
      "loss": 2.4134,
      "theoretical_loss": 3.5005698967901164,
      "tokens_seen": 1577517056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003088897295252506,
      "loss": 2.5795,
      "theoretical_loss": 3.5005571985271455,
      "tokens_seen": 1577582592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030887317949687915,
      "loss": 2.5683,
      "theoretical_loss": 3.5005445009393723,
      "tokens_seen": 1577648128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030885662946850763,
      "loss": 2.4987,
      "theoretical_loss": 3.5005318040267315,
      "tokens_seen": 1577713664
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030884007944013617,
      "loss": 2.6319,
      "theoretical_loss": 3.50051910778916,
      "tokens_seen": 1577779200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003088235294117647,
      "loss": 2.6691,
      "theoretical_loss": 3.500506412226594,
      "tokens_seen": 1577844736
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030880697938339324,
      "loss": 2.7726,
      "theoretical_loss": 3.500493717338969,
      "tokens_seen": 1577910272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003087904293550217,
      "loss": 2.6351,
      "theoretical_loss": 3.5004810231262216,
      "tokens_seen": 1577975808
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030877387932665026,
      "loss": 2.5774,
      "theoretical_loss": 3.500468329588288,
      "tokens_seen": 1578041344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003087573292982788,
      "loss": 2.5365,
      "theoretical_loss": 3.5004556367251043,
      "tokens_seen": 1578106880
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030874077926990734,
      "loss": 2.7556,
      "theoretical_loss": 3.500442944536606,
      "tokens_seen": 1578172416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003087242292415358,
      "loss": 2.6585,
      "theoretical_loss": 3.500430253022729,
      "tokens_seen": 1578237952
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030870767921316436,
      "loss": 2.4187,
      "theoretical_loss": 3.5004175621834106,
      "tokens_seen": 1578303488
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003086911291847929,
      "loss": 1.975,
      "theoretical_loss": 3.5004048720185863,
      "tokens_seen": 1578369024
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030867457915642144,
      "loss": 2.6888,
      "theoretical_loss": 3.5003921825281923,
      "tokens_seen": 1578434560
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003086580291280499,
      "loss": 2.6894,
      "theoretical_loss": 3.500379493712165,
      "tokens_seen": 1578500096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030864147909967846,
      "loss": 2.7399,
      "theoretical_loss": 3.50036680557044,
      "tokens_seen": 1578565632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000308624929071307,
      "loss": 2.7224,
      "theoretical_loss": 3.500354118102954,
      "tokens_seen": 1578631168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030860837904293553,
      "loss": 2.7351,
      "theoretical_loss": 3.500341431309643,
      "tokens_seen": 1578696704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000308591829014564,
      "loss": 2.5921,
      "theoretical_loss": 3.5003287451904423,
      "tokens_seen": 1578762240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030857527898619255,
      "loss": 2.6634,
      "theoretical_loss": 3.5003160597452903,
      "tokens_seen": 1578827776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003085587289578211,
      "loss": 2.4717,
      "theoretical_loss": 3.500303374974121,
      "tokens_seen": 1578893312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030854217892944963,
      "loss": 2.6495,
      "theoretical_loss": 3.500290690876872,
      "tokens_seen": 1578958848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003085256289010781,
      "loss": 2.6013,
      "theoretical_loss": 3.5002780074534785,
      "tokens_seen": 1579024384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030850907887270665,
      "loss": 2.5688,
      "theoretical_loss": 3.5002653247038777,
      "tokens_seen": 1579089920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003084925288443352,
      "loss": 2.7742,
      "theoretical_loss": 3.5002526426280047,
      "tokens_seen": 1579155456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003084759788159637,
      "loss": 2.6076,
      "theoretical_loss": 3.5002399612257973,
      "tokens_seen": 1579220992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003084594287875922,
      "loss": 2.8227,
      "theoretical_loss": 3.5002272804971906,
      "tokens_seen": 1579286528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003084428787592207,
      "loss": 2.7478,
      "theoretical_loss": 3.5002146004421215,
      "tokens_seen": 1579352064
    },
    {
      "epoch": 0.56,
      "objective/train/avg_token_score": 0.019712278619408607,
      "objective/train/avg_weight": 0.9768515229225159,
      "objective/train/docs_used": 894462,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.770419120788574,
      "objective/train/std_weight": 0.1042497456073761,
      "objective/train/theoretical_loss": 3.5002019210605253,
      "objective/train/tokens_used": 1599877600,
      "theoretical_loss": 3.5002019210605253,
      "tokens_seen": 1579417600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030842632873084923,
      "loss": 2.7891,
      "theoretical_loss": 3.5002019210605253,
      "tokens_seen": 1579417600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030840977870247777,
      "loss": 2.6969,
      "theoretical_loss": 3.5001892423523397,
      "tokens_seen": 1579483136
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003083932286741063,
      "loss": 2.5801,
      "theoretical_loss": 3.5001765643174996,
      "tokens_seen": 1579548672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003083766786457348,
      "loss": 2.5444,
      "theoretical_loss": 3.500163886955942,
      "tokens_seen": 1579614208
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003083601286173633,
      "loss": 2.7966,
      "theoretical_loss": 3.500151210267604,
      "tokens_seen": 1579679744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030834357858899186,
      "loss": 2.5703,
      "theoretical_loss": 3.5001385342524203,
      "tokens_seen": 1579745280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003083270285606204,
      "loss": 2.7533,
      "theoretical_loss": 3.5001258589103283,
      "tokens_seen": 1579810816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003083104785322489,
      "loss": 2.2854,
      "theoretical_loss": 3.500113184241264,
      "tokens_seen": 1579876352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003082939285038774,
      "loss": 2.7224,
      "theoretical_loss": 3.5001005102451637,
      "tokens_seen": 1579941888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030827737847550596,
      "loss": 2.7607,
      "theoretical_loss": 3.500087836921964,
      "tokens_seen": 1580007424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003082608284471345,
      "loss": 2.6499,
      "theoretical_loss": 3.500075164271601,
      "tokens_seen": 1580072960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000308244278418763,
      "loss": 2.6396,
      "theoretical_loss": 3.500062492294011,
      "tokens_seen": 1580138496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003082277283903915,
      "loss": 2.4585,
      "theoretical_loss": 3.5000498209891306,
      "tokens_seen": 1580204032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030821117836202005,
      "loss": 2.9447,
      "theoretical_loss": 3.5000371503568966,
      "tokens_seen": 1580269568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003081946283336486,
      "loss": 2.7669,
      "theoretical_loss": 3.5000244803972453,
      "tokens_seen": 1580335104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003081780783052771,
      "loss": 2.4924,
      "theoretical_loss": 3.500011811110112,
      "tokens_seen": 1580400640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003081615282769056,
      "loss": 2.6707,
      "theoretical_loss": 3.499999142495434,
      "tokens_seen": 1580466176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030814497824853415,
      "loss": 2.3441,
      "theoretical_loss": 3.499986474553148,
      "tokens_seen": 1580531712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003081284282201627,
      "loss": 2.8644,
      "theoretical_loss": 3.49997380728319,
      "tokens_seen": 1580597248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030811187819179117,
      "loss": 2.76,
      "theoretical_loss": 3.499961140685496,
      "tokens_seen": 1580662784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003080953281634197,
      "loss": 2.6567,
      "theoretical_loss": 3.4999484747600036,
      "tokens_seen": 1580728320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030807877813504824,
      "loss": 2.462,
      "theoretical_loss": 3.499935809506648,
      "tokens_seen": 1580793856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003080622281066768,
      "loss": 2.5447,
      "theoretical_loss": 3.499923144925367,
      "tokens_seen": 1580859392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030804567807830527,
      "loss": 2.6955,
      "theoretical_loss": 3.499910481016096,
      "tokens_seen": 1580924928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003080291280499338,
      "loss": 2.7743,
      "theoretical_loss": 3.4998978177787716,
      "tokens_seen": 1580990464
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030801257802156234,
      "loss": 2.9432,
      "theoretical_loss": 3.499885155213331,
      "tokens_seen": 1581056000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003079960279931909,
      "loss": 2.6755,
      "theoretical_loss": 3.4998724933197103,
      "tokens_seen": 1581121536
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030797947796481936,
      "loss": 2.5187,
      "theoretical_loss": 3.4998598320978456,
      "tokens_seen": 1581187072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030796292793644785,
      "loss": 2.6577,
      "theoretical_loss": 3.4998471715476738,
      "tokens_seen": 1581252608
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003079463779080764,
      "loss": 2.5186,
      "theoretical_loss": 3.4998345116691314,
      "tokens_seen": 1581318144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003079298278797049,
      "loss": 2.8177,
      "theoretical_loss": 3.499821852462155,
      "tokens_seen": 1581383680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030791327785133346,
      "loss": 2.6531,
      "theoretical_loss": 3.499809193926681,
      "tokens_seen": 1581449216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030789672782296194,
      "loss": 2.5801,
      "theoretical_loss": 3.4997965360626466,
      "tokens_seen": 1581514752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003078801777945905,
      "loss": 2.5917,
      "theoretical_loss": 3.4997838788699873,
      "tokens_seen": 1581580288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000307863627766219,
      "loss": 2.7459,
      "theoretical_loss": 3.49977122234864,
      "tokens_seen": 1581645824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030784707773784755,
      "loss": 2.5805,
      "theoretical_loss": 3.499758566498542,
      "tokens_seen": 1581711360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030783052770947604,
      "loss": 2.3387,
      "theoretical_loss": 3.499745911319629,
      "tokens_seen": 1581776896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003078139776811046,
      "loss": 2.7388,
      "theoretical_loss": 3.499733256811838,
      "tokens_seen": 1581842432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0003077974276527331,
      "loss": 2.7268,
      "theoretical_loss": 3.4997206029751053,
      "tokens_seen": 1581907968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00030778087762436165,
      "loss": 2.6451,
      "theoretical_loss": 3.4997079498093684,
      "tokens_seen": 1581973504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030776432759599013,
      "loss": 2.7303,
      "theoretical_loss": 3.499695297314563,
      "tokens_seen": 1582039040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030774777756761867,
      "loss": 2.5399,
      "theoretical_loss": 3.4996826454906262,
      "tokens_seen": 1582104576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003077312275392472,
      "loss": 2.7256,
      "theoretical_loss": 3.499669994337494,
      "tokens_seen": 1582170112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030771467751087575,
      "loss": 2.5846,
      "theoretical_loss": 3.499657343855104,
      "tokens_seen": 1582235648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030769812748250423,
      "loss": 2.5025,
      "theoretical_loss": 3.499644694043392,
      "tokens_seen": 1582301184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030768157745413277,
      "loss": 2.556,
      "theoretical_loss": 3.4996320449022953,
      "tokens_seen": 1582366720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003076650274257613,
      "loss": 2.7227,
      "theoretical_loss": 3.4996193964317506,
      "tokens_seen": 1582432256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030764847739738984,
      "loss": 2.8127,
      "theoretical_loss": 3.4996067486316935,
      "tokens_seen": 1582497792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003076319273690183,
      "loss": 2.5623,
      "theoretical_loss": 3.499594101502062,
      "tokens_seen": 1582563328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030761537734064686,
      "loss": 2.6687,
      "theoretical_loss": 3.499581455042793,
      "tokens_seen": 1582628864
    },
    {
      "epoch": 0.57,
      "objective/train/avg_token_score": 0.02583514153957367,
      "objective/train/avg_weight": 0.9696108102798462,
      "objective/train/docs_used": 896326,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.4638614654541016,
      "objective/train/std_weight": 0.12554612755775452,
      "objective/train/theoretical_loss": 3.4995688092538213,
      "objective/train/tokens_used": 1603154400,
      "theoretical_loss": 3.4995688092538213,
      "tokens_seen": 1582694400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003075988273122754,
      "loss": 2.5646,
      "theoretical_loss": 3.4995688092538213,
      "tokens_seen": 1582694400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030758227728390394,
      "loss": 2.4141,
      "theoretical_loss": 3.499556164135085,
      "tokens_seen": 1582759936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003075657272555324,
      "loss": 2.3835,
      "theoretical_loss": 3.4995435196865214,
      "tokens_seen": 1582825472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030754917722716096,
      "loss": 2.4347,
      "theoretical_loss": 3.499530875908066,
      "tokens_seen": 1582891008
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003075326271987895,
      "loss": 2.6058,
      "theoretical_loss": 3.499518232799656,
      "tokens_seen": 1582956544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030751607717041803,
      "loss": 2.5319,
      "theoretical_loss": 3.4995055903612284,
      "tokens_seen": 1583022080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003074995271420465,
      "loss": 2.4266,
      "theoretical_loss": 3.49949294859272,
      "tokens_seen": 1583087616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030748297711367505,
      "loss": 2.6325,
      "theoretical_loss": 3.499480307494067,
      "tokens_seen": 1583153152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030746642708530354,
      "loss": 2.7242,
      "theoretical_loss": 3.4994676670652067,
      "tokens_seen": 1583218688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003074498770569321,
      "loss": 2.725,
      "theoretical_loss": 3.499455027306076,
      "tokens_seen": 1583284224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003074333270285606,
      "loss": 2.6017,
      "theoretical_loss": 3.49944238821661,
      "tokens_seen": 1583349760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030741677700018915,
      "loss": 2.7392,
      "theoretical_loss": 3.4994297497967484,
      "tokens_seen": 1583415296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030740022697181763,
      "loss": 2.643,
      "theoretical_loss": 3.4994171120464257,
      "tokens_seen": 1583480832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030738367694344617,
      "loss": 2.667,
      "theoretical_loss": 3.49940447496558,
      "tokens_seen": 1583546368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003073671269150747,
      "loss": 2.7126,
      "theoretical_loss": 3.499391838554147,
      "tokens_seen": 1583611904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030735057688670325,
      "loss": 2.5816,
      "theoretical_loss": 3.499379202812065,
      "tokens_seen": 1583677440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030733402685833173,
      "loss": 2.5263,
      "theoretical_loss": 3.49936656773927,
      "tokens_seen": 1583742976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030731747682996027,
      "loss": 2.6052,
      "theoretical_loss": 3.499353933335698,
      "tokens_seen": 1583808512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003073009268015888,
      "loss": 2.4496,
      "theoretical_loss": 3.4993412996012876,
      "tokens_seen": 1583874048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030728437677321734,
      "loss": 2.4749,
      "theoretical_loss": 3.4993286665359746,
      "tokens_seen": 1583939584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003072678267448458,
      "loss": 2.6984,
      "theoretical_loss": 3.4993160341396963,
      "tokens_seen": 1584005120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030725127671647436,
      "loss": 2.4078,
      "theoretical_loss": 3.499303402412389,
      "tokens_seen": 1584070656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003072347266881029,
      "loss": 2.5268,
      "theoretical_loss": 3.4992907713539907,
      "tokens_seen": 1584136192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030721817665973144,
      "loss": 2.8148,
      "theoretical_loss": 3.499278140964437,
      "tokens_seen": 1584201728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003072016266313599,
      "loss": 2.5949,
      "theoretical_loss": 3.4992655112436655,
      "tokens_seen": 1584267264
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030718507660298846,
      "loss": 2.5526,
      "theoretical_loss": 3.4992528821916133,
      "tokens_seen": 1584332800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000307168526574617,
      "loss": 2.5878,
      "theoretical_loss": 3.4992402538082166,
      "tokens_seen": 1584398336
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030715197654624553,
      "loss": 2.5193,
      "theoretical_loss": 3.4992276260934134,
      "tokens_seen": 1584463872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000307135426517874,
      "loss": 2.9922,
      "theoretical_loss": 3.49921499904714,
      "tokens_seen": 1584529408
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030711887648950256,
      "loss": 2.5226,
      "theoretical_loss": 3.499202372669333,
      "tokens_seen": 1584594944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003071023264611311,
      "loss": 2.46,
      "theoretical_loss": 3.4991897469599307,
      "tokens_seen": 1584660480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030708577643275963,
      "loss": 2.6002,
      "theoretical_loss": 3.4991771219188683,
      "tokens_seen": 1584726016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003070692264043881,
      "loss": 2.6778,
      "theoretical_loss": 3.4991644975460843,
      "tokens_seen": 1584791552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030705267637601665,
      "loss": 2.517,
      "theoretical_loss": 3.4991518738415146,
      "tokens_seen": 1584857088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003070361263476452,
      "loss": 2.4822,
      "theoretical_loss": 3.499139250805097,
      "tokens_seen": 1584922624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003070195763192737,
      "loss": 2.6764,
      "theoretical_loss": 3.499126628436768,
      "tokens_seen": 1584988160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003070030262909022,
      "loss": 2.5757,
      "theoretical_loss": 3.4991140067364648,
      "tokens_seen": 1585053696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003069864762625307,
      "loss": 2.6397,
      "theoretical_loss": 3.499101385704124,
      "tokens_seen": 1585119232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030696992623415923,
      "loss": 2.56,
      "theoretical_loss": 3.499088765339684,
      "tokens_seen": 1585184768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030695337620578777,
      "loss": 2.7138,
      "theoretical_loss": 3.4990761456430803,
      "tokens_seen": 1585250304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003069368261774163,
      "loss": 2.4908,
      "theoretical_loss": 3.4990635266142505,
      "tokens_seen": 1585315840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003069202761490448,
      "loss": 2.617,
      "theoretical_loss": 3.4990509082531323,
      "tokens_seen": 1585381376
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003069037261206733,
      "loss": 2.6268,
      "theoretical_loss": 3.4990382905596618,
      "tokens_seen": 1585446912
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030688717609230186,
      "loss": 2.7023,
      "theoretical_loss": 3.4990256735337764,
      "tokens_seen": 1585512448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003068706260639304,
      "loss": 2.7378,
      "theoretical_loss": 3.4990130571754134,
      "tokens_seen": 1585577984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003068540760355589,
      "loss": 2.6941,
      "theoretical_loss": 3.49900044148451,
      "tokens_seen": 1585643520
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003068375260071874,
      "loss": 2.6284,
      "theoretical_loss": 3.4989878264610024,
      "tokens_seen": 1585709056
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030682097597881596,
      "loss": 2.4393,
      "theoretical_loss": 3.4989752121048285,
      "tokens_seen": 1585774592
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003068044259504445,
      "loss": 2.4,
      "theoretical_loss": 3.4989625984159254,
      "tokens_seen": 1585840128
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000306787875922073,
      "loss": 2.5243,
      "theoretical_loss": 3.4989499853942307,
      "tokens_seen": 1585905664
    },
    {
      "epoch": 0.57,
      "objective/train/avg_token_score": 0.018824368715286255,
      "objective/train/avg_weight": 0.9778434634208679,
      "objective/train/docs_used": 897668,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.00390625,
      "objective/train/std_weight": 0.10168220847845078,
      "objective/train/theoretical_loss": 3.49893737303968,
      "objective/train/tokens_used": 1606431200,
      "theoretical_loss": 3.49893737303968,
      "tokens_seen": 1585971200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003067713258937015,
      "loss": 2.3288,
      "theoretical_loss": 3.49893737303968,
      "tokens_seen": 1585971200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030675477586533006,
      "loss": 2.5274,
      "theoretical_loss": 3.498924761352212,
      "tokens_seen": 1586036736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003067382258369586,
      "loss": 2.7226,
      "theoretical_loss": 3.498912150331763,
      "tokens_seen": 1586102272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003067216758085871,
      "loss": 2.2758,
      "theoretical_loss": 3.4988995399782707,
      "tokens_seen": 1586167808
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003067051257802156,
      "loss": 2.4081,
      "theoretical_loss": 3.498886930291672,
      "tokens_seen": 1586233344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030668857575184415,
      "loss": 2.5647,
      "theoretical_loss": 3.498874321271904,
      "tokens_seen": 1586298880
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003066720257234727,
      "loss": 2.5016,
      "theoretical_loss": 3.4988617129189037,
      "tokens_seen": 1586364416
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030665547569510117,
      "loss": 2.699,
      "theoretical_loss": 3.4988491052326087,
      "tokens_seen": 1586429952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003066389256667297,
      "loss": 2.623,
      "theoretical_loss": 3.498836498212956,
      "tokens_seen": 1586495488
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030662237563835825,
      "loss": 2.4767,
      "theoretical_loss": 3.498823891859883,
      "tokens_seen": 1586561024
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003066058256099868,
      "loss": 2.7199,
      "theoretical_loss": 3.4988112861733267,
      "tokens_seen": 1586626560
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030658927558161527,
      "loss": 2.633,
      "theoretical_loss": 3.498798681153225,
      "tokens_seen": 1586692096
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003065727255532438,
      "loss": 2.4507,
      "theoretical_loss": 3.4987860767995143,
      "tokens_seen": 1586757632
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030655617552487234,
      "loss": 2.4218,
      "theoretical_loss": 3.4987734731121316,
      "tokens_seen": 1586823168
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003065396254965009,
      "loss": 2.5273,
      "theoretical_loss": 3.498760870091015,
      "tokens_seen": 1586888704
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030652307546812936,
      "loss": 2.4154,
      "theoretical_loss": 3.4987482677361017,
      "tokens_seen": 1586954240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003065065254397579,
      "loss": 2.8128,
      "theoretical_loss": 3.4987356660473283,
      "tokens_seen": 1587019776
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003064899754113864,
      "loss": 2.6455,
      "theoretical_loss": 3.498723065024633,
      "tokens_seen": 1587085312
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003064734253830149,
      "loss": 2.467,
      "theoretical_loss": 3.498710464667952,
      "tokens_seen": 1587150848
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030645687535464346,
      "loss": 2.6787,
      "theoretical_loss": 3.4986978649772236,
      "tokens_seen": 1587216384
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030644032532627194,
      "loss": 2.706,
      "theoretical_loss": 3.4986852659523846,
      "tokens_seen": 1587281920
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003064237752979005,
      "loss": 2.5853,
      "theoretical_loss": 3.4986726675933726,
      "tokens_seen": 1587347456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000306407225269529,
      "loss": 2.2336,
      "theoretical_loss": 3.498660069900124,
      "tokens_seen": 1587412992
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030639067524115756,
      "loss": 2.5205,
      "theoretical_loss": 3.4986474728725776,
      "tokens_seen": 1587478528
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030637412521278604,
      "loss": 2.4106,
      "theoretical_loss": 3.4986348765106703,
      "tokens_seen": 1587544064
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003063575751844146,
      "loss": 2.7629,
      "theoretical_loss": 3.4986222808143386,
      "tokens_seen": 1587609600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003063410251560431,
      "loss": 2.6765,
      "theoretical_loss": 3.4986096857835207,
      "tokens_seen": 1587675136
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030632447512767165,
      "loss": 2.3999,
      "theoretical_loss": 3.4985970914181532,
      "tokens_seen": 1587740672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030630792509930014,
      "loss": 2.693,
      "theoretical_loss": 3.4985844977181744,
      "tokens_seen": 1587806208
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003062913750709287,
      "loss": 2.6495,
      "theoretical_loss": 3.4985719046835215,
      "tokens_seen": 1587871744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003062748250425572,
      "loss": 2.6666,
      "theoretical_loss": 3.498559312314131,
      "tokens_seen": 1587937280
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030625827501418575,
      "loss": 2.4882,
      "theoretical_loss": 3.4985467206099417,
      "tokens_seen": 1588002816
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003062417249858143,
      "loss": 2.3087,
      "theoretical_loss": 3.4985341295708903,
      "tokens_seen": 1588068352
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030622517495744277,
      "loss": 2.5337,
      "theoretical_loss": 3.4985215391969136,
      "tokens_seen": 1588133888
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003062086249290713,
      "loss": 2.4166,
      "theoretical_loss": 3.4985089494879498,
      "tokens_seen": 1588199424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030619207490069984,
      "loss": 2.5627,
      "theoretical_loss": 3.4984963604439363,
      "tokens_seen": 1588264960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003061755248723284,
      "loss": 2.5817,
      "theoretical_loss": 3.49848377206481,
      "tokens_seen": 1588330496
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030615897484395687,
      "loss": 2.5832,
      "theoretical_loss": 3.4984711843505094,
      "tokens_seen": 1588396032
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003061424248155854,
      "loss": 2.6343,
      "theoretical_loss": 3.4984585973009708,
      "tokens_seen": 1588461568
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030612587478721394,
      "loss": 2.6439,
      "theoretical_loss": 3.4984460109161324,
      "tokens_seen": 1588527104
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003061093247588425,
      "loss": 2.8371,
      "theoretical_loss": 3.4984334251959313,
      "tokens_seen": 1588592640
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030609277473047096,
      "loss": 2.7745,
      "theoretical_loss": 3.498420840140305,
      "tokens_seen": 1588658176
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003060762247020995,
      "loss": 2.5564,
      "theoretical_loss": 3.498408255749191,
      "tokens_seen": 1588723712
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030605967467372804,
      "loss": 2.6609,
      "theoretical_loss": 3.4983956720225278,
      "tokens_seen": 1588789248
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003060431246453566,
      "loss": 2.6773,
      "theoretical_loss": 3.498383088960251,
      "tokens_seen": 1588854784
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030602657461698506,
      "loss": 2.4973,
      "theoretical_loss": 3.4983705065623,
      "tokens_seen": 1588920320
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003060100245886136,
      "loss": 2.5043,
      "theoretical_loss": 3.4983579248286114,
      "tokens_seen": 1588985856
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003059934745602421,
      "loss": 2.6407,
      "theoretical_loss": 3.4983453437591225,
      "tokens_seen": 1589051392
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003059769245318706,
      "loss": 2.6146,
      "theoretical_loss": 3.498332763353771,
      "tokens_seen": 1589116928
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030596037450349915,
      "loss": 2.5826,
      "theoretical_loss": 3.498320183612495,
      "tokens_seen": 1589182464
    },
    {
      "epoch": 0.57,
      "objective/train/avg_token_score": 0.05539172142744064,
      "objective/train/avg_weight": 0.9348219037055969,
      "objective/train/docs_used": 899481,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8025293350219727,
      "objective/train/std_weight": 0.22229690849781036,
      "objective/train/theoretical_loss": 3.4983076045352317,
      "objective/train/tokens_used": 1609708000,
      "theoretical_loss": 3.4983076045352317,
      "tokens_seen": 1589248000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030594382447512764,
      "loss": 2.3671,
      "theoretical_loss": 3.4983076045352317,
      "tokens_seen": 1589248000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003059272744467562,
      "loss": 2.3763,
      "theoretical_loss": 3.4982950261219186,
      "tokens_seen": 1589313536
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003059107244183847,
      "loss": 2.6544,
      "theoretical_loss": 3.4982824483724935,
      "tokens_seen": 1589379072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030589417439001325,
      "loss": 2.6095,
      "theoretical_loss": 3.498269871286894,
      "tokens_seen": 1589444608
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030587762436164173,
      "loss": 2.442,
      "theoretical_loss": 3.4982572948650565,
      "tokens_seen": 1589510144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030586107433327027,
      "loss": 2.5359,
      "theoretical_loss": 3.4982447191069204,
      "tokens_seen": 1589575680
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003058445243048988,
      "loss": 2.3928,
      "theoretical_loss": 3.4982321440124227,
      "tokens_seen": 1589641216
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030582797427652735,
      "loss": 2.4165,
      "theoretical_loss": 3.498219569581501,
      "tokens_seen": 1589706752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030581142424815583,
      "loss": 2.6434,
      "theoretical_loss": 3.4982069958140922,
      "tokens_seen": 1589772288
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030579487421978437,
      "loss": 2.6253,
      "theoretical_loss": 3.4981944227101347,
      "tokens_seen": 1589837824
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003057783241914129,
      "loss": 2.4963,
      "theoretical_loss": 3.498181850269566,
      "tokens_seen": 1589903360
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030576177416304144,
      "loss": 2.3492,
      "theoretical_loss": 3.498169278492324,
      "tokens_seen": 1589968896
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003057452241346699,
      "loss": 2.7212,
      "theoretical_loss": 3.498156707378346,
      "tokens_seen": 1590034432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030572867410629846,
      "loss": 2.5622,
      "theoretical_loss": 3.4981441369275696,
      "tokens_seen": 1590099968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000305712124077927,
      "loss": 2.6609,
      "theoretical_loss": 3.4981315671399327,
      "tokens_seen": 1590165504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030569557404955554,
      "loss": 2.5083,
      "theoretical_loss": 3.498118998015373,
      "tokens_seen": 1590231040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000305679024021184,
      "loss": 2.7041,
      "theoretical_loss": 3.498106429553828,
      "tokens_seen": 1590296576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030566247399281256,
      "loss": 2.797,
      "theoretical_loss": 3.498093861755236,
      "tokens_seen": 1590362112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003056459239644411,
      "loss": 2.3715,
      "theoretical_loss": 3.4980812946195337,
      "tokens_seen": 1590427648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030562937393606963,
      "loss": 2.3879,
      "theoretical_loss": 3.4980687281466594,
      "tokens_seen": 1590493184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003056128239076981,
      "loss": 2.6582,
      "theoretical_loss": 3.4980561623365514,
      "tokens_seen": 1590558720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030559627387932665,
      "loss": 2.6168,
      "theoretical_loss": 3.498043597189146,
      "tokens_seen": 1590624256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003055797238509552,
      "loss": 2.7231,
      "theoretical_loss": 3.4980310327043824,
      "tokens_seen": 1590689792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030556317382258373,
      "loss": 2.4368,
      "theoretical_loss": 3.498018468882198,
      "tokens_seen": 1590755328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003055466237942122,
      "loss": 2.7312,
      "theoretical_loss": 3.49800590572253,
      "tokens_seen": 1590820864
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030553007376584075,
      "loss": 2.4014,
      "theoretical_loss": 3.4979933432253163,
      "tokens_seen": 1590886400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003055135237374693,
      "loss": 2.811,
      "theoretical_loss": 3.497980781390495,
      "tokens_seen": 1590951936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030549697370909777,
      "loss": 2.559,
      "theoretical_loss": 3.497968220218004,
      "tokens_seen": 1591017472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003054804236807263,
      "loss": 2.7129,
      "theoretical_loss": 3.49795565970778,
      "tokens_seen": 1591083008
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003054638736523548,
      "loss": 2.4411,
      "theoretical_loss": 3.4979430998597625,
      "tokens_seen": 1591148544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030544732362398333,
      "loss": 2.5644,
      "theoretical_loss": 3.4979305406738885,
      "tokens_seen": 1591214080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030543077359561187,
      "loss": 2.5617,
      "theoretical_loss": 3.4979179821500956,
      "tokens_seen": 1591279616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003054142235672404,
      "loss": 2.599,
      "theoretical_loss": 3.497905424288322,
      "tokens_seen": 1591345152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003053976735388689,
      "loss": 2.7662,
      "theoretical_loss": 3.497892867088505,
      "tokens_seen": 1591410688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003053811235104974,
      "loss": 2.7996,
      "theoretical_loss": 3.4978803105505833,
      "tokens_seen": 1591476224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030536457348212596,
      "loss": 2.2804,
      "theoretical_loss": 3.497867754674494,
      "tokens_seen": 1591541760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003053480234537545,
      "loss": 2.5844,
      "theoretical_loss": 3.4978551994601754,
      "tokens_seen": 1591607296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000305331473425383,
      "loss": 2.6624,
      "theoretical_loss": 3.497842644907565,
      "tokens_seen": 1591672832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003053149233970115,
      "loss": 2.6985,
      "theoretical_loss": 3.4978300910166014,
      "tokens_seen": 1591738368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030529837336864006,
      "loss": 2.721,
      "theoretical_loss": 3.4978175377872214,
      "tokens_seen": 1591803904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003052818233402686,
      "loss": 2.4091,
      "theoretical_loss": 3.497804985219364,
      "tokens_seen": 1591869440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003052652733118971,
      "loss": 2.4354,
      "theoretical_loss": 3.4977924333129664,
      "tokens_seen": 1591934976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003052487232835256,
      "loss": 2.5702,
      "theoretical_loss": 3.4977798820679666,
      "tokens_seen": 1592000512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030523217325515415,
      "loss": 2.5114,
      "theoretical_loss": 3.4977673314843027,
      "tokens_seen": 1592066048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003052156232267827,
      "loss": 2.6127,
      "theoretical_loss": 3.4977547815619126,
      "tokens_seen": 1592131584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003051990731984112,
      "loss": 2.5039,
      "theoretical_loss": 3.497742232300734,
      "tokens_seen": 1592197120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003051825231700397,
      "loss": 2.47,
      "theoretical_loss": 3.4977296837007055,
      "tokens_seen": 1592262656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030516597314166825,
      "loss": 2.7215,
      "theoretical_loss": 3.4977171357617642,
      "tokens_seen": 1592328192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003051494231132968,
      "loss": 2.7227,
      "theoretical_loss": 3.497704588483849,
      "tokens_seen": 1592393728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030513287308492527,
      "loss": 2.5616,
      "theoretical_loss": 3.497692041866897,
      "tokens_seen": 1592459264
    },
    {
      "epoch": 0.57,
      "objective/train/avg_token_score": 0.008998654782772064,
      "objective/train/avg_weight": 0.9894756078720093,
      "objective/train/docs_used": 902104,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4315249919891357,
      "objective/train/std_weight": 0.05187848582863808,
      "objective/train/theoretical_loss": 3.4976794959108464,
      "objective/train/tokens_used": 1612984800,
      "theoretical_loss": 3.4976794959108464,
      "tokens_seen": 1592524800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003051163230565538,
      "loss": 2.2196,
      "theoretical_loss": 3.4976794959108464,
      "tokens_seen": 1592524800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030509977302818235,
      "loss": 2.732,
      "theoretical_loss": 3.4976669506156357,
      "tokens_seen": 1592590336
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003050832229998109,
      "loss": 2.5443,
      "theoretical_loss": 3.4976544059812023,
      "tokens_seen": 1592655872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003050832229998109,
      "loss": 4.9737,
      "theoretical_loss": 3.4976418620074847,
      "tokens_seen": 1592721408
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030506667297143937,
      "loss": 2.8284,
      "theoretical_loss": 3.49762931869442,
      "tokens_seen": 1592786944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003050501229430679,
      "loss": 3.7164,
      "theoretical_loss": 3.497616776041948,
      "tokens_seen": 1592852480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030503357291469644,
      "loss": 2.6576,
      "theoretical_loss": 3.497604234050005,
      "tokens_seen": 1592918016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003050170228863249,
      "loss": 2.5661,
      "theoretical_loss": 3.4975916927185295,
      "tokens_seen": 1592983552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030500047285795346,
      "loss": 2.6621,
      "theoretical_loss": 3.49757915204746,
      "tokens_seen": 1593049088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000304983922829582,
      "loss": 2.7039,
      "theoretical_loss": 3.497566612036734,
      "tokens_seen": 1593114624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003049673728012105,
      "loss": 2.718,
      "theoretical_loss": 3.49755407268629,
      "tokens_seen": 1593180160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000304950822772839,
      "loss": 2.3344,
      "theoretical_loss": 3.497541533996066,
      "tokens_seen": 1593245696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030493427274446756,
      "loss": 2.5503,
      "theoretical_loss": 3.497528995966,
      "tokens_seen": 1593311232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003049177227160961,
      "loss": 2.5098,
      "theoretical_loss": 3.4975164585960297,
      "tokens_seen": 1593376768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003049011726877246,
      "loss": 2.6554,
      "theoretical_loss": 3.497503921886094,
      "tokens_seen": 1593442304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003048846226593531,
      "loss": 2.789,
      "theoretical_loss": 3.4974913858361303,
      "tokens_seen": 1593507840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030486807263098166,
      "loss": 2.7346,
      "theoretical_loss": 3.497478850446077,
      "tokens_seen": 1593573376
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003048515226026102,
      "loss": 2.5642,
      "theoretical_loss": 3.4974663157158727,
      "tokens_seen": 1593638912
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003048349725742387,
      "loss": 2.8208,
      "theoretical_loss": 3.497453781645455,
      "tokens_seen": 1593704448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003048184225458672,
      "loss": 2.7464,
      "theoretical_loss": 3.497441248234762,
      "tokens_seen": 1593769984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030480187251749575,
      "loss": 2.7571,
      "theoretical_loss": 3.4974287154837316,
      "tokens_seen": 1593835520
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003047853224891243,
      "loss": 2.5692,
      "theoretical_loss": 3.4974161833923025,
      "tokens_seen": 1593901056
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030476877246075277,
      "loss": 2.7087,
      "theoretical_loss": 3.497403651960413,
      "tokens_seen": 1593966592
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003047522224323813,
      "loss": 2.7311,
      "theoretical_loss": 3.4973911211880004,
      "tokens_seen": 1594032128
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030473567240400985,
      "loss": 2.6801,
      "theoretical_loss": 3.4973785910750035,
      "tokens_seen": 1594097664
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003047191223756384,
      "loss": 2.5672,
      "theoretical_loss": 3.497366061621361,
      "tokens_seen": 1594163200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030470257234726687,
      "loss": 2.8377,
      "theoretical_loss": 3.4973535328270104,
      "tokens_seen": 1594228736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003046860223188954,
      "loss": 2.6539,
      "theoretical_loss": 3.4973410046918896,
      "tokens_seen": 1594294272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030466947229052394,
      "loss": 2.7374,
      "theoretical_loss": 3.4973284772159374,
      "tokens_seen": 1594359808
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003046529222621525,
      "loss": 2.4439,
      "theoretical_loss": 3.4973159503990923,
      "tokens_seen": 1594425344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030463637223378096,
      "loss": 2.5569,
      "theoretical_loss": 3.4973034242412915,
      "tokens_seen": 1594490880
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003046198222054095,
      "loss": 2.7812,
      "theoretical_loss": 3.4972908987424747,
      "tokens_seen": 1594556416
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030460327217703804,
      "loss": 2.7093,
      "theoretical_loss": 3.4972783739025783,
      "tokens_seen": 1594621952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003045867221486666,
      "loss": 2.6591,
      "theoretical_loss": 3.497265849721542,
      "tokens_seen": 1594687488
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030457017212029506,
      "loss": 2.7505,
      "theoretical_loss": 3.4972533261993037,
      "tokens_seen": 1594753024
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003045536220919236,
      "loss": 2.397,
      "theoretical_loss": 3.497240803335802,
      "tokens_seen": 1594818560
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030453707206355214,
      "loss": 2.6312,
      "theoretical_loss": 3.4972282811309743,
      "tokens_seen": 1594884096
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003045205220351806,
      "loss": 2.7383,
      "theoretical_loss": 3.4972157595847593,
      "tokens_seen": 1594949632
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030450397200680916,
      "loss": 2.5425,
      "theoretical_loss": 3.4972032386970957,
      "tokens_seen": 1595015168
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030448742197843764,
      "loss": 2.6729,
      "theoretical_loss": 3.4971907184679214,
      "tokens_seen": 1595080704
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003044708719500662,
      "loss": 2.4701,
      "theoretical_loss": 3.4971781988971746,
      "tokens_seen": 1595146240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003044543219216947,
      "loss": 2.7104,
      "theoretical_loss": 3.497165679984794,
      "tokens_seen": 1595211776
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030443777189332325,
      "loss": 2.6397,
      "theoretical_loss": 3.4971531617307177,
      "tokens_seen": 1595277312
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030442122186495174,
      "loss": 2.5526,
      "theoretical_loss": 3.497140644134884,
      "tokens_seen": 1595342848
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030440467183658027,
      "loss": 2.4987,
      "theoretical_loss": 3.4971281271972314,
      "tokens_seen": 1595408384
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003043881218082088,
      "loss": 2.7637,
      "theoretical_loss": 3.4971156109176986,
      "tokens_seen": 1595473920
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030437157177983735,
      "loss": 2.5407,
      "theoretical_loss": 3.4971030952962234,
      "tokens_seen": 1595539456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030435502175146583,
      "loss": 2.7224,
      "theoretical_loss": 3.497090580332744,
      "tokens_seen": 1595604992
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030433847172309437,
      "loss": 2.6716,
      "theoretical_loss": 3.4970780660271994,
      "tokens_seen": 1595670528
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003043219216947229,
      "loss": 2.7959,
      "theoretical_loss": 3.497065552379528,
      "tokens_seen": 1595736064
    },
    {
      "epoch": 0.57,
      "objective/train/avg_token_score": 0.013247211463749409,
      "objective/train/avg_weight": 0.9844453930854797,
      "objective/train/docs_used": 903830,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.37144136428833,
      "objective/train/std_weight": 0.05648886784911156,
      "objective/train/theoretical_loss": 3.4970530393896673,
      "objective/train/tokens_used": 1616261600,
      "theoretical_loss": 3.4970530393896673,
      "tokens_seen": 1595801600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030430537166635144,
      "loss": 2.5078,
      "theoretical_loss": 3.4970530393896673,
      "tokens_seen": 1595801600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030428882163797993,
      "loss": 2.4868,
      "theoretical_loss": 3.497040527057557,
      "tokens_seen": 1595867136
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030427227160960846,
      "loss": 2.3832,
      "theoretical_loss": 3.497028015383134,
      "tokens_seen": 1595932672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000304255721581237,
      "loss": 2.4731,
      "theoretical_loss": 3.4970155043663382,
      "tokens_seen": 1595998208
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030423917155286554,
      "loss": 2.5947,
      "theoretical_loss": 3.4970029940071075,
      "tokens_seen": 1596063744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000304222621524494,
      "loss": 3.563,
      "theoretical_loss": 3.49699048430538,
      "tokens_seen": 1596129280
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030420607149612256,
      "loss": 2.6518,
      "theoretical_loss": 3.4969779752610943,
      "tokens_seen": 1596194816
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003041895214677511,
      "loss": 2.7479,
      "theoretical_loss": 3.4969654668741894,
      "tokens_seen": 1596260352
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030417297143937964,
      "loss": 2.6625,
      "theoretical_loss": 3.496952959144603,
      "tokens_seen": 1596325888
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003041564214110081,
      "loss": 2.6806,
      "theoretical_loss": 3.496940452072274,
      "tokens_seen": 1596391424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030413987138263666,
      "loss": 2.7818,
      "theoretical_loss": 3.4969279456571405,
      "tokens_seen": 1596456960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003041233213542652,
      "loss": 2.5748,
      "theoretical_loss": 3.4969154398991416,
      "tokens_seen": 1596522496
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030410677132589373,
      "loss": 2.5872,
      "theoretical_loss": 3.4969029347982152,
      "tokens_seen": 1596588032
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003040902212975222,
      "loss": 2.673,
      "theoretical_loss": 3.4968904303543002,
      "tokens_seen": 1596653568
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030407367126915075,
      "loss": 2.4005,
      "theoretical_loss": 3.4968779265673353,
      "tokens_seen": 1596719104
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003040571212407793,
      "loss": 2.6717,
      "theoretical_loss": 3.4968654234372583,
      "tokens_seen": 1596784640
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030404057121240783,
      "loss": 2.8028,
      "theoretical_loss": 3.4968529209640087,
      "tokens_seen": 1596850176
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003040240211840363,
      "loss": 2.4705,
      "theoretical_loss": 3.4968404191475235,
      "tokens_seen": 1596915712
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003040074711556648,
      "loss": 2.8359,
      "theoretical_loss": 3.4968279179877433,
      "tokens_seen": 1596981248
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030399092112729333,
      "loss": 2.4059,
      "theoretical_loss": 3.496815417484605,
      "tokens_seen": 1597046784
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030397437109892187,
      "loss": 2.4988,
      "theoretical_loss": 3.496802917638048,
      "tokens_seen": 1597112320
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003039578210705504,
      "loss": 2.8091,
      "theoretical_loss": 3.4967904184480103,
      "tokens_seen": 1597177856
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003039412710421789,
      "loss": 2.6892,
      "theoretical_loss": 3.4967779199144307,
      "tokens_seen": 1597243392
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030392472101380743,
      "loss": 2.6786,
      "theoretical_loss": 3.4967654220372486,
      "tokens_seen": 1597308928
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030390817098543597,
      "loss": 2.6923,
      "theoretical_loss": 3.496752924816401,
      "tokens_seen": 1597374464
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003038916209570645,
      "loss": 2.6058,
      "theoretical_loss": 3.496740428251828,
      "tokens_seen": 1597440000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000303875070928693,
      "loss": 2.648,
      "theoretical_loss": 3.4967279323434677,
      "tokens_seen": 1597505536
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003038585209003215,
      "loss": 2.7074,
      "theoretical_loss": 3.4967154370912583,
      "tokens_seen": 1597571072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030384197087195006,
      "loss": 2.7388,
      "theoretical_loss": 3.4967029424951392,
      "tokens_seen": 1597636608
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003038254208435786,
      "loss": 2.426,
      "theoretical_loss": 3.4966904485550483,
      "tokens_seen": 1597702144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003038088708152071,
      "loss": 2.4517,
      "theoretical_loss": 3.4966779552709246,
      "tokens_seen": 1597767680
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003037923207868356,
      "loss": 2.5492,
      "theoretical_loss": 3.496665462642707,
      "tokens_seen": 1597833216
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030377577075846416,
      "loss": 2.4795,
      "theoretical_loss": 3.4966529706703335,
      "tokens_seen": 1597898752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003037592207300927,
      "loss": 2.5871,
      "theoretical_loss": 3.4966404793537436,
      "tokens_seen": 1597964288
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003037426707017212,
      "loss": 2.8442,
      "theoretical_loss": 3.4966279886928753,
      "tokens_seen": 1598029824
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003037261206733497,
      "loss": 2.658,
      "theoretical_loss": 3.4966154986876674,
      "tokens_seen": 1598095360
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030370957064497825,
      "loss": 2.4824,
      "theoretical_loss": 3.4966030093380587,
      "tokens_seen": 1598160896
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003036930206166068,
      "loss": 2.4931,
      "theoretical_loss": 3.496590520643988,
      "tokens_seen": 1598226432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003036764705882353,
      "loss": 2.8629,
      "theoretical_loss": 3.496578032605394,
      "tokens_seen": 1598291968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003036599205598638,
      "loss": 2.4565,
      "theoretical_loss": 3.4965655452222153,
      "tokens_seen": 1598357504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030364337053149235,
      "loss": 2.7401,
      "theoretical_loss": 3.4965530584943907,
      "tokens_seen": 1598423040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003036268205031209,
      "loss": 2.6304,
      "theoretical_loss": 3.496540572421859,
      "tokens_seen": 1598488576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030361027047474937,
      "loss": 2.652,
      "theoretical_loss": 3.4965280870045587,
      "tokens_seen": 1598554112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003035937204463779,
      "loss": 2.5245,
      "theoretical_loss": 3.496515602242429,
      "tokens_seen": 1598619648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030357717041800645,
      "loss": 2.6288,
      "theoretical_loss": 3.4965031181354083,
      "tokens_seen": 1598685184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000303560620389635,
      "loss": 2.8436,
      "theoretical_loss": 3.4964906346834352,
      "tokens_seen": 1598750720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030354407036126347,
      "loss": 2.6876,
      "theoretical_loss": 3.496478151886449,
      "tokens_seen": 1598816256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000303527520332892,
      "loss": 2.6192,
      "theoretical_loss": 3.4964656697443885,
      "tokens_seen": 1598881792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003035109703045205,
      "loss": 2.4428,
      "theoretical_loss": 3.496453188257192,
      "tokens_seen": 1598947328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000303494420276149,
      "loss": 2.4503,
      "theoretical_loss": 3.4964407074247985,
      "tokens_seen": 1599012864
    },
    {
      "epoch": 0.57,
      "objective/train/avg_token_score": 0.009716270491480827,
      "objective/train/avg_weight": 0.9885661602020264,
      "objective/train/docs_used": 906083,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8931713104248047,
      "objective/train/std_weight": 0.057287152856588364,
      "objective/train/theoretical_loss": 3.4964282272471467,
      "objective/train/tokens_used": 1619538400,
      "theoretical_loss": 3.4964282272471467,
      "tokens_seen": 1599078400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030347787024777756,
      "loss": 2.6067,
      "theoretical_loss": 3.4964282272471467,
      "tokens_seen": 1599078400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003034613202194061,
      "loss": 2.5333,
      "theoretical_loss": 3.4964157477241757,
      "tokens_seen": 1599143936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003034447701910346,
      "loss": 2.6148,
      "theoretical_loss": 3.4964032688558246,
      "tokens_seen": 1599209472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003034282201626631,
      "loss": 2.5485,
      "theoretical_loss": 3.4963907906420313,
      "tokens_seen": 1599275008
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030341167013429166,
      "loss": 2.6644,
      "theoretical_loss": 3.4963783130827353,
      "tokens_seen": 1599340544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003033951201059202,
      "loss": 2.5301,
      "theoretical_loss": 3.4963658361778753,
      "tokens_seen": 1599406080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003033785700775487,
      "loss": 2.3573,
      "theoretical_loss": 3.49635335992739,
      "tokens_seen": 1599471616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003033620200491772,
      "loss": 2.7335,
      "theoretical_loss": 3.496340884331219,
      "tokens_seen": 1599537152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030334547002080575,
      "loss": 2.68,
      "theoretical_loss": 3.4963284093893003,
      "tokens_seen": 1599602688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003033289199924343,
      "loss": 2.5356,
      "theoretical_loss": 3.4963159351015736,
      "tokens_seen": 1599668224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003033123699640628,
      "loss": 2.6191,
      "theoretical_loss": 3.496303461467977,
      "tokens_seen": 1599733760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003032958199356913,
      "loss": 2.7423,
      "theoretical_loss": 3.4962909884884494,
      "tokens_seen": 1599799296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030327926990731985,
      "loss": 2.5489,
      "theoretical_loss": 3.4962785161629304,
      "tokens_seen": 1599864832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003032627198789484,
      "loss": 2.6711,
      "theoretical_loss": 3.4962660444913585,
      "tokens_seen": 1599930368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030324616985057687,
      "loss": 2.6699,
      "theoretical_loss": 3.4962535734736733,
      "tokens_seen": 1599995904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003032296198222054,
      "loss": 2.4866,
      "theoretical_loss": 3.496241103109812,
      "tokens_seen": 1600061440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030321306979383395,
      "loss": 2.5446,
      "theoretical_loss": 3.4962286333997152,
      "tokens_seen": 1600126976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003031965197654625,
      "loss": 2.6128,
      "theoretical_loss": 3.4962161643433216,
      "tokens_seen": 1600192512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030317996973709097,
      "loss": 2.5329,
      "theoretical_loss": 3.4962036959405696,
      "tokens_seen": 1600258048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003031634197087195,
      "loss": 2.6147,
      "theoretical_loss": 3.4961912281913987,
      "tokens_seen": 1600323584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030314686968034804,
      "loss": 2.4394,
      "theoretical_loss": 3.496178761095747,
      "tokens_seen": 1600389120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003031303196519766,
      "loss": 2.6361,
      "theoretical_loss": 3.4961662946535546,
      "tokens_seen": 1600454656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030311376962360506,
      "loss": 2.8582,
      "theoretical_loss": 3.49615382886476,
      "tokens_seen": 1600520192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003030972195952336,
      "loss": 2.6645,
      "theoretical_loss": 3.496141363729302,
      "tokens_seen": 1600585728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030308066956686214,
      "loss": 2.66,
      "theoretical_loss": 3.49612889924712,
      "tokens_seen": 1600651264
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003030641195384907,
      "loss": 2.7422,
      "theoretical_loss": 3.496116435418153,
      "tokens_seen": 1600716800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030304756951011916,
      "loss": 2.5016,
      "theoretical_loss": 3.4961039722423397,
      "tokens_seen": 1600782336
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030303101948174764,
      "loss": 2.5796,
      "theoretical_loss": 3.4960915097196192,
      "tokens_seen": 1600847872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003030144694533762,
      "loss": 2.4094,
      "theoretical_loss": 3.49607904784993,
      "tokens_seen": 1600913408
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003029979194250047,
      "loss": 2.5217,
      "theoretical_loss": 3.496066586633213,
      "tokens_seen": 1600978944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030298136939663325,
      "loss": 2.7992,
      "theoretical_loss": 3.4960541260694056,
      "tokens_seen": 1601044480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030296481936826174,
      "loss": 2.5798,
      "theoretical_loss": 3.496041666158447,
      "tokens_seen": 1601110016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003029482693398903,
      "loss": 2.5423,
      "theoretical_loss": 3.4960292069002765,
      "tokens_seen": 1601175552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003029317193115188,
      "loss": 2.5693,
      "theoretical_loss": 3.496016748294834,
      "tokens_seen": 1601241088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030291516928314735,
      "loss": 2.5675,
      "theoretical_loss": 3.496004290342057,
      "tokens_seen": 1601306624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030289861925477583,
      "loss": 2.733,
      "theoretical_loss": 3.495991833041886,
      "tokens_seen": 1601372160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030288206922640437,
      "loss": 2.6946,
      "theoretical_loss": 3.49597937639426,
      "tokens_seen": 1601437696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003028655191980329,
      "loss": 2.5643,
      "theoretical_loss": 3.495966920399117,
      "tokens_seen": 1601503232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030284896916966145,
      "loss": 2.6058,
      "theoretical_loss": 3.495954465056397,
      "tokens_seen": 1601568768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030283241914128993,
      "loss": 2.5085,
      "theoretical_loss": 3.495942010366039,
      "tokens_seen": 1601634304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030281586911291847,
      "loss": 2.4654,
      "theoretical_loss": 3.4959295563279817,
      "tokens_seen": 1601699840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000302799319084547,
      "loss": 2.6711,
      "theoretical_loss": 3.495917102942165,
      "tokens_seen": 1601765376
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030278276905617554,
      "loss": 2.5263,
      "theoretical_loss": 3.4959046502085274,
      "tokens_seen": 1601830912
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000302766219027804,
      "loss": 2.6488,
      "theoretical_loss": 3.4958921981270086,
      "tokens_seen": 1601896448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030274966899943256,
      "loss": 2.6477,
      "theoretical_loss": 3.4958797466975478,
      "tokens_seen": 1601961984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003027331189710611,
      "loss": 2.7217,
      "theoretical_loss": 3.4958672959200836,
      "tokens_seen": 1602027520
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030271656894268964,
      "loss": 2.2293,
      "theoretical_loss": 3.4958548457945553,
      "tokens_seen": 1602093056
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003027000189143181,
      "loss": 2.5143,
      "theoretical_loss": 3.495842396320903,
      "tokens_seen": 1602158592
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030268346888594666,
      "loss": 2.2643,
      "theoretical_loss": 3.4958299474990646,
      "tokens_seen": 1602224128
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003026669188575752,
      "loss": 2.6072,
      "theoretical_loss": 3.4958174993289797,
      "tokens_seen": 1602289664
    },
    {
      "epoch": 0.57,
      "objective/train/avg_token_score": 0.010142363607883453,
      "objective/train/avg_weight": 0.9880617260932922,
      "objective/train/docs_used": 907881,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6639885902404785,
      "objective/train/std_weight": 0.06037627533078194,
      "objective/train/theoretical_loss": 3.495805051810588,
      "objective/train/tokens_used": 1622815200,
      "theoretical_loss": 3.495805051810588,
      "tokens_seen": 1602355200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030265036882920373,
      "loss": 2.5417,
      "theoretical_loss": 3.495805051810588,
      "tokens_seen": 1602355200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003026338188008322,
      "loss": 2.6271,
      "theoretical_loss": 3.4957926049438286,
      "tokens_seen": 1602420736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030261726877246076,
      "loss": 2.6417,
      "theoretical_loss": 3.495780158728641,
      "tokens_seen": 1602486272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003026007187440893,
      "loss": 2.5176,
      "theoretical_loss": 3.4957677131649634,
      "tokens_seen": 1602551808
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030258416871571783,
      "loss": 2.6602,
      "theoretical_loss": 3.4957552682527355,
      "tokens_seen": 1602617344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003025676186873463,
      "loss": 2.4947,
      "theoretical_loss": 3.4957428239918973,
      "tokens_seen": 1602682880
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003025510686589748,
      "loss": 2.5217,
      "theoretical_loss": 3.4957303803823874,
      "tokens_seen": 1602748416
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030253451863060333,
      "loss": 2.6061,
      "theoretical_loss": 3.4957179374241454,
      "tokens_seen": 1602813952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030251796860223187,
      "loss": 2.7017,
      "theoretical_loss": 3.49570549511711,
      "tokens_seen": 1602879488
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003025014185738604,
      "loss": 2.6918,
      "theoretical_loss": 3.4956930534612214,
      "tokens_seen": 1602945024
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003024848685454889,
      "loss": 2.6473,
      "theoretical_loss": 3.495680612456418,
      "tokens_seen": 1603010560
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030246831851711743,
      "loss": 2.757,
      "theoretical_loss": 3.49566817210264,
      "tokens_seen": 1603076096
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030245176848874597,
      "loss": 2.6079,
      "theoretical_loss": 3.4956557323998254,
      "tokens_seen": 1603141632
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003024352184603745,
      "loss": 2.4806,
      "theoretical_loss": 3.495643293347915,
      "tokens_seen": 1603207168
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000302418668432003,
      "loss": 2.4606,
      "theoretical_loss": 3.4956308549468478,
      "tokens_seen": 1603272704
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003024021184036315,
      "loss": 2.5939,
      "theoretical_loss": 3.495618417196562,
      "tokens_seen": 1603338240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030238556837526006,
      "loss": 2.4826,
      "theoretical_loss": 3.4956059800969985,
      "tokens_seen": 1603403776
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003023690183468886,
      "loss": 2.2915,
      "theoretical_loss": 3.4955935436480954,
      "tokens_seen": 1603469312
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003023524683185171,
      "loss": 2.7424,
      "theoretical_loss": 3.4955811078497927,
      "tokens_seen": 1603534848
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003023359182901456,
      "loss": 2.7238,
      "theoretical_loss": 3.49556867270203,
      "tokens_seen": 1603600384
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030231936826177416,
      "loss": 2.6446,
      "theoretical_loss": 3.4955562382047463,
      "tokens_seen": 1603665920
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003023028182334027,
      "loss": 2.3805,
      "theoretical_loss": 3.4955438043578813,
      "tokens_seen": 1603731456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003022862682050312,
      "loss": 2.4261,
      "theoretical_loss": 3.495531371161374,
      "tokens_seen": 1603796992
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003022697181766597,
      "loss": 2.3368,
      "theoretical_loss": 3.495518938615164,
      "tokens_seen": 1603862528
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030225316814828826,
      "loss": 2.7868,
      "theoretical_loss": 3.4955065067191904,
      "tokens_seen": 1603928064
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003022366181199168,
      "loss": 2.4991,
      "theoretical_loss": 3.4954940754733936,
      "tokens_seen": 1603993600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030222006809154533,
      "loss": 2.5333,
      "theoretical_loss": 3.4954816448777115,
      "tokens_seen": 1604059136
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003022035180631738,
      "loss": 2.6168,
      "theoretical_loss": 3.495469214932085,
      "tokens_seen": 1604124672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030218696803480235,
      "loss": 2.6137,
      "theoretical_loss": 3.4954567856364527,
      "tokens_seen": 1604190208
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003021704180064309,
      "loss": 2.7194,
      "theoretical_loss": 3.4954443569907543,
      "tokens_seen": 1604255744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030215386797805943,
      "loss": 2.7098,
      "theoretical_loss": 3.49543192899493,
      "tokens_seen": 1604321280
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003021373179496879,
      "loss": 2.5606,
      "theoretical_loss": 3.495419501648918,
      "tokens_seen": 1604386816
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030212076792131645,
      "loss": 2.5051,
      "theoretical_loss": 3.495407074952658,
      "tokens_seen": 1604452352
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000302104217892945,
      "loss": 2.5652,
      "theoretical_loss": 3.4953946489060903,
      "tokens_seen": 1604517888
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003020876678645735,
      "loss": 2.566,
      "theoretical_loss": 3.4953822235091536,
      "tokens_seen": 1604583424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000302071117836202,
      "loss": 2.5434,
      "theoretical_loss": 3.495369798761788,
      "tokens_seen": 1604648960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003020545678078305,
      "loss": 2.2209,
      "theoretical_loss": 3.4953573746639326,
      "tokens_seen": 1604714496
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030203801777945903,
      "loss": 2.4459,
      "theoretical_loss": 3.4953449512155275,
      "tokens_seen": 1604780032
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030202146775108757,
      "loss": 2.9416,
      "theoretical_loss": 3.4953325284165113,
      "tokens_seen": 1604845568
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003020049177227161,
      "loss": 2.7237,
      "theoretical_loss": 3.495320106266824,
      "tokens_seen": 1604911104
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003019883676943446,
      "loss": 2.4715,
      "theoretical_loss": 3.4953076847664057,
      "tokens_seen": 1604976640
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003019718176659731,
      "loss": 2.571,
      "theoretical_loss": 3.495295263915195,
      "tokens_seen": 1605042176
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030195526763760166,
      "loss": 2.6572,
      "theoretical_loss": 3.4952828437131322,
      "tokens_seen": 1605107712
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003019387176092302,
      "loss": 2.3896,
      "theoretical_loss": 3.4952704241601564,
      "tokens_seen": 1605173248
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003019221675808587,
      "loss": 2.5254,
      "theoretical_loss": 3.4952580052562077,
      "tokens_seen": 1605238784
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003019056175524872,
      "loss": 2.8131,
      "theoretical_loss": 3.4952455870012247,
      "tokens_seen": 1605304320
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030188906752411576,
      "loss": 2.6164,
      "theoretical_loss": 3.495233169395148,
      "tokens_seen": 1605369856
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003018725174957443,
      "loss": 2.6253,
      "theoretical_loss": 3.4952207524379166,
      "tokens_seen": 1605435392
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003018559674673728,
      "loss": 2.6465,
      "theoretical_loss": 3.495208336129471,
      "tokens_seen": 1605500928
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003018394174390013,
      "loss": 2.5081,
      "theoretical_loss": 3.4951959204697496,
      "tokens_seen": 1605566464
    },
    {
      "epoch": 0.57,
      "objective/train/avg_token_score": 0.010045407339930534,
      "objective/train/avg_weight": 0.9881762266159058,
      "objective/train/docs_used": 909847,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1227049827575684,
      "objective/train/std_weight": 0.058170385658741,
      "objective/train/theoretical_loss": 3.4951835054586926,
      "objective/train/tokens_used": 1626092000,
      "theoretical_loss": 3.4951835054586926,
      "tokens_seen": 1605632000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030182286741062985,
      "loss": 2.7839,
      "theoretical_loss": 3.4951835054586926,
      "tokens_seen": 1605632000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003018063173822584,
      "loss": 2.6016,
      "theoretical_loss": 3.49517109109624,
      "tokens_seen": 1605697536
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003017897673538869,
      "loss": 2.7222,
      "theoretical_loss": 3.4951586773823307,
      "tokens_seen": 1605763072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003017732173255154,
      "loss": 2.448,
      "theoretical_loss": 3.4951462643169053,
      "tokens_seen": 1605828608
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030175666729714395,
      "loss": 2.6834,
      "theoretical_loss": 3.4951338518999027,
      "tokens_seen": 1605894144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003017401172687725,
      "loss": 2.5473,
      "theoretical_loss": 3.4951214401312622,
      "tokens_seen": 1605959680
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030172356724040097,
      "loss": 2.6529,
      "theoretical_loss": 3.4951090290109246,
      "tokens_seen": 1606025216
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003017070172120295,
      "loss": 2.1883,
      "theoretical_loss": 3.495096618538829,
      "tokens_seen": 1606090752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030169046718365804,
      "loss": 2.3416,
      "theoretical_loss": 3.495084208714915,
      "tokens_seen": 1606156288
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003016739171552866,
      "loss": 2.8201,
      "theoretical_loss": 3.4950717995391223,
      "tokens_seen": 1606221824
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030165736712691507,
      "loss": 2.604,
      "theoretical_loss": 3.4950593910113907,
      "tokens_seen": 1606287360
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003016408170985436,
      "loss": 2.5447,
      "theoretical_loss": 3.49504698313166,
      "tokens_seen": 1606352896
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030162426707017214,
      "loss": 2.5163,
      "theoretical_loss": 3.4950345758998704,
      "tokens_seen": 1606418432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003016077170418007,
      "loss": 2.2867,
      "theoretical_loss": 3.4950221693159604,
      "tokens_seen": 1606483968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030159116701342916,
      "loss": 2.8723,
      "theoretical_loss": 3.495009763379871,
      "tokens_seen": 1606549504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030157461698505765,
      "loss": 2.4694,
      "theoretical_loss": 3.494997358091541,
      "tokens_seen": 1606615040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003015580669566862,
      "loss": 2.3959,
      "theoretical_loss": 3.4949849534509108,
      "tokens_seen": 1606680576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003015415169283147,
      "loss": 2.552,
      "theoretical_loss": 3.49497254945792,
      "tokens_seen": 1606746112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030152496689994326,
      "loss": 2.7021,
      "theoretical_loss": 3.4949601461125077,
      "tokens_seen": 1606811648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030150841687157174,
      "loss": 2.4687,
      "theoretical_loss": 3.494947743414615,
      "tokens_seen": 1606877184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003014918668432003,
      "loss": 2.6558,
      "theoretical_loss": 3.4949353413641804,
      "tokens_seen": 1606942720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003014753168148288,
      "loss": 2.6389,
      "theoretical_loss": 3.4949229399611443,
      "tokens_seen": 1607008256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030145876678645735,
      "loss": 2.5988,
      "theoretical_loss": 3.494910539205447,
      "tokens_seen": 1607073792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030144221675808584,
      "loss": 2.4577,
      "theoretical_loss": 3.4948981390970273,
      "tokens_seen": 1607139328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003014256667297144,
      "loss": 2.7659,
      "theoretical_loss": 3.4948857396358255,
      "tokens_seen": 1607204864
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003014091167013429,
      "loss": 2.4722,
      "theoretical_loss": 3.4948733408217816,
      "tokens_seen": 1607270400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030139256667297145,
      "loss": 2.6354,
      "theoretical_loss": 3.494860942654835,
      "tokens_seen": 1607335936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030137601664459993,
      "loss": 2.6417,
      "theoretical_loss": 3.4948485451349263,
      "tokens_seen": 1607401472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030135946661622847,
      "loss": 2.4023,
      "theoretical_loss": 3.494836148261995,
      "tokens_seen": 1607467008
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000301342916587857,
      "loss": 2.7304,
      "theoretical_loss": 3.4948237520359804,
      "tokens_seen": 1607532544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030132636655948555,
      "loss": 2.4828,
      "theoretical_loss": 3.4948113564568226,
      "tokens_seen": 1607598080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030130981653111403,
      "loss": 2.561,
      "theoretical_loss": 3.494798961524462,
      "tokens_seen": 1607663616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030129326650274257,
      "loss": 2.5883,
      "theoretical_loss": 3.4947865672388376,
      "tokens_seen": 1607729152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003012767164743711,
      "loss": 2.4451,
      "theoretical_loss": 3.4947741735998905,
      "tokens_seen": 1607794688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030126016644599964,
      "loss": 2.5995,
      "theoretical_loss": 3.49476178060756,
      "tokens_seen": 1607860224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003012436164176281,
      "loss": 2.7529,
      "theoretical_loss": 3.494749388261786,
      "tokens_seen": 1607925760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030122706638925666,
      "loss": 2.5814,
      "theoretical_loss": 3.4947369965625077,
      "tokens_seen": 1607991296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003012105163608852,
      "loss": 2.8425,
      "theoretical_loss": 3.494724605509666,
      "tokens_seen": 1608056832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030119396633251374,
      "loss": 2.5802,
      "theoretical_loss": 3.494712215103201,
      "tokens_seen": 1608122368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003011774163041422,
      "loss": 2.7742,
      "theoretical_loss": 3.494699825343052,
      "tokens_seen": 1608187904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030116086627577076,
      "loss": 2.6121,
      "theoretical_loss": 3.494687436229159,
      "tokens_seen": 1608253440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003011443162473993,
      "loss": 2.6602,
      "theoretical_loss": 3.4946750477614623,
      "tokens_seen": 1608318976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030112776621902783,
      "loss": 2.6929,
      "theoretical_loss": 3.494662659939901,
      "tokens_seen": 1608384512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003011112161906563,
      "loss": 2.6383,
      "theoretical_loss": 3.4946502727644164,
      "tokens_seen": 1608450048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030109466616228485,
      "loss": 2.6512,
      "theoretical_loss": 3.4946378862349476,
      "tokens_seen": 1608515584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030107811613391334,
      "loss": 2.432,
      "theoretical_loss": 3.494625500351435,
      "tokens_seen": 1608581120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003010615661055419,
      "loss": 2.769,
      "theoretical_loss": 3.4946131151138182,
      "tokens_seen": 1608646656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003010450160771704,
      "loss": 2.7604,
      "theoretical_loss": 3.4946007305220377,
      "tokens_seen": 1608712192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003010284660487989,
      "loss": 2.3307,
      "theoretical_loss": 3.494588346576033,
      "tokens_seen": 1608777728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030101191602042743,
      "loss": 2.6209,
      "theoretical_loss": 3.4945759632757443,
      "tokens_seen": 1608843264
    },
    {
      "epoch": 0.57,
      "objective/train/avg_token_score": 0.011269571259617805,
      "objective/train/avg_weight": 0.9867509007453918,
      "objective/train/docs_used": 911828,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.136337995529175,
      "objective/train/std_weight": 0.058813486248254776,
      "objective/train/theoretical_loss": 3.4945635806211115,
      "objective/train/tokens_used": 1629368800,
      "theoretical_loss": 3.4945635806211115,
      "tokens_seen": 1608908800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030099536599205597,
      "loss": 2.7239,
      "theoretical_loss": 3.4945635806211115,
      "tokens_seen": 1608908800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003009788159636845,
      "loss": 2.6944,
      "theoretical_loss": 3.494551198612075,
      "tokens_seen": 1608974336
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030096226593531305,
      "loss": 2.4976,
      "theoretical_loss": 3.494538817248575,
      "tokens_seen": 1609039872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030094571590694153,
      "loss": 2.7104,
      "theoretical_loss": 3.494526436530551,
      "tokens_seen": 1609105408
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030092916587857007,
      "loss": 2.2273,
      "theoretical_loss": 3.494514056457943,
      "tokens_seen": 1609170944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003009126158501986,
      "loss": 2.7109,
      "theoretical_loss": 3.4945016770306916,
      "tokens_seen": 1609236480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030089606582182714,
      "loss": 2.4938,
      "theoretical_loss": 3.4944892982487366,
      "tokens_seen": 1609302016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003008795157934556,
      "loss": 2.4989,
      "theoretical_loss": 3.4944769201120183,
      "tokens_seen": 1609367552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030086296576508416,
      "loss": 2.7423,
      "theoretical_loss": 3.4944645426204763,
      "tokens_seen": 1609433088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003008464157367127,
      "loss": 2.6549,
      "theoretical_loss": 3.4944521657740513,
      "tokens_seen": 1609498624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030082986570834124,
      "loss": 2.7536,
      "theoretical_loss": 3.494439789572683,
      "tokens_seen": 1609564160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003008133156799697,
      "loss": 2.3705,
      "theoretical_loss": 3.494427414016312,
      "tokens_seen": 1609629696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030079676565159826,
      "loss": 2.6617,
      "theoretical_loss": 3.494415039104877,
      "tokens_seen": 1609695232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003007802156232268,
      "loss": 2.7431,
      "theoretical_loss": 3.4944026648383204,
      "tokens_seen": 1609760768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030076366559485533,
      "loss": 2.5394,
      "theoretical_loss": 3.4943902912165807,
      "tokens_seen": 1609826304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0003007471155664838,
      "loss": 2.5539,
      "theoretical_loss": 3.494377918239598,
      "tokens_seen": 1609891840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00030073056553811236,
      "loss": 2.7307,
      "theoretical_loss": 3.4943655459073133,
      "tokens_seen": 1609957376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003007140155097409,
      "loss": 2.6225,
      "theoretical_loss": 3.4943531742196665,
      "tokens_seen": 1610022912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030069746548136943,
      "loss": 2.4856,
      "theoretical_loss": 3.4943408031765983,
      "tokens_seen": 1610088448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003006809154529979,
      "loss": 2.6989,
      "theoretical_loss": 3.4943284327780475,
      "tokens_seen": 1610153984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030066436542462645,
      "loss": 2.7479,
      "theoretical_loss": 3.4943160630239554,
      "tokens_seen": 1610219520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000300647815396255,
      "loss": 2.5977,
      "theoretical_loss": 3.4943036939142615,
      "tokens_seen": 1610285056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003006312653678835,
      "loss": 2.621,
      "theoretical_loss": 3.4942913254489065,
      "tokens_seen": 1610350592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000300614715339512,
      "loss": 2.6999,
      "theoretical_loss": 3.494278957627831,
      "tokens_seen": 1610416128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030059816531114055,
      "loss": 2.7595,
      "theoretical_loss": 3.494266590450974,
      "tokens_seen": 1610481664
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030058161528276903,
      "loss": 2.7632,
      "theoretical_loss": 3.494254223918277,
      "tokens_seen": 1610547200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030056506525439757,
      "loss": 2.4485,
      "theoretical_loss": 3.4942418580296795,
      "tokens_seen": 1610612736
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003005485152260261,
      "loss": 2.7055,
      "theoretical_loss": 3.494229492785122,
      "tokens_seen": 1610678272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003005319651976546,
      "loss": 2.4891,
      "theoretical_loss": 3.4942171281845447,
      "tokens_seen": 1610743808
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003005154151692831,
      "loss": 2.5435,
      "theoretical_loss": 3.494204764227887,
      "tokens_seen": 1610809344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030049886514091166,
      "loss": 2.4279,
      "theoretical_loss": 3.4941924009150913,
      "tokens_seen": 1610874880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003004823151125402,
      "loss": 2.9372,
      "theoretical_loss": 3.494180038246096,
      "tokens_seen": 1610940416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003004657650841687,
      "loss": 2.3711,
      "theoretical_loss": 3.494167676220842,
      "tokens_seen": 1611005952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003004492150557972,
      "loss": 2.6681,
      "theoretical_loss": 3.4941553148392694,
      "tokens_seen": 1611071488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030043266502742576,
      "loss": 2.5329,
      "theoretical_loss": 3.494142954101319,
      "tokens_seen": 1611137024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003004161149990543,
      "loss": 2.4986,
      "theoretical_loss": 3.4941305940069305,
      "tokens_seen": 1611202560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003003995649706828,
      "loss": 2.756,
      "theoretical_loss": 3.494118234556044,
      "tokens_seen": 1611268096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003003830149423113,
      "loss": 2.6663,
      "theoretical_loss": 3.4941058757486014,
      "tokens_seen": 1611333632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030036646491393986,
      "loss": 2.5259,
      "theoretical_loss": 3.494093517584542,
      "tokens_seen": 1611399168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003003499148855684,
      "loss": 2.9756,
      "theoretical_loss": 3.4940811600638053,
      "tokens_seen": 1611464704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003003333648571969,
      "loss": 2.576,
      "theoretical_loss": 3.4940688031863325,
      "tokens_seen": 1611530240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003003168148288254,
      "loss": 2.5082,
      "theoretical_loss": 3.4940564469520643,
      "tokens_seen": 1611595776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030030026480045395,
      "loss": 2.8287,
      "theoretical_loss": 3.4940440913609407,
      "tokens_seen": 1611661312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003002837147720825,
      "loss": 2.7244,
      "theoretical_loss": 3.494031736412902,
      "tokens_seen": 1611726848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030026716474371097,
      "loss": 2.7102,
      "theoretical_loss": 3.494019382107888,
      "tokens_seen": 1611792384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003002506147153395,
      "loss": 2.7814,
      "theoretical_loss": 3.49400702844584,
      "tokens_seen": 1611857920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030023406468696805,
      "loss": 2.7193,
      "theoretical_loss": 3.4939946754266984,
      "tokens_seen": 1611923456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003002175146585966,
      "loss": 2.2644,
      "theoretical_loss": 3.4939823230504037,
      "tokens_seen": 1611988992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030020096463022507,
      "loss": 2.6699,
      "theoretical_loss": 3.4939699713168952,
      "tokens_seen": 1612054528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003001844146018536,
      "loss": 2.6458,
      "theoretical_loss": 3.493957620226115,
      "tokens_seen": 1612120064
    },
    {
      "epoch": 0.58,
      "objective/train/avg_token_score": 0.016005031764507294,
      "objective/train/avg_weight": 0.9811685681343079,
      "objective/train/docs_used": 913095,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8946657180786133,
      "objective/train/std_weight": 0.0756552442908287,
      "objective/train/theoretical_loss": 3.493945269778002,
      "objective/train/tokens_used": 1632645600,
      "theoretical_loss": 3.493945269778002,
      "tokens_seen": 1612185600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030016786457348214,
      "loss": 2.7489,
      "theoretical_loss": 3.493945269778002,
      "tokens_seen": 1612185600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003001513145451107,
      "loss": 2.3205,
      "theoretical_loss": 3.4939329199724964,
      "tokens_seen": 1612251136
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030013476451673916,
      "loss": 2.6985,
      "theoretical_loss": 3.4939205708095407,
      "tokens_seen": 1612316672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003001182144883677,
      "loss": 2.5786,
      "theoretical_loss": 3.4939082222890736,
      "tokens_seen": 1612382208
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003001016644599962,
      "loss": 2.588,
      "theoretical_loss": 3.493895874411036,
      "tokens_seen": 1612447744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003000851144316247,
      "loss": 2.8064,
      "theoretical_loss": 3.4938835271753685,
      "tokens_seen": 1612513280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030006856440325326,
      "loss": 2.7337,
      "theoretical_loss": 3.493871180582012,
      "tokens_seen": 1612578816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030005201437488174,
      "loss": 2.4979,
      "theoretical_loss": 3.493858834630906,
      "tokens_seen": 1612644352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003000354643465103,
      "loss": 2.3283,
      "theoretical_loss": 3.4938464893219914,
      "tokens_seen": 1612709888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003000189143181388,
      "loss": 2.4857,
      "theoretical_loss": 3.4938341446552093,
      "tokens_seen": 1612775424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00030000236428976736,
      "loss": 2.642,
      "theoretical_loss": 3.4938218006304993,
      "tokens_seen": 1612840960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029998581426139584,
      "loss": 2.4412,
      "theoretical_loss": 3.493809457247803,
      "tokens_seen": 1612906496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002999692642330244,
      "loss": 2.6938,
      "theoretical_loss": 3.4937971145070597,
      "tokens_seen": 1612972032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002999527142046529,
      "loss": 2.8711,
      "theoretical_loss": 3.4937847724082105,
      "tokens_seen": 1613037568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029993616417628145,
      "loss": 2.5572,
      "theoretical_loss": 3.493772430951196,
      "tokens_seen": 1613103104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029991961414790994,
      "loss": 2.5337,
      "theoretical_loss": 3.4937600901359565,
      "tokens_seen": 1613168640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002999030641195385,
      "loss": 2.378,
      "theoretical_loss": 3.493747749962433,
      "tokens_seen": 1613234176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000299886514091167,
      "loss": 2.6844,
      "theoretical_loss": 3.4937354104305656,
      "tokens_seen": 1613299712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029986996406279555,
      "loss": 2.7888,
      "theoretical_loss": 3.4937230715402956,
      "tokens_seen": 1613365248
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029985341403442403,
      "loss": 2.66,
      "theoretical_loss": 3.4937107332915622,
      "tokens_seen": 1613430784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029983686400605257,
      "loss": 2.5796,
      "theoretical_loss": 3.4936983956843073,
      "tokens_seen": 1613496320
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002998203139776811,
      "loss": 2.5092,
      "theoretical_loss": 3.4936860587184713,
      "tokens_seen": 1613561856
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029980376394930964,
      "loss": 2.5357,
      "theoretical_loss": 3.4936737223939938,
      "tokens_seen": 1613627392
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029978721392093813,
      "loss": 2.823,
      "theoretical_loss": 3.4936613867108166,
      "tokens_seen": 1613692928
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029977066389256667,
      "loss": 2.7878,
      "theoretical_loss": 3.4936490516688803,
      "tokens_seen": 1613758464
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002997541138641952,
      "loss": 2.827,
      "theoretical_loss": 3.4936367172681244,
      "tokens_seen": 1613824000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029973756383582374,
      "loss": 2.9347,
      "theoretical_loss": 3.4936243835084904,
      "tokens_seen": 1613889536
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002997210138074522,
      "loss": 2.5222,
      "theoretical_loss": 3.493612050389919,
      "tokens_seen": 1613955072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029970446377908076,
      "loss": 2.4996,
      "theoretical_loss": 3.4935997179123506,
      "tokens_seen": 1614020608
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002996879137507093,
      "loss": 2.5509,
      "theoretical_loss": 3.493587386075726,
      "tokens_seen": 1614086144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029967136372233784,
      "loss": 2.6651,
      "theoretical_loss": 3.493575054879985,
      "tokens_seen": 1614151680
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002996548136939663,
      "loss": 2.6742,
      "theoretical_loss": 3.49356272432507,
      "tokens_seen": 1614217216
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029963826366559486,
      "loss": 2.4725,
      "theoretical_loss": 3.49355039441092,
      "tokens_seen": 1614282752
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002996217136372234,
      "loss": 2.5789,
      "theoretical_loss": 3.493538065137477,
      "tokens_seen": 1614348288
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002996051636088519,
      "loss": 2.8776,
      "theoretical_loss": 3.4935257365046803,
      "tokens_seen": 1614413824
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002995886135804804,
      "loss": 2.8205,
      "theoretical_loss": 3.493513408512472,
      "tokens_seen": 1614479360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029957206355210895,
      "loss": 2.8143,
      "theoretical_loss": 3.493501081160792,
      "tokens_seen": 1614544896
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029955551352373744,
      "loss": 2.6136,
      "theoretical_loss": 3.493488754449581,
      "tokens_seen": 1614610432
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000299538963495366,
      "loss": 2.415,
      "theoretical_loss": 3.4934764283787807,
      "tokens_seen": 1614675968
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002995224134669945,
      "loss": 2.728,
      "theoretical_loss": 3.493464102948331,
      "tokens_seen": 1614741504
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029950586343862305,
      "loss": 2.5473,
      "theoretical_loss": 3.493451778158172,
      "tokens_seen": 1614807040
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029948931341025153,
      "loss": 2.6787,
      "theoretical_loss": 3.493439454008246,
      "tokens_seen": 1614872576
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029947276338188007,
      "loss": 2.6063,
      "theoretical_loss": 3.4934271304984925,
      "tokens_seen": 1614938112
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002994562133535086,
      "loss": 2.5651,
      "theoretical_loss": 3.493414807628853,
      "tokens_seen": 1615003648
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029943966332513715,
      "loss": 2.3171,
      "theoretical_loss": 3.493402485399268,
      "tokens_seen": 1615069184
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029942311329676563,
      "loss": 2.6045,
      "theoretical_loss": 3.493390163809678,
      "tokens_seen": 1615134720
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029940656326839417,
      "loss": 2.5264,
      "theoretical_loss": 3.493377842860024,
      "tokens_seen": 1615200256
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002993900132400227,
      "loss": 2.7298,
      "theoretical_loss": 3.4933655225502473,
      "tokens_seen": 1615265792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029937346321165124,
      "loss": 2.6552,
      "theoretical_loss": 3.4933532028802885,
      "tokens_seen": 1615331328
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002993569131832797,
      "loss": 2.6252,
      "theoretical_loss": 3.493340883850088,
      "tokens_seen": 1615396864
    },
    {
      "epoch": 0.58,
      "objective/train/avg_token_score": 0.033233772963285446,
      "objective/train/avg_weight": 0.9608977437019348,
      "objective/train/docs_used": 915136,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.030107498168945,
      "objective/train/std_weight": 0.1508164405822754,
      "objective/train/theoretical_loss": 3.4933285654595867,
      "objective/train/tokens_used": 1635922400,
      "theoretical_loss": 3.4933285654595867,
      "tokens_seen": 1615462400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029934036315490826,
      "loss": 2.6264,
      "theoretical_loss": 3.4933285654595867,
      "tokens_seen": 1615462400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002993238131265368,
      "loss": 2.5195,
      "theoretical_loss": 3.4933162477087256,
      "tokens_seen": 1615527936
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029930726309816534,
      "loss": 2.6308,
      "theoretical_loss": 3.493303930597446,
      "tokens_seen": 1615593472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002992907130697938,
      "loss": 2.5878,
      "theoretical_loss": 3.493291614125688,
      "tokens_seen": 1615659008
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029927416304142236,
      "loss": 2.3872,
      "theoretical_loss": 3.493279298293392,
      "tokens_seen": 1615724544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002992576130130509,
      "loss": 2.5369,
      "theoretical_loss": 3.4932669831005008,
      "tokens_seen": 1615790080
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029924106298467943,
      "loss": 2.7598,
      "theoretical_loss": 3.493254668546953,
      "tokens_seen": 1615855616
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002992245129563079,
      "loss": 2.4589,
      "theoretical_loss": 3.4932423546326907,
      "tokens_seen": 1615921152
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029920796292793645,
      "loss": 2.6535,
      "theoretical_loss": 3.4932300413576556,
      "tokens_seen": 1615986688
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000299191412899565,
      "loss": 2.6325,
      "theoretical_loss": 3.4932177287217865,
      "tokens_seen": 1616052224
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029917486287119353,
      "loss": 2.5903,
      "theoretical_loss": 3.4932054167250257,
      "tokens_seen": 1616117760
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000299158312842822,
      "loss": 2.7795,
      "theoretical_loss": 3.4931931053673146,
      "tokens_seen": 1616183296
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029914176281445055,
      "loss": 2.5402,
      "theoretical_loss": 3.4931807946485924,
      "tokens_seen": 1616248832
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002991252127860791,
      "loss": 2.7669,
      "theoretical_loss": 3.4931684845688014,
      "tokens_seen": 1616314368
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029910866275770757,
      "loss": 2.6094,
      "theoretical_loss": 3.4931561751278823,
      "tokens_seen": 1616379904
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002990921127293361,
      "loss": 2.8869,
      "theoretical_loss": 3.4931438663257754,
      "tokens_seen": 1616445440
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002990755627009646,
      "loss": 2.6897,
      "theoretical_loss": 3.4931315581624225,
      "tokens_seen": 1616510976
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029905901267259313,
      "loss": 2.5233,
      "theoretical_loss": 3.493119250637764,
      "tokens_seen": 1616576512
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029904246264422167,
      "loss": 2.4878,
      "theoretical_loss": 3.493106943751741,
      "tokens_seen": 1616642048
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002990259126158502,
      "loss": 2.6473,
      "theoretical_loss": 3.493094637504295,
      "tokens_seen": 1616707584
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002990093625874787,
      "loss": 2.5227,
      "theoretical_loss": 3.4930823318953657,
      "tokens_seen": 1616773120
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002989928125591072,
      "loss": 2.7283,
      "theoretical_loss": 3.4930700269248955,
      "tokens_seen": 1616838656
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029897626253073576,
      "loss": 2.4471,
      "theoretical_loss": 3.493057722592824,
      "tokens_seen": 1616904192
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002989597125023643,
      "loss": 2.675,
      "theoretical_loss": 3.493045418899094,
      "tokens_seen": 1616969728
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002989431624739928,
      "loss": 2.6515,
      "theoretical_loss": 3.493033115843645,
      "tokens_seen": 1617035264
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002989266124456213,
      "loss": 2.7106,
      "theoretical_loss": 3.4930208134264182,
      "tokens_seen": 1617100800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029891006241724986,
      "loss": 2.5464,
      "theoretical_loss": 3.4930085116473553,
      "tokens_seen": 1617166336
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002988935123888784,
      "loss": 2.2606,
      "theoretical_loss": 3.492996210506397,
      "tokens_seen": 1617231872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002988769623605069,
      "loss": 2.5673,
      "theoretical_loss": 3.4929839100034847,
      "tokens_seen": 1617297408
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002988604123321354,
      "loss": 2.9227,
      "theoretical_loss": 3.4929716101385586,
      "tokens_seen": 1617362944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029884386230376395,
      "loss": 2.3712,
      "theoretical_loss": 3.4929593109115604,
      "tokens_seen": 1617428480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002988273122753925,
      "loss": 2.6348,
      "theoretical_loss": 3.492947012322431,
      "tokens_seen": 1617494016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000298810762247021,
      "loss": 2.5752,
      "theoretical_loss": 3.4929347143711116,
      "tokens_seen": 1617559552
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002987942122186495,
      "loss": 2.4315,
      "theoretical_loss": 3.492922417057543,
      "tokens_seen": 1617625088
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029877766219027805,
      "loss": 2.7233,
      "theoretical_loss": 3.4929101203816666,
      "tokens_seen": 1617690624
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002987611121619066,
      "loss": 2.5162,
      "theoretical_loss": 3.492897824343423,
      "tokens_seen": 1617756160
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029874456213353507,
      "loss": 2.7232,
      "theoretical_loss": 3.492885528942754,
      "tokens_seen": 1617821696
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002987280121051636,
      "loss": 2.4417,
      "theoretical_loss": 3.4928732341796005,
      "tokens_seen": 1617887232
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029871146207679215,
      "loss": 2.5336,
      "theoretical_loss": 3.4928609400539035,
      "tokens_seen": 1617952768
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002986949120484207,
      "loss": 2.4986,
      "theoretical_loss": 3.4928486465656037,
      "tokens_seen": 1618018304
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029867836202004917,
      "loss": 2.7598,
      "theoretical_loss": 3.492836353714643,
      "tokens_seen": 1618083840
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002986618119916777,
      "loss": 2.6697,
      "theoretical_loss": 3.4928240615009623,
      "tokens_seen": 1618149376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029864526196330624,
      "loss": 2.5509,
      "theoretical_loss": 3.4928117699245025,
      "tokens_seen": 1618214912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002986287119349347,
      "loss": 2.6742,
      "theoretical_loss": 3.492799478985205,
      "tokens_seen": 1618280448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029861216190656326,
      "loss": 2.7946,
      "theoretical_loss": 3.492787188683011,
      "tokens_seen": 1618345984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029859561187819175,
      "loss": 2.5909,
      "theoretical_loss": 3.4927748990178613,
      "tokens_seen": 1618411520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002985790618498203,
      "loss": 2.5409,
      "theoretical_loss": 3.492762609989698,
      "tokens_seen": 1618477056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002985625118214488,
      "loss": 2.7473,
      "theoretical_loss": 3.4927503215984608,
      "tokens_seen": 1618542592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029854596179307736,
      "loss": 2.4131,
      "theoretical_loss": 3.492738033844092,
      "tokens_seen": 1618608128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029852941176470584,
      "loss": 2.6708,
      "theoretical_loss": 3.492725746726533,
      "tokens_seen": 1618673664
    },
    {
      "epoch": 0.58,
      "objective/train/avg_token_score": 0.010029289871454239,
      "objective/train/avg_weight": 0.9881942868232727,
      "objective/train/docs_used": 916739,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.8876163959503174,
      "objective/train/std_weight": 0.07255935668945312,
      "objective/train/theoretical_loss": 3.4927134602457244,
      "objective/train/tokens_used": 1639199200,
      "theoretical_loss": 3.4927134602457244,
      "tokens_seen": 1618739200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002985128617363344,
      "loss": 2.7365,
      "theoretical_loss": 3.4927134602457244,
      "tokens_seen": 1618739200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002984963117079629,
      "loss": 2.44,
      "theoretical_loss": 3.492701174401608,
      "tokens_seen": 1618804736
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029847976167959146,
      "loss": 2.2589,
      "theoretical_loss": 3.492688889194124,
      "tokens_seen": 1618870272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029846321165121994,
      "loss": 2.8107,
      "theoretical_loss": 3.492676604623215,
      "tokens_seen": 1618935808
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002984466616228485,
      "loss": 2.542,
      "theoretical_loss": 3.492664320688821,
      "tokens_seen": 1619001344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000298430111594477,
      "loss": 2.7391,
      "theoretical_loss": 3.492652037390884,
      "tokens_seen": 1619066880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029841356156610555,
      "loss": 2.8646,
      "theoretical_loss": 3.4926397547293444,
      "tokens_seen": 1619132416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029839701153773403,
      "loss": 2.5859,
      "theoretical_loss": 3.492627472704145,
      "tokens_seen": 1619197952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029838046150936257,
      "loss": 2.3856,
      "theoretical_loss": 3.4926151913152257,
      "tokens_seen": 1619263488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002983639114809911,
      "loss": 2.8094,
      "theoretical_loss": 3.4926029105625283,
      "tokens_seen": 1619329024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029834736145261965,
      "loss": 2.7319,
      "theoretical_loss": 3.492590630445995,
      "tokens_seen": 1619394560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029833081142424813,
      "loss": 2.6494,
      "theoretical_loss": 3.492578350965565,
      "tokens_seen": 1619460096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029831426139587667,
      "loss": 2.8151,
      "theoretical_loss": 3.492566072121181,
      "tokens_seen": 1619525632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002982977113675052,
      "loss": 2.697,
      "theoretical_loss": 3.492553793912785,
      "tokens_seen": 1619591168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029828116133913374,
      "loss": 2.5865,
      "theoretical_loss": 3.4925415163403164,
      "tokens_seen": 1619656704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002982646113107622,
      "loss": 2.6877,
      "theoretical_loss": 3.4925292394037184,
      "tokens_seen": 1619722240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029824806128239076,
      "loss": 2.5002,
      "theoretical_loss": 3.492516963102931,
      "tokens_seen": 1619787776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002982315112540193,
      "loss": 2.6322,
      "theoretical_loss": 3.492504687437896,
      "tokens_seen": 1619853312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029821496122564784,
      "loss": 2.4617,
      "theoretical_loss": 3.4924924124085557,
      "tokens_seen": 1619918848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002981984111972764,
      "loss": 2.6132,
      "theoretical_loss": 3.49248013801485,
      "tokens_seen": 1619984384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029818186116890486,
      "loss": 2.8501,
      "theoretical_loss": 3.4924678642567204,
      "tokens_seen": 1620049920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002981653111405334,
      "loss": 2.6767,
      "theoretical_loss": 3.4924555911341093,
      "tokens_seen": 1620115456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029814876111216194,
      "loss": 2.7083,
      "theoretical_loss": 3.4924433186469575,
      "tokens_seen": 1620180992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002981322110837904,
      "loss": 2.3741,
      "theoretical_loss": 3.492431046795206,
      "tokens_seen": 1620246528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029811566105541896,
      "loss": 2.5908,
      "theoretical_loss": 3.4924187755787974,
      "tokens_seen": 1620312064
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029809911102704744,
      "loss": 2.4432,
      "theoretical_loss": 3.492406504997672,
      "tokens_seen": 1620377600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000298082560998676,
      "loss": 2.599,
      "theoretical_loss": 3.4923942350517714,
      "tokens_seen": 1620443136
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002980660109703045,
      "loss": 2.851,
      "theoretical_loss": 3.4923819657410373,
      "tokens_seen": 1620508672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029804946094193305,
      "loss": 2.7106,
      "theoretical_loss": 3.492369697065411,
      "tokens_seen": 1620574208
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029803291091356154,
      "loss": 2.815,
      "theoretical_loss": 3.4923574290248336,
      "tokens_seen": 1620639744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029801636088519007,
      "loss": 2.7267,
      "theoretical_loss": 3.492345161619248,
      "tokens_seen": 1620705280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002979998108568186,
      "loss": 2.7262,
      "theoretical_loss": 3.4923328948485937,
      "tokens_seen": 1620770816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029798326082844715,
      "loss": 2.3518,
      "theoretical_loss": 3.492320628712813,
      "tokens_seen": 1620836352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029796671080007563,
      "loss": 2.5706,
      "theoretical_loss": 3.4923083632118477,
      "tokens_seen": 1620901888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029795016077170417,
      "loss": 2.7649,
      "theoretical_loss": 3.492296098345639,
      "tokens_seen": 1620967424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002979336107433327,
      "loss": 2.6485,
      "theoretical_loss": 3.4922838341141285,
      "tokens_seen": 1621032960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029791706071496124,
      "loss": 2.4843,
      "theoretical_loss": 3.492271570517257,
      "tokens_seen": 1621098496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029790051068658973,
      "loss": 2.5901,
      "theoretical_loss": 3.492259307554967,
      "tokens_seen": 1621164032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029788396065821826,
      "loss": 2.6743,
      "theoretical_loss": 3.4922470452272,
      "tokens_seen": 1621229568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002978674106298468,
      "loss": 2.74,
      "theoretical_loss": 3.4922347835338967,
      "tokens_seen": 1621295104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029785086060147534,
      "loss": 2.3762,
      "theoretical_loss": 3.492222522474999,
      "tokens_seen": 1621360640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002978343105731038,
      "loss": 2.6061,
      "theoretical_loss": 3.492210262050449,
      "tokens_seen": 1621426176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029781776054473236,
      "loss": 2.4534,
      "theoretical_loss": 3.492198002260187,
      "tokens_seen": 1621491712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002978012105163609,
      "loss": 2.5218,
      "theoretical_loss": 3.4921857431041556,
      "tokens_seen": 1621557248
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029778466048798944,
      "loss": 2.7854,
      "theoretical_loss": 3.4921734845822963,
      "tokens_seen": 1621622784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002977681104596179,
      "loss": 2.8095,
      "theoretical_loss": 3.4921612266945496,
      "tokens_seen": 1621688320
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029775156043124646,
      "loss": 2.6283,
      "theoretical_loss": 3.4921489694408585,
      "tokens_seen": 1621753856
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000297735010402875,
      "loss": 2.3593,
      "theoretical_loss": 3.4921367128211642,
      "tokens_seen": 1621819392
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029771846037450353,
      "loss": 2.7799,
      "theoretical_loss": 3.4921244568354077,
      "tokens_seen": 1621884928
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000297701910346132,
      "loss": 2.4977,
      "theoretical_loss": 3.492112201483531,
      "tokens_seen": 1621950464
    },
    {
      "epoch": 0.58,
      "objective/train/avg_token_score": 0.01991870440542698,
      "objective/train/avg_weight": 0.9765791893005371,
      "objective/train/docs_used": 918332,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6183078289031982,
      "objective/train/std_weight": 0.07276356220245361,
      "objective/train/theoretical_loss": 3.4920999467654754,
      "objective/train/tokens_used": 1642476000,
      "theoretical_loss": 3.4920999467654754,
      "tokens_seen": 1622016000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029768536031776055,
      "loss": 2.3314,
      "theoretical_loss": 3.4920999467654754,
      "tokens_seen": 1622016000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002976688102893891,
      "loss": 2.318,
      "theoretical_loss": 3.4920876926811832,
      "tokens_seen": 1622081536
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029765226026101763,
      "loss": 2.5448,
      "theoretical_loss": 3.4920754392305957,
      "tokens_seen": 1622147072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002976357102326461,
      "loss": 2.5794,
      "theoretical_loss": 3.492063186413654,
      "tokens_seen": 1622212608
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002976191602042746,
      "loss": 2.4674,
      "theoretical_loss": 3.4920509342303,
      "tokens_seen": 1622278144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029760261017590313,
      "loss": 2.7256,
      "theoretical_loss": 3.492038682680476,
      "tokens_seen": 1622343680
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029758606014753167,
      "loss": 2.6791,
      "theoretical_loss": 3.492026431764123,
      "tokens_seen": 1622409216
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002975695101191602,
      "loss": 2.4958,
      "theoretical_loss": 3.4920141814811827,
      "tokens_seen": 1622474752
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002975529600907887,
      "loss": 2.3194,
      "theoretical_loss": 3.492001931831597,
      "tokens_seen": 1622540288
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029753641006241723,
      "loss": 2.7549,
      "theoretical_loss": 3.4919896828153076,
      "tokens_seen": 1622605824
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029751986003404577,
      "loss": 2.7757,
      "theoretical_loss": 3.491977434432256,
      "tokens_seen": 1622671360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002975033100056743,
      "loss": 2.646,
      "theoretical_loss": 3.4919651866823833,
      "tokens_seen": 1622736896
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002974867599773028,
      "loss": 2.4327,
      "theoretical_loss": 3.4919529395656324,
      "tokens_seen": 1622802432
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002974702099489313,
      "loss": 2.8655,
      "theoretical_loss": 3.4919406930819443,
      "tokens_seen": 1622867968
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029745365992055986,
      "loss": 2.7411,
      "theoretical_loss": 3.4919284472312606,
      "tokens_seen": 1622933504
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002974371098921884,
      "loss": 2.7033,
      "theoretical_loss": 3.491916202013524,
      "tokens_seen": 1622999040
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002974205598638169,
      "loss": 2.588,
      "theoretical_loss": 3.4919039574286748,
      "tokens_seen": 1623064576
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002974040098354454,
      "loss": 2.7484,
      "theoretical_loss": 3.4918917134766554,
      "tokens_seen": 1623130112
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029738745980707396,
      "loss": 2.3867,
      "theoretical_loss": 3.491879470157408,
      "tokens_seen": 1623195648
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002973709097787025,
      "loss": 2.5692,
      "theoretical_loss": 3.4918672274708737,
      "tokens_seen": 1623261184
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000297354359750331,
      "loss": 2.7049,
      "theoretical_loss": 3.4918549854169947,
      "tokens_seen": 1623326720
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002973378097219595,
      "loss": 2.4938,
      "theoretical_loss": 3.4918427439957123,
      "tokens_seen": 1623392256
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029732125969358805,
      "loss": 2.7066,
      "theoretical_loss": 3.4918305032069687,
      "tokens_seen": 1623457792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002973047096652166,
      "loss": 2.8277,
      "theoretical_loss": 3.4918182630507055,
      "tokens_seen": 1623523328
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002972881596368451,
      "loss": 2.6176,
      "theoretical_loss": 3.4918060235268644,
      "tokens_seen": 1623588864
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002972716096084736,
      "loss": 2.6529,
      "theoretical_loss": 3.4917937846353873,
      "tokens_seen": 1623654400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029725505958010215,
      "loss": 2.8085,
      "theoretical_loss": 3.491781546376216,
      "tokens_seen": 1623719936
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002972385095517307,
      "loss": 2.6437,
      "theoretical_loss": 3.4917693087492925,
      "tokens_seen": 1623785472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029722195952335917,
      "loss": 2.433,
      "theoretical_loss": 3.491757071754558,
      "tokens_seen": 1623851008
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002972054094949877,
      "loss": 2.4742,
      "theoretical_loss": 3.4917448353919553,
      "tokens_seen": 1623916544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029718885946661625,
      "loss": 2.6506,
      "theoretical_loss": 3.4917325996614257,
      "tokens_seen": 1623982080
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002971723094382448,
      "loss": 2.6575,
      "theoretical_loss": 3.4917203645629105,
      "tokens_seen": 1624047616
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029715575940987327,
      "loss": 2.6612,
      "theoretical_loss": 3.4917081300963524,
      "tokens_seen": 1624113152
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029713920938150175,
      "loss": 2.83,
      "theoretical_loss": 3.4916958962616933,
      "tokens_seen": 1624178688
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002971226593531303,
      "loss": 2.3554,
      "theoretical_loss": 3.4916836630588746,
      "tokens_seen": 1624244224
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002971061093247588,
      "loss": 2.35,
      "theoretical_loss": 3.491671430487838,
      "tokens_seen": 1624309760
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029708955929638736,
      "loss": 2.5626,
      "theoretical_loss": 3.491659198548526,
      "tokens_seen": 1624375296
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029707300926801585,
      "loss": 2.7411,
      "theoretical_loss": 3.49164696724088,
      "tokens_seen": 1624440832
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002970564592396444,
      "loss": 2.6756,
      "theoretical_loss": 3.4916347365648424,
      "tokens_seen": 1624506368
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002970399092112729,
      "loss": 2.4966,
      "theoretical_loss": 3.491622506520354,
      "tokens_seen": 1624571904
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029702335918290146,
      "loss": 2.7289,
      "theoretical_loss": 3.491610277107358,
      "tokens_seen": 1624637440
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029700680915452994,
      "loss": 2.6457,
      "theoretical_loss": 3.4915980483257965,
      "tokens_seen": 1624702976
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002969902591261585,
      "loss": 2.6601,
      "theoretical_loss": 3.49158582017561,
      "tokens_seen": 1624768512
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000296973709097787,
      "loss": 2.6289,
      "theoretical_loss": 3.4915735926567413,
      "tokens_seen": 1624834048
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029695715906941555,
      "loss": 2.505,
      "theoretical_loss": 3.4915613657691322,
      "tokens_seen": 1624899584
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002969406090410441,
      "loss": 2.6924,
      "theoretical_loss": 3.4915491395127245,
      "tokens_seen": 1624965120
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002969240590126726,
      "loss": 2.9473,
      "theoretical_loss": 3.491536913887461,
      "tokens_seen": 1625030656
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002969075089843011,
      "loss": 2.6772,
      "theoretical_loss": 3.491524688893282,
      "tokens_seen": 1625096192
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029689095895592965,
      "loss": 2.6351,
      "theoretical_loss": 3.4915124645301314,
      "tokens_seen": 1625161728
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002968744089275582,
      "loss": 2.3366,
      "theoretical_loss": 3.49150024079795,
      "tokens_seen": 1625227264
    },
    {
      "epoch": 0.58,
      "objective/train/avg_token_score": 0.03361232206225395,
      "objective/train/avg_weight": 0.9604467749595642,
      "objective/train/docs_used": 920327,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.819654941558838,
      "objective/train/std_weight": 0.10831819474697113,
      "objective/train/theoretical_loss": 3.49148801769668,
      "objective/train/tokens_used": 1645752800,
      "theoretical_loss": 3.49148801769668,
      "tokens_seen": 1625292800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029685785889918667,
      "loss": 2.3024,
      "theoretical_loss": 3.49148801769668,
      "tokens_seen": 1625292800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002968413088708152,
      "loss": 2.924,
      "theoretical_loss": 3.4914757952262634,
      "tokens_seen": 1625358336
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029682475884244375,
      "loss": 2.7308,
      "theoretical_loss": 3.4914635733866426,
      "tokens_seen": 1625423872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002968082088140723,
      "loss": 3.2808,
      "theoretical_loss": 3.4914513521777586,
      "tokens_seen": 1625489408
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029679165878570077,
      "loss": 2.657,
      "theoretical_loss": 3.491439131599555,
      "tokens_seen": 1625554944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002967751087573293,
      "loss": 2.6716,
      "theoretical_loss": 3.4914269116519723,
      "tokens_seen": 1625620480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029675855872895784,
      "loss": 2.5745,
      "theoretical_loss": 3.4914146923349536,
      "tokens_seen": 1625686016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002967420087005864,
      "loss": 2.3927,
      "theoretical_loss": 3.49140247364844,
      "tokens_seen": 1625751552
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029672545867221486,
      "loss": 2.5422,
      "theoretical_loss": 3.491390255592374,
      "tokens_seen": 1625817088
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002967089086438434,
      "loss": 2.7045,
      "theoretical_loss": 3.4913780381666983,
      "tokens_seen": 1625882624
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029669235861547194,
      "loss": 2.654,
      "theoretical_loss": 3.491365821371354,
      "tokens_seen": 1625948160
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002966758085871005,
      "loss": 2.8461,
      "theoretical_loss": 3.491353605206284,
      "tokens_seen": 1626013696
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029665925855872896,
      "loss": 2.5421,
      "theoretical_loss": 3.4913413896714296,
      "tokens_seen": 1626079232
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029664270853035744,
      "loss": 2.8169,
      "theoretical_loss": 3.491329174766733,
      "tokens_seen": 1626144768
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000296626158501986,
      "loss": 2.7395,
      "theoretical_loss": 3.491316960492137,
      "tokens_seen": 1626210304
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002966096084736145,
      "loss": 2.4204,
      "theoretical_loss": 3.4913047468475833,
      "tokens_seen": 1626275840
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029659305844524305,
      "loss": 2.7686,
      "theoretical_loss": 3.491292533833014,
      "tokens_seen": 1626341376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029657650841687154,
      "loss": 2.4863,
      "theoretical_loss": 3.4912803214483707,
      "tokens_seen": 1626406912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002965599583885001,
      "loss": 2.658,
      "theoretical_loss": 3.4912681096935962,
      "tokens_seen": 1626472448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002965434083601286,
      "loss": 2.8165,
      "theoretical_loss": 3.491255898568632,
      "tokens_seen": 1626537984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029652685833175715,
      "loss": 2.7912,
      "theoretical_loss": 3.4912436880734212,
      "tokens_seen": 1626603520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029651030830338563,
      "loss": 2.7843,
      "theoretical_loss": 3.4912314782079057,
      "tokens_seen": 1626669056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029649375827501417,
      "loss": 2.6115,
      "theoretical_loss": 3.491219268972027,
      "tokens_seen": 1626734592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002964772082466427,
      "loss": 2.8546,
      "theoretical_loss": 3.4912070603657277,
      "tokens_seen": 1626800128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029646065821827125,
      "loss": 2.9226,
      "theoretical_loss": 3.49119485238895,
      "tokens_seen": 1626865664
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029644410818989973,
      "loss": 2.6453,
      "theoretical_loss": 3.4911826450416363,
      "tokens_seen": 1626931200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029642755816152827,
      "loss": 2.554,
      "theoretical_loss": 3.491170438323728,
      "tokens_seen": 1626996736
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002964110081331568,
      "loss": 2.8111,
      "theoretical_loss": 3.4911582322351684,
      "tokens_seen": 1627062272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029639445810478534,
      "loss": 2.9531,
      "theoretical_loss": 3.4911460267758985,
      "tokens_seen": 1627127808
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002963779080764138,
      "loss": 2.8339,
      "theoretical_loss": 3.491133821945861,
      "tokens_seen": 1627193344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029636135804804236,
      "loss": 2.7417,
      "theoretical_loss": 3.491121617744999,
      "tokens_seen": 1627258880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002963448080196709,
      "loss": 2.5614,
      "theoretical_loss": 3.4911094141732537,
      "tokens_seen": 1627324416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029632825799129944,
      "loss": 2.6524,
      "theoretical_loss": 3.4910972112305676,
      "tokens_seen": 1627389952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002963117079629279,
      "loss": 2.867,
      "theoretical_loss": 3.491085008916883,
      "tokens_seen": 1627455488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029629515793455646,
      "loss": 2.7126,
      "theoretical_loss": 3.4910728072321415,
      "tokens_seen": 1627521024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000296278607906185,
      "loss": 2.5192,
      "theoretical_loss": 3.491060606176287,
      "tokens_seen": 1627586560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029626205787781353,
      "loss": 2.4787,
      "theoretical_loss": 3.49104840574926,
      "tokens_seen": 1627652096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000296245507849442,
      "loss": 2.5961,
      "theoretical_loss": 3.491036205951003,
      "tokens_seen": 1627717632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029622895782107056,
      "loss": 2.5726,
      "theoretical_loss": 3.49102400678146,
      "tokens_seen": 1627783168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002962124077926991,
      "loss": 2.6869,
      "theoretical_loss": 3.4910118082405717,
      "tokens_seen": 1627848704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029619585776432763,
      "loss": 2.5518,
      "theoretical_loss": 3.4909996103282803,
      "tokens_seen": 1627914240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002961793077359561,
      "loss": 2.2718,
      "theoretical_loss": 3.490987413044529,
      "tokens_seen": 1627979776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002961627577075846,
      "loss": 2.9637,
      "theoretical_loss": 3.490975216389259,
      "tokens_seen": 1628045312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029614620767921313,
      "loss": 2.6849,
      "theoretical_loss": 3.4909630203624142,
      "tokens_seen": 1628110848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029612965765084167,
      "loss": 2.6814,
      "theoretical_loss": 3.490950824963936,
      "tokens_seen": 1628176384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002961131076224702,
      "loss": 2.5873,
      "theoretical_loss": 3.4909386301937664,
      "tokens_seen": 1628241920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002960965575940987,
      "loss": 2.6448,
      "theoretical_loss": 3.490926436051848,
      "tokens_seen": 1628307456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029608000756572723,
      "loss": 2.5622,
      "theoretical_loss": 3.4909142425381234,
      "tokens_seen": 1628372992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029606345753735577,
      "loss": 2.5971,
      "theoretical_loss": 3.4909020496525347,
      "tokens_seen": 1628438528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002960469075089843,
      "loss": 2.7224,
      "theoretical_loss": 3.4908898573950244,
      "tokens_seen": 1628504064
    },
    {
      "epoch": 0.58,
      "objective/train/avg_token_score": 0.009216820821166039,
      "objective/train/avg_weight": 0.9891709089279175,
      "objective/train/docs_used": 922281,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.5686891078948975,
      "objective/train/std_weight": 0.07094916701316833,
      "objective/train/theoretical_loss": 3.4908776657655354,
      "objective/train/tokens_used": 1649029600,
      "theoretical_loss": 3.4908776657655354,
      "tokens_seen": 1628569600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002960303574806128,
      "loss": 2.6947,
      "theoretical_loss": 3.4908776657655354,
      "tokens_seen": 1628569600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002960138074522413,
      "loss": 2.3119,
      "theoretical_loss": 3.490865474764009,
      "tokens_seen": 1628635136
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029599725742386986,
      "loss": 2.7153,
      "theoretical_loss": 3.490853284390388,
      "tokens_seen": 1628700672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002959807073954984,
      "loss": 2.6106,
      "theoretical_loss": 3.4908410946446153,
      "tokens_seen": 1628766208
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002959641573671269,
      "loss": 2.4587,
      "theoretical_loss": 3.490828905526633,
      "tokens_seen": 1628831744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002959476073387554,
      "loss": 2.6274,
      "theoretical_loss": 3.490816717036383,
      "tokens_seen": 1628897280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029593105731038396,
      "loss": 2.5331,
      "theoretical_loss": 3.490804529173808,
      "tokens_seen": 1628962816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002959145072820125,
      "loss": 2.245,
      "theoretical_loss": 3.4907923419388514,
      "tokens_seen": 1629028352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000295897957253641,
      "loss": 2.4921,
      "theoretical_loss": 3.4907801553314544,
      "tokens_seen": 1629093888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002958814072252695,
      "loss": 2.292,
      "theoretical_loss": 3.4907679693515594,
      "tokens_seen": 1629159424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029586485719689806,
      "loss": 2.4121,
      "theoretical_loss": 3.4907557839991097,
      "tokens_seen": 1629224960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002958483071685266,
      "loss": 2.4927,
      "theoretical_loss": 3.4907435992740474,
      "tokens_seen": 1629290496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002958317571401551,
      "loss": 2.6829,
      "theoretical_loss": 3.490731415176315,
      "tokens_seen": 1629356032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002958152071117836,
      "loss": 2.5542,
      "theoretical_loss": 3.4907192317058544,
      "tokens_seen": 1629421568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029579865708341215,
      "loss": 2.4481,
      "theoretical_loss": 3.4907070488626086,
      "tokens_seen": 1629487104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002957821070550407,
      "loss": 2.7392,
      "theoretical_loss": 3.4906948666465203,
      "tokens_seen": 1629552640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002957655570266692,
      "loss": 2.5432,
      "theoretical_loss": 3.490682685057532,
      "tokens_seen": 1629618176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002957490069982977,
      "loss": 2.7508,
      "theoretical_loss": 3.490670504095586,
      "tokens_seen": 1629683712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029573245696992625,
      "loss": 2.6899,
      "theoretical_loss": 3.4906583237606243,
      "tokens_seen": 1629749248
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002957159069415548,
      "loss": 2.5406,
      "theoretical_loss": 3.49064614405259,
      "tokens_seen": 1629814784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029569935691318327,
      "loss": 2.7027,
      "theoretical_loss": 3.4906339649714258,
      "tokens_seen": 1629880320
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002956828068848118,
      "loss": 2.7404,
      "theoretical_loss": 3.4906217865170737,
      "tokens_seen": 1629945856
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002956662568564403,
      "loss": 2.6651,
      "theoretical_loss": 3.4906096086894767,
      "tokens_seen": 1630011392
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029564970682806883,
      "loss": 2.4879,
      "theoretical_loss": 3.4905974314885766,
      "tokens_seen": 1630076928
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029563315679969736,
      "loss": 2.5907,
      "theoretical_loss": 3.4905852549143166,
      "tokens_seen": 1630142464
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002956166067713259,
      "loss": 2.8218,
      "theoretical_loss": 3.49057307896664,
      "tokens_seen": 1630208000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002956000567429544,
      "loss": 2.407,
      "theoretical_loss": 3.4905609036454877,
      "tokens_seen": 1630273536
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002955835067145829,
      "loss": 2.7594,
      "theoretical_loss": 3.490548728950803,
      "tokens_seen": 1630339072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029556695668621146,
      "loss": 2.2955,
      "theoretical_loss": 3.4905365548825285,
      "tokens_seen": 1630404608
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029555040665784,
      "loss": 2.5261,
      "theoretical_loss": 3.4905243814406073,
      "tokens_seen": 1630470144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002955338566294685,
      "loss": 2.6496,
      "theoretical_loss": 3.490512208624981,
      "tokens_seen": 1630535680
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000295517306601097,
      "loss": 2.7116,
      "theoretical_loss": 3.4905000364355936,
      "tokens_seen": 1630601216
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029550075657272556,
      "loss": 2.5416,
      "theoretical_loss": 3.490487864872386,
      "tokens_seen": 1630666752
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002954842065443541,
      "loss": 2.7134,
      "theoretical_loss": 3.4904756939353025,
      "tokens_seen": 1630732288
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002954676565159826,
      "loss": 2.5713,
      "theoretical_loss": 3.4904635236242845,
      "tokens_seen": 1630797824
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002954511064876111,
      "loss": 2.818,
      "theoretical_loss": 3.4904513539392754,
      "tokens_seen": 1630863360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029543455645923965,
      "loss": 2.5847,
      "theoretical_loss": 3.490439184880217,
      "tokens_seen": 1630928896
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002954180064308682,
      "loss": 2.6303,
      "theoretical_loss": 3.490427016447053,
      "tokens_seen": 1630994432
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002954014564024967,
      "loss": 2.4748,
      "theoretical_loss": 3.490414848639725,
      "tokens_seen": 1631059968
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002953849063741252,
      "loss": 2.2486,
      "theoretical_loss": 3.490402681458177,
      "tokens_seen": 1631125504
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029536835634575375,
      "loss": 2.3171,
      "theoretical_loss": 3.4903905149023506,
      "tokens_seen": 1631191040
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002953518063173823,
      "loss": 2.5297,
      "theoretical_loss": 3.4903783489721882,
      "tokens_seen": 1631256576
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029533525628901077,
      "loss": 2.207,
      "theoretical_loss": 3.490366183667634,
      "tokens_seen": 1631322112
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002953187062606393,
      "loss": 2.3375,
      "theoretical_loss": 3.4903540189886284,
      "tokens_seen": 1631387648
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029530215623226784,
      "loss": 2.5184,
      "theoretical_loss": 3.4903418549351164,
      "tokens_seen": 1631453184
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002952856062038964,
      "loss": 2.451,
      "theoretical_loss": 3.49032969150704,
      "tokens_seen": 1631518720
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029526905617552487,
      "loss": 2.4815,
      "theoretical_loss": 3.490317528704341,
      "tokens_seen": 1631584256
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002952525061471534,
      "loss": 2.6115,
      "theoretical_loss": 3.4903053665269637,
      "tokens_seen": 1631649792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029523595611878194,
      "loss": 2.4112,
      "theoretical_loss": 3.490293204974849,
      "tokens_seen": 1631715328
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002952194060904105,
      "loss": 2.5883,
      "theoretical_loss": 3.490281044047941,
      "tokens_seen": 1631780864
    },
    {
      "epoch": 0.58,
      "objective/train/avg_token_score": 0.029144512489438057,
      "objective/train/avg_weight": 0.96570885181427,
      "objective/train/docs_used": 923944,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.812119483947754,
      "objective/train/std_weight": 0.12124200165271759,
      "objective/train/theoretical_loss": 3.4902688837461824,
      "objective/train/tokens_used": 1652306400,
      "theoretical_loss": 3.4902688837461824,
      "tokens_seen": 1631846400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029520285606203896,
      "loss": 2.2736,
      "theoretical_loss": 3.4902688837461824,
      "tokens_seen": 1631846400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029518630603366745,
      "loss": 2.3647,
      "theoretical_loss": 3.490256724069515,
      "tokens_seen": 1631911936
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000295169756005296,
      "loss": 2.3502,
      "theoretical_loss": 3.490244565017883,
      "tokens_seen": 1631977472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002951532059769245,
      "loss": 2.6399,
      "theoretical_loss": 3.490232406591227,
      "tokens_seen": 1632043008
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029513665594855306,
      "loss": 2.6036,
      "theoretical_loss": 3.4902202487894924,
      "tokens_seen": 1632108544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029512010592018154,
      "loss": 2.8089,
      "theoretical_loss": 3.49020809161262,
      "tokens_seen": 1632174080
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002951035558918101,
      "loss": 2.7086,
      "theoretical_loss": 3.490195935060554,
      "tokens_seen": 1632239616
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002950870058634386,
      "loss": 2.4584,
      "theoretical_loss": 3.490183779133236,
      "tokens_seen": 1632305152
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029507045583506715,
      "loss": 2.5976,
      "theoretical_loss": 3.4901716238306095,
      "tokens_seen": 1632370688
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029505390580669564,
      "loss": 2.4441,
      "theoretical_loss": 3.490159469152617,
      "tokens_seen": 1632436224
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002950373557783242,
      "loss": 2.7008,
      "theoretical_loss": 3.4901473150992017,
      "tokens_seen": 1632501760
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002950208057499527,
      "loss": 2.5441,
      "theoretical_loss": 3.490135161670306,
      "tokens_seen": 1632567296
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029500425572158125,
      "loss": 2.7296,
      "theoretical_loss": 3.490123008865873,
      "tokens_seen": 1632632832
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029498770569320973,
      "loss": 2.7115,
      "theoretical_loss": 3.490110856685846,
      "tokens_seen": 1632698368
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029497115566483827,
      "loss": 2.49,
      "theoretical_loss": 3.4900987051301664,
      "tokens_seen": 1632763904
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002949546056364668,
      "loss": 2.7277,
      "theoretical_loss": 3.490086554198779,
      "tokens_seen": 1632829440
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029493805560809535,
      "loss": 2.0961,
      "theoretical_loss": 3.490074403891625,
      "tokens_seen": 1632894976
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029492150557972383,
      "loss": 2.6279,
      "theoretical_loss": 3.490062254208649,
      "tokens_seen": 1632960512
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029490495555135237,
      "loss": 2.6396,
      "theoretical_loss": 3.490050105149792,
      "tokens_seen": 1633026048
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002948884055229809,
      "loss": 2.5947,
      "theoretical_loss": 3.490037956714998,
      "tokens_seen": 1633091584
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029487185549460944,
      "loss": 2.4846,
      "theoretical_loss": 3.4900258089042095,
      "tokens_seen": 1633157120
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002948553054662379,
      "loss": 2.3441,
      "theoretical_loss": 3.490013661717369,
      "tokens_seen": 1633222656
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029483875543786646,
      "loss": 2.4338,
      "theoretical_loss": 3.490001515154421,
      "tokens_seen": 1633288192
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000294822205409495,
      "loss": 2.6236,
      "theoretical_loss": 3.4899893692153072,
      "tokens_seen": 1633353728
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029480565538112354,
      "loss": 2.509,
      "theoretical_loss": 3.4899772238999707,
      "tokens_seen": 1633419264
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000294789105352752,
      "loss": 2.7086,
      "theoretical_loss": 3.4899650792083543,
      "tokens_seen": 1633484800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029477255532438056,
      "loss": 2.4014,
      "theoretical_loss": 3.4899529351404013,
      "tokens_seen": 1633550336
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002947560052960091,
      "loss": 2.5861,
      "theoretical_loss": 3.4899407916960543,
      "tokens_seen": 1633615872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029473945526763763,
      "loss": 2.5333,
      "theoretical_loss": 3.4899286488752566,
      "tokens_seen": 1633681408
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002947229052392661,
      "loss": 2.6523,
      "theoretical_loss": 3.489916506677951,
      "tokens_seen": 1633746944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029470635521089465,
      "loss": 2.4646,
      "theoretical_loss": 3.4899043651040804,
      "tokens_seen": 1633812480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029468980518252314,
      "loss": 2.6594,
      "theoretical_loss": 3.4898922241535884,
      "tokens_seen": 1633878016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002946732551541517,
      "loss": 2.2667,
      "theoretical_loss": 3.489880083826417,
      "tokens_seen": 1633943552
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002946567051257802,
      "loss": 2.5499,
      "theoretical_loss": 3.4898679441225093,
      "tokens_seen": 1634009088
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002946401550974087,
      "loss": 2.3974,
      "theoretical_loss": 3.4898558050418096,
      "tokens_seen": 1634074624
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029462360506903723,
      "loss": 2.5277,
      "theoretical_loss": 3.4898436665842594,
      "tokens_seen": 1634140160
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029460705504066577,
      "loss": 2.4165,
      "theoretical_loss": 3.4898315287498023,
      "tokens_seen": 1634205696
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002945905050122943,
      "loss": 2.4185,
      "theoretical_loss": 3.489819391538381,
      "tokens_seen": 1634271232
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002945739549839228,
      "loss": 2.6741,
      "theoretical_loss": 3.4898072549499397,
      "tokens_seen": 1634336768
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029455740495555133,
      "loss": 2.2907,
      "theoretical_loss": 3.4897951189844205,
      "tokens_seen": 1634402304
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029454085492717987,
      "loss": 2.5783,
      "theoretical_loss": 3.489782983641766,
      "tokens_seen": 1634467840
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002945243048988084,
      "loss": 2.5699,
      "theoretical_loss": 3.4897708489219204,
      "tokens_seen": 1634533376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002945077548704369,
      "loss": 2.807,
      "theoretical_loss": 3.489758714824826,
      "tokens_seen": 1634598912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002944912048420654,
      "loss": 2.7465,
      "theoretical_loss": 3.4897465813504263,
      "tokens_seen": 1634664448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029447465481369396,
      "loss": 2.5587,
      "theoretical_loss": 3.489734448498664,
      "tokens_seen": 1634729984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002944581047853225,
      "loss": 2.5786,
      "theoretical_loss": 3.489722316269482,
      "tokens_seen": 1634795520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000294441554756951,
      "loss": 2.6754,
      "theoretical_loss": 3.4897101846628242,
      "tokens_seen": 1634861056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002944250047285795,
      "loss": 2.4273,
      "theoretical_loss": 3.4896980536786333,
      "tokens_seen": 1634926592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029440845470020806,
      "loss": 2.6227,
      "theoretical_loss": 3.489685923316852,
      "tokens_seen": 1634992128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002943919046718366,
      "loss": 2.4184,
      "theoretical_loss": 3.489673793577424,
      "tokens_seen": 1635057664
    },
    {
      "epoch": 0.58,
      "objective/train/avg_token_score": 0.015594095923006535,
      "objective/train/avg_weight": 0.9816436767578125,
      "objective/train/docs_used": 925798,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7319135665893555,
      "objective/train/std_weight": 0.06854657083749771,
      "objective/train/theoretical_loss": 3.489661664460292,
      "objective/train/tokens_used": 1655583200,
      "theoretical_loss": 3.489661664460292,
      "tokens_seen": 1635123200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002943753546434651,
      "loss": 2.3795,
      "theoretical_loss": 3.489661664460292,
      "tokens_seen": 1635123200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002943588046150936,
      "loss": 2.4982,
      "theoretical_loss": 3.4896495359653996,
      "tokens_seen": 1635188736
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029434225458672216,
      "loss": 2.6038,
      "theoretical_loss": 3.4896374080926895,
      "tokens_seen": 1635254272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002943257045583507,
      "loss": 2.4468,
      "theoretical_loss": 3.4896252808421053,
      "tokens_seen": 1635319808
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002943091545299792,
      "loss": 2.548,
      "theoretical_loss": 3.4896131542135898,
      "tokens_seen": 1635385344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002942926045016077,
      "loss": 2.9555,
      "theoretical_loss": 3.4896010282070864,
      "tokens_seen": 1635450880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029427605447323625,
      "loss": 2.5515,
      "theoretical_loss": 3.489588902822538,
      "tokens_seen": 1635516416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002942595044448648,
      "loss": 2.7835,
      "theoretical_loss": 3.4895767780598876,
      "tokens_seen": 1635581952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002942429544164933,
      "loss": 2.5237,
      "theoretical_loss": 3.489564653919079,
      "tokens_seen": 1635647488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002942264043881218,
      "loss": 2.524,
      "theoretical_loss": 3.489552530400055,
      "tokens_seen": 1635713024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029420985435975035,
      "loss": 2.6244,
      "theoretical_loss": 3.489540407502759,
      "tokens_seen": 1635778560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029419330433137883,
      "loss": 2.5434,
      "theoretical_loss": 3.489528285227134,
      "tokens_seen": 1635844096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029417675430300737,
      "loss": 2.796,
      "theoretical_loss": 3.489516163573123,
      "tokens_seen": 1635909632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002941602042746359,
      "loss": 2.4698,
      "theoretical_loss": 3.4895040425406703,
      "tokens_seen": 1635975168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002941436542462644,
      "loss": 2.5927,
      "theoretical_loss": 3.4894919221297176,
      "tokens_seen": 1636040704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002941271042178929,
      "loss": 2.4414,
      "theoretical_loss": 3.4894798023402096,
      "tokens_seen": 1636106240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029411055418952146,
      "loss": 2.6808,
      "theoretical_loss": 3.4894676831720886,
      "tokens_seen": 1636171776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029409400416115,
      "loss": 2.5908,
      "theoretical_loss": 3.489455564625298,
      "tokens_seen": 1636237312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002940774541327785,
      "loss": 2.8434,
      "theoretical_loss": 3.489443446699781,
      "tokens_seen": 1636302848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000294060904104407,
      "loss": 2.4964,
      "theoretical_loss": 3.4894313293954813,
      "tokens_seen": 1636368384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029404435407603556,
      "loss": 2.5381,
      "theoretical_loss": 3.489419212712342,
      "tokens_seen": 1636433920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002940278040476641,
      "loss": 2.478,
      "theoretical_loss": 3.4894070966503064,
      "tokens_seen": 1636499456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002940112540192926,
      "loss": 2.5285,
      "theoretical_loss": 3.4893949812093172,
      "tokens_seen": 1636564992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002939947039909211,
      "loss": 2.4929,
      "theoretical_loss": 3.4893828663893185,
      "tokens_seen": 1636630528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029397815396254966,
      "loss": 2.718,
      "theoretical_loss": 3.4893707521902533,
      "tokens_seen": 1636696064
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002939616039341782,
      "loss": 2.6121,
      "theoretical_loss": 3.4893586386120647,
      "tokens_seen": 1636761600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002939450539058067,
      "loss": 2.7757,
      "theoretical_loss": 3.489346525654696,
      "tokens_seen": 1636827136
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002939285038774352,
      "loss": 2.5864,
      "theoretical_loss": 3.4893344133180912,
      "tokens_seen": 1636892672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029391195384906375,
      "loss": 2.6818,
      "theoretical_loss": 3.4893223016021935,
      "tokens_seen": 1636958208
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002938954038206923,
      "loss": 2.5081,
      "theoretical_loss": 3.4893101905069455,
      "tokens_seen": 1637023744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029387885379232077,
      "loss": 2.6916,
      "theoretical_loss": 3.489298080032291,
      "tokens_seen": 1637089280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002938623037639493,
      "loss": 2.5818,
      "theoretical_loss": 3.4892859701781735,
      "tokens_seen": 1637154816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029384575373557785,
      "loss": 2.6475,
      "theoretical_loss": 3.489273860944536,
      "tokens_seen": 1637220352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002938292037072064,
      "loss": 2.179,
      "theoretical_loss": 3.4892617523313225,
      "tokens_seen": 1637285888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029381265367883487,
      "loss": 2.6185,
      "theoretical_loss": 3.4892496443384755,
      "tokens_seen": 1637351424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002937961036504634,
      "loss": 2.701,
      "theoretical_loss": 3.4892375369659394,
      "tokens_seen": 1637416960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029377955362209194,
      "loss": 2.6307,
      "theoretical_loss": 3.489225430213656,
      "tokens_seen": 1637482496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002937630035937205,
      "loss": 2.3021,
      "theoretical_loss": 3.489213324081571,
      "tokens_seen": 1637548032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029374645356534896,
      "loss": 2.4933,
      "theoretical_loss": 3.489201218569626,
      "tokens_seen": 1637613568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002937299035369775,
      "loss": 2.5216,
      "theoretical_loss": 3.4891891136777646,
      "tokens_seen": 1637679104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000293713353508606,
      "loss": 2.1927,
      "theoretical_loss": 3.489177009405931,
      "tokens_seen": 1637744640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002936968034802345,
      "loss": 2.4224,
      "theoretical_loss": 3.4891649057540683,
      "tokens_seen": 1637810176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029368025345186306,
      "loss": 2.5483,
      "theoretical_loss": 3.4891528027221197,
      "tokens_seen": 1637875712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00029366370342349154,
      "loss": 2.4772,
      "theoretical_loss": 3.489140700310029,
      "tokens_seen": 1637941248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002936471533951201,
      "loss": 2.5767,
      "theoretical_loss": 3.4891285985177394,
      "tokens_seen": 1638006784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002936306033667486,
      "loss": 2.5823,
      "theoretical_loss": 3.4891164973451945,
      "tokens_seen": 1638072320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029361405333837716,
      "loss": 2.6504,
      "theoretical_loss": 3.4891043967923374,
      "tokens_seen": 1638137856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029359750331000564,
      "loss": 2.6367,
      "theoretical_loss": 3.4890922968591123,
      "tokens_seen": 1638203392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002935809532816342,
      "loss": 2.8548,
      "theoretical_loss": 3.489080197545462,
      "tokens_seen": 1638268928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002935644032532627,
      "loss": 2.6807,
      "theoretical_loss": 3.4890680988513303,
      "tokens_seen": 1638334464
    },
    {
      "debugging/Self-BLEU-5": 0.4862537472725578,
      "debugging/distinct-1-grams": 0.7771784690107355,
      "debugging/distinct-2-grams": 0.9668669456652452,
      "debugging/entropy-1-grams": 6.202714981817353,
      "debugging/entropy-2-grams": 7.211883548596637,
      "debugging/length": 471.0,
      "debugging/num_segments": 21,
      "epoch": 0.59,
      "objective/train/avg_token_score": 0.013972360640764236,
      "objective/train/avg_weight": 0.9835533499717712,
      "objective/train/docs_used": 927402,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.951176166534424,
      "objective/train/std_weight": 0.07358939200639725,
      "objective/train/theoretical_loss": 3.4890560007766607,
      "objective/train/tokens_used": 1658860000,
      "theoretical_loss": 3.4890560007766607,
      "tokens_seen": 1638400000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029354785322489125,
      "loss": 2.7015,
      "theoretical_loss": 3.4890560007766607,
      "tokens_seen": 1638400000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029353130319651974,
      "loss": 2.4527,
      "theoretical_loss": 3.489043903321397,
      "tokens_seen": 1638465536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002935147531681483,
      "loss": 2.5682,
      "theoretical_loss": 3.4890318064854817,
      "tokens_seen": 1638531072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002934982031397768,
      "loss": 2.6105,
      "theoretical_loss": 3.4890197102688596,
      "tokens_seen": 1638596608
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029348165311140535,
      "loss": 2.6519,
      "theoretical_loss": 3.489007614671473,
      "tokens_seen": 1638662144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029346510308303383,
      "loss": 2.7225,
      "theoretical_loss": 3.4889955196932663,
      "tokens_seen": 1638727680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029344855305466237,
      "loss": 2.6628,
      "theoretical_loss": 3.488983425334183,
      "tokens_seen": 1638793216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002934320030262909,
      "loss": 2.6745,
      "theoretical_loss": 3.4889713315941666,
      "tokens_seen": 1638858752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029341545299791944,
      "loss": 2.6036,
      "theoretical_loss": 3.4889592384731603,
      "tokens_seen": 1638924288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029339890296954793,
      "loss": 2.593,
      "theoretical_loss": 3.4889471459711077,
      "tokens_seen": 1638989824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029338235294117647,
      "loss": 2.733,
      "theoretical_loss": 3.4889350540879533,
      "tokens_seen": 1639055360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000293365802912805,
      "loss": 2.6457,
      "theoretical_loss": 3.488922962823639,
      "tokens_seen": 1639120896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029334925288443354,
      "loss": 2.5993,
      "theoretical_loss": 3.48891087217811,
      "tokens_seen": 1639186432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000293332702856062,
      "loss": 2.6069,
      "theoretical_loss": 3.488898782151309,
      "tokens_seen": 1639251968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029331615282769056,
      "loss": 2.6449,
      "theoretical_loss": 3.4888866927431796,
      "tokens_seen": 1639317504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002932996027993191,
      "loss": 2.5656,
      "theoretical_loss": 3.4888746039536658,
      "tokens_seen": 1639383040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029328305277094764,
      "loss": 2.6992,
      "theoretical_loss": 3.488862515782711,
      "tokens_seen": 1639448576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002932665027425761,
      "loss": 2.5654,
      "theoretical_loss": 3.4888504282302595,
      "tokens_seen": 1639514112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029324995271420466,
      "loss": 2.6969,
      "theoretical_loss": 3.4888383412962534,
      "tokens_seen": 1639579648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002932334026858332,
      "loss": 2.3946,
      "theoretical_loss": 3.488826254980638,
      "tokens_seen": 1639645184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002932168526574617,
      "loss": 2.5996,
      "theoretical_loss": 3.4888141692833554,
      "tokens_seen": 1639710720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002932003026290902,
      "loss": 2.3653,
      "theoretical_loss": 3.488802084204351,
      "tokens_seen": 1639776256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002931837526007187,
      "loss": 2.6453,
      "theoretical_loss": 3.4887899997435667,
      "tokens_seen": 1639841792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029316720257234724,
      "loss": 2.8028,
      "theoretical_loss": 3.488777915900947,
      "tokens_seen": 1639907328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002931506525439758,
      "loss": 2.435,
      "theoretical_loss": 3.488765832676436,
      "tokens_seen": 1639972864
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002931341025156043,
      "loss": 2.6445,
      "theoretical_loss": 3.488753750069977,
      "tokens_seen": 1640038400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002931175524872328,
      "loss": 2.6301,
      "theoretical_loss": 3.488741668081513,
      "tokens_seen": 1640103936
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029310100245886133,
      "loss": 2.4933,
      "theoretical_loss": 3.4887295867109884,
      "tokens_seen": 1640169472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029308445243048987,
      "loss": 2.374,
      "theoretical_loss": 3.4887175059583475,
      "tokens_seen": 1640235008
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002930679024021184,
      "loss": 2.5007,
      "theoretical_loss": 3.4887054258235333,
      "tokens_seen": 1640300544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002930513523737469,
      "loss": 2.5906,
      "theoretical_loss": 3.488693346306489,
      "tokens_seen": 1640366080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029303480234537543,
      "loss": 2.5779,
      "theoretical_loss": 3.4886812674071592,
      "tokens_seen": 1640431616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029301825231700397,
      "loss": 2.4963,
      "theoretical_loss": 3.4886691891254875,
      "tokens_seen": 1640497152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002930017022886325,
      "loss": 2.2092,
      "theoretical_loss": 3.488657111461417,
      "tokens_seen": 1640562688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029298515226026104,
      "loss": 2.6108,
      "theoretical_loss": 3.4886450344148923,
      "tokens_seen": 1640628224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002929686022318895,
      "loss": 2.7508,
      "theoretical_loss": 3.488632957985857,
      "tokens_seen": 1640693760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029295205220351806,
      "loss": 2.5303,
      "theoretical_loss": 3.488620882174254,
      "tokens_seen": 1640759296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002929355021751466,
      "loss": 2.6771,
      "theoretical_loss": 3.4886088069800283,
      "tokens_seen": 1640824832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029291895214677514,
      "loss": 2.2848,
      "theoretical_loss": 3.488596732403123,
      "tokens_seen": 1640890368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002929024021184036,
      "loss": 2.3948,
      "theoretical_loss": 3.488584658443482,
      "tokens_seen": 1640955904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029288585209003216,
      "loss": 2.4497,
      "theoretical_loss": 3.488572585101049,
      "tokens_seen": 1641021440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002928693020616607,
      "loss": 2.5601,
      "theoretical_loss": 3.4885605123757677,
      "tokens_seen": 1641086976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029285275203328923,
      "loss": 2.8004,
      "theoretical_loss": 3.488548440267582,
      "tokens_seen": 1641152512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002928362020049177,
      "loss": 2.636,
      "theoretical_loss": 3.4885363687764364,
      "tokens_seen": 1641218048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029281965197654625,
      "loss": 2.5468,
      "theoretical_loss": 3.488524297902274,
      "tokens_seen": 1641283584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002928031019481748,
      "loss": 2.6092,
      "theoretical_loss": 3.488512227645039,
      "tokens_seen": 1641349120
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029278655191980333,
      "loss": 2.5093,
      "theoretical_loss": 3.4885001580046744,
      "tokens_seen": 1641414656
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002927700018914318,
      "loss": 2.2832,
      "theoretical_loss": 3.488488088981125,
      "tokens_seen": 1641480192
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029275345186306035,
      "loss": 2.503,
      "theoretical_loss": 3.4884760205743337,
      "tokens_seen": 1641545728
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002927369018346889,
      "loss": 2.6165,
      "theoretical_loss": 3.4884639527842456,
      "tokens_seen": 1641611264
    },
    {
      "epoch": 0.59,
      "objective/train/avg_token_score": 0.010949091985821724,
      "objective/train/avg_weight": 0.9871118068695068,
      "objective/train/docs_used": 928835,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.970010757446289,
      "objective/train/std_weight": 0.06599122285842896,
      "objective/train/theoretical_loss": 3.4884518856108038,
      "objective/train/tokens_used": 1662136800,
      "theoretical_loss": 3.4884518856108038,
      "tokens_seen": 1641676800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029272035180631737,
      "loss": 2.5145,
      "theoretical_loss": 3.4884518856108038,
      "tokens_seen": 1641676800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002927038017779459,
      "loss": 2.5797,
      "theoretical_loss": 3.4884398190539523,
      "tokens_seen": 1641742336
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002926872517495744,
      "loss": 2.5968,
      "theoretical_loss": 3.488427753113635,
      "tokens_seen": 1641807872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029267070172120293,
      "loss": 2.4763,
      "theoretical_loss": 3.488415687789796,
      "tokens_seen": 1641873408
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029265415169283147,
      "loss": 2.6859,
      "theoretical_loss": 3.4884036230823785,
      "tokens_seen": 1641938944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029263760166446,
      "loss": 2.4911,
      "theoretical_loss": 3.4883915589913275,
      "tokens_seen": 1642004480
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002926210516360885,
      "loss": 2.5571,
      "theoretical_loss": 3.488379495516586,
      "tokens_seen": 1642070016
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000292604501607717,
      "loss": 2.3916,
      "theoretical_loss": 3.488367432658098,
      "tokens_seen": 1642135552
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029258795157934556,
      "loss": 2.3337,
      "theoretical_loss": 3.488355370415808,
      "tokens_seen": 1642201088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002925714015509741,
      "loss": 2.9601,
      "theoretical_loss": 3.488343308789659,
      "tokens_seen": 1642266624
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002925548515226026,
      "loss": 2.6921,
      "theoretical_loss": 3.4883312477795965,
      "tokens_seen": 1642332160
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002925383014942311,
      "loss": 2.3664,
      "theoretical_loss": 3.488319187385563,
      "tokens_seen": 1642397696
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029252175146585966,
      "loss": 2.8446,
      "theoretical_loss": 3.488307127607503,
      "tokens_seen": 1642463232
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002925052014374882,
      "loss": 2.7261,
      "theoretical_loss": 3.4882950684453604,
      "tokens_seen": 1642528768
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002924886514091167,
      "loss": 2.5871,
      "theoretical_loss": 3.488283009899079,
      "tokens_seen": 1642594304
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002924721013807452,
      "loss": 2.6547,
      "theoretical_loss": 3.4882709519686026,
      "tokens_seen": 1642659840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029245555135237375,
      "loss": 2.8582,
      "theoretical_loss": 3.488258894653876,
      "tokens_seen": 1642725376
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002924390013240023,
      "loss": 2.5397,
      "theoretical_loss": 3.4882468379548426,
      "tokens_seen": 1642790912
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002924224512956308,
      "loss": 2.4485,
      "theoretical_loss": 3.4882347818714465,
      "tokens_seen": 1642856448
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002924059012672593,
      "loss": 2.6435,
      "theoretical_loss": 3.488222726403632,
      "tokens_seen": 1642921984
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029238935123888785,
      "loss": 2.4726,
      "theoretical_loss": 3.488210671551342,
      "tokens_seen": 1642987520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002923728012105164,
      "loss": 2.3234,
      "theoretical_loss": 3.4881986173145227,
      "tokens_seen": 1643053056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029235625118214487,
      "loss": 2.637,
      "theoretical_loss": 3.488186563693116,
      "tokens_seen": 1643118592
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002923397011537734,
      "loss": 2.3492,
      "theoretical_loss": 3.4881745106870667,
      "tokens_seen": 1643184128
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029232315112540195,
      "loss": 2.6419,
      "theoretical_loss": 3.488162458296319,
      "tokens_seen": 1643249664
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002923066010970305,
      "loss": 2.5861,
      "theoretical_loss": 3.488150406520817,
      "tokens_seen": 1643315200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029229005106865897,
      "loss": 2.5608,
      "theoretical_loss": 3.4881383553605043,
      "tokens_seen": 1643380736
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002922735010402875,
      "loss": 2.7016,
      "theoretical_loss": 3.4881263048153253,
      "tokens_seen": 1643446272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029225695101191604,
      "loss": 2.5561,
      "theoretical_loss": 3.4881142548852244,
      "tokens_seen": 1643511808
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002922404009835446,
      "loss": 2.5282,
      "theoretical_loss": 3.488102205570145,
      "tokens_seen": 1643577344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029222385095517306,
      "loss": 2.6966,
      "theoretical_loss": 3.4880901568700313,
      "tokens_seen": 1643642880
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029220730092680155,
      "loss": 2.6479,
      "theoretical_loss": 3.488078108784828,
      "tokens_seen": 1643708416
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002921907508984301,
      "loss": 2.7002,
      "theoretical_loss": 3.4880660613144787,
      "tokens_seen": 1643773952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002921742008700586,
      "loss": 2.7148,
      "theoretical_loss": 3.4880540144589274,
      "tokens_seen": 1643839488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029215765084168716,
      "loss": 2.4289,
      "theoretical_loss": 3.488041968218118,
      "tokens_seen": 1643905024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029214110081331564,
      "loss": 2.8302,
      "theoretical_loss": 3.4880299225919957,
      "tokens_seen": 1643970560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002921245507849442,
      "loss": 2.473,
      "theoretical_loss": 3.488017877580504,
      "tokens_seen": 1644036096
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002921080007565727,
      "loss": 2.4076,
      "theoretical_loss": 3.4880058331835873,
      "tokens_seen": 1644101632
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029209145072820126,
      "loss": 2.3338,
      "theoretical_loss": 3.4879937894011888,
      "tokens_seen": 1644167168
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029207490069982974,
      "loss": 3.0319,
      "theoretical_loss": 3.4879817462332534,
      "tokens_seen": 1644232704
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002920583506714583,
      "loss": 2.3696,
      "theoretical_loss": 3.4879697036797257,
      "tokens_seen": 1644298240
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002920418006430868,
      "loss": 2.5736,
      "theoretical_loss": 3.4879576617405488,
      "tokens_seen": 1644363776
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029202525061471535,
      "loss": 2.5211,
      "theoretical_loss": 3.4879456204156676,
      "tokens_seen": 1644429312
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029200870058634383,
      "loss": 2.6682,
      "theoretical_loss": 3.487933579705026,
      "tokens_seen": 1644494848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029199215055797237,
      "loss": 2.6006,
      "theoretical_loss": 3.4879215396085685,
      "tokens_seen": 1644560384
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002919756005296009,
      "loss": 2.5372,
      "theoretical_loss": 3.4879095001262392,
      "tokens_seen": 1644625920
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029195905050122945,
      "loss": 2.5177,
      "theoretical_loss": 3.487897461257982,
      "tokens_seen": 1644691456
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029194250047285793,
      "loss": 2.5452,
      "theoretical_loss": 3.487885423003741,
      "tokens_seen": 1644756992
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029192595044448647,
      "loss": 2.5676,
      "theoretical_loss": 3.487873385363461,
      "tokens_seen": 1644822528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000291909400416115,
      "loss": 2.5914,
      "theoretical_loss": 3.487861348337086,
      "tokens_seen": 1644888064
    },
    {
      "epoch": 0.59,
      "objective/train/avg_token_score": 0.012742018327116966,
      "objective/train/avg_weight": 0.9850109815597534,
      "objective/train/docs_used": 930288,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.954169273376465,
      "objective/train/std_weight": 0.07522138208150864,
      "objective/train/theoretical_loss": 3.4878493119245606,
      "objective/train/tokens_used": 1665413600,
      "theoretical_loss": 3.4878493119245606,
      "tokens_seen": 1644953600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029189285038774354,
      "loss": 2.4499,
      "theoretical_loss": 3.4878493119245606,
      "tokens_seen": 1644953600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000291876300359372,
      "loss": 2.8628,
      "theoretical_loss": 3.487837276125828,
      "tokens_seen": 1645019136
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029185975033100056,
      "loss": 2.6451,
      "theoretical_loss": 3.4878252409408335,
      "tokens_seen": 1645084672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002918432003026291,
      "loss": 2.6215,
      "theoretical_loss": 3.4878132063695206,
      "tokens_seen": 1645150208
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029182665027425764,
      "loss": 2.7567,
      "theoretical_loss": 3.4878011724118343,
      "tokens_seen": 1645215744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002918101002458861,
      "loss": 2.5861,
      "theoretical_loss": 3.487789139067718,
      "tokens_seen": 1645281280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029179355021751466,
      "loss": 2.7453,
      "theoretical_loss": 3.487777106337117,
      "tokens_seen": 1645346816
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002917770001891432,
      "loss": 2.391,
      "theoretical_loss": 3.4877650742199746,
      "tokens_seen": 1645412352
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029176045016077174,
      "loss": 2.4213,
      "theoretical_loss": 3.4877530427162355,
      "tokens_seen": 1645477888
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002917439001324002,
      "loss": 2.5237,
      "theoretical_loss": 3.487741011825844,
      "tokens_seen": 1645543424
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029172735010402876,
      "loss": 2.761,
      "theoretical_loss": 3.4877289815487447,
      "tokens_seen": 1645608960
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029171080007565724,
      "loss": 2.5878,
      "theoretical_loss": 3.487716951884882,
      "tokens_seen": 1645674496
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002916942500472858,
      "loss": 2.6182,
      "theoretical_loss": 3.487704922834199,
      "tokens_seen": 1645740032
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002916777000189143,
      "loss": 2.6004,
      "theoretical_loss": 3.487692894396641,
      "tokens_seen": 1645805568
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029166114999054285,
      "loss": 2.4592,
      "theoretical_loss": 3.487680866572153,
      "tokens_seen": 1645871104
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029164459996217134,
      "loss": 2.3222,
      "theoretical_loss": 3.487668839360678,
      "tokens_seen": 1645936640
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029162804993379987,
      "loss": 2.7126,
      "theoretical_loss": 3.487656812762161,
      "tokens_seen": 1646002176
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002916114999054284,
      "loss": 2.6985,
      "theoretical_loss": 3.4876447867765465,
      "tokens_seen": 1646067712
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029159494987705695,
      "loss": 2.6221,
      "theoretical_loss": 3.4876327614037788,
      "tokens_seen": 1646133248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029157839984868543,
      "loss": 2.7215,
      "theoretical_loss": 3.4876207366438017,
      "tokens_seen": 1646198784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029156184982031397,
      "loss": 2.2196,
      "theoretical_loss": 3.4876087124965602,
      "tokens_seen": 1646264320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002915452997919425,
      "loss": 2.7591,
      "theoretical_loss": 3.4875966889619985,
      "tokens_seen": 1646329856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029152874976357104,
      "loss": 2.7448,
      "theoretical_loss": 3.4875846660400605,
      "tokens_seen": 1646395392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029151219973519953,
      "loss": 2.572,
      "theoretical_loss": 3.487572643730692,
      "tokens_seen": 1646460928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029149564970682806,
      "loss": 2.6766,
      "theoretical_loss": 3.4875606220338358,
      "tokens_seen": 1646526464
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002914790996784566,
      "loss": 2.4986,
      "theoretical_loss": 3.487548600949437,
      "tokens_seen": 1646592000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029146254965008514,
      "loss": 2.4921,
      "theoretical_loss": 3.4875365804774408,
      "tokens_seen": 1646657536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002914459996217136,
      "loss": 2.6474,
      "theoretical_loss": 3.48752456061779,
      "tokens_seen": 1646723072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029142944959334216,
      "loss": 2.5841,
      "theoretical_loss": 3.48751254137043,
      "tokens_seen": 1646788608
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002914128995649707,
      "loss": 2.3232,
      "theoretical_loss": 3.4875005227353055,
      "tokens_seen": 1646854144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029139634953659924,
      "loss": 2.7015,
      "theoretical_loss": 3.4874885047123607,
      "tokens_seen": 1646919680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002913797995082277,
      "loss": 2.3121,
      "theoretical_loss": 3.4874764873015396,
      "tokens_seen": 1646985216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029136324947985626,
      "loss": 2.8525,
      "theoretical_loss": 3.487464470502787,
      "tokens_seen": 1647050752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002913466994514848,
      "loss": 2.3238,
      "theoretical_loss": 3.487452454316048,
      "tokens_seen": 1647116288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029133014942311333,
      "loss": 2.2778,
      "theoretical_loss": 3.487440438741266,
      "tokens_seen": 1647181824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002913135993947418,
      "loss": 2.5323,
      "theoretical_loss": 3.4874284237783857,
      "tokens_seen": 1647247360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029129704936637035,
      "loss": 2.8724,
      "theoretical_loss": 3.487416409427352,
      "tokens_seen": 1647312896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002912804993379989,
      "loss": 2.5944,
      "theoretical_loss": 3.4874043956881096,
      "tokens_seen": 1647378432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029126394930962743,
      "loss": 2.5246,
      "theoretical_loss": 3.4873923825606026,
      "tokens_seen": 1647443968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002912473992812559,
      "loss": 2.3132,
      "theoretical_loss": 3.4873803700447756,
      "tokens_seen": 1647509504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002912308492528844,
      "loss": 2.5107,
      "theoretical_loss": 3.487368358140573,
      "tokens_seen": 1647575040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029121429922451293,
      "loss": 2.6149,
      "theoretical_loss": 3.4873563468479394,
      "tokens_seen": 1647640576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029119774919614147,
      "loss": 2.8041,
      "theoretical_loss": 3.4873443361668195,
      "tokens_seen": 1647706112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029118119916777,
      "loss": 2.4349,
      "theoretical_loss": 3.4873323260971576,
      "tokens_seen": 1647771648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002911646491393985,
      "loss": 2.4637,
      "theoretical_loss": 3.4873203166388986,
      "tokens_seen": 1647837184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029114809911102703,
      "loss": 2.5859,
      "theoretical_loss": 3.487308307791987,
      "tokens_seen": 1647902720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029113154908265557,
      "loss": 2.5371,
      "theoretical_loss": 3.487296299556367,
      "tokens_seen": 1647968256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002911149990542841,
      "loss": 2.2905,
      "theoretical_loss": 3.4872842919319833,
      "tokens_seen": 1648033792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002910984490259126,
      "loss": 2.5597,
      "theoretical_loss": 3.4872722849187805,
      "tokens_seen": 1648099328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002910818989975411,
      "loss": 2.5755,
      "theoretical_loss": 3.4872602785167035,
      "tokens_seen": 1648164864
    },
    {
      "epoch": 0.59,
      "objective/train/avg_token_score": 0.016771705821156502,
      "objective/train/avg_weight": 0.9802923798561096,
      "objective/train/docs_used": 932466,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4086556434631348,
      "objective/train/std_weight": 0.06443489342927933,
      "objective/train/theoretical_loss": 3.4872482727256964,
      "objective/train/tokens_used": 1668690400,
      "theoretical_loss": 3.4872482727256964,
      "tokens_seen": 1648230400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029106534896916966,
      "loss": 2.5211,
      "theoretical_loss": 3.4872482727256964,
      "tokens_seen": 1648230400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002910487989407982,
      "loss": 2.8336,
      "theoretical_loss": 3.4872362675457045,
      "tokens_seen": 1648295936
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002910322489124267,
      "loss": 2.7547,
      "theoretical_loss": 3.487224262976672,
      "tokens_seen": 1648361472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002910156988840552,
      "loss": 2.4032,
      "theoretical_loss": 3.487212259018543,
      "tokens_seen": 1648427008
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029099914885568376,
      "loss": 2.6969,
      "theoretical_loss": 3.487200255671263,
      "tokens_seen": 1648492544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002909825988273123,
      "loss": 2.4706,
      "theoretical_loss": 3.487188252934776,
      "tokens_seen": 1648558080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002909660487989408,
      "loss": 2.8691,
      "theoretical_loss": 3.4871762508090267,
      "tokens_seen": 1648623616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002909494987705693,
      "loss": 2.539,
      "theoretical_loss": 3.4871642492939605,
      "tokens_seen": 1648689152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029093294874219785,
      "loss": 2.7843,
      "theoretical_loss": 3.4871522483895214,
      "tokens_seen": 1648754688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002909163987138264,
      "loss": 2.6162,
      "theoretical_loss": 3.4871402480956535,
      "tokens_seen": 1648820224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002908998486854549,
      "loss": 2.3566,
      "theoretical_loss": 3.487128248412303,
      "tokens_seen": 1648885760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002908832986570834,
      "loss": 2.6671,
      "theoretical_loss": 3.4871162493394134,
      "tokens_seen": 1648951296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029086674862871195,
      "loss": 2.4968,
      "theoretical_loss": 3.4871042508769294,
      "tokens_seen": 1649016832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002908501986003405,
      "loss": 2.7161,
      "theoretical_loss": 3.4870922530247963,
      "tokens_seen": 1649082368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029083364857196897,
      "loss": 2.4348,
      "theoretical_loss": 3.4870802557829585,
      "tokens_seen": 1649147904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002908170985435975,
      "loss": 2.8312,
      "theoretical_loss": 3.4870682591513606,
      "tokens_seen": 1649213440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029080054851522605,
      "loss": 2.5905,
      "theoretical_loss": 3.4870562631299475,
      "tokens_seen": 1649278976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002907839984868546,
      "loss": 2.8254,
      "theoretical_loss": 3.4870442677186637,
      "tokens_seen": 1649344512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029076744845848307,
      "loss": 2.8797,
      "theoretical_loss": 3.487032272917454,
      "tokens_seen": 1649410048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029075089843011155,
      "loss": 2.5127,
      "theoretical_loss": 3.4870202787262636,
      "tokens_seen": 1649475584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002907343484017401,
      "loss": 2.3918,
      "theoretical_loss": 3.4870082851450364,
      "tokens_seen": 1649541120
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002907177983733686,
      "loss": 2.5505,
      "theoretical_loss": 3.4869962921737176,
      "tokens_seen": 1649606656
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029070124834499716,
      "loss": 2.5564,
      "theoretical_loss": 3.486984299812252,
      "tokens_seen": 1649672192
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029068469831662565,
      "loss": 2.7249,
      "theoretical_loss": 3.486972308060584,
      "tokens_seen": 1649737728
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002906681482882542,
      "loss": 2.5142,
      "theoretical_loss": 3.486960316918659,
      "tokens_seen": 1649803264
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002906515982598827,
      "loss": 2.0581,
      "theoretical_loss": 3.4869483263864214,
      "tokens_seen": 1649868800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029063504823151126,
      "loss": 2.5624,
      "theoretical_loss": 3.486936336463816,
      "tokens_seen": 1649934336
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029061849820313974,
      "loss": 2.5411,
      "theoretical_loss": 3.4869243471507874,
      "tokens_seen": 1649999872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002906019481747683,
      "loss": 2.623,
      "theoretical_loss": 3.486912358447281,
      "tokens_seen": 1650065408
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002905853981463968,
      "loss": 2.571,
      "theoretical_loss": 3.486900370353241,
      "tokens_seen": 1650130944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029056884811802535,
      "loss": 2.5689,
      "theoretical_loss": 3.486888382868612,
      "tokens_seen": 1650196480
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029055229808965384,
      "loss": 2.6802,
      "theoretical_loss": 3.48687639599334,
      "tokens_seen": 1650262016
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002905357480612824,
      "loss": 2.6248,
      "theoretical_loss": 3.4868644097273687,
      "tokens_seen": 1650327552
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002905191980329109,
      "loss": 2.687,
      "theoretical_loss": 3.4868524240706433,
      "tokens_seen": 1650393088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029050264800453945,
      "loss": 2.8592,
      "theoretical_loss": 3.4868404390231085,
      "tokens_seen": 1650458624
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029048609797616793,
      "loss": 2.714,
      "theoretical_loss": 3.4868284545847095,
      "tokens_seen": 1650524160
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029046954794779647,
      "loss": 2.5346,
      "theoretical_loss": 3.4868164707553913,
      "tokens_seen": 1650589696
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000290452997919425,
      "loss": 2.3881,
      "theoretical_loss": 3.486804487535098,
      "tokens_seen": 1650655232
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029043644789105355,
      "loss": 2.5326,
      "theoretical_loss": 3.486792504923775,
      "tokens_seen": 1650720768
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029041989786268203,
      "loss": 2.56,
      "theoretical_loss": 3.486780522921367,
      "tokens_seen": 1650786304
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029040334783431057,
      "loss": 2.676,
      "theoretical_loss": 3.486768541527819,
      "tokens_seen": 1650851840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002903867978059391,
      "loss": 2.6538,
      "theoretical_loss": 3.4867565607430757,
      "tokens_seen": 1650917376
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029037024777756764,
      "loss": 2.591,
      "theoretical_loss": 3.4867445805670827,
      "tokens_seen": 1650982912
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002903536977491961,
      "loss": 2.5155,
      "theoretical_loss": 3.486732600999784,
      "tokens_seen": 1651048448
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029033714772082466,
      "loss": 2.7207,
      "theoretical_loss": 3.4867206220411244,
      "tokens_seen": 1651113984
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002903205976924532,
      "loss": 2.5034,
      "theoretical_loss": 3.48670864369105,
      "tokens_seen": 1651179520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029030404766408174,
      "loss": 2.6664,
      "theoretical_loss": 3.4866966659495047,
      "tokens_seen": 1651245056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002902874976357102,
      "loss": 2.4838,
      "theoretical_loss": 3.4866846888164336,
      "tokens_seen": 1651310592
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029027094760733876,
      "loss": 2.438,
      "theoretical_loss": 3.486672712291782,
      "tokens_seen": 1651376128
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029025439757896724,
      "loss": 2.6948,
      "theoretical_loss": 3.4866607363754945,
      "tokens_seen": 1651441664
    },
    {
      "epoch": 0.59,
      "objective/train/avg_token_score": 0.012210376560688019,
      "objective/train/avg_weight": 0.9856298565864563,
      "objective/train/docs_used": 934348,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.625027656555176,
      "objective/train/std_weight": 0.07082774490118027,
      "objective/train/theoretical_loss": 3.4866487610675163,
      "objective/train/tokens_used": 1671967200,
      "theoretical_loss": 3.4866487610675163,
      "tokens_seen": 1651507200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002902378475505958,
      "loss": 2.5192,
      "theoretical_loss": 3.4866487610675163,
      "tokens_seen": 1651507200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002902212975222243,
      "loss": 2.6563,
      "theoretical_loss": 3.4866367863677925,
      "tokens_seen": 1651572736
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029020474749385285,
      "loss": 2.5879,
      "theoretical_loss": 3.4866248122762675,
      "tokens_seen": 1651638272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029018819746548134,
      "loss": 2.3401,
      "theoretical_loss": 3.4866128387928867,
      "tokens_seen": 1651703808
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002901716474371099,
      "loss": 2.7545,
      "theoretical_loss": 3.4866008659175955,
      "tokens_seen": 1651769344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002901550974087384,
      "loss": 2.4657,
      "theoretical_loss": 3.486588893650338,
      "tokens_seen": 1651834880
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029013854738036695,
      "loss": 2.7547,
      "theoretical_loss": 3.486576921991059,
      "tokens_seen": 1651900416
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029012199735199543,
      "loss": 2.5297,
      "theoretical_loss": 3.4865649509397048,
      "tokens_seen": 1651965952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029010544732362397,
      "loss": 2.8536,
      "theoretical_loss": 3.4865529804962194,
      "tokens_seen": 1652031488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002900888972952525,
      "loss": 2.8162,
      "theoretical_loss": 3.4865410106605483,
      "tokens_seen": 1652097024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029007234726688105,
      "loss": 2.5962,
      "theoretical_loss": 3.486529041432636,
      "tokens_seen": 1652162560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029005579723850953,
      "loss": 2.3504,
      "theoretical_loss": 3.4865170728124286,
      "tokens_seen": 1652228096
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029003924721013807,
      "loss": 2.5944,
      "theoretical_loss": 3.4865051047998703,
      "tokens_seen": 1652293632
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002900226971817666,
      "loss": 2.6583,
      "theoretical_loss": 3.4864931373949055,
      "tokens_seen": 1652359168
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029000614715339514,
      "loss": 2.6075,
      "theoretical_loss": 3.486481170597481,
      "tokens_seen": 1652424704
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002899895971250236,
      "loss": 2.4858,
      "theoretical_loss": 3.4864692044075403,
      "tokens_seen": 1652490240
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028997304709665216,
      "loss": 2.762,
      "theoretical_loss": 3.4864572388250292,
      "tokens_seen": 1652555776
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002899564970682807,
      "loss": 2.5383,
      "theoretical_loss": 3.486445273849893,
      "tokens_seen": 1652621312
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028993994703990924,
      "loss": 2.5322,
      "theoretical_loss": 3.4864333094820763,
      "tokens_seen": 1652686848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002899233970115377,
      "loss": 2.6279,
      "theoretical_loss": 3.4864213457215243,
      "tokens_seen": 1652752384
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028990684698316626,
      "loss": 2.5379,
      "theoretical_loss": 3.4864093825681817,
      "tokens_seen": 1652817920
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002898902969547948,
      "loss": 2.4734,
      "theoretical_loss": 3.4863974200219943,
      "tokens_seen": 1652883456
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028987374692642333,
      "loss": 2.53,
      "theoretical_loss": 3.486385458082907,
      "tokens_seen": 1652948992
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002898571968980518,
      "loss": 2.482,
      "theoretical_loss": 3.486373496750865,
      "tokens_seen": 1653014528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028984064686968036,
      "loss": 2.5941,
      "theoretical_loss": 3.4863615360258136,
      "tokens_seen": 1653080064
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002898240968413089,
      "loss": 2.8216,
      "theoretical_loss": 3.4863495759076972,
      "tokens_seen": 1653145600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028980754681293743,
      "loss": 2.5305,
      "theoretical_loss": 3.4863376163964617,
      "tokens_seen": 1653211136
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002897909967845659,
      "loss": 2.5678,
      "theoretical_loss": 3.4863256574920514,
      "tokens_seen": 1653276672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002897744467561944,
      "loss": 2.4173,
      "theoretical_loss": 3.4863136991944126,
      "tokens_seen": 1653342208
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028975789672782293,
      "loss": 2.5156,
      "theoretical_loss": 3.486301741503489,
      "tokens_seen": 1653407744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028974134669945147,
      "loss": 2.5174,
      "theoretical_loss": 3.4862897844192275,
      "tokens_seen": 1653473280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028972479667108,
      "loss": 2.6034,
      "theoretical_loss": 3.486277827941572,
      "tokens_seen": 1653538816
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002897082466427085,
      "loss": 2.5587,
      "theoretical_loss": 3.4862658720704682,
      "tokens_seen": 1653604352
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028969169661433703,
      "loss": 2.7528,
      "theoretical_loss": 3.486253916805861,
      "tokens_seen": 1653669888
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028967514658596557,
      "loss": 2.441,
      "theoretical_loss": 3.4862419621476963,
      "tokens_seen": 1653735424
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002896585965575941,
      "loss": 2.372,
      "theoretical_loss": 3.4862300080959185,
      "tokens_seen": 1653800960
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002896420465292226,
      "loss": 2.7,
      "theoretical_loss": 3.4862180546504726,
      "tokens_seen": 1653866496
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002896254965008511,
      "loss": 2.8765,
      "theoretical_loss": 3.4862061018113053,
      "tokens_seen": 1653932032
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028960894647247966,
      "loss": 2.5006,
      "theoretical_loss": 3.48619414957836,
      "tokens_seen": 1653997568
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002895923964441082,
      "loss": 2.3523,
      "theoretical_loss": 3.4861821979515835,
      "tokens_seen": 1654063104
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002895758464157367,
      "loss": 2.8421,
      "theoretical_loss": 3.4861702469309197,
      "tokens_seen": 1654128640
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002895592963873652,
      "loss": 2.5605,
      "theoretical_loss": 3.4861582965163147,
      "tokens_seen": 1654194176
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028954274635899376,
      "loss": 2.7197,
      "theoretical_loss": 3.486146346707714,
      "tokens_seen": 1654259712
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002895261963306223,
      "loss": 2.6216,
      "theoretical_loss": 3.4861343975050616,
      "tokens_seen": 1654325248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002895096463022508,
      "loss": 2.5031,
      "theoretical_loss": 3.4861224489083042,
      "tokens_seen": 1654390784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002894930962738793,
      "loss": 2.7946,
      "theoretical_loss": 3.486110500917386,
      "tokens_seen": 1654456320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028947654624550786,
      "loss": 2.7017,
      "theoretical_loss": 3.4860985535322526,
      "tokens_seen": 1654521856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002894599962171364,
      "loss": 2.7483,
      "theoretical_loss": 3.48608660675285,
      "tokens_seen": 1654587392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002894434461887649,
      "loss": 2.7651,
      "theoretical_loss": 3.4860746605791224,
      "tokens_seen": 1654652928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002894268961603934,
      "loss": 2.582,
      "theoretical_loss": 3.486062715011016,
      "tokens_seen": 1654718464
    },
    {
      "epoch": 0.59,
      "objective/train/avg_token_score": 0.02494867518544197,
      "objective/train/avg_weight": 0.9706562161445618,
      "objective/train/docs_used": 936266,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.647510528564453,
      "objective/train/std_weight": 0.09701001644134521,
      "objective/train/theoretical_loss": 3.4860507700484757,
      "objective/train/tokens_used": 1675244000,
      "theoretical_loss": 3.4860507700484757,
      "tokens_seen": 1654784000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028941034613202195,
      "loss": 2.686,
      "theoretical_loss": 3.4860507700484757,
      "tokens_seen": 1654784000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002893937961036505,
      "loss": 2.373,
      "theoretical_loss": 3.486038825691447,
      "tokens_seen": 1654849536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000289377246075279,
      "loss": 2.5337,
      "theoretical_loss": 3.4860268819398748,
      "tokens_seen": 1654915072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002893606960469075,
      "loss": 2.1416,
      "theoretical_loss": 3.486014938793705,
      "tokens_seen": 1654980608
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028934414601853605,
      "loss": 2.7541,
      "theoretical_loss": 3.4860029962528825,
      "tokens_seen": 1655046144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002893275959901646,
      "loss": 2.4795,
      "theoretical_loss": 3.4859910543173527,
      "tokens_seen": 1655111680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028931104596179307,
      "loss": 2.6765,
      "theoretical_loss": 3.485979112987061,
      "tokens_seen": 1655177216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002892944959334216,
      "loss": 2.5354,
      "theoretical_loss": 3.4859671722619536,
      "tokens_seen": 1655242752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002892779459050501,
      "loss": 2.7182,
      "theoretical_loss": 3.4859552321419742,
      "tokens_seen": 1655308288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028926139587667863,
      "loss": 2.7089,
      "theoretical_loss": 3.4859432926270695,
      "tokens_seen": 1655373824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028924484584830716,
      "loss": 2.4378,
      "theoretical_loss": 3.4859313537171843,
      "tokens_seen": 1655439360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028922829581993565,
      "loss": 2.7941,
      "theoretical_loss": 3.4859194154122646,
      "tokens_seen": 1655504896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002892117457915642,
      "loss": 2.568,
      "theoretical_loss": 3.4859074777122547,
      "tokens_seen": 1655570432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002891951957631927,
      "loss": 2.3739,
      "theoretical_loss": 3.485895540617101,
      "tokens_seen": 1655635968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028917864573482126,
      "loss": 2.2853,
      "theoretical_loss": 3.485883604126749,
      "tokens_seen": 1655701504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028916209570644974,
      "loss": 2.4417,
      "theoretical_loss": 3.485871668241143,
      "tokens_seen": 1655767040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002891455456780783,
      "loss": 2.2726,
      "theoretical_loss": 3.4858597329602294,
      "tokens_seen": 1655832576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002891289956497068,
      "loss": 2.5922,
      "theoretical_loss": 3.4858477982839533,
      "tokens_seen": 1655898112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028911244562133536,
      "loss": 2.4097,
      "theoretical_loss": 3.4858358642122607,
      "tokens_seen": 1655963648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028909589559296384,
      "loss": 2.7561,
      "theoretical_loss": 3.485823930745096,
      "tokens_seen": 1656029184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002890793455645924,
      "loss": 2.7344,
      "theoretical_loss": 3.485811997882405,
      "tokens_seen": 1656094720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002890627955362209,
      "loss": 2.7431,
      "theoretical_loss": 3.485800065624134,
      "tokens_seen": 1656160256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028904624550784945,
      "loss": 2.5448,
      "theoretical_loss": 3.4857881339702272,
      "tokens_seen": 1656225792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028902969547947794,
      "loss": 2.3601,
      "theoretical_loss": 3.4857762029206314,
      "tokens_seen": 1656291328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002890131454511065,
      "loss": 2.396,
      "theoretical_loss": 3.485764272475291,
      "tokens_seen": 1656356864
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000288996595422735,
      "loss": 2.1656,
      "theoretical_loss": 3.485752342634152,
      "tokens_seen": 1656422400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028898004539436355,
      "loss": 2.3175,
      "theoretical_loss": 3.48574041339716,
      "tokens_seen": 1656487936
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002889634953659921,
      "loss": 2.5885,
      "theoretical_loss": 3.48572848476426,
      "tokens_seen": 1656553472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028894694533762057,
      "loss": 2.4628,
      "theoretical_loss": 3.4857165567353974,
      "tokens_seen": 1656619008
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002889303953092491,
      "loss": 2.5615,
      "theoretical_loss": 3.4857046293105185,
      "tokens_seen": 1656684544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028891384528087764,
      "loss": 2.4242,
      "theoretical_loss": 3.4856927024895685,
      "tokens_seen": 1656750080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002888972952525062,
      "loss": 2.8831,
      "theoretical_loss": 3.485680776272493,
      "tokens_seen": 1656815616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028888074522413467,
      "loss": 2.3823,
      "theoretical_loss": 3.4856688506592373,
      "tokens_seen": 1656881152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002888641951957632,
      "loss": 2.6767,
      "theoretical_loss": 3.4856569256497467,
      "tokens_seen": 1656946688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028884764516739174,
      "loss": 2.543,
      "theoretical_loss": 3.4856450012439675,
      "tokens_seen": 1657012224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002888310951390203,
      "loss": 2.4969,
      "theoretical_loss": 3.485633077441845,
      "tokens_seen": 1657077760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028881454511064876,
      "loss": 2.6791,
      "theoretical_loss": 3.485621154243325,
      "tokens_seen": 1657143296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028879799508227724,
      "loss": 2.5154,
      "theoretical_loss": 3.485609231648352,
      "tokens_seen": 1657208832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002887814450539058,
      "loss": 2.6623,
      "theoretical_loss": 3.4855973096568724,
      "tokens_seen": 1657274368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002887648950255343,
      "loss": 2.745,
      "theoretical_loss": 3.4855853882688317,
      "tokens_seen": 1657339904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028874834499716286,
      "loss": 2.6363,
      "theoretical_loss": 3.485573467484176,
      "tokens_seen": 1657405440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028873179496879134,
      "loss": 2.8206,
      "theoretical_loss": 3.48556154730285,
      "tokens_seen": 1657470976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002887152449404199,
      "loss": 2.7045,
      "theoretical_loss": 3.4855496277248,
      "tokens_seen": 1657536512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002886986949120484,
      "loss": 2.6747,
      "theoretical_loss": 3.4855377087499706,
      "tokens_seen": 1657602048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028868214488367695,
      "loss": 2.5331,
      "theoretical_loss": 3.4855257903783086,
      "tokens_seen": 1657667584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028866559485530544,
      "loss": 2.4542,
      "theoretical_loss": 3.4855138726097596,
      "tokens_seen": 1657733120
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000288649044826934,
      "loss": 2.5659,
      "theoretical_loss": 3.485501955444268,
      "tokens_seen": 1657798656
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002886324947985625,
      "loss": 2.7193,
      "theoretical_loss": 3.485490038881781,
      "tokens_seen": 1657864192
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028861594477019105,
      "loss": 2.4551,
      "theoretical_loss": 3.485478122922243,
      "tokens_seen": 1657929728
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028859939474181953,
      "loss": 2.3768,
      "theoretical_loss": 3.485466207565601,
      "tokens_seen": 1657995264
    },
    {
      "epoch": 0.59,
      "objective/train/avg_token_score": 0.008360104635357857,
      "objective/train/avg_weight": 0.9901723265647888,
      "objective/train/docs_used": 938319,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.571950912475586,
      "objective/train/std_weight": 0.05304324999451637,
      "objective/train/theoretical_loss": 3.4854542928117986,
      "objective/train/tokens_used": 1678520800,
      "theoretical_loss": 3.4854542928117986,
      "tokens_seen": 1658060800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028858284471344807,
      "loss": 2.6085,
      "theoretical_loss": 3.4854542928117986,
      "tokens_seen": 1658060800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002885662946850766,
      "loss": 2.3705,
      "theoretical_loss": 3.4854423786607835,
      "tokens_seen": 1658126336
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028854974465670515,
      "loss": 2.4406,
      "theoretical_loss": 3.4854304651125005,
      "tokens_seen": 1658191872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028853319462833363,
      "loss": 2.4498,
      "theoretical_loss": 3.4854185521668954,
      "tokens_seen": 1658257408
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028851664459996217,
      "loss": 2.6453,
      "theoretical_loss": 3.485406639823914,
      "tokens_seen": 1658322944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002885000945715907,
      "loss": 2.6994,
      "theoretical_loss": 3.485394728083502,
      "tokens_seen": 1658388480
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028848354454321924,
      "loss": 2.6547,
      "theoretical_loss": 3.485382816945605,
      "tokens_seen": 1658454016
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002884669945148477,
      "loss": 2.5344,
      "theoretical_loss": 3.485370906410168,
      "tokens_seen": 1658519552
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028845044448647626,
      "loss": 2.5594,
      "theoretical_loss": 3.485358996477138,
      "tokens_seen": 1658585088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002884338944581048,
      "loss": 2.7539,
      "theoretical_loss": 3.4853470871464602,
      "tokens_seen": 1658650624
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028841734442973334,
      "loss": 2.64,
      "theoretical_loss": 3.4853351784180804,
      "tokens_seen": 1658716160
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002884007944013618,
      "loss": 2.5358,
      "theoretical_loss": 3.4853232702919437,
      "tokens_seen": 1658781696
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028838424437299036,
      "loss": 2.6509,
      "theoretical_loss": 3.4853113627679972,
      "tokens_seen": 1658847232
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002883676943446189,
      "loss": 2.5865,
      "theoretical_loss": 3.485299455846185,
      "tokens_seen": 1658912768
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028835114431624743,
      "loss": 2.7231,
      "theoretical_loss": 3.485287549526454,
      "tokens_seen": 1658978304
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002883345942878759,
      "loss": 2.1243,
      "theoretical_loss": 3.4852756438087504,
      "tokens_seen": 1659043840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028831804425950445,
      "loss": 2.8026,
      "theoretical_loss": 3.4852637386930185,
      "tokens_seen": 1659109376
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028830149423113294,
      "loss": 2.6159,
      "theoretical_loss": 3.4852518341792056,
      "tokens_seen": 1659174912
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002882849442027615,
      "loss": 2.3242,
      "theoretical_loss": 3.485239930267256,
      "tokens_seen": 1659240448
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028826839417439,
      "loss": 2.6237,
      "theoretical_loss": 3.4852280269571168,
      "tokens_seen": 1659305984
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002882518441460185,
      "loss": 2.3009,
      "theoretical_loss": 3.4852161242487325,
      "tokens_seen": 1659371520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028823529411764703,
      "loss": 2.4861,
      "theoretical_loss": 3.4852042221420505,
      "tokens_seen": 1659437056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028821874408927557,
      "loss": 2.2411,
      "theoretical_loss": 3.4851923206370157,
      "tokens_seen": 1659502592
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002882021940609041,
      "loss": 2.5827,
      "theoretical_loss": 3.485180419733574,
      "tokens_seen": 1659568128
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002881856440325326,
      "loss": 2.5698,
      "theoretical_loss": 3.485168519431671,
      "tokens_seen": 1659633664
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028816909400416113,
      "loss": 2.6425,
      "theoretical_loss": 3.485156619731253,
      "tokens_seen": 1659699200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028815254397578967,
      "loss": 2.6005,
      "theoretical_loss": 3.4851447206322654,
      "tokens_seen": 1659764736
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002881359939474182,
      "loss": 2.5953,
      "theoretical_loss": 3.4851328221346547,
      "tokens_seen": 1659830272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002881194439190467,
      "loss": 2.3775,
      "theoretical_loss": 3.485120924238366,
      "tokens_seen": 1659895808
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002881028938906752,
      "loss": 2.5492,
      "theoretical_loss": 3.485109026943346,
      "tokens_seen": 1659961344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028808634386230376,
      "loss": 2.4167,
      "theoretical_loss": 3.4850971302495397,
      "tokens_seen": 1660026880
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002880697938339323,
      "loss": 2.5764,
      "theoretical_loss": 3.4850852341568936,
      "tokens_seen": 1660092416
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002880532438055608,
      "loss": 2.6205,
      "theoretical_loss": 3.485073338665354,
      "tokens_seen": 1660157952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002880366937771893,
      "loss": 2.6011,
      "theoretical_loss": 3.4850614437748657,
      "tokens_seen": 1660223488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028802014374881786,
      "loss": 2.4547,
      "theoretical_loss": 3.4850495494853746,
      "tokens_seen": 1660289024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002880035937204464,
      "loss": 2.531,
      "theoretical_loss": 3.4850376557968277,
      "tokens_seen": 1660354560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002879870436920749,
      "loss": 2.6405,
      "theoretical_loss": 3.4850257627091707,
      "tokens_seen": 1660420096
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002879704936637034,
      "loss": 2.6569,
      "theoretical_loss": 3.4850138702223488,
      "tokens_seen": 1660485632
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028795394363533195,
      "loss": 2.7154,
      "theoretical_loss": 3.485001978336308,
      "tokens_seen": 1660551168
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002879373936069605,
      "loss": 2.439,
      "theoretical_loss": 3.4849900870509947,
      "tokens_seen": 1660616704
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000287920843578589,
      "loss": 2.5153,
      "theoretical_loss": 3.484978196366355,
      "tokens_seen": 1660682240
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002879042935502175,
      "loss": 2.5701,
      "theoretical_loss": 3.4849663062823346,
      "tokens_seen": 1660747776
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028788774352184605,
      "loss": 2.5592,
      "theoretical_loss": 3.484954416798879,
      "tokens_seen": 1660813312
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002878711934934746,
      "loss": 2.4555,
      "theoretical_loss": 3.4849425279159347,
      "tokens_seen": 1660878848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028785464346510307,
      "loss": 2.6128,
      "theoretical_loss": 3.484930639633448,
      "tokens_seen": 1660944384
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002878380934367316,
      "loss": 2.5769,
      "theoretical_loss": 3.484918751951364,
      "tokens_seen": 1661009920
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028782154340836015,
      "loss": 2.7205,
      "theoretical_loss": 3.484906864869629,
      "tokens_seen": 1661075456
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028780499337998863,
      "loss": 2.6652,
      "theoretical_loss": 3.4848949783881893,
      "tokens_seen": 1661140992
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028778844335161717,
      "loss": 2.4795,
      "theoretical_loss": 3.484883092506991,
      "tokens_seen": 1661206528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028777189332324565,
      "loss": 2.3762,
      "theoretical_loss": 3.4848712072259795,
      "tokens_seen": 1661272064
    },
    {
      "epoch": 0.59,
      "objective/train/avg_token_score": 0.013685280457139015,
      "objective/train/avg_weight": 0.9838877320289612,
      "objective/train/docs_used": 940185,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.863860607147217,
      "objective/train/std_weight": 0.07989224046468735,
      "objective/train/theoretical_loss": 3.484859322545101,
      "objective/train/tokens_used": 1681797600,
      "theoretical_loss": 3.484859322545101,
      "tokens_seen": 1661337600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002877553432948742,
      "loss": 2.6702,
      "theoretical_loss": 3.484859322545101,
      "tokens_seen": 1661337600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002877387932665027,
      "loss": 2.649,
      "theoretical_loss": 3.4848474384643024,
      "tokens_seen": 1661403136
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028772224323813126,
      "loss": 2.4802,
      "theoretical_loss": 3.484835554983529,
      "tokens_seen": 1661468672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002877056932097598,
      "loss": 2.7128,
      "theoretical_loss": 3.484823672102726,
      "tokens_seen": 1661534208
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002876891431813883,
      "loss": 2.553,
      "theoretical_loss": 3.484811789821841,
      "tokens_seen": 1661599744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002876725931530168,
      "loss": 2.5278,
      "theoretical_loss": 3.484799908140819,
      "tokens_seen": 1661665280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028765604312464536,
      "loss": 2.7355,
      "theoretical_loss": 3.4847880270596066,
      "tokens_seen": 1661730816
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002876394930962739,
      "loss": 2.4403,
      "theoretical_loss": 3.4847761465781497,
      "tokens_seen": 1661796352
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002876229430679024,
      "loss": 2.6827,
      "theoretical_loss": 3.484764266696394,
      "tokens_seen": 1661861888
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002876063930395309,
      "loss": 2.7207,
      "theoretical_loss": 3.4847523874142867,
      "tokens_seen": 1661927424
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028758984301115946,
      "loss": 2.4247,
      "theoretical_loss": 3.4847405087317727,
      "tokens_seen": 1661992960
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000287573292982788,
      "loss": 2.6259,
      "theoretical_loss": 3.4847286306487986,
      "tokens_seen": 1662058496
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002875567429544165,
      "loss": 2.5726,
      "theoretical_loss": 3.48471675316531,
      "tokens_seen": 1662124032
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000287540192926045,
      "loss": 2.5761,
      "theoretical_loss": 3.4847048762812536,
      "tokens_seen": 1662189568
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028752364289767355,
      "loss": 2.7004,
      "theoretical_loss": 3.4846929999965757,
      "tokens_seen": 1662255104
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002875070928693021,
      "loss": 2.3589,
      "theoretical_loss": 3.484681124311222,
      "tokens_seen": 1662320640
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028749054284093057,
      "loss": 2.4736,
      "theoretical_loss": 3.4846692492251385,
      "tokens_seen": 1662386176
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002874739928125591,
      "loss": 2.607,
      "theoretical_loss": 3.484657374738272,
      "tokens_seen": 1662451712
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028745744278418765,
      "loss": 2.7973,
      "theoretical_loss": 3.4846455008505677,
      "tokens_seen": 1662517248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002874408927558162,
      "loss": 2.3684,
      "theoretical_loss": 3.484633627561972,
      "tokens_seen": 1662582784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028742434272744467,
      "loss": 2.5037,
      "theoretical_loss": 3.484621754872432,
      "tokens_seen": 1662648320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002874077926990732,
      "loss": 2.3173,
      "theoretical_loss": 3.4846098827818928,
      "tokens_seen": 1662713856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028739124267070174,
      "loss": 2.5743,
      "theoretical_loss": 3.4845980112903003,
      "tokens_seen": 1662779392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002873746926423303,
      "loss": 2.6262,
      "theoretical_loss": 3.484586140397602,
      "tokens_seen": 1662844928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028735814261395876,
      "loss": 2.4874,
      "theoretical_loss": 3.484574270103743,
      "tokens_seen": 1662910464
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002873415925855873,
      "loss": 2.5534,
      "theoretical_loss": 3.484562400408671,
      "tokens_seen": 1662976000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028732504255721584,
      "loss": 2.5602,
      "theoretical_loss": 3.4845505313123297,
      "tokens_seen": 1663041536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002873084925288443,
      "loss": 2.6576,
      "theoretical_loss": 3.4845386628146677,
      "tokens_seen": 1663107072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028729194250047286,
      "loss": 2.6982,
      "theoretical_loss": 3.4845267949156296,
      "tokens_seen": 1663172608
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028727539247210134,
      "loss": 2.5703,
      "theoretical_loss": 3.4845149276151623,
      "tokens_seen": 1663238144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002872588424437299,
      "loss": 2.5861,
      "theoretical_loss": 3.484503060913212,
      "tokens_seen": 1663303680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002872422924153584,
      "loss": 2.3947,
      "theoretical_loss": 3.484491194809725,
      "tokens_seen": 1663369216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028722574238698696,
      "loss": 2.5885,
      "theoretical_loss": 3.484479329304647,
      "tokens_seen": 1663434752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028720919235861544,
      "loss": 2.8066,
      "theoretical_loss": 3.484467464397925,
      "tokens_seen": 1663500288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000287192642330244,
      "loss": 2.6828,
      "theoretical_loss": 3.484455600089505,
      "tokens_seen": 1663565824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002871760923018725,
      "loss": 2.4573,
      "theoretical_loss": 3.4844437363793332,
      "tokens_seen": 1663631360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028715954227350105,
      "loss": 2.2124,
      "theoretical_loss": 3.4844318732673556,
      "tokens_seen": 1663696896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028714299224512954,
      "loss": 2.5141,
      "theoretical_loss": 3.484420010753519,
      "tokens_seen": 1663762432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002871264422167581,
      "loss": 2.6645,
      "theoretical_loss": 3.484408148837769,
      "tokens_seen": 1663827968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002871098921883866,
      "loss": 2.5952,
      "theoretical_loss": 3.4843962875200525,
      "tokens_seen": 1663893504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028709334216001515,
      "loss": 2.5672,
      "theoretical_loss": 3.484384426800316,
      "tokens_seen": 1663959040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028707679213164363,
      "loss": 2.6502,
      "theoretical_loss": 3.4843725666785046,
      "tokens_seen": 1664024576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028706024210327217,
      "loss": 2.5557,
      "theoretical_loss": 3.484360707154566,
      "tokens_seen": 1664090112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002870436920749007,
      "loss": 2.4036,
      "theoretical_loss": 3.484348848228446,
      "tokens_seen": 1664155648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028702714204652924,
      "loss": 2.5399,
      "theoretical_loss": 3.4843369899000907,
      "tokens_seen": 1664221184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028701059201815773,
      "loss": 2.5911,
      "theoretical_loss": 3.484325132169446,
      "tokens_seen": 1664286720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028699404198978627,
      "loss": 2.7469,
      "theoretical_loss": 3.4843132750364596,
      "tokens_seen": 1664352256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002869774919614148,
      "loss": 2.5538,
      "theoretical_loss": 3.4843014185010768,
      "tokens_seen": 1664417792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028696094193304334,
      "loss": 2.7593,
      "theoretical_loss": 3.4842895625632444,
      "tokens_seen": 1664483328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002869443919046718,
      "loss": 2.2269,
      "theoretical_loss": 3.4842777072229083,
      "tokens_seen": 1664548864
    },
    {
      "epoch": 0.59,
      "objective/train/avg_token_score": 0.03458588197827339,
      "objective/train/avg_weight": 0.9593294858932495,
      "objective/train/docs_used": 942299,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.991722583770752,
      "objective/train/std_weight": 0.14357082545757294,
      "objective/train/theoretical_loss": 3.484265852480015,
      "objective/train/tokens_used": 1685074400,
      "theoretical_loss": 3.484265852480015,
      "tokens_seen": 1664614400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028692784187630036,
      "loss": 2.2876,
      "theoretical_loss": 3.484265852480015,
      "tokens_seen": 1664614400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002869112918479289,
      "loss": 2.4907,
      "theoretical_loss": 3.4842539983345118,
      "tokens_seen": 1664679936
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028689474181955744,
      "loss": 2.37,
      "theoretical_loss": 3.484242144786344,
      "tokens_seen": 1664745472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002868781917911859,
      "loss": 2.5577,
      "theoretical_loss": 3.4842302918354577,
      "tokens_seen": 1664811008
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028686164176281446,
      "loss": 2.4533,
      "theoretical_loss": 3.4842184394818,
      "tokens_seen": 1664876544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000286845091734443,
      "loss": 2.8064,
      "theoretical_loss": 3.4842065877253177,
      "tokens_seen": 1664942080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002868285417060715,
      "loss": 2.1509,
      "theoretical_loss": 3.4841947365659562,
      "tokens_seen": 1665007616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002868119916777,
      "loss": 2.6095,
      "theoretical_loss": 3.4841828860036625,
      "tokens_seen": 1665073152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002867954416493285,
      "loss": 2.5525,
      "theoretical_loss": 3.484171036038383,
      "tokens_seen": 1665138688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028677889162095704,
      "loss": 2.4587,
      "theoretical_loss": 3.4841591866700643,
      "tokens_seen": 1665204224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002867623415925856,
      "loss": 2.3863,
      "theoretical_loss": 3.4841473378986523,
      "tokens_seen": 1665269760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002867457915642141,
      "loss": 2.7571,
      "theoretical_loss": 3.4841354897240935,
      "tokens_seen": 1665335296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002867292415358426,
      "loss": 2.5407,
      "theoretical_loss": 3.4841236421463346,
      "tokens_seen": 1665400832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028671269150747113,
      "loss": 2.6868,
      "theoretical_loss": 3.4841117951653224,
      "tokens_seen": 1665466368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028669614147909967,
      "loss": 2.697,
      "theoretical_loss": 3.4840999487810027,
      "tokens_seen": 1665531904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002866795914507282,
      "loss": 2.3999,
      "theoretical_loss": 3.4840881029933226,
      "tokens_seen": 1665597440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002866630414223567,
      "loss": 2.338,
      "theoretical_loss": 3.484076257802228,
      "tokens_seen": 1665662976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028664649139398523,
      "loss": 2.5177,
      "theoretical_loss": 3.4840644132076655,
      "tokens_seen": 1665728512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00028662994136561377,
      "loss": 2.5841,
      "theoretical_loss": 3.484052569209582,
      "tokens_seen": 1665794048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002866133913372423,
      "loss": 2.6769,
      "theoretical_loss": 3.4840407258079233,
      "tokens_seen": 1665859584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002865968413088708,
      "loss": 2.4925,
      "theoretical_loss": 3.4840288830026367,
      "tokens_seen": 1665925120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002865802912804993,
      "loss": 2.5784,
      "theoretical_loss": 3.484017040793668,
      "tokens_seen": 1665990656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028656374125212786,
      "loss": 2.4415,
      "theoretical_loss": 3.4840051991809644,
      "tokens_seen": 1666056192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002865471912237564,
      "loss": 2.4505,
      "theoretical_loss": 3.4839933581644718,
      "tokens_seen": 1666121728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002865306411953849,
      "loss": 2.5531,
      "theoretical_loss": 3.4839815177441373,
      "tokens_seen": 1666187264
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002865140911670134,
      "loss": 2.6124,
      "theoretical_loss": 3.4839696779199065,
      "tokens_seen": 1666252800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028649754113864196,
      "loss": 2.6907,
      "theoretical_loss": 3.483957838691727,
      "tokens_seen": 1666318336
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002864809911102705,
      "loss": 2.6499,
      "theoretical_loss": 3.483946000059545,
      "tokens_seen": 1666383872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000286464441081899,
      "loss": 2.5786,
      "theoretical_loss": 3.483934162023307,
      "tokens_seen": 1666449408
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002864478910535275,
      "loss": 2.3971,
      "theoretical_loss": 3.4839223245829594,
      "tokens_seen": 1666514944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028643134102515605,
      "loss": 2.8735,
      "theoretical_loss": 3.4839104877384486,
      "tokens_seen": 1666580480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002864147909967846,
      "loss": 2.6412,
      "theoretical_loss": 3.483898651489722,
      "tokens_seen": 1666646016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002863982409684131,
      "loss": 2.6936,
      "theoretical_loss": 3.4838868158367253,
      "tokens_seen": 1666711552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002863816909400416,
      "loss": 2.6257,
      "theoretical_loss": 3.4838749807794063,
      "tokens_seen": 1666777088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028636514091167015,
      "loss": 2.6686,
      "theoretical_loss": 3.4838631463177103,
      "tokens_seen": 1666842624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002863485908832987,
      "loss": 2.7381,
      "theoretical_loss": 3.483851312451584,
      "tokens_seen": 1666908160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028633204085492717,
      "loss": 2.4476,
      "theoretical_loss": 3.4838394791809746,
      "tokens_seen": 1666973696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002863154908265557,
      "loss": 2.6165,
      "theoretical_loss": 3.4838276465058287,
      "tokens_seen": 1667039232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002862989407981842,
      "loss": 2.6563,
      "theoretical_loss": 3.4838158144260927,
      "tokens_seen": 1667104768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028628239076981273,
      "loss": 2.4944,
      "theoretical_loss": 3.483803982941714,
      "tokens_seen": 1667170304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028626584074144127,
      "loss": 2.8944,
      "theoretical_loss": 3.4837921520526374,
      "tokens_seen": 1667235840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002862492907130698,
      "loss": 2.7706,
      "theoretical_loss": 3.483780321758811,
      "tokens_seen": 1667301376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002862327406846983,
      "loss": 2.7272,
      "theoretical_loss": 3.4837684920601815,
      "tokens_seen": 1667366912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002862161906563268,
      "loss": 2.7416,
      "theoretical_loss": 3.483756662956695,
      "tokens_seen": 1667432448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028619964062795536,
      "loss": 2.7145,
      "theoretical_loss": 3.483744834448298,
      "tokens_seen": 1667497984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002861830905995839,
      "loss": 2.8058,
      "theoretical_loss": 3.483733006534938,
      "tokens_seen": 1667563520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002861665405712124,
      "loss": 2.6824,
      "theoretical_loss": 3.483721179216561,
      "tokens_seen": 1667629056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002861499905428409,
      "loss": 2.627,
      "theoretical_loss": 3.4837093524931144,
      "tokens_seen": 1667694592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028613344051446946,
      "loss": 2.9017,
      "theoretical_loss": 3.4836975263645438,
      "tokens_seen": 1667760128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000286116890486098,
      "loss": 2.4952,
      "theoretical_loss": 3.483685700830797,
      "tokens_seen": 1667825664
    },
    {
      "epoch": 0.6,
      "objective/train/avg_token_score": 0.009160331450402737,
      "objective/train/avg_weight": 0.9892209768295288,
      "objective/train/docs_used": 944133,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8806910514831543,
      "objective/train/std_weight": 0.039626069366931915,
      "objective/train/theoretical_loss": 3.48367387589182,
      "objective/train/tokens_used": 1688351200,
      "theoretical_loss": 3.48367387589182,
      "tokens_seen": 1667891200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002861003404577265,
      "loss": 2.6709,
      "theoretical_loss": 3.48367387589182,
      "tokens_seen": 1667891200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000286083790429355,
      "loss": 2.4921,
      "theoretical_loss": 3.48366205154756,
      "tokens_seen": 1667956736
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028606724040098355,
      "loss": 2.6956,
      "theoretical_loss": 3.4836502277979626,
      "tokens_seen": 1668022272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002860506903726121,
      "loss": 2.697,
      "theoretical_loss": 3.4836384046429765,
      "tokens_seen": 1668087808
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002860341403442406,
      "loss": 2.6984,
      "theoretical_loss": 3.4836265820825467,
      "tokens_seen": 1668153344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002860175903158691,
      "loss": 2.3793,
      "theoretical_loss": 3.483614760116621,
      "tokens_seen": 1668218880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028600104028749765,
      "loss": 2.7747,
      "theoretical_loss": 3.4836029387451455,
      "tokens_seen": 1668284416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002859844902591262,
      "loss": 2.4184,
      "theoretical_loss": 3.483591117968067,
      "tokens_seen": 1668349952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028596794023075467,
      "loss": 2.9222,
      "theoretical_loss": 3.483579297785333,
      "tokens_seen": 1668415488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002859513902023832,
      "loss": 2.7553,
      "theoretical_loss": 3.483567478196889,
      "tokens_seen": 1668481024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028593484017401175,
      "loss": 2.6085,
      "theoretical_loss": 3.483555659202683,
      "tokens_seen": 1668546560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002859182901456403,
      "loss": 2.7515,
      "theoretical_loss": 3.4835438408026613,
      "tokens_seen": 1668612096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028590174011726877,
      "loss": 2.7236,
      "theoretical_loss": 3.483532022996771,
      "tokens_seen": 1668677632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002858851900888973,
      "loss": 2.573,
      "theoretical_loss": 3.483520205784958,
      "tokens_seen": 1668743168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028586864006052584,
      "loss": 2.6542,
      "theoretical_loss": 3.48350838916717,
      "tokens_seen": 1668808704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002858520900321544,
      "loss": 2.5076,
      "theoretical_loss": 3.4834965731433534,
      "tokens_seen": 1668874240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028583554000378286,
      "loss": 2.7622,
      "theoretical_loss": 3.4834847577134553,
      "tokens_seen": 1668939776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028581898997541135,
      "loss": 2.3024,
      "theoretical_loss": 3.4834729428774223,
      "tokens_seen": 1669005312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002858024399470399,
      "loss": 2.6365,
      "theoretical_loss": 3.4834611286352013,
      "tokens_seen": 1669070848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002857858899186684,
      "loss": 2.7407,
      "theoretical_loss": 3.4834493149867387,
      "tokens_seen": 1669136384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028576933989029696,
      "loss": 2.5755,
      "theoretical_loss": 3.483437501931982,
      "tokens_seen": 1669201920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028575278986192544,
      "loss": 2.5597,
      "theoretical_loss": 3.483425689470878,
      "tokens_seen": 1669267456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000285736239833554,
      "loss": 2.4299,
      "theoretical_loss": 3.4834138776033736,
      "tokens_seen": 1669332992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002857196898051825,
      "loss": 2.6444,
      "theoretical_loss": 3.483402066329415,
      "tokens_seen": 1669398528
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028570313977681106,
      "loss": 2.7298,
      "theoretical_loss": 3.48339025564895,
      "tokens_seen": 1669464064
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028568658974843954,
      "loss": 2.6569,
      "theoretical_loss": 3.4833784455619248,
      "tokens_seen": 1669529600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002856700397200681,
      "loss": 2.7448,
      "theoretical_loss": 3.4833666360682867,
      "tokens_seen": 1669595136
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002856534896916966,
      "loss": 2.3801,
      "theoretical_loss": 3.4833548271679824,
      "tokens_seen": 1669660672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028563693966332515,
      "loss": 2.5663,
      "theoretical_loss": 3.4833430188609587,
      "tokens_seen": 1669726208
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028562038963495363,
      "loss": 2.642,
      "theoretical_loss": 3.4833312111471626,
      "tokens_seen": 1669791744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028560383960658217,
      "loss": 2.488,
      "theoretical_loss": 3.483319404026541,
      "tokens_seen": 1669857280
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002855872895782107,
      "loss": 2.7352,
      "theoretical_loss": 3.4833075974990404,
      "tokens_seen": 1669922816
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028557073954983925,
      "loss": 2.6031,
      "theoretical_loss": 3.4832957915646086,
      "tokens_seen": 1669988352
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028555418952146773,
      "loss": 2.5943,
      "theoretical_loss": 3.4832839862231926,
      "tokens_seen": 1670053888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028553763949309627,
      "loss": 2.8299,
      "theoretical_loss": 3.4832721814747383,
      "tokens_seen": 1670119424
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002855210894647248,
      "loss": 2.5958,
      "theoretical_loss": 3.4832603773191932,
      "tokens_seen": 1670184960
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028550453943635334,
      "loss": 2.7593,
      "theoretical_loss": 3.4832485737565047,
      "tokens_seen": 1670250496
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002854879894079818,
      "loss": 2.5423,
      "theoretical_loss": 3.483236770786619,
      "tokens_seen": 1670316032
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028547143937961036,
      "loss": 2.6286,
      "theoretical_loss": 3.4832249684094836,
      "tokens_seen": 1670381568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002854548893512389,
      "loss": 2.4684,
      "theoretical_loss": 3.4832131666250454,
      "tokens_seen": 1670447104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028543833932286744,
      "loss": 2.496,
      "theoretical_loss": 3.483201365433251,
      "tokens_seen": 1670512640
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002854217892944959,
      "loss": 2.7819,
      "theoretical_loss": 3.4831895648340474,
      "tokens_seen": 1670578176
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028540523926612446,
      "loss": 2.6084,
      "theoretical_loss": 3.4831777648273823,
      "tokens_seen": 1670643712
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000285388689237753,
      "loss": 2.5027,
      "theoretical_loss": 3.4831659654132023,
      "tokens_seen": 1670709248
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028537213920938154,
      "loss": 2.6493,
      "theoretical_loss": 3.4831541665914543,
      "tokens_seen": 1670774784
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028535558918101,
      "loss": 2.6616,
      "theoretical_loss": 3.4831423683620852,
      "tokens_seen": 1670840320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002853390391526385,
      "loss": 2.4222,
      "theoretical_loss": 3.4831305707250424,
      "tokens_seen": 1670905856
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028532248912426704,
      "loss": 2.3339,
      "theoretical_loss": 3.483118773680273,
      "tokens_seen": 1670971392
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002853059390958956,
      "loss": 2.6248,
      "theoretical_loss": 3.4831069772277234,
      "tokens_seen": 1671036928
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002852893890675241,
      "loss": 2.6738,
      "theoretical_loss": 3.483095181367341,
      "tokens_seen": 1671102464
    },
    {
      "epoch": 0.6,
      "objective/train/avg_token_score": 0.015720685943961143,
      "objective/train/avg_weight": 0.9814952611923218,
      "objective/train/docs_used": 945841,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2386341094970703,
      "objective/train/std_weight": 0.08218727260828018,
      "objective/train/theoretical_loss": 3.4830833860990733,
      "objective/train/tokens_used": 1691628000,
      "theoretical_loss": 3.4830833860990733,
      "tokens_seen": 1671168000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002852728390391526,
      "loss": 2.4594,
      "theoretical_loss": 3.4830833860990733,
      "tokens_seen": 1671168000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028525628901078114,
      "loss": 2.5603,
      "theoretical_loss": 3.4830715914228665,
      "tokens_seen": 1671233536
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028523973898240967,
      "loss": 2.6314,
      "theoretical_loss": 3.4830597973386683,
      "tokens_seen": 1671299072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002852231889540382,
      "loss": 2.8783,
      "theoretical_loss": 3.4830480038464255,
      "tokens_seen": 1671364608
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002852066389256667,
      "loss": 2.5009,
      "theoretical_loss": 3.4830362109460857,
      "tokens_seen": 1671430144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028519008889729523,
      "loss": 2.4981,
      "theoretical_loss": 3.483024418637595,
      "tokens_seen": 1671495680
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028517353886892377,
      "loss": 2.3357,
      "theoretical_loss": 3.4830126269209014,
      "tokens_seen": 1671561216
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002851569888405523,
      "loss": 2.9364,
      "theoretical_loss": 3.4830008357959517,
      "tokens_seen": 1671626752
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002851404388121808,
      "loss": 2.6016,
      "theoretical_loss": 3.4829890452626926,
      "tokens_seen": 1671692288
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028512388878380933,
      "loss": 2.7592,
      "theoretical_loss": 3.482977255321072,
      "tokens_seen": 1671757824
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028510733875543786,
      "loss": 2.5528,
      "theoretical_loss": 3.482965465971036,
      "tokens_seen": 1671823360
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002850907887270664,
      "loss": 2.5622,
      "theoretical_loss": 3.4829536772125325,
      "tokens_seen": 1671888896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002850742386986949,
      "loss": 2.6944,
      "theoretical_loss": 3.482941889045509,
      "tokens_seen": 1671954432
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002850576886703234,
      "loss": 2.621,
      "theoretical_loss": 3.4829301014699117,
      "tokens_seen": 1672019968
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028504113864195196,
      "loss": 2.7454,
      "theoretical_loss": 3.482918314485688,
      "tokens_seen": 1672085504
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002850245886135805,
      "loss": 2.8161,
      "theoretical_loss": 3.4829065280927853,
      "tokens_seen": 1672151040
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000285008038585209,
      "loss": 2.723,
      "theoretical_loss": 3.482894742291151,
      "tokens_seen": 1672216576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002849914885568375,
      "loss": 2.8108,
      "theoretical_loss": 3.4828829570807316,
      "tokens_seen": 1672282112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028497493852846606,
      "loss": 2.6607,
      "theoretical_loss": 3.4828711724614747,
      "tokens_seen": 1672347648
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002849583885000946,
      "loss": 2.7619,
      "theoretical_loss": 3.4828593884333277,
      "tokens_seen": 1672413184
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028494183847172313,
      "loss": 2.5505,
      "theoretical_loss": 3.4828476049962367,
      "tokens_seen": 1672478720
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002849252884433516,
      "loss": 2.3857,
      "theoretical_loss": 3.4828358221501503,
      "tokens_seen": 1672544256
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028490873841498015,
      "loss": 2.7154,
      "theoretical_loss": 3.4828240398950143,
      "tokens_seen": 1672609792
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002848921883866087,
      "loss": 2.5369,
      "theoretical_loss": 3.4828122582307777,
      "tokens_seen": 1672675328
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028487563835823723,
      "loss": 2.6354,
      "theoretical_loss": 3.482800477157386,
      "tokens_seen": 1672740864
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002848590883298657,
      "loss": 2.5246,
      "theoretical_loss": 3.4827886966747874,
      "tokens_seen": 1672806400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002848425383014942,
      "loss": 2.1704,
      "theoretical_loss": 3.482776916782929,
      "tokens_seen": 1672871936
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028482598827312273,
      "loss": 2.5098,
      "theoretical_loss": 3.482765137481757,
      "tokens_seen": 1672937472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028480943824475127,
      "loss": 2.5037,
      "theoretical_loss": 3.4827533587712205,
      "tokens_seen": 1673003008
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002847928882163798,
      "loss": 2.6486,
      "theoretical_loss": 3.4827415806512656,
      "tokens_seen": 1673068544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002847763381880083,
      "loss": 2.5048,
      "theoretical_loss": 3.4827298031218397,
      "tokens_seen": 1673134080
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028475978815963683,
      "loss": 2.5981,
      "theoretical_loss": 3.48271802618289,
      "tokens_seen": 1673199616
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028474323813126537,
      "loss": 2.3959,
      "theoretical_loss": 3.4827062498343637,
      "tokens_seen": 1673265152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002847266881028939,
      "loss": 2.5896,
      "theoretical_loss": 3.4826944740762085,
      "tokens_seen": 1673330688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002847101380745224,
      "loss": 2.6504,
      "theoretical_loss": 3.4826826989083712,
      "tokens_seen": 1673396224
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002846935880461509,
      "loss": 2.7876,
      "theoretical_loss": 3.4826709243307996,
      "tokens_seen": 1673461760
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028467703801777946,
      "loss": 2.7831,
      "theoretical_loss": 3.4826591503434408,
      "tokens_seen": 1673527296
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000284660487989408,
      "loss": 2.5519,
      "theoretical_loss": 3.4826473769462414,
      "tokens_seen": 1673592832
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002846439379610365,
      "loss": 2.765,
      "theoretical_loss": 3.4826356041391495,
      "tokens_seen": 1673658368
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000284627387932665,
      "loss": 2.3426,
      "theoretical_loss": 3.482623831922113,
      "tokens_seen": 1673723904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028461083790429356,
      "loss": 2.3273,
      "theoretical_loss": 3.4826120602950774,
      "tokens_seen": 1673789440
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002845942878759221,
      "loss": 2.492,
      "theoretical_loss": 3.4826002892579915,
      "tokens_seen": 1673854976
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002845777378475506,
      "loss": 2.6528,
      "theoretical_loss": 3.482588518810802,
      "tokens_seen": 1673920512
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002845611878191791,
      "loss": 2.6742,
      "theoretical_loss": 3.4825767489534574,
      "tokens_seen": 1673986048
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028454463779080765,
      "loss": 2.7865,
      "theoretical_loss": 3.482564979685903,
      "tokens_seen": 1674051584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002845280877624362,
      "loss": 2.6239,
      "theoretical_loss": 3.482553211008088,
      "tokens_seen": 1674117120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002845115377340647,
      "loss": 2.6276,
      "theoretical_loss": 3.482541442919959,
      "tokens_seen": 1674182656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002844949877056932,
      "loss": 2.6634,
      "theoretical_loss": 3.482529675421463,
      "tokens_seen": 1674248192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028447843767732175,
      "loss": 2.5584,
      "theoretical_loss": 3.4825179085125484,
      "tokens_seen": 1674313728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002844618876489503,
      "loss": 2.7016,
      "theoretical_loss": 3.4825061421931616,
      "tokens_seen": 1674379264
    },
    {
      "epoch": 0.6,
      "objective/train/avg_token_score": 0.024898286908864975,
      "objective/train/avg_weight": 0.9706918001174927,
      "objective/train/docs_used": 947805,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.895209312438965,
      "objective/train/std_weight": 0.10977835208177567,
      "objective/train/theoretical_loss": 3.4824943764632503,
      "objective/train/tokens_used": 1694904800,
      "theoretical_loss": 3.4824943764632503,
      "tokens_seen": 1674444800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028444533762057877,
      "loss": 2.4894,
      "theoretical_loss": 3.4824943764632503,
      "tokens_seen": 1674444800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002844287875922073,
      "loss": 2.5492,
      "theoretical_loss": 3.4824826113227623,
      "tokens_seen": 1674510336
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028441223756383585,
      "loss": 1.9896,
      "theoretical_loss": 3.4824708467716445,
      "tokens_seen": 1674575872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002843956875354644,
      "loss": 2.6119,
      "theoretical_loss": 3.4824590828098447,
      "tokens_seen": 1674641408
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028437913750709287,
      "loss": 2.6283,
      "theoretical_loss": 3.48244731943731,
      "tokens_seen": 1674706944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028436258747872135,
      "loss": 2.4425,
      "theoretical_loss": 3.4824355566539875,
      "tokens_seen": 1674772480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002843460374503499,
      "loss": 2.5584,
      "theoretical_loss": 3.4824237944598257,
      "tokens_seen": 1674838016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002843294874219784,
      "loss": 2.5152,
      "theoretical_loss": 3.482412032854771,
      "tokens_seen": 1674903552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028431293739360696,
      "loss": 2.8823,
      "theoretical_loss": 3.482400271838771,
      "tokens_seen": 1674969088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028429638736523545,
      "loss": 2.3693,
      "theoretical_loss": 3.482388511411774,
      "tokens_seen": 1675034624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000284279837336864,
      "loss": 2.5613,
      "theoretical_loss": 3.4823767515737267,
      "tokens_seen": 1675100160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002842632873084925,
      "loss": 2.4164,
      "theoretical_loss": 3.4823649923245767,
      "tokens_seen": 1675165696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028424673728012106,
      "loss": 2.6473,
      "theoretical_loss": 3.4823532336642717,
      "tokens_seen": 1675231232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028423018725174954,
      "loss": 2.6931,
      "theoretical_loss": 3.482341475592759,
      "tokens_seen": 1675296768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002842136372233781,
      "loss": 2.5159,
      "theoretical_loss": 3.4823297181099857,
      "tokens_seen": 1675362304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002841970871950066,
      "loss": 2.4579,
      "theoretical_loss": 3.4823179612159,
      "tokens_seen": 1675427840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028418053716663515,
      "loss": 2.9283,
      "theoretical_loss": 3.4823062049104485,
      "tokens_seen": 1675493376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028416398713826364,
      "loss": 2.5591,
      "theoretical_loss": 3.4822944491935797,
      "tokens_seen": 1675558912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002841474371098922,
      "loss": 2.436,
      "theoretical_loss": 3.4822826940652405,
      "tokens_seen": 1675624448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002841308870815207,
      "loss": 2.664,
      "theoretical_loss": 3.482270939525379,
      "tokens_seen": 1675689984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028411433705314925,
      "loss": 2.745,
      "theoretical_loss": 3.482259185573942,
      "tokens_seen": 1675755520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028409778702477773,
      "loss": 2.48,
      "theoretical_loss": 3.4822474322108774,
      "tokens_seen": 1675821056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028408123699640627,
      "loss": 2.5141,
      "theoretical_loss": 3.4822356794361324,
      "tokens_seen": 1675886592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002840646869680348,
      "loss": 2.607,
      "theoretical_loss": 3.482223927249655,
      "tokens_seen": 1675952128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028404813693966335,
      "loss": 2.7913,
      "theoretical_loss": 3.482212175651393,
      "tokens_seen": 1676017664
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028403158691129183,
      "loss": 2.7487,
      "theoretical_loss": 3.4822004246412934,
      "tokens_seen": 1676083200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028401503688292037,
      "loss": 2.52,
      "theoretical_loss": 3.4821886742193042,
      "tokens_seen": 1676148736
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002839984868545489,
      "loss": 2.6211,
      "theoretical_loss": 3.4821769243853717,
      "tokens_seen": 1676214272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028398193682617744,
      "loss": 2.4177,
      "theoretical_loss": 3.4821651751394453,
      "tokens_seen": 1676279808
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002839653867978059,
      "loss": 2.7631,
      "theoretical_loss": 3.4821534264814717,
      "tokens_seen": 1676345344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028394883676943446,
      "loss": 2.5828,
      "theoretical_loss": 3.4821416784113985,
      "tokens_seen": 1676410880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000283932286741063,
      "loss": 2.6791,
      "theoretical_loss": 3.4821299309291733,
      "tokens_seen": 1676476416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028391573671269154,
      "loss": 2.5818,
      "theoretical_loss": 3.4821181840347437,
      "tokens_seen": 1676541952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028389918668432,
      "loss": 2.7477,
      "theoretical_loss": 3.4821064377280577,
      "tokens_seen": 1676607488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002838826366559485,
      "loss": 2.1664,
      "theoretical_loss": 3.4820946920090625,
      "tokens_seen": 1676673024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028386608662757704,
      "loss": 2.5858,
      "theoretical_loss": 3.4820829468777057,
      "tokens_seen": 1676738560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002838495365992056,
      "loss": 2.3971,
      "theoretical_loss": 3.482071202333935,
      "tokens_seen": 1676804096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002838329865708341,
      "loss": 2.7756,
      "theoretical_loss": 3.482059458377698,
      "tokens_seen": 1676869632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002838164365424626,
      "loss": 2.5125,
      "theoretical_loss": 3.4820477150089433,
      "tokens_seen": 1676935168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028379988651409114,
      "loss": 2.6332,
      "theoretical_loss": 3.4820359722276164,
      "tokens_seen": 1677000704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002837833364857197,
      "loss": 2.7973,
      "theoretical_loss": 3.4820242300336677,
      "tokens_seen": 1677066240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002837667864573482,
      "loss": 2.6115,
      "theoretical_loss": 3.4820124884270425,
      "tokens_seen": 1677131776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028375023642897675,
      "loss": 2.7901,
      "theoretical_loss": 3.4820007474076893,
      "tokens_seen": 1677197312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028373368640060523,
      "loss": 2.5978,
      "theoretical_loss": 3.481989006975556,
      "tokens_seen": 1677262848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028371713637223377,
      "loss": 2.6766,
      "theoretical_loss": 3.4819772671305906,
      "tokens_seen": 1677328384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002837005863438623,
      "loss": 2.7691,
      "theoretical_loss": 3.48196552787274,
      "tokens_seen": 1677393920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028368403631549085,
      "loss": 2.6727,
      "theoretical_loss": 3.4819537892019525,
      "tokens_seen": 1677459456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028366748628711933,
      "loss": 2.587,
      "theoretical_loss": 3.4819420511181756,
      "tokens_seen": 1677524992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028365093625874787,
      "loss": 2.7131,
      "theoretical_loss": 3.481930313621357,
      "tokens_seen": 1677590528
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002836343862303764,
      "loss": 2.6939,
      "theoretical_loss": 3.4819185767114442,
      "tokens_seen": 1677656064
    },
    {
      "epoch": 0.6,
      "objective/train/avg_token_score": 0.014005700126290321,
      "objective/train/avg_weight": 0.9835160374641418,
      "objective/train/docs_used": 948902,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.680532932281494,
      "objective/train/std_weight": 0.07203130424022675,
      "objective/train/theoretical_loss": 3.4819068403883855,
      "objective/train/tokens_used": 1698181600,
      "theoretical_loss": 3.4819068403883855,
      "tokens_seen": 1677721600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028361783620200494,
      "loss": 2.646,
      "theoretical_loss": 3.4819068403883855,
      "tokens_seen": 1677721600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002836012861736334,
      "loss": 2.4985,
      "theoretical_loss": 3.4818951046521276,
      "tokens_seen": 1677787136
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028358473614526196,
      "loss": 2.5326,
      "theoretical_loss": 3.4818833695026195,
      "tokens_seen": 1677852672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002835681861168905,
      "loss": 2.528,
      "theoretical_loss": 3.4818716349398082,
      "tokens_seen": 1677918208
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028355163608851904,
      "loss": 2.6982,
      "theoretical_loss": 3.481859900963642,
      "tokens_seen": 1677983744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002835350860601475,
      "loss": 2.6213,
      "theoretical_loss": 3.481848167574068,
      "tokens_seen": 1678049280
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028351853603177606,
      "loss": 2.7949,
      "theoretical_loss": 3.481836434771034,
      "tokens_seen": 1678114816
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002835019860034046,
      "loss": 2.8478,
      "theoretical_loss": 3.4818247025544884,
      "tokens_seen": 1678180352
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028348543597503313,
      "loss": 2.7773,
      "theoretical_loss": 3.481812970924379,
      "tokens_seen": 1678245888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002834688859466616,
      "loss": 2.6864,
      "theoretical_loss": 3.481801239880652,
      "tokens_seen": 1678311424
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028345233591829016,
      "loss": 2.8032,
      "theoretical_loss": 3.4817895094232574,
      "tokens_seen": 1678376960
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002834357858899187,
      "loss": 2.5525,
      "theoretical_loss": 3.481777779552141,
      "tokens_seen": 1678442496
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028341923586154723,
      "loss": 2.5662,
      "theoretical_loss": 3.481766050267253,
      "tokens_seen": 1678508032
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002834026858331757,
      "loss": 2.372,
      "theoretical_loss": 3.481754321568539,
      "tokens_seen": 1678573568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002833861358048042,
      "loss": 2.6695,
      "theoretical_loss": 3.4817425934559476,
      "tokens_seen": 1678639104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028336958577643273,
      "loss": 2.7848,
      "theoretical_loss": 3.4817308659294266,
      "tokens_seen": 1678704640
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028335303574806127,
      "loss": 2.6431,
      "theoretical_loss": 3.4817191389889244,
      "tokens_seen": 1678770176
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002833364857196898,
      "loss": 2.367,
      "theoretical_loss": 3.481707412634388,
      "tokens_seen": 1678835712
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002833199356913183,
      "loss": 2.8894,
      "theoretical_loss": 3.481695686865766,
      "tokens_seen": 1678901248
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028330338566294683,
      "loss": 2.7998,
      "theoretical_loss": 3.481683961683005,
      "tokens_seen": 1678966784
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028328683563457537,
      "loss": 2.5465,
      "theoretical_loss": 3.4816722370860544,
      "tokens_seen": 1679032320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002832702856062039,
      "loss": 2.5512,
      "theoretical_loss": 3.4816605130748615,
      "tokens_seen": 1679097856
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002832537355778324,
      "loss": 2.7492,
      "theoretical_loss": 3.4816487896493733,
      "tokens_seen": 1679163392
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002832371855494609,
      "loss": 2.5828,
      "theoretical_loss": 3.4816370668095393,
      "tokens_seen": 1679228928
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028322063552108946,
      "loss": 2.7628,
      "theoretical_loss": 3.4816253445553063,
      "tokens_seen": 1679294464
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000283204085492718,
      "loss": 2.3424,
      "theoretical_loss": 3.4816136228866217,
      "tokens_seen": 1679360000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002831875354643465,
      "loss": 2.5945,
      "theoretical_loss": 3.481601901803435,
      "tokens_seen": 1679425536
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000283170985435975,
      "loss": 2.5137,
      "theoretical_loss": 3.481590181305693,
      "tokens_seen": 1679491072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028315443540760356,
      "loss": 2.3352,
      "theoretical_loss": 3.4815784613933434,
      "tokens_seen": 1679556608
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002831378853792321,
      "loss": 2.589,
      "theoretical_loss": 3.481566742066335,
      "tokens_seen": 1679622144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002831213353508606,
      "loss": 2.5554,
      "theoretical_loss": 3.4815550233246153,
      "tokens_seen": 1679687680
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002831047853224891,
      "loss": 2.7012,
      "theoretical_loss": 3.481543305168132,
      "tokens_seen": 1679753216
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028308823529411766,
      "loss": 2.2265,
      "theoretical_loss": 3.4815315875968333,
      "tokens_seen": 1679818752
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002830716852657462,
      "loss": 2.6345,
      "theoretical_loss": 3.4815198706106676,
      "tokens_seen": 1679884288
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002830551352373747,
      "loss": 2.6362,
      "theoretical_loss": 3.481508154209582,
      "tokens_seen": 1679949824
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002830385852090032,
      "loss": 2.4938,
      "theoretical_loss": 3.4814964383935245,
      "tokens_seen": 1680015360
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028302203518063175,
      "loss": 2.6768,
      "theoretical_loss": 3.4814847231624437,
      "tokens_seen": 1680080896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002830054851522603,
      "loss": 2.3178,
      "theoretical_loss": 3.4814730085162875,
      "tokens_seen": 1680146432
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002829889351238888,
      "loss": 2.8339,
      "theoretical_loss": 3.4814612944550034,
      "tokens_seen": 1680211968
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002829723850955173,
      "loss": 2.4532,
      "theoretical_loss": 3.4814495809785395,
      "tokens_seen": 1680277504
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028295583506714585,
      "loss": 2.4263,
      "theoretical_loss": 3.4814378680868447,
      "tokens_seen": 1680343040
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002829392850387744,
      "loss": 2.2922,
      "theoretical_loss": 3.4814261557798654,
      "tokens_seen": 1680408576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028292273501040287,
      "loss": 2.7049,
      "theoretical_loss": 3.4814144440575507,
      "tokens_seen": 1680474112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002829061849820314,
      "loss": 2.6665,
      "theoretical_loss": 3.4814027329198485,
      "tokens_seen": 1680539648
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002828896349536599,
      "loss": 2.5604,
      "theoretical_loss": 3.481391022366707,
      "tokens_seen": 1680605184
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028287308492528843,
      "loss": 2.8844,
      "theoretical_loss": 3.481379312398073,
      "tokens_seen": 1680670720
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028285653489691696,
      "loss": 2.6039,
      "theoretical_loss": 3.4813676030138963,
      "tokens_seen": 1680736256
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028283998486854545,
      "loss": 2.5823,
      "theoretical_loss": 3.481355894214124,
      "tokens_seen": 1680801792
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000282823434840174,
      "loss": 2.7358,
      "theoretical_loss": 3.481344185998704,
      "tokens_seen": 1680867328
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002828068848118025,
      "loss": 2.5233,
      "theoretical_loss": 3.4813324783675847,
      "tokens_seen": 1680932864
    },
    {
      "epoch": 0.6,
      "objective/train/avg_token_score": 0.026763740926980972,
      "objective/train/avg_weight": 0.9685049057006836,
      "objective/train/docs_used": 950589,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8561525344848633,
      "objective/train/std_weight": 0.10420725494623184,
      "objective/train/theoretical_loss": 3.481320771320714,
      "objective/train/tokens_used": 1701458400,
      "theoretical_loss": 3.481320771320714,
      "tokens_seen": 1680998400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028279033478343106,
      "loss": 2.6929,
      "theoretical_loss": 3.481320771320714,
      "tokens_seen": 1680998400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028277378475505954,
      "loss": 2.2875,
      "theoretical_loss": 3.4813090648580403,
      "tokens_seen": 1681063936
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002827572347266881,
      "loss": 2.6257,
      "theoretical_loss": 3.4812973589795106,
      "tokens_seen": 1681129472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002827406846983166,
      "loss": 2.7365,
      "theoretical_loss": 3.4812856536850747,
      "tokens_seen": 1681195008
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028272413466994516,
      "loss": 2.6411,
      "theoretical_loss": 3.4812739489746796,
      "tokens_seen": 1681260544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028270758464157364,
      "loss": 2.6411,
      "theoretical_loss": 3.4812622448482733,
      "tokens_seen": 1681326080
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002826910346132022,
      "loss": 2.5933,
      "theoretical_loss": 3.481250541305804,
      "tokens_seen": 1681391616
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002826744845848307,
      "loss": 2.7834,
      "theoretical_loss": 3.4812388383472204,
      "tokens_seen": 1681457152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028265793455645925,
      "loss": 2.5568,
      "theoretical_loss": 3.4812271359724702,
      "tokens_seen": 1681522688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028264138452808774,
      "loss": 2.8024,
      "theoretical_loss": 3.481215434181501,
      "tokens_seen": 1681588224
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002826248344997163,
      "loss": 2.6357,
      "theoretical_loss": 3.481203732974262,
      "tokens_seen": 1681653760
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002826082844713448,
      "loss": 2.7196,
      "theoretical_loss": 3.481192032350701,
      "tokens_seen": 1681719296
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028259173444297335,
      "loss": 2.6287,
      "theoretical_loss": 3.481180332310765,
      "tokens_seen": 1681784832
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028257518441460183,
      "loss": 2.7866,
      "theoretical_loss": 3.481168632854404,
      "tokens_seen": 1681850368
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028255863438623037,
      "loss": 2.558,
      "theoretical_loss": 3.4811569339815644,
      "tokens_seen": 1681915904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002825420843578589,
      "loss": 2.608,
      "theoretical_loss": 3.4811452356921953,
      "tokens_seen": 1681981440
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028252553432948744,
      "loss": 2.6161,
      "theoretical_loss": 3.4811335379862456,
      "tokens_seen": 1682046976
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028250898430111593,
      "loss": 2.3526,
      "theoretical_loss": 3.4811218408636617,
      "tokens_seen": 1682112512
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028249243427274447,
      "loss": 2.5155,
      "theoretical_loss": 3.481110144324393,
      "tokens_seen": 1682178048
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000282475884244373,
      "loss": 2.5684,
      "theoretical_loss": 3.481098448368387,
      "tokens_seen": 1682243584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028245933421600154,
      "loss": 2.4748,
      "theoretical_loss": 3.4810867529955933,
      "tokens_seen": 1682309120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028244278418763,
      "loss": 2.6779,
      "theoretical_loss": 3.481075058205958,
      "tokens_seen": 1682374656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028242623415925856,
      "loss": 2.2303,
      "theoretical_loss": 3.4810633639994313,
      "tokens_seen": 1682440192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002824096841308871,
      "loss": 2.3413,
      "theoretical_loss": 3.48105167037596,
      "tokens_seen": 1682505728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002823931341025156,
      "loss": 2.5056,
      "theoretical_loss": 3.4810399773354925,
      "tokens_seen": 1682571264
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002823765840741441,
      "loss": 2.8642,
      "theoretical_loss": 3.4810282848779783,
      "tokens_seen": 1682636800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028236003404577266,
      "loss": 2.7219,
      "theoretical_loss": 3.481016593003364,
      "tokens_seen": 1682702336
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028234348401740114,
      "loss": 2.4616,
      "theoretical_loss": 3.4810049017115983,
      "tokens_seen": 1682767872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002823269339890297,
      "loss": 2.5348,
      "theoretical_loss": 3.4809932110026303,
      "tokens_seen": 1682833408
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002823103839606582,
      "loss": 2.6201,
      "theoretical_loss": 3.4809815208764077,
      "tokens_seen": 1682898944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028229383393228675,
      "loss": 2.8912,
      "theoretical_loss": 3.480969831332878,
      "tokens_seen": 1682964480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028227728390391524,
      "loss": 2.4764,
      "theoretical_loss": 3.4809581423719904,
      "tokens_seen": 1683030016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002822607338755438,
      "loss": 2.5614,
      "theoretical_loss": 3.480946453993693,
      "tokens_seen": 1683095552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002822441838471723,
      "loss": 2.7137,
      "theoretical_loss": 3.480934766197934,
      "tokens_seen": 1683161088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028222763381880085,
      "loss": 2.5989,
      "theoretical_loss": 3.480923078984662,
      "tokens_seen": 1683226624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028221108379042933,
      "loss": 2.6577,
      "theoretical_loss": 3.480911392353825,
      "tokens_seen": 1683292160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028219453376205787,
      "loss": 2.7165,
      "theoretical_loss": 3.4808997063053706,
      "tokens_seen": 1683357696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002821779837336864,
      "loss": 2.6059,
      "theoretical_loss": 3.480888020839248,
      "tokens_seen": 1683423232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028216143370531495,
      "loss": 2.6299,
      "theoretical_loss": 3.480876335955406,
      "tokens_seen": 1683488768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028214488367694343,
      "loss": 2.9732,
      "theoretical_loss": 3.4808646516537918,
      "tokens_seen": 1683554304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028212833364857197,
      "loss": 2.7982,
      "theoretical_loss": 3.4808529679343536,
      "tokens_seen": 1683619840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002821117836202005,
      "loss": 2.5534,
      "theoretical_loss": 3.480841284797041,
      "tokens_seen": 1683685376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028209523359182904,
      "loss": 2.6766,
      "theoretical_loss": 3.4808296022418013,
      "tokens_seen": 1683750912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002820786835634575,
      "loss": 2.6488,
      "theoretical_loss": 3.480817920268583,
      "tokens_seen": 1683816448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028206213353508606,
      "loss": 2.6786,
      "theoretical_loss": 3.480806238877335,
      "tokens_seen": 1683881984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002820455835067146,
      "loss": 2.2949,
      "theoretical_loss": 3.4807945580680055,
      "tokens_seen": 1683947520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028202903347834314,
      "loss": 2.5673,
      "theoretical_loss": 3.480782877840542,
      "tokens_seen": 1684013056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002820124834499716,
      "loss": 2.5222,
      "theoretical_loss": 3.480771198194894,
      "tokens_seen": 1684078592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028199593342160016,
      "loss": 2.4059,
      "theoretical_loss": 3.4807595191310092,
      "tokens_seen": 1684144128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002819793833932287,
      "loss": 2.5717,
      "theoretical_loss": 3.4807478406488364,
      "tokens_seen": 1684209664
    },
    {
      "epoch": 0.6,
      "objective/train/avg_token_score": 0.014429762959480286,
      "objective/train/avg_weight": 0.9830335378646851,
      "objective/train/docs_used": 952447,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5195178985595703,
      "objective/train/std_weight": 0.057609185576438904,
      "objective/train/theoretical_loss": 3.4807361627483235,
      "objective/train/tokens_used": 1704735200,
      "theoretical_loss": 3.4807361627483235,
      "tokens_seen": 1684275200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028196283336485723,
      "loss": 2.4088,
      "theoretical_loss": 3.4807361627483235,
      "tokens_seen": 1684275200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002819462833364857,
      "loss": 2.6788,
      "theoretical_loss": 3.4807244854294197,
      "tokens_seen": 1684340736
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028192973330811425,
      "loss": 2.7915,
      "theoretical_loss": 3.480712808692072,
      "tokens_seen": 1684406272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028191318327974274,
      "loss": 2.4688,
      "theoretical_loss": 3.4807011325362307,
      "tokens_seen": 1684471808
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002818966332513713,
      "loss": 2.4516,
      "theoretical_loss": 3.4806894569618425,
      "tokens_seen": 1684537344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002818800832229998,
      "loss": 2.4603,
      "theoretical_loss": 3.480677781968857,
      "tokens_seen": 1684602880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002818635331946283,
      "loss": 2.6596,
      "theoretical_loss": 3.480666107557221,
      "tokens_seen": 1684668416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028184698316625683,
      "loss": 2.4884,
      "theoretical_loss": 3.4806544337268854,
      "tokens_seen": 1684733952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028183043313788537,
      "loss": 2.5599,
      "theoretical_loss": 3.4806427604777967,
      "tokens_seen": 1684799488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002818138831095139,
      "loss": 2.7025,
      "theoretical_loss": 3.4806310878099045,
      "tokens_seen": 1684865024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002817973330811424,
      "loss": 2.7106,
      "theoretical_loss": 3.4806194157231567,
      "tokens_seen": 1684930560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028178078305277093,
      "loss": 2.5984,
      "theoretical_loss": 3.480607744217502,
      "tokens_seen": 1684996096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028176423302439947,
      "loss": 2.7773,
      "theoretical_loss": 3.4805960732928876,
      "tokens_seen": 1685061632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000281747682996028,
      "loss": 2.722,
      "theoretical_loss": 3.480584402949264,
      "tokens_seen": 1685127168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002817311329676565,
      "loss": 2.6123,
      "theoretical_loss": 3.4805727331865786,
      "tokens_seen": 1685192704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000281714582939285,
      "loss": 2.6882,
      "theoretical_loss": 3.48056106400478,
      "tokens_seen": 1685258240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028169803291091356,
      "loss": 2.9752,
      "theoretical_loss": 3.480549395403817,
      "tokens_seen": 1685323776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002816814828825421,
      "loss": 2.616,
      "theoretical_loss": 3.4805377273836378,
      "tokens_seen": 1685389312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002816649328541706,
      "loss": 2.5082,
      "theoretical_loss": 3.4805260599441903,
      "tokens_seen": 1685454848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002816483828257991,
      "loss": 2.5596,
      "theoretical_loss": 3.480514393085424,
      "tokens_seen": 1685520384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028163183279742766,
      "loss": 2.4453,
      "theoretical_loss": 3.4805027268072877,
      "tokens_seen": 1685585920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002816152827690562,
      "loss": 2.6038,
      "theoretical_loss": 3.4804910611097286,
      "tokens_seen": 1685651456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002815987327406847,
      "loss": 2.3761,
      "theoretical_loss": 3.4804793959926963,
      "tokens_seen": 1685716992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002815821827123132,
      "loss": 2.7828,
      "theoretical_loss": 3.480467731456139,
      "tokens_seen": 1685782528
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028156563268394175,
      "loss": 2.464,
      "theoretical_loss": 3.4804560675000054,
      "tokens_seen": 1685848064
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002815490826555703,
      "loss": 2.7107,
      "theoretical_loss": 3.4804444041242437,
      "tokens_seen": 1685913600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002815325326271988,
      "loss": 2.6089,
      "theoretical_loss": 3.480432741328803,
      "tokens_seen": 1685979136
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002815159825988273,
      "loss": 2.7836,
      "theoretical_loss": 3.4804210791136314,
      "tokens_seen": 1686044672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028149943257045585,
      "loss": 2.8085,
      "theoretical_loss": 3.4804094174786773,
      "tokens_seen": 1686110208
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002814828825420844,
      "loss": 2.4714,
      "theoretical_loss": 3.4803977564238897,
      "tokens_seen": 1686175744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028146633251371287,
      "loss": 2.2492,
      "theoretical_loss": 3.4803860959492177,
      "tokens_seen": 1686241280
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002814497824853414,
      "loss": 2.8507,
      "theoretical_loss": 3.4803744360546087,
      "tokens_seen": 1686306816
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028143323245696995,
      "loss": 2.8141,
      "theoretical_loss": 3.4803627767400123,
      "tokens_seen": 1686372352
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028141668242859843,
      "loss": 2.4603,
      "theoretical_loss": 3.480351118005376,
      "tokens_seen": 1686437888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028140013240022697,
      "loss": 2.96,
      "theoretical_loss": 3.48033945985065,
      "tokens_seen": 1686503424
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028138358237185545,
      "loss": 2.6737,
      "theoretical_loss": 3.4803278022757818,
      "tokens_seen": 1686568960
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000281367032343484,
      "loss": 2.9309,
      "theoretical_loss": 3.4803161452807196,
      "tokens_seen": 1686634496
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002813504823151125,
      "loss": 2.2661,
      "theoretical_loss": 3.4803044888654133,
      "tokens_seen": 1686700032
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028133393228674106,
      "loss": 3.0255,
      "theoretical_loss": 3.480292833029811,
      "tokens_seen": 1686765568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028131738225836955,
      "loss": 2.6643,
      "theoretical_loss": 3.480281177773861,
      "tokens_seen": 1686831104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002813008322299981,
      "loss": 2.6522,
      "theoretical_loss": 3.4802695230975127,
      "tokens_seen": 1686896640
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002812842822016266,
      "loss": 2.7278,
      "theoretical_loss": 3.480257869000714,
      "tokens_seen": 1686962176
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028126773217325516,
      "loss": 2.487,
      "theoretical_loss": 3.4802462154834135,
      "tokens_seen": 1687027712
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028125118214488364,
      "loss": 2.753,
      "theoretical_loss": 3.4802345625455606,
      "tokens_seen": 1687093248
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002812346321165122,
      "loss": 2.4801,
      "theoretical_loss": 3.480222910187104,
      "tokens_seen": 1687158784
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002812180820881407,
      "loss": 2.6994,
      "theoretical_loss": 3.4802112584079916,
      "tokens_seen": 1687224320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028120153205976926,
      "loss": 2.5866,
      "theoretical_loss": 3.4801996072081725,
      "tokens_seen": 1687289856
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028118498203139774,
      "loss": 2.605,
      "theoretical_loss": 3.4801879565875957,
      "tokens_seen": 1687355392
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002811684320030263,
      "loss": 2.7449,
      "theoretical_loss": 3.4801763065462095,
      "tokens_seen": 1687420928
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002811518819746548,
      "loss": 2.6814,
      "theoretical_loss": 3.4801646570839626,
      "tokens_seen": 1687486464
    },
    {
      "epoch": 0.6,
      "objective/train/avg_token_score": 0.010340344160795212,
      "objective/train/avg_weight": 0.9878295063972473,
      "objective/train/docs_used": 954180,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.913677215576172,
      "objective/train/std_weight": 0.05597910284996033,
      "objective/train/theoretical_loss": 3.480153008200804,
      "objective/train/tokens_used": 1708012000,
      "theoretical_loss": 3.480153008200804,
      "tokens_seen": 1687552000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028113533194628335,
      "loss": 2.7387,
      "theoretical_loss": 3.480153008200804,
      "tokens_seen": 1687552000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028111878191791183,
      "loss": 2.5794,
      "theoretical_loss": 3.480141359896683,
      "tokens_seen": 1687617536
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028110223188954037,
      "loss": 2.8029,
      "theoretical_loss": 3.4801297121715464,
      "tokens_seen": 1687683072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002810856818611689,
      "loss": 2.9272,
      "theoretical_loss": 3.4801180650253447,
      "tokens_seen": 1687748608
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028106913183279745,
      "loss": 2.5372,
      "theoretical_loss": 3.480106418458026,
      "tokens_seen": 1687814144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028105258180442593,
      "loss": 2.1132,
      "theoretical_loss": 3.4800947724695392,
      "tokens_seen": 1687879680
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028103603177605447,
      "loss": 2.7178,
      "theoretical_loss": 3.4800831270598334,
      "tokens_seen": 1687945216
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000281019481747683,
      "loss": 2.4874,
      "theoretical_loss": 3.4800714822288565,
      "tokens_seen": 1688010752
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028100293171931154,
      "loss": 2.6243,
      "theoretical_loss": 3.4800598379765577,
      "tokens_seen": 1688076288
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002809863816909401,
      "loss": 2.5825,
      "theoretical_loss": 3.4800481943028863,
      "tokens_seen": 1688141824
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028096983166256856,
      "loss": 2.5589,
      "theoretical_loss": 3.48003655120779,
      "tokens_seen": 1688207360
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002809532816341971,
      "loss": 2.4937,
      "theoretical_loss": 3.4800249086912185,
      "tokens_seen": 1688272896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028093673160582564,
      "loss": 2.552,
      "theoretical_loss": 3.4800132667531205,
      "tokens_seen": 1688338432
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002809201815774541,
      "loss": 2.626,
      "theoretical_loss": 3.4800016253934447,
      "tokens_seen": 1688403968
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028090363154908266,
      "loss": 2.6887,
      "theoretical_loss": 3.4799899846121396,
      "tokens_seen": 1688469504
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028088708152071114,
      "loss": 2.5945,
      "theoretical_loss": 3.4799783444091545,
      "tokens_seen": 1688535040
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002808705314923397,
      "loss": 2.7233,
      "theoretical_loss": 3.4799667047844376,
      "tokens_seen": 1688600576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002808539814639682,
      "loss": 2.6876,
      "theoretical_loss": 3.479955065737938,
      "tokens_seen": 1688666112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028083743143559676,
      "loss": 2.5964,
      "theoretical_loss": 3.479943427269605,
      "tokens_seen": 1688731648
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028082088140722524,
      "loss": 2.6803,
      "theoretical_loss": 3.4799317893793873,
      "tokens_seen": 1688797184
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002808043313788538,
      "loss": 2.6885,
      "theoretical_loss": 3.479920152067233,
      "tokens_seen": 1688862720
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002807877813504823,
      "loss": 2.4802,
      "theoretical_loss": 3.4799085153330918,
      "tokens_seen": 1688928256
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028077123132211085,
      "loss": 2.5955,
      "theoretical_loss": 3.479896879176912,
      "tokens_seen": 1688993792
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028075468129373934,
      "loss": 2.5333,
      "theoretical_loss": 3.479885243598643,
      "tokens_seen": 1689059328
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002807381312653679,
      "loss": 2.6628,
      "theoretical_loss": 3.4798736085982336,
      "tokens_seen": 1689124864
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002807215812369964,
      "loss": 2.5925,
      "theoretical_loss": 3.479861974175632,
      "tokens_seen": 1689190400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028070503120862495,
      "loss": 2.5423,
      "theoretical_loss": 3.4798503403307874,
      "tokens_seen": 1689255936
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028068848118025343,
      "loss": 2.8141,
      "theoretical_loss": 3.4798387070636494,
      "tokens_seen": 1689321472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028067193115188197,
      "loss": 2.7686,
      "theoretical_loss": 3.479827074374166,
      "tokens_seen": 1689387008
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002806553811235105,
      "loss": 2.4075,
      "theoretical_loss": 3.4798154422622867,
      "tokens_seen": 1689452544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028063883109513904,
      "loss": 2.5833,
      "theoretical_loss": 3.4798038107279603,
      "tokens_seen": 1689518080
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028062228106676753,
      "loss": 2.819,
      "theoretical_loss": 3.4797921797711355,
      "tokens_seen": 1689583616
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028060573103839607,
      "loss": 2.6354,
      "theoretical_loss": 3.4797805493917613,
      "tokens_seen": 1689649152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002805891810100246,
      "loss": 2.688,
      "theoretical_loss": 3.4797689195897865,
      "tokens_seen": 1689714688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028057263098165314,
      "loss": 2.592,
      "theoretical_loss": 3.4797572903651606,
      "tokens_seen": 1689780224
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002805560809532816,
      "loss": 2.7505,
      "theoretical_loss": 3.479745661717832,
      "tokens_seen": 1689845760
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028053953092491016,
      "loss": 2.5481,
      "theoretical_loss": 3.47973403364775,
      "tokens_seen": 1689911296
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002805229808965387,
      "loss": 2.386,
      "theoretical_loss": 3.4797224061548633,
      "tokens_seen": 1689976832
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028050643086816724,
      "loss": 2.4875,
      "theoretical_loss": 3.4797107792391206,
      "tokens_seen": 1690042368
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002804898808397957,
      "loss": 2.6605,
      "theoretical_loss": 3.4796991529004715,
      "tokens_seen": 1690107904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028047333081142426,
      "loss": 2.3597,
      "theoretical_loss": 3.4796875271388648,
      "tokens_seen": 1690173440
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002804567807830528,
      "loss": 2.5504,
      "theoretical_loss": 3.4796759019542494,
      "tokens_seen": 1690238976
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002804402307546813,
      "loss": 2.532,
      "theoretical_loss": 3.479664277346574,
      "tokens_seen": 1690304512
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002804236807263098,
      "loss": 2.447,
      "theoretical_loss": 3.479652653315788,
      "tokens_seen": 1690370048
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002804071306979383,
      "loss": 2.6205,
      "theoretical_loss": 3.4796410298618405,
      "tokens_seen": 1690435584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028039058066956684,
      "loss": 2.1383,
      "theoretical_loss": 3.4796294069846803,
      "tokens_seen": 1690501120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002803740306411954,
      "loss": 2.7683,
      "theoretical_loss": 3.4796177846842564,
      "tokens_seen": 1690566656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002803574806128239,
      "loss": 2.649,
      "theoretical_loss": 3.479606162960518,
      "tokens_seen": 1690632192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002803409305844524,
      "loss": 2.6126,
      "theoretical_loss": 3.4795945418134133,
      "tokens_seen": 1690697728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028032438055608093,
      "loss": 2.716,
      "theoretical_loss": 3.479582921242893,
      "tokens_seen": 1690763264
    },
    {
      "epoch": 0.6,
      "objective/train/avg_token_score": 0.013309408910572529,
      "objective/train/avg_weight": 0.9843361973762512,
      "objective/train/docs_used": 956253,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6116442680358887,
      "objective/train/std_weight": 0.0651121437549591,
      "objective/train/theoretical_loss": 3.4795713012489045,
      "objective/train/tokens_used": 1711288800,
      "theoretical_loss": 3.4795713012489045,
      "tokens_seen": 1690828800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028030783052770947,
      "loss": 2.616,
      "theoretical_loss": 3.4795713012489045,
      "tokens_seen": 1690828800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000280291280499338,
      "loss": 2.7101,
      "theoretical_loss": 3.4795596818313976,
      "tokens_seen": 1690894336
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002802747304709665,
      "loss": 2.5376,
      "theoretical_loss": 3.4795480629903217,
      "tokens_seen": 1690959872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028025818044259503,
      "loss": 2.3906,
      "theoretical_loss": 3.479536444725625,
      "tokens_seen": 1691025408
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028024163041422357,
      "loss": 2.6242,
      "theoretical_loss": 3.4795248270372574,
      "tokens_seen": 1691090944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002802250803858521,
      "loss": 2.884,
      "theoretical_loss": 3.479513209925167,
      "tokens_seen": 1691156480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002802085303574806,
      "loss": 2.4785,
      "theoretical_loss": 3.479501593389304,
      "tokens_seen": 1691222016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002801919803291091,
      "loss": 2.3948,
      "theoretical_loss": 3.479489977429617,
      "tokens_seen": 1691287552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028017543030073766,
      "loss": 2.7318,
      "theoretical_loss": 3.4794783620460548,
      "tokens_seen": 1691353088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002801588802723662,
      "loss": 2.1357,
      "theoretical_loss": 3.479466747238567,
      "tokens_seen": 1691418624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002801423302439947,
      "loss": 2.5986,
      "theoretical_loss": 3.4794551330071024,
      "tokens_seen": 1691484160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002801257802156232,
      "loss": 2.5965,
      "theoretical_loss": 3.47944351935161,
      "tokens_seen": 1691549696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028010923018725176,
      "loss": 2.5679,
      "theoretical_loss": 3.4794319062720396,
      "tokens_seen": 1691615232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002800926801588803,
      "loss": 2.4036,
      "theoretical_loss": 3.479420293768339,
      "tokens_seen": 1691680768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002800761301305088,
      "loss": 2.4593,
      "theoretical_loss": 3.479408681840459,
      "tokens_seen": 1691746304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002800595801021373,
      "loss": 2.4996,
      "theoretical_loss": 3.4793970704883477,
      "tokens_seen": 1691811840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00028004303007376585,
      "loss": 2.8079,
      "theoretical_loss": 3.4793854597119545,
      "tokens_seen": 1691877376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002800264800453944,
      "loss": 2.6101,
      "theoretical_loss": 3.4793738495112283,
      "tokens_seen": 1691942912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002800099300170229,
      "loss": 2.8008,
      "theoretical_loss": 3.4793622398861186,
      "tokens_seen": 1692008448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002799933799886514,
      "loss": 2.7471,
      "theoretical_loss": 3.4793506308365747,
      "tokens_seen": 1692073984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027997682996027995,
      "loss": 2.5471,
      "theoretical_loss": 3.479339022362545,
      "tokens_seen": 1692139520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002799602799319085,
      "loss": 2.7534,
      "theoretical_loss": 3.47932741446398,
      "tokens_seen": 1692205056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027994372990353697,
      "loss": 2.4891,
      "theoretical_loss": 3.4793158071408277,
      "tokens_seen": 1692270592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027992717987516545,
      "loss": 2.6913,
      "theoretical_loss": 3.4793042003930377,
      "tokens_seen": 1692336128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000279910629846794,
      "loss": 2.4611,
      "theoretical_loss": 3.479292594220559,
      "tokens_seen": 1692401664
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027989407981842253,
      "loss": 2.3073,
      "theoretical_loss": 3.4792809886233416,
      "tokens_seen": 1692467200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027987752979005107,
      "loss": 2.5691,
      "theoretical_loss": 3.4792693836013333,
      "tokens_seen": 1692532736
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027986097976167955,
      "loss": 2.4794,
      "theoretical_loss": 3.4792577791544845,
      "tokens_seen": 1692598272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002798444297333081,
      "loss": 2.5991,
      "theoretical_loss": 3.4792461752827446,
      "tokens_seen": 1692663808
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002798278797049366,
      "loss": 2.5595,
      "theoretical_loss": 3.479234571986062,
      "tokens_seen": 1692729344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027981132967656516,
      "loss": 2.7731,
      "theoretical_loss": 3.479222969264386,
      "tokens_seen": 1692794880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027979477964819365,
      "loss": 2.4418,
      "theoretical_loss": 3.4792113671176663,
      "tokens_seen": 1692860416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002797782296198222,
      "loss": 2.5968,
      "theoretical_loss": 3.479199765545852,
      "tokens_seen": 1692925952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002797616795914507,
      "loss": 2.6,
      "theoretical_loss": 3.4791881645488916,
      "tokens_seen": 1692991488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027974512956307926,
      "loss": 2.4389,
      "theoretical_loss": 3.479176564126736,
      "tokens_seen": 1693057024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002797285795347078,
      "loss": 2.7087,
      "theoretical_loss": 3.4791649642793328,
      "tokens_seen": 1693122560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002797120295063363,
      "loss": 2.6955,
      "theoretical_loss": 3.479153365006632,
      "tokens_seen": 1693188096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002796954794779648,
      "loss": 2.6811,
      "theoretical_loss": 3.479141766308584,
      "tokens_seen": 1693253632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027967892944959335,
      "loss": 2.4434,
      "theoretical_loss": 3.479130168185136,
      "tokens_seen": 1693319168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002796623794212219,
      "loss": 2.5585,
      "theoretical_loss": 3.4791185706362384,
      "tokens_seen": 1693384704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002796458293928504,
      "loss": 2.6395,
      "theoretical_loss": 3.4791069736618407,
      "tokens_seen": 1693450240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002796292793644789,
      "loss": 2.552,
      "theoretical_loss": 3.479095377261892,
      "tokens_seen": 1693515776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027961272933610745,
      "loss": 2.5144,
      "theoretical_loss": 3.4790837814363407,
      "tokens_seen": 1693581312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000279596179307736,
      "loss": 2.8355,
      "theoretical_loss": 3.479072186185138,
      "tokens_seen": 1693646848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027957962927936447,
      "loss": 2.3161,
      "theoretical_loss": 3.4790605915082313,
      "tokens_seen": 1693712384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000279563079250993,
      "loss": 2.6052,
      "theoretical_loss": 3.479048997405571,
      "tokens_seen": 1693777920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027954652922262155,
      "loss": 2.7177,
      "theoretical_loss": 3.479037403877107,
      "tokens_seen": 1693843456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002795299791942501,
      "loss": 2.9096,
      "theoretical_loss": 3.4790258109227867,
      "tokens_seen": 1693908992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00027951342916587857,
      "loss": 2.754,
      "theoretical_loss": 3.479014218542561,
      "tokens_seen": 1693974528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002794968791375071,
      "loss": 2.6754,
      "theoretical_loss": 3.4790026267363796,
      "tokens_seen": 1694040064
    },
    {
      "epoch": 0.61,
      "objective/train/avg_token_score": 0.013803130015730858,
      "objective/train/avg_weight": 0.9837790727615356,
      "objective/train/docs_used": 958497,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0191261768341064,
      "objective/train/std_weight": 0.09993963688611984,
      "objective/train/theoretical_loss": 3.478991035504191,
      "objective/train/tokens_used": 1714565600,
      "theoretical_loss": 3.478991035504191,
      "tokens_seen": 1694105600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027948032910913564,
      "loss": 2.4399,
      "theoretical_loss": 3.478991035504191,
      "tokens_seen": 1694105600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002794637790807642,
      "loss": 2.7066,
      "theoretical_loss": 3.4789794448459443,
      "tokens_seen": 1694171136
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027944722905239266,
      "loss": 2.6078,
      "theoretical_loss": 3.4789678547615894,
      "tokens_seen": 1694236672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027943067902402115,
      "loss": 2.5088,
      "theoretical_loss": 3.478956265251076,
      "tokens_seen": 1694302208
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002794141289956497,
      "loss": 2.4594,
      "theoretical_loss": 3.478944676314353,
      "tokens_seen": 1694367744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002793975789672782,
      "loss": 2.7889,
      "theoretical_loss": 3.4789330879513694,
      "tokens_seen": 1694433280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027938102893890676,
      "loss": 2.2882,
      "theoretical_loss": 3.478921500162076,
      "tokens_seen": 1694498816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027936447891053524,
      "loss": 2.7996,
      "theoretical_loss": 3.4789099129464205,
      "tokens_seen": 1694564352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002793479288821638,
      "loss": 2.7402,
      "theoretical_loss": 3.4788983263043534,
      "tokens_seen": 1694629888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002793313788537923,
      "loss": 2.6777,
      "theoretical_loss": 3.478886740235824,
      "tokens_seen": 1694695424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027931482882542086,
      "loss": 2.761,
      "theoretical_loss": 3.4788751547407815,
      "tokens_seen": 1694760960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027929827879704934,
      "loss": 3.0236,
      "theoretical_loss": 3.478863569819176,
      "tokens_seen": 1694826496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002792817287686779,
      "loss": 2.6387,
      "theoretical_loss": 3.478851985470956,
      "tokens_seen": 1694892032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002792651787403064,
      "loss": 2.4426,
      "theoretical_loss": 3.478840401696071,
      "tokens_seen": 1694957568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027924862871193495,
      "loss": 2.773,
      "theoretical_loss": 3.4788288184944713,
      "tokens_seen": 1695023104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027923207868356343,
      "loss": 2.6058,
      "theoretical_loss": 3.4788172358661056,
      "tokens_seen": 1695088640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027921552865519197,
      "loss": 2.6822,
      "theoretical_loss": 3.478805653810924,
      "tokens_seen": 1695154176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002791989786268205,
      "loss": 2.699,
      "theoretical_loss": 3.478794072328875,
      "tokens_seen": 1695219712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027918242859844905,
      "loss": 2.4593,
      "theoretical_loss": 3.478782491419909,
      "tokens_seen": 1695285248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027916587857007753,
      "loss": 2.9418,
      "theoretical_loss": 3.4787709110839753,
      "tokens_seen": 1695350784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027914932854170607,
      "loss": 2.6337,
      "theoretical_loss": 3.478759331321023,
      "tokens_seen": 1695416320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002791327785133346,
      "loss": 2.9277,
      "theoretical_loss": 3.478747752131002,
      "tokens_seen": 1695481856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027911622848496314,
      "loss": 2.67,
      "theoretical_loss": 3.4787361735138624,
      "tokens_seen": 1695547392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002790996784565916,
      "loss": 2.7579,
      "theoretical_loss": 3.478724595469552,
      "tokens_seen": 1695612928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027908312842822016,
      "loss": 2.7153,
      "theoretical_loss": 3.4787130179980217,
      "tokens_seen": 1695678464
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002790665783998487,
      "loss": 2.7348,
      "theoretical_loss": 3.478701441099221,
      "tokens_seen": 1695744000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027905002837147724,
      "loss": 2.6345,
      "theoretical_loss": 3.4786898647730986,
      "tokens_seen": 1695809536
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002790334783431057,
      "loss": 2.8096,
      "theoretical_loss": 3.478678289019605,
      "tokens_seen": 1695875072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027901692831473426,
      "loss": 2.9222,
      "theoretical_loss": 3.478666713838689,
      "tokens_seen": 1695940608
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002790003782863628,
      "loss": 2.6921,
      "theoretical_loss": 3.478655139230301,
      "tokens_seen": 1696006144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027898382825799134,
      "loss": 2.6023,
      "theoretical_loss": 3.478643565194389,
      "tokens_seen": 1696071680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002789672782296198,
      "loss": 2.7962,
      "theoretical_loss": 3.478631991730904,
      "tokens_seen": 1696137216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002789507282012483,
      "loss": 2.4556,
      "theoretical_loss": 3.478620418839795,
      "tokens_seen": 1696202752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027893417817287684,
      "loss": 2.7432,
      "theoretical_loss": 3.4786088465210123,
      "tokens_seen": 1696268288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002789176281445054,
      "loss": 2.494,
      "theoretical_loss": 3.478597274774504,
      "tokens_seen": 1696333824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002789010781161339,
      "loss": 2.6093,
      "theoretical_loss": 3.4785857036002215,
      "tokens_seen": 1696399360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002788845280877624,
      "loss": 2.5536,
      "theoretical_loss": 3.478574132998113,
      "tokens_seen": 1696464896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027886797805939094,
      "loss": 2.6662,
      "theoretical_loss": 3.478562562968129,
      "tokens_seen": 1696530432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027885142803101947,
      "loss": 2.6546,
      "theoretical_loss": 3.4785509935102183,
      "tokens_seen": 1696595968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000278834878002648,
      "loss": 2.7415,
      "theoretical_loss": 3.4785394246243317,
      "tokens_seen": 1696661504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002788183279742765,
      "loss": 2.4167,
      "theoretical_loss": 3.478527856310417,
      "tokens_seen": 1696727040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027880177794590503,
      "loss": 2.5223,
      "theoretical_loss": 3.4785162885684255,
      "tokens_seen": 1696792576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027878522791753357,
      "loss": 2.5977,
      "theoretical_loss": 3.478504721398306,
      "tokens_seen": 1696858112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002787686778891621,
      "loss": 2.5875,
      "theoretical_loss": 3.4784931548000086,
      "tokens_seen": 1696923648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002787521278607906,
      "loss": 2.5701,
      "theoretical_loss": 3.4784815887734823,
      "tokens_seen": 1696989184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027873557783241913,
      "loss": 2.7751,
      "theoretical_loss": 3.4784700233186774,
      "tokens_seen": 1697054720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027871902780404766,
      "loss": 2.8852,
      "theoretical_loss": 3.4784584584355436,
      "tokens_seen": 1697120256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002787024777756762,
      "loss": 2.7969,
      "theoretical_loss": 3.47844689412403,
      "tokens_seen": 1697185792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002786859277473047,
      "loss": 2.7539,
      "theoretical_loss": 3.4784353303840865,
      "tokens_seen": 1697251328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002786693777189332,
      "loss": 2.5629,
      "theoretical_loss": 3.478423767215663,
      "tokens_seen": 1697316864
    },
    {
      "epoch": 0.61,
      "objective/train/avg_token_score": 0.01632816717028618,
      "objective/train/avg_weight": 0.9807875156402588,
      "objective/train/docs_used": 960356,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6844584941864014,
      "objective/train/std_weight": 0.08871375769376755,
      "objective/train/theoretical_loss": 3.478412204618709,
      "objective/train/tokens_used": 1717842400,
      "theoretical_loss": 3.478412204618709,
      "tokens_seen": 1697382400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027865282769056176,
      "loss": 2.4921,
      "theoretical_loss": 3.478412204618709,
      "tokens_seen": 1697382400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002786362776621903,
      "loss": 2.8932,
      "theoretical_loss": 3.4784006425931744,
      "tokens_seen": 1697447936
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002786197276338188,
      "loss": 2.6438,
      "theoretical_loss": 3.478389081139009,
      "tokens_seen": 1697513472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002786031776054473,
      "loss": 2.5167,
      "theoretical_loss": 3.4783775202561618,
      "tokens_seen": 1697579008
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027858662757707586,
      "loss": 2.696,
      "theoretical_loss": 3.4783659599445826,
      "tokens_seen": 1697644544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002785700775487044,
      "loss": 2.5834,
      "theoretical_loss": 3.478354400204222,
      "tokens_seen": 1697710080
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002785535275203329,
      "loss": 2.7103,
      "theoretical_loss": 3.4783428410350297,
      "tokens_seen": 1697775616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002785369774919614,
      "loss": 2.5658,
      "theoretical_loss": 3.4783312824369537,
      "tokens_seen": 1697841152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027852042746358995,
      "loss": 2.6934,
      "theoretical_loss": 3.478319724409946,
      "tokens_seen": 1697906688
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002785038774352185,
      "loss": 2.6431,
      "theoretical_loss": 3.478308166953955,
      "tokens_seen": 1697972224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000278487327406847,
      "loss": 2.615,
      "theoretical_loss": 3.4782966100689308,
      "tokens_seen": 1698037760
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002784707773784755,
      "loss": 2.6202,
      "theoretical_loss": 3.4782850537548233,
      "tokens_seen": 1698103296
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000278454227350104,
      "loss": 2.8672,
      "theoretical_loss": 3.478273498011582,
      "tokens_seen": 1698168832
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027843767732173253,
      "loss": 2.487,
      "theoretical_loss": 3.4782619428391564,
      "tokens_seen": 1698234368
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027842112729336107,
      "loss": 2.6634,
      "theoretical_loss": 3.478250388237497,
      "tokens_seen": 1698299904
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002784045772649896,
      "loss": 2.6681,
      "theoretical_loss": 3.478238834206553,
      "tokens_seen": 1698365440
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002783880272366181,
      "loss": 2.4588,
      "theoretical_loss": 3.4782272807462746,
      "tokens_seen": 1698430976
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027837147720824663,
      "loss": 2.5625,
      "theoretical_loss": 3.4782157278566115,
      "tokens_seen": 1698496512
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027835492717987517,
      "loss": 2.365,
      "theoretical_loss": 3.4782041755375133,
      "tokens_seen": 1698562048
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002783383771515037,
      "loss": 2.6835,
      "theoretical_loss": 3.47819262378893,
      "tokens_seen": 1698627584
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002783218271231322,
      "loss": 2.749,
      "theoretical_loss": 3.478181072610811,
      "tokens_seen": 1698693120
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002783052770947607,
      "loss": 2.5723,
      "theoretical_loss": 3.478169522003107,
      "tokens_seen": 1698758656
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027828872706638926,
      "loss": 2.5638,
      "theoretical_loss": 3.478157971965767,
      "tokens_seen": 1698824192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002782721770380178,
      "loss": 2.6888,
      "theoretical_loss": 3.4781464224987415,
      "tokens_seen": 1698889728
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002782556270096463,
      "loss": 2.3921,
      "theoretical_loss": 3.478134873601979,
      "tokens_seen": 1698955264
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002782390769812748,
      "loss": 2.5737,
      "theoretical_loss": 3.4781233252754316,
      "tokens_seen": 1699020800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027822252695290336,
      "loss": 2.6466,
      "theoretical_loss": 3.4781117775190467,
      "tokens_seen": 1699086336
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002782059769245319,
      "loss": 2.3432,
      "theoretical_loss": 3.4781002303327764,
      "tokens_seen": 1699151872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002781894268961604,
      "loss": 2.7549,
      "theoretical_loss": 3.4780886837165688,
      "tokens_seen": 1699217408
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002781728768677889,
      "loss": 2.6481,
      "theoretical_loss": 3.4780771376703745,
      "tokens_seen": 1699282944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027815632683941745,
      "loss": 2.6613,
      "theoretical_loss": 3.4780655921941435,
      "tokens_seen": 1699348480
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000278139776811046,
      "loss": 2.6737,
      "theoretical_loss": 3.4780540472878254,
      "tokens_seen": 1699414016
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002781232267826745,
      "loss": 2.701,
      "theoretical_loss": 3.47804250295137,
      "tokens_seen": 1699479552
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000278106676754303,
      "loss": 2.692,
      "theoretical_loss": 3.478030959184728,
      "tokens_seen": 1699545088
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027809012672593155,
      "loss": 2.7569,
      "theoretical_loss": 3.4780194159878484,
      "tokens_seen": 1699610624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002780735766975601,
      "loss": 2.6204,
      "theoretical_loss": 3.4780078733606814,
      "tokens_seen": 1699676160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027805702666918857,
      "loss": 2.2227,
      "theoretical_loss": 3.477996331303177,
      "tokens_seen": 1699741696
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002780404766408171,
      "loss": 2.704,
      "theoretical_loss": 3.4779847898152854,
      "tokens_seen": 1699807232
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027802392661244565,
      "loss": 2.7357,
      "theoretical_loss": 3.477973248896956,
      "tokens_seen": 1699872768
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002780073765840742,
      "loss": 2.5041,
      "theoretical_loss": 3.4779617085481385,
      "tokens_seen": 1699938304
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027799082655570267,
      "loss": 2.6412,
      "theoretical_loss": 3.4779501687687837,
      "tokens_seen": 1700003840
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027797427652733115,
      "loss": 2.1421,
      "theoretical_loss": 3.477938629558841,
      "tokens_seen": 1700069376
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002779577264989597,
      "loss": 2.7933,
      "theoretical_loss": 3.4779270909182607,
      "tokens_seen": 1700134912
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002779411764705882,
      "loss": 2.5084,
      "theoretical_loss": 3.4779155528469925,
      "tokens_seen": 1700200448
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027792462644221676,
      "loss": 2.4604,
      "theoretical_loss": 3.477904015344986,
      "tokens_seen": 1700265984
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027790807641384525,
      "loss": 2.3792,
      "theoretical_loss": 3.4778924784121914,
      "tokens_seen": 1700331520
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002778915263854738,
      "loss": 2.6001,
      "theoretical_loss": 3.47788094204856,
      "tokens_seen": 1700397056
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002778749763571023,
      "loss": 2.5308,
      "theoretical_loss": 3.4778694062540394,
      "tokens_seen": 1700462592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027785842632873086,
      "loss": 2.6773,
      "theoretical_loss": 3.4778578710285815,
      "tokens_seen": 1700528128
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027784187630035934,
      "loss": 2.5909,
      "theoretical_loss": 3.477846336372136,
      "tokens_seen": 1700593664
    },
    {
      "epoch": 0.61,
      "objective/train/avg_token_score": 0.0201057568192482,
      "objective/train/avg_weight": 0.9763284921646118,
      "objective/train/docs_used": 961887,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9306869506835938,
      "objective/train/std_weight": 0.08547782152891159,
      "objective/train/theoretical_loss": 3.4778348022846517,
      "objective/train/tokens_used": 1721119200,
      "theoretical_loss": 3.4778348022846517,
      "tokens_seen": 1700659200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002778253262719879,
      "loss": 2.3048,
      "theoretical_loss": 3.4778348022846517,
      "tokens_seen": 1700659200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002778087762436164,
      "loss": 2.3863,
      "theoretical_loss": 3.47782326876608,
      "tokens_seen": 1700724736
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027779222621524495,
      "loss": 2.4646,
      "theoretical_loss": 3.4778117358163705,
      "tokens_seen": 1700790272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027777567618687344,
      "loss": 2.727,
      "theoretical_loss": 3.4778002034354722,
      "tokens_seen": 1700855808
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000277759126158502,
      "loss": 2.8205,
      "theoretical_loss": 3.477788671623337,
      "tokens_seen": 1700921344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002777425761301305,
      "loss": 2.4586,
      "theoretical_loss": 3.4777771403799136,
      "tokens_seen": 1700986880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027772602610175905,
      "loss": 2.7006,
      "theoretical_loss": 3.4777656097051524,
      "tokens_seen": 1701052416
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027770947607338753,
      "loss": 2.7794,
      "theoretical_loss": 3.4777540795990034,
      "tokens_seen": 1701117952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027769292604501607,
      "loss": 2.6041,
      "theoretical_loss": 3.477742550061417,
      "tokens_seen": 1701183488
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002776763760166446,
      "loss": 2.4741,
      "theoretical_loss": 3.4777310210923424,
      "tokens_seen": 1701249024
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027765982598827315,
      "loss": 2.6649,
      "theoretical_loss": 3.4777194926917305,
      "tokens_seen": 1701314560
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027764327595990163,
      "loss": 2.532,
      "theoretical_loss": 3.4777079648595315,
      "tokens_seen": 1701380096
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027762672593153017,
      "loss": 2.1045,
      "theoretical_loss": 3.4776964375956947,
      "tokens_seen": 1701445632
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002776101759031587,
      "loss": 2.6275,
      "theoretical_loss": 3.477684910900171,
      "tokens_seen": 1701511168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027759362587478724,
      "loss": 2.7151,
      "theoretical_loss": 3.4776733847729098,
      "tokens_seen": 1701576704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002775770758464157,
      "loss": 2.8748,
      "theoretical_loss": 3.4776618592138613,
      "tokens_seen": 1701642240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027756052581804426,
      "loss": 2.5527,
      "theoretical_loss": 3.477650334222976,
      "tokens_seen": 1701707776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002775439757896728,
      "loss": 2.4511,
      "theoretical_loss": 3.477638809800204,
      "tokens_seen": 1701773312
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027752742576130134,
      "loss": 2.5398,
      "theoretical_loss": 3.477627285945495,
      "tokens_seen": 1701838848
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002775108757329298,
      "loss": 2.5775,
      "theoretical_loss": 3.4776157626587993,
      "tokens_seen": 1701904384
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027749432570455836,
      "loss": 2.7948,
      "theoretical_loss": 3.4776042399400673,
      "tokens_seen": 1701969920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027747777567618684,
      "loss": 2.6379,
      "theoretical_loss": 3.477592717789248,
      "tokens_seen": 1702035456
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002774612256478154,
      "loss": 2.6361,
      "theoretical_loss": 3.4775811962062937,
      "tokens_seen": 1702100992
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002774446756194439,
      "loss": 2.7517,
      "theoretical_loss": 3.4775696751911527,
      "tokens_seen": 1702166528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002774281255910724,
      "loss": 2.7369,
      "theoretical_loss": 3.4775581547437757,
      "tokens_seen": 1702232064
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027741157556270094,
      "loss": 2.7697,
      "theoretical_loss": 3.477546634864113,
      "tokens_seen": 1702297600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002773950255343295,
      "loss": 2.6629,
      "theoretical_loss": 3.477535115552115,
      "tokens_seen": 1702363136
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000277378475505958,
      "loss": 2.8302,
      "theoretical_loss": 3.4775235968077314,
      "tokens_seen": 1702428672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002773619254775865,
      "loss": 2.4171,
      "theoretical_loss": 3.4775120786309124,
      "tokens_seen": 1702494208
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027734537544921503,
      "loss": 2.6646,
      "theoretical_loss": 3.4775005610216088,
      "tokens_seen": 1702559744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027732882542084357,
      "loss": 2.2539,
      "theoretical_loss": 3.4774890439797694,
      "tokens_seen": 1702625280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002773122753924721,
      "loss": 2.6932,
      "theoretical_loss": 3.4774775275053464,
      "tokens_seen": 1702690816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002772957253641006,
      "loss": 2.7597,
      "theoretical_loss": 3.4774660115982883,
      "tokens_seen": 1702756352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027727917533572913,
      "loss": 2.7,
      "theoretical_loss": 3.477454496258546,
      "tokens_seen": 1702821888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027726262530735767,
      "loss": 2.6745,
      "theoretical_loss": 3.47744298148607,
      "tokens_seen": 1702887424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002772460752789862,
      "loss": 2.5545,
      "theoretical_loss": 3.4774314672808098,
      "tokens_seen": 1702952960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002772295252506147,
      "loss": 2.9158,
      "theoretical_loss": 3.477419953642716,
      "tokens_seen": 1703018496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002772129752222432,
      "loss": 2.5914,
      "theoretical_loss": 3.477408440571739,
      "tokens_seen": 1703084032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027719642519387176,
      "loss": 2.4898,
      "theoretical_loss": 3.4773969280678294,
      "tokens_seen": 1703149568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002771798751655003,
      "loss": 2.7967,
      "theoretical_loss": 3.4773854161309363,
      "tokens_seen": 1703215104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002771633251371288,
      "loss": 2.791,
      "theoretical_loss": 3.477373904761011,
      "tokens_seen": 1703280640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002771467751087573,
      "loss": 2.5133,
      "theoretical_loss": 3.4773623939580034,
      "tokens_seen": 1703346176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027713022508038586,
      "loss": 2.632,
      "theoretical_loss": 3.477350883721863,
      "tokens_seen": 1703411712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002771136750520144,
      "loss": 2.7304,
      "theoretical_loss": 3.477339374052542,
      "tokens_seen": 1703477248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002770971250236429,
      "loss": 2.7257,
      "theoretical_loss": 3.477327864949989,
      "tokens_seen": 1703542784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002770805749952714,
      "loss": 2.7393,
      "theoretical_loss": 3.4773163564141547,
      "tokens_seen": 1703608320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027706402496689996,
      "loss": 2.589,
      "theoretical_loss": 3.4773048484449895,
      "tokens_seen": 1703673856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002770474749385285,
      "loss": 2.6797,
      "theoretical_loss": 3.4772933410424436,
      "tokens_seen": 1703739392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000277030924910157,
      "loss": 2.7162,
      "theoretical_loss": 3.477281834206468,
      "tokens_seen": 1703804928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002770143748817855,
      "loss": 2.7234,
      "theoretical_loss": 3.4772703279370116,
      "tokens_seen": 1703870464
    },
    {
      "epoch": 0.61,
      "objective/train/avg_token_score": 0.02506786584854126,
      "objective/train/avg_weight": 0.9704902172088623,
      "objective/train/docs_used": 963663,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8932735919952393,
      "objective/train/std_weight": 0.0963016226887703,
      "objective/train/theoretical_loss": 3.477258822234026,
      "objective/train/tokens_used": 1724396000,
      "theoretical_loss": 3.477258822234026,
      "tokens_seen": 1703936000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000276997824853414,
      "loss": 2.8331,
      "theoretical_loss": 3.477258822234026,
      "tokens_seen": 1703936000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027698127482504253,
      "loss": 2.5132,
      "theoretical_loss": 3.4772473170974614,
      "tokens_seen": 1704001536
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027696472479667107,
      "loss": 2.3706,
      "theoretical_loss": 3.477235812527267,
      "tokens_seen": 1704067072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002769481747682996,
      "loss": 2.3228,
      "theoretical_loss": 3.4772243085233945,
      "tokens_seen": 1704132608
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002769316247399281,
      "loss": 2.7881,
      "theoretical_loss": 3.4772128050857933,
      "tokens_seen": 1704198144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027691507471155663,
      "loss": 2.7268,
      "theoretical_loss": 3.4772013022144144,
      "tokens_seen": 1704263680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027689852468318517,
      "loss": 2.3812,
      "theoretical_loss": 3.4771897999092083,
      "tokens_seen": 1704329216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002768819746548137,
      "loss": 2.7581,
      "theoretical_loss": 3.4771782981701245,
      "tokens_seen": 1704394752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002768654246264422,
      "loss": 2.5734,
      "theoretical_loss": 3.4771667969971136,
      "tokens_seen": 1704460288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002768488745980707,
      "loss": 2.736,
      "theoretical_loss": 3.4771552963901264,
      "tokens_seen": 1704525824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027683232456969926,
      "loss": 2.7239,
      "theoretical_loss": 3.477143796349113,
      "tokens_seen": 1704591360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002768157745413278,
      "loss": 2.6785,
      "theoretical_loss": 3.477132296874024,
      "tokens_seen": 1704656896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002767992245129563,
      "loss": 2.4212,
      "theoretical_loss": 3.47712079796481,
      "tokens_seen": 1704722432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002767826744845848,
      "loss": 2.5041,
      "theoretical_loss": 3.4771092996214206,
      "tokens_seen": 1704787968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027676612445621336,
      "loss": 2.7398,
      "theoretical_loss": 3.4770978018438066,
      "tokens_seen": 1704853504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002767495744278419,
      "loss": 2.6221,
      "theoretical_loss": 3.4770863046319187,
      "tokens_seen": 1704919040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002767330243994704,
      "loss": 2.7956,
      "theoretical_loss": 3.4770748079857072,
      "tokens_seen": 1704984576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002767164743710989,
      "loss": 2.3995,
      "theoretical_loss": 3.4770633119051224,
      "tokens_seen": 1705050112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027669992434272746,
      "loss": 2.6763,
      "theoretical_loss": 3.4770518163901145,
      "tokens_seen": 1705115648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000276683374314356,
      "loss": 2.8128,
      "theoretical_loss": 3.4770403214406347,
      "tokens_seen": 1705181184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002766668242859845,
      "loss": 2.5746,
      "theoretical_loss": 3.4770288270566327,
      "tokens_seen": 1705246720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000276650274257613,
      "loss": 2.6125,
      "theoretical_loss": 3.477017333238059,
      "tokens_seen": 1705312256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027663372422924155,
      "loss": 2.6267,
      "theoretical_loss": 3.4770058399848645,
      "tokens_seen": 1705377792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002766171742008701,
      "loss": 2.724,
      "theoretical_loss": 3.476994347296999,
      "tokens_seen": 1705443328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002766006241724986,
      "loss": 2.6505,
      "theoretical_loss": 3.476982855174414,
      "tokens_seen": 1705508864
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002765840741441271,
      "loss": 2.8419,
      "theoretical_loss": 3.476971363617059,
      "tokens_seen": 1705574400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027656752411575565,
      "loss": 2.6794,
      "theoretical_loss": 3.476959872624885,
      "tokens_seen": 1705639936
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002765509740873842,
      "loss": 2.5198,
      "theoretical_loss": 3.4769483821978424,
      "tokens_seen": 1705705472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027653442405901267,
      "loss": 2.6878,
      "theoretical_loss": 3.4769368923358814,
      "tokens_seen": 1705771008
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002765178740306412,
      "loss": 2.4284,
      "theoretical_loss": 3.476925403038953,
      "tokens_seen": 1705836544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002765013240022697,
      "loss": 2.7043,
      "theoretical_loss": 3.4769139143070076,
      "tokens_seen": 1705902080
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027648477397389823,
      "loss": 2.5611,
      "theoretical_loss": 3.476902426139995,
      "tokens_seen": 1705967616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027646822394552676,
      "loss": 2.425,
      "theoretical_loss": 3.4768909385378666,
      "tokens_seen": 1706033152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027645167391715525,
      "loss": 2.5182,
      "theoretical_loss": 3.4768794515005723,
      "tokens_seen": 1706098688
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002764351238887838,
      "loss": 2.5724,
      "theoretical_loss": 3.476867965028063,
      "tokens_seen": 1706164224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002764185738604123,
      "loss": 2.3063,
      "theoretical_loss": 3.47685647912029,
      "tokens_seen": 1706229760
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027640202383204086,
      "loss": 2.5955,
      "theoretical_loss": 3.476844993777202,
      "tokens_seen": 1706295296
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027638547380366934,
      "loss": 2.38,
      "theoretical_loss": 3.476833508998751,
      "tokens_seen": 1706360832
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002763689237752979,
      "loss": 2.598,
      "theoretical_loss": 3.4768220247848873,
      "tokens_seen": 1706426368
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002763523737469264,
      "loss": 2.5379,
      "theoretical_loss": 3.4768105411355608,
      "tokens_seen": 1706491904
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027633582371855496,
      "loss": 2.7362,
      "theoretical_loss": 3.476799058050723,
      "tokens_seen": 1706557440
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027631927369018344,
      "loss": 2.7046,
      "theoretical_loss": 3.4767875755303237,
      "tokens_seen": 1706622976
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000276302723661812,
      "loss": 2.5955,
      "theoretical_loss": 3.476776093574314,
      "tokens_seen": 1706688512
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002762861736334405,
      "loss": 2.5572,
      "theoretical_loss": 3.4767646121826443,
      "tokens_seen": 1706754048
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027626962360506905,
      "loss": 2.5761,
      "theoretical_loss": 3.4767531313552658,
      "tokens_seen": 1706819584
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027625307357669754,
      "loss": 2.6504,
      "theoretical_loss": 3.4767416510921274,
      "tokens_seen": 1706885120
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002762365235483261,
      "loss": 2.3459,
      "theoretical_loss": 3.4767301713931813,
      "tokens_seen": 1706950656
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002762199735199546,
      "loss": 2.3242,
      "theoretical_loss": 3.4767186922583777,
      "tokens_seen": 1707016192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027620342349158315,
      "loss": 2.5059,
      "theoretical_loss": 3.4767072136876673,
      "tokens_seen": 1707081728
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027618687346321163,
      "loss": 2.5586,
      "theoretical_loss": 3.4766957356810004,
      "tokens_seen": 1707147264
    },
    {
      "epoch": 0.61,
      "objective/train/avg_token_score": 0.022932233288884163,
      "objective/train/avg_weight": 0.9730276465415955,
      "objective/train/docs_used": 965784,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8175392150878906,
      "objective/train/std_weight": 0.09276441484689713,
      "objective/train/theoretical_loss": 3.4766842582383277,
      "objective/train/tokens_used": 1727672800,
      "theoretical_loss": 3.4766842582383277,
      "tokens_seen": 1707212800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027617032343484017,
      "loss": 2.3551,
      "theoretical_loss": 3.4766842582383277,
      "tokens_seen": 1707212800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002761537734064687,
      "loss": 2.6071,
      "theoretical_loss": 3.4766727813596,
      "tokens_seen": 1707278336
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027613722337809724,
      "loss": 2.6786,
      "theoretical_loss": 3.476661305044768,
      "tokens_seen": 1707343872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027612067334972573,
      "loss": 2.7801,
      "theoretical_loss": 3.476649829293782,
      "tokens_seen": 1707409408
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027610412332135427,
      "loss": 2.3479,
      "theoretical_loss": 3.4766383541065933,
      "tokens_seen": 1707474944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002760875732929828,
      "loss": 2.6295,
      "theoretical_loss": 3.4766268794831516,
      "tokens_seen": 1707540480
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027607102326461134,
      "loss": 2.3606,
      "theoretical_loss": 3.4766154054234084,
      "tokens_seen": 1707606016
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002760544732362398,
      "loss": 2.7442,
      "theoretical_loss": 3.4766039319273143,
      "tokens_seen": 1707671552
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027603792320786836,
      "loss": 2.6118,
      "theoretical_loss": 3.4765924589948196,
      "tokens_seen": 1707737088
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002760213731794969,
      "loss": 2.8302,
      "theoretical_loss": 3.4765809866258754,
      "tokens_seen": 1707802624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002760048231511254,
      "loss": 2.4884,
      "theoretical_loss": 3.476569514820432,
      "tokens_seen": 1707868160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002759882731227539,
      "loss": 2.4723,
      "theoretical_loss": 3.4765580435784402,
      "tokens_seen": 1707933696
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002759717230943824,
      "loss": 2.7788,
      "theoretical_loss": 3.4765465728998506,
      "tokens_seen": 1707999232
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027595517306601094,
      "loss": 2.551,
      "theoretical_loss": 3.476535102784614,
      "tokens_seen": 1708064768
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002759386230376395,
      "loss": 2.5748,
      "theoretical_loss": 3.4765236332326817,
      "tokens_seen": 1708130304
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000275922073009268,
      "loss": 2.6319,
      "theoretical_loss": 3.4765121642440038,
      "tokens_seen": 1708195840
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002759055229808965,
      "loss": 2.4584,
      "theoretical_loss": 3.476500695818531,
      "tokens_seen": 1708261376
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027588897295252504,
      "loss": 2.3798,
      "theoretical_loss": 3.476489227956214,
      "tokens_seen": 1708326912
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002758724229241536,
      "loss": 2.3866,
      "theoretical_loss": 3.476477760657004,
      "tokens_seen": 1708392448
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002758558728957821,
      "loss": 2.4774,
      "theoretical_loss": 3.4764662939208515,
      "tokens_seen": 1708457984
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002758393228674106,
      "loss": 2.6077,
      "theoretical_loss": 3.476454827747707,
      "tokens_seen": 1708523520
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027582277283903913,
      "loss": 2.4218,
      "theoretical_loss": 3.476443362137522,
      "tokens_seen": 1708589056
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027580622281066767,
      "loss": 2.5428,
      "theoretical_loss": 3.4764318970902464,
      "tokens_seen": 1708654592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002757896727822962,
      "loss": 2.3714,
      "theoretical_loss": 3.4764204326058317,
      "tokens_seen": 1708720128
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002757731227539247,
      "loss": 2.4071,
      "theoretical_loss": 3.476408968684228,
      "tokens_seen": 1708785664
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027575657272555323,
      "loss": 2.6647,
      "theoretical_loss": 3.4763975053253864,
      "tokens_seen": 1708851200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027574002269718177,
      "loss": 2.4919,
      "theoretical_loss": 3.4763860425292576,
      "tokens_seen": 1708916736
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002757234726688103,
      "loss": 2.483,
      "theoretical_loss": 3.4763745802957926,
      "tokens_seen": 1708982272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027570692264043884,
      "loss": 2.5744,
      "theoretical_loss": 3.476363118624943,
      "tokens_seen": 1709047808
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002756903726120673,
      "loss": 2.5833,
      "theoretical_loss": 3.4763516575166573,
      "tokens_seen": 1709113344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027567382258369586,
      "loss": 2.7465,
      "theoretical_loss": 3.476340196970889,
      "tokens_seen": 1709178880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002756572725553244,
      "loss": 2.664,
      "theoretical_loss": 3.476328736987587,
      "tokens_seen": 1709244416
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027564072252695294,
      "loss": 2.6168,
      "theoretical_loss": 3.476317277566703,
      "tokens_seen": 1709309952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002756241724985814,
      "loss": 2.6414,
      "theoretical_loss": 3.4763058187081874,
      "tokens_seen": 1709375488
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027560762247020996,
      "loss": 2.6557,
      "theoretical_loss": 3.476294360411991,
      "tokens_seen": 1709441024
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002755910724418385,
      "loss": 2.3603,
      "theoretical_loss": 3.4762829026780655,
      "tokens_seen": 1709506560
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027557452241346703,
      "loss": 2.5622,
      "theoretical_loss": 3.4762714455063612,
      "tokens_seen": 1709572096
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002755579723850955,
      "loss": 2.8375,
      "theoretical_loss": 3.4762599888968286,
      "tokens_seen": 1709637632
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027554142235672405,
      "loss": 2.6421,
      "theoretical_loss": 3.476248532849419,
      "tokens_seen": 1709703168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027552487232835254,
      "loss": 2.7829,
      "theoretical_loss": 3.4762370773640834,
      "tokens_seen": 1709768704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002755083222999811,
      "loss": 2.7844,
      "theoretical_loss": 3.476225622440772,
      "tokens_seen": 1709834240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002754917722716096,
      "loss": 2.5341,
      "theoretical_loss": 3.4762141680794367,
      "tokens_seen": 1709899776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002754752222432381,
      "loss": 2.5884,
      "theoretical_loss": 3.476202714280028,
      "tokens_seen": 1709965312
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027545867221486663,
      "loss": 2.6846,
      "theoretical_loss": 3.476191261042496,
      "tokens_seen": 1710030848
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027544212218649517,
      "loss": 2.6168,
      "theoretical_loss": 3.4761798083667923,
      "tokens_seen": 1710096384
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002754255721581237,
      "loss": 2.6347,
      "theoretical_loss": 3.476168356252868,
      "tokens_seen": 1710161920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002754090221297522,
      "loss": 2.6855,
      "theoretical_loss": 3.476156904700674,
      "tokens_seen": 1710227456
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027539247210138073,
      "loss": 2.6073,
      "theoretical_loss": 3.476145453710161,
      "tokens_seen": 1710292992
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027537592207300927,
      "loss": 2.6115,
      "theoretical_loss": 3.4761340032812793,
      "tokens_seen": 1710358528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002753593720446378,
      "loss": 2.1588,
      "theoretical_loss": 3.4761225534139806,
      "tokens_seen": 1710424064
    },
    {
      "epoch": 0.61,
      "objective/train/avg_token_score": 0.007394811604171991,
      "objective/train/avg_weight": 0.9912950396537781,
      "objective/train/docs_used": 967686,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0445175170898438,
      "objective/train/std_weight": 0.053407568484544754,
      "objective/train/theoretical_loss": 3.476111104108216,
      "objective/train/tokens_used": 1730949600,
      "theoretical_loss": 3.476111104108216,
      "tokens_seen": 1710489600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002753428220162663,
      "loss": 2.388,
      "theoretical_loss": 3.476111104108216,
      "tokens_seen": 1710489600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002753262719878948,
      "loss": 2.467,
      "theoretical_loss": 3.476099655363936,
      "tokens_seen": 1710555136
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027530972195952336,
      "loss": 2.371,
      "theoretical_loss": 3.476088207181092,
      "tokens_seen": 1710620672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002752931719311519,
      "loss": 2.9346,
      "theoretical_loss": 3.4760767595596347,
      "tokens_seen": 1710686208
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002752766219027804,
      "loss": 2.6861,
      "theoretical_loss": 3.4760653124995144,
      "tokens_seen": 1710751744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002752600718744089,
      "loss": 2.6455,
      "theoretical_loss": 3.4760538660006834,
      "tokens_seen": 1710817280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027524352184603746,
      "loss": 2.6968,
      "theoretical_loss": 3.4760424200630915,
      "tokens_seen": 1710882816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000275226971817666,
      "loss": 2.5651,
      "theoretical_loss": 3.4760309746866906,
      "tokens_seen": 1710948352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002752104217892945,
      "loss": 2.6618,
      "theoretical_loss": 3.476019529871431,
      "tokens_seen": 1711013888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000275193871760923,
      "loss": 2.538,
      "theoretical_loss": 3.4760080856172637,
      "tokens_seen": 1711079424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027517732173255155,
      "loss": 2.6513,
      "theoretical_loss": 3.475996641924141,
      "tokens_seen": 1711144960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002751607717041801,
      "loss": 2.7284,
      "theoretical_loss": 3.475985198792012,
      "tokens_seen": 1711210496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002751442216758086,
      "loss": 2.7887,
      "theoretical_loss": 3.4759737562208284,
      "tokens_seen": 1711276032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002751276716474371,
      "loss": 2.6883,
      "theoretical_loss": 3.4759623142105416,
      "tokens_seen": 1711341568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027511112161906565,
      "loss": 2.6021,
      "theoretical_loss": 3.4759508727611026,
      "tokens_seen": 1711407104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002750945715906942,
      "loss": 2.6894,
      "theoretical_loss": 3.475939431872462,
      "tokens_seen": 1711472640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027507802156232267,
      "loss": 2.5718,
      "theoretical_loss": 3.4759279915445713,
      "tokens_seen": 1711538176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002750614715339512,
      "loss": 2.618,
      "theoretical_loss": 3.4759165517773813,
      "tokens_seen": 1711603712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027504492150557975,
      "loss": 2.4664,
      "theoretical_loss": 3.4759051125708433,
      "tokens_seen": 1711669248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027502837147720823,
      "loss": 2.7352,
      "theoretical_loss": 3.475893673924908,
      "tokens_seen": 1711734784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027501182144883677,
      "loss": 2.3719,
      "theoretical_loss": 3.475882235839527,
      "tokens_seen": 1711800320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027499527142046525,
      "loss": 2.5676,
      "theoretical_loss": 3.4758707983146504,
      "tokens_seen": 1711865856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002749787213920938,
      "loss": 2.6154,
      "theoretical_loss": 3.47585936135023,
      "tokens_seen": 1711931392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002749621713637223,
      "loss": 2.5448,
      "theoretical_loss": 3.475847924946217,
      "tokens_seen": 1711996928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027494562133535086,
      "loss": 2.4584,
      "theoretical_loss": 3.475836489102562,
      "tokens_seen": 1712062464
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027492907130697935,
      "loss": 2.8407,
      "theoretical_loss": 3.4758250538192166,
      "tokens_seen": 1712128000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002749125212786079,
      "loss": 2.289,
      "theoretical_loss": 3.4758136190961313,
      "tokens_seen": 1712193536
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002748959712502364,
      "loss": 2.5536,
      "theoretical_loss": 3.475802184933258,
      "tokens_seen": 1712259072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027487942122186496,
      "loss": 2.4512,
      "theoretical_loss": 3.475790751330547,
      "tokens_seen": 1712324608
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027486287119349344,
      "loss": 2.6646,
      "theoretical_loss": 3.47577931828795,
      "tokens_seen": 1712390144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000274846321165122,
      "loss": 2.3402,
      "theoretical_loss": 3.4757678858054177,
      "tokens_seen": 1712455680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002748297711367505,
      "loss": 2.6897,
      "theoretical_loss": 3.4757564538829016,
      "tokens_seen": 1712521216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027481322110837906,
      "loss": 2.6684,
      "theoretical_loss": 3.4757450225203526,
      "tokens_seen": 1712586752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027479667108000754,
      "loss": 2.7451,
      "theoretical_loss": 3.475733591717722,
      "tokens_seen": 1712652288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002747801210516361,
      "loss": 2.5861,
      "theoretical_loss": 3.4757221614749607,
      "tokens_seen": 1712717824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002747635710232646,
      "loss": 2.6151,
      "theoretical_loss": 3.47571073179202,
      "tokens_seen": 1712783360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027474702099489315,
      "loss": 2.624,
      "theoretical_loss": 3.4756993026688514,
      "tokens_seen": 1712848896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027473047096652163,
      "loss": 2.148,
      "theoretical_loss": 3.475687874105406,
      "tokens_seen": 1712914432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027471392093815017,
      "loss": 2.456,
      "theoretical_loss": 3.4756764461016347,
      "tokens_seen": 1712979968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002746973709097787,
      "loss": 2.5639,
      "theoretical_loss": 3.4756650186574882,
      "tokens_seen": 1713045504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027468082088140725,
      "loss": 2.7239,
      "theoretical_loss": 3.475653591772918,
      "tokens_seen": 1713111040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027466427085303573,
      "loss": 2.4376,
      "theoretical_loss": 3.4756421654478764,
      "tokens_seen": 1713176576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027464772082466427,
      "loss": 2.7779,
      "theoretical_loss": 3.475630739682313,
      "tokens_seen": 1713242112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002746311707962928,
      "loss": 2.3722,
      "theoretical_loss": 3.47561931447618,
      "tokens_seen": 1713307648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027461462076792134,
      "loss": 2.695,
      "theoretical_loss": 3.475607889829428,
      "tokens_seen": 1713373184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002745980707395498,
      "loss": 2.6237,
      "theoretical_loss": 3.475596465742009,
      "tokens_seen": 1713438720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027458152071117836,
      "loss": 2.7532,
      "theoretical_loss": 3.475585042213874,
      "tokens_seen": 1713504256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002745649706828069,
      "loss": 2.4507,
      "theoretical_loss": 3.4755736192449733,
      "tokens_seen": 1713569792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027454842065443544,
      "loss": 2.4601,
      "theoretical_loss": 3.4755621968352592,
      "tokens_seen": 1713635328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002745318706260639,
      "loss": 2.659,
      "theoretical_loss": 3.475550774984682,
      "tokens_seen": 1713700864
    },
    {
      "epoch": 0.61,
      "objective/train/avg_token_score": 0.017948132008314133,
      "objective/train/avg_weight": 0.9788703918457031,
      "objective/train/docs_used": 969481,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.717637300491333,
      "objective/train/std_weight": 0.07632562518119812,
      "objective/train/theoretical_loss": 3.4755393536931942,
      "objective/train/tokens_used": 1734226400,
      "theoretical_loss": 3.4755393536931942,
      "tokens_seen": 1713766400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002745153205976924,
      "loss": 2.7116,
      "theoretical_loss": 3.4755393536931942,
      "tokens_seen": 1713766400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027449877056932094,
      "loss": 2.5564,
      "theoretical_loss": 3.4755279329607465,
      "tokens_seen": 1713831936
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002744822205409495,
      "loss": 2.5443,
      "theoretical_loss": 3.4755165127872893,
      "tokens_seen": 1713897472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000274465670512578,
      "loss": 2.5247,
      "theoretical_loss": 3.4755050931727753,
      "tokens_seen": 1713963008
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027444912048420656,
      "loss": 2.6071,
      "theoretical_loss": 3.4754936741171543,
      "tokens_seen": 1714028544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027443257045583504,
      "loss": 2.5494,
      "theoretical_loss": 3.475482255620379,
      "tokens_seen": 1714094080
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002744160204274636,
      "loss": 2.4776,
      "theoretical_loss": 3.4754708376823995,
      "tokens_seen": 1714159616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002743994703990921,
      "loss": 2.8846,
      "theoretical_loss": 3.475459420303168,
      "tokens_seen": 1714225152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027438292037072065,
      "loss": 2.4372,
      "theoretical_loss": 3.4754480034826356,
      "tokens_seen": 1714290688
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027436637034234914,
      "loss": 2.6783,
      "theoretical_loss": 3.475436587220753,
      "tokens_seen": 1714356224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002743498203139777,
      "loss": 2.5844,
      "theoretical_loss": 3.4754251715174727,
      "tokens_seen": 1714421760
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002743332702856062,
      "loss": 2.4396,
      "theoretical_loss": 3.4754137563727445,
      "tokens_seen": 1714487296
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027431672025723475,
      "loss": 2.5564,
      "theoretical_loss": 3.4754023417865207,
      "tokens_seen": 1714552832
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027430017022886323,
      "loss": 2.6125,
      "theoretical_loss": 3.4753909277587525,
      "tokens_seen": 1714618368
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027428362020049177,
      "loss": 2.3539,
      "theoretical_loss": 3.475379514289391,
      "tokens_seen": 1714683904
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002742670701721203,
      "loss": 2.5997,
      "theoretical_loss": 3.475368101378388,
      "tokens_seen": 1714749440
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027425052014374884,
      "loss": 2.6585,
      "theoretical_loss": 3.475356689025694,
      "tokens_seen": 1714814976
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027423397011537733,
      "loss": 2.6209,
      "theoretical_loss": 3.4753452772312614,
      "tokens_seen": 1714880512
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027421742008700587,
      "loss": 2.7586,
      "theoretical_loss": 3.475333865995041,
      "tokens_seen": 1714946048
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002742008700586344,
      "loss": 2.5611,
      "theoretical_loss": 3.4753224553169835,
      "tokens_seen": 1715011584
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027418432003026294,
      "loss": 2.7514,
      "theoretical_loss": 3.4753110451970417,
      "tokens_seen": 1715077120
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002741677700018914,
      "loss": 2.581,
      "theoretical_loss": 3.475299635635166,
      "tokens_seen": 1715142656
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027415121997351996,
      "loss": 2.7169,
      "theoretical_loss": 3.475288226631308,
      "tokens_seen": 1715208192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002741346699451485,
      "loss": 2.6315,
      "theoretical_loss": 3.475276818185419,
      "tokens_seen": 1715273728
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027411811991677704,
      "loss": 2.5382,
      "theoretical_loss": 3.475265410297451,
      "tokens_seen": 1715339264
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002741015698884055,
      "loss": 2.4985,
      "theoretical_loss": 3.4752540029673544,
      "tokens_seen": 1715404800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027408501986003406,
      "loss": 2.7188,
      "theoretical_loss": 3.4752425961950815,
      "tokens_seen": 1715470336
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002740684698316626,
      "loss": 2.7681,
      "theoretical_loss": 3.4752311899805832,
      "tokens_seen": 1715535872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002740519198032911,
      "loss": 2.301,
      "theoretical_loss": 3.475219784323811,
      "tokens_seen": 1715601408
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002740353697749196,
      "loss": 2.507,
      "theoretical_loss": 3.475208379224716,
      "tokens_seen": 1715666944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002740188197465481,
      "loss": 2.751,
      "theoretical_loss": 3.475196974683251,
      "tokens_seen": 1715732480
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027400226971817664,
      "loss": 2.6746,
      "theoretical_loss": 3.4751855706993657,
      "tokens_seen": 1715798016
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002739857196898052,
      "loss": 2.7269,
      "theoretical_loss": 3.4751741672730123,
      "tokens_seen": 1715863552
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002739691696614337,
      "loss": 2.6598,
      "theoretical_loss": 3.4751627644041427,
      "tokens_seen": 1715929088
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002739526196330622,
      "loss": 2.7812,
      "theoretical_loss": 3.4751513620927073,
      "tokens_seen": 1715994624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027393606960469073,
      "loss": 2.6504,
      "theoretical_loss": 3.475139960338659,
      "tokens_seen": 1716060160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027391951957631927,
      "loss": 2.837,
      "theoretical_loss": 3.4751285591419476,
      "tokens_seen": 1716125696
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002739029695479478,
      "loss": 2.5524,
      "theoretical_loss": 3.475117158502526,
      "tokens_seen": 1716191232
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002738864195195763,
      "loss": 2.6819,
      "theoretical_loss": 3.4751057584203444,
      "tokens_seen": 1716256768
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027386986949120483,
      "loss": 2.7185,
      "theoretical_loss": 3.4750943588953556,
      "tokens_seen": 1716322304
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027385331946283337,
      "loss": 2.6916,
      "theoretical_loss": 3.47508295992751,
      "tokens_seen": 1716387840
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002738367694344619,
      "loss": 2.2404,
      "theoretical_loss": 3.47507156151676,
      "tokens_seen": 1716453376
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002738202194060904,
      "loss": 2.5543,
      "theoretical_loss": 3.4750601636630565,
      "tokens_seen": 1716518912
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002738036693777189,
      "loss": 2.5201,
      "theoretical_loss": 3.475048766366351,
      "tokens_seen": 1716584448
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027378711934934746,
      "loss": 2.7578,
      "theoretical_loss": 3.4750373696265955,
      "tokens_seen": 1716649984
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000273770569320976,
      "loss": 2.7984,
      "theoretical_loss": 3.4750259734437408,
      "tokens_seen": 1716715520
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002737540192926045,
      "loss": 2.5132,
      "theoretical_loss": 3.4750145778177393,
      "tokens_seen": 1716781056
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000273737469264233,
      "loss": 2.4717,
      "theoretical_loss": 3.4750031827485417,
      "tokens_seen": 1716846592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027372091923586156,
      "loss": 2.9293,
      "theoretical_loss": 3.4749917882360997,
      "tokens_seen": 1716912128
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002737043692074901,
      "loss": 2.629,
      "theoretical_loss": 3.4749803942803656,
      "tokens_seen": 1716977664
    },
    {
      "epoch": 0.61,
      "objective/train/avg_token_score": 0.033936116844415665,
      "objective/train/avg_weight": 0.9600928425788879,
      "objective/train/docs_used": 970822,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.164966106414795,
      "objective/train/std_weight": 0.12184659391641617,
      "objective/train/theoretical_loss": 3.4749690008812903,
      "objective/train/tokens_used": 1737503200,
      "theoretical_loss": 3.4749690008812903,
      "tokens_seen": 1717043200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002736878191791186,
      "loss": 2.633,
      "theoretical_loss": 3.4749690008812903,
      "tokens_seen": 1717043200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002736712691507471,
      "loss": 2.74,
      "theoretical_loss": 3.4749576080388254,
      "tokens_seen": 1717108736
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027365471912237565,
      "loss": 2.7275,
      "theoretical_loss": 3.4749462157529227,
      "tokens_seen": 1717174272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002736381690940042,
      "loss": 2.6668,
      "theoretical_loss": 3.4749348240235336,
      "tokens_seen": 1717239808
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002736216190656327,
      "loss": 2.6916,
      "theoretical_loss": 3.47492343285061,
      "tokens_seen": 1717305344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002736050690372612,
      "loss": 2.6639,
      "theoretical_loss": 3.474912042234102,
      "tokens_seen": 1717370880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027358851900888975,
      "loss": 2.4904,
      "theoretical_loss": 3.4749006521739636,
      "tokens_seen": 1717436416
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002735719689805183,
      "loss": 2.5733,
      "theoretical_loss": 3.4748892626701444,
      "tokens_seen": 1717501952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027355541895214677,
      "loss": 2.749,
      "theoretical_loss": 3.4748778737225976,
      "tokens_seen": 1717567488
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027353886892377525,
      "loss": 2.5994,
      "theoretical_loss": 3.4748664853312734,
      "tokens_seen": 1717633024
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002735223188954038,
      "loss": 2.547,
      "theoretical_loss": 3.474855097496124,
      "tokens_seen": 1717698560
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027350576886703233,
      "loss": 2.6272,
      "theoretical_loss": 3.474843710217101,
      "tokens_seen": 1717764096
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027348921883866087,
      "loss": 2.4773,
      "theoretical_loss": 3.474832323494156,
      "tokens_seen": 1717829632
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027347266881028935,
      "loss": 2.9004,
      "theoretical_loss": 3.474820937327241,
      "tokens_seen": 1717895168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002734561187819179,
      "loss": 2.4638,
      "theoretical_loss": 3.474809551716307,
      "tokens_seen": 1717960704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002734395687535464,
      "loss": 2.6944,
      "theoretical_loss": 3.474798166661306,
      "tokens_seen": 1718026240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027342301872517496,
      "loss": 2.7226,
      "theoretical_loss": 3.4747867821621896,
      "tokens_seen": 1718091776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027340646869680345,
      "loss": 2.0654,
      "theoretical_loss": 3.4747753982189096,
      "tokens_seen": 1718157312
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000273389918668432,
      "loss": 2.5379,
      "theoretical_loss": 3.474764014831417,
      "tokens_seen": 1718222848
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002733733686400605,
      "loss": 2.6003,
      "theoretical_loss": 3.4747526319996647,
      "tokens_seen": 1718288384
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027335681861168906,
      "loss": 2.6596,
      "theoretical_loss": 3.474741249723603,
      "tokens_seen": 1718353920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027334026858331754,
      "loss": 2.4706,
      "theoretical_loss": 3.4747298680031844,
      "tokens_seen": 1718419456
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002733237185549461,
      "loss": 2.7454,
      "theoretical_loss": 3.4747184868383605,
      "tokens_seen": 1718484992
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002733071685265746,
      "loss": 2.5615,
      "theoretical_loss": 3.474707106229083,
      "tokens_seen": 1718550528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027329061849820315,
      "loss": 2.1974,
      "theoretical_loss": 3.4746957261753035,
      "tokens_seen": 1718616064
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027327406846983164,
      "loss": 2.6259,
      "theoretical_loss": 3.4746843466769737,
      "tokens_seen": 1718681600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002732575184414602,
      "loss": 2.4537,
      "theoretical_loss": 3.4746729677340453,
      "tokens_seen": 1718747136
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002732409684130887,
      "loss": 2.7608,
      "theoretical_loss": 3.47466158934647,
      "tokens_seen": 1718812672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027322441838471725,
      "loss": 2.9749,
      "theoretical_loss": 3.4746502115141995,
      "tokens_seen": 1718878208
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027320786835634573,
      "loss": 2.3794,
      "theoretical_loss": 3.4746388342371857,
      "tokens_seen": 1718943744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027319131832797427,
      "loss": 2.5489,
      "theoretical_loss": 3.47462745751538,
      "tokens_seen": 1719009280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002731747682996028,
      "loss": 2.3249,
      "theoretical_loss": 3.474616081348735,
      "tokens_seen": 1719074816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027315821827123135,
      "loss": 2.6461,
      "theoretical_loss": 3.4746047057372014,
      "tokens_seen": 1719140352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027314166824285983,
      "loss": 2.6714,
      "theoretical_loss": 3.4745933306807313,
      "tokens_seen": 1719205888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027312511821448837,
      "loss": 2.3343,
      "theoretical_loss": 3.474581956179277,
      "tokens_seen": 1719271424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002731085681861169,
      "loss": 2.617,
      "theoretical_loss": 3.474570582232789,
      "tokens_seen": 1719336960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027309201815774544,
      "loss": 2.6168,
      "theoretical_loss": 3.4745592088412205,
      "tokens_seen": 1719402496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002730754681293739,
      "loss": 2.5102,
      "theoretical_loss": 3.4745478360045228,
      "tokens_seen": 1719468032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027305891810100246,
      "loss": 2.5216,
      "theoretical_loss": 3.4745364637226466,
      "tokens_seen": 1719533568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027304236807263095,
      "loss": 2.6553,
      "theoretical_loss": 3.4745250919955453,
      "tokens_seen": 1719599104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002730258180442595,
      "loss": 2.7004,
      "theoretical_loss": 3.47451372082317,
      "tokens_seen": 1719664640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000273009268015888,
      "loss": 2.5498,
      "theoretical_loss": 3.474502350205472,
      "tokens_seen": 1719730176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027299271798751656,
      "loss": 2.8443,
      "theoretical_loss": 3.474490980142404,
      "tokens_seen": 1719795712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027297616795914504,
      "loss": 2.5267,
      "theoretical_loss": 3.4744796106339173,
      "tokens_seen": 1719861248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002729596179307736,
      "loss": 2.4937,
      "theoretical_loss": 3.474468241679964,
      "tokens_seen": 1719926784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002729430679024021,
      "loss": 2.6746,
      "theoretical_loss": 3.474456873280496,
      "tokens_seen": 1719992320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027292651787403066,
      "loss": 2.5986,
      "theoretical_loss": 3.4744455054354644,
      "tokens_seen": 1720057856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027290996784565914,
      "loss": 2.3089,
      "theoretical_loss": 3.474434138144822,
      "tokens_seen": 1720123392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002728934178172877,
      "loss": 2.7457,
      "theoretical_loss": 3.47442277140852,
      "tokens_seen": 1720188928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002728768677889162,
      "loss": 2.7962,
      "theoretical_loss": 3.47441140522651,
      "tokens_seen": 1720254464
    },
    {
      "epoch": 0.61,
      "objective/train/avg_token_score": 0.0231667198240757,
      "objective/train/avg_weight": 0.9727231860160828,
      "objective/train/docs_used": 972105,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.048185348510742,
      "objective/train/std_weight": 0.11588402092456818,
      "objective/train/theoretical_loss": 3.474400039598745,
      "objective/train/tokens_used": 1740780000,
      "theoretical_loss": 3.474400039598745,
      "tokens_seen": 1720320000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027286031776054475,
      "loss": 2.5199,
      "theoretical_loss": 3.474400039598745,
      "tokens_seen": 1720320000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027284376773217323,
      "loss": 2.5801,
      "theoretical_loss": 3.4743886745251755,
      "tokens_seen": 1720385536
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027282721770380177,
      "loss": 2.548,
      "theoretical_loss": 3.4743773100057544,
      "tokens_seen": 1720451072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002728106676754303,
      "loss": 2.6666,
      "theoretical_loss": 3.4743659460404333,
      "tokens_seen": 1720516608
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027279411764705885,
      "loss": 2.6859,
      "theoretical_loss": 3.4743545826291635,
      "tokens_seen": 1720582144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027277756761868733,
      "loss": 2.7676,
      "theoretical_loss": 3.4743432197718978,
      "tokens_seen": 1720647680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027276101759031587,
      "loss": 2.4916,
      "theoretical_loss": 3.4743318574685875,
      "tokens_seen": 1720713216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002727444675619444,
      "loss": 2.7985,
      "theoretical_loss": 3.4743204957191844,
      "tokens_seen": 1720778752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027272791753357294,
      "loss": 2.5437,
      "theoretical_loss": 3.474309134523641,
      "tokens_seen": 1720844288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002727113675052014,
      "loss": 2.6214,
      "theoretical_loss": 3.4742977738819087,
      "tokens_seen": 1720909824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027269481747682996,
      "loss": 2.7283,
      "theoretical_loss": 3.4742864137939398,
      "tokens_seen": 1720975360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002726782674484585,
      "loss": 2.7957,
      "theoretical_loss": 3.474275054259685,
      "tokens_seen": 1721040896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027266171742008704,
      "loss": 2.4385,
      "theoretical_loss": 3.4742636952790984,
      "tokens_seen": 1721106432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002726451673917155,
      "loss": 2.7926,
      "theoretical_loss": 3.47425233685213,
      "tokens_seen": 1721171968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027262861736334406,
      "loss": 2.708,
      "theoretical_loss": 3.474240978978733,
      "tokens_seen": 1721237504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002726120673349726,
      "loss": 2.6054,
      "theoretical_loss": 3.474229621658859,
      "tokens_seen": 1721303040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027259551730660113,
      "loss": 2.9962,
      "theoretical_loss": 3.474218264892459,
      "tokens_seen": 1721368576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002725789672782296,
      "loss": 2.6892,
      "theoretical_loss": 3.4742069086794864,
      "tokens_seen": 1721434112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002725624172498581,
      "loss": 2.6202,
      "theoretical_loss": 3.474195553019892,
      "tokens_seen": 1721499648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027254586722148664,
      "loss": 2.675,
      "theoretical_loss": 3.4741841979136288,
      "tokens_seen": 1721565184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002725293171931152,
      "loss": 2.787,
      "theoretical_loss": 3.4741728433606482,
      "tokens_seen": 1721630720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002725127671647437,
      "loss": 2.9332,
      "theoretical_loss": 3.4741614893609016,
      "tokens_seen": 1721696256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002724962171363722,
      "loss": 2.841,
      "theoretical_loss": 3.4741501359143423,
      "tokens_seen": 1721761792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027247966710800074,
      "loss": 2.8573,
      "theoretical_loss": 3.4741387830209214,
      "tokens_seen": 1721827328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00027246311707962927,
      "loss": 2.6655,
      "theoretical_loss": 3.474127430680591,
      "tokens_seen": 1721892864
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002724465670512578,
      "loss": 2.699,
      "theoretical_loss": 3.474116078893303,
      "tokens_seen": 1721958400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002724300170228863,
      "loss": 2.5543,
      "theoretical_loss": 3.47410472765901,
      "tokens_seen": 1722023936
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027241346699451483,
      "loss": 2.4856,
      "theoretical_loss": 3.474093376977664,
      "tokens_seen": 1722089472
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027239691696614337,
      "loss": 2.6181,
      "theoretical_loss": 3.474082026849216,
      "tokens_seen": 1722155008
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002723803669377719,
      "loss": 2.6067,
      "theoretical_loss": 3.4740706772736187,
      "tokens_seen": 1722220544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002723638169094004,
      "loss": 2.6347,
      "theoretical_loss": 3.4740593282508243,
      "tokens_seen": 1722286080
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002723472668810289,
      "loss": 2.3032,
      "theoretical_loss": 3.474047979780785,
      "tokens_seen": 1722351616
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027233071685265746,
      "loss": 2.7501,
      "theoretical_loss": 3.474036631863452,
      "tokens_seen": 1722417152
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000272314166824286,
      "loss": 2.5519,
      "theoretical_loss": 3.4740252844987785,
      "tokens_seen": 1722482688
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002722976167959145,
      "loss": 2.5013,
      "theoretical_loss": 3.474013937686715,
      "tokens_seen": 1722548224
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000272281066767543,
      "loss": 2.3292,
      "theoretical_loss": 3.4740025914272152,
      "tokens_seen": 1722613760
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027226451673917156,
      "loss": 2.5744,
      "theoretical_loss": 3.4739912457202307,
      "tokens_seen": 1722679296
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002722479667108001,
      "loss": 2.2092,
      "theoretical_loss": 3.4739799005657126,
      "tokens_seen": 1722744832
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002722314166824286,
      "loss": 2.5645,
      "theoretical_loss": 3.473968555963614,
      "tokens_seen": 1722810368
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002722148666540571,
      "loss": 2.7531,
      "theoretical_loss": 3.473957211913887,
      "tokens_seen": 1722875904
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027219831662568566,
      "loss": 2.6789,
      "theoretical_loss": 3.4739458684164832,
      "tokens_seen": 1722941440
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002721817665973142,
      "loss": 2.7001,
      "theoretical_loss": 3.473934525471355,
      "tokens_seen": 1723006976
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002721652165689427,
      "loss": 2.4871,
      "theoretical_loss": 3.4739231830784543,
      "tokens_seen": 1723072512
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002721486665405712,
      "loss": 2.5697,
      "theoretical_loss": 3.473911841237733,
      "tokens_seen": 1723138048
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027213211651219975,
      "loss": 2.611,
      "theoretical_loss": 3.473900499949144,
      "tokens_seen": 1723203584
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002721155664838283,
      "loss": 2.8237,
      "theoretical_loss": 3.4738891592126393,
      "tokens_seen": 1723269120
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002720990164554568,
      "loss": 2.5383,
      "theoretical_loss": 3.4738778190281705,
      "tokens_seen": 1723334656
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027208246642708526,
      "loss": 2.5734,
      "theoretical_loss": 3.4738664793956895,
      "tokens_seen": 1723400192
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002720659163987138,
      "loss": 2.6211,
      "theoretical_loss": 3.473855140315149,
      "tokens_seen": 1723465728
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027204936637034233,
      "loss": 2.3201,
      "theoretical_loss": 3.473843801786501,
      "tokens_seen": 1723531264
    },
    {
      "epoch": 0.62,
      "objective/train/avg_token_score": 0.009342223405838013,
      "objective/train/avg_weight": 0.9890127182006836,
      "objective/train/docs_used": 974104,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3293700218200684,
      "objective/train/std_weight": 0.057375483214855194,
      "objective/train/theoretical_loss": 3.473832463809698,
      "objective/train/tokens_used": 1744056800,
      "theoretical_loss": 3.473832463809698,
      "tokens_seen": 1723596800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027203281634197087,
      "loss": 2.7453,
      "theoretical_loss": 3.473832463809698,
      "tokens_seen": 1723596800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027201626631359935,
      "loss": 2.4391,
      "theoretical_loss": 3.4738211263846916,
      "tokens_seen": 1723662336
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002719997162852279,
      "loss": 2.6898,
      "theoretical_loss": 3.473809789511434,
      "tokens_seen": 1723727872
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027198316625685643,
      "loss": 2.7805,
      "theoretical_loss": 3.473798453189878,
      "tokens_seen": 1723793408
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027196661622848497,
      "loss": 2.7753,
      "theoretical_loss": 3.4737871174199757,
      "tokens_seen": 1723858944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027195006620011345,
      "loss": 2.652,
      "theoretical_loss": 3.4737757822016784,
      "tokens_seen": 1723924480
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000271933516171742,
      "loss": 2.5395,
      "theoretical_loss": 3.473764447534939,
      "tokens_seen": 1723990016
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002719169661433705,
      "loss": 2.963,
      "theoretical_loss": 3.4737531134197095,
      "tokens_seen": 1724055552
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027190041611499906,
      "loss": 2.8135,
      "theoretical_loss": 3.473741779855942,
      "tokens_seen": 1724121088
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027188386608662754,
      "loss": 2.4477,
      "theoretical_loss": 3.4737304468435894,
      "tokens_seen": 1724186624
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002718673160582561,
      "loss": 2.6833,
      "theoretical_loss": 3.473719114382603,
      "tokens_seen": 1724252160
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002718507660298846,
      "loss": 2.8204,
      "theoretical_loss": 3.4737077824729354,
      "tokens_seen": 1724317696
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027183421600151316,
      "loss": 2.5196,
      "theoretical_loss": 3.4736964511145385,
      "tokens_seen": 1724383232
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027181766597314164,
      "loss": 2.5987,
      "theoretical_loss": 3.473685120307365,
      "tokens_seen": 1724448768
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002718011159447702,
      "loss": 2.6653,
      "theoretical_loss": 3.4736737900513672,
      "tokens_seen": 1724514304
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002717845659163987,
      "loss": 2.6314,
      "theoretical_loss": 3.473662460346497,
      "tokens_seen": 1724579840
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027176801588802725,
      "loss": 2.5534,
      "theoretical_loss": 3.4736511311927067,
      "tokens_seen": 1724645376
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027175146585965574,
      "loss": 2.4399,
      "theoretical_loss": 3.4736398025899486,
      "tokens_seen": 1724710912
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002717349158312843,
      "loss": 2.8995,
      "theoretical_loss": 3.4736284745381756,
      "tokens_seen": 1724776448
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002717183658029128,
      "loss": 2.5473,
      "theoretical_loss": 3.4736171470373387,
      "tokens_seen": 1724841984
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027170181577454135,
      "loss": 2.737,
      "theoretical_loss": 3.473605820087391,
      "tokens_seen": 1724907520
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002716852657461699,
      "loss": 2.6513,
      "theoretical_loss": 3.4735944936882848,
      "tokens_seen": 1724973056
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027166871571779837,
      "loss": 2.6872,
      "theoretical_loss": 3.473583167839972,
      "tokens_seen": 1725038592
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002716521656894269,
      "loss": 2.5981,
      "theoretical_loss": 3.473571842542405,
      "tokens_seen": 1725104128
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027163561566105545,
      "loss": 2.7587,
      "theoretical_loss": 3.473560517795536,
      "tokens_seen": 1725169664
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000271619065632684,
      "loss": 2.5938,
      "theoretical_loss": 3.473549193599318,
      "tokens_seen": 1725235200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027160251560431247,
      "loss": 2.6502,
      "theoretical_loss": 3.4735378699537027,
      "tokens_seen": 1725300736
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027158596557594095,
      "loss": 2.4574,
      "theoretical_loss": 3.4735265468586425,
      "tokens_seen": 1725366272
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002715694155475695,
      "loss": 2.7194,
      "theoretical_loss": 3.4735152243140894,
      "tokens_seen": 1725431808
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000271552865519198,
      "loss": 2.5978,
      "theoretical_loss": 3.4735039023199965,
      "tokens_seen": 1725497344
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027153631549082656,
      "loss": 2.3788,
      "theoretical_loss": 3.4734925808763153,
      "tokens_seen": 1725562880
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027151976546245505,
      "loss": 2.7454,
      "theoretical_loss": 3.473481259982999,
      "tokens_seen": 1725628416
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002715032154340836,
      "loss": 2.4542,
      "theoretical_loss": 3.4734699396399993,
      "tokens_seen": 1725693952
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002714866654057121,
      "loss": 2.618,
      "theoretical_loss": 3.4734586198472686,
      "tokens_seen": 1725759488
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027147011537734066,
      "loss": 3.0105,
      "theoretical_loss": 3.4734473006047595,
      "tokens_seen": 1725825024
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027145356534896914,
      "loss": 2.703,
      "theoretical_loss": 3.4734359819124245,
      "tokens_seen": 1725890560
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002714370153205977,
      "loss": 2.3746,
      "theoretical_loss": 3.4734246637702153,
      "tokens_seen": 1725956096
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002714204652922262,
      "loss": 2.5986,
      "theoretical_loss": 3.4734133461780847,
      "tokens_seen": 1726021632
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027140391526385475,
      "loss": 2.7506,
      "theoretical_loss": 3.4734020291359853,
      "tokens_seen": 1726087168
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027138736523548324,
      "loss": 2.4095,
      "theoretical_loss": 3.473390712643869,
      "tokens_seen": 1726152704
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002713708152071118,
      "loss": 2.334,
      "theoretical_loss": 3.473379396701689,
      "tokens_seen": 1726218240
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002713542651787403,
      "loss": 2.5475,
      "theoretical_loss": 3.473368081309397,
      "tokens_seen": 1726283776
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027133771515036885,
      "loss": 2.4594,
      "theoretical_loss": 3.473356766466945,
      "tokens_seen": 1726349312
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027132116512199733,
      "loss": 2.6097,
      "theoretical_loss": 3.4733454521742866,
      "tokens_seen": 1726414848
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027130461509362587,
      "loss": 2.5969,
      "theoretical_loss": 3.4733341384313734,
      "tokens_seen": 1726480384
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002712880650652544,
      "loss": 2.7139,
      "theoretical_loss": 3.4733228252381574,
      "tokens_seen": 1726545920
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027127151503688295,
      "loss": 2.7538,
      "theoretical_loss": 3.4733115125945924,
      "tokens_seen": 1726611456
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027125496500851143,
      "loss": 2.6371,
      "theoretical_loss": 3.47330020050063,
      "tokens_seen": 1726676992
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027123841498013997,
      "loss": 2.8741,
      "theoretical_loss": 3.4732888889562226,
      "tokens_seen": 1726742528
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002712218649517685,
      "loss": 2.5243,
      "theoretical_loss": 3.4732775779613227,
      "tokens_seen": 1726808064
    },
    {
      "epoch": 0.62,
      "objective/train/avg_token_score": 0.020880253985524178,
      "objective/train/avg_weight": 0.9754216074943542,
      "objective/train/docs_used": 976014,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.482815742492676,
      "objective/train/std_weight": 0.10294607281684875,
      "objective/train/theoretical_loss": 3.473266267515883,
      "objective/train/tokens_used": 1747333600,
      "theoretical_loss": 3.473266267515883,
      "tokens_seen": 1726873600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027120531492339704,
      "loss": 2.6616,
      "theoretical_loss": 3.473266267515883,
      "tokens_seen": 1726873600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002711887648950255,
      "loss": 2.8284,
      "theoretical_loss": 3.4732549576198553,
      "tokens_seen": 1726939136
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027117221486665406,
      "loss": 2.9084,
      "theoretical_loss": 3.4732436482731925,
      "tokens_seen": 1727004672
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002711556648382826,
      "loss": 2.7297,
      "theoretical_loss": 3.4732323394758478,
      "tokens_seen": 1727070208
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027113911480991114,
      "loss": 2.6246,
      "theoretical_loss": 3.473221031227773,
      "tokens_seen": 1727135744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002711225647815396,
      "loss": 2.6127,
      "theoretical_loss": 3.47320972352892,
      "tokens_seen": 1727201280
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027110601475316816,
      "loss": 2.9664,
      "theoretical_loss": 3.473198416379242,
      "tokens_seen": 1727266816
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027108946472479664,
      "loss": 2.5663,
      "theoretical_loss": 3.4731871097786913,
      "tokens_seen": 1727332352
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002710729146964252,
      "loss": 2.669,
      "theoretical_loss": 3.4731758037272202,
      "tokens_seen": 1727397888
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002710563646680537,
      "loss": 2.3973,
      "theoretical_loss": 3.473164498224782,
      "tokens_seen": 1727463424
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002710398146396822,
      "loss": 2.6834,
      "theoretical_loss": 3.4731531932713287,
      "tokens_seen": 1727528960
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027102326461131074,
      "loss": 2.6407,
      "theoretical_loss": 3.473141888866812,
      "tokens_seen": 1727594496
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002710067145829393,
      "loss": 2.6799,
      "theoretical_loss": 3.4731305850111855,
      "tokens_seen": 1727660032
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002709901645545678,
      "loss": 2.3791,
      "theoretical_loss": 3.473119281704402,
      "tokens_seen": 1727725568
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002709736145261963,
      "loss": 2.9709,
      "theoretical_loss": 3.473107978946413,
      "tokens_seen": 1727791104
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027095706449782483,
      "loss": 2.6106,
      "theoretical_loss": 3.4730966767371716,
      "tokens_seen": 1727856640
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027094051446945337,
      "loss": 2.699,
      "theoretical_loss": 3.4730853750766304,
      "tokens_seen": 1727922176
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002709239644410819,
      "loss": 2.5508,
      "theoretical_loss": 3.4730740739647414,
      "tokens_seen": 1727987712
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002709074144127104,
      "loss": 2.7131,
      "theoretical_loss": 3.4730627734014576,
      "tokens_seen": 1728053248
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027089086438433893,
      "loss": 2.6625,
      "theoretical_loss": 3.4730514733867315,
      "tokens_seen": 1728118784
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027087431435596747,
      "loss": 2.773,
      "theoretical_loss": 3.4730401739205163,
      "tokens_seen": 1728184320
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000270857764327596,
      "loss": 2.9772,
      "theoretical_loss": 3.4730288750027634,
      "tokens_seen": 1728249856
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002708412142992245,
      "loss": 2.9669,
      "theoretical_loss": 3.473017576633426,
      "tokens_seen": 1728315392
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000270824664270853,
      "loss": 2.7724,
      "theoretical_loss": 3.4730062788124565,
      "tokens_seen": 1728380928
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027080811424248156,
      "loss": 2.6566,
      "theoretical_loss": 3.472994981539808,
      "tokens_seen": 1728446464
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002707915642141101,
      "loss": 2.2414,
      "theoretical_loss": 3.4729836848154325,
      "tokens_seen": 1728512000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002707750141857386,
      "loss": 2.7955,
      "theoretical_loss": 3.472972388639283,
      "tokens_seen": 1728577536
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002707584641573671,
      "loss": 2.4197,
      "theoretical_loss": 3.4729610930113117,
      "tokens_seen": 1728643072
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027074191412899566,
      "loss": 2.6662,
      "theoretical_loss": 3.472949797931472,
      "tokens_seen": 1728708608
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002707253641006242,
      "loss": 2.6101,
      "theoretical_loss": 3.472938503399715,
      "tokens_seen": 1728774144
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002707088140722527,
      "loss": 2.3823,
      "theoretical_loss": 3.472927209415995,
      "tokens_seen": 1728839680
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002706922640438812,
      "loss": 2.5683,
      "theoretical_loss": 3.4729159159802636,
      "tokens_seen": 1728905216
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027067571401550976,
      "loss": 2.7812,
      "theoretical_loss": 3.4729046230924743,
      "tokens_seen": 1728970752
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002706591639871383,
      "loss": 2.6753,
      "theoretical_loss": 3.472893330752579,
      "tokens_seen": 1729036288
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002706426139587668,
      "loss": 2.3815,
      "theoretical_loss": 3.4728820389605306,
      "tokens_seen": 1729101824
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002706260639303953,
      "loss": 2.7405,
      "theoretical_loss": 3.4728707477162817,
      "tokens_seen": 1729167360
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002706095139020238,
      "loss": 2.7626,
      "theoretical_loss": 3.472859457019785,
      "tokens_seen": 1729232896
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027059296387365233,
      "loss": 2.6121,
      "theoretical_loss": 3.4728481668709934,
      "tokens_seen": 1729298432
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027057641384528087,
      "loss": 2.5528,
      "theoretical_loss": 3.4728368772698586,
      "tokens_seen": 1729363968
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027055986381690936,
      "loss": 2.9207,
      "theoretical_loss": 3.4728255882163346,
      "tokens_seen": 1729429504
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002705433137885379,
      "loss": 2.4405,
      "theoretical_loss": 3.4728142997103735,
      "tokens_seen": 1729495040
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027052676376016643,
      "loss": 2.4759,
      "theoretical_loss": 3.472803011751928,
      "tokens_seen": 1729560576
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027051021373179497,
      "loss": 2.3891,
      "theoretical_loss": 3.472791724340951,
      "tokens_seen": 1729626112
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002704936637034235,
      "loss": 2.6239,
      "theoretical_loss": 3.472780437477395,
      "tokens_seen": 1729691648
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000270477113675052,
      "loss": 2.4347,
      "theoretical_loss": 3.472769151161213,
      "tokens_seen": 1729757184
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002704605636466805,
      "loss": 2.8229,
      "theoretical_loss": 3.4727578653923565,
      "tokens_seen": 1729822720
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027044401361830906,
      "loss": 2.5176,
      "theoretical_loss": 3.4727465801707798,
      "tokens_seen": 1729888256
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002704274635899376,
      "loss": 2.6945,
      "theoretical_loss": 3.472735295496435,
      "tokens_seen": 1729953792
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002704109135615661,
      "loss": 2.6194,
      "theoretical_loss": 3.4727240113692748,
      "tokens_seen": 1730019328
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002703943635331946,
      "loss": 2.5796,
      "theoretical_loss": 3.472712727789252,
      "tokens_seen": 1730084864
    },
    {
      "epoch": 0.62,
      "objective/train/avg_token_score": 0.018559280782938004,
      "objective/train/avg_weight": 0.9781551361083984,
      "objective/train/docs_used": 978029,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.7889537811279297,
      "objective/train/std_weight": 0.09386657178401947,
      "objective/train/theoretical_loss": 3.4727014447563196,
      "objective/train/tokens_used": 1750610400,
      "theoretical_loss": 3.4727014447563196,
      "tokens_seen": 1730150400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027037781350482316,
      "loss": 2.5747,
      "theoretical_loss": 3.4727014447563196,
      "tokens_seen": 1730150400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002703612634764517,
      "loss": 2.8233,
      "theoretical_loss": 3.472690162270429,
      "tokens_seen": 1730215936
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002703447134480802,
      "loss": 2.4833,
      "theoretical_loss": 3.472678880331535,
      "tokens_seen": 1730281472
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002703281634197087,
      "loss": 2.618,
      "theoretical_loss": 3.4726675989395894,
      "tokens_seen": 1730347008
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027031161339133726,
      "loss": 2.7454,
      "theoretical_loss": 3.472656318094545,
      "tokens_seen": 1730412544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002702950633629658,
      "loss": 2.9354,
      "theoretical_loss": 3.4726450377963545,
      "tokens_seen": 1730478080
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002702785133345943,
      "loss": 2.8667,
      "theoretical_loss": 3.47263375804497,
      "tokens_seen": 1730543616
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002702619633062228,
      "loss": 2.6794,
      "theoretical_loss": 3.472622478840346,
      "tokens_seen": 1730609152
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027024541327785135,
      "loss": 2.5379,
      "theoretical_loss": 3.472611200182434,
      "tokens_seen": 1730674688
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002702288632494799,
      "loss": 2.6921,
      "theoretical_loss": 3.472599922071187,
      "tokens_seen": 1730740224
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002702123132211084,
      "loss": 2.6506,
      "theoretical_loss": 3.472588644506558,
      "tokens_seen": 1730805760
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002701957631927369,
      "loss": 2.5983,
      "theoretical_loss": 3.4725773674884994,
      "tokens_seen": 1730871296
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027017921316436545,
      "loss": 2.4486,
      "theoretical_loss": 3.4725660910169647,
      "tokens_seen": 1730936832
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000270162663135994,
      "loss": 2.4922,
      "theoretical_loss": 3.472554815091906,
      "tokens_seen": 1731002368
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027014611310762247,
      "loss": 2.5896,
      "theoretical_loss": 3.472543539713277,
      "tokens_seen": 1731067904
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000270129563079251,
      "loss": 2.6476,
      "theoretical_loss": 3.4725322648810297,
      "tokens_seen": 1731133440
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002701130130508795,
      "loss": 2.6516,
      "theoretical_loss": 3.4725209905951173,
      "tokens_seen": 1731198976
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027009646302250803,
      "loss": 2.6104,
      "theoretical_loss": 3.472509716855493,
      "tokens_seen": 1731264512
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027007991299413656,
      "loss": 2.8095,
      "theoretical_loss": 3.4724984436621087,
      "tokens_seen": 1731330048
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027006336296576505,
      "loss": 2.5678,
      "theoretical_loss": 3.4724871710149183,
      "tokens_seen": 1731395584
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002700468129373936,
      "loss": 2.5544,
      "theoretical_loss": 3.4724758989138738,
      "tokens_seen": 1731461120
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002700302629090221,
      "loss": 2.7051,
      "theoretical_loss": 3.472464627358929,
      "tokens_seen": 1731526656
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00027001371288065066,
      "loss": 2.4803,
      "theoretical_loss": 3.4724533563500355,
      "tokens_seen": 1731592192
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026999716285227914,
      "loss": 2.578,
      "theoretical_loss": 3.472442085887147,
      "tokens_seen": 1731657728
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002699806128239077,
      "loss": 2.6207,
      "theoretical_loss": 3.472430815970217,
      "tokens_seen": 1731723264
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002699640627955362,
      "loss": 2.559,
      "theoretical_loss": 3.4724195465991974,
      "tokens_seen": 1731788800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026994751276716476,
      "loss": 2.7243,
      "theoretical_loss": 3.4724082777740413,
      "tokens_seen": 1731854336
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026993096273879324,
      "loss": 2.6334,
      "theoretical_loss": 3.472397009494702,
      "tokens_seen": 1731919872
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002699144127104218,
      "loss": 2.5891,
      "theoretical_loss": 3.4723857417611317,
      "tokens_seen": 1731985408
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002698978626820503,
      "loss": 2.6202,
      "theoretical_loss": 3.4723744745732836,
      "tokens_seen": 1732050944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026988131265367885,
      "loss": 2.5284,
      "theoretical_loss": 3.472363207931111,
      "tokens_seen": 1732116480
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026986476262530734,
      "loss": 2.9393,
      "theoretical_loss": 3.4723519418345665,
      "tokens_seen": 1732182016
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002698482125969359,
      "loss": 2.6275,
      "theoretical_loss": 3.4723406762836033,
      "tokens_seen": 1732247552
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002698316625685644,
      "loss": 2.8887,
      "theoretical_loss": 3.472329411278174,
      "tokens_seen": 1732313088
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026981511254019295,
      "loss": 2.6295,
      "theoretical_loss": 3.472318146818232,
      "tokens_seen": 1732378624
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026979856251182143,
      "loss": 2.9992,
      "theoretical_loss": 3.4723068829037294,
      "tokens_seen": 1732444160
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026978201248344997,
      "loss": 2.9153,
      "theoretical_loss": 3.47229561953462,
      "tokens_seen": 1732509696
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002697654624550785,
      "loss": 2.8337,
      "theoretical_loss": 3.4722843567108566,
      "tokens_seen": 1732575232
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026974891242670704,
      "loss": 2.7,
      "theoretical_loss": 3.4722730944323916,
      "tokens_seen": 1732640768
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026973236239833553,
      "loss": 2.3143,
      "theoretical_loss": 3.4722618326991785,
      "tokens_seen": 1732706304
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026971581236996407,
      "loss": 2.2368,
      "theoretical_loss": 3.4722505715111707,
      "tokens_seen": 1732771840
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002696992623415926,
      "loss": 2.4889,
      "theoretical_loss": 3.4722393108683205,
      "tokens_seen": 1732837376
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026968271231322114,
      "loss": 2.6664,
      "theoretical_loss": 3.472228050770581,
      "tokens_seen": 1732902912
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002696661622848496,
      "loss": 2.5115,
      "theoretical_loss": 3.472216791217905,
      "tokens_seen": 1732968448
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026964961225647816,
      "loss": 2.6083,
      "theoretical_loss": 3.4722055322102463,
      "tokens_seen": 1733033984
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002696330622281067,
      "loss": 2.8171,
      "theoretical_loss": 3.472194273747557,
      "tokens_seen": 1733099520
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002696165121997352,
      "loss": 2.4007,
      "theoretical_loss": 3.47218301582979,
      "tokens_seen": 1733165056
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002695999621713637,
      "loss": 2.677,
      "theoretical_loss": 3.4721717584569003,
      "tokens_seen": 1733230592
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002695834121429922,
      "loss": 2.6526,
      "theoretical_loss": 3.472160501628838,
      "tokens_seen": 1733296128
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026956686211462074,
      "loss": 2.6764,
      "theoretical_loss": 3.4721492453455585,
      "tokens_seen": 1733361664
    },
    {
      "epoch": 0.62,
      "objective/train/avg_token_score": 0.016294091939926147,
      "objective/train/avg_weight": 0.9808323979377747,
      "objective/train/docs_used": 979550,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.674853801727295,
      "objective/train/std_weight": 0.07974858582019806,
      "objective/train/theoretical_loss": 3.4721379896070133,
      "objective/train/tokens_used": 1753887200,
      "theoretical_loss": 3.4721379896070133,
      "tokens_seen": 1733427200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002695503120862493,
      "loss": 2.575,
      "theoretical_loss": 3.4721379896070133,
      "tokens_seen": 1733427200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002695337620578778,
      "loss": 2.7798,
      "theoretical_loss": 3.4721267344131563,
      "tokens_seen": 1733492736
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002695172120295063,
      "loss": 2.7334,
      "theoretical_loss": 3.4721154797639406,
      "tokens_seen": 1733558272
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026950066200113484,
      "loss": 2.661,
      "theoretical_loss": 3.4721042256593186,
      "tokens_seen": 1733623808
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002694841119727634,
      "loss": 3.1073,
      "theoretical_loss": 3.472092972099244,
      "tokens_seen": 1733689344
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002694675619443919,
      "loss": 2.816,
      "theoretical_loss": 3.4720817190836692,
      "tokens_seen": 1733754880
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002694510119160204,
      "loss": 2.6651,
      "theoretical_loss": 3.472070466612548,
      "tokens_seen": 1733820416
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026943446188764893,
      "loss": 2.6791,
      "theoretical_loss": 3.4720592146858333,
      "tokens_seen": 1733885952
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026941791185927747,
      "loss": 2.553,
      "theoretical_loss": 3.4720479633034778,
      "tokens_seen": 1733951488
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000269401361830906,
      "loss": 2.6522,
      "theoretical_loss": 3.472036712465435,
      "tokens_seen": 1734017024
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002693848118025345,
      "loss": 2.5969,
      "theoretical_loss": 3.4720254621716578,
      "tokens_seen": 1734082560
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026936826177416303,
      "loss": 2.4258,
      "theoretical_loss": 3.472014212422099,
      "tokens_seen": 1734148096
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026935171174579157,
      "loss": 2.7741,
      "theoretical_loss": 3.4720029632167124,
      "tokens_seen": 1734213632
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002693351617174201,
      "loss": 2.8227,
      "theoretical_loss": 3.471991714555451,
      "tokens_seen": 1734279168
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002693186116890486,
      "loss": 2.3603,
      "theoretical_loss": 3.471980466438267,
      "tokens_seen": 1734344704
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002693020616606771,
      "loss": 2.8821,
      "theoretical_loss": 3.471969218865115,
      "tokens_seen": 1734410240
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026928551163230566,
      "loss": 2.6928,
      "theoretical_loss": 3.4719579718359466,
      "tokens_seen": 1734475776
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002692689616039342,
      "loss": 2.5228,
      "theoretical_loss": 3.4719467253507164,
      "tokens_seen": 1734541312
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002692524115755627,
      "loss": 2.8364,
      "theoretical_loss": 3.471935479409376,
      "tokens_seen": 1734606848
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002692358615471912,
      "loss": 2.4425,
      "theoretical_loss": 3.4719242340118797,
      "tokens_seen": 1734672384
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026921931151881976,
      "loss": 2.8307,
      "theoretical_loss": 3.471912989158181,
      "tokens_seen": 1734737920
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002692027614904483,
      "loss": 2.8996,
      "theoretical_loss": 3.471901744848231,
      "tokens_seen": 1734803456
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002691862114620768,
      "loss": 2.5855,
      "theoretical_loss": 3.471890501081985,
      "tokens_seen": 1734868992
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002691696614337053,
      "loss": 2.6873,
      "theoretical_loss": 3.4718792578593956,
      "tokens_seen": 1734934528
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026915311140533385,
      "loss": 2.7364,
      "theoretical_loss": 3.4718680151804158,
      "tokens_seen": 1735000064
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002691365613769624,
      "loss": 2.6417,
      "theoretical_loss": 3.4718567730449985,
      "tokens_seen": 1735065600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002691200113485909,
      "loss": 2.6293,
      "theoretical_loss": 3.4718455314530976,
      "tokens_seen": 1735131136
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002691034613202194,
      "loss": 2.7613,
      "theoretical_loss": 3.471834290404665,
      "tokens_seen": 1735196672
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002690869112918479,
      "loss": 2.5584,
      "theoretical_loss": 3.4718230498996556,
      "tokens_seen": 1735262208
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026907036126347643,
      "loss": 2.7385,
      "theoretical_loss": 3.4718118099380213,
      "tokens_seen": 1735327744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026905381123510497,
      "loss": 2.5841,
      "theoretical_loss": 3.4718005705197164,
      "tokens_seen": 1735393280
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002690372612067335,
      "loss": 2.9526,
      "theoretical_loss": 3.4717893316446933,
      "tokens_seen": 1735458816
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000269020711178362,
      "loss": 2.7048,
      "theoretical_loss": 3.471778093312905,
      "tokens_seen": 1735524352
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026900416114999053,
      "loss": 2.8458,
      "theoretical_loss": 3.4717668555243053,
      "tokens_seen": 1735589888
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026898761112161907,
      "loss": 2.5841,
      "theoretical_loss": 3.4717556182788476,
      "tokens_seen": 1735655424
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002689710610932476,
      "loss": 2.7077,
      "theoretical_loss": 3.471744381576485,
      "tokens_seen": 1735720960
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002689545110648761,
      "loss": 2.7347,
      "theoretical_loss": 3.4717331454171703,
      "tokens_seen": 1735786496
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002689379610365046,
      "loss": 2.4723,
      "theoretical_loss": 3.471721909800857,
      "tokens_seen": 1735852032
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026892141100813316,
      "loss": 2.4619,
      "theoretical_loss": 3.4717106747274986,
      "tokens_seen": 1735917568
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002689048609797617,
      "loss": 2.7388,
      "theoretical_loss": 3.471699440197048,
      "tokens_seen": 1735983104
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002688883109513902,
      "loss": 2.7023,
      "theoretical_loss": 3.4716882062094587,
      "tokens_seen": 1736048640
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002688717609230187,
      "loss": 2.9439,
      "theoretical_loss": 3.4716769727646835,
      "tokens_seen": 1736114176
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026885521089464726,
      "loss": 2.8149,
      "theoretical_loss": 3.4716657398626767,
      "tokens_seen": 1736179712
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002688386608662758,
      "loss": 2.769,
      "theoretical_loss": 3.471654507503391,
      "tokens_seen": 1736245248
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002688221108379043,
      "loss": 2.5672,
      "theoretical_loss": 3.4716432756867794,
      "tokens_seen": 1736310784
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002688055608095328,
      "loss": 2.3312,
      "theoretical_loss": 3.4716320444127957,
      "tokens_seen": 1736376320
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026878901078116135,
      "loss": 2.7145,
      "theoretical_loss": 3.471620813681393,
      "tokens_seen": 1736441856
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002687724607527899,
      "loss": 2.6499,
      "theoretical_loss": 3.4716095834925245,
      "tokens_seen": 1736507392
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002687559107244184,
      "loss": 2.4767,
      "theoretical_loss": 3.4715983538461432,
      "tokens_seen": 1736572928
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002687393606960469,
      "loss": 2.5146,
      "theoretical_loss": 3.4715871247422037,
      "tokens_seen": 1736638464
    },
    {
      "epoch": 0.62,
      "objective/train/avg_token_score": 0.013804541900753975,
      "objective/train/avg_weight": 0.9837714433670044,
      "objective/train/docs_used": 981478,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8631362915039062,
      "objective/train/std_weight": 0.06586441397666931,
      "objective/train/theoretical_loss": 3.4715758961806573,
      "objective/train/tokens_used": 1757164000,
      "theoretical_loss": 3.4715758961806573,
      "tokens_seen": 1736704000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026872281066767545,
      "loss": 2.4765,
      "theoretical_loss": 3.4715758961806573,
      "tokens_seen": 1736704000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000268706260639304,
      "loss": 2.6539,
      "theoretical_loss": 3.4715646681614594,
      "tokens_seen": 1736769536
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026868971061093247,
      "loss": 2.6484,
      "theoretical_loss": 3.4715534406845627,
      "tokens_seen": 1736835072
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000268673160582561,
      "loss": 2.6363,
      "theoretical_loss": 3.4715422137499194,
      "tokens_seen": 1736900608
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026865661055418955,
      "loss": 2.8463,
      "theoretical_loss": 3.4715309873574847,
      "tokens_seen": 1736966144
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026864006052581803,
      "loss": 2.6825,
      "theoretical_loss": 3.4715197615072104,
      "tokens_seen": 1737031680
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026862351049744657,
      "loss": 2.4877,
      "theoretical_loss": 3.4715085361990505,
      "tokens_seen": 1737097216
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026860696046907505,
      "loss": 2.6599,
      "theoretical_loss": 3.471497311432959,
      "tokens_seen": 1737162752
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002685904104407036,
      "loss": 2.5145,
      "theoretical_loss": 3.4714860872088877,
      "tokens_seen": 1737228288
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002685738604123321,
      "loss": 2.8995,
      "theoretical_loss": 3.471474863526791,
      "tokens_seen": 1737293824
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026855731038396066,
      "loss": 2.9889,
      "theoretical_loss": 3.471463640386623,
      "tokens_seen": 1737359360
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026854076035558915,
      "loss": 2.704,
      "theoretical_loss": 3.4714524177883352,
      "tokens_seen": 1737424896
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002685242103272177,
      "loss": 2.7328,
      "theoretical_loss": 3.4714411957318827,
      "tokens_seen": 1737490432
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002685076602988462,
      "loss": 2.469,
      "theoretical_loss": 3.4714299742172186,
      "tokens_seen": 1737555968
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026849111027047476,
      "loss": 2.63,
      "theoretical_loss": 3.4714187532442953,
      "tokens_seen": 1737621504
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026847456024210324,
      "loss": 2.3794,
      "theoretical_loss": 3.471407532813067,
      "tokens_seen": 1737687040
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002684580102137318,
      "loss": 2.6454,
      "theoretical_loss": 3.4713963129234875,
      "tokens_seen": 1737752576
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002684414601853603,
      "loss": 2.5255,
      "theoretical_loss": 3.4713850935755093,
      "tokens_seen": 1737818112
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026842491015698886,
      "loss": 2.5922,
      "theoretical_loss": 3.4713738747690863,
      "tokens_seen": 1737883648
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026840836012861734,
      "loss": 2.4906,
      "theoretical_loss": 3.471362656504172,
      "tokens_seen": 1737949184
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002683918101002459,
      "loss": 2.5704,
      "theoretical_loss": 3.4713514387807196,
      "tokens_seen": 1738014720
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002683752600718744,
      "loss": 2.7323,
      "theoretical_loss": 3.471340221598683,
      "tokens_seen": 1738080256
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026835871004350295,
      "loss": 2.8666,
      "theoretical_loss": 3.4713290049580157,
      "tokens_seen": 1738145792
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026834216001513143,
      "loss": 2.6735,
      "theoretical_loss": 3.4713177888586704,
      "tokens_seen": 1738211328
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026832560998675997,
      "loss": 2.7189,
      "theoretical_loss": 3.471306573300601,
      "tokens_seen": 1738276864
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002683090599583885,
      "loss": 2.5021,
      "theoretical_loss": 3.4712953582837613,
      "tokens_seen": 1738342400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026829250993001705,
      "loss": 2.6247,
      "theoretical_loss": 3.471284143808104,
      "tokens_seen": 1738407936
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026827595990164553,
      "loss": 2.5111,
      "theoretical_loss": 3.4712729298735834,
      "tokens_seen": 1738473472
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026825940987327407,
      "loss": 2.8421,
      "theoretical_loss": 3.471261716480152,
      "tokens_seen": 1738539008
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002682428598449026,
      "loss": 2.6186,
      "theoretical_loss": 3.4712505036277648,
      "tokens_seen": 1738604544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026822630981653114,
      "loss": 2.5678,
      "theoretical_loss": 3.4712392913163734,
      "tokens_seen": 1738670080
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002682097597881596,
      "loss": 2.6871,
      "theoretical_loss": 3.4712280795459334,
      "tokens_seen": 1738735616
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026819320975978816,
      "loss": 2.5889,
      "theoretical_loss": 3.4712168683163966,
      "tokens_seen": 1738801152
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002681766597314167,
      "loss": 2.611,
      "theoretical_loss": 3.4712056576277175,
      "tokens_seen": 1738866688
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026816010970304524,
      "loss": 2.7334,
      "theoretical_loss": 3.4711944474798493,
      "tokens_seen": 1738932224
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002681435596746737,
      "loss": 2.5468,
      "theoretical_loss": 3.4711832378727454,
      "tokens_seen": 1738997760
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002681270096463022,
      "loss": 2.52,
      "theoretical_loss": 3.4711720288063592,
      "tokens_seen": 1739063296
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026811045961793074,
      "loss": 2.6106,
      "theoretical_loss": 3.4711608202806445,
      "tokens_seen": 1739128832
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002680939095895593,
      "loss": 2.6748,
      "theoretical_loss": 3.471149612295555,
      "tokens_seen": 1739194368
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002680773595611878,
      "loss": 2.5762,
      "theoretical_loss": 3.4711384048510445,
      "tokens_seen": 1739259904
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002680608095328163,
      "loss": 2.7618,
      "theoretical_loss": 3.471127197947066,
      "tokens_seen": 1739325440
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026804425950444484,
      "loss": 2.7113,
      "theoretical_loss": 3.4711159915835728,
      "tokens_seen": 1739390976
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002680277094760734,
      "loss": 2.3839,
      "theoretical_loss": 3.4711047857605193,
      "tokens_seen": 1739456512
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002680111594477019,
      "loss": 2.4105,
      "theoretical_loss": 3.4710935804778584,
      "tokens_seen": 1739522048
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002679946094193304,
      "loss": 2.6521,
      "theoretical_loss": 3.471082375735544,
      "tokens_seen": 1739587584
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026797805939095894,
      "loss": 2.747,
      "theoretical_loss": 3.47107117153353,
      "tokens_seen": 1739653120
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002679615093625875,
      "loss": 2.7028,
      "theoretical_loss": 3.471059967871769,
      "tokens_seen": 1739718656
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000267944959334216,
      "loss": 2.5924,
      "theoretical_loss": 3.4710487647502157,
      "tokens_seen": 1739784192
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002679284093058445,
      "loss": 2.7565,
      "theoretical_loss": 3.4710375621688234,
      "tokens_seen": 1739849728
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026791185927747303,
      "loss": 2.6386,
      "theoretical_loss": 3.471026360127545,
      "tokens_seen": 1739915264
    },
    {
      "epoch": 0.62,
      "objective/train/avg_token_score": 0.029783912003040314,
      "objective/train/avg_weight": 0.9649562835693359,
      "objective/train/docs_used": 983277,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8072361946105957,
      "objective/train/std_weight": 0.10289938747882843,
      "objective/train/theoretical_loss": 3.471015158626335,
      "objective/train/tokens_used": 1760440800,
      "theoretical_loss": 3.471015158626335,
      "tokens_seen": 1739980800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026789530924910157,
      "loss": 2.6632,
      "theoretical_loss": 3.471015158626335,
      "tokens_seen": 1739980800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002678787592207301,
      "loss": 2.4347,
      "theoretical_loss": 3.471003957665147,
      "tokens_seen": 1740046336
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002678622091923586,
      "loss": 2.6854,
      "theoretical_loss": 3.470992757243934,
      "tokens_seen": 1740111872
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026784565916398713,
      "loss": 2.8208,
      "theoretical_loss": 3.4709815573626495,
      "tokens_seen": 1740177408
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026782910913561567,
      "loss": 2.6065,
      "theoretical_loss": 3.4709703580212485,
      "tokens_seen": 1740242944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002678125591072442,
      "loss": 2.8496,
      "theoretical_loss": 3.470959159219683,
      "tokens_seen": 1740308480
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002677960090788727,
      "loss": 2.879,
      "theoretical_loss": 3.4709479609579077,
      "tokens_seen": 1740374016
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002677794590505012,
      "loss": 2.8408,
      "theoretical_loss": 3.470936763235876,
      "tokens_seen": 1740439552
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026776290902212976,
      "loss": 2.6507,
      "theoretical_loss": 3.470925566053541,
      "tokens_seen": 1740505088
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002677463589937583,
      "loss": 2.4659,
      "theoretical_loss": 3.4709143694108575,
      "tokens_seen": 1740570624
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026772980896538684,
      "loss": 2.5015,
      "theoretical_loss": 3.4709031733077786,
      "tokens_seen": 1740636160
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002677132589370153,
      "loss": 2.7578,
      "theoretical_loss": 3.4708919777442575,
      "tokens_seen": 1740701696
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026769670890864386,
      "loss": 2.7493,
      "theoretical_loss": 3.470880782720249,
      "tokens_seen": 1740767232
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002676801588802724,
      "loss": 2.5421,
      "theoretical_loss": 3.470869588235705,
      "tokens_seen": 1740832768
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026766360885190093,
      "loss": 2.4963,
      "theoretical_loss": 3.4708583942905813,
      "tokens_seen": 1740898304
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002676470588235294,
      "loss": 2.8704,
      "theoretical_loss": 3.4708472008848297,
      "tokens_seen": 1740963840
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002676305087951579,
      "loss": 2.729,
      "theoretical_loss": 3.4708360080184057,
      "tokens_seen": 1741029376
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026761395876678644,
      "loss": 2.4048,
      "theoretical_loss": 3.4708248156912616,
      "tokens_seen": 1741094912
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000267597408738415,
      "loss": 2.5364,
      "theoretical_loss": 3.470813623903352,
      "tokens_seen": 1741160448
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002675808587100435,
      "loss": 2.6164,
      "theoretical_loss": 3.47080243265463,
      "tokens_seen": 1741225984
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000267564308681672,
      "loss": 2.5008,
      "theoretical_loss": 3.47079124194505,
      "tokens_seen": 1741291520
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026754775865330053,
      "loss": 2.6586,
      "theoretical_loss": 3.4707800517745646,
      "tokens_seen": 1741357056
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026753120862492907,
      "loss": 2.5399,
      "theoretical_loss": 3.470768862143129,
      "tokens_seen": 1741422592
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002675146585965576,
      "loss": 2.7033,
      "theoretical_loss": 3.470757673050696,
      "tokens_seen": 1741488128
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002674981085681861,
      "loss": 2.8282,
      "theoretical_loss": 3.4707464844972193,
      "tokens_seen": 1741553664
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026748155853981463,
      "loss": 2.4647,
      "theoretical_loss": 3.470735296482653,
      "tokens_seen": 1741619200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026746500851144317,
      "loss": 2.9314,
      "theoretical_loss": 3.470724109006951,
      "tokens_seen": 1741684736
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002674484584830717,
      "loss": 2.7746,
      "theoretical_loss": 3.470712922070067,
      "tokens_seen": 1741750272
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002674319084547002,
      "loss": 2.8571,
      "theoretical_loss": 3.4707017356719545,
      "tokens_seen": 1741815808
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002674153584263287,
      "loss": 3.014,
      "theoretical_loss": 3.4706905498125673,
      "tokens_seen": 1741881344
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026739880839795726,
      "loss": 2.4511,
      "theoretical_loss": 3.47067936449186,
      "tokens_seen": 1741946880
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002673822583695858,
      "loss": 2.6246,
      "theoretical_loss": 3.4706681797097847,
      "tokens_seen": 1742012416
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002673657083412143,
      "loss": 2.737,
      "theoretical_loss": 3.470656995466297,
      "tokens_seen": 1742077952
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002673491583128428,
      "loss": 2.6261,
      "theoretical_loss": 3.47064581176135,
      "tokens_seen": 1742143488
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026733260828447136,
      "loss": 2.5167,
      "theoretical_loss": 3.470634628594897,
      "tokens_seen": 1742209024
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002673160582560999,
      "loss": 2.4793,
      "theoretical_loss": 3.4706234459668925,
      "tokens_seen": 1742274560
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002672995082277284,
      "loss": 2.5937,
      "theoretical_loss": 3.47061226387729,
      "tokens_seen": 1742340096
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002672829581993569,
      "loss": 2.6879,
      "theoretical_loss": 3.470601082326043,
      "tokens_seen": 1742405632
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026726640817098545,
      "loss": 2.6462,
      "theoretical_loss": 3.4705899013131063,
      "tokens_seen": 1742471168
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000267249858142614,
      "loss": 2.4842,
      "theoretical_loss": 3.470578720838433,
      "tokens_seen": 1742536704
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002672333081142425,
      "loss": 2.3665,
      "theoretical_loss": 3.4705675409019774,
      "tokens_seen": 1742602240
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000267216758085871,
      "loss": 2.6693,
      "theoretical_loss": 3.4705563615036925,
      "tokens_seen": 1742667776
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026720020805749955,
      "loss": 2.713,
      "theoretical_loss": 3.4705451826435327,
      "tokens_seen": 1742733312
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002671836580291281,
      "loss": 2.9391,
      "theoretical_loss": 3.4705340043214523,
      "tokens_seen": 1742798848
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026716710800075657,
      "loss": 2.5328,
      "theoretical_loss": 3.4705228265374046,
      "tokens_seen": 1742864384
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026715055797238505,
      "loss": 2.7161,
      "theoretical_loss": 3.470511649291344,
      "tokens_seen": 1742929920
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002671340079440136,
      "loss": 2.7483,
      "theoretical_loss": 3.4705004725832236,
      "tokens_seen": 1742995456
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026711745791564213,
      "loss": 2.5924,
      "theoretical_loss": 3.470489296412998,
      "tokens_seen": 1743060992
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026710090788727067,
      "loss": 2.589,
      "theoretical_loss": 3.470478120780621,
      "tokens_seen": 1743126528
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026708435785889915,
      "loss": 2.8153,
      "theoretical_loss": 3.4704669456860455,
      "tokens_seen": 1743192064
    },
    {
      "epoch": 0.62,
      "objective/train/avg_token_score": 0.008784908801317215,
      "objective/train/avg_weight": 0.989657998085022,
      "objective/train/docs_used": 985168,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4265832901000977,
      "objective/train/std_weight": 0.046620722860097885,
      "objective/train/theoretical_loss": 3.4704557711292265,
      "objective/train/tokens_used": 1763717600,
      "theoretical_loss": 3.4704557711292265,
      "tokens_seen": 1743257600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002670678078305277,
      "loss": 2.4448,
      "theoretical_loss": 3.4704557711292265,
      "tokens_seen": 1743257600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002670512578021562,
      "loss": 2.7221,
      "theoretical_loss": 3.470444597110118,
      "tokens_seen": 1743323136
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026703470777378476,
      "loss": 2.4798,
      "theoretical_loss": 3.4704334236286734,
      "tokens_seen": 1743388672
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026701815774541325,
      "loss": 2.728,
      "theoretical_loss": 3.4704222506848463,
      "tokens_seen": 1743454208
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002670016077170418,
      "loss": 2.6388,
      "theoretical_loss": 3.470411078278591,
      "tokens_seen": 1743519744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002669850576886703,
      "loss": 2.5972,
      "theoretical_loss": 3.4703999064098623,
      "tokens_seen": 1743585280
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026696850766029886,
      "loss": 2.5561,
      "theoretical_loss": 3.470388735078613,
      "tokens_seen": 1743650816
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026695195763192734,
      "loss": 2.6491,
      "theoretical_loss": 3.4703775642847967,
      "tokens_seen": 1743716352
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002669354076035559,
      "loss": 2.7853,
      "theoretical_loss": 3.4703663940283684,
      "tokens_seen": 1743781888
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002669188575751844,
      "loss": 2.6136,
      "theoretical_loss": 3.470355224309282,
      "tokens_seen": 1743847424
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026690230754681295,
      "loss": 2.7829,
      "theoretical_loss": 3.470344055127491,
      "tokens_seen": 1743912960
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026688575751844144,
      "loss": 2.7391,
      "theoretical_loss": 3.4703328864829492,
      "tokens_seen": 1743978496
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026686920749007,
      "loss": 2.5655,
      "theoretical_loss": 3.4703217183756108,
      "tokens_seen": 1744044032
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002668526574616985,
      "loss": 2.972,
      "theoretical_loss": 3.47031055080543,
      "tokens_seen": 1744109568
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026683610743332705,
      "loss": 2.5597,
      "theoretical_loss": 3.470299383772361,
      "tokens_seen": 1744175104
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026681955740495553,
      "loss": 2.5481,
      "theoretical_loss": 3.470288217276357,
      "tokens_seen": 1744240640
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026680300737658407,
      "loss": 2.5282,
      "theoretical_loss": 3.4702770513173724,
      "tokens_seen": 1744306176
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002667864573482126,
      "loss": 2.5924,
      "theoretical_loss": 3.470265885895361,
      "tokens_seen": 1744371712
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026676990731984115,
      "loss": 2.7037,
      "theoretical_loss": 3.4702547210102774,
      "tokens_seen": 1744437248
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026675335729146963,
      "loss": 2.7095,
      "theoretical_loss": 3.470243556662075,
      "tokens_seen": 1744502784
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026673680726309817,
      "loss": 2.6528,
      "theoretical_loss": 3.470232392850708,
      "tokens_seen": 1744568320
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002667202572347267,
      "loss": 2.5937,
      "theoretical_loss": 3.470221229576131,
      "tokens_seen": 1744633856
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026670370720635524,
      "loss": 2.7921,
      "theoretical_loss": 3.4702100668382965,
      "tokens_seen": 1744699392
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002666871571779837,
      "loss": 2.6803,
      "theoretical_loss": 3.47019890463716,
      "tokens_seen": 1744764928
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002666706071496122,
      "loss": 2.5254,
      "theoretical_loss": 3.470187742972675,
      "tokens_seen": 1744830464
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026665405712124075,
      "loss": 2.3898,
      "theoretical_loss": 3.470176581844796,
      "tokens_seen": 1744896000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002666375070928693,
      "loss": 2.8164,
      "theoretical_loss": 3.4701654212534763,
      "tokens_seen": 1744961536
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002666209570644978,
      "loss": 2.6097,
      "theoretical_loss": 3.4701542611986698,
      "tokens_seen": 1745027072
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002666044070361263,
      "loss": 2.6033,
      "theoretical_loss": 3.470143101680332,
      "tokens_seen": 1745092608
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026658785700775484,
      "loss": 2.5668,
      "theoretical_loss": 3.4701319426984156,
      "tokens_seen": 1745158144
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002665713069793834,
      "loss": 2.5901,
      "theoretical_loss": 3.4701207842528747,
      "tokens_seen": 1745223680
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002665547569510119,
      "loss": 2.3489,
      "theoretical_loss": 3.470109626343664,
      "tokens_seen": 1745289216
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002665382069226404,
      "loss": 2.5419,
      "theoretical_loss": 3.470098468970738,
      "tokens_seen": 1745354752
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026652165689426894,
      "loss": 2.9568,
      "theoretical_loss": 3.470087312134049,
      "tokens_seen": 1745420288
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002665051068658975,
      "loss": 2.73,
      "theoretical_loss": 3.470076155833553,
      "tokens_seen": 1745485824
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000266488556837526,
      "loss": 2.6517,
      "theoretical_loss": 3.4700650000692033,
      "tokens_seen": 1745551360
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026647200680915455,
      "loss": 2.8967,
      "theoretical_loss": 3.4700538448409537,
      "tokens_seen": 1745616896
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026645545678078303,
      "loss": 2.7541,
      "theoretical_loss": 3.470042690148759,
      "tokens_seen": 1745682432
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026643890675241157,
      "loss": 2.677,
      "theoretical_loss": 3.4700315359925726,
      "tokens_seen": 1745747968
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002664223567240401,
      "loss": 2.6656,
      "theoretical_loss": 3.470020382372349,
      "tokens_seen": 1745813504
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026640580669566865,
      "loss": 2.6017,
      "theoretical_loss": 3.470009229288043,
      "tokens_seen": 1745879040
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026638925666729713,
      "loss": 2.7569,
      "theoretical_loss": 3.469998076739607,
      "tokens_seen": 1745944576
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026637270663892567,
      "loss": 2.5449,
      "theoretical_loss": 3.469986924726997,
      "tokens_seen": 1746010112
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002663561566105542,
      "loss": 2.7292,
      "theoretical_loss": 3.469975773250166,
      "tokens_seen": 1746075648
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026633960658218274,
      "loss": 2.774,
      "theoretical_loss": 3.4699646223090688,
      "tokens_seen": 1746141184
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002663230565538112,
      "loss": 2.6351,
      "theoretical_loss": 3.469953471903659,
      "tokens_seen": 1746206720
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026630650652543976,
      "loss": 2.6618,
      "theoretical_loss": 3.469942322033891,
      "tokens_seen": 1746272256
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002662899564970683,
      "loss": 2.6812,
      "theoretical_loss": 3.4699311726997193,
      "tokens_seen": 1746337792
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026627340646869684,
      "loss": 2.8801,
      "theoretical_loss": 3.469920023901097,
      "tokens_seen": 1746403328
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002662568564403253,
      "loss": 2.4766,
      "theoretical_loss": 3.46990887563798,
      "tokens_seen": 1746468864
    },
    {
      "epoch": 0.62,
      "objective/train/avg_token_score": 0.023078057914972305,
      "objective/train/avg_weight": 0.9728617072105408,
      "objective/train/docs_used": 987031,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8473329544067383,
      "objective/train/std_weight": 0.13332003355026245,
      "objective/train/theoretical_loss": 3.469897727910321,
      "objective/train/tokens_used": 1766994400,
      "theoretical_loss": 3.469897727910321,
      "tokens_seen": 1746534400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026624030641195386,
      "loss": 2.6364,
      "theoretical_loss": 3.469897727910321,
      "tokens_seen": 1746534400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002662237563835824,
      "loss": 2.6662,
      "theoretical_loss": 3.4698865807180748,
      "tokens_seen": 1746599936
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026620720635521093,
      "loss": 2.9229,
      "theoretical_loss": 3.4698754340611955,
      "tokens_seen": 1746665472
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002661906563268394,
      "loss": 2.9411,
      "theoretical_loss": 3.4698642879396373,
      "tokens_seen": 1746731008
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002661741062984679,
      "loss": 2.729,
      "theoretical_loss": 3.4698531423533545,
      "tokens_seen": 1746796544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026615755627009644,
      "loss": 2.7192,
      "theoretical_loss": 3.469841997302301,
      "tokens_seen": 1746862080
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000266141006241725,
      "loss": 2.523,
      "theoretical_loss": 3.4698308527864317,
      "tokens_seen": 1746927616
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002661244562133535,
      "loss": 2.674,
      "theoretical_loss": 3.4698197088056997,
      "tokens_seen": 1746993152
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000266107906184982,
      "loss": 2.5959,
      "theoretical_loss": 3.4698085653600605,
      "tokens_seen": 1747058688
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026609135615661054,
      "loss": 2.7354,
      "theoretical_loss": 3.469797422449467,
      "tokens_seen": 1747124224
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026607480612823907,
      "loss": 2.7305,
      "theoretical_loss": 3.4697862800738752,
      "tokens_seen": 1747189760
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002660582560998676,
      "loss": 2.6636,
      "theoretical_loss": 3.4697751382332376,
      "tokens_seen": 1747255296
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002660417060714961,
      "loss": 2.4408,
      "theoretical_loss": 3.46976399692751,
      "tokens_seen": 1747320832
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026602515604312463,
      "loss": 2.8431,
      "theoretical_loss": 3.469752856156645,
      "tokens_seen": 1747386368
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026600860601475317,
      "loss": 2.603,
      "theoretical_loss": 3.469741715920598,
      "tokens_seen": 1747451904
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002659920559863817,
      "loss": 2.7315,
      "theoretical_loss": 3.4697305762193227,
      "tokens_seen": 1747517440
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002659755059580102,
      "loss": 2.694,
      "theoretical_loss": 3.469719437052774,
      "tokens_seen": 1747582976
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002659589559296387,
      "loss": 2.4533,
      "theoretical_loss": 3.4697082984209056,
      "tokens_seen": 1747648512
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026594240590126726,
      "loss": 2.9317,
      "theoretical_loss": 3.469697160323672,
      "tokens_seen": 1747714048
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002659258558728958,
      "loss": 2.793,
      "theoretical_loss": 3.469686022761028,
      "tokens_seen": 1747779584
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002659093058445243,
      "loss": 2.6206,
      "theoretical_loss": 3.469674885732927,
      "tokens_seen": 1747845120
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002658927558161528,
      "loss": 2.6656,
      "theoretical_loss": 3.4696637492393236,
      "tokens_seen": 1747910656
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026587620578778136,
      "loss": 2.6382,
      "theoretical_loss": 3.4696526132801724,
      "tokens_seen": 1747976192
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002658596557594099,
      "loss": 2.7602,
      "theoretical_loss": 3.4696414778554274,
      "tokens_seen": 1748041728
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002658431057310384,
      "loss": 2.8477,
      "theoretical_loss": 3.4696303429650435,
      "tokens_seen": 1748107264
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002658265557026669,
      "loss": 2.7991,
      "theoretical_loss": 3.469619208608974,
      "tokens_seen": 1748172800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026581000567429546,
      "loss": 2.7645,
      "theoretical_loss": 3.469608074787174,
      "tokens_seen": 1748238336
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000265793455645924,
      "loss": 2.7959,
      "theoretical_loss": 3.4695969414995975,
      "tokens_seen": 1748303872
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002657769056175525,
      "loss": 2.6343,
      "theoretical_loss": 3.4695858087461993,
      "tokens_seen": 1748369408
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000265760355589181,
      "loss": 2.7186,
      "theoretical_loss": 3.469574676526933,
      "tokens_seen": 1748434944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026574380556080955,
      "loss": 2.5119,
      "theoretical_loss": 3.469563544841754,
      "tokens_seen": 1748500480
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002657272555324381,
      "loss": 2.8509,
      "theoretical_loss": 3.4695524136906153,
      "tokens_seen": 1748566016
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002657107055040666,
      "loss": 2.7755,
      "theoretical_loss": 3.4695412830734726,
      "tokens_seen": 1748631552
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026569415547569506,
      "loss": 2.4062,
      "theoretical_loss": 3.46953015299028,
      "tokens_seen": 1748697088
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002656776054473236,
      "loss": 2.3973,
      "theoretical_loss": 3.469519023440991,
      "tokens_seen": 1748762624
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026566105541895213,
      "loss": 2.5955,
      "theoretical_loss": 3.4695078944255604,
      "tokens_seen": 1748828160
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026564450539058067,
      "loss": 2.9632,
      "theoretical_loss": 3.4694967659439433,
      "tokens_seen": 1748893696
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026562795536220915,
      "loss": 2.7061,
      "theoretical_loss": 3.469485637996093,
      "tokens_seen": 1748959232
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002656114053338377,
      "loss": 2.7662,
      "theoretical_loss": 3.4694745105819647,
      "tokens_seen": 1749024768
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026559485530546623,
      "loss": 2.7377,
      "theoretical_loss": 3.4694633837015125,
      "tokens_seen": 1749090304
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026557830527709477,
      "loss": 2.8145,
      "theoretical_loss": 3.469452257354691,
      "tokens_seen": 1749155840
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026556175524872325,
      "loss": 2.4527,
      "theoretical_loss": 3.469441131541454,
      "tokens_seen": 1749221376
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002655452052203518,
      "loss": 2.6434,
      "theoretical_loss": 3.4694300062617565,
      "tokens_seen": 1749286912
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002655286551919803,
      "loss": 2.5293,
      "theoretical_loss": 3.469418881515553,
      "tokens_seen": 1749352448
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026551210516360886,
      "loss": 2.693,
      "theoretical_loss": 3.4694077573027977,
      "tokens_seen": 1749417984
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026549555513523734,
      "loss": 2.7624,
      "theoretical_loss": 3.469396633623445,
      "tokens_seen": 1749483520
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002654790051068659,
      "loss": 2.6946,
      "theoretical_loss": 3.4693855104774496,
      "tokens_seen": 1749549056
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002654624550784944,
      "loss": 2.1379,
      "theoretical_loss": 3.4693743878647654,
      "tokens_seen": 1749614592
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026544590505012296,
      "loss": 2.5729,
      "theoretical_loss": 3.469363265785348,
      "tokens_seen": 1749680128
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026542935502175144,
      "loss": 2.6081,
      "theoretical_loss": 3.46935214423915,
      "tokens_seen": 1749745664
    },
    {
      "epoch": 0.62,
      "objective/train/avg_token_score": 0.014407171867787838,
      "objective/train/avg_weight": 0.983039379119873,
      "objective/train/docs_used": 989075,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.976675510406494,
      "objective/train/std_weight": 0.07632457464933395,
      "objective/train/theoretical_loss": 3.4693410232261277,
      "objective/train/tokens_used": 1770271200,
      "theoretical_loss": 3.4693410232261277,
      "tokens_seen": 1749811200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026541280499338,
      "loss": 2.8041,
      "theoretical_loss": 3.4693410232261277,
      "tokens_seen": 1749811200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002653962549650085,
      "loss": 2.5499,
      "theoretical_loss": 3.4693299027462343,
      "tokens_seen": 1749876736
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00026537970493663705,
      "loss": 2.6004,
      "theoretical_loss": 3.469318782799425,
      "tokens_seen": 1749942272
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026536315490826554,
      "loss": 2.7005,
      "theoretical_loss": 3.4693076633856545,
      "tokens_seen": 1750007808
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002653466048798941,
      "loss": 3.0108,
      "theoretical_loss": 3.4692965445048767,
      "tokens_seen": 1750073344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002653300548515226,
      "loss": 2.7543,
      "theoretical_loss": 3.469285426157046,
      "tokens_seen": 1750138880
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026531350482315115,
      "loss": 2.5612,
      "theoretical_loss": 3.469274308342117,
      "tokens_seen": 1750204416
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026529695479477963,
      "loss": 2.5911,
      "theoretical_loss": 3.469263191060045,
      "tokens_seen": 1750269952
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026528040476640817,
      "loss": 2.0876,
      "theoretical_loss": 3.4692520743107833,
      "tokens_seen": 1750335488
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002652638547380367,
      "loss": 2.8016,
      "theoretical_loss": 3.469240958094287,
      "tokens_seen": 1750401024
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026524730470966525,
      "loss": 2.6844,
      "theoretical_loss": 3.469229842410511,
      "tokens_seen": 1750466560
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026523075468129373,
      "loss": 2.6556,
      "theoretical_loss": 3.4692187272594093,
      "tokens_seen": 1750532096
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026521420465292227,
      "loss": 2.7675,
      "theoretical_loss": 3.469207612640936,
      "tokens_seen": 1750597632
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026519765462455075,
      "loss": 2.5744,
      "theoretical_loss": 3.469196498555047,
      "tokens_seen": 1750663168
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002651811045961793,
      "loss": 2.4952,
      "theoretical_loss": 3.469185385001696,
      "tokens_seen": 1750728704
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002651645545678078,
      "loss": 2.6496,
      "theoretical_loss": 3.4691742719808376,
      "tokens_seen": 1750794240
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026514800453943636,
      "loss": 2.5835,
      "theoretical_loss": 3.469163159492426,
      "tokens_seen": 1750859776
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026513145451106485,
      "loss": 2.5829,
      "theoretical_loss": 3.469152047536416,
      "tokens_seen": 1750925312
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002651149044826934,
      "loss": 2.686,
      "theoretical_loss": 3.4691409361127628,
      "tokens_seen": 1750990848
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002650983544543219,
      "loss": 3.07,
      "theoretical_loss": 3.46912982522142,
      "tokens_seen": 1751056384
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026508180442595046,
      "loss": 2.6128,
      "theoretical_loss": 3.469118714862343,
      "tokens_seen": 1751121920
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026506525439757894,
      "loss": 2.5061,
      "theoretical_loss": 3.469107605035486,
      "tokens_seen": 1751187456
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002650487043692075,
      "loss": 2.8806,
      "theoretical_loss": 3.4690964957408035,
      "tokens_seen": 1751252992
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000265032154340836,
      "loss": 2.6998,
      "theoretical_loss": 3.46908538697825,
      "tokens_seen": 1751318528
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026501560431246455,
      "loss": 2.7471,
      "theoretical_loss": 3.4690742787477804,
      "tokens_seen": 1751384064
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026499905428409304,
      "loss": 2.1913,
      "theoretical_loss": 3.4690631710493496,
      "tokens_seen": 1751449600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002649825042557216,
      "loss": 2.7491,
      "theoretical_loss": 3.469052063882912,
      "tokens_seen": 1751515136
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002649659542273501,
      "loss": 2.2142,
      "theoretical_loss": 3.4690409572484215,
      "tokens_seen": 1751580672
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026494940419897865,
      "loss": 2.6502,
      "theoretical_loss": 3.469029851145833,
      "tokens_seen": 1751646208
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026493285417060713,
      "loss": 2.8502,
      "theoretical_loss": 3.469018745575102,
      "tokens_seen": 1751711744
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026491630414223567,
      "loss": 2.8482,
      "theoretical_loss": 3.469007640536182,
      "tokens_seen": 1751777280
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002648997541138642,
      "loss": 2.5326,
      "theoretical_loss": 3.4689965360290285,
      "tokens_seen": 1751842816
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026488320408549275,
      "loss": 2.6753,
      "theoretical_loss": 3.4689854320535956,
      "tokens_seen": 1751908352
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026486665405712123,
      "loss": 2.746,
      "theoretical_loss": 3.4689743286098382,
      "tokens_seen": 1751973888
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026485010402874977,
      "loss": 2.8015,
      "theoretical_loss": 3.468963225697711,
      "tokens_seen": 1752039424
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002648335540003783,
      "loss": 2.8148,
      "theoretical_loss": 3.4689521233171687,
      "tokens_seen": 1752104960
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026481700397200684,
      "loss": 2.6307,
      "theoretical_loss": 3.468941021468166,
      "tokens_seen": 1752170496
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002648004539436353,
      "loss": 2.6094,
      "theoretical_loss": 3.468929920150657,
      "tokens_seen": 1752236032
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026478390391526386,
      "loss": 2.5473,
      "theoretical_loss": 3.468918819364597,
      "tokens_seen": 1752301568
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002647673538868924,
      "loss": 2.805,
      "theoretical_loss": 3.4689077191099402,
      "tokens_seen": 1752367104
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026475080385852094,
      "loss": 2.7298,
      "theoretical_loss": 3.4688966193866415,
      "tokens_seen": 1752432640
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002647342538301494,
      "loss": 2.6057,
      "theoretical_loss": 3.468885520194656,
      "tokens_seen": 1752498176
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026471770380177796,
      "loss": 2.7677,
      "theoretical_loss": 3.468874421533938,
      "tokens_seen": 1752563712
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026470115377340644,
      "loss": 2.6859,
      "theoretical_loss": 3.4688633234044426,
      "tokens_seen": 1752629248
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000264684603745035,
      "loss": 2.5272,
      "theoretical_loss": 3.468852225806123,
      "tokens_seen": 1752694784
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002646680537166635,
      "loss": 2.6306,
      "theoretical_loss": 3.4688411287389362,
      "tokens_seen": 1752760320
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000264651503688292,
      "loss": 2.4835,
      "theoretical_loss": 3.4688300322028356,
      "tokens_seen": 1752825856
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026463495365992054,
      "loss": 2.6301,
      "theoretical_loss": 3.468818936197776,
      "tokens_seen": 1752891392
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002646184036315491,
      "loss": 2.6944,
      "theoretical_loss": 3.4688078407237124,
      "tokens_seen": 1752956928
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002646018536031776,
      "loss": 2.8341,
      "theoretical_loss": 3.4687967457805993,
      "tokens_seen": 1753022464
    },
    {
      "epoch": 0.63,
      "objective/train/avg_token_score": 0.02127518132328987,
      "objective/train/avg_weight": 0.9749829769134521,
      "objective/train/docs_used": 991265,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 8.53312873840332,
      "objective/train/std_weight": 0.13962100446224213,
      "objective/train/theoretical_loss": 3.4687856513683917,
      "objective/train/tokens_used": 1773548000,
      "theoretical_loss": 3.4687856513683917,
      "tokens_seen": 1753088000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002645853035748061,
      "loss": 2.2543,
      "theoretical_loss": 3.4687856513683917,
      "tokens_seen": 1753088000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026456875354643463,
      "loss": 2.4892,
      "theoretical_loss": 3.468774557487044,
      "tokens_seen": 1753153536
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026455220351806317,
      "loss": 2.8567,
      "theoretical_loss": 3.4687634641365115,
      "tokens_seen": 1753219072
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002645356534896917,
      "loss": 2.9138,
      "theoretical_loss": 3.4687523713167483,
      "tokens_seen": 1753284608
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002645191034613202,
      "loss": 2.5571,
      "theoretical_loss": 3.468741279027709,
      "tokens_seen": 1753350144
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026450255343294873,
      "loss": 2.7294,
      "theoretical_loss": 3.4687301872693497,
      "tokens_seen": 1753415680
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026448600340457727,
      "loss": 2.8107,
      "theoretical_loss": 3.4687190960416245,
      "tokens_seen": 1753481216
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002644694533762058,
      "loss": 2.8285,
      "theoretical_loss": 3.468708005344487,
      "tokens_seen": 1753546752
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002644529033478343,
      "loss": 2.6268,
      "theoretical_loss": 3.468696915177894,
      "tokens_seen": 1753612288
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002644363533194628,
      "loss": 2.7864,
      "theoretical_loss": 3.4686858255417987,
      "tokens_seen": 1753677824
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026441980329109136,
      "loss": 2.8292,
      "theoretical_loss": 3.4686747364361565,
      "tokens_seen": 1753743360
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002644032532627199,
      "loss": 2.6501,
      "theoretical_loss": 3.4686636478609225,
      "tokens_seen": 1753808896
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002643867032343484,
      "loss": 2.7766,
      "theoretical_loss": 3.4686525598160514,
      "tokens_seen": 1753874432
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002643701532059769,
      "loss": 2.6621,
      "theoretical_loss": 3.4686414723014973,
      "tokens_seen": 1753939968
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026435360317760546,
      "loss": 2.6173,
      "theoretical_loss": 3.4686303853172165,
      "tokens_seen": 1754005504
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000264337053149234,
      "loss": 2.4558,
      "theoretical_loss": 3.468619298863162,
      "tokens_seen": 1754071040
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002643205031208625,
      "loss": 2.7939,
      "theoretical_loss": 3.46860821293929,
      "tokens_seen": 1754136576
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000264303953092491,
      "loss": 2.7378,
      "theoretical_loss": 3.4685971275455545,
      "tokens_seen": 1754202112
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026428740306411956,
      "loss": 2.8209,
      "theoretical_loss": 3.4685860426819106,
      "tokens_seen": 1754267648
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002642708530357481,
      "loss": 2.3809,
      "theoretical_loss": 3.468574958348314,
      "tokens_seen": 1754333184
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002642543030073766,
      "loss": 2.7015,
      "theoretical_loss": 3.468563874544718,
      "tokens_seen": 1754398720
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002642377529790051,
      "loss": 2.8976,
      "theoretical_loss": 3.4685527912710787,
      "tokens_seen": 1754464256
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026422120295063365,
      "loss": 2.7767,
      "theoretical_loss": 3.4685417085273507,
      "tokens_seen": 1754529792
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026420465292226213,
      "loss": 2.7883,
      "theoretical_loss": 3.468530626313488,
      "tokens_seen": 1754595328
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026418810289389067,
      "loss": 2.5516,
      "theoretical_loss": 3.468519544629447,
      "tokens_seen": 1754660864
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026417155286551916,
      "loss": 2.7987,
      "theoretical_loss": 3.468508463475181,
      "tokens_seen": 1754726400
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002641550028371477,
      "loss": 2.6693,
      "theoretical_loss": 3.4684973828506465,
      "tokens_seen": 1754791936
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026413845280877623,
      "loss": 2.1557,
      "theoretical_loss": 3.468486302755797,
      "tokens_seen": 1754857472
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026412190278040477,
      "loss": 2.482,
      "theoretical_loss": 3.4684752231905875,
      "tokens_seen": 1754923008
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026410535275203325,
      "loss": 2.4389,
      "theoretical_loss": 3.468464144154974,
      "tokens_seen": 1754988544
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002640888027236618,
      "loss": 2.8167,
      "theoretical_loss": 3.468453065648911,
      "tokens_seen": 1755054080
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002640722526952903,
      "loss": 2.7038,
      "theoretical_loss": 3.468441987672353,
      "tokens_seen": 1755119616
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026405570266691886,
      "loss": 2.7477,
      "theoretical_loss": 3.468430910225255,
      "tokens_seen": 1755185152
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026403915263854735,
      "loss": 2.7703,
      "theoretical_loss": 3.468419833307572,
      "tokens_seen": 1755250688
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002640226026101759,
      "loss": 2.4705,
      "theoretical_loss": 3.468408756919259,
      "tokens_seen": 1755316224
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002640060525818044,
      "loss": 2.6087,
      "theoretical_loss": 3.468397681060271,
      "tokens_seen": 1755381760
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026398950255343296,
      "loss": 2.7427,
      "theoretical_loss": 3.4683866057305623,
      "tokens_seen": 1755447296
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026397295252506144,
      "loss": 2.5523,
      "theoretical_loss": 3.468375530930089,
      "tokens_seen": 1755512832
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026395640249669,
      "loss": 2.7269,
      "theoretical_loss": 3.468364456658805,
      "tokens_seen": 1755578368
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002639398524683185,
      "loss": 2.6571,
      "theoretical_loss": 3.4683533829166664,
      "tokens_seen": 1755643904
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026392330243994706,
      "loss": 2.7059,
      "theoretical_loss": 3.468342309703627,
      "tokens_seen": 1755709440
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026390675241157554,
      "loss": 2.6034,
      "theoretical_loss": 3.4683312370196426,
      "tokens_seen": 1755774976
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002638902023832041,
      "loss": 2.7259,
      "theoretical_loss": 3.4683201648646675,
      "tokens_seen": 1755840512
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002638736523548326,
      "loss": 2.7995,
      "theoretical_loss": 3.4683090932386573,
      "tokens_seen": 1755906048
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026385710232646115,
      "loss": 2.5456,
      "theoretical_loss": 3.4682980221415662,
      "tokens_seen": 1755971584
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026384055229808964,
      "loss": 2.707,
      "theoretical_loss": 3.46828695157335,
      "tokens_seen": 1756037120
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026382400226971817,
      "loss": 2.7373,
      "theoretical_loss": 3.4682758815339634,
      "tokens_seen": 1756102656
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002638074522413467,
      "loss": 2.603,
      "theoretical_loss": 3.4682648120233615,
      "tokens_seen": 1756168192
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026379090221297525,
      "loss": 2.6396,
      "theoretical_loss": 3.4682537430414992,
      "tokens_seen": 1756233728
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026377435218460373,
      "loss": 2.7092,
      "theoretical_loss": 3.468242674588331,
      "tokens_seen": 1756299264
    },
    {
      "epoch": 0.63,
      "objective/train/avg_token_score": 0.013334047049283981,
      "objective/train/avg_weight": 0.9843087196350098,
      "objective/train/docs_used": 993214,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.974987030029297,
      "objective/train/std_weight": 0.07203251868486404,
      "objective/train/theoretical_loss": 3.468231606663813,
      "objective/train/tokens_used": 1776824800,
      "theoretical_loss": 3.468231606663813,
      "tokens_seen": 1756364800
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026375780215623227,
      "loss": 2.88,
      "theoretical_loss": 3.468231606663813,
      "tokens_seen": 1756364800
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002637412521278608,
      "loss": 2.9094,
      "theoretical_loss": 3.4682205392678993,
      "tokens_seen": 1756430336
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002637247020994893,
      "loss": 2.6188,
      "theoretical_loss": 3.4682094724005457,
      "tokens_seen": 1756495872
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026370815207111783,
      "loss": 2.5925,
      "theoretical_loss": 3.4681984060617066,
      "tokens_seen": 1756561408
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026369160204274636,
      "loss": 2.8647,
      "theoretical_loss": 3.4681873402513372,
      "tokens_seen": 1756626944
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026367505201437485,
      "loss": 2.6872,
      "theoretical_loss": 3.4681762749693927,
      "tokens_seen": 1756692480
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002636585019860034,
      "loss": 2.6205,
      "theoretical_loss": 3.4681652102158282,
      "tokens_seen": 1756758016
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002636419519576319,
      "loss": 2.4389,
      "theoretical_loss": 3.4681541459905985,
      "tokens_seen": 1756823552
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026362540192926046,
      "loss": 2.8089,
      "theoretical_loss": 3.468143082293659,
      "tokens_seen": 1756889088
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026360885190088894,
      "loss": 2.652,
      "theoretical_loss": 3.4681320191249645,
      "tokens_seen": 1756954624
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002635923018725175,
      "loss": 2.6068,
      "theoretical_loss": 3.46812095648447,
      "tokens_seen": 1757020160
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000263575751844146,
      "loss": 2.8169,
      "theoretical_loss": 3.468109894372131,
      "tokens_seen": 1757085696
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026355920181577456,
      "loss": 2.765,
      "theoretical_loss": 3.468098832787902,
      "tokens_seen": 1757151232
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026354265178740304,
      "loss": 2.5329,
      "theoretical_loss": 3.4680877717317387,
      "tokens_seen": 1757216768
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002635261017590316,
      "loss": 2.6479,
      "theoretical_loss": 3.4680767112035955,
      "tokens_seen": 1757282304
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002635095517306601,
      "loss": 2.5146,
      "theoretical_loss": 3.4680656512034282,
      "tokens_seen": 1757347840
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026349300170228865,
      "loss": 2.6406,
      "theoretical_loss": 3.4680545917311916,
      "tokens_seen": 1757413376
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026347645167391714,
      "loss": 2.8481,
      "theoretical_loss": 3.468043532786841,
      "tokens_seen": 1757478912
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002634599016455457,
      "loss": 2.6349,
      "theoretical_loss": 3.4680324743703315,
      "tokens_seen": 1757544448
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002634433516171742,
      "loss": 2.7709,
      "theoretical_loss": 3.4680214164816174,
      "tokens_seen": 1757609984
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026342680158880275,
      "loss": 2.654,
      "theoretical_loss": 3.468010359120655,
      "tokens_seen": 1757675520
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026341025156043123,
      "loss": 2.8241,
      "theoretical_loss": 3.467999302287399,
      "tokens_seen": 1757741056
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026339370153205977,
      "loss": 2.652,
      "theoretical_loss": 3.4679882459818043,
      "tokens_seen": 1757806592
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002633771515036883,
      "loss": 2.6681,
      "theoretical_loss": 3.4679771902038263,
      "tokens_seen": 1757872128
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026336060147531684,
      "loss": 2.5301,
      "theoretical_loss": 3.4679661349534197,
      "tokens_seen": 1757937664
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026334405144694533,
      "loss": 2.5047,
      "theoretical_loss": 3.4679550802305408,
      "tokens_seen": 1758003200
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026332750141857387,
      "loss": 2.7973,
      "theoretical_loss": 3.467944026035144,
      "tokens_seen": 1758068736
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002633109513902024,
      "loss": 2.4419,
      "theoretical_loss": 3.4679329723671835,
      "tokens_seen": 1758134272
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026329440136183094,
      "loss": 2.6497,
      "theoretical_loss": 3.467921919226616,
      "tokens_seen": 1758199808
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002632778513334594,
      "loss": 2.5964,
      "theoretical_loss": 3.4679108666133964,
      "tokens_seen": 1758265344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026326130130508796,
      "loss": 2.7423,
      "theoretical_loss": 3.4678998145274793,
      "tokens_seen": 1758330880
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002632447512767165,
      "loss": 2.2031,
      "theoretical_loss": 3.467888762968821,
      "tokens_seen": 1758396416
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000263228201248345,
      "loss": 2.5817,
      "theoretical_loss": 3.467877711937375,
      "tokens_seen": 1758461952
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002632116512199735,
      "loss": 2.6571,
      "theoretical_loss": 3.4678666614330975,
      "tokens_seen": 1758527488
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000263195101191602,
      "loss": 2.7935,
      "theoretical_loss": 3.467855611455944,
      "tokens_seen": 1758593024
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026317855116323054,
      "loss": 2.5563,
      "theoretical_loss": 3.467844562005869,
      "tokens_seen": 1758658560
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002631620011348591,
      "loss": 2.6537,
      "theoretical_loss": 3.4678335130828284,
      "tokens_seen": 1758724096
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002631454511064876,
      "loss": 2.7802,
      "theoretical_loss": 3.4678224646867766,
      "tokens_seen": 1758789632
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002631289010781161,
      "loss": 2.6609,
      "theoretical_loss": 3.4678114168176695,
      "tokens_seen": 1758855168
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026311235104974464,
      "loss": 2.6305,
      "theoretical_loss": 3.467800369475462,
      "tokens_seen": 1758920704
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002630958010213732,
      "loss": 2.6149,
      "theoretical_loss": 3.46778932266011,
      "tokens_seen": 1758986240
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002630792509930017,
      "loss": 2.7564,
      "theoretical_loss": 3.467778276371568,
      "tokens_seen": 1759051776
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002630627009646302,
      "loss": 2.6542,
      "theoretical_loss": 3.467767230609791,
      "tokens_seen": 1759117312
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026304615093625873,
      "loss": 2.7396,
      "theoretical_loss": 3.4677561853747356,
      "tokens_seen": 1759182848
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026302960090788727,
      "loss": 2.4345,
      "theoretical_loss": 3.4677451406663558,
      "tokens_seen": 1759248384
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002630130508795158,
      "loss": 2.6244,
      "theoretical_loss": 3.467734096484607,
      "tokens_seen": 1759313920
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002629965008511443,
      "loss": 2.7684,
      "theoretical_loss": 3.467723052829445,
      "tokens_seen": 1759379456
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026297995082277283,
      "loss": 2.6213,
      "theoretical_loss": 3.4677120097008247,
      "tokens_seen": 1759444992
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026296340079440137,
      "loss": 2.4699,
      "theoretical_loss": 3.4677009670987013,
      "tokens_seen": 1759510528
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002629468507660299,
      "loss": 2.7313,
      "theoretical_loss": 3.4676899250230306,
      "tokens_seen": 1759576064
    },
    {
      "epoch": 0.63,
      "objective/train/avg_token_score": 0.011843722313642502,
      "objective/train/avg_weight": 0.9860578775405884,
      "objective/train/docs_used": 995165,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.632314682006836,
      "objective/train/std_weight": 0.06644388288259506,
      "objective/train/theoretical_loss": 3.4676788834737677,
      "objective/train/tokens_used": 1780101600,
      "theoretical_loss": 3.4676788834737677,
      "tokens_seen": 1759641600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002629303007376584,
      "loss": 2.543,
      "theoretical_loss": 3.4676788834737677,
      "tokens_seen": 1759641600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002629137507092869,
      "loss": 2.3913,
      "theoretical_loss": 3.467667842450868,
      "tokens_seen": 1759707136
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026289720068091546,
      "loss": 2.6932,
      "theoretical_loss": 3.467656801954286,
      "tokens_seen": 1759772672
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000262880650652544,
      "loss": 2.3676,
      "theoretical_loss": 3.4676457619839773,
      "tokens_seen": 1759838208
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002628641006241725,
      "loss": 2.7627,
      "theoretical_loss": 3.4676347225398985,
      "tokens_seen": 1759903744
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000262847550595801,
      "loss": 2.8831,
      "theoretical_loss": 3.467623683622003,
      "tokens_seen": 1759969280
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026283100056742956,
      "loss": 2.4403,
      "theoretical_loss": 3.467612645230248,
      "tokens_seen": 1760034816
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002628144505390581,
      "loss": 2.5202,
      "theoretical_loss": 3.4676016073645872,
      "tokens_seen": 1760100352
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002627979005106866,
      "loss": 2.7938,
      "theoretical_loss": 3.467590570024977,
      "tokens_seen": 1760165888
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002627813504823151,
      "loss": 2.7006,
      "theoretical_loss": 3.4675795332113726,
      "tokens_seen": 1760231424
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026276480045394365,
      "loss": 2.7179,
      "theoretical_loss": 3.467568496923729,
      "tokens_seen": 1760296960
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002627482504255722,
      "loss": 2.6466,
      "theoretical_loss": 3.4675574611620013,
      "tokens_seen": 1760362496
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002627317003972007,
      "loss": 2.6512,
      "theoretical_loss": 3.4675464259261455,
      "tokens_seen": 1760428032
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026271515036882916,
      "loss": 2.7441,
      "theoretical_loss": 3.467535391216116,
      "tokens_seen": 1760493568
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002626986003404577,
      "loss": 2.1806,
      "theoretical_loss": 3.46752435703187,
      "tokens_seen": 1760559104
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026268205031208623,
      "loss": 2.1626,
      "theoretical_loss": 3.4675133233733613,
      "tokens_seen": 1760624640
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026266550028371477,
      "loss": 2.7024,
      "theoretical_loss": 3.467502290240546,
      "tokens_seen": 1760690176
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026264895025534325,
      "loss": 2.7377,
      "theoretical_loss": 3.4674912576333785,
      "tokens_seen": 1760755712
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002626324002269718,
      "loss": 2.4768,
      "theoretical_loss": 3.4674802255518156,
      "tokens_seen": 1760821248
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026261585019860033,
      "loss": 2.9,
      "theoretical_loss": 3.467469193995812,
      "tokens_seen": 1760886784
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026259930017022887,
      "loss": 2.7697,
      "theoretical_loss": 3.4674581629653227,
      "tokens_seen": 1760952320
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026258275014185735,
      "loss": 2.723,
      "theoretical_loss": 3.467447132460304,
      "tokens_seen": 1761017856
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002625662001134859,
      "loss": 2.6618,
      "theoretical_loss": 3.467436102480711,
      "tokens_seen": 1761083392
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002625496500851144,
      "loss": 2.8525,
      "theoretical_loss": 3.4674250730264986,
      "tokens_seen": 1761148928
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026253310005674296,
      "loss": 2.5963,
      "theoretical_loss": 3.4674140440976227,
      "tokens_seen": 1761214464
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026251655002837145,
      "loss": 2.5539,
      "theoretical_loss": 3.467403015694039,
      "tokens_seen": 1761280000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002625,
      "loss": 2.4742,
      "theoretical_loss": 3.467391987815702,
      "tokens_seen": 1761345536
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002624834499716285,
      "loss": 2.733,
      "theoretical_loss": 3.4673809604625676,
      "tokens_seen": 1761411072
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026246689994325706,
      "loss": 2.6743,
      "theoretical_loss": 3.4673699336345916,
      "tokens_seen": 1761476608
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002624503499148856,
      "loss": 2.5626,
      "theoretical_loss": 3.4673589073317297,
      "tokens_seen": 1761542144
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002624337998865141,
      "loss": 2.5679,
      "theoretical_loss": 3.4673478815539363,
      "tokens_seen": 1761607680
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002624172498581426,
      "loss": 2.6443,
      "theoretical_loss": 3.4673368563011677,
      "tokens_seen": 1761673216
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026240069982977115,
      "loss": 2.5506,
      "theoretical_loss": 3.467325831573379,
      "tokens_seen": 1761738752
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002623841498013997,
      "loss": 2.6088,
      "theoretical_loss": 3.467314807370526,
      "tokens_seen": 1761804288
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002623675997730282,
      "loss": 2.7131,
      "theoretical_loss": 3.4673037836925635,
      "tokens_seen": 1761869824
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002623510497446567,
      "loss": 2.5802,
      "theoretical_loss": 3.467292760539448,
      "tokens_seen": 1761935360
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026233449971628525,
      "loss": 2.7488,
      "theoretical_loss": 3.467281737911134,
      "tokens_seen": 1762000896
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002623179496879138,
      "loss": 2.5856,
      "theoretical_loss": 3.4672707158075777,
      "tokens_seen": 1762066432
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026230139965954227,
      "loss": 2.5674,
      "theoretical_loss": 3.467259694228734,
      "tokens_seen": 1762131968
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002622848496311708,
      "loss": 2.2975,
      "theoretical_loss": 3.467248673174559,
      "tokens_seen": 1762197504
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026226829960279935,
      "loss": 2.323,
      "theoretical_loss": 3.467237652645008,
      "tokens_seen": 1762263040
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026225174957442783,
      "loss": 2.6252,
      "theoretical_loss": 3.4672266326400365,
      "tokens_seen": 1762328576
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026223519954605637,
      "loss": 2.7256,
      "theoretical_loss": 3.4672156131595995,
      "tokens_seen": 1762394112
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026221864951768485,
      "loss": 2.6542,
      "theoretical_loss": 3.4672045942036536,
      "tokens_seen": 1762459648
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002622020994893134,
      "loss": 2.3175,
      "theoretical_loss": 3.4671935757721535,
      "tokens_seen": 1762525184
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002621855494609419,
      "loss": 2.6056,
      "theoretical_loss": 3.467182557865055,
      "tokens_seen": 1762590720
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026216899943257046,
      "loss": 2.8122,
      "theoretical_loss": 3.4671715404823136,
      "tokens_seen": 1762656256
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026215244940419895,
      "loss": 2.5926,
      "theoretical_loss": 3.4671605236238854,
      "tokens_seen": 1762721792
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002621358993758275,
      "loss": 2.6705,
      "theoretical_loss": 3.467149507289725,
      "tokens_seen": 1762787328
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000262119349347456,
      "loss": 2.6199,
      "theoretical_loss": 3.4671384914797887,
      "tokens_seen": 1762852864
    },
    {
      "epoch": 0.63,
      "objective/train/avg_token_score": 0.014797152951359749,
      "objective/train/avg_weight": 0.9825850129127502,
      "objective/train/docs_used": 997036,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.660623550415039,
      "objective/train/std_weight": 0.06329334527254105,
      "objective/train/theoretical_loss": 3.4671274761940314,
      "objective/train/tokens_used": 1783378400,
      "theoretical_loss": 3.4671274761940314,
      "tokens_seen": 1762918400
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026210279931908456,
      "loss": 2.706,
      "theoretical_loss": 3.4671274761940314,
      "tokens_seen": 1762918400
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026208624929071304,
      "loss": 2.506,
      "theoretical_loss": 3.4671164614324095,
      "tokens_seen": 1762983936
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002620696992623416,
      "loss": 2.4547,
      "theoretical_loss": 3.4671054471948777,
      "tokens_seen": 1763049472
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002620531492339701,
      "loss": 2.7675,
      "theoretical_loss": 3.467094433481392,
      "tokens_seen": 1763115008
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026203659920559866,
      "loss": 2.8892,
      "theoretical_loss": 3.4670834202919085,
      "tokens_seen": 1763180544
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026202004917722714,
      "loss": 2.3573,
      "theoretical_loss": 3.4670724076263815,
      "tokens_seen": 1763246080
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002620034991488557,
      "loss": 2.6156,
      "theoretical_loss": 3.4670613954847678,
      "tokens_seen": 1763311616
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002619869491204842,
      "loss": 2.6167,
      "theoretical_loss": 3.467050383867023,
      "tokens_seen": 1763377152
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026197039909211275,
      "loss": 2.809,
      "theoretical_loss": 3.467039372773102,
      "tokens_seen": 1763442688
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026195384906374123,
      "loss": 2.7295,
      "theoretical_loss": 3.467028362202961,
      "tokens_seen": 1763508224
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026193729903536977,
      "loss": 2.6129,
      "theoretical_loss": 3.467017352156555,
      "tokens_seen": 1763573760
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002619207490069983,
      "loss": 2.6234,
      "theoretical_loss": 3.46700634263384,
      "tokens_seen": 1763639296
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026190419897862685,
      "loss": 2.6666,
      "theoretical_loss": 3.466995333634772,
      "tokens_seen": 1763704832
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026188764895025533,
      "loss": 2.5525,
      "theoretical_loss": 3.4669843251593058,
      "tokens_seen": 1763770368
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026187109892188387,
      "loss": 2.5875,
      "theoretical_loss": 3.4669733172073975,
      "tokens_seen": 1763835904
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002618545488935124,
      "loss": 2.3867,
      "theoretical_loss": 3.466962309779003,
      "tokens_seen": 1763901440
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026183799886514094,
      "loss": 2.629,
      "theoretical_loss": 3.4669513028740777,
      "tokens_seen": 1763966976
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002618214488367694,
      "loss": 2.4484,
      "theoretical_loss": 3.466940296492577,
      "tokens_seen": 1764032512
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026180489880839796,
      "loss": 2.4584,
      "theoretical_loss": 3.466929290634457,
      "tokens_seen": 1764098048
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002617883487800265,
      "loss": 2.8055,
      "theoretical_loss": 3.466918285299673,
      "tokens_seen": 1764163584
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026177179875165504,
      "loss": 2.6418,
      "theoretical_loss": 3.466907280488181,
      "tokens_seen": 1764229120
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002617552487232835,
      "loss": 2.6262,
      "theoretical_loss": 3.4668962761999365,
      "tokens_seen": 1764294656
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000261738698694912,
      "loss": 2.6824,
      "theoretical_loss": 3.4668852724348955,
      "tokens_seen": 1764360192
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026172214866654054,
      "loss": 2.4443,
      "theoretical_loss": 3.4668742691930126,
      "tokens_seen": 1764425728
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002617055986381691,
      "loss": 2.6271,
      "theoretical_loss": 3.4668632664742454,
      "tokens_seen": 1764491264
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002616890486097976,
      "loss": 2.633,
      "theoretical_loss": 3.4668522642785478,
      "tokens_seen": 1764556800
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002616724985814261,
      "loss": 2.6337,
      "theoretical_loss": 3.466841262605876,
      "tokens_seen": 1764622336
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026165594855305464,
      "loss": 2.5789,
      "theoretical_loss": 3.466830261456187,
      "tokens_seen": 1764687872
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002616393985246832,
      "loss": 2.7364,
      "theoretical_loss": 3.4668192608294346,
      "tokens_seen": 1764753408
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002616228484963117,
      "loss": 2.8171,
      "theoretical_loss": 3.4668082607255757,
      "tokens_seen": 1764818944
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002616062984679402,
      "loss": 2.5969,
      "theoretical_loss": 3.4667972611445657,
      "tokens_seen": 1764884480
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026158974843956874,
      "loss": 2.5157,
      "theoretical_loss": 3.46678626208636,
      "tokens_seen": 1764950016
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002615731984111973,
      "loss": 2.5237,
      "theoretical_loss": 3.466775263550915,
      "tokens_seen": 1765015552
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002615566483828258,
      "loss": 2.6566,
      "theoretical_loss": 3.466764265538186,
      "tokens_seen": 1765081088
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002615400983544543,
      "loss": 2.811,
      "theoretical_loss": 3.466753268048129,
      "tokens_seen": 1765146624
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026152354832608283,
      "loss": 2.8039,
      "theoretical_loss": 3.4667422710806997,
      "tokens_seen": 1765212160
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026150699829771137,
      "loss": 2.8596,
      "theoretical_loss": 3.466731274635854,
      "tokens_seen": 1765277696
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002614904482693399,
      "loss": 2.6069,
      "theoretical_loss": 3.466720278713547,
      "tokens_seen": 1765343232
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002614738982409684,
      "loss": 2.654,
      "theoretical_loss": 3.466709283313735,
      "tokens_seen": 1765408768
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026145734821259693,
      "loss": 2.7343,
      "theoretical_loss": 3.4666982884363735,
      "tokens_seen": 1765474304
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026144079818422547,
      "loss": 2.4844,
      "theoretical_loss": 3.466687294081419,
      "tokens_seen": 1765539840
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000261424248155854,
      "loss": 2.5449,
      "theoretical_loss": 3.466676300248827,
      "tokens_seen": 1765605376
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002614076981274825,
      "loss": 2.7028,
      "theoretical_loss": 3.4666653069385527,
      "tokens_seen": 1765670912
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000261391148099111,
      "loss": 2.7284,
      "theoretical_loss": 3.466654314150552,
      "tokens_seen": 1765736448
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026137459807073956,
      "loss": 2.6074,
      "theoretical_loss": 3.4666433218847814,
      "tokens_seen": 1765801984
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002613580480423681,
      "loss": 2.5623,
      "theoretical_loss": 3.466632330141196,
      "tokens_seen": 1765867520
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002613414980139966,
      "loss": 2.5714,
      "theoretical_loss": 3.4666213389197518,
      "tokens_seen": 1765933056
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002613249479856251,
      "loss": 2.6493,
      "theoretical_loss": 3.4666103482204056,
      "tokens_seen": 1765998592
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026130839795725366,
      "loss": 2.7803,
      "theoretical_loss": 3.466599358043111,
      "tokens_seen": 1766064128
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002612918479288822,
      "loss": 2.731,
      "theoretical_loss": 3.4665883683878262,
      "tokens_seen": 1766129664
    },
    {
      "epoch": 0.63,
      "objective/train/avg_token_score": 0.020840119570493698,
      "objective/train/avg_weight": 0.9754711985588074,
      "objective/train/docs_used": 998803,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1803293228149414,
      "objective/train/std_weight": 0.11695574223995209,
      "objective/train/theoretical_loss": 3.466577379254506,
      "objective/train/tokens_used": 1786655200,
      "theoretical_loss": 3.466577379254506,
      "tokens_seen": 1766195200
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002612752979005107,
      "loss": 2.6214,
      "theoretical_loss": 3.466577379254506,
      "tokens_seen": 1766195200
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026125874787213916,
      "loss": 2.7991,
      "theoretical_loss": 3.4665663906431057,
      "tokens_seen": 1766260736
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002612421978437677,
      "loss": 2.6504,
      "theoretical_loss": 3.466555402553582,
      "tokens_seen": 1766326272
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026122564781539624,
      "loss": 2.4378,
      "theoretical_loss": 3.466544414985891,
      "tokens_seen": 1766391808
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002612090977870248,
      "loss": 2.6078,
      "theoretical_loss": 3.4665334279399875,
      "tokens_seen": 1766457344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002611925477586533,
      "loss": 2.7767,
      "theoretical_loss": 3.4665224414158278,
      "tokens_seen": 1766522880
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002611759977302818,
      "loss": 2.5881,
      "theoretical_loss": 3.466511455413368,
      "tokens_seen": 1766588416
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026115944770191033,
      "loss": 2.7725,
      "theoretical_loss": 3.466500469932564,
      "tokens_seen": 1766653952
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026114289767353887,
      "loss": 2.4847,
      "theoretical_loss": 3.466489484973371,
      "tokens_seen": 1766719488
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002611263476451674,
      "loss": 2.7204,
      "theoretical_loss": 3.4664785005357457,
      "tokens_seen": 1766785024
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002611097976167959,
      "loss": 2.4862,
      "theoretical_loss": 3.4664675166196437,
      "tokens_seen": 1766850560
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026109324758842443,
      "loss": 2.6164,
      "theoretical_loss": 3.4664565332250215,
      "tokens_seen": 1766916096
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026107669756005297,
      "loss": 2.504,
      "theoretical_loss": 3.4664455503518337,
      "tokens_seen": 1766981632
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002610601475316815,
      "loss": 2.6464,
      "theoretical_loss": 3.466434568000037,
      "tokens_seen": 1767047168
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026104359750331,
      "loss": 2.5648,
      "theoretical_loss": 3.4664235861695873,
      "tokens_seen": 1767112704
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002610270474749385,
      "loss": 2.4739,
      "theoretical_loss": 3.46641260486044,
      "tokens_seen": 1767178240
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026101049744656706,
      "loss": 2.5502,
      "theoretical_loss": 3.466401624072552,
      "tokens_seen": 1767243776
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002609939474181956,
      "loss": 2.6591,
      "theoretical_loss": 3.4663906438058785,
      "tokens_seen": 1767309312
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002609773973898241,
      "loss": 2.6663,
      "theoretical_loss": 3.4663796640603755,
      "tokens_seen": 1767374848
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002609608473614526,
      "loss": 2.8722,
      "theoretical_loss": 3.466368684835999,
      "tokens_seen": 1767440384
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026094429733308116,
      "loss": 2.7437,
      "theoretical_loss": 3.466357706132705,
      "tokens_seen": 1767505920
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002609277473047097,
      "loss": 2.5172,
      "theoretical_loss": 3.46634672795045,
      "tokens_seen": 1767571456
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002609111972763382,
      "loss": 2.7086,
      "theoretical_loss": 3.466335750289189,
      "tokens_seen": 1767636992
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002608946472479667,
      "loss": 2.4481,
      "theoretical_loss": 3.466324773148878,
      "tokens_seen": 1767702528
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026087809721959525,
      "loss": 2.8179,
      "theoretical_loss": 3.466313796529474,
      "tokens_seen": 1767768064
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002608615471912238,
      "loss": 2.5368,
      "theoretical_loss": 3.466302820430932,
      "tokens_seen": 1767833600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002608449971628523,
      "loss": 2.5593,
      "theoretical_loss": 3.466291844853208,
      "tokens_seen": 1767899136
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002608284471344808,
      "loss": 2.7434,
      "theoretical_loss": 3.4662808697962584,
      "tokens_seen": 1767964672
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026081189710610935,
      "loss": 2.5818,
      "theoretical_loss": 3.466269895260039,
      "tokens_seen": 1768030208
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002607953470777379,
      "loss": 2.7374,
      "theoretical_loss": 3.466258921244506,
      "tokens_seen": 1768095744
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026077879704936637,
      "loss": 2.6718,
      "theoretical_loss": 3.466247947749615,
      "tokens_seen": 1768161280
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026076224702099485,
      "loss": 2.5583,
      "theoretical_loss": 3.466236974775322,
      "tokens_seen": 1768226816
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002607456969926234,
      "loss": 2.4482,
      "theoretical_loss": 3.4662260023215836,
      "tokens_seen": 1768292352
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026072914696425193,
      "loss": 2.6094,
      "theoretical_loss": 3.466215030388356,
      "tokens_seen": 1768357888
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026071259693588047,
      "loss": 2.5358,
      "theoretical_loss": 3.4662040589755936,
      "tokens_seen": 1768423424
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026069604690750895,
      "loss": 2.1918,
      "theoretical_loss": 3.466193088083254,
      "tokens_seen": 1768488960
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002606794968791375,
      "loss": 2.5645,
      "theoretical_loss": 3.466182117711293,
      "tokens_seen": 1768554496
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000260662946850766,
      "loss": 2.5501,
      "theoretical_loss": 3.466171147859666,
      "tokens_seen": 1768620032
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026064639682239456,
      "loss": 2.7388,
      "theoretical_loss": 3.46616017852833,
      "tokens_seen": 1768685568
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026062984679402305,
      "loss": 2.7018,
      "theoretical_loss": 3.4661492097172397,
      "tokens_seen": 1768751104
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002606132967656516,
      "loss": 2.3216,
      "theoretical_loss": 3.466138241426352,
      "tokens_seen": 1768816640
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002605967467372801,
      "loss": 2.3736,
      "theoretical_loss": 3.466127273655623,
      "tokens_seen": 1768882176
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026058019670890866,
      "loss": 2.4435,
      "theoretical_loss": 3.466116306405009,
      "tokens_seen": 1768947712
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026056364668053714,
      "loss": 2.4787,
      "theoretical_loss": 3.466105339674465,
      "tokens_seen": 1769013248
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002605470966521657,
      "loss": 2.4207,
      "theoretical_loss": 3.466094373463948,
      "tokens_seen": 1769078784
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002605305466237942,
      "loss": 2.539,
      "theoretical_loss": 3.4660834077734144,
      "tokens_seen": 1769144320
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026051399659542275,
      "loss": 2.5601,
      "theoretical_loss": 3.466072442602819,
      "tokens_seen": 1769209856
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026049744656705124,
      "loss": 2.751,
      "theoretical_loss": 3.4660614779521186,
      "tokens_seen": 1769275392
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002604808965386798,
      "loss": 2.7412,
      "theoretical_loss": 3.4660505138212696,
      "tokens_seen": 1769340928
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002604643465103083,
      "loss": 2.5943,
      "theoretical_loss": 3.466039550210228,
      "tokens_seen": 1769406464
    },
    {
      "epoch": 0.63,
      "objective/train/avg_token_score": 0.021811235696077347,
      "objective/train/avg_weight": 0.974335253238678,
      "objective/train/docs_used": 999986,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.665508508682251,
      "objective/train/std_weight": 0.10797236114740372,
      "objective/train/theoretical_loss": 3.4660285871189496,
      "objective/train/tokens_used": 1789932000,
      "theoretical_loss": 3.4660285871189496,
      "tokens_seen": 1769472000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026044779648193685,
      "loss": 2.5308,
      "theoretical_loss": 3.4660285871189496,
      "tokens_seen": 1769472000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026043124645356533,
      "loss": 2.6637,
      "theoretical_loss": 3.4660176245473906,
      "tokens_seen": 1769537536
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026041469642519387,
      "loss": 2.5358,
      "theoretical_loss": 3.4660066624955066,
      "tokens_seen": 1769603072
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002603981463968224,
      "loss": 2.5831,
      "theoretical_loss": 3.465995700963255,
      "tokens_seen": 1769668608
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026038159636845095,
      "loss": 2.7102,
      "theoretical_loss": 3.4659847399505903,
      "tokens_seen": 1769734144
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026036504634007943,
      "loss": 2.8532,
      "theoretical_loss": 3.46597377945747,
      "tokens_seen": 1769799680
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026034849631170797,
      "loss": 2.8831,
      "theoretical_loss": 3.46596281948385,
      "tokens_seen": 1769865216
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002603319462833365,
      "loss": 1.8883,
      "theoretical_loss": 3.465951860029686,
      "tokens_seen": 1769930752
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026031539625496504,
      "loss": 2.5713,
      "theoretical_loss": 3.4659409010949345,
      "tokens_seen": 1769996288
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002602988462265935,
      "loss": 2.6623,
      "theoretical_loss": 3.4659299426795513,
      "tokens_seen": 1770061824
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000260282296198222,
      "loss": 2.6349,
      "theoretical_loss": 3.4659189847834924,
      "tokens_seen": 1770127360
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026026574616985055,
      "loss": 2.6972,
      "theoretical_loss": 3.465908027406715,
      "tokens_seen": 1770192896
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002602491961414791,
      "loss": 2.7847,
      "theoretical_loss": 3.465897070549174,
      "tokens_seen": 1770258432
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002602326461131076,
      "loss": 2.7211,
      "theoretical_loss": 3.4658861142108264,
      "tokens_seen": 1770323968
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002602160960847361,
      "loss": 2.6296,
      "theoretical_loss": 3.4658751583916283,
      "tokens_seen": 1770389504
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026019954605636464,
      "loss": 2.747,
      "theoretical_loss": 3.4658642030915354,
      "tokens_seen": 1770455040
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002601829960279932,
      "loss": 2.6187,
      "theoretical_loss": 3.4658532483105042,
      "tokens_seen": 1770520576
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002601664459996217,
      "loss": 2.6532,
      "theoretical_loss": 3.4658422940484908,
      "tokens_seen": 1770586112
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002601498959712502,
      "loss": 2.7291,
      "theoretical_loss": 3.465831340305452,
      "tokens_seen": 1770651648
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026013334594287874,
      "loss": 2.7669,
      "theoretical_loss": 3.4658203870813433,
      "tokens_seen": 1770717184
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002601167959145073,
      "loss": 2.8409,
      "theoretical_loss": 3.465809434376121,
      "tokens_seen": 1770782720
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002601002458861358,
      "loss": 2.1548,
      "theoretical_loss": 3.465798482189742,
      "tokens_seen": 1770848256
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002600836958577643,
      "loss": 2.7685,
      "theoretical_loss": 3.465787530522161,
      "tokens_seen": 1770913792
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026006714582939283,
      "loss": 2.5502,
      "theoretical_loss": 3.4657765793733355,
      "tokens_seen": 1770979328
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026005059580102137,
      "loss": 2.6636,
      "theoretical_loss": 3.4657656287432212,
      "tokens_seen": 1771044864
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002600340457726499,
      "loss": 2.5882,
      "theoretical_loss": 3.465754678631775,
      "tokens_seen": 1771110400
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002600174957442784,
      "loss": 2.7505,
      "theoretical_loss": 3.4657437290389526,
      "tokens_seen": 1771175936
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00026000094571590693,
      "loss": 2.4762,
      "theoretical_loss": 3.4657327799647106,
      "tokens_seen": 1771241472
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025998439568753547,
      "loss": 2.4928,
      "theoretical_loss": 3.4657218314090046,
      "tokens_seen": 1771307008
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000259967845659164,
      "loss": 2.6775,
      "theoretical_loss": 3.4657108833717913,
      "tokens_seen": 1771372544
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002599512956307925,
      "loss": 2.7937,
      "theoretical_loss": 3.465699935853027,
      "tokens_seen": 1771438080
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000259934745602421,
      "loss": 2.6689,
      "theoretical_loss": 3.465688988852668,
      "tokens_seen": 1771503616
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025991819557404956,
      "loss": 2.598,
      "theoretical_loss": 3.46567804237067,
      "tokens_seen": 1771569152
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002599016455456781,
      "loss": 2.6742,
      "theoretical_loss": 3.4656670964069907,
      "tokens_seen": 1771634688
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002598850955173066,
      "loss": 2.3957,
      "theoretical_loss": 3.4656561509615846,
      "tokens_seen": 1771700224
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002598685454889351,
      "loss": 2.5112,
      "theoretical_loss": 3.4656452060344085,
      "tokens_seen": 1771765760
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025985199546056366,
      "loss": 2.797,
      "theoretical_loss": 3.4656342616254197,
      "tokens_seen": 1771831296
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002598354454321922,
      "loss": 2.7906,
      "theoretical_loss": 3.4656233177345737,
      "tokens_seen": 1771896832
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002598188954038207,
      "loss": 2.8343,
      "theoretical_loss": 3.465612374361827,
      "tokens_seen": 1771962368
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002598023453754492,
      "loss": 2.6833,
      "theoretical_loss": 3.4656014315071357,
      "tokens_seen": 1772027904
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002597857953470777,
      "loss": 2.753,
      "theoretical_loss": 3.4655904891704563,
      "tokens_seen": 1772093440
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025976924531870624,
      "loss": 1.9475,
      "theoretical_loss": 3.4655795473517452,
      "tokens_seen": 1772158976
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002597526952903348,
      "loss": 2.6079,
      "theoretical_loss": 3.465568606050958,
      "tokens_seen": 1772224512
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002597361452619633,
      "loss": 2.2572,
      "theoretical_loss": 3.465557665268052,
      "tokens_seen": 1772290048
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002597195952335918,
      "loss": 2.6314,
      "theoretical_loss": 3.4655467250029837,
      "tokens_seen": 1772355584
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025970304520522034,
      "loss": 2.41,
      "theoretical_loss": 3.4655357852557085,
      "tokens_seen": 1772421120
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025968649517684887,
      "loss": 2.6251,
      "theoretical_loss": 3.4655248460261827,
      "tokens_seen": 1772486656
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002596699451484774,
      "loss": 2.5997,
      "theoretical_loss": 3.4655139073143637,
      "tokens_seen": 1772552192
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002596533951201059,
      "loss": 2.5737,
      "theoretical_loss": 3.465502969120207,
      "tokens_seen": 1772617728
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025963684509173443,
      "loss": 2.6534,
      "theoretical_loss": 3.4654920314436697,
      "tokens_seen": 1772683264
    },
    {
      "epoch": 0.63,
      "objective/train/avg_token_score": 0.028194651007652283,
      "objective/train/avg_weight": 0.9668104648590088,
      "objective/train/docs_used": 1001522,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.273062705993652,
      "objective/train/std_weight": 0.11671003699302673,
      "objective/train/theoretical_loss": 3.465481094284707,
      "objective/train/tokens_used": 1793208800,
      "theoretical_loss": 3.465481094284707,
      "tokens_seen": 1772748800
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025962029506336297,
      "loss": 2.4171,
      "theoretical_loss": 3.465481094284707,
      "tokens_seen": 1772748800
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002596037450349915,
      "loss": 2.5312,
      "theoretical_loss": 3.465470157643277,
      "tokens_seen": 1772814336
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025958719500662,
      "loss": 2.8181,
      "theoretical_loss": 3.4654592215193345,
      "tokens_seen": 1772879872
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002595706449782485,
      "loss": 2.6956,
      "theoretical_loss": 3.465448285912836,
      "tokens_seen": 1772945408
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025955409494987706,
      "loss": 2.5114,
      "theoretical_loss": 3.465437350823739,
      "tokens_seen": 1773010944
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002595375449215056,
      "loss": 2.587,
      "theoretical_loss": 3.465426416251999,
      "tokens_seen": 1773076480
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002595209948931341,
      "loss": 2.5723,
      "theoretical_loss": 3.4654154821975727,
      "tokens_seen": 1773142016
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002595044448647626,
      "loss": 2.5177,
      "theoretical_loss": 3.465404548660416,
      "tokens_seen": 1773207552
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025948789483639116,
      "loss": 2.6377,
      "theoretical_loss": 3.4653936156404868,
      "tokens_seen": 1773273088
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002594713448080197,
      "loss": 2.6619,
      "theoretical_loss": 3.4653826831377397,
      "tokens_seen": 1773338624
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002594547947796482,
      "loss": 2.5099,
      "theoretical_loss": 3.465371751152132,
      "tokens_seen": 1773404160
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002594382447512767,
      "loss": 2.5105,
      "theoretical_loss": 3.46536081968362,
      "tokens_seen": 1773469696
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025942169472290526,
      "loss": 2.7718,
      "theoretical_loss": 3.4653498887321605,
      "tokens_seen": 1773535232
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002594051446945338,
      "loss": 2.4686,
      "theoretical_loss": 3.465338958297709,
      "tokens_seen": 1773600768
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002593885946661623,
      "loss": 2.7019,
      "theoretical_loss": 3.465328028380223,
      "tokens_seen": 1773666304
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002593720446377908,
      "loss": 2.3237,
      "theoretical_loss": 3.465317098979659,
      "tokens_seen": 1773731840
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025935549460941935,
      "loss": 2.7691,
      "theoretical_loss": 3.465306170095972,
      "tokens_seen": 1773797376
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002593389445810479,
      "loss": 2.7408,
      "theoretical_loss": 3.46529524172912,
      "tokens_seen": 1773862912
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002593223945526764,
      "loss": 2.7575,
      "theoretical_loss": 3.4652843138790583,
      "tokens_seen": 1773928448
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002593058445243049,
      "loss": 2.549,
      "theoretical_loss": 3.4652733865457446,
      "tokens_seen": 1773993984
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002592892944959334,
      "loss": 2.6879,
      "theoretical_loss": 3.465262459729135,
      "tokens_seen": 1774059520
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025927274446756193,
      "loss": 2.8941,
      "theoretical_loss": 3.4652515334291847,
      "tokens_seen": 1774125056
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025925619443919047,
      "loss": 2.6867,
      "theoretical_loss": 3.4652406076458515,
      "tokens_seen": 1774190592
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025923964441081895,
      "loss": 2.7431,
      "theoretical_loss": 3.465229682379092,
      "tokens_seen": 1774256128
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002592230943824475,
      "loss": 2.7156,
      "theoretical_loss": 3.465218757628862,
      "tokens_seen": 1774321664
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025920654435407603,
      "loss": 2.4085,
      "theoretical_loss": 3.4652078333951186,
      "tokens_seen": 1774387200
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025918999432570457,
      "loss": 2.7158,
      "theoretical_loss": 3.4651969096778177,
      "tokens_seen": 1774452736
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025917344429733305,
      "loss": 2.5217,
      "theoretical_loss": 3.465185986476916,
      "tokens_seen": 1774518272
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002591568942689616,
      "loss": 2.7227,
      "theoretical_loss": 3.465175063792371,
      "tokens_seen": 1774583808
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002591403442405901,
      "loss": 2.6977,
      "theoretical_loss": 3.465164141624137,
      "tokens_seen": 1774649344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025912379421221866,
      "loss": 2.6639,
      "theoretical_loss": 3.465153219972173,
      "tokens_seen": 1774714880
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025910724418384714,
      "loss": 2.2893,
      "theoretical_loss": 3.4651422988364335,
      "tokens_seen": 1774780416
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002590906941554757,
      "loss": 2.767,
      "theoretical_loss": 3.4651313782168764,
      "tokens_seen": 1774845952
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002590741441271042,
      "loss": 2.8715,
      "theoretical_loss": 3.465120458113458,
      "tokens_seen": 1774911488
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025905759409873276,
      "loss": 2.6007,
      "theoretical_loss": 3.4651095385261343,
      "tokens_seen": 1774977024
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025904104407036124,
      "loss": 2.8377,
      "theoretical_loss": 3.4650986194548623,
      "tokens_seen": 1775042560
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002590244940419898,
      "loss": 2.5131,
      "theoretical_loss": 3.4650877008995984,
      "tokens_seen": 1775108096
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002590079440136183,
      "loss": 2.5134,
      "theoretical_loss": 3.4650767828602995,
      "tokens_seen": 1775173632
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025899139398524685,
      "loss": 2.6394,
      "theoretical_loss": 3.4650658653369213,
      "tokens_seen": 1775239168
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025897484395687534,
      "loss": 2.5457,
      "theoretical_loss": 3.4650549483294215,
      "tokens_seen": 1775304704
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002589582939285039,
      "loss": 2.8778,
      "theoretical_loss": 3.4650440318377562,
      "tokens_seen": 1775370240
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002589417439001324,
      "loss": 2.8547,
      "theoretical_loss": 3.465033115861882,
      "tokens_seen": 1775435776
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025892519387176095,
      "loss": 2.3558,
      "theoretical_loss": 3.465022200401755,
      "tokens_seen": 1775501312
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025890864384338943,
      "loss": 2.5493,
      "theoretical_loss": 3.465011285457332,
      "tokens_seen": 1775566848
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025889209381501797,
      "loss": 2.9623,
      "theoretical_loss": 3.4650003710285704,
      "tokens_seen": 1775632384
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002588755437866465,
      "loss": 1.9149,
      "theoretical_loss": 3.4649894571154265,
      "tokens_seen": 1775697920
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025885899375827505,
      "loss": 2.5703,
      "theoretical_loss": 3.464978543717856,
      "tokens_seen": 1775763456
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025884244372990353,
      "loss": 2.6818,
      "theoretical_loss": 3.464967630835816,
      "tokens_seen": 1775828992
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025882589370153207,
      "loss": 2.5603,
      "theoretical_loss": 3.464956718469264,
      "tokens_seen": 1775894528
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025880934367316055,
      "loss": 2.455,
      "theoretical_loss": 3.4649458066181555,
      "tokens_seen": 1775960064
    },
    {
      "epoch": 0.63,
      "objective/train/avg_token_score": 0.023502476513385773,
      "objective/train/avg_weight": 0.9723361134529114,
      "objective/train/docs_used": 1003498,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.756737232208252,
      "objective/train/std_weight": 0.09095840156078339,
      "objective/train/theoretical_loss": 3.4649348952824477,
      "objective/train/tokens_used": 1796485600,
      "theoretical_loss": 3.4649348952824477,
      "tokens_seen": 1776025600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002587927936447891,
      "loss": 2.2087,
      "theoretical_loss": 3.4649348952824477,
      "tokens_seen": 1776025600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002587762436164176,
      "loss": 2.4009,
      "theoretical_loss": 3.4649239844620974,
      "tokens_seen": 1776091136
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002587596935880461,
      "loss": 2.5223,
      "theoretical_loss": 3.46491307415706,
      "tokens_seen": 1776156672
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025874314355967465,
      "loss": 2.3269,
      "theoretical_loss": 3.464902164367294,
      "tokens_seen": 1776222208
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002587265935313032,
      "loss": 2.7487,
      "theoretical_loss": 3.4648912550927546,
      "tokens_seen": 1776287744
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002587100435029317,
      "loss": 1.5232,
      "theoretical_loss": 3.4648803463333993,
      "tokens_seen": 1776353280
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002586934934745602,
      "loss": 2.689,
      "theoretical_loss": 3.4648694380891847,
      "tokens_seen": 1776418816
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025867694344618874,
      "loss": 2.4841,
      "theoretical_loss": 3.4648585303600665,
      "tokens_seen": 1776484352
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002586603934178173,
      "loss": 2.5634,
      "theoretical_loss": 3.464847623146003,
      "tokens_seen": 1776549888
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002586438433894458,
      "loss": 2.6354,
      "theoretical_loss": 3.464836716446949,
      "tokens_seen": 1776615424
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002586272933610743,
      "loss": 2.5383,
      "theoretical_loss": 3.464825810262863,
      "tokens_seen": 1776680960
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025861074333270284,
      "loss": 2.5505,
      "theoretical_loss": 3.4648149045937005,
      "tokens_seen": 1776746496
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002585941933043314,
      "loss": 2.7926,
      "theoretical_loss": 3.464803999439418,
      "tokens_seen": 1776812032
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002585776432759599,
      "loss": 2.5837,
      "theoretical_loss": 3.4647930947999743,
      "tokens_seen": 1776877568
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002585610932475884,
      "loss": 2.526,
      "theoretical_loss": 3.464782190675323,
      "tokens_seen": 1776943104
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025854454321921693,
      "loss": 2.929,
      "theoretical_loss": 3.4647712870654237,
      "tokens_seen": 1777008640
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025852799319084547,
      "loss": 2.7347,
      "theoretical_loss": 3.464760383970231,
      "tokens_seen": 1777074176
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000258511443162474,
      "loss": 2.4666,
      "theoretical_loss": 3.4647494813897026,
      "tokens_seen": 1777139712
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025849489313410255,
      "loss": 2.5352,
      "theoretical_loss": 3.464738579323795,
      "tokens_seen": 1777205248
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025849489313410255,
      "loss": 18.5715,
      "theoretical_loss": 3.4647276777724647,
      "tokens_seen": 1777270784
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025847834310573103,
      "loss": 2.7113,
      "theoretical_loss": 3.4647167767356692,
      "tokens_seen": 1777336320
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025846179307735957,
      "loss": 2.5309,
      "theoretical_loss": 3.4647058762133645,
      "tokens_seen": 1777401856
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002584452430489881,
      "loss": 2.6503,
      "theoretical_loss": 3.4646949762055077,
      "tokens_seen": 1777467392
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025842869302061664,
      "loss": 2.685,
      "theoretical_loss": 3.4646840767120555,
      "tokens_seen": 1777532928
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002584121429922451,
      "loss": 2.5829,
      "theoretical_loss": 3.4646731777329647,
      "tokens_seen": 1777598464
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025839559296387366,
      "loss": 2.5059,
      "theoretical_loss": 3.4646622792681923,
      "tokens_seen": 1777664000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002583790429355022,
      "loss": 2.457,
      "theoretical_loss": 3.4646513813176942,
      "tokens_seen": 1777729536
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025836249290713074,
      "loss": 2.705,
      "theoretical_loss": 3.464640483881428,
      "tokens_seen": 1777795072
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002583459428787592,
      "loss": 2.4346,
      "theoretical_loss": 3.46462958695935,
      "tokens_seen": 1777860608
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00025832939285038776,
      "loss": 2.5594,
      "theoretical_loss": 3.4646186905514176,
      "tokens_seen": 1777926144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025831284282201624,
      "loss": 2.6831,
      "theoretical_loss": 3.464607794657587,
      "tokens_seen": 1777991680
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002582962927936448,
      "loss": 2.6571,
      "theoretical_loss": 3.464596899277815,
      "tokens_seen": 1778057216
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002582797427652733,
      "loss": 2.4614,
      "theoretical_loss": 3.464586004412059,
      "tokens_seen": 1778122752
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002582631927369018,
      "loss": 2.7331,
      "theoretical_loss": 3.464575110060275,
      "tokens_seen": 1778188288
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025824664270853034,
      "loss": 2.0005,
      "theoretical_loss": 3.4645642162224206,
      "tokens_seen": 1778253824
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002582300926801589,
      "loss": 2.3813,
      "theoretical_loss": 3.4645533228984515,
      "tokens_seen": 1778319360
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002582135426517874,
      "loss": 2.689,
      "theoretical_loss": 3.464542430088326,
      "tokens_seen": 1778384896
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002581969926234159,
      "loss": 2.2855,
      "theoretical_loss": 3.4645315377920003,
      "tokens_seen": 1778450432
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025818044259504443,
      "loss": 2.6938,
      "theoretical_loss": 3.46452064600943,
      "tokens_seen": 1778515968
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025816389256667297,
      "loss": 2.8873,
      "theoretical_loss": 3.4645097547405745,
      "tokens_seen": 1778581504
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002581473425383015,
      "loss": 2.65,
      "theoretical_loss": 3.4644988639853884,
      "tokens_seen": 1778647040
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025813079250993,
      "loss": 2.5756,
      "theoretical_loss": 3.4644879737438297,
      "tokens_seen": 1778712576
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025811424248155853,
      "loss": 2.5323,
      "theoretical_loss": 3.4644770840158543,
      "tokens_seen": 1778778112
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025809769245318707,
      "loss": 2.4546,
      "theoretical_loss": 3.46446619480142,
      "tokens_seen": 1778843648
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002580811424248156,
      "loss": 2.559,
      "theoretical_loss": 3.4644553061004832,
      "tokens_seen": 1778909184
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002580645923964441,
      "loss": 2.8668,
      "theoretical_loss": 3.4644444179130014,
      "tokens_seen": 1778974720
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002580480423680726,
      "loss": 2.5796,
      "theoretical_loss": 3.46443353023893,
      "tokens_seen": 1779040256
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025803149233970116,
      "loss": 2.6543,
      "theoretical_loss": 3.464422643078228,
      "tokens_seen": 1779105792
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002580149423113297,
      "loss": 2.5046,
      "theoretical_loss": 3.4644117564308505,
      "tokens_seen": 1779171328
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002579983922829582,
      "loss": 2.583,
      "theoretical_loss": 3.464400870296755,
      "tokens_seen": 1779236864
    },
    {
      "epoch": 0.64,
      "objective/train/avg_token_score": 0.008760835975408554,
      "objective/train/avg_weight": 0.9896875023841858,
      "objective/train/docs_used": 1005457,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5327885150909424,
      "objective/train/std_weight": 0.054491907358169556,
      "objective/train/theoretical_loss": 3.4643899846758983,
      "objective/train/tokens_used": 1799762400,
      "theoretical_loss": 3.4643899846758983,
      "tokens_seen": 1779302400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002579818422545867,
      "loss": 2.6581,
      "theoretical_loss": 3.4643899846758983,
      "tokens_seen": 1779302400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025796529222621526,
      "loss": 2.4783,
      "theoretical_loss": 3.4643790995682373,
      "tokens_seen": 1779367936
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002579487421978438,
      "loss": 2.4881,
      "theoretical_loss": 3.46436821497373,
      "tokens_seen": 1779433472
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002579321921694723,
      "loss": 2.627,
      "theoretical_loss": 3.464357330892331,
      "tokens_seen": 1779499008
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002579156421411008,
      "loss": 2.6046,
      "theoretical_loss": 3.464346447323999,
      "tokens_seen": 1779564544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025789909211272936,
      "loss": 2.4362,
      "theoretical_loss": 3.4643355642686906,
      "tokens_seen": 1779630080
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002578825420843579,
      "loss": 2.6153,
      "theoretical_loss": 3.464324681726363,
      "tokens_seen": 1779695616
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002578659920559864,
      "loss": 2.7145,
      "theoretical_loss": 3.4643137996969724,
      "tokens_seen": 1779761152
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002578494420276149,
      "loss": 2.5035,
      "theoretical_loss": 3.4643029181804756,
      "tokens_seen": 1779826688
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025783289199924345,
      "loss": 2.7155,
      "theoretical_loss": 3.464292037176831,
      "tokens_seen": 1779892224
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025781634197087193,
      "loss": 2.2604,
      "theoretical_loss": 3.4642811566859937,
      "tokens_seen": 1779957760
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025779979194250047,
      "loss": 2.3924,
      "theoretical_loss": 3.4642702767079214,
      "tokens_seen": 1780023296
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025778324191412896,
      "loss": 2.4942,
      "theoretical_loss": 3.4642593972425715,
      "tokens_seen": 1780088832
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002577666918857575,
      "loss": 2.4262,
      "theoretical_loss": 3.4642485182899008,
      "tokens_seen": 1780154368
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025775014185738603,
      "loss": 2.6503,
      "theoretical_loss": 3.464237639849866,
      "tokens_seen": 1780219904
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025773359182901457,
      "loss": 2.6775,
      "theoretical_loss": 3.464226761922424,
      "tokens_seen": 1780285440
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025771704180064305,
      "loss": 2.7026,
      "theoretical_loss": 3.464215884507532,
      "tokens_seen": 1780350976
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002577004917722716,
      "loss": 2.6009,
      "theoretical_loss": 3.464205007605147,
      "tokens_seen": 1780416512
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002576839417439001,
      "loss": 2.3181,
      "theoretical_loss": 3.464194131215226,
      "tokens_seen": 1780482048
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025766739171552866,
      "loss": 2.6045,
      "theoretical_loss": 3.4641832553377254,
      "tokens_seen": 1780547584
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025765084168715715,
      "loss": 2.4818,
      "theoretical_loss": 3.464172379972603,
      "tokens_seen": 1780613120
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002576342916587857,
      "loss": 2.5068,
      "theoretical_loss": 3.4641615051198156,
      "tokens_seen": 1780678656
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002576177416304142,
      "loss": 2.5193,
      "theoretical_loss": 3.4641506307793204,
      "tokens_seen": 1780744192
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025760119160204276,
      "loss": 2.8313,
      "theoretical_loss": 3.4641397569510737,
      "tokens_seen": 1780809728
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025758464157367124,
      "loss": 2.4344,
      "theoretical_loss": 3.464128883635033,
      "tokens_seen": 1780875264
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002575680915452998,
      "loss": 2.5542,
      "theoretical_loss": 3.4641180108311556,
      "tokens_seen": 1780940800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002575515415169283,
      "loss": 2.6352,
      "theoretical_loss": 3.4641071385393976,
      "tokens_seen": 1781006336
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025753499148855686,
      "loss": 2.2991,
      "theoretical_loss": 3.464096266759717,
      "tokens_seen": 1781071872
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025751844146018534,
      "loss": 2.7257,
      "theoretical_loss": 3.4640853954920705,
      "tokens_seen": 1781137408
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002575018914318139,
      "loss": 2.5441,
      "theoretical_loss": 3.4640745247364153,
      "tokens_seen": 1781202944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002574853414034424,
      "loss": 2.3502,
      "theoretical_loss": 3.4640636544927075,
      "tokens_seen": 1781268480
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025746879137507095,
      "loss": 2.7901,
      "theoretical_loss": 3.4640527847609057,
      "tokens_seen": 1781334016
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025745224134669944,
      "loss": 2.6468,
      "theoretical_loss": 3.4640419155409665,
      "tokens_seen": 1781399552
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025743569131832797,
      "loss": 2.5434,
      "theoretical_loss": 3.4640310468328455,
      "tokens_seen": 1781465088
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002574191412899565,
      "loss": 2.4941,
      "theoretical_loss": 3.4640201786365017,
      "tokens_seen": 1781530624
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025740259126158505,
      "loss": 2.6553,
      "theoretical_loss": 3.4640093109518917,
      "tokens_seen": 1781596160
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025738604123321353,
      "loss": 2.6994,
      "theoretical_loss": 3.463998443778972,
      "tokens_seen": 1781661696
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025736949120484207,
      "loss": 2.7066,
      "theoretical_loss": 3.4639875771177,
      "tokens_seen": 1781727232
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002573529411764706,
      "loss": 2.6106,
      "theoretical_loss": 3.4639767109680326,
      "tokens_seen": 1781792768
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002573363911480991,
      "loss": 2.7829,
      "theoretical_loss": 3.463965845329927,
      "tokens_seen": 1781858304
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025731984111972763,
      "loss": 2.2795,
      "theoretical_loss": 3.46395498020334,
      "tokens_seen": 1781923840
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002573032910913561,
      "loss": 2.4918,
      "theoretical_loss": 3.4639441155882302,
      "tokens_seen": 1781989376
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025728674106298465,
      "loss": 2.3948,
      "theoretical_loss": 3.463933251484553,
      "tokens_seen": 1782054912
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002572701910346132,
      "loss": 2.5155,
      "theoretical_loss": 3.463922387892266,
      "tokens_seen": 1782120448
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002572536410062417,
      "loss": 2.5602,
      "theoretical_loss": 3.4639115248113264,
      "tokens_seen": 1782185984
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025723709097787026,
      "loss": 2.8329,
      "theoretical_loss": 3.4639006622416915,
      "tokens_seen": 1782251520
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025722054094949874,
      "loss": 2.6085,
      "theoretical_loss": 3.4638898001833183,
      "tokens_seen": 1782317056
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002572039909211273,
      "loss": 2.6263,
      "theoretical_loss": 3.463878938636164,
      "tokens_seen": 1782382592
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002571874408927558,
      "loss": 2.4315,
      "theoretical_loss": 3.4638680776001856,
      "tokens_seen": 1782448128
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025717089086438436,
      "loss": 2.7307,
      "theoretical_loss": 3.4638572170753408,
      "tokens_seen": 1782513664
    },
    {
      "epoch": 0.64,
      "objective/train/avg_token_score": 0.02747834101319313,
      "objective/train/avg_weight": 0.9677461385726929,
      "objective/train/docs_used": 1007648,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.5650196075439453,
      "objective/train/std_weight": 0.12933802604675293,
      "objective/train/theoretical_loss": 3.4638463570615854,
      "objective/train/tokens_used": 1803039200,
      "theoretical_loss": 3.4638463570615854,
      "tokens_seen": 1782579200
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025715434083601284,
      "loss": 2.2997,
      "theoretical_loss": 3.4638463570615854,
      "tokens_seen": 1782579200
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002571377908076414,
      "loss": 2.5992,
      "theoretical_loss": 3.463835497558878,
      "tokens_seen": 1782644736
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002571212407792699,
      "loss": 2.3946,
      "theoretical_loss": 3.463824638567175,
      "tokens_seen": 1782710272
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025710469075089845,
      "loss": 2.8026,
      "theoretical_loss": 3.4638137800864337,
      "tokens_seen": 1782775808
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025708814072252694,
      "loss": 2.6807,
      "theoretical_loss": 3.463802922116612,
      "tokens_seen": 1782841344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002570715906941555,
      "loss": 2.4802,
      "theoretical_loss": 3.4637920646576656,
      "tokens_seen": 1782906880
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000257055040665784,
      "loss": 2.6798,
      "theoretical_loss": 3.463781207709553,
      "tokens_seen": 1782972416
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025703849063741255,
      "loss": 2.6376,
      "theoretical_loss": 3.463770351272231,
      "tokens_seen": 1783037952
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025702194060904103,
      "loss": 2.3438,
      "theoretical_loss": 3.4637594953456565,
      "tokens_seen": 1783103488
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025700539058066957,
      "loss": 2.6854,
      "theoretical_loss": 3.463748639929787,
      "tokens_seen": 1783169024
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002569888405522981,
      "loss": 2.6326,
      "theoretical_loss": 3.4637377850245796,
      "tokens_seen": 1783234560
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025697229052392664,
      "loss": 2.4596,
      "theoretical_loss": 3.4637269306299916,
      "tokens_seen": 1783300096
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025695574049555513,
      "loss": 2.3152,
      "theoretical_loss": 3.46371607674598,
      "tokens_seen": 1783365632
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025693919046718367,
      "loss": 2.5496,
      "theoretical_loss": 3.463705223372502,
      "tokens_seen": 1783431168
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002569226404388122,
      "loss": 2.1342,
      "theoretical_loss": 3.463694370509515,
      "tokens_seen": 1783496704
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025690609041044074,
      "loss": 2.8546,
      "theoretical_loss": 3.4636835181569765,
      "tokens_seen": 1783562240
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002568895403820692,
      "loss": 2.5906,
      "theoretical_loss": 3.4636726663148436,
      "tokens_seen": 1783627776
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025687299035369776,
      "loss": 2.7115,
      "theoretical_loss": 3.463661814983073,
      "tokens_seen": 1783693312
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002568564403253263,
      "loss": 2.7096,
      "theoretical_loss": 3.463650964161623,
      "tokens_seen": 1783758848
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002568398902969548,
      "loss": 2.8269,
      "theoretical_loss": 3.4636401138504493,
      "tokens_seen": 1783824384
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002568233402685833,
      "loss": 2.8171,
      "theoretical_loss": 3.4636292640495103,
      "tokens_seen": 1783889920
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002568067902402118,
      "loss": 2.4005,
      "theoretical_loss": 3.4636184147587636,
      "tokens_seen": 1783955456
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025679024021184034,
      "loss": 2.433,
      "theoretical_loss": 3.4636075659781653,
      "tokens_seen": 1784020992
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002567736901834689,
      "loss": 2.7835,
      "theoretical_loss": 3.4635967177076736,
      "tokens_seen": 1784086528
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002567571401550974,
      "loss": 2.4408,
      "theoretical_loss": 3.463585869947245,
      "tokens_seen": 1784152064
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002567405901267259,
      "loss": 2.4766,
      "theoretical_loss": 3.463575022696838,
      "tokens_seen": 1784217600
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025672404009835444,
      "loss": 2.59,
      "theoretical_loss": 3.4635641759564084,
      "tokens_seen": 1784283136
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000256707490069983,
      "loss": 2.4368,
      "theoretical_loss": 3.4635533297259142,
      "tokens_seen": 1784348672
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002566909400416115,
      "loss": 2.7271,
      "theoretical_loss": 3.463542484005313,
      "tokens_seen": 1784414208
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025667439001324,
      "loss": 2.4614,
      "theoretical_loss": 3.463531638794562,
      "tokens_seen": 1784479744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025665783998486853,
      "loss": 2.3829,
      "theoretical_loss": 3.463520794093618,
      "tokens_seen": 1784545280
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025664128995649707,
      "loss": 2.4771,
      "theoretical_loss": 3.4635099499024387,
      "tokens_seen": 1784610816
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002566247399281256,
      "loss": 2.6461,
      "theoretical_loss": 3.463499106220981,
      "tokens_seen": 1784676352
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002566081898997541,
      "loss": 2.4379,
      "theoretical_loss": 3.463488263049203,
      "tokens_seen": 1784741888
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025659163987138263,
      "loss": 2.6605,
      "theoretical_loss": 3.4634774203870617,
      "tokens_seen": 1784807424
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025657508984301117,
      "loss": 2.6406,
      "theoretical_loss": 3.463466578234514,
      "tokens_seen": 1784872960
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002565585398146397,
      "loss": 2.3811,
      "theoretical_loss": 3.463455736591518,
      "tokens_seen": 1784938496
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002565419897862682,
      "loss": 2.7082,
      "theoretical_loss": 3.4634448954580304,
      "tokens_seen": 1785004032
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002565254397578967,
      "loss": 2.7859,
      "theoretical_loss": 3.463434054834009,
      "tokens_seen": 1785069568
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025650888972952526,
      "loss": 2.626,
      "theoretical_loss": 3.463423214719411,
      "tokens_seen": 1785135104
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002564923397011538,
      "loss": 2.7592,
      "theoretical_loss": 3.463412375114193,
      "tokens_seen": 1785200640
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002564757896727823,
      "loss": 2.8687,
      "theoretical_loss": 3.4634015360183135,
      "tokens_seen": 1785266176
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002564592396444108,
      "loss": 2.7556,
      "theoretical_loss": 3.4633906974317292,
      "tokens_seen": 1785331712
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025644268961603936,
      "loss": 2.7617,
      "theoretical_loss": 3.463379859354398,
      "tokens_seen": 1785397248
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002564261395876679,
      "loss": 2.5734,
      "theoretical_loss": 3.463369021786277,
      "tokens_seen": 1785462784
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002564095895592964,
      "loss": 2.6557,
      "theoretical_loss": 3.463358184727323,
      "tokens_seen": 1785528320
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002563930395309249,
      "loss": 2.5637,
      "theoretical_loss": 3.4633473481774946,
      "tokens_seen": 1785593856
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025637648950255345,
      "loss": 2.4817,
      "theoretical_loss": 3.4633365121367485,
      "tokens_seen": 1785659392
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000256359939474182,
      "loss": 2.6648,
      "theoretical_loss": 3.463325676605042,
      "tokens_seen": 1785724928
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002563433894458105,
      "loss": 2.5495,
      "theoretical_loss": 3.4633148415823323,
      "tokens_seen": 1785790464
    },
    {
      "epoch": 0.64,
      "objective/train/avg_token_score": 0.02141863852739334,
      "objective/train/avg_weight": 0.9748120903968811,
      "objective/train/docs_used": 1009512,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6133341789245605,
      "objective/train/std_weight": 0.08480720967054367,
      "objective/train/theoretical_loss": 3.4633040070685777,
      "objective/train/tokens_used": 1806316000,
      "theoretical_loss": 3.4633040070685777,
      "tokens_seen": 1785856000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025632683941743896,
      "loss": 2.6554,
      "theoretical_loss": 3.4633040070685777,
      "tokens_seen": 1785856000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002563102893890675,
      "loss": 2.6323,
      "theoretical_loss": 3.463293173063735,
      "tokens_seen": 1785921536
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025629373936069603,
      "loss": 2.852,
      "theoretical_loss": 3.463282339567762,
      "tokens_seen": 1785987072
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025627718933232457,
      "loss": 2.6064,
      "theoretical_loss": 3.4632715065806154,
      "tokens_seen": 1786052608
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025626063930395305,
      "loss": 2.4303,
      "theoretical_loss": 3.463260674102253,
      "tokens_seen": 1786118144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002562440892755816,
      "loss": 2.5303,
      "theoretical_loss": 3.4632498421326328,
      "tokens_seen": 1786183680
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025622753924721013,
      "loss": 2.6356,
      "theoretical_loss": 3.4632390106717112,
      "tokens_seen": 1786249216
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025621098921883867,
      "loss": 2.6588,
      "theoretical_loss": 3.4632281797194464,
      "tokens_seen": 1786314752
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025619443919046715,
      "loss": 2.4961,
      "theoretical_loss": 3.463217349275796,
      "tokens_seen": 1786380288
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002561778891620957,
      "loss": 2.576,
      "theoretical_loss": 3.4632065193407167,
      "tokens_seen": 1786445824
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002561613391337242,
      "loss": 2.4235,
      "theoretical_loss": 3.4631956899141665,
      "tokens_seen": 1786511360
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025614478910535276,
      "loss": 2.6918,
      "theoretical_loss": 3.463184860996103,
      "tokens_seen": 1786576896
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025612823907698125,
      "loss": 2.665,
      "theoretical_loss": 3.4631740325864833,
      "tokens_seen": 1786642432
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002561116890486098,
      "loss": 2.6566,
      "theoretical_loss": 3.463163204685265,
      "tokens_seen": 1786707968
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002560951390202383,
      "loss": 2.6719,
      "theoretical_loss": 3.4631523772924053,
      "tokens_seen": 1786773504
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025607858899186686,
      "loss": 2.6238,
      "theoretical_loss": 3.463141550407862,
      "tokens_seen": 1786839040
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025606203896349534,
      "loss": 2.7068,
      "theoretical_loss": 3.4631307240315934,
      "tokens_seen": 1786904576
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002560454889351239,
      "loss": 2.7748,
      "theoretical_loss": 3.463119898163555,
      "tokens_seen": 1786970112
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002560289389067524,
      "loss": 2.6646,
      "theoretical_loss": 3.463109072803706,
      "tokens_seen": 1787035648
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025601238887838095,
      "loss": 2.3202,
      "theoretical_loss": 3.4630982479520034,
      "tokens_seen": 1787101184
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025599583885000944,
      "loss": 2.6271,
      "theoretical_loss": 3.4630874236084046,
      "tokens_seen": 1787166720
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000255979288821638,
      "loss": 2.609,
      "theoretical_loss": 3.4630765997728674,
      "tokens_seen": 1787232256
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002559627387932665,
      "loss": 2.699,
      "theoretical_loss": 3.463065776445349,
      "tokens_seen": 1787297792
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025594618876489505,
      "loss": 2.6015,
      "theoretical_loss": 3.463054953625807,
      "tokens_seen": 1787363328
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025592963873652353,
      "loss": 2.6124,
      "theoretical_loss": 3.4630441313141986,
      "tokens_seen": 1787428864
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025591308870815207,
      "loss": 2.3414,
      "theoretical_loss": 3.4630333095104824,
      "tokens_seen": 1787494400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002558965386797806,
      "loss": 2.7415,
      "theoretical_loss": 3.463022488214615,
      "tokens_seen": 1787559936
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025587998865140915,
      "loss": 2.596,
      "theoretical_loss": 3.4630116674265548,
      "tokens_seen": 1787625472
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025586343862303763,
      "loss": 2.1683,
      "theoretical_loss": 3.463000847146258,
      "tokens_seen": 1787691008
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025584688859466617,
      "loss": 2.4773,
      "theoretical_loss": 3.4629900273736833,
      "tokens_seen": 1787756544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025583033856629465,
      "loss": 2.4359,
      "theoretical_loss": 3.462979208108788,
      "tokens_seen": 1787822080
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002558137885379232,
      "loss": 2.6854,
      "theoretical_loss": 3.462968389351529,
      "tokens_seen": 1787887616
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002557972385095517,
      "loss": 2.343,
      "theoretical_loss": 3.4629575711018648,
      "tokens_seen": 1787953152
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025578068848118026,
      "loss": 2.4779,
      "theoretical_loss": 3.462946753359753,
      "tokens_seen": 1788018688
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025576413845280875,
      "loss": 2.6201,
      "theoretical_loss": 3.4629359361251506,
      "tokens_seen": 1788084224
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002557475884244373,
      "loss": 2.6709,
      "theoretical_loss": 3.4629251193980153,
      "tokens_seen": 1788149760
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002557310383960658,
      "loss": 2.6123,
      "theoretical_loss": 3.462914303178305,
      "tokens_seen": 1788215296
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025571448836769436,
      "loss": 2.5233,
      "theoretical_loss": 3.4629034874659768,
      "tokens_seen": 1788280832
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025569793833932284,
      "loss": 2.6017,
      "theoretical_loss": 3.4628926722609887,
      "tokens_seen": 1788346368
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002556813883109514,
      "loss": 1.9524,
      "theoretical_loss": 3.4628818575632985,
      "tokens_seen": 1788411904
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002556648382825799,
      "loss": 2.45,
      "theoretical_loss": 3.4628710433728633,
      "tokens_seen": 1788477440
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025564828825420846,
      "loss": 2.4803,
      "theoretical_loss": 3.462860229689641,
      "tokens_seen": 1788542976
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025563173822583694,
      "loss": 2.7463,
      "theoretical_loss": 3.4628494165135892,
      "tokens_seen": 1788608512
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002556151881974655,
      "loss": 2.6181,
      "theoretical_loss": 3.4628386038446655,
      "tokens_seen": 1788674048
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000255598638169094,
      "loss": 2.6584,
      "theoretical_loss": 3.4628277916828276,
      "tokens_seen": 1788739584
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025558208814072255,
      "loss": 2.7097,
      "theoretical_loss": 3.462816980028033,
      "tokens_seen": 1788805120
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025556553811235103,
      "loss": 2.7329,
      "theoretical_loss": 3.46280616888024,
      "tokens_seen": 1788870656
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025554898808397957,
      "loss": 2.5886,
      "theoretical_loss": 3.462795358239405,
      "tokens_seen": 1788936192
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002555324380556081,
      "loss": 2.6127,
      "theoretical_loss": 3.462784548105487,
      "tokens_seen": 1789001728
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025551588802723665,
      "loss": 2.9131,
      "theoretical_loss": 3.4627737384784423,
      "tokens_seen": 1789067264
    },
    {
      "epoch": 0.64,
      "objective/train/avg_token_score": 0.011972409673035145,
      "objective/train/avg_weight": 0.9859357476234436,
      "objective/train/docs_used": 1011320,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6411986351013184,
      "objective/train/std_weight": 0.08135899901390076,
      "objective/train/theoretical_loss": 3.46276292935823,
      "objective/train/tokens_used": 1809592800,
      "theoretical_loss": 3.46276292935823,
      "tokens_seen": 1789132800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025549933799886513,
      "loss": 2.6787,
      "theoretical_loss": 3.46276292935823,
      "tokens_seen": 1789132800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025548278797049367,
      "loss": 2.7062,
      "theoretical_loss": 3.4627521207448066,
      "tokens_seen": 1789198336
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002554662379421222,
      "loss": 2.67,
      "theoretical_loss": 3.4627413126381303,
      "tokens_seen": 1789263872
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025544968791375074,
      "loss": 2.4756,
      "theoretical_loss": 3.462730505038159,
      "tokens_seen": 1789329408
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002554331378853792,
      "loss": 2.8046,
      "theoretical_loss": 3.46271969794485,
      "tokens_seen": 1789394944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025541658785700776,
      "loss": 2.6866,
      "theoretical_loss": 3.4627088913581607,
      "tokens_seen": 1789460480
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002554000378286363,
      "loss": 2.3663,
      "theoretical_loss": 3.4626980852780496,
      "tokens_seen": 1789526016
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025538348780026484,
      "loss": 2.3258,
      "theoretical_loss": 3.4626872797044737,
      "tokens_seen": 1789591552
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002553669377718933,
      "loss": 2.6114,
      "theoretical_loss": 3.4626764746373913,
      "tokens_seen": 1789657088
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002553503877435218,
      "loss": 2.579,
      "theoretical_loss": 3.4626656700767597,
      "tokens_seen": 1789722624
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025533383771515034,
      "loss": 2.8208,
      "theoretical_loss": 3.462654866022537,
      "tokens_seen": 1789788160
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002553172876867789,
      "loss": 2.4736,
      "theoretical_loss": 3.4626440624746797,
      "tokens_seen": 1789853696
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002553007376584074,
      "loss": 2.5896,
      "theoretical_loss": 3.4626332594331473,
      "tokens_seen": 1789919232
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002552841876300359,
      "loss": 2.4149,
      "theoretical_loss": 3.462622456897897,
      "tokens_seen": 1789984768
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025526763760166444,
      "loss": 2.2897,
      "theoretical_loss": 3.4626116548688857,
      "tokens_seen": 1790050304
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000255251087573293,
      "loss": 2.8336,
      "theoretical_loss": 3.462600853346072,
      "tokens_seen": 1790115840
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002552345375449215,
      "loss": 2.3666,
      "theoretical_loss": 3.4625900523294133,
      "tokens_seen": 1790181376
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025521798751655,
      "loss": 2.4542,
      "theoretical_loss": 3.462579251818867,
      "tokens_seen": 1790246912
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025520143748817854,
      "loss": 2.5686,
      "theoretical_loss": 3.4625684518143918,
      "tokens_seen": 1790312448
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002551848874598071,
      "loss": 2.6504,
      "theoretical_loss": 3.4625576523159447,
      "tokens_seen": 1790377984
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002551683374314356,
      "loss": 2.6909,
      "theoretical_loss": 3.4625468533234836,
      "tokens_seen": 1790443520
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002551517874030641,
      "loss": 2.5822,
      "theoretical_loss": 3.4625360548369666,
      "tokens_seen": 1790509056
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025513523737469263,
      "loss": 2.5868,
      "theoretical_loss": 3.462525256856351,
      "tokens_seen": 1790574592
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025511868734632117,
      "loss": 2.7407,
      "theoretical_loss": 3.462514459381595,
      "tokens_seen": 1790640128
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002551021373179497,
      "loss": 2.4706,
      "theoretical_loss": 3.462503662412656,
      "tokens_seen": 1790705664
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002550855872895782,
      "loss": 2.1484,
      "theoretical_loss": 3.4624928659494927,
      "tokens_seen": 1790771200
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025506903726120673,
      "loss": 2.3771,
      "theoretical_loss": 3.4624820699920615,
      "tokens_seen": 1790836736
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025505248723283526,
      "loss": 2.6041,
      "theoretical_loss": 3.462471274540321,
      "tokens_seen": 1790902272
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002550359372044638,
      "loss": 2.6677,
      "theoretical_loss": 3.4624604795942293,
      "tokens_seen": 1790967808
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002550193871760923,
      "loss": 2.4356,
      "theoretical_loss": 3.462449685153744,
      "tokens_seen": 1791033344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002550028371477208,
      "loss": 2.5403,
      "theoretical_loss": 3.462438891218822,
      "tokens_seen": 1791098880
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025498628711934936,
      "loss": 2.4594,
      "theoretical_loss": 3.4624280977894224,
      "tokens_seen": 1791164416
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002549697370909779,
      "loss": 2.5792,
      "theoretical_loss": 3.462417304865502,
      "tokens_seen": 1791229952
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002549531870626064,
      "loss": 2.8631,
      "theoretical_loss": 3.4624065124470196,
      "tokens_seen": 1791295488
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002549366370342349,
      "loss": 2.5515,
      "theoretical_loss": 3.462395720533933,
      "tokens_seen": 1791361024
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025492008700586346,
      "loss": 2.7013,
      "theoretical_loss": 3.462384929126199,
      "tokens_seen": 1791426560
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000254903536977492,
      "loss": 2.5577,
      "theoretical_loss": 3.462374138223776,
      "tokens_seen": 1791492096
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002548869869491205,
      "loss": 2.7479,
      "theoretical_loss": 3.4623633478266225,
      "tokens_seen": 1791557632
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025487043692074896,
      "loss": 2.8309,
      "theoretical_loss": 3.4623525579346954,
      "tokens_seen": 1791623168
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002548538868923775,
      "loss": 2.5645,
      "theoretical_loss": 3.462341768547953,
      "tokens_seen": 1791688704
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025483733686400604,
      "loss": 2.5166,
      "theoretical_loss": 3.462330979666353,
      "tokens_seen": 1791754240
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002548207868356346,
      "loss": 2.7532,
      "theoretical_loss": 3.462320191289854,
      "tokens_seen": 1791819776
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025480423680726306,
      "loss": 2.5088,
      "theoretical_loss": 3.4623094034184128,
      "tokens_seen": 1791885312
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002547876867788916,
      "loss": 2.7723,
      "theoretical_loss": 3.4622986160519877,
      "tokens_seen": 1791950848
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025477113675052013,
      "loss": 2.5585,
      "theoretical_loss": 3.4622878291905366,
      "tokens_seen": 1792016384
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025475458672214867,
      "loss": 2.5286,
      "theoretical_loss": 3.4622770428340184,
      "tokens_seen": 1792081920
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025473803669377715,
      "loss": 2.447,
      "theoretical_loss": 3.4622662569823888,
      "tokens_seen": 1792147456
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002547214866654057,
      "loss": 2.5843,
      "theoretical_loss": 3.4622554716356078,
      "tokens_seen": 1792212992
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025470493663703423,
      "loss": 2.7327,
      "theoretical_loss": 3.462244686793632,
      "tokens_seen": 1792278528
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025468838660866277,
      "loss": 2.3262,
      "theoretical_loss": 3.46223390245642,
      "tokens_seen": 1792344064
    },
    {
      "epoch": 0.64,
      "objective/train/avg_token_score": 0.015291470102965832,
      "objective/train/avg_weight": 0.9820017218589783,
      "objective/train/docs_used": 1013493,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 6.476727485656738,
      "objective/train/std_weight": 0.10320126265287399,
      "objective/train/theoretical_loss": 3.4622231186239296,
      "objective/train/tokens_used": 1812869600,
      "theoretical_loss": 3.4622231186239296,
      "tokens_seen": 1792409600
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025467183658029125,
      "loss": 2.3851,
      "theoretical_loss": 3.4622231186239296,
      "tokens_seen": 1792409600
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002546552865519198,
      "loss": 2.6962,
      "theoretical_loss": 3.4622123352961185,
      "tokens_seen": 1792475136
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002546387365235483,
      "loss": 2.6923,
      "theoretical_loss": 3.462201552472945,
      "tokens_seen": 1792540672
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025462218649517686,
      "loss": 2.4703,
      "theoretical_loss": 3.4621907701543666,
      "tokens_seen": 1792606208
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025460563646680535,
      "loss": 2.7854,
      "theoretical_loss": 3.462179988340341,
      "tokens_seen": 1792671744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002545890864384339,
      "loss": 2.6741,
      "theoretical_loss": 3.4621692070308274,
      "tokens_seen": 1792737280
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002545725364100624,
      "loss": 2.9281,
      "theoretical_loss": 3.4621584262257823,
      "tokens_seen": 1792802816
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025455598638169096,
      "loss": 2.6091,
      "theoretical_loss": 3.4621476459251648,
      "tokens_seen": 1792868352
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025453943635331944,
      "loss": 2.3704,
      "theoretical_loss": 3.4621368661289322,
      "tokens_seen": 1792933888
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000254522886324948,
      "loss": 2.5961,
      "theoretical_loss": 3.462126086837042,
      "tokens_seen": 1792999424
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002545063362965765,
      "loss": 2.7403,
      "theoretical_loss": 3.4621153080494538,
      "tokens_seen": 1793064960
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025448978626820505,
      "loss": 2.5879,
      "theoretical_loss": 3.462104529766124,
      "tokens_seen": 1793130496
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002544732362398336,
      "loss": 2.4729,
      "theoretical_loss": 3.462093751987011,
      "tokens_seen": 1793196032
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002544566862114621,
      "loss": 2.3507,
      "theoretical_loss": 3.4620829747120734,
      "tokens_seen": 1793261568
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002544401361830906,
      "loss": 2.4646,
      "theoretical_loss": 3.4620721979412687,
      "tokens_seen": 1793327104
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025442358615471915,
      "loss": 1.7827,
      "theoretical_loss": 3.462061421674554,
      "tokens_seen": 1793392640
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002544070361263477,
      "loss": 2.7976,
      "theoretical_loss": 3.46205064591189,
      "tokens_seen": 1793458176
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025439048609797617,
      "loss": 2.5257,
      "theoretical_loss": 3.462039870653231,
      "tokens_seen": 1793523712
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025437393606960465,
      "loss": 2.4936,
      "theoretical_loss": 3.4620290958985382,
      "tokens_seen": 1793589248
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002543573860412332,
      "loss": 2.4695,
      "theoretical_loss": 3.462018321647768,
      "tokens_seen": 1793654784
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025434083601286173,
      "loss": 2.6155,
      "theoretical_loss": 3.462007547900879,
      "tokens_seen": 1793720320
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025432428598449027,
      "loss": 2.7401,
      "theoretical_loss": 3.4619967746578286,
      "tokens_seen": 1793785856
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025430773595611875,
      "loss": 2.371,
      "theoretical_loss": 3.4619860019185755,
      "tokens_seen": 1793851392
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002542911859277473,
      "loss": 2.4909,
      "theoretical_loss": 3.4619752296830772,
      "tokens_seen": 1793916928
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002542746358993758,
      "loss": 2.5751,
      "theoretical_loss": 3.461964457951292,
      "tokens_seen": 1793982464
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025425808587100436,
      "loss": 2.5073,
      "theoretical_loss": 3.461953686723178,
      "tokens_seen": 1794048000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025424153584263285,
      "loss": 2.5731,
      "theoretical_loss": 3.461942915998694,
      "tokens_seen": 1794113536
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002542249858142614,
      "loss": 2.5483,
      "theoretical_loss": 3.4619321457777965,
      "tokens_seen": 1794179072
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002542084357858899,
      "loss": 2.5438,
      "theoretical_loss": 3.4619213760604444,
      "tokens_seen": 1794244608
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025419188575751846,
      "loss": 2.6811,
      "theoretical_loss": 3.4619106068465957,
      "tokens_seen": 1794310144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025417533572914694,
      "loss": 2.6937,
      "theoretical_loss": 3.4618998381362087,
      "tokens_seen": 1794375680
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002541587857007755,
      "loss": 2.6472,
      "theoretical_loss": 3.4618890699292413,
      "tokens_seen": 1794441216
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000254142235672404,
      "loss": 2.5097,
      "theoretical_loss": 3.461878302225651,
      "tokens_seen": 1794506752
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025412568564403255,
      "loss": 2.6149,
      "theoretical_loss": 3.4618675350253967,
      "tokens_seen": 1794572288
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025410913561566104,
      "loss": 2.4375,
      "theoretical_loss": 3.4618567683284365,
      "tokens_seen": 1794637824
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002540925855872896,
      "loss": 2.7651,
      "theoretical_loss": 3.4618460021347275,
      "tokens_seen": 1794703360
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002540760355589181,
      "loss": 2.5997,
      "theoretical_loss": 3.4618352364442293,
      "tokens_seen": 1794768896
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025405948553054665,
      "loss": 2.2768,
      "theoretical_loss": 3.4618244712568984,
      "tokens_seen": 1794834432
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025404293550217513,
      "loss": 2.7674,
      "theoretical_loss": 3.4618137065726944,
      "tokens_seen": 1794899968
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025402638547380367,
      "loss": 2.5099,
      "theoretical_loss": 3.4618029423915746,
      "tokens_seen": 1794965504
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002540098354454322,
      "loss": 2.5178,
      "theoretical_loss": 3.461792178713497,
      "tokens_seen": 1795031040
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025399328541706075,
      "loss": 2.4806,
      "theoretical_loss": 3.46178141553842,
      "tokens_seen": 1795096576
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025397673538868923,
      "loss": 2.3675,
      "theoretical_loss": 3.461770652866302,
      "tokens_seen": 1795162112
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025396018536031777,
      "loss": 2.5273,
      "theoretical_loss": 3.4617598906971008,
      "tokens_seen": 1795227648
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002539436353319463,
      "loss": 2.6582,
      "theoretical_loss": 3.4617491290307743,
      "tokens_seen": 1795293184
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025392708530357484,
      "loss": 2.4869,
      "theoretical_loss": 3.461738367867281,
      "tokens_seen": 1795358720
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002539105352752033,
      "loss": 2.7526,
      "theoretical_loss": 3.461727607206579,
      "tokens_seen": 1795424256
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002538939852468318,
      "loss": 2.8148,
      "theoretical_loss": 3.461716847048627,
      "tokens_seen": 1795489792
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025387743521846035,
      "loss": 2.5559,
      "theoretical_loss": 3.461706087393382,
      "tokens_seen": 1795555328
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002538608851900889,
      "loss": 2.5236,
      "theoretical_loss": 3.461695328240803,
      "tokens_seen": 1795620864
    },
    {
      "epoch": 0.64,
      "objective/train/avg_token_score": 0.02414952591061592,
      "objective/train/avg_weight": 0.9715981483459473,
      "objective/train/docs_used": 1015450,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.148921489715576,
      "objective/train/std_weight": 0.09672360867261887,
      "objective/train/theoretical_loss": 3.461684569590848,
      "objective/train/tokens_used": 1816146400,
      "theoretical_loss": 3.461684569590848,
      "tokens_seen": 1795686400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002538443351617174,
      "loss": 2.5204,
      "theoretical_loss": 3.461684569590848,
      "tokens_seen": 1795686400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002538277851333459,
      "loss": 2.6436,
      "theoretical_loss": 3.461673811443475,
      "tokens_seen": 1795751936
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025381123510497444,
      "loss": 2.4454,
      "theoretical_loss": 3.4616630537986426,
      "tokens_seen": 1795817472
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000253794685076603,
      "loss": 2.5249,
      "theoretical_loss": 3.4616522966563084,
      "tokens_seen": 1795883008
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002537781350482315,
      "loss": 2.8079,
      "theoretical_loss": 3.461641540016431,
      "tokens_seen": 1795948544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025376158501986,
      "loss": 2.6132,
      "theoretical_loss": 3.4616307838789684,
      "tokens_seen": 1796014080
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025374503499148854,
      "loss": 2.5311,
      "theoretical_loss": 3.4616200282438783,
      "tokens_seen": 1796079616
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002537284849631171,
      "loss": 2.3275,
      "theoretical_loss": 3.4616092731111205,
      "tokens_seen": 1796145152
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002537119349347456,
      "loss": 2.4743,
      "theoretical_loss": 3.4615985184806517,
      "tokens_seen": 1796210688
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002536953849063741,
      "loss": 2.6404,
      "theoretical_loss": 3.461587764352431,
      "tokens_seen": 1796276224
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025367883487800263,
      "loss": 2.7691,
      "theoretical_loss": 3.4615770107264154,
      "tokens_seen": 1796341760
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025366228484963117,
      "loss": 2.6087,
      "theoretical_loss": 3.461566257602565,
      "tokens_seen": 1796407296
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002536457348212597,
      "loss": 2.4122,
      "theoretical_loss": 3.4615555049808364,
      "tokens_seen": 1796472832
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002536291847928882,
      "loss": 1.3956,
      "theoretical_loss": 3.4615447528611885,
      "tokens_seen": 1796538368
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025361263476451673,
      "loss": 2.4064,
      "theoretical_loss": 3.46153400124358,
      "tokens_seen": 1796603904
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025359608473614527,
      "loss": 2.6209,
      "theoretical_loss": 3.461523250127968,
      "tokens_seen": 1796669440
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002535795347077738,
      "loss": 2.6836,
      "theoretical_loss": 3.4615124995143116,
      "tokens_seen": 1796734976
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002535629846794023,
      "loss": 2.3736,
      "theoretical_loss": 3.4615017494025686,
      "tokens_seen": 1796800512
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002535464346510308,
      "loss": 2.6723,
      "theoretical_loss": 3.461490999792698,
      "tokens_seen": 1796866048
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025352988462265936,
      "loss": 2.6368,
      "theoretical_loss": 3.461480250684657,
      "tokens_seen": 1796931584
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002535133345942879,
      "loss": 2.6273,
      "theoretical_loss": 3.4614695020784048,
      "tokens_seen": 1796997120
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002534967845659164,
      "loss": 2.3467,
      "theoretical_loss": 3.4614587539738992,
      "tokens_seen": 1797062656
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002534802345375449,
      "loss": 2.3936,
      "theoretical_loss": 3.4614480063710986,
      "tokens_seen": 1797128192
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025346368450917346,
      "loss": 2.8846,
      "theoretical_loss": 3.461437259269961,
      "tokens_seen": 1797193728
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000253447134480802,
      "loss": 2.4288,
      "theoretical_loss": 3.4614265126704455,
      "tokens_seen": 1797259264
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002534305844524305,
      "loss": 2.4461,
      "theoretical_loss": 3.46141576657251,
      "tokens_seen": 1797324800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000253414034424059,
      "loss": 2.7114,
      "theoretical_loss": 3.461405020976112,
      "tokens_seen": 1797390336
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002533974843956875,
      "loss": 2.6188,
      "theoretical_loss": 3.4613942758812106,
      "tokens_seen": 1797455872
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025338093436731604,
      "loss": 2.6076,
      "theoretical_loss": 3.4613835312877645,
      "tokens_seen": 1797521408
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002533643843389446,
      "loss": 2.1443,
      "theoretical_loss": 3.461372787195731,
      "tokens_seen": 1797586944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025334783431057306,
      "loss": 2.6253,
      "theoretical_loss": 3.461362043605069,
      "tokens_seen": 1797652480
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002533312842822016,
      "loss": 2.55,
      "theoretical_loss": 3.4613513005157373,
      "tokens_seen": 1797718016
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025331473425383014,
      "loss": 2.5194,
      "theoretical_loss": 3.4613405579276932,
      "tokens_seen": 1797783552
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025329818422545867,
      "loss": 2.548,
      "theoretical_loss": 3.4613298158408954,
      "tokens_seen": 1797849088
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025328163419708716,
      "loss": 2.5049,
      "theoretical_loss": 3.461319074255303,
      "tokens_seen": 1797914624
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002532650841687157,
      "loss": 2.4328,
      "theoretical_loss": 3.461308333170873,
      "tokens_seen": 1797980160
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025324853414034423,
      "loss": 2.6768,
      "theoretical_loss": 3.461297592587565,
      "tokens_seen": 1798045696
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025323198411197277,
      "loss": 2.5356,
      "theoretical_loss": 3.461286852505336,
      "tokens_seen": 1798111232
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002532154340836013,
      "loss": 2.837,
      "theoretical_loss": 3.4612761129241463,
      "tokens_seen": 1798176768
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002531988840552298,
      "loss": 2.7116,
      "theoretical_loss": 3.4612653738439523,
      "tokens_seen": 1798242304
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002531823340268583,
      "loss": 2.055,
      "theoretical_loss": 3.4612546352647136,
      "tokens_seen": 1798307840
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025316578399848686,
      "loss": 2.5141,
      "theoretical_loss": 3.461243897186388,
      "tokens_seen": 1798373376
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002531492339701154,
      "loss": 1.727,
      "theoretical_loss": 3.461233159608934,
      "tokens_seen": 1798438912
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002531326839417439,
      "loss": 2.582,
      "theoretical_loss": 3.4612224225323103,
      "tokens_seen": 1798504448
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002531161339133724,
      "loss": 2.4998,
      "theoretical_loss": 3.461211685956475,
      "tokens_seen": 1798569984
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025309958388500096,
      "loss": 2.264,
      "theoretical_loss": 3.4612009498813867,
      "tokens_seen": 1798635520
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002530830338566295,
      "loss": 2.8228,
      "theoretical_loss": 3.4611902143070035,
      "tokens_seen": 1798701056
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000253066483828258,
      "loss": 2.5569,
      "theoretical_loss": 3.4611794792332837,
      "tokens_seen": 1798766592
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002530499337998865,
      "loss": 2.6682,
      "theoretical_loss": 3.4611687446601866,
      "tokens_seen": 1798832128
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025303338377151506,
      "loss": 2.4487,
      "theoretical_loss": 3.4611580105876696,
      "tokens_seen": 1798897664
    },
    {
      "epoch": 0.64,
      "objective/train/avg_token_score": 0.01910346746444702,
      "objective/train/avg_weight": 0.9775212407112122,
      "objective/train/docs_used": 1017085,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.819819211959839,
      "objective/train/std_weight": 0.09031400829553604,
      "objective/train/theoretical_loss": 3.4611472770156917,
      "objective/train/tokens_used": 1819423200,
      "theoretical_loss": 3.4611472770156917,
      "tokens_seen": 1798963200
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002530168337431436,
      "loss": 2.6207,
      "theoretical_loss": 3.4611472770156917,
      "tokens_seen": 1798963200
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002530002837147721,
      "loss": 2.8278,
      "theoretical_loss": 3.4611365439442103,
      "tokens_seen": 1799028736
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002529837336864006,
      "loss": 2.5648,
      "theoretical_loss": 3.4611258113731855,
      "tokens_seen": 1799094272
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025296718365802915,
      "loss": 2.7341,
      "theoretical_loss": 3.4611150793025747,
      "tokens_seen": 1799159808
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002529506336296577,
      "loss": 2.701,
      "theoretical_loss": 3.461104347732337,
      "tokens_seen": 1799225344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002529340836012862,
      "loss": 2.7668,
      "theoretical_loss": 3.4610936166624295,
      "tokens_seen": 1799290880
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002529175335729147,
      "loss": 2.7876,
      "theoretical_loss": 3.4610828860928122,
      "tokens_seen": 1799356416
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002529009835445432,
      "loss": 2.3373,
      "theoretical_loss": 3.4610721560234428,
      "tokens_seen": 1799421952
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025288443351617173,
      "loss": 2.5215,
      "theoretical_loss": 3.46106142645428,
      "tokens_seen": 1799487488
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025286788348780027,
      "loss": 2.6496,
      "theoretical_loss": 3.4610506973852813,
      "tokens_seen": 1799553024
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025285133345942875,
      "loss": 2.7631,
      "theoretical_loss": 3.4610399688164066,
      "tokens_seen": 1799618560
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002528347834310573,
      "loss": 2.6834,
      "theoretical_loss": 3.4610292407476138,
      "tokens_seen": 1799684096
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025281823340268583,
      "loss": 2.3944,
      "theoretical_loss": 3.4610185131788613,
      "tokens_seen": 1799749632
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025280168337431437,
      "loss": 2.416,
      "theoretical_loss": 3.461007786110108,
      "tokens_seen": 1799815168
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025278513334594285,
      "loss": 2.611,
      "theoretical_loss": 3.4609970595413113,
      "tokens_seen": 1799880704
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002527685833175714,
      "loss": 2.7377,
      "theoretical_loss": 3.4609863334724307,
      "tokens_seen": 1799946240
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002527520332891999,
      "loss": 2.6974,
      "theoretical_loss": 3.460975607903425,
      "tokens_seen": 1800011776
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025273548326082846,
      "loss": 2.5745,
      "theoretical_loss": 3.460964882834251,
      "tokens_seen": 1800077312
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025271893323245694,
      "loss": 2.5712,
      "theoretical_loss": 3.460954158264869,
      "tokens_seen": 1800142848
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002527023832040855,
      "loss": 2.4288,
      "theoretical_loss": 3.460943434195237,
      "tokens_seen": 1800208384
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000252685833175714,
      "loss": 2.4101,
      "theoretical_loss": 3.460932710625313,
      "tokens_seen": 1800273920
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025266928314734256,
      "loss": 2.4644,
      "theoretical_loss": 3.4609219875550563,
      "tokens_seen": 1800339456
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025265273311897104,
      "loss": 1.7667,
      "theoretical_loss": 3.460911264984425,
      "tokens_seen": 1800404992
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002526361830905996,
      "loss": 2.6747,
      "theoretical_loss": 3.460900542913377,
      "tokens_seen": 1800470528
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002526196330622281,
      "loss": 2.77,
      "theoretical_loss": 3.460889821341872,
      "tokens_seen": 1800536064
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025260308303385665,
      "loss": 2.5189,
      "theoretical_loss": 3.4608791002698682,
      "tokens_seen": 1800601600
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025258653300548514,
      "loss": 2.5145,
      "theoretical_loss": 3.460868379697324,
      "tokens_seen": 1800667136
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002525699829771137,
      "loss": 2.5665,
      "theoretical_loss": 3.460857659624198,
      "tokens_seen": 1800732672
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002525534329487422,
      "loss": 2.497,
      "theoretical_loss": 3.4608469400504482,
      "tokens_seen": 1800798208
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025253688292037075,
      "loss": 2.6319,
      "theoretical_loss": 3.4608362209760335,
      "tokens_seen": 1800863744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025252033289199923,
      "loss": 2.5112,
      "theoretical_loss": 3.4608255024009136,
      "tokens_seen": 1800929280
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025250378286362777,
      "loss": 2.5304,
      "theoretical_loss": 3.4608147843250454,
      "tokens_seen": 1800994816
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002524872328352563,
      "loss": 2.3352,
      "theoretical_loss": 3.4608040667483886,
      "tokens_seen": 1801060352
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025247068280688485,
      "loss": 2.5208,
      "theoretical_loss": 3.460793349670901,
      "tokens_seen": 1801125888
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025245413277851333,
      "loss": 2.4488,
      "theoretical_loss": 3.460782633092542,
      "tokens_seen": 1801191424
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025243758275014187,
      "loss": 2.3617,
      "theoretical_loss": 3.4607719170132696,
      "tokens_seen": 1801256960
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025242103272177035,
      "loss": 2.2978,
      "theoretical_loss": 3.4607612014330424,
      "tokens_seen": 1801322496
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002524044826933989,
      "loss": 2.5694,
      "theoretical_loss": 3.4607504863518193,
      "tokens_seen": 1801388032
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002523879326650274,
      "loss": 2.6699,
      "theoretical_loss": 3.460739771769559,
      "tokens_seen": 1801453568
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002523713826366559,
      "loss": 2.6756,
      "theoretical_loss": 3.46072905768622,
      "tokens_seen": 1801519104
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025235483260828445,
      "loss": 2.7162,
      "theoretical_loss": 3.4607183441017604,
      "tokens_seen": 1801584640
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000252338282579913,
      "loss": 2.6442,
      "theoretical_loss": 3.460707631016139,
      "tokens_seen": 1801650176
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002523217325515415,
      "loss": 2.5245,
      "theoretical_loss": 3.4606969184293153,
      "tokens_seen": 1801715712
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025230518252317,
      "loss": 2.5319,
      "theoretical_loss": 3.460686206341247,
      "tokens_seen": 1801781248
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025228863249479854,
      "loss": 2.5207,
      "theoretical_loss": 3.4606754947518934,
      "tokens_seen": 1801846784
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002522720824664271,
      "loss": 2.669,
      "theoretical_loss": 3.4606647836612128,
      "tokens_seen": 1801912320
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002522555324380556,
      "loss": 2.6633,
      "theoretical_loss": 3.460654073069163,
      "tokens_seen": 1801977856
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002522389824096841,
      "loss": 2.7069,
      "theoretical_loss": 3.4606433629757043,
      "tokens_seen": 1802043392
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025222243238131264,
      "loss": 2.587,
      "theoretical_loss": 3.4606326533807943,
      "tokens_seen": 1802108928
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002522058823529412,
      "loss": 2.6372,
      "theoretical_loss": 3.4606219442843917,
      "tokens_seen": 1802174464
    },
    {
      "epoch": 0.64,
      "objective/train/avg_token_score": 0.013695425353944302,
      "objective/train/avg_weight": 0.9838804006576538,
      "objective/train/docs_used": 1018281,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.118741989135742,
      "objective/train/std_weight": 0.07616222649812698,
      "objective/train/theoretical_loss": 3.460611235686456,
      "objective/train/tokens_used": 1822700000,
      "theoretical_loss": 3.460611235686456,
      "tokens_seen": 1802240000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002521893323245697,
      "loss": 2.5381,
      "theoretical_loss": 3.460611235686456,
      "tokens_seen": 1802240000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002521727822961982,
      "loss": 2.6422,
      "theoretical_loss": 3.4606005275869447,
      "tokens_seen": 1802305536
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025215623226782673,
      "loss": 2.5062,
      "theoretical_loss": 3.460589819985817,
      "tokens_seen": 1802371072
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025213968223945527,
      "loss": 2.7904,
      "theoretical_loss": 3.460579112883032,
      "tokens_seen": 1802436608
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002521231322110838,
      "loss": 2.543,
      "theoretical_loss": 3.460568406278548,
      "tokens_seen": 1802502144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002521065821827123,
      "loss": 2.4549,
      "theoretical_loss": 3.4605577001723233,
      "tokens_seen": 1802567680
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025209003215434083,
      "loss": 2.5364,
      "theoretical_loss": 3.460546994564317,
      "tokens_seen": 1802633216
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025207348212596937,
      "loss": 2.6195,
      "theoretical_loss": 3.460536289454488,
      "tokens_seen": 1802698752
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002520569320975979,
      "loss": 2.4459,
      "theoretical_loss": 3.460525584842795,
      "tokens_seen": 1802764288
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002520403820692264,
      "loss": 2.68,
      "theoretical_loss": 3.4605148807291966,
      "tokens_seen": 1802829824
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002520238320408549,
      "loss": 2.4253,
      "theoretical_loss": 3.460504177113651,
      "tokens_seen": 1802895360
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025200728201248346,
      "loss": 2.5588,
      "theoretical_loss": 3.4604934739961175,
      "tokens_seen": 1802960896
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000251990731984112,
      "loss": 2.4853,
      "theoretical_loss": 3.460482771376555,
      "tokens_seen": 1803026432
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002519741819557405,
      "loss": 2.5624,
      "theoretical_loss": 3.460472069254922,
      "tokens_seen": 1803091968
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000251957631927369,
      "loss": 2.5185,
      "theoretical_loss": 3.4604613676311766,
      "tokens_seen": 1803157504
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025194108189899756,
      "loss": 2.627,
      "theoretical_loss": 3.4604506665052783,
      "tokens_seen": 1803223040
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025192453187062604,
      "loss": 2.5262,
      "theoretical_loss": 3.4604399658771863,
      "tokens_seen": 1803288576
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002519079818422546,
      "loss": 2.6475,
      "theoretical_loss": 3.460429265746858,
      "tokens_seen": 1803354112
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002518914318138831,
      "loss": 2.6581,
      "theoretical_loss": 3.460418566114253,
      "tokens_seen": 1803419648
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002518748817855116,
      "loss": 2.6918,
      "theoretical_loss": 3.4604078669793297,
      "tokens_seen": 1803485184
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025185833175714014,
      "loss": 2.5061,
      "theoretical_loss": 3.4603971683420474,
      "tokens_seen": 1803550720
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002518417817287687,
      "loss": 2.6148,
      "theoretical_loss": 3.4603864702023643,
      "tokens_seen": 1803616256
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002518252317003972,
      "loss": 2.7352,
      "theoretical_loss": 3.4603757725602398,
      "tokens_seen": 1803681792
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002518086816720257,
      "loss": 2.4529,
      "theoretical_loss": 3.4603650754156323,
      "tokens_seen": 1803747328
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025179213164365423,
      "loss": 2.6198,
      "theoretical_loss": 3.4603543787685,
      "tokens_seen": 1803812864
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025177558161528277,
      "loss": 2.5862,
      "theoretical_loss": 3.460343682618803,
      "tokens_seen": 1803878400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002517590315869113,
      "loss": 2.5861,
      "theoretical_loss": 3.460332986966499,
      "tokens_seen": 1803943936
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002517424815585398,
      "loss": 2.3746,
      "theoretical_loss": 3.4603222918115475,
      "tokens_seen": 1804009472
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025172593153016833,
      "loss": 2.5842,
      "theoretical_loss": 3.4603115971539067,
      "tokens_seen": 1804075008
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025170938150179687,
      "loss": 2.5666,
      "theoretical_loss": 3.460300902993536,
      "tokens_seen": 1804140544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002516928314734254,
      "loss": 2.3837,
      "theoretical_loss": 3.4602902093303936,
      "tokens_seen": 1804206080
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002516762814450539,
      "loss": 2.626,
      "theoretical_loss": 3.460279516164439,
      "tokens_seen": 1804271616
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002516597314166824,
      "loss": 2.3656,
      "theoretical_loss": 3.460268823495631,
      "tokens_seen": 1804337152
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025164318138831096,
      "loss": 2.2827,
      "theoretical_loss": 3.4602581313239273,
      "tokens_seen": 1804402688
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002516266313599395,
      "loss": 2.4601,
      "theoretical_loss": 3.460247439649288,
      "tokens_seen": 1804468224
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000251610081331568,
      "loss": 2.4829,
      "theoretical_loss": 3.4602367484716714,
      "tokens_seen": 1804533760
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002515935313031965,
      "loss": 2.6186,
      "theoretical_loss": 3.4602260577910364,
      "tokens_seen": 1804599296
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025157698127482506,
      "loss": 2.3793,
      "theoretical_loss": 3.4602153676073417,
      "tokens_seen": 1804664832
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002515604312464536,
      "loss": 2.5992,
      "theoretical_loss": 3.4602046779205464,
      "tokens_seen": 1804730368
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002515438812180821,
      "loss": 2.6514,
      "theoretical_loss": 3.4601939887306097,
      "tokens_seen": 1804795904
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002515273311897106,
      "loss": 2.5701,
      "theoretical_loss": 3.4601833000374898,
      "tokens_seen": 1804861440
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025151078116133916,
      "loss": 2.6772,
      "theoretical_loss": 3.4601726118411458,
      "tokens_seen": 1804926976
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002514942311329677,
      "loss": 2.558,
      "theoretical_loss": 3.460161924141537,
      "tokens_seen": 1804992512
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002514776811045962,
      "loss": 2.5551,
      "theoretical_loss": 3.4601512369386214,
      "tokens_seen": 1805058048
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002514611310762247,
      "loss": 2.6566,
      "theoretical_loss": 3.4601405502323583,
      "tokens_seen": 1805123584
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025144458104785325,
      "loss": 2.3606,
      "theoretical_loss": 3.460129864022707,
      "tokens_seen": 1805189120
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025142803101948173,
      "loss": 2.5592,
      "theoretical_loss": 3.460119178309626,
      "tokens_seen": 1805254656
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025141148099111027,
      "loss": 2.4214,
      "theoretical_loss": 3.4601084930930743,
      "tokens_seen": 1805320192
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025139493096273876,
      "loss": 2.5909,
      "theoretical_loss": 3.4600978083730105,
      "tokens_seen": 1805385728
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002513783809343673,
      "loss": 2.6497,
      "theoretical_loss": 3.460087124149394,
      "tokens_seen": 1805451264
    },
    {
      "epoch": 0.64,
      "objective/train/avg_token_score": 0.006307967938482761,
      "objective/train/avg_weight": 0.9925777912139893,
      "objective/train/docs_used": 1020317,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9873039722442627,
      "objective/train/std_weight": 0.04272739589214325,
      "objective/train/theoretical_loss": 3.4600764404221835,
      "objective/train/tokens_used": 1825976800,
      "theoretical_loss": 3.4600764404221835,
      "tokens_seen": 1805516800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025136183090599583,
      "loss": 2.709,
      "theoretical_loss": 3.4600764404221835,
      "tokens_seen": 1805516800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025134528087762437,
      "loss": 2.7655,
      "theoretical_loss": 3.460065757191338,
      "tokens_seen": 1805582336
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025132873084925285,
      "loss": 2.5849,
      "theoretical_loss": 3.460055074456816,
      "tokens_seen": 1805647872
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002513121808208814,
      "loss": 2.3761,
      "theoretical_loss": 3.4600443922185766,
      "tokens_seen": 1805713408
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002512956307925099,
      "loss": 2.5972,
      "theoretical_loss": 3.4600337104765795,
      "tokens_seen": 1805778944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025127908076413846,
      "loss": 2.5733,
      "theoretical_loss": 3.4600230292307828,
      "tokens_seen": 1805844480
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00025126253073576695,
      "loss": 2.6317,
      "theoretical_loss": 3.4600123484811456,
      "tokens_seen": 1805910016
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002512459807073955,
      "loss": 2.4832,
      "theoretical_loss": 3.460001668227627,
      "tokens_seen": 1805975552
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000251229430679024,
      "loss": 2.7237,
      "theoretical_loss": 3.459990988470186,
      "tokens_seen": 1806041088
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025121288065065256,
      "loss": 2.6218,
      "theoretical_loss": 3.459980309208781,
      "tokens_seen": 1806106624
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025119633062228104,
      "loss": 2.5589,
      "theoretical_loss": 3.459969630443372,
      "tokens_seen": 1806172160
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002511797805939096,
      "loss": 2.7692,
      "theoretical_loss": 3.459958952173917,
      "tokens_seen": 1806237696
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002511632305655381,
      "loss": 2.7008,
      "theoretical_loss": 3.459948274400375,
      "tokens_seen": 1806303232
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025114668053716666,
      "loss": 2.0587,
      "theoretical_loss": 3.459937597122706,
      "tokens_seen": 1806368768
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025113013050879514,
      "loss": 2.6953,
      "theoretical_loss": 3.459926920340868,
      "tokens_seen": 1806434304
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002511135804804237,
      "loss": 2.5235,
      "theoretical_loss": 3.4599162440548197,
      "tokens_seen": 1806499840
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002510970304520522,
      "loss": 2.5729,
      "theoretical_loss": 3.4599055682645217,
      "tokens_seen": 1806565376
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025108048042368075,
      "loss": 2.5894,
      "theoretical_loss": 3.459894892969931,
      "tokens_seen": 1806630912
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025106393039530924,
      "loss": 2.3992,
      "theoretical_loss": 3.459884218171008,
      "tokens_seen": 1806696448
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025104738036693777,
      "loss": 2.6797,
      "theoretical_loss": 3.4598735438677117,
      "tokens_seen": 1806761984
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002510308303385663,
      "loss": 2.4914,
      "theoretical_loss": 3.4598628700600003,
      "tokens_seen": 1806827520
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025101428031019485,
      "loss": 2.691,
      "theoretical_loss": 3.4598521967478337,
      "tokens_seen": 1806893056
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025099773028182333,
      "loss": 2.616,
      "theoretical_loss": 3.4598415239311695,
      "tokens_seen": 1806958592
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025098118025345187,
      "loss": 2.2893,
      "theoretical_loss": 3.4598308516099685,
      "tokens_seen": 1807024128
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002509646302250804,
      "loss": 2.6243,
      "theoretical_loss": 3.4598201797841885,
      "tokens_seen": 1807089664
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025094808019670894,
      "loss": 2.5766,
      "theoretical_loss": 3.459809508453789,
      "tokens_seen": 1807155200
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025093153016833743,
      "loss": 2.4434,
      "theoretical_loss": 3.459798837618729,
      "tokens_seen": 1807220736
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002509149801399659,
      "loss": 2.5476,
      "theoretical_loss": 3.4597881672789677,
      "tokens_seen": 1807286272
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025089843011159445,
      "loss": 2.5157,
      "theoretical_loss": 3.4597774974344637,
      "tokens_seen": 1807351808
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000250881880083223,
      "loss": 2.5115,
      "theoretical_loss": 3.4597668280851765,
      "tokens_seen": 1807417344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002508653300548515,
      "loss": 2.6373,
      "theoretical_loss": 3.4597561592310644,
      "tokens_seen": 1807482880
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025084878002648,
      "loss": 2.4011,
      "theoretical_loss": 3.4597454908720877,
      "tokens_seen": 1807548416
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025083222999810854,
      "loss": 2.561,
      "theoretical_loss": 3.459734823008205,
      "tokens_seen": 1807613952
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002508156799697371,
      "loss": 2.61,
      "theoretical_loss": 3.4597241556393747,
      "tokens_seen": 1807679488
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002507991299413656,
      "loss": 2.5477,
      "theoretical_loss": 3.4597134887655567,
      "tokens_seen": 1807745024
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002507825799129941,
      "loss": 2.335,
      "theoretical_loss": 3.459702822386709,
      "tokens_seen": 1807810560
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025076602988462264,
      "loss": 2.5233,
      "theoretical_loss": 3.4596921565027925,
      "tokens_seen": 1807876096
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002507494798562512,
      "loss": 2.4842,
      "theoretical_loss": 3.4596814911137646,
      "tokens_seen": 1807941632
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002507329298278797,
      "loss": 2.2536,
      "theoretical_loss": 3.459670826219585,
      "tokens_seen": 1808007168
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002507163797995082,
      "loss": 2.583,
      "theoretical_loss": 3.459660161820213,
      "tokens_seen": 1808072704
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025069982977113674,
      "loss": 2.7088,
      "theoretical_loss": 3.459649497915608,
      "tokens_seen": 1808138240
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002506832797427653,
      "loss": 2.5407,
      "theoretical_loss": 3.459638834505728,
      "tokens_seen": 1808203776
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002506667297143938,
      "loss": 2.8156,
      "theoretical_loss": 3.4596281715905333,
      "tokens_seen": 1808269312
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002506501796860223,
      "loss": 2.4606,
      "theoretical_loss": 3.4596175091699823,
      "tokens_seen": 1808334848
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025063362965765083,
      "loss": 2.6049,
      "theoretical_loss": 3.459606847244034,
      "tokens_seen": 1808400384
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025061707962927937,
      "loss": 2.6168,
      "theoretical_loss": 3.4595961858126483,
      "tokens_seen": 1808465920
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002506005296009079,
      "loss": 1.4624,
      "theoretical_loss": 3.459585524875784,
      "tokens_seen": 1808531456
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002505839795725364,
      "loss": 2.7361,
      "theoretical_loss": 3.4595748644334,
      "tokens_seen": 1808596992
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025056742954416493,
      "loss": 2.4994,
      "theoretical_loss": 3.4595642044854555,
      "tokens_seen": 1808662528
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025055087951579347,
      "loss": 2.2129,
      "theoretical_loss": 3.45955354503191,
      "tokens_seen": 1808728064
    },
    {
      "epoch": 0.65,
      "objective/train/avg_token_score": 0.021012721583247185,
      "objective/train/avg_weight": 0.975270688533783,
      "objective/train/docs_used": 1022235,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7896814346313477,
      "objective/train/std_weight": 0.0834147036075592,
      "objective/train/theoretical_loss": 3.4595428860727218,
      "objective/train/tokens_used": 1829253600,
      "theoretical_loss": 3.4595428860727218,
      "tokens_seen": 1808793600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000250534329487422,
      "loss": 2.6892,
      "theoretical_loss": 3.4595428860727218,
      "tokens_seen": 1808793600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002505177794590505,
      "loss": 2.3882,
      "theoretical_loss": 3.4595322276078515,
      "tokens_seen": 1808859136
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000250501229430679,
      "loss": 2.3534,
      "theoretical_loss": 3.459521569637257,
      "tokens_seen": 1808924672
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025048467940230756,
      "loss": 2.3194,
      "theoretical_loss": 3.459510912160898,
      "tokens_seen": 1808990208
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002504681293739361,
      "loss": 2.7419,
      "theoretical_loss": 3.4595002551787335,
      "tokens_seen": 1809055744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002504515793455646,
      "loss": 2.5175,
      "theoretical_loss": 3.4594895986907233,
      "tokens_seen": 1809121280
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002504350293171931,
      "loss": 2.4071,
      "theoretical_loss": 3.4594789426968253,
      "tokens_seen": 1809186816
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002504184792888216,
      "loss": 2.4389,
      "theoretical_loss": 3.4594682871969997,
      "tokens_seen": 1809252352
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025040192926045014,
      "loss": 2.4882,
      "theoretical_loss": 3.4594576321912056,
      "tokens_seen": 1809317888
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002503853792320787,
      "loss": 2.4858,
      "theoretical_loss": 3.459446977679402,
      "tokens_seen": 1809383424
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002503688292037072,
      "loss": 2.6548,
      "theoretical_loss": 3.4594363236615484,
      "tokens_seen": 1809448960
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002503522791753357,
      "loss": 2.5634,
      "theoretical_loss": 3.4594256701376036,
      "tokens_seen": 1809514496
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025033572914696424,
      "loss": 2.6088,
      "theoretical_loss": 3.459415017107527,
      "tokens_seen": 1809580032
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002503191791185928,
      "loss": 2.5034,
      "theoretical_loss": 3.4594043645712773,
      "tokens_seen": 1809645568
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002503026290902213,
      "loss": 2.4478,
      "theoretical_loss": 3.459393712528815,
      "tokens_seen": 1809711104
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002502860790618498,
      "loss": 2.5123,
      "theoretical_loss": 3.459383060980098,
      "tokens_seen": 1809776640
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025026952903347833,
      "loss": 3.7477,
      "theoretical_loss": 3.459372409925087,
      "tokens_seen": 1809842176
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025025297900510687,
      "loss": 2.295,
      "theoretical_loss": 3.4593617593637394,
      "tokens_seen": 1809907712
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002502364289767354,
      "loss": 2.6221,
      "theoretical_loss": 3.4593511092960156,
      "tokens_seen": 1809973248
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002502198789483639,
      "loss": 2.8403,
      "theoretical_loss": 3.459340459721875,
      "tokens_seen": 1810038784
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025020332891999243,
      "loss": 2.6399,
      "theoretical_loss": 3.459329810641276,
      "tokens_seen": 1810104320
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025018677889162097,
      "loss": 2.5926,
      "theoretical_loss": 3.4593191620541783,
      "tokens_seen": 1810169856
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002501702288632495,
      "loss": 2.7211,
      "theoretical_loss": 3.4593085139605417,
      "tokens_seen": 1810235392
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000250153678834878,
      "loss": 2.527,
      "theoretical_loss": 3.459297866360325,
      "tokens_seen": 1810300928
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002501371288065065,
      "loss": 2.5066,
      "theoretical_loss": 3.459287219253487,
      "tokens_seen": 1810366464
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025012057877813506,
      "loss": 2.4589,
      "theoretical_loss": 3.459276572639988,
      "tokens_seen": 1810432000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002501040287497636,
      "loss": 2.4879,
      "theoretical_loss": 3.459265926519786,
      "tokens_seen": 1810497536
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002500874787213921,
      "loss": 2.4972,
      "theoretical_loss": 3.459255280892841,
      "tokens_seen": 1810563072
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002500709286930206,
      "loss": 2.5895,
      "theoretical_loss": 3.459244635759113,
      "tokens_seen": 1810628608
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00025005437866464916,
      "loss": 2.4264,
      "theoretical_loss": 3.459233991118561,
      "tokens_seen": 1810694144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002500378286362777,
      "loss": 2.5024,
      "theoretical_loss": 3.4592233469711426,
      "tokens_seen": 1810759680
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002500212786079062,
      "loss": 2.3546,
      "theoretical_loss": 3.459212703316819,
      "tokens_seen": 1810825216
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002500047285795347,
      "loss": 2.3361,
      "theoretical_loss": 3.4592020601555493,
      "tokens_seen": 1810890752
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024998817855116325,
      "loss": 2.4187,
      "theoretical_loss": 3.4591914174872915,
      "tokens_seen": 1810956288
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002499716285227918,
      "loss": 2.582,
      "theoretical_loss": 3.4591807753120065,
      "tokens_seen": 1811021824
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002499550784944203,
      "loss": 2.7319,
      "theoretical_loss": 3.4591701336296534,
      "tokens_seen": 1811087360
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024993852846604876,
      "loss": 2.6518,
      "theoretical_loss": 3.4591594924401905,
      "tokens_seen": 1811152896
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002499219784376773,
      "loss": 2.4862,
      "theoretical_loss": 3.459148851743578,
      "tokens_seen": 1811218432
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024990542840930583,
      "loss": 2.7441,
      "theoretical_loss": 3.459138211539775,
      "tokens_seen": 1811283968
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024988887838093437,
      "loss": 2.5554,
      "theoretical_loss": 3.4591275718287404,
      "tokens_seen": 1811349504
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024987232835256285,
      "loss": 2.2617,
      "theoretical_loss": 3.459116932610435,
      "tokens_seen": 1811415040
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002498557783241914,
      "loss": 2.4206,
      "theoretical_loss": 3.4591062938848163,
      "tokens_seen": 1811480576
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024983922829581993,
      "loss": 2.8402,
      "theoretical_loss": 3.4590956556518444,
      "tokens_seen": 1811546112
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024982267826744847,
      "loss": 2.6679,
      "theoretical_loss": 3.4590850179114794,
      "tokens_seen": 1811611648
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024980612823907695,
      "loss": 2.2236,
      "theoretical_loss": 3.45907438066368,
      "tokens_seen": 1811677184
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002497895782107055,
      "loss": 2.55,
      "theoretical_loss": 3.4590637439084055,
      "tokens_seen": 1811742720
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000249773028182334,
      "loss": 2.1953,
      "theoretical_loss": 3.459053107645615,
      "tokens_seen": 1811808256
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024975647815396256,
      "loss": 2.6471,
      "theoretical_loss": 3.459042471875269,
      "tokens_seen": 1811873792
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024973992812559105,
      "loss": 2.4269,
      "theoretical_loss": 3.4590318365973256,
      "tokens_seen": 1811939328
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002497233780972196,
      "loss": 2.5735,
      "theoretical_loss": 3.459021201811745,
      "tokens_seen": 1812004864
    },
    {
      "epoch": 0.65,
      "objective/train/avg_token_score": 0.024345165118575096,
      "objective/train/avg_weight": 0.9713844656944275,
      "objective/train/docs_used": 1024230,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.4497203826904297,
      "objective/train/std_weight": 0.12102269381284714,
      "objective/train/theoretical_loss": 3.4590105675184866,
      "objective/train/tokens_used": 1832530400,
      "theoretical_loss": 3.4590105675184866,
      "tokens_seen": 1812070400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002497068280688481,
      "loss": 2.3052,
      "theoretical_loss": 3.4590105675184866,
      "tokens_seen": 1812070400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024969027804047666,
      "loss": 2.5862,
      "theoretical_loss": 3.458999933717509,
      "tokens_seen": 1812135936
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024967372801210514,
      "loss": 2.6156,
      "theoretical_loss": 3.4589893004087733,
      "tokens_seen": 1812201472
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002496571779837337,
      "loss": 2.2113,
      "theoretical_loss": 3.4589786675922367,
      "tokens_seen": 1812267008
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002496406279553622,
      "loss": 2.6423,
      "theoretical_loss": 3.4589680352678602,
      "tokens_seen": 1812332544
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024962407792699075,
      "loss": 2.8823,
      "theoretical_loss": 3.4589574034356025,
      "tokens_seen": 1812398080
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024960752789861924,
      "loss": 2.5881,
      "theoretical_loss": 3.4589467720954232,
      "tokens_seen": 1812463616
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002495909778702478,
      "loss": 2.4655,
      "theoretical_loss": 3.4589361412472828,
      "tokens_seen": 1812529152
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002495744278418763,
      "loss": 2.5356,
      "theoretical_loss": 3.4589255108911385,
      "tokens_seen": 1812594688
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024955787781350485,
      "loss": 2.5792,
      "theoretical_loss": 3.458914881026952,
      "tokens_seen": 1812660224
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024954132778513333,
      "loss": 2.457,
      "theoretical_loss": 3.458904251654681,
      "tokens_seen": 1812725760
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024952477775676187,
      "loss": 2.6654,
      "theoretical_loss": 3.458893622774286,
      "tokens_seen": 1812791296
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002495082277283904,
      "loss": 2.504,
      "theoretical_loss": 3.4588829943857258,
      "tokens_seen": 1812856832
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024949167770001895,
      "loss": 2.8071,
      "theoretical_loss": 3.4588723664889605,
      "tokens_seen": 1812922368
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024947512767164743,
      "loss": 2.5229,
      "theoretical_loss": 3.4588617390839493,
      "tokens_seen": 1812987904
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002494585776432759,
      "loss": 1.6933,
      "theoretical_loss": 3.4588511121706516,
      "tokens_seen": 1813053440
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024944202761490445,
      "loss": 2.6253,
      "theoretical_loss": 3.458840485749027,
      "tokens_seen": 1813118976
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000249425477586533,
      "loss": 2.4837,
      "theoretical_loss": 3.4588298598190352,
      "tokens_seen": 1813184512
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002494089275581615,
      "loss": 2.6291,
      "theoretical_loss": 3.458819234380635,
      "tokens_seen": 1813250048
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024939237752979,
      "loss": 2.4695,
      "theoretical_loss": 3.4588086094337864,
      "tokens_seen": 1813315584
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024937582750141855,
      "loss": 2.3948,
      "theoretical_loss": 3.4587979849784487,
      "tokens_seen": 1813381120
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002493592774730471,
      "loss": 2.5838,
      "theoretical_loss": 3.458787361014582,
      "tokens_seen": 1813446656
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002493427274446756,
      "loss": 2.7792,
      "theoretical_loss": 3.4587767375421445,
      "tokens_seen": 1813512192
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002493261774163041,
      "loss": 2.6344,
      "theoretical_loss": 3.458766114561097,
      "tokens_seen": 1813577728
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024930962738793264,
      "loss": 2.5094,
      "theoretical_loss": 3.458755492071398,
      "tokens_seen": 1813643264
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002492930773595612,
      "loss": 2.6809,
      "theoretical_loss": 3.458744870073008,
      "tokens_seen": 1813708800
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002492765273311897,
      "loss": 2.3112,
      "theoretical_loss": 3.458734248565886,
      "tokens_seen": 1813774336
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002492599773028182,
      "loss": 2.6462,
      "theoretical_loss": 3.4587236275499915,
      "tokens_seen": 1813839872
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024924342727444674,
      "loss": 2.6217,
      "theoretical_loss": 3.458713007025284,
      "tokens_seen": 1813905408
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002492268772460753,
      "loss": 2.3254,
      "theoretical_loss": 3.4587023869917237,
      "tokens_seen": 1813970944
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002492103272177038,
      "loss": 2.5199,
      "theoretical_loss": 3.458691767449269,
      "tokens_seen": 1814036480
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024919377718933235,
      "loss": 2.883,
      "theoretical_loss": 3.45868114839788,
      "tokens_seen": 1814102016
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024917722716096083,
      "loss": 2.3247,
      "theoretical_loss": 3.4586705298375167,
      "tokens_seen": 1814167552
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024916067713258937,
      "loss": 2.4532,
      "theoretical_loss": 3.4586599117681383,
      "tokens_seen": 1814233088
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002491441271042179,
      "loss": 2.8494,
      "theoretical_loss": 3.458649294189704,
      "tokens_seen": 1814298624
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024912757707584645,
      "loss": 2.4695,
      "theoretical_loss": 3.458638677102174,
      "tokens_seen": 1814364160
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024911102704747493,
      "loss": 2.4415,
      "theoretical_loss": 3.4586280605055073,
      "tokens_seen": 1814429696
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024909447701910347,
      "loss": 2.7008,
      "theoretical_loss": 3.458617444399664,
      "tokens_seen": 1814495232
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000249077926990732,
      "loss": 2.6431,
      "theoretical_loss": 3.4586068287846032,
      "tokens_seen": 1814560768
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024906137696236054,
      "loss": 2.553,
      "theoretical_loss": 3.458596213660285,
      "tokens_seen": 1814626304
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000249044826933989,
      "loss": 2.6965,
      "theoretical_loss": 3.4585855990266685,
      "tokens_seen": 1814691840
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024902827690561756,
      "loss": 2.4703,
      "theoretical_loss": 3.4585749848837133,
      "tokens_seen": 1814757376
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002490117268772461,
      "loss": 2.3364,
      "theoretical_loss": 3.45856437123138,
      "tokens_seen": 1814822912
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024899517684887464,
      "loss": 2.7346,
      "theoretical_loss": 3.4585537580696264,
      "tokens_seen": 1814888448
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002489786268205031,
      "loss": 2.5639,
      "theoretical_loss": 3.458543145398414,
      "tokens_seen": 1814953984
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002489620767921316,
      "loss": 2.588,
      "theoretical_loss": 3.458532533217701,
      "tokens_seen": 1815019520
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024894552676376014,
      "loss": 2.1313,
      "theoretical_loss": 3.4585219215274474,
      "tokens_seen": 1815085056
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002489289767353887,
      "loss": 2.5517,
      "theoretical_loss": 3.458511310327613,
      "tokens_seen": 1815150592
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002489124267070172,
      "loss": 2.8248,
      "theoretical_loss": 3.458500699618158,
      "tokens_seen": 1815216128
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002488958766786457,
      "loss": 2.4255,
      "theoretical_loss": 3.458490089399041,
      "tokens_seen": 1815281664
    },
    {
      "epoch": 0.65,
      "objective/train/avg_token_score": 0.010590584948658943,
      "objective/train/avg_weight": 0.9875693917274475,
      "objective/train/docs_used": 1026340,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.005767583847046,
      "objective/train/std_weight": 0.0732852891087532,
      "objective/train/theoretical_loss": 3.4584794796702223,
      "objective/train/tokens_used": 1835807200,
      "theoretical_loss": 3.4584794796702223,
      "tokens_seen": 1815347200
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024887932665027424,
      "loss": 2.852,
      "theoretical_loss": 3.4584794796702223,
      "tokens_seen": 1815347200
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002488627766219028,
      "loss": 2.6792,
      "theoretical_loss": 3.458468870431661,
      "tokens_seen": 1815412736
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002488462265935313,
      "loss": 1.4876,
      "theoretical_loss": 3.4584582616833175,
      "tokens_seen": 1815478272
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002488296765651598,
      "loss": 2.5829,
      "theoretical_loss": 3.458447653425151,
      "tokens_seen": 1815543808
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024881312653678834,
      "loss": 2.6042,
      "theoretical_loss": 3.4584370456571207,
      "tokens_seen": 1815609344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002487965765084169,
      "loss": 2.5526,
      "theoretical_loss": 3.4584264383791874,
      "tokens_seen": 1815674880
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002487800264800454,
      "loss": 2.5114,
      "theoretical_loss": 3.4584158315913096,
      "tokens_seen": 1815740416
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002487634764516739,
      "loss": 2.2746,
      "theoretical_loss": 3.458405225293448,
      "tokens_seen": 1815805952
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024874692642330243,
      "loss": 2.4082,
      "theoretical_loss": 3.458394619485561,
      "tokens_seen": 1815871488
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024873037639493097,
      "loss": 2.7751,
      "theoretical_loss": 3.4583840141676094,
      "tokens_seen": 1815937024
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002487138263665595,
      "loss": 2.7306,
      "theoretical_loss": 3.458373409339553,
      "tokens_seen": 1816002560
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000248697276338188,
      "loss": 2.0095,
      "theoretical_loss": 3.4583628050013506,
      "tokens_seen": 1816068096
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024868072630981653,
      "loss": 2.6999,
      "theoretical_loss": 3.4583522011529624,
      "tokens_seen": 1816133632
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024866417628144506,
      "loss": 2.5068,
      "theoretical_loss": 3.458341597794348,
      "tokens_seen": 1816199168
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002486476262530736,
      "loss": 2.2464,
      "theoretical_loss": 3.4583309949254675,
      "tokens_seen": 1816264704
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002486310762247021,
      "loss": 2.3653,
      "theoretical_loss": 3.4583203925462795,
      "tokens_seen": 1816330240
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002486145261963306,
      "loss": 2.648,
      "theoretical_loss": 3.4583097906567453,
      "tokens_seen": 1816395776
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024859797616795916,
      "loss": 2.6236,
      "theoretical_loss": 3.458299189256823,
      "tokens_seen": 1816461312
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002485814261395877,
      "loss": 2.4558,
      "theoretical_loss": 3.458288588346474,
      "tokens_seen": 1816526848
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002485648761112162,
      "loss": 2.8523,
      "theoretical_loss": 3.4582779879256567,
      "tokens_seen": 1816592384
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002485483260828447,
      "loss": 2.5732,
      "theoretical_loss": 3.458267387994331,
      "tokens_seen": 1816657920
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024853177605447326,
      "loss": 2.6376,
      "theoretical_loss": 3.458256788552457,
      "tokens_seen": 1816723456
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002485152260261018,
      "loss": 2.693,
      "theoretical_loss": 3.4582461895999947,
      "tokens_seen": 1816788992
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002484986759977303,
      "loss": 2.3385,
      "theoretical_loss": 3.458235591136903,
      "tokens_seen": 1816854528
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024848212596935876,
      "loss": 2.3721,
      "theoretical_loss": 3.458224993163143,
      "tokens_seen": 1816920064
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002484655759409873,
      "loss": 2.4009,
      "theoretical_loss": 3.458214395678673,
      "tokens_seen": 1816985600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024844902591261584,
      "loss": 2.9067,
      "theoretical_loss": 3.4582037986834537,
      "tokens_seen": 1817051136
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002484324758842444,
      "loss": 2.6655,
      "theoretical_loss": 3.458193202177444,
      "tokens_seen": 1817116672
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024841592585587286,
      "loss": 2.4641,
      "theoretical_loss": 3.4581826061606047,
      "tokens_seen": 1817182208
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002483993758275014,
      "loss": 2.3763,
      "theoretical_loss": 3.458172010632895,
      "tokens_seen": 1817247744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024838282579912993,
      "loss": 2.6966,
      "theoretical_loss": 3.458161415594275,
      "tokens_seen": 1817313280
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024836627577075847,
      "loss": 2.5491,
      "theoretical_loss": 3.458150821044704,
      "tokens_seen": 1817378816
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024834972574238695,
      "loss": 2.5226,
      "theoretical_loss": 3.458140226984142,
      "tokens_seen": 1817444352
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002483331757140155,
      "loss": 2.5069,
      "theoretical_loss": 3.458129633412549,
      "tokens_seen": 1817509888
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024831662568564403,
      "loss": 2.2811,
      "theoretical_loss": 3.4581190403298847,
      "tokens_seen": 1817575424
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024830007565727257,
      "loss": 2.908,
      "theoretical_loss": 3.458108447736109,
      "tokens_seen": 1817640960
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024828352562890105,
      "loss": 2.5635,
      "theoretical_loss": 3.4580978556311814,
      "tokens_seen": 1817706496
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002482669756005296,
      "loss": 2.6194,
      "theoretical_loss": 3.458087264015062,
      "tokens_seen": 1817772032
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002482504255721581,
      "loss": 2.6493,
      "theoretical_loss": 3.45807667288771,
      "tokens_seen": 1817837568
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024823387554378666,
      "loss": 2.4419,
      "theoretical_loss": 3.4580660822490863,
      "tokens_seen": 1817903104
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024821732551541514,
      "loss": 4.2022,
      "theoretical_loss": 3.45805549209915,
      "tokens_seen": 1817968640
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002482007754870437,
      "loss": 2.5699,
      "theoretical_loss": 3.4580449024378614,
      "tokens_seen": 1818034176
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002481842254586722,
      "loss": 2.49,
      "theoretical_loss": 3.4580343132651796,
      "tokens_seen": 1818099712
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024816767543030076,
      "loss": 2.6027,
      "theoretical_loss": 3.458023724581065,
      "tokens_seen": 1818165248
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024815112540192924,
      "loss": 2.7058,
      "theoretical_loss": 3.4580131363854782,
      "tokens_seen": 1818230784
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002481345753735578,
      "loss": 2.6843,
      "theoretical_loss": 3.458002548678377,
      "tokens_seen": 1818296320
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002481180253451863,
      "loss": 2.6262,
      "theoretical_loss": 3.457991961459723,
      "tokens_seen": 1818361856
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024810147531681485,
      "loss": 2.6035,
      "theoretical_loss": 3.4579813747294756,
      "tokens_seen": 1818427392
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024808492528844334,
      "loss": 2.7498,
      "theoretical_loss": 3.4579707884875948,
      "tokens_seen": 1818492928
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002480683752600719,
      "loss": 2.8681,
      "theoretical_loss": 3.4579602027340393,
      "tokens_seen": 1818558464
    },
    {
      "epoch": 0.65,
      "objective/train/avg_token_score": 0.008346556685864925,
      "objective/train/avg_weight": 0.9901807308197021,
      "objective/train/docs_used": 1027487,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.096323013305664,
      "objective/train/std_weight": 0.048452503979206085,
      "objective/train/theoretical_loss": 3.457949617468771,
      "objective/train/tokens_used": 1839084000,
      "theoretical_loss": 3.457949617468771,
      "tokens_seen": 1818624000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002480518252317004,
      "loss": 2.3389,
      "theoretical_loss": 3.457949617468771,
      "tokens_seen": 1818624000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024803527520332895,
      "loss": 2.6432,
      "theoretical_loss": 3.457939032691748,
      "tokens_seen": 1818689536
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024801872517495743,
      "loss": 2.5614,
      "theoretical_loss": 3.4579284484029316,
      "tokens_seen": 1818755072
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024800217514658597,
      "loss": 2.4705,
      "theoretical_loss": 3.45791786460228,
      "tokens_seen": 1818820608
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024798562511821445,
      "loss": 2.8665,
      "theoretical_loss": 3.4579072812897547,
      "tokens_seen": 1818886144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000247969075089843,
      "loss": 2.6779,
      "theoretical_loss": 3.457896698465315,
      "tokens_seen": 1818951680
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024795252506147153,
      "loss": 2.2242,
      "theoretical_loss": 3.457886116128921,
      "tokens_seen": 1819017216
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024793597503310007,
      "loss": 2.7332,
      "theoretical_loss": 3.4578755342805323,
      "tokens_seen": 1819082752
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024791942500472855,
      "loss": 2.7013,
      "theoretical_loss": 3.457864952920109,
      "tokens_seen": 1819148288
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002479028749763571,
      "loss": 2.4632,
      "theoretical_loss": 3.4578543720476103,
      "tokens_seen": 1819213824
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002478863249479856,
      "loss": 2.7137,
      "theoretical_loss": 3.4578437916629974,
      "tokens_seen": 1819279360
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024786977491961416,
      "loss": 2.5555,
      "theoretical_loss": 3.4578332117662294,
      "tokens_seen": 1819344896
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024785322489124265,
      "loss": 2.7501,
      "theoretical_loss": 3.4578226323572663,
      "tokens_seen": 1819410432
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002478366748628712,
      "loss": 2.7158,
      "theoretical_loss": 3.4578120534360686,
      "tokens_seen": 1819475968
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002478201248344997,
      "loss": 2.6753,
      "theoretical_loss": 3.4578014750025954,
      "tokens_seen": 1819541504
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024780357480612826,
      "loss": 2.6598,
      "theoretical_loss": 3.457790897056807,
      "tokens_seen": 1819607040
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024778702477775674,
      "loss": 2.4788,
      "theoretical_loss": 3.4577803195986636,
      "tokens_seen": 1819672576
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002477704747493853,
      "loss": 2.5285,
      "theoretical_loss": 3.457769742628125,
      "tokens_seen": 1819738112
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002477539247210138,
      "loss": 2.6148,
      "theoretical_loss": 3.4577591661451517,
      "tokens_seen": 1819803648
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024773737469264235,
      "loss": 2.7578,
      "theoretical_loss": 3.457748590149703,
      "tokens_seen": 1819869184
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024772082466427084,
      "loss": 2.5252,
      "theoretical_loss": 3.457738014641738,
      "tokens_seen": 1819934720
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002477042746358994,
      "loss": 2.6293,
      "theoretical_loss": 3.4577274396212183,
      "tokens_seen": 1820000256
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002476877246075279,
      "loss": 2.6069,
      "theoretical_loss": 3.457716865088103,
      "tokens_seen": 1820065792
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024767117457915645,
      "loss": 2.7889,
      "theoretical_loss": 3.457706291042353,
      "tokens_seen": 1820131328
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024765462455078493,
      "loss": 2.3938,
      "theoretical_loss": 3.4576957174839267,
      "tokens_seen": 1820196864
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024763807452241347,
      "loss": 2.7724,
      "theoretical_loss": 3.4576851444127854,
      "tokens_seen": 1820262400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000247621524494042,
      "loss": 2.6925,
      "theoretical_loss": 3.457674571828889,
      "tokens_seen": 1820327936
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024760497446567055,
      "loss": 2.6784,
      "theoretical_loss": 3.457663999732197,
      "tokens_seen": 1820393472
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024758842443729903,
      "loss": 2.6251,
      "theoretical_loss": 3.4576534281226694,
      "tokens_seen": 1820459008
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024757187440892757,
      "loss": 2.3159,
      "theoretical_loss": 3.457642857000267,
      "tokens_seen": 1820524544
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002475553243805561,
      "loss": 2.744,
      "theoretical_loss": 3.457632286364949,
      "tokens_seen": 1820590080
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024753877435218464,
      "loss": 2.6233,
      "theoretical_loss": 3.4576217162166754,
      "tokens_seen": 1820655616
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002475222243238131,
      "loss": 2.7363,
      "theoretical_loss": 3.457611146555407,
      "tokens_seen": 1820721152
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002475056742954416,
      "loss": 2.1088,
      "theoretical_loss": 3.457600577381103,
      "tokens_seen": 1820786688
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024748912426707015,
      "loss": 2.7069,
      "theoretical_loss": 3.4575900086937237,
      "tokens_seen": 1820852224
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002474725742386987,
      "loss": 2.6333,
      "theoretical_loss": 3.4575794404932294,
      "tokens_seen": 1820917760
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002474560242103272,
      "loss": 2.4987,
      "theoretical_loss": 3.4575688727795804,
      "tokens_seen": 1820983296
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002474394741819557,
      "loss": 2.6123,
      "theoretical_loss": 3.4575583055527357,
      "tokens_seen": 1821048832
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024742292415358424,
      "loss": 2.5484,
      "theoretical_loss": 3.4575477388126563,
      "tokens_seen": 1821114368
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002474063741252128,
      "loss": 2.4926,
      "theoretical_loss": 3.457537172559302,
      "tokens_seen": 1821179904
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002473898240968413,
      "loss": 2.6037,
      "theoretical_loss": 3.4575266067926327,
      "tokens_seen": 1821245440
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002473732740684698,
      "loss": 2.5628,
      "theoretical_loss": 3.4575160415126085,
      "tokens_seen": 1821310976
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024735672404009834,
      "loss": 2.7177,
      "theoretical_loss": 3.4575054767191897,
      "tokens_seen": 1821376512
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002473401740117269,
      "loss": 2.6583,
      "theoretical_loss": 3.4574949124123364,
      "tokens_seen": 1821442048
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002473236239833554,
      "loss": 2.6545,
      "theoretical_loss": 3.457484348592008,
      "tokens_seen": 1821507584
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002473070739549839,
      "loss": 2.6639,
      "theoretical_loss": 3.4574737852581654,
      "tokens_seen": 1821573120
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024729052392661243,
      "loss": 2.4909,
      "theoretical_loss": 3.4574632224107686,
      "tokens_seen": 1821638656
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024727397389824097,
      "loss": 2.8988,
      "theoretical_loss": 3.4574526600497775,
      "tokens_seen": 1821704192
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002472574238698695,
      "loss": 2.8066,
      "theoretical_loss": 3.457442098175152,
      "tokens_seen": 1821769728
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000247240873841498,
      "loss": 2.6343,
      "theoretical_loss": 3.457431536786853,
      "tokens_seen": 1821835264
    },
    {
      "epoch": 0.65,
      "objective/train/avg_token_score": 0.030278533697128296,
      "objective/train/avg_weight": 0.9643910527229309,
      "objective/train/docs_used": 1029511,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.366235733032227,
      "objective/train/std_weight": 0.15441840887069702,
      "objective/train/theoretical_loss": 3.4574209758848395,
      "objective/train/tokens_used": 1842360800,
      "theoretical_loss": 3.4574209758848395,
      "tokens_seen": 1821900800
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024722432381312653,
      "loss": 2.3996,
      "theoretical_loss": 3.4574209758848395,
      "tokens_seen": 1821900800
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024720777378475507,
      "loss": 2.5397,
      "theoretical_loss": 3.457410415469072,
      "tokens_seen": 1821966336
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002471912237563836,
      "loss": 2.5912,
      "theoretical_loss": 3.4573998555395113,
      "tokens_seen": 1822031872
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002471746737280121,
      "loss": 2.6615,
      "theoretical_loss": 3.4573892960961166,
      "tokens_seen": 1822097408
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002471581236996406,
      "loss": 2.4042,
      "theoretical_loss": 3.457378737138849,
      "tokens_seen": 1822162944
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024714157367126916,
      "loss": 2.5632,
      "theoretical_loss": 3.4573681786676675,
      "tokens_seen": 1822228480
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002471250236428977,
      "loss": 2.5917,
      "theoretical_loss": 3.457357620682533,
      "tokens_seen": 1822294016
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002471084736145262,
      "loss": 2.7398,
      "theoretical_loss": 3.4573470631834056,
      "tokens_seen": 1822359552
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002470919235861547,
      "loss": 2.6774,
      "theoretical_loss": 3.457336506170245,
      "tokens_seen": 1822425088
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024707537355778326,
      "loss": 2.6878,
      "theoretical_loss": 3.457325949643012,
      "tokens_seen": 1822490624
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002470588235294118,
      "loss": 2.624,
      "theoretical_loss": 3.4573153936016663,
      "tokens_seen": 1822556160
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002470422735010403,
      "loss": 2.5554,
      "theoretical_loss": 3.4573048380461686,
      "tokens_seen": 1822621696
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002470257234726688,
      "loss": 2.7465,
      "theoretical_loss": 3.4572942829764783,
      "tokens_seen": 1822687232
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002470091734442973,
      "loss": 2.4972,
      "theoretical_loss": 3.4572837283925564,
      "tokens_seen": 1822752768
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024699262341592584,
      "loss": 1.8631,
      "theoretical_loss": 3.457273174294362,
      "tokens_seen": 1822818304
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002469760733875544,
      "loss": 2.6046,
      "theoretical_loss": 3.457262620681856,
      "tokens_seen": 1822883840
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024695952335918286,
      "loss": 2.4491,
      "theoretical_loss": 3.4572520675549985,
      "tokens_seen": 1822949376
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002469429733308114,
      "loss": 2.5521,
      "theoretical_loss": 3.45724151491375,
      "tokens_seen": 1823014912
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024692642330243993,
      "loss": 2.5512,
      "theoretical_loss": 3.4572309627580706,
      "tokens_seen": 1823080448
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024690987327406847,
      "loss": 2.5847,
      "theoretical_loss": 3.4572204110879197,
      "tokens_seen": 1823145984
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024689332324569696,
      "loss": 2.6337,
      "theoretical_loss": 3.4572098599032586,
      "tokens_seen": 1823211520
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002468767732173255,
      "loss": 2.4337,
      "theoretical_loss": 3.4571993092040465,
      "tokens_seen": 1823277056
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024686022318895403,
      "loss": 2.4998,
      "theoretical_loss": 3.4571887589902444,
      "tokens_seen": 1823342592
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024684367316058257,
      "loss": 2.3996,
      "theoretical_loss": 3.457178209261812,
      "tokens_seen": 1823408128
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024682712313221105,
      "loss": 2.5512,
      "theoretical_loss": 3.457167660018711,
      "tokens_seen": 1823473664
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002468105731038396,
      "loss": 2.7676,
      "theoretical_loss": 3.4571571112608988,
      "tokens_seen": 1823539200
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002467940230754681,
      "loss": 2.202,
      "theoretical_loss": 3.4571465629883384,
      "tokens_seen": 1823604736
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024677747304709666,
      "loss": 2.5201,
      "theoretical_loss": 3.4571360152009882,
      "tokens_seen": 1823670272
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024676092301872515,
      "loss": 2.5904,
      "theoretical_loss": 3.4571254678988095,
      "tokens_seen": 1823735808
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002467443729903537,
      "loss": 2.3666,
      "theoretical_loss": 3.4571149210817618,
      "tokens_seen": 1823801344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002467278229619822,
      "loss": 2.7392,
      "theoretical_loss": 3.4571043747498065,
      "tokens_seen": 1823866880
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024671127293361076,
      "loss": 2.4767,
      "theoretical_loss": 3.4570938289029023,
      "tokens_seen": 1823932416
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024669472290523924,
      "loss": 2.65,
      "theoretical_loss": 3.457083283541011,
      "tokens_seen": 1823997952
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002466781728768678,
      "loss": 2.2253,
      "theoretical_loss": 3.4570727386640914,
      "tokens_seen": 1824063488
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002466616228484963,
      "loss": 2.705,
      "theoretical_loss": 3.4570621942721047,
      "tokens_seen": 1824129024
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024664507282012486,
      "loss": 2.3632,
      "theoretical_loss": 3.4570516503650115,
      "tokens_seen": 1824194560
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024662852279175334,
      "loss": 2.5476,
      "theoretical_loss": 3.4570411069427713,
      "tokens_seen": 1824260096
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002466119727633819,
      "loss": 2.347,
      "theoretical_loss": 3.4570305640053443,
      "tokens_seen": 1824325632
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002465954227350104,
      "loss": 2.3877,
      "theoretical_loss": 3.457020021552691,
      "tokens_seen": 1824391168
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024657887270663895,
      "loss": 2.6299,
      "theoretical_loss": 3.4570094795847726,
      "tokens_seen": 1824456704
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024656232267826744,
      "loss": 2.3032,
      "theoretical_loss": 3.456998938101548,
      "tokens_seen": 1824522240
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000246545772649896,
      "loss": 2.4582,
      "theoretical_loss": 3.456988397102979,
      "tokens_seen": 1824587776
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002465292226215245,
      "loss": 2.7377,
      "theoretical_loss": 3.4569778565890243,
      "tokens_seen": 1824653312
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000246512672593153,
      "loss": 2.7293,
      "theoretical_loss": 3.456967316559645,
      "tokens_seen": 1824718848
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024649612256478153,
      "loss": 2.5014,
      "theoretical_loss": 3.4569567770148018,
      "tokens_seen": 1824784384
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024647957253641007,
      "loss": 2.6785,
      "theoretical_loss": 3.4569462379544547,
      "tokens_seen": 1824849920
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024646302250803855,
      "loss": 2.483,
      "theoretical_loss": 3.4569356993785636,
      "tokens_seen": 1824915456
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002464464724796671,
      "loss": 2.4369,
      "theoretical_loss": 3.456925161287089,
      "tokens_seen": 1824980992
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024642992245129563,
      "loss": 2.5386,
      "theoretical_loss": 3.456914623679992,
      "tokens_seen": 1825046528
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024641337242292417,
      "loss": 2.6945,
      "theoretical_loss": 3.4569040865572322,
      "tokens_seen": 1825112064
    },
    {
      "epoch": 0.65,
      "objective/train/avg_token_score": 0.014862995594739914,
      "objective/train/avg_weight": 0.9825058579444885,
      "objective/train/docs_used": 1031319,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3459205627441406,
      "objective/train/std_weight": 0.08411923795938492,
      "objective/train/theoretical_loss": 3.4568935499187696,
      "objective/train/tokens_used": 1845637600,
      "theoretical_loss": 3.4568935499187696,
      "tokens_seen": 1825177600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024639682239455265,
      "loss": 2.7306,
      "theoretical_loss": 3.4568935499187696,
      "tokens_seen": 1825177600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002463802723661812,
      "loss": 2.6172,
      "theoretical_loss": 3.456883013764566,
      "tokens_seen": 1825243136
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002463637223378097,
      "loss": 2.7162,
      "theoretical_loss": 3.4568724780945796,
      "tokens_seen": 1825308672
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024634717230943826,
      "loss": 2.6791,
      "theoretical_loss": 3.456861942908773,
      "tokens_seen": 1825374208
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024633062228106674,
      "loss": 2.8764,
      "theoretical_loss": 3.456851408207105,
      "tokens_seen": 1825439744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002463140722526953,
      "loss": 2.5226,
      "theoretical_loss": 3.456840873989537,
      "tokens_seen": 1825505280
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002462975222243238,
      "loss": 2.6621,
      "theoretical_loss": 3.4568303402560288,
      "tokens_seen": 1825570816
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024628097219595236,
      "loss": 2.2502,
      "theoretical_loss": 3.4568198070065406,
      "tokens_seen": 1825636352
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024626442216758084,
      "loss": 2.6318,
      "theoretical_loss": 3.4568092742410332,
      "tokens_seen": 1825701888
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002462478721392094,
      "loss": 2.6026,
      "theoretical_loss": 3.456798741959467,
      "tokens_seen": 1825767424
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002462313221108379,
      "loss": 2.2159,
      "theoretical_loss": 3.456788210161802,
      "tokens_seen": 1825832960
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024621477208246645,
      "loss": 2.7554,
      "theoretical_loss": 3.456777678847999,
      "tokens_seen": 1825898496
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024619822205409494,
      "loss": 1.9252,
      "theoretical_loss": 3.4567671480180184,
      "tokens_seen": 1825964032
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002461816720257235,
      "loss": 2.3287,
      "theoretical_loss": 3.4567566176718203,
      "tokens_seen": 1826029568
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000246165121997352,
      "loss": 2.6792,
      "theoretical_loss": 3.456746087809365,
      "tokens_seen": 1826095104
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024614857196898055,
      "loss": 2.4104,
      "theoretical_loss": 3.456735558430614,
      "tokens_seen": 1826160640
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024613202194060903,
      "loss": 2.3047,
      "theoretical_loss": 3.4567250295355265,
      "tokens_seen": 1826226176
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024611547191223757,
      "loss": 2.2835,
      "theoretical_loss": 3.456714501124063,
      "tokens_seen": 1826291712
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002460989218838661,
      "loss": 2.4909,
      "theoretical_loss": 3.4567039731961846,
      "tokens_seen": 1826357248
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024608237185549465,
      "loss": 2.5596,
      "theoretical_loss": 3.4566934457518514,
      "tokens_seen": 1826422784
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024606582182712313,
      "loss": 2.5898,
      "theoretical_loss": 3.456682918791024,
      "tokens_seen": 1826488320
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024604927179875167,
      "loss": 2.798,
      "theoretical_loss": 3.4566723923136626,
      "tokens_seen": 1826553856
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002460327217703802,
      "loss": 2.6492,
      "theoretical_loss": 3.4566618663197275,
      "tokens_seen": 1826619392
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002460161717420087,
      "loss": 2.6424,
      "theoretical_loss": 3.45665134080918,
      "tokens_seen": 1826684928
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002459996217136372,
      "loss": 2.5236,
      "theoretical_loss": 3.4566408157819795,
      "tokens_seen": 1826750464
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002459830716852657,
      "loss": 2.7089,
      "theoretical_loss": 3.456630291238087,
      "tokens_seen": 1826816000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024596652165689425,
      "loss": 2.3569,
      "theoretical_loss": 3.4566197671774628,
      "tokens_seen": 1826881536
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002459499716285228,
      "loss": 2.6333,
      "theoretical_loss": 3.456609243600068,
      "tokens_seen": 1826947072
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002459334216001513,
      "loss": 2.4369,
      "theoretical_loss": 3.456598720505862,
      "tokens_seen": 1827012608
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002459168715717798,
      "loss": 2.4907,
      "theoretical_loss": 3.4565881978948063,
      "tokens_seen": 1827078144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024590032154340834,
      "loss": 2.7487,
      "theoretical_loss": 3.4565776757668605,
      "tokens_seen": 1827143680
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002458837715150369,
      "loss": 2.5088,
      "theoretical_loss": 3.456567154121986,
      "tokens_seen": 1827209216
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002458672214866654,
      "loss": 2.7587,
      "theoretical_loss": 3.4565566329601425,
      "tokens_seen": 1827274752
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002458506714582939,
      "loss": 2.6037,
      "theoretical_loss": 3.4565461122812904,
      "tokens_seen": 1827340288
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024583412142992244,
      "loss": 2.5179,
      "theoretical_loss": 3.4565355920853915,
      "tokens_seen": 1827405824
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000245817571401551,
      "loss": 1.9643,
      "theoretical_loss": 3.456525072372405,
      "tokens_seen": 1827471360
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002458010213731795,
      "loss": 2.4644,
      "theoretical_loss": 3.456514553142292,
      "tokens_seen": 1827536896
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000245784471344808,
      "loss": 2.5646,
      "theoretical_loss": 3.456504034395013,
      "tokens_seen": 1827602432
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024576792131643653,
      "loss": 2.5949,
      "theoretical_loss": 3.4564935161305277,
      "tokens_seen": 1827667968
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024575137128806507,
      "loss": 2.6267,
      "theoretical_loss": 3.4564829983487977,
      "tokens_seen": 1827733504
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002457348212596936,
      "loss": 2.7367,
      "theoretical_loss": 3.4564724810497838,
      "tokens_seen": 1827799040
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002457182712313221,
      "loss": 2.622,
      "theoretical_loss": 3.456461964233445,
      "tokens_seen": 1827864576
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024570172120295063,
      "loss": 2.7205,
      "theoretical_loss": 3.4564514478997435,
      "tokens_seen": 1827930112
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024568517117457917,
      "loss": 2.3274,
      "theoretical_loss": 3.456440932048639,
      "tokens_seen": 1827995648
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002456686211462077,
      "loss": 2.8516,
      "theoretical_loss": 3.4564304166800914,
      "tokens_seen": 1828061184
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002456520711178362,
      "loss": 1.9296,
      "theoretical_loss": 3.456419901794063,
      "tokens_seen": 1828126720
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002456355210894647,
      "loss": 2.3926,
      "theoretical_loss": 3.4564093873905124,
      "tokens_seen": 1828192256
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024561897106109326,
      "loss": 2.6101,
      "theoretical_loss": 3.4563988734694018,
      "tokens_seen": 1828257792
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002456024210327218,
      "loss": 2.0501,
      "theoretical_loss": 3.456388360030691,
      "tokens_seen": 1828323328
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002455858710043503,
      "loss": 2.3565,
      "theoretical_loss": 3.4563778470743403,
      "tokens_seen": 1828388864
    },
    {
      "epoch": 0.65,
      "objective/train/avg_token_score": 0.01669861003756523,
      "objective/train/avg_weight": 0.9803463220596313,
      "objective/train/docs_used": 1033212,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4837794303894043,
      "objective/train/std_weight": 0.06467600911855698,
      "objective/train/theoretical_loss": 3.456367334600311,
      "objective/train/tokens_used": 1848914400,
      "theoretical_loss": 3.456367334600311,
      "tokens_seen": 1828454400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002455693209759788,
      "loss": 2.3805,
      "theoretical_loss": 3.456367334600311,
      "tokens_seen": 1828454400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024555277094760736,
      "loss": 2.5029,
      "theoretical_loss": 3.4563568226085635,
      "tokens_seen": 1828519936
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024553622091923584,
      "loss": 2.4992,
      "theoretical_loss": 3.456346311099058,
      "tokens_seen": 1828585472
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002455196708908644,
      "loss": 2.881,
      "theoretical_loss": 3.4563358000717552,
      "tokens_seen": 1828651008
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024550312086249286,
      "loss": 2.7462,
      "theoretical_loss": 3.456325289526616,
      "tokens_seen": 1828716544
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002454865708341214,
      "loss": 2.6298,
      "theoretical_loss": 3.456314779463601,
      "tokens_seen": 1828782080
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024547002080574994,
      "loss": 2.1662,
      "theoretical_loss": 3.4563042698826707,
      "tokens_seen": 1828847616
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002454534707773785,
      "loss": 2.6596,
      "theoretical_loss": 3.456293760783786,
      "tokens_seen": 1828913152
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024543692074900696,
      "loss": 2.5914,
      "theoretical_loss": 3.456283252166906,
      "tokens_seen": 1828978688
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002454203707206355,
      "loss": 1.9877,
      "theoretical_loss": 3.4562727440319936,
      "tokens_seen": 1829044224
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024540382069226403,
      "loss": 2.3613,
      "theoretical_loss": 3.4562622363790076,
      "tokens_seen": 1829109760
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024538727066389257,
      "loss": 2.6037,
      "theoretical_loss": 3.45625172920791,
      "tokens_seen": 1829175296
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024537072063552105,
      "loss": 2.4252,
      "theoretical_loss": 3.45624122251866,
      "tokens_seen": 1829240832
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002453541706071496,
      "loss": 2.6121,
      "theoretical_loss": 3.45623071631122,
      "tokens_seen": 1829306368
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024533762057877813,
      "loss": 2.821,
      "theoretical_loss": 3.456220210585549,
      "tokens_seen": 1829371904
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024532107055040667,
      "loss": 2.4163,
      "theoretical_loss": 3.4562097053416085,
      "tokens_seen": 1829437440
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024530452052203515,
      "loss": 2.1544,
      "theoretical_loss": 3.456199200579359,
      "tokens_seen": 1829502976
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002452879704936637,
      "loss": 2.8484,
      "theoretical_loss": 3.456188696298761,
      "tokens_seen": 1829568512
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002452714204652922,
      "loss": 2.6494,
      "theoretical_loss": 3.4561781924997756,
      "tokens_seen": 1829634048
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024525487043692076,
      "loss": 2.0776,
      "theoretical_loss": 3.456167689182363,
      "tokens_seen": 1829699584
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002452383204085493,
      "loss": 2.7283,
      "theoretical_loss": 3.456157186346484,
      "tokens_seen": 1829765120
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002452217703801778,
      "loss": 2.302,
      "theoretical_loss": 3.4561466839921,
      "tokens_seen": 1829830656
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002452052203518063,
      "loss": 2.4669,
      "theoretical_loss": 3.4561361821191703,
      "tokens_seen": 1829896192
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024518867032343486,
      "loss": 2.4161,
      "theoretical_loss": 3.456125680727656,
      "tokens_seen": 1829961728
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002451721202950634,
      "loss": 1.9379,
      "theoretical_loss": 3.4561151798175187,
      "tokens_seen": 1830027264
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002451555702666919,
      "loss": 2.6001,
      "theoretical_loss": 3.4561046793887185,
      "tokens_seen": 1830092800
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002451390202383204,
      "loss": 2.4861,
      "theoretical_loss": 3.456094179441216,
      "tokens_seen": 1830158336
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024512247020994896,
      "loss": 2.5876,
      "theoretical_loss": 3.456083679974972,
      "tokens_seen": 1830223872
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002451059201815775,
      "loss": 2.5704,
      "theoretical_loss": 3.456073180989947,
      "tokens_seen": 1830289408
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000245089370153206,
      "loss": 2.6077,
      "theoretical_loss": 3.456062682486102,
      "tokens_seen": 1830354944
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002450728201248345,
      "loss": 2.2588,
      "theoretical_loss": 3.4560521844633976,
      "tokens_seen": 1830420480
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024505627009646305,
      "loss": 2.7099,
      "theoretical_loss": 3.456041686921795,
      "tokens_seen": 1830486016
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024503972006809153,
      "loss": 2.4436,
      "theoretical_loss": 3.456031189861254,
      "tokens_seen": 1830551552
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024502317003972007,
      "loss": 2.56,
      "theoretical_loss": 3.4560206932817357,
      "tokens_seen": 1830617088
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024500662001134856,
      "loss": 2.6756,
      "theoretical_loss": 3.4560101971832013,
      "tokens_seen": 1830682624
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002449900699829771,
      "loss": 2.5735,
      "theoretical_loss": 3.455999701565611,
      "tokens_seen": 1830748160
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024497351995460563,
      "loss": 2.5551,
      "theoretical_loss": 3.455989206428926,
      "tokens_seen": 1830813696
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024495696992623417,
      "loss": 2.6951,
      "theoretical_loss": 3.455978711773107,
      "tokens_seen": 1830879232
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024494041989786265,
      "loss": 2.7878,
      "theoretical_loss": 3.455968217598114,
      "tokens_seen": 1830944768
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002449238698694912,
      "loss": 2.4391,
      "theoretical_loss": 3.455957723903909,
      "tokens_seen": 1831010304
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002449073198411197,
      "loss": 2.7674,
      "theoretical_loss": 3.4559472306904517,
      "tokens_seen": 1831075840
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024489076981274826,
      "loss": 2.5053,
      "theoretical_loss": 3.4559367379577033,
      "tokens_seen": 1831141376
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024487421978437675,
      "loss": 2.5665,
      "theoretical_loss": 3.4559262457056246,
      "tokens_seen": 1831206912
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002448576697560053,
      "loss": 2.3922,
      "theoretical_loss": 3.4559157539341765,
      "tokens_seen": 1831272448
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002448411197276338,
      "loss": 2.7035,
      "theoretical_loss": 3.455905262643319,
      "tokens_seen": 1831337984
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024482456969926236,
      "loss": 2.4815,
      "theoretical_loss": 3.4558947718330137,
      "tokens_seen": 1831403520
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024480801967089084,
      "loss": 2.266,
      "theoretical_loss": 3.4558842815032214,
      "tokens_seen": 1831469056
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002447914696425194,
      "loss": 2.7164,
      "theoretical_loss": 3.455873791653903,
      "tokens_seen": 1831534592
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002447749196141479,
      "loss": 2.5621,
      "theoretical_loss": 3.4558633022850183,
      "tokens_seen": 1831600128
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024475836958577646,
      "loss": 2.4154,
      "theoretical_loss": 3.455852813396529,
      "tokens_seen": 1831665664
    },
    {
      "epoch": 0.65,
      "objective/train/avg_token_score": 0.009757228195667267,
      "objective/train/avg_weight": 0.9885177612304688,
      "objective/train/docs_used": 1034888,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8865416049957275,
      "objective/train/std_weight": 0.05738809332251549,
      "objective/train/theoretical_loss": 3.455842324988396,
      "objective/train/tokens_used": 1852191200,
      "theoretical_loss": 3.455842324988396,
      "tokens_seen": 1831731200
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024474181955740494,
      "loss": 2.3733,
      "theoretical_loss": 3.455842324988396,
      "tokens_seen": 1831731200
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002447252695290335,
      "loss": 2.641,
      "theoretical_loss": 3.455831837060579,
      "tokens_seen": 1831796736
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000244708719500662,
      "loss": 2.6076,
      "theoretical_loss": 3.45582134961304,
      "tokens_seen": 1831862272
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024469216947229055,
      "loss": 2.6141,
      "theoretical_loss": 3.4558108626457402,
      "tokens_seen": 1831927808
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024467561944391904,
      "loss": 2.6941,
      "theoretical_loss": 3.455800376158639,
      "tokens_seen": 1831993344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024465906941554757,
      "loss": 2.6352,
      "theoretical_loss": 3.4557898901516984,
      "tokens_seen": 1832058880
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002446425193871761,
      "loss": 2.7607,
      "theoretical_loss": 3.455779404624878,
      "tokens_seen": 1832124416
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024462596935880465,
      "loss": 2.6454,
      "theoretical_loss": 3.45576891957814,
      "tokens_seen": 1832189952
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024460941933043313,
      "loss": 2.7329,
      "theoretical_loss": 3.4557584350114445,
      "tokens_seen": 1832255488
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024459286930206167,
      "loss": 2.6005,
      "theoretical_loss": 3.4557479509247524,
      "tokens_seen": 1832321024
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002445763192736902,
      "loss": 2.8564,
      "theoretical_loss": 3.455737467318025,
      "tokens_seen": 1832386560
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024455976924531874,
      "loss": 2.6812,
      "theoretical_loss": 3.4557269841912226,
      "tokens_seen": 1832452096
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024454321921694723,
      "loss": 2.6392,
      "theoretical_loss": 3.4557165015443063,
      "tokens_seen": 1832517632
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002445266691885757,
      "loss": 2.5419,
      "theoretical_loss": 3.455706019377237,
      "tokens_seen": 1832583168
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024451011916020425,
      "loss": 2.5299,
      "theoretical_loss": 3.4556955376899756,
      "tokens_seen": 1832648704
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002444935691318328,
      "loss": 2.7234,
      "theoretical_loss": 3.4556850564824826,
      "tokens_seen": 1832714240
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002444770191034613,
      "loss": 2.5577,
      "theoretical_loss": 3.4556745757547196,
      "tokens_seen": 1832779776
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002444604690750898,
      "loss": 2.6118,
      "theoretical_loss": 3.455664095506647,
      "tokens_seen": 1832845312
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024444391904671834,
      "loss": 2.6138,
      "theoretical_loss": 3.4556536157382256,
      "tokens_seen": 1832910848
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002444273690183469,
      "loss": 2.7038,
      "theoretical_loss": 3.455643136449417,
      "tokens_seen": 1832976384
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002444108189899754,
      "loss": 2.4118,
      "theoretical_loss": 3.455632657640181,
      "tokens_seen": 1833041920
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002443942689616039,
      "loss": 2.7661,
      "theoretical_loss": 3.4556221793104793,
      "tokens_seen": 1833107456
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024437771893323244,
      "loss": 2.5256,
      "theoretical_loss": 3.4556117014602727,
      "tokens_seen": 1833172992
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000244361168904861,
      "loss": 2.5133,
      "theoretical_loss": 3.455601224089522,
      "tokens_seen": 1833238528
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002443446188764895,
      "loss": 2.9304,
      "theoretical_loss": 3.4555907471981886,
      "tokens_seen": 1833304064
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000244328068848118,
      "loss": 2.4221,
      "theoretical_loss": 3.455580270786232,
      "tokens_seen": 1833369600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024431151881974654,
      "loss": 2.4978,
      "theoretical_loss": 3.455569794853615,
      "tokens_seen": 1833435136
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002442949687913751,
      "loss": 2.5116,
      "theoretical_loss": 3.4555593194002974,
      "tokens_seen": 1833500672
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002442784187630036,
      "loss": 2.3898,
      "theoretical_loss": 3.4555488444262403,
      "tokens_seen": 1833566208
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002442618687346321,
      "loss": 2.5487,
      "theoretical_loss": 3.4555383699314044,
      "tokens_seen": 1833631744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024424531870626063,
      "loss": 2.2363,
      "theoretical_loss": 3.4555278959157514,
      "tokens_seen": 1833697280
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024422876867788917,
      "loss": 2.4602,
      "theoretical_loss": 3.455517422379242,
      "tokens_seen": 1833762816
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002442122186495177,
      "loss": 2.4643,
      "theoretical_loss": 3.455506949321837,
      "tokens_seen": 1833828352
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002441956686211462,
      "loss": 2.2317,
      "theoretical_loss": 3.455496476743497,
      "tokens_seen": 1833893888
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00024417911859277473,
      "loss": 2.8781,
      "theoretical_loss": 3.4554860046441833,
      "tokens_seen": 1833959424
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024416256856440327,
      "loss": 2.5157,
      "theoretical_loss": 3.455475533023857,
      "tokens_seen": 1834024960
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002441460185360318,
      "loss": 2.3307,
      "theoretical_loss": 3.455465061882479,
      "tokens_seen": 1834090496
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002441294685076603,
      "loss": 2.7688,
      "theoretical_loss": 3.4554545912200103,
      "tokens_seen": 1834156032
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024411291847928882,
      "loss": 2.4524,
      "theoretical_loss": 3.455444121036412,
      "tokens_seen": 1834221568
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024409636845091736,
      "loss": 2.6433,
      "theoretical_loss": 3.455433651331645,
      "tokens_seen": 1834287104
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024407981842254587,
      "loss": 2.4632,
      "theoretical_loss": 3.45542318210567,
      "tokens_seen": 1834352640
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024406326839417436,
      "loss": 2.5024,
      "theoretical_loss": 3.455412713358448,
      "tokens_seen": 1834418176
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002440467183658029,
      "loss": 2.3583,
      "theoretical_loss": 3.45540224508994,
      "tokens_seen": 1834483712
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002440301683374314,
      "loss": 2.327,
      "theoretical_loss": 3.455391777300108,
      "tokens_seen": 1834549248
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024401361830905994,
      "loss": 2.609,
      "theoretical_loss": 3.455381309988912,
      "tokens_seen": 1834614784
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024399706828068848,
      "loss": 2.6627,
      "theoretical_loss": 3.455370843156313,
      "tokens_seen": 1834680320
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000243980518252317,
      "loss": 2.4567,
      "theoretical_loss": 3.4553603768022727,
      "tokens_seen": 1834745856
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024396396822394553,
      "loss": 2.6844,
      "theoretical_loss": 3.4553499109267514,
      "tokens_seen": 1834811392
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024394741819557404,
      "loss": 2.5977,
      "theoretical_loss": 3.4553394455297104,
      "tokens_seen": 1834876928
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024393086816720257,
      "loss": 2.7309,
      "theoretical_loss": 3.455328980611111,
      "tokens_seen": 1834942464
    },
    {
      "epoch": 0.66,
      "objective/train/avg_token_score": 0.016779616475105286,
      "objective/train/avg_weight": 0.9802452325820923,
      "objective/train/docs_used": 1036847,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.6809425354003906,
      "objective/train/std_weight": 0.10503675043582916,
      "objective/train/theoretical_loss": 3.455318516170914,
      "objective/train/tokens_used": 1855468000,
      "theoretical_loss": 3.455318516170914,
      "tokens_seen": 1835008000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024391431813883108,
      "loss": 2.3895,
      "theoretical_loss": 3.455318516170914,
      "tokens_seen": 1835008000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024389776811045962,
      "loss": 2.4987,
      "theoretical_loss": 3.4553080522090807,
      "tokens_seen": 1835073536
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024388121808208813,
      "loss": 2.3958,
      "theoretical_loss": 3.4552975887255717,
      "tokens_seen": 1835139072
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024386466805371667,
      "loss": 2.3819,
      "theoretical_loss": 3.455287125720348,
      "tokens_seen": 1835204608
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024384811802534518,
      "loss": 2.5747,
      "theoretical_loss": 3.455276663193372,
      "tokens_seen": 1835270144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024383156799697372,
      "loss": 2.3148,
      "theoretical_loss": 3.4552662011446023,
      "tokens_seen": 1835335680
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024381501796860223,
      "loss": 2.5587,
      "theoretical_loss": 3.4552557395740022,
      "tokens_seen": 1835401216
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024379846794023077,
      "loss": 2.4803,
      "theoretical_loss": 3.4552452784815317,
      "tokens_seen": 1835466752
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024378191791185928,
      "loss": 2.4378,
      "theoretical_loss": 3.455234817867152,
      "tokens_seen": 1835532288
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024376536788348781,
      "loss": 2.718,
      "theoretical_loss": 3.455224357730825,
      "tokens_seen": 1835597824
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024374881785511632,
      "loss": 2.5264,
      "theoretical_loss": 3.4552138980725102,
      "tokens_seen": 1835663360
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024373226782674486,
      "loss": 2.4155,
      "theoretical_loss": 3.4552034388921697,
      "tokens_seen": 1835728896
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024371571779837337,
      "loss": 2.562,
      "theoretical_loss": 3.4551929801897647,
      "tokens_seen": 1835794432
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002436991677700019,
      "loss": 2.6587,
      "theoretical_loss": 3.455182521965256,
      "tokens_seen": 1835859968
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024368261774163042,
      "loss": 2.4057,
      "theoretical_loss": 3.455172064218605,
      "tokens_seen": 1835925504
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024366606771325896,
      "loss": 1.8401,
      "theoretical_loss": 3.4551616069497726,
      "tokens_seen": 1835991040
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024364951768488747,
      "loss": 2.5163,
      "theoretical_loss": 3.4551511501587195,
      "tokens_seen": 1836056576
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000243632967656516,
      "loss": 2.5633,
      "theoretical_loss": 3.4551406938454075,
      "tokens_seen": 1836122112
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024361641762814452,
      "loss": 2.6642,
      "theoretical_loss": 3.4551302380097972,
      "tokens_seen": 1836187648
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024359986759977305,
      "loss": 2.3709,
      "theoretical_loss": 3.4551197826518507,
      "tokens_seen": 1836253184
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024358331757140156,
      "loss": 2.3712,
      "theoretical_loss": 3.4551093277715275,
      "tokens_seen": 1836318720
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024356676754303005,
      "loss": 2.7893,
      "theoretical_loss": 3.45509887336879,
      "tokens_seen": 1836384256
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024355021751465859,
      "loss": 2.5839,
      "theoretical_loss": 3.455088419443599,
      "tokens_seen": 1836449792
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002435336674862871,
      "loss": 2.5349,
      "theoretical_loss": 3.4550779659959154,
      "tokens_seen": 1836515328
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024351711745791563,
      "loss": 2.3712,
      "theoretical_loss": 3.455067513025701,
      "tokens_seen": 1836580864
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024350056742954414,
      "loss": 2.4875,
      "theoretical_loss": 3.4550570605329165,
      "tokens_seen": 1836646400
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024348401740117268,
      "loss": 2.6758,
      "theoretical_loss": 3.455046608517523,
      "tokens_seen": 1836711936
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002434674673728012,
      "loss": 2.7541,
      "theoretical_loss": 3.4550361569794816,
      "tokens_seen": 1836777472
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024345091734442973,
      "loss": 2.6483,
      "theoretical_loss": 3.4550257059187537,
      "tokens_seen": 1836843008
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024343436731605824,
      "loss": 2.6374,
      "theoretical_loss": 3.4550152553353,
      "tokens_seen": 1836908544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024341781728768678,
      "loss": 2.6169,
      "theoretical_loss": 3.4550048052290823,
      "tokens_seen": 1836974080
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002434012672593153,
      "loss": 2.653,
      "theoretical_loss": 3.454994355600062,
      "tokens_seen": 1837039616
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024338471723094383,
      "loss": 2.5064,
      "theoretical_loss": 3.4549839064481995,
      "tokens_seen": 1837105152
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024336816720257234,
      "loss": 2.5053,
      "theoretical_loss": 3.4549734577734563,
      "tokens_seen": 1837170688
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024335161717420087,
      "loss": 2.6008,
      "theoretical_loss": 3.4549630095757937,
      "tokens_seen": 1837236224
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024333506714582938,
      "loss": 2.5282,
      "theoretical_loss": 3.4549525618551726,
      "tokens_seen": 1837301760
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024331851711745792,
      "loss": 1.4921,
      "theoretical_loss": 3.454942114611555,
      "tokens_seen": 1837367296
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024330196708908643,
      "loss": 2.732,
      "theoretical_loss": 3.454931667844901,
      "tokens_seen": 1837432832
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024328541706071497,
      "loss": 2.4136,
      "theoretical_loss": 3.4549212215551726,
      "tokens_seen": 1837498368
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024326886703234348,
      "loss": 2.4763,
      "theoretical_loss": 3.4549107757423307,
      "tokens_seen": 1837563904
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024325231700397202,
      "loss": 2.6422,
      "theoretical_loss": 3.4549003304063364,
      "tokens_seen": 1837629440
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024323576697560053,
      "loss": 2.7504,
      "theoretical_loss": 3.4548898855471517,
      "tokens_seen": 1837694976
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024321921694722907,
      "loss": 2.757,
      "theoretical_loss": 3.4548794411647368,
      "tokens_seen": 1837760512
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024320266691885758,
      "loss": 2.6651,
      "theoretical_loss": 3.4548689972590534,
      "tokens_seen": 1837826048
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002431861168904861,
      "loss": 2.5527,
      "theoretical_loss": 3.4548585538300625,
      "tokens_seen": 1837891584
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024316956686211462,
      "loss": 2.8062,
      "theoretical_loss": 3.454848110877726,
      "tokens_seen": 1837957120
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024315301683374316,
      "loss": 2.3087,
      "theoretical_loss": 3.4548376684020043,
      "tokens_seen": 1838022656
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024313646680537167,
      "loss": 2.4029,
      "theoretical_loss": 3.45482722640286,
      "tokens_seen": 1838088192
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002431199167770002,
      "loss": 2.8167,
      "theoretical_loss": 3.4548167848802525,
      "tokens_seen": 1838153728
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024310336674862872,
      "loss": 2.7654,
      "theoretical_loss": 3.454806343834144,
      "tokens_seen": 1838219264
    },
    {
      "epoch": 0.66,
      "objective/train/avg_token_score": 0.013940539211034775,
      "objective/train/avg_weight": 0.9835913777351379,
      "objective/train/docs_used": 1038918,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7710862159729004,
      "objective/train/std_weight": 0.0666249468922615,
      "objective/train/theoretical_loss": 3.454795903264496,
      "objective/train/tokens_used": 1858744800,
      "theoretical_loss": 3.454795903264496,
      "tokens_seen": 1838284800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024308681672025726,
      "loss": 2.8262,
      "theoretical_loss": 3.454795903264496,
      "tokens_seen": 1838284800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024307026669188574,
      "loss": 2.5994,
      "theoretical_loss": 3.4547854631712696,
      "tokens_seen": 1838350336
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024305371666351425,
      "loss": 2.63,
      "theoretical_loss": 3.454775023554426,
      "tokens_seen": 1838415872
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002430371666351428,
      "loss": 2.7138,
      "theoretical_loss": 3.454764584413927,
      "tokens_seen": 1838481408
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002430206166067713,
      "loss": 2.4228,
      "theoretical_loss": 3.4547541457497326,
      "tokens_seen": 1838546944
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024300406657839984,
      "loss": 2.6311,
      "theoretical_loss": 3.4547437075618053,
      "tokens_seen": 1838612480
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024298751655002835,
      "loss": 2.6579,
      "theoretical_loss": 3.454733269850106,
      "tokens_seen": 1838678016
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024297096652165688,
      "loss": 2.6545,
      "theoretical_loss": 3.4547228326145953,
      "tokens_seen": 1838743552
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002429544164932854,
      "loss": 2.5711,
      "theoretical_loss": 3.454712395855236,
      "tokens_seen": 1838809088
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024293786646491393,
      "loss": 2.1641,
      "theoretical_loss": 3.454701959571988,
      "tokens_seen": 1838874624
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024292131643654244,
      "loss": 2.2699,
      "theoretical_loss": 3.4546915237648137,
      "tokens_seen": 1838940160
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024290476640817098,
      "loss": 2.9746,
      "theoretical_loss": 3.4546810884336736,
      "tokens_seen": 1839005696
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002428882163797995,
      "loss": 2.4916,
      "theoretical_loss": 3.454670653578529,
      "tokens_seen": 1839071232
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024287166635142803,
      "loss": 2.6959,
      "theoretical_loss": 3.4546602191993427,
      "tokens_seen": 1839136768
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024285511632305654,
      "loss": 2.4852,
      "theoretical_loss": 3.4546497852960742,
      "tokens_seen": 1839202304
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024283856629468508,
      "loss": 2.6384,
      "theoretical_loss": 3.4546393518686855,
      "tokens_seen": 1839267840
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002428220162663136,
      "loss": 2.6863,
      "theoretical_loss": 3.4546289189171375,
      "tokens_seen": 1839333376
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024280546623794212,
      "loss": 2.5744,
      "theoretical_loss": 3.454618486441393,
      "tokens_seen": 1839398912
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024278891620957063,
      "loss": 2.422,
      "theoretical_loss": 3.4546080544414117,
      "tokens_seen": 1839464448
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024277236618119917,
      "loss": 2.7404,
      "theoretical_loss": 3.4545976229171558,
      "tokens_seen": 1839529984
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024275581615282768,
      "loss": 2.5813,
      "theoretical_loss": 3.454587191868587,
      "tokens_seen": 1839595520
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024273926612445622,
      "loss": 2.5244,
      "theoretical_loss": 3.454576761295665,
      "tokens_seen": 1839661056
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024272271609608473,
      "loss": 2.7638,
      "theoretical_loss": 3.454566331198353,
      "tokens_seen": 1839726592
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024270616606771327,
      "loss": 2.4423,
      "theoretical_loss": 3.4545559015766116,
      "tokens_seen": 1839792128
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024268961603934178,
      "loss": 2.7072,
      "theoretical_loss": 3.4545454724304028,
      "tokens_seen": 1839857664
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024267306601097032,
      "loss": 2.4169,
      "theoretical_loss": 3.4545350437596865,
      "tokens_seen": 1839923200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024265651598259883,
      "loss": 2.6918,
      "theoretical_loss": 3.4545246155644254,
      "tokens_seen": 1839988736
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024263996595422736,
      "loss": 2.6766,
      "theoretical_loss": 3.4545141878445804,
      "tokens_seen": 1840054272
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024262341592585587,
      "loss": 2.7595,
      "theoretical_loss": 3.4545037606001134,
      "tokens_seen": 1840119808
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002426068658974844,
      "loss": 2.182,
      "theoretical_loss": 3.4544933338309853,
      "tokens_seen": 1840185344
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002425903158691129,
      "loss": 2.5068,
      "theoretical_loss": 3.4544829075371575,
      "tokens_seen": 1840250880
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024257376584074143,
      "loss": 3.9635,
      "theoretical_loss": 3.4544724817185912,
      "tokens_seen": 1840316416
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024255721581236994,
      "loss": 2.7235,
      "theoretical_loss": 3.454462056375249,
      "tokens_seen": 1840381952
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024254066578399848,
      "loss": 11.2587,
      "theoretical_loss": 3.454451631507091,
      "tokens_seen": 1840447488
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000242524115755627,
      "loss": 22.3206,
      "theoretical_loss": 3.454441207114079,
      "tokens_seen": 1840513024
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024250756572725553,
      "loss": 2.8333,
      "theoretical_loss": 3.454430783196174,
      "tokens_seen": 1840578560
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024249101569888404,
      "loss": 2.1786,
      "theoretical_loss": 3.4544203597533385,
      "tokens_seen": 1840644096
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024247446567051258,
      "loss": 2.6281,
      "theoretical_loss": 3.454409936785533,
      "tokens_seen": 1840709632
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002424579156421411,
      "loss": 2.5896,
      "theoretical_loss": 3.45439951429272,
      "tokens_seen": 1840775168
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024244136561376963,
      "loss": 2.5522,
      "theoretical_loss": 3.4543890922748597,
      "tokens_seen": 1840840704
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024242481558539814,
      "loss": 2.7407,
      "theoretical_loss": 3.454378670731914,
      "tokens_seen": 1840906240
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024240826555702667,
      "loss": 2.7423,
      "theoretical_loss": 3.454368249663845,
      "tokens_seen": 1840971776
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024239171552865518,
      "loss": 2.491,
      "theoretical_loss": 3.454357829070613,
      "tokens_seen": 1841037312
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024237516550028372,
      "loss": 2.661,
      "theoretical_loss": 3.4543474089521804,
      "tokens_seen": 1841102848
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024235861547191223,
      "loss": 2.651,
      "theoretical_loss": 3.4543369893085085,
      "tokens_seen": 1841168384
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024234206544354077,
      "loss": 2.6493,
      "theoretical_loss": 3.4543265701395582,
      "tokens_seen": 1841233920
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024232551541516928,
      "loss": 2.4181,
      "theoretical_loss": 3.4543161514452914,
      "tokens_seen": 1841299456
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024230896538679782,
      "loss": 2.6491,
      "theoretical_loss": 3.45430573322567,
      "tokens_seen": 1841364992
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024229241535842633,
      "loss": 2.6224,
      "theoretical_loss": 3.4542953154806546,
      "tokens_seen": 1841430528
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024227586533005486,
      "loss": 2.595,
      "theoretical_loss": 3.4542848982102075,
      "tokens_seen": 1841496064
    },
    {
      "epoch": 0.66,
      "objective/train/avg_token_score": 0.01446022093296051,
      "objective/train/avg_weight": 0.9829805493354797,
      "objective/train/docs_used": 1041004,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.3724875450134277,
      "objective/train/std_weight": 0.06687811762094498,
      "objective/train/theoretical_loss": 3.4542744814142896,
      "objective/train/tokens_used": 1862021600,
      "theoretical_loss": 3.4542744814142896,
      "tokens_seen": 1841561600
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024225931530168338,
      "loss": 2.4405,
      "theoretical_loss": 3.4542744814142896,
      "tokens_seen": 1841561600
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002422427652733119,
      "loss": 2.4878,
      "theoretical_loss": 3.454264065092863,
      "tokens_seen": 1841627136
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024222621524494042,
      "loss": 2.6572,
      "theoretical_loss": 3.454253649245888,
      "tokens_seen": 1841692672
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024220966521656896,
      "loss": 2.559,
      "theoretical_loss": 3.4542432338733278,
      "tokens_seen": 1841758208
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024219311518819747,
      "loss": 2.5174,
      "theoretical_loss": 3.4542328189751426,
      "tokens_seen": 1841823744
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000242176565159826,
      "loss": 2.4331,
      "theoretical_loss": 3.4542224045512944,
      "tokens_seen": 1841889280
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024216001513145452,
      "loss": 2.6379,
      "theoretical_loss": 3.4542119906017446,
      "tokens_seen": 1841954816
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024214346510308306,
      "loss": 2.689,
      "theoretical_loss": 3.4542015771264554,
      "tokens_seen": 1842020352
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024212691507471157,
      "loss": 2.8545,
      "theoretical_loss": 3.454191164125387,
      "tokens_seen": 1842085888
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002421103650463401,
      "loss": 2.5831,
      "theoretical_loss": 3.4541807515985017,
      "tokens_seen": 1842151424
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002420938150179686,
      "loss": 2.6966,
      "theoretical_loss": 3.4541703395457617,
      "tokens_seen": 1842216960
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002420772649895971,
      "loss": 2.3475,
      "theoretical_loss": 3.4541599279671273,
      "tokens_seen": 1842282496
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024206071496122564,
      "loss": 2.5712,
      "theoretical_loss": 3.454149516862561,
      "tokens_seen": 1842348032
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024204416493285415,
      "loss": 2.8106,
      "theoretical_loss": 3.4541391062320237,
      "tokens_seen": 1842413568
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024202761490448268,
      "loss": 2.7944,
      "theoretical_loss": 3.454128696075477,
      "tokens_seen": 1842479104
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002420110648761112,
      "loss": 2.7725,
      "theoretical_loss": 3.454118286392883,
      "tokens_seen": 1842544640
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024199451484773973,
      "loss": 2.4932,
      "theoretical_loss": 3.454107877184203,
      "tokens_seen": 1842610176
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024197796481936824,
      "loss": 2.6535,
      "theoretical_loss": 3.4540974684493984,
      "tokens_seen": 1842675712
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024196141479099678,
      "loss": 2.2108,
      "theoretical_loss": 3.4540870601884306,
      "tokens_seen": 1842741248
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002419448647626253,
      "loss": 2.8081,
      "theoretical_loss": 3.4540766524012616,
      "tokens_seen": 1842806784
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024192831473425383,
      "loss": 2.4716,
      "theoretical_loss": 3.454066245087853,
      "tokens_seen": 1842872320
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024191176470588234,
      "loss": 2.7012,
      "theoretical_loss": 3.4540558382481663,
      "tokens_seen": 1842937856
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024189521467751088,
      "loss": 2.7812,
      "theoretical_loss": 3.4540454318821627,
      "tokens_seen": 1843003392
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002418786646491394,
      "loss": 2.5839,
      "theoretical_loss": 3.4540350259898043,
      "tokens_seen": 1843068928
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024186211462076792,
      "loss": 2.6824,
      "theoretical_loss": 3.4540246205710528,
      "tokens_seen": 1843134464
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024184556459239643,
      "loss": 2.6155,
      "theoretical_loss": 3.454014215625869,
      "tokens_seen": 1843200000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024182901456402497,
      "loss": 2.7853,
      "theoretical_loss": 3.4540038111542155,
      "tokens_seen": 1843265536
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024181246453565348,
      "loss": 2.7357,
      "theoretical_loss": 3.4539934071560534,
      "tokens_seen": 1843331072
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024179591450728202,
      "loss": 2.6126,
      "theoretical_loss": 3.453983003631344,
      "tokens_seen": 1843396608
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024177936447891053,
      "loss": 2.6924,
      "theoretical_loss": 3.45397260058005,
      "tokens_seen": 1843462144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024176281445053907,
      "loss": 2.7664,
      "theoretical_loss": 3.453962198002132,
      "tokens_seen": 1843527680
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024174626442216758,
      "loss": 2.4853,
      "theoretical_loss": 3.4539517958975514,
      "tokens_seen": 1843593216
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024172971439379612,
      "loss": 2.7003,
      "theoretical_loss": 3.4539413942662707,
      "tokens_seen": 1843658752
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024171316436542463,
      "loss": 2.6885,
      "theoretical_loss": 3.4539309931082514,
      "tokens_seen": 1843724288
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024169661433705316,
      "loss": 2.6194,
      "theoretical_loss": 3.4539205924234553,
      "tokens_seen": 1843789824
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024168006430868167,
      "loss": 2.8658,
      "theoretical_loss": 3.4539101922118434,
      "tokens_seen": 1843855360
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002416635142803102,
      "loss": 2.3615,
      "theoretical_loss": 3.4538997924733774,
      "tokens_seen": 1843920896
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024164696425193872,
      "loss": 2.6901,
      "theoretical_loss": 3.4538893932080197,
      "tokens_seen": 1843986432
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024163041422356726,
      "loss": 2.7129,
      "theoretical_loss": 3.4538789944157315,
      "tokens_seen": 1844051968
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024161386419519577,
      "loss": 2.5446,
      "theoretical_loss": 3.4538685960964743,
      "tokens_seen": 1844117504
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024159731416682425,
      "loss": 2.9071,
      "theoretical_loss": 3.45385819825021,
      "tokens_seen": 1844183040
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002415807641384528,
      "loss": 2.6693,
      "theoretical_loss": 3.4538478008769,
      "tokens_seen": 1844248576
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002415642141100813,
      "loss": 2.7598,
      "theoretical_loss": 3.4538374039765065,
      "tokens_seen": 1844314112
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024154766408170984,
      "loss": 2.6441,
      "theoretical_loss": 3.4538270075489907,
      "tokens_seen": 1844379648
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024153111405333835,
      "loss": 2.6508,
      "theoretical_loss": 3.4538166115943145,
      "tokens_seen": 1844445184
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002415145640249669,
      "loss": 2.6245,
      "theoretical_loss": 3.45380621611244,
      "tokens_seen": 1844510720
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002414980139965954,
      "loss": 2.7538,
      "theoretical_loss": 3.453795821103328,
      "tokens_seen": 1844576256
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024148146396822394,
      "loss": 2.6332,
      "theoretical_loss": 3.4537854265669408,
      "tokens_seen": 1844641792
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024146491393985245,
      "loss": 2.6026,
      "theoretical_loss": 3.4537750325032404,
      "tokens_seen": 1844707328
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024144836391148098,
      "loss": 2.4154,
      "theoretical_loss": 3.4537646389121877,
      "tokens_seen": 1844772864
    },
    {
      "epoch": 0.66,
      "objective/train/avg_token_score": 0.008798511698842049,
      "objective/train/avg_weight": 0.9896441698074341,
      "objective/train/docs_used": 1042471,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8529844284057617,
      "objective/train/std_weight": 0.06752355396747589,
      "objective/train/theoretical_loss": 3.4537542457937453,
      "objective/train/tokens_used": 1865298400,
      "theoretical_loss": 3.4537542457937453,
      "tokens_seen": 1844838400
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002414318138831095,
      "loss": 2.6585,
      "theoretical_loss": 3.4537542457937453,
      "tokens_seen": 1844838400
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024141526385473803,
      "loss": 2.6331,
      "theoretical_loss": 3.4537438531478735,
      "tokens_seen": 1844903936
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024139871382636654,
      "loss": 2.4454,
      "theoretical_loss": 3.4537334609745356,
      "tokens_seen": 1844969472
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024138216379799508,
      "loss": 2.8625,
      "theoretical_loss": 3.4537230692736927,
      "tokens_seen": 1845035008
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002413656137696236,
      "loss": 2.4492,
      "theoretical_loss": 3.4537126780453065,
      "tokens_seen": 1845100544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024134906374125213,
      "loss": 2.6632,
      "theoretical_loss": 3.453702287289339,
      "tokens_seen": 1845166080
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024133251371288064,
      "loss": 2.7762,
      "theoretical_loss": 3.4536918970057515,
      "tokens_seen": 1845231616
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024131596368450918,
      "loss": 2.5872,
      "theoretical_loss": 3.453681507194506,
      "tokens_seen": 1845297152
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024129941365613769,
      "loss": 2.7676,
      "theoretical_loss": 3.4536711178555644,
      "tokens_seen": 1845362688
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024128286362776622,
      "loss": 2.6332,
      "theoretical_loss": 3.453660728988888,
      "tokens_seen": 1845428224
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024126631359939473,
      "loss": 2.79,
      "theoretical_loss": 3.453650340594439,
      "tokens_seen": 1845493760
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024124976357102327,
      "loss": 2.7775,
      "theoretical_loss": 3.453639952672179,
      "tokens_seen": 1845559296
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002412332135426518,
      "loss": 2.621,
      "theoretical_loss": 3.4536295652220694,
      "tokens_seen": 1845624832
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024121666351428032,
      "loss": 2.5918,
      "theoretical_loss": 3.453619178244073,
      "tokens_seen": 1845690368
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024120011348590886,
      "loss": 2.7066,
      "theoretical_loss": 3.4536087917381506,
      "tokens_seen": 1845755904
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024118356345753737,
      "loss": 2.6286,
      "theoretical_loss": 3.4535984057042644,
      "tokens_seen": 1845821440
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002411670134291659,
      "loss": 2.4371,
      "theoretical_loss": 3.4535880201423765,
      "tokens_seen": 1845886976
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024115046340079442,
      "loss": 2.6761,
      "theoretical_loss": 3.4535776350524476,
      "tokens_seen": 1845952512
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024113391337242295,
      "loss": 2.7417,
      "theoretical_loss": 3.453567250434441,
      "tokens_seen": 1846018048
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024111736334405146,
      "loss": 2.628,
      "theoretical_loss": 3.453556866288317,
      "tokens_seen": 1846083584
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024110081331567995,
      "loss": 2.6176,
      "theoretical_loss": 3.453546482614038,
      "tokens_seen": 1846149120
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024108426328730848,
      "loss": 2.5447,
      "theoretical_loss": 3.4535360994115667,
      "tokens_seen": 1846214656
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000241067713258937,
      "loss": 2.654,
      "theoretical_loss": 3.453525716680863,
      "tokens_seen": 1846280192
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024105116323056553,
      "loss": 2.6032,
      "theoretical_loss": 3.453515334421891,
      "tokens_seen": 1846345728
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024103461320219404,
      "loss": 2.598,
      "theoretical_loss": 3.453504952634611,
      "tokens_seen": 1846411264
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024101806317382258,
      "loss": 2.588,
      "theoretical_loss": 3.453494571318985,
      "tokens_seen": 1846476800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002410015131454511,
      "loss": 2.6774,
      "theoretical_loss": 3.453484190474976,
      "tokens_seen": 1846542336
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024098496311707963,
      "loss": 2.8219,
      "theoretical_loss": 3.4534738101025435,
      "tokens_seen": 1846607872
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024096841308870814,
      "loss": 2.6526,
      "theoretical_loss": 3.4534634302016514,
      "tokens_seen": 1846673408
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024095186306033668,
      "loss": 2.5349,
      "theoretical_loss": 3.4534530507722607,
      "tokens_seen": 1846738944
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024093531303196519,
      "loss": 2.6452,
      "theoretical_loss": 3.453442671814334,
      "tokens_seen": 1846804480
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024091876300359372,
      "loss": 2.4899,
      "theoretical_loss": 3.4534322933278316,
      "tokens_seen": 1846870016
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024090221297522223,
      "loss": 2.6498,
      "theoretical_loss": 3.453421915312717,
      "tokens_seen": 1846935552
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024088566294685077,
      "loss": 2.7406,
      "theoretical_loss": 3.453411537768951,
      "tokens_seen": 1847001088
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024086911291847928,
      "loss": 2.8563,
      "theoretical_loss": 3.4534011606964965,
      "tokens_seen": 1847066624
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024085256289010782,
      "loss": 2.6439,
      "theoretical_loss": 3.4533907840953146,
      "tokens_seen": 1847132160
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024083601286173633,
      "loss": 2.6754,
      "theoretical_loss": 3.453380407965367,
      "tokens_seen": 1847197696
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024081946283336487,
      "loss": 2.7102,
      "theoretical_loss": 3.4533700323066157,
      "tokens_seen": 1847263232
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024080291280499338,
      "loss": 2.6232,
      "theoretical_loss": 3.4533596571190235,
      "tokens_seen": 1847328768
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024078636277662192,
      "loss": 2.6453,
      "theoretical_loss": 3.4533492824025513,
      "tokens_seen": 1847394304
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024076981274825043,
      "loss": 2.6523,
      "theoretical_loss": 3.453338908157161,
      "tokens_seen": 1847459840
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024075326271987896,
      "loss": 2.6623,
      "theoretical_loss": 3.4533285343828153,
      "tokens_seen": 1847525376
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024073671269150747,
      "loss": 2.7899,
      "theoretical_loss": 3.453318161079475,
      "tokens_seen": 1847590912
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000240720162663136,
      "loss": 2.6118,
      "theoretical_loss": 3.453307788247103,
      "tokens_seen": 1847656448
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024070361263476452,
      "loss": 2.6343,
      "theoretical_loss": 3.45329741588566,
      "tokens_seen": 1847721984
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024068706260639306,
      "loss": 2.6169,
      "theoretical_loss": 3.4532870439951098,
      "tokens_seen": 1847787520
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024067051257802157,
      "loss": 2.3889,
      "theoretical_loss": 3.453276672575413,
      "tokens_seen": 1847853056
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002406539625496501,
      "loss": 2.5022,
      "theoretical_loss": 3.4532663016265315,
      "tokens_seen": 1847918592
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024063741252127862,
      "loss": 2.5889,
      "theoretical_loss": 3.4532559311484277,
      "tokens_seen": 1847984128
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002406208624929071,
      "loss": 2.7587,
      "theoretical_loss": 3.4532455611410633,
      "tokens_seen": 1848049664
    },
    {
      "epoch": 0.66,
      "objective/train/avg_token_score": 0.022241879254579544,
      "objective/train/avg_weight": 0.9738261103630066,
      "objective/train/docs_used": 1043966,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.102886199951172,
      "objective/train/std_weight": 0.11464962363243103,
      "objective/train/theoretical_loss": 3.4532351916044,
      "objective/train/tokens_used": 1868575200,
      "theoretical_loss": 3.4532351916044,
      "tokens_seen": 1848115200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024060431246453564,
      "loss": 2.7863,
      "theoretical_loss": 3.4532351916044,
      "tokens_seen": 1848115200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024058776243616415,
      "loss": 2.7531,
      "theoretical_loss": 3.4532248225384006,
      "tokens_seen": 1848180736
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002405712124077927,
      "loss": 2.4863,
      "theoretical_loss": 3.453214453943026,
      "tokens_seen": 1848246272
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002405546623794212,
      "loss": 2.7981,
      "theoretical_loss": 3.4532040858182387,
      "tokens_seen": 1848311808
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024053811235104973,
      "loss": 2.7113,
      "theoretical_loss": 3.453193718164001,
      "tokens_seen": 1848377344
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024052156232267825,
      "loss": 2.8596,
      "theoretical_loss": 3.4531833509802734,
      "tokens_seen": 1848442880
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024050501229430678,
      "loss": 2.7669,
      "theoretical_loss": 3.45317298426702,
      "tokens_seen": 1848508416
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002404884622659353,
      "loss": 2.6545,
      "theoretical_loss": 3.453162618024201,
      "tokens_seen": 1848573952
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024047191223756383,
      "loss": 2.7158,
      "theoretical_loss": 3.4531522522517797,
      "tokens_seen": 1848639488
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024045536220919234,
      "loss": 2.5958,
      "theoretical_loss": 3.453141886949717,
      "tokens_seen": 1848705024
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024043881218082088,
      "loss": 2.7216,
      "theoretical_loss": 3.453131522117975,
      "tokens_seen": 1848770560
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002404222621524494,
      "loss": 2.561,
      "theoretical_loss": 3.4531211577565166,
      "tokens_seen": 1848836096
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024040571212407793,
      "loss": 2.5134,
      "theoretical_loss": 3.453110793865303,
      "tokens_seen": 1848901632
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024038916209570644,
      "loss": 2.7594,
      "theoretical_loss": 3.4531004304442963,
      "tokens_seen": 1848967168
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024037261206733497,
      "loss": 2.5773,
      "theoretical_loss": 3.4530900674934584,
      "tokens_seen": 1849032704
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024035606203896349,
      "loss": 2.7016,
      "theoretical_loss": 3.453079705012752,
      "tokens_seen": 1849098240
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024033951201059202,
      "loss": 2.6957,
      "theoretical_loss": 3.4530693430021384,
      "tokens_seen": 1849163776
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024032296198222053,
      "loss": 2.6361,
      "theoretical_loss": 3.45305898146158,
      "tokens_seen": 1849229312
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024030641195384907,
      "loss": 2.7738,
      "theoretical_loss": 3.453048620391038,
      "tokens_seen": 1849294848
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024028986192547758,
      "loss": 2.8329,
      "theoretical_loss": 3.4530382597904756,
      "tokens_seen": 1849360384
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024027331189710612,
      "loss": 2.7599,
      "theoretical_loss": 3.4530278996598547,
      "tokens_seen": 1849425920
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024025676186873463,
      "loss": 2.5542,
      "theoretical_loss": 3.453017539999136,
      "tokens_seen": 1849491456
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024024021184036317,
      "loss": 2.5613,
      "theoretical_loss": 3.4530071808082834,
      "tokens_seen": 1849556992
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024022366181199168,
      "loss": 2.5011,
      "theoretical_loss": 3.4529968220872576,
      "tokens_seen": 1849622528
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024020711178362021,
      "loss": 2.5689,
      "theoretical_loss": 3.4529864638360213,
      "tokens_seen": 1849688064
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024019056175524873,
      "loss": 2.7511,
      "theoretical_loss": 3.452976106054536,
      "tokens_seen": 1849753600
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024017401172687726,
      "loss": 2.6094,
      "theoretical_loss": 3.4529657487427636,
      "tokens_seen": 1849819136
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024015746169850577,
      "loss": 2.7315,
      "theoretical_loss": 3.4529553919006672,
      "tokens_seen": 1849884672
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002401409116701343,
      "loss": 2.5843,
      "theoretical_loss": 3.4529450355282085,
      "tokens_seen": 1849950208
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002401243616417628,
      "loss": 2.3298,
      "theoretical_loss": 3.452934679625349,
      "tokens_seen": 1850015744
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002401078116133913,
      "loss": 2.7419,
      "theoretical_loss": 3.4529243241920513,
      "tokens_seen": 1850081280
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024009126158501984,
      "loss": 2.6164,
      "theoretical_loss": 3.452913969228277,
      "tokens_seen": 1850146816
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024007471155664835,
      "loss": 2.6343,
      "theoretical_loss": 3.4529036147339887,
      "tokens_seen": 1850212352
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002400581615282769,
      "loss": 2.5076,
      "theoretical_loss": 3.452893260709148,
      "tokens_seen": 1850277888
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002400416114999054,
      "loss": 2.7102,
      "theoretical_loss": 3.4528829071537177,
      "tokens_seen": 1850343424
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024002506147153394,
      "loss": 2.6881,
      "theoretical_loss": 3.4528725540676595,
      "tokens_seen": 1850408960
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00024000851144316245,
      "loss": 2.5518,
      "theoretical_loss": 3.452862201450935,
      "tokens_seen": 1850474496
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023999196141479099,
      "loss": 2.7843,
      "theoretical_loss": 3.452851849303507,
      "tokens_seen": 1850540032
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023997541138641952,
      "loss": 2.7056,
      "theoretical_loss": 3.452841497625337,
      "tokens_seen": 1850605568
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023995886135804803,
      "loss": 2.5859,
      "theoretical_loss": 3.452831146416388,
      "tokens_seen": 1850671104
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023994231132967657,
      "loss": 2.2894,
      "theoretical_loss": 3.4528207956766215,
      "tokens_seen": 1850736640
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023992576130130508,
      "loss": 2.6518,
      "theoretical_loss": 3.4528104454059996,
      "tokens_seen": 1850802176
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023990921127293362,
      "loss": 2.4752,
      "theoretical_loss": 3.4528000956044846,
      "tokens_seen": 1850867712
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023989266124456213,
      "loss": 2.6491,
      "theoretical_loss": 3.4527897462720385,
      "tokens_seen": 1850933248
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023987611121619067,
      "loss": 2.6531,
      "theoretical_loss": 3.452779397408624,
      "tokens_seen": 1850998784
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023985956118781918,
      "loss": 2.6388,
      "theoretical_loss": 3.452769049014202,
      "tokens_seen": 1851064320
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023984301115944772,
      "loss": 2.6482,
      "theoretical_loss": 3.4527587010887357,
      "tokens_seen": 1851129856
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023982646113107623,
      "loss": 2.6632,
      "theoretical_loss": 3.4527483536321864,
      "tokens_seen": 1851195392
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023980991110270476,
      "loss": 2.6951,
      "theoretical_loss": 3.4527380066445175,
      "tokens_seen": 1851260928
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023979336107433327,
      "loss": 2.7123,
      "theoretical_loss": 3.4527276601256904,
      "tokens_seen": 1851326464
    },
    {
      "epoch": 0.66,
      "objective/train/avg_token_score": 0.009790003299713135,
      "objective/train/avg_weight": 0.9884960651397705,
      "objective/train/docs_used": 1045702,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.417029619216919,
      "objective/train/std_weight": 0.05439963564276695,
      "objective/train/theoretical_loss": 3.4527173140756666,
      "objective/train/tokens_used": 1871852000,
      "theoretical_loss": 3.4527173140756666,
      "tokens_seen": 1851392000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002397768110459618,
      "loss": 2.6163,
      "theoretical_loss": 3.4527173140756666,
      "tokens_seen": 1851392000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023976026101759032,
      "loss": 2.5458,
      "theoretical_loss": 3.4527069684944096,
      "tokens_seen": 1851457536
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023974371098921886,
      "loss": 2.6353,
      "theoretical_loss": 3.4526966233818808,
      "tokens_seen": 1851523072
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023972716096084737,
      "loss": 2.5146,
      "theoretical_loss": 3.4526862787380423,
      "tokens_seen": 1851588608
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002397106109324759,
      "loss": 2.6357,
      "theoretical_loss": 3.4526759345628566,
      "tokens_seen": 1851654144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023969406090410442,
      "loss": 2.6332,
      "theoretical_loss": 3.452665590856286,
      "tokens_seen": 1851719680
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023967751087573296,
      "loss": 2.5454,
      "theoretical_loss": 3.452655247618292,
      "tokens_seen": 1851785216
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023966096084736147,
      "loss": 2.7089,
      "theoretical_loss": 3.452644904848838,
      "tokens_seen": 1851850752
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023964441081899,
      "loss": 2.8305,
      "theoretical_loss": 3.4526345625478845,
      "tokens_seen": 1851916288
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002396278607906185,
      "loss": 2.1084,
      "theoretical_loss": 3.4526242207153954,
      "tokens_seen": 1851981824
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000239611310762247,
      "loss": 2.7093,
      "theoretical_loss": 3.4526138793513317,
      "tokens_seen": 1852047360
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023959476073387553,
      "loss": 2.6684,
      "theoretical_loss": 3.452603538455656,
      "tokens_seen": 1852112896
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023957821070550405,
      "loss": 2.8145,
      "theoretical_loss": 3.452593198028331,
      "tokens_seen": 1852178432
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023956166067713258,
      "loss": 2.6142,
      "theoretical_loss": 3.4525828580693183,
      "tokens_seen": 1852243968
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002395451106487611,
      "loss": 2.6445,
      "theoretical_loss": 3.45257251857858,
      "tokens_seen": 1852309504
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023952856062038963,
      "loss": 2.7131,
      "theoretical_loss": 3.452562179556079,
      "tokens_seen": 1852375040
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023951201059201814,
      "loss": 2.9403,
      "theoretical_loss": 3.452551841001777,
      "tokens_seen": 1852440576
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023949546056364668,
      "loss": 2.8121,
      "theoretical_loss": 3.4525415029156363,
      "tokens_seen": 1852506112
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002394789105352752,
      "loss": 2.6493,
      "theoretical_loss": 3.452531165297619,
      "tokens_seen": 1852571648
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023946236050690373,
      "loss": 2.7207,
      "theoretical_loss": 3.452520828147688,
      "tokens_seen": 1852637184
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023944581047853224,
      "loss": 2.4691,
      "theoretical_loss": 3.452510491465805,
      "tokens_seen": 1852702720
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023942926045016077,
      "loss": 2.6346,
      "theoretical_loss": 3.4525001552519323,
      "tokens_seen": 1852768256
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023941271042178929,
      "loss": 2.7301,
      "theoretical_loss": 3.4524898195060323,
      "tokens_seen": 1852833792
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023939616039341782,
      "loss": 2.6769,
      "theoretical_loss": 3.4524794842280673,
      "tokens_seen": 1852899328
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023937961036504633,
      "loss": 2.773,
      "theoretical_loss": 3.4524691494179995,
      "tokens_seen": 1852964864
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023936306033667487,
      "loss": 2.6063,
      "theoretical_loss": 3.452458815075791,
      "tokens_seen": 1853030400
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023934651030830338,
      "loss": 2.7364,
      "theoretical_loss": 3.452448481201404,
      "tokens_seen": 1853095936
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023932996027993192,
      "loss": 2.726,
      "theoretical_loss": 3.452438147794801,
      "tokens_seen": 1853161472
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023931341025156043,
      "loss": 2.5109,
      "theoretical_loss": 3.4524278148559446,
      "tokens_seen": 1853227008
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023929686022318897,
      "loss": 2.7295,
      "theoretical_loss": 3.4524174823847966,
      "tokens_seen": 1853292544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023928031019481748,
      "loss": 2.6098,
      "theoretical_loss": 3.452407150381319,
      "tokens_seen": 1853358080
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023926376016644601,
      "loss": 2.541,
      "theoretical_loss": 3.4523968188454752,
      "tokens_seen": 1853423616
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023924721013807452,
      "loss": 2.4986,
      "theoretical_loss": 3.452386487777227,
      "tokens_seen": 1853489152
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023923066010970306,
      "loss": 2.7167,
      "theoretical_loss": 3.452376157176536,
      "tokens_seen": 1853554688
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023921411008133157,
      "loss": 2.6961,
      "theoretical_loss": 3.452365827043365,
      "tokens_seen": 1853620224
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002391975600529601,
      "loss": 1.9837,
      "theoretical_loss": 3.4523554973776767,
      "tokens_seen": 1853685760
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023918101002458862,
      "loss": 2.7535,
      "theoretical_loss": 3.452345168179433,
      "tokens_seen": 1853751296
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023916445999621716,
      "loss": 2.547,
      "theoretical_loss": 3.4523348394485964,
      "tokens_seen": 1853816832
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023914790996784564,
      "loss": 2.7098,
      "theoretical_loss": 3.452324511185129,
      "tokens_seen": 1853882368
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023913135993947415,
      "loss": 2.6024,
      "theoretical_loss": 3.452314183388993,
      "tokens_seen": 1853947904
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002391148099111027,
      "loss": 2.6445,
      "theoretical_loss": 3.4523038560601513,
      "tokens_seen": 1854013440
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002390982598827312,
      "loss": 2.6406,
      "theoretical_loss": 3.452293529198566,
      "tokens_seen": 1854078976
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023908170985435974,
      "loss": 2.595,
      "theoretical_loss": 3.4522832028041988,
      "tokens_seen": 1854144512
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023906515982598825,
      "loss": 2.5889,
      "theoretical_loss": 3.4522728768770135,
      "tokens_seen": 1854210048
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023904860979761679,
      "loss": 2.8698,
      "theoretical_loss": 3.452262551416971,
      "tokens_seen": 1854275584
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002390320597692453,
      "loss": 2.6718,
      "theoretical_loss": 3.4522522264240347,
      "tokens_seen": 1854341120
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023901550974087383,
      "loss": 2.6934,
      "theoretical_loss": 3.452241901898166,
      "tokens_seen": 1854406656
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023899895971250234,
      "loss": 2.5964,
      "theoretical_loss": 3.452231577839328,
      "tokens_seen": 1854472192
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023898240968413088,
      "loss": 2.8162,
      "theoretical_loss": 3.452221254247483,
      "tokens_seen": 1854537728
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002389658596557594,
      "loss": 2.7929,
      "theoretical_loss": 3.4522109311225933,
      "tokens_seen": 1854603264
    },
    {
      "epoch": 0.66,
      "objective/train/avg_token_score": 0.011345956474542618,
      "objective/train/avg_weight": 0.9866488575935364,
      "objective/train/docs_used": 1047742,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0004372596740723,
      "objective/train/std_weight": 0.06199072301387787,
      "objective/train/theoretical_loss": 3.4522006084646204,
      "objective/train/tokens_used": 1875128800,
      "theoretical_loss": 3.4522006084646204,
      "tokens_seen": 1854668800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023894930962738793,
      "loss": 2.6706,
      "theoretical_loss": 3.4522006084646204,
      "tokens_seen": 1854668800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023893275959901644,
      "loss": 3.7591,
      "theoretical_loss": 3.4521902862735283,
      "tokens_seen": 1854734336
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023891620957064498,
      "loss": 2.5339,
      "theoretical_loss": 3.4521799645492783,
      "tokens_seen": 1854799872
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002388996595422735,
      "loss": 2.913,
      "theoretical_loss": 3.452169643291833,
      "tokens_seen": 1854865408
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023888310951390203,
      "loss": 2.4409,
      "theoretical_loss": 3.452159322501155,
      "tokens_seen": 1854930944
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023886655948553054,
      "loss": 2.652,
      "theoretical_loss": 3.4521490021772063,
      "tokens_seen": 1854996480
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023885000945715907,
      "loss": 2.4042,
      "theoretical_loss": 3.45213868231995,
      "tokens_seen": 1855062016
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023883345942878758,
      "loss": 2.6032,
      "theoretical_loss": 3.452128362929348,
      "tokens_seen": 1855127552
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023881690940041612,
      "loss": 2.6699,
      "theoretical_loss": 3.4521180440053625,
      "tokens_seen": 1855193088
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023880035937204463,
      "loss": 2.4203,
      "theoretical_loss": 3.452107725547956,
      "tokens_seen": 1855258624
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023878380934367317,
      "loss": 2.6768,
      "theoretical_loss": 3.4520974075570914,
      "tokens_seen": 1855324160
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023876725931530168,
      "loss": 2.8165,
      "theoretical_loss": 3.452087090032731,
      "tokens_seen": 1855389696
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023875070928693022,
      "loss": 2.7793,
      "theoretical_loss": 3.452076772974837,
      "tokens_seen": 1855455232
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023873415925855873,
      "loss": 2.7673,
      "theoretical_loss": 3.4520664563833723,
      "tokens_seen": 1855520768
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023871760923018727,
      "loss": 2.4971,
      "theoretical_loss": 3.452056140258298,
      "tokens_seen": 1855586304
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023870105920181578,
      "loss": 2.2626,
      "theoretical_loss": 3.4520458245995784,
      "tokens_seen": 1855651840
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023868450917344431,
      "loss": 2.6653,
      "theoretical_loss": 3.4520355094071746,
      "tokens_seen": 1855717376
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023866795914507282,
      "loss": 2.6676,
      "theoretical_loss": 3.45202519468105,
      "tokens_seen": 1855782912
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023865140911670133,
      "loss": 2.8711,
      "theoretical_loss": 3.452014880421166,
      "tokens_seen": 1855848448
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023863485908832984,
      "loss": 1.5686,
      "theoretical_loss": 3.452004566627486,
      "tokens_seen": 1855913984
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023861830905995838,
      "loss": 3.0243,
      "theoretical_loss": 3.4519942532999712,
      "tokens_seen": 1855979520
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002386017590315869,
      "loss": 2.7967,
      "theoretical_loss": 3.451983940438586,
      "tokens_seen": 1856045056
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023858520900321543,
      "loss": 2.8306,
      "theoretical_loss": 3.4519736280432918,
      "tokens_seen": 1856110592
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023856865897484394,
      "loss": 2.5829,
      "theoretical_loss": 3.4519633161140506,
      "tokens_seen": 1856176128
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023855210894647248,
      "loss": 2.5615,
      "theoretical_loss": 3.4519530046508256,
      "tokens_seen": 1856241664
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000238535558918101,
      "loss": 2.6862,
      "theoretical_loss": 3.451942693653579,
      "tokens_seen": 1856307200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023851900888972953,
      "loss": 2.8209,
      "theoretical_loss": 3.4519323831222737,
      "tokens_seen": 1856372736
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023850245886135804,
      "loss": 2.9042,
      "theoretical_loss": 3.4519220730568714,
      "tokens_seen": 1856438272
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023848590883298657,
      "loss": 2.5892,
      "theoretical_loss": 3.4519117634573355,
      "tokens_seen": 1856503808
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023846935880461508,
      "loss": 2.8712,
      "theoretical_loss": 3.451901454323628,
      "tokens_seen": 1856569344
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023845280877624362,
      "loss": 2.6512,
      "theoretical_loss": 3.4518911456557113,
      "tokens_seen": 1856634880
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023843625874787213,
      "loss": 2.7218,
      "theoretical_loss": 3.4518808374535483,
      "tokens_seen": 1856700416
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023841970871950067,
      "loss": 3.0722,
      "theoretical_loss": 3.4518705297171013,
      "tokens_seen": 1856765952
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023840315869112918,
      "loss": 2.7524,
      "theoretical_loss": 3.4518602224463324,
      "tokens_seen": 1856831488
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023838660866275772,
      "loss": 2.6034,
      "theoretical_loss": 3.4518499156412052,
      "tokens_seen": 1856897024
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023837005863438623,
      "loss": 2.6723,
      "theoretical_loss": 3.4518396093016817,
      "tokens_seen": 1856962560
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023835350860601477,
      "loss": 2.6831,
      "theoretical_loss": 3.451829303427724,
      "tokens_seen": 1857028096
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023833695857764328,
      "loss": 2.6935,
      "theoretical_loss": 3.451818998019295,
      "tokens_seen": 1857093632
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023832040854927181,
      "loss": 2.7459,
      "theoretical_loss": 3.451808693076357,
      "tokens_seen": 1857159168
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023830385852090032,
      "loss": 2.5817,
      "theoretical_loss": 3.4517983885988732,
      "tokens_seen": 1857224704
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023828730849252886,
      "loss": 2.7392,
      "theoretical_loss": 3.4517880845868056,
      "tokens_seen": 1857290240
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023827075846415737,
      "loss": 2.9743,
      "theoretical_loss": 3.4517777810401165,
      "tokens_seen": 1857355776
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002382542084357859,
      "loss": 2.6441,
      "theoretical_loss": 3.4517674779587693,
      "tokens_seen": 1857421312
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023823765840741442,
      "loss": 2.4615,
      "theoretical_loss": 3.451757175342726,
      "tokens_seen": 1857486848
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023822110837904296,
      "loss": 2.6777,
      "theoretical_loss": 3.451746873191949,
      "tokens_seen": 1857552384
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023820455835067147,
      "loss": 2.8173,
      "theoretical_loss": 3.4517365715064017,
      "tokens_seen": 1857617920
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002381880083223,
      "loss": 2.7338,
      "theoretical_loss": 3.451726270286046,
      "tokens_seen": 1857683456
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023817145829392852,
      "loss": 2.5869,
      "theoretical_loss": 3.451715969530844,
      "tokens_seen": 1857748992
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000238154908265557,
      "loss": 2.6427,
      "theoretical_loss": 3.4517056692407597,
      "tokens_seen": 1857814528
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023813835823718554,
      "loss": 2.6879,
      "theoretical_loss": 3.451695369415755,
      "tokens_seen": 1857880064
    },
    {
      "epoch": 0.66,
      "objective/train/avg_token_score": 0.022884123027324677,
      "objective/train/avg_weight": 0.9731055498123169,
      "objective/train/docs_used": 1049474,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0237350463867188,
      "objective/train/std_weight": 0.09664279967546463,
      "objective/train/theoretical_loss": 3.4516850700557917,
      "objective/train/tokens_used": 1878405600,
      "theoretical_loss": 3.4516850700557917,
      "tokens_seen": 1857945600
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023812180820881405,
      "loss": 2.8532,
      "theoretical_loss": 3.4516850700557917,
      "tokens_seen": 1857945600
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023810525818044259,
      "loss": 2.5419,
      "theoretical_loss": 3.451674771160833,
      "tokens_seen": 1858011136
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002380887081520711,
      "loss": 2.6955,
      "theoretical_loss": 3.451664472730842,
      "tokens_seen": 1858076672
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023807215812369963,
      "loss": 2.8206,
      "theoretical_loss": 3.451654174765781,
      "tokens_seen": 1858142208
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023805560809532814,
      "loss": 2.6882,
      "theoretical_loss": 3.451643877265613,
      "tokens_seen": 1858207744
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023803905806695668,
      "loss": 2.8556,
      "theoretical_loss": 3.451633580230299,
      "tokens_seen": 1858273280
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002380225080385852,
      "loss": 2.1002,
      "theoretical_loss": 3.4516232836598033,
      "tokens_seen": 1858338816
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023800595801021373,
      "loss": 2.6144,
      "theoretical_loss": 3.451612987554088,
      "tokens_seen": 1858404352
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023798940798184224,
      "loss": 2.6176,
      "theoretical_loss": 3.451602691913116,
      "tokens_seen": 1858469888
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023797285795347078,
      "loss": 2.6085,
      "theoretical_loss": 3.4515923967368494,
      "tokens_seen": 1858535424
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002379563079250993,
      "loss": 2.2158,
      "theoretical_loss": 3.4515821020252506,
      "tokens_seen": 1858600960
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023793975789672783,
      "loss": 2.6523,
      "theoretical_loss": 3.4515718077782838,
      "tokens_seen": 1858666496
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023792320786835634,
      "loss": 2.6491,
      "theoretical_loss": 3.4515615139959097,
      "tokens_seen": 1858732032
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023790665783998487,
      "loss": 2.9516,
      "theoretical_loss": 3.451551220678092,
      "tokens_seen": 1858797568
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023789010781161338,
      "loss": 2.7677,
      "theoretical_loss": 3.4515409278247935,
      "tokens_seen": 1858863104
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023787355778324192,
      "loss": 2.7712,
      "theoretical_loss": 3.4515306354359767,
      "tokens_seen": 1858928640
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023785700775487043,
      "loss": 2.656,
      "theoretical_loss": 3.4515203435116035,
      "tokens_seen": 1858994176
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023784045772649897,
      "loss": 2.7059,
      "theoretical_loss": 3.4515100520516375,
      "tokens_seen": 1859059712
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023782390769812748,
      "loss": 2.7655,
      "theoretical_loss": 3.4514997610560414,
      "tokens_seen": 1859125248
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023780735766975602,
      "loss": 2.7329,
      "theoretical_loss": 3.451489470524777,
      "tokens_seen": 1859190784
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023779080764138453,
      "loss": 2.9425,
      "theoretical_loss": 3.4514791804578078,
      "tokens_seen": 1859256320
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023777425761301307,
      "loss": 2.6674,
      "theoretical_loss": 3.4514688908550966,
      "tokens_seen": 1859321856
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023775770758464158,
      "loss": 2.5954,
      "theoretical_loss": 3.451458601716605,
      "tokens_seen": 1859387392
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002377411575562701,
      "loss": 2.7056,
      "theoretical_loss": 3.451448313042297,
      "tokens_seen": 1859452928
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023772460752789862,
      "loss": 2.6462,
      "theoretical_loss": 3.451438024832134,
      "tokens_seen": 1859518464
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023770805749952716,
      "loss": 2.857,
      "theoretical_loss": 3.45142773708608,
      "tokens_seen": 1859584000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023769150747115567,
      "loss": 2.5812,
      "theoretical_loss": 3.451417449804097,
      "tokens_seen": 1859649536
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023767495744278416,
      "loss": 2.9662,
      "theoretical_loss": 3.451407162986148,
      "tokens_seen": 1859715072
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002376584074144127,
      "loss": 2.5922,
      "theoretical_loss": 3.451396876632195,
      "tokens_seen": 1859780608
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002376418573860412,
      "loss": 2.8187,
      "theoretical_loss": 3.4513865907422017,
      "tokens_seen": 1859846144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023762530735766974,
      "loss": 2.7081,
      "theoretical_loss": 3.4513763053161304,
      "tokens_seen": 1859911680
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023760875732929825,
      "loss": 2.4792,
      "theoretical_loss": 3.4513660203539436,
      "tokens_seen": 1859977216
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002375922073009268,
      "loss": 2.7582,
      "theoretical_loss": 3.4513557358556044,
      "tokens_seen": 1860042752
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002375756572725553,
      "loss": 2.6539,
      "theoretical_loss": 3.4513454518210755,
      "tokens_seen": 1860108288
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023755910724418384,
      "loss": 2.6423,
      "theoretical_loss": 3.4513351682503193,
      "tokens_seen": 1860173824
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023754255721581235,
      "loss": 2.783,
      "theoretical_loss": 3.4513248851432987,
      "tokens_seen": 1860239360
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023752600718744088,
      "loss": 2.8737,
      "theoretical_loss": 3.451314602499977,
      "tokens_seen": 1860304896
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002375094571590694,
      "loss": 2.6,
      "theoretical_loss": 3.451304320320316,
      "tokens_seen": 1860370432
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023749290713069793,
      "loss": 2.772,
      "theoretical_loss": 3.451294038604279,
      "tokens_seen": 1860435968
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023747635710232644,
      "loss": 2.9361,
      "theoretical_loss": 3.4512837573518294,
      "tokens_seen": 1860501504
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023745980707395498,
      "loss": 2.6767,
      "theoretical_loss": 3.4512734765629283,
      "tokens_seen": 1860567040
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002374432570455835,
      "loss": 2.8677,
      "theoretical_loss": 3.45126319623754,
      "tokens_seen": 1860632576
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023742670701721203,
      "loss": 2.4931,
      "theoretical_loss": 3.451252916375627,
      "tokens_seen": 1860698112
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023741015698884054,
      "loss": 2.873,
      "theoretical_loss": 3.4512426369771516,
      "tokens_seen": 1860763648
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023739360696046908,
      "loss": 2.6732,
      "theoretical_loss": 3.4512323580420765,
      "tokens_seen": 1860829184
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002373770569320976,
      "loss": 2.8521,
      "theoretical_loss": 3.451222079570365,
      "tokens_seen": 1860894720
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023736050690372612,
      "loss": 2.687,
      "theoretical_loss": 3.4512118015619797,
      "tokens_seen": 1860960256
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023734395687535463,
      "loss": 2.9171,
      "theoretical_loss": 3.4512015240168834,
      "tokens_seen": 1861025792
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023732740684698317,
      "loss": 2.5096,
      "theoretical_loss": 3.451191246935039,
      "tokens_seen": 1861091328
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023731085681861168,
      "loss": 2.4897,
      "theoretical_loss": 3.4511809703164094,
      "tokens_seen": 1861156864
    },
    {
      "epoch": 0.66,
      "objective/train/avg_token_score": 0.015696238726377487,
      "objective/train/avg_weight": 0.9815456867218018,
      "objective/train/docs_used": 1051410,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1908035278320312,
      "objective/train/std_weight": 0.08212626725435257,
      "objective/train/theoretical_loss": 3.4511706941609566,
      "objective/train/tokens_used": 1881682400,
      "theoretical_loss": 3.4511706941609566,
      "tokens_seen": 1861222400
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023729430679024022,
      "loss": 2.8676,
      "theoretical_loss": 3.4511706941609566,
      "tokens_seen": 1861222400
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023727775676186873,
      "loss": 2.5847,
      "theoretical_loss": 3.4511604184686444,
      "tokens_seen": 1861287936
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023726120673349727,
      "loss": 2.6748,
      "theoretical_loss": 3.4511501432394356,
      "tokens_seen": 1861353472
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023724465670512578,
      "loss": 2.6732,
      "theoretical_loss": 3.451139868473292,
      "tokens_seen": 1861419008
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023722810667675432,
      "loss": 2.6651,
      "theoretical_loss": 3.4511295941701774,
      "tokens_seen": 1861484544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023721155664838285,
      "loss": 2.5259,
      "theoretical_loss": 3.4511193203300543,
      "tokens_seen": 1861550080
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023719500662001136,
      "loss": 2.637,
      "theoretical_loss": 3.451109046952886,
      "tokens_seen": 1861615616
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023717845659163985,
      "loss": 2.6425,
      "theoretical_loss": 3.451098774038634,
      "tokens_seen": 1861681152
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023716190656326839,
      "loss": 2.5666,
      "theoretical_loss": 3.4510885015872628,
      "tokens_seen": 1861746688
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002371453565348969,
      "loss": 2.3348,
      "theoretical_loss": 3.4510782295987346,
      "tokens_seen": 1861812224
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023712880650652543,
      "loss": 2.4675,
      "theoretical_loss": 3.4510679580730117,
      "tokens_seen": 1861877760
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00023711225647815394,
      "loss": 2.6752,
      "theoretical_loss": 3.4510576870100573,
      "tokens_seen": 1861943296
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023709570644978248,
      "loss": 2.5888,
      "theoretical_loss": 3.4510474164098346,
      "tokens_seen": 1862008832
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000237079156421411,
      "loss": 2.4931,
      "theoretical_loss": 3.4510371462723066,
      "tokens_seen": 1862074368
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023706260639303953,
      "loss": 2.5898,
      "theoretical_loss": 3.4510268765974357,
      "tokens_seen": 1862139904
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023704605636466804,
      "loss": 2.6605,
      "theoretical_loss": 3.451016607385185,
      "tokens_seen": 1862205440
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023702950633629658,
      "loss": 2.493,
      "theoretical_loss": 3.451006338635517,
      "tokens_seen": 1862270976
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002370129563079251,
      "loss": 2.794,
      "theoretical_loss": 3.4509960703483946,
      "tokens_seen": 1862336512
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023699640627955363,
      "loss": 2.6509,
      "theoretical_loss": 3.450985802523782,
      "tokens_seen": 1862402048
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023697985625118214,
      "loss": 2.6293,
      "theoretical_loss": 3.45097553516164,
      "tokens_seen": 1862467584
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023696330622281067,
      "loss": 2.8905,
      "theoretical_loss": 3.4509652682619327,
      "tokens_seen": 1862533120
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023694675619443918,
      "loss": 2.703,
      "theoretical_loss": 3.4509550018246236,
      "tokens_seen": 1862598656
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023693020616606772,
      "loss": 2.7063,
      "theoretical_loss": 3.4509447358496743,
      "tokens_seen": 1862664192
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023691365613769623,
      "loss": 2.6759,
      "theoretical_loss": 3.450934470337048,
      "tokens_seen": 1862729728
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023689710610932477,
      "loss": 2.4595,
      "theoretical_loss": 3.450924205286708,
      "tokens_seen": 1862795264
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023688055608095328,
      "loss": 2.63,
      "theoretical_loss": 3.4509139406986176,
      "tokens_seen": 1862860800
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023686400605258182,
      "loss": 2.4839,
      "theoretical_loss": 3.450903676572739,
      "tokens_seen": 1862926336
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023684745602421033,
      "loss": 2.7165,
      "theoretical_loss": 3.4508934129090356,
      "tokens_seen": 1862991872
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023683090599583887,
      "loss": 2.6654,
      "theoretical_loss": 3.4508831497074697,
      "tokens_seen": 1863057408
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023681435596746738,
      "loss": 2.5312,
      "theoretical_loss": 3.4508728869680043,
      "tokens_seen": 1863122944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002367978059390959,
      "loss": 2.7198,
      "theoretical_loss": 3.450862624690603,
      "tokens_seen": 1863188480
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023678125591072442,
      "loss": 2.6976,
      "theoretical_loss": 3.450852362875229,
      "tokens_seen": 1863254016
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023676470588235296,
      "loss": 2.8007,
      "theoretical_loss": 3.450842101521844,
      "tokens_seen": 1863319552
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023674815585398147,
      "loss": 2.6461,
      "theoretical_loss": 3.4508318406304115,
      "tokens_seen": 1863385088
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023673160582561,
      "loss": 2.5802,
      "theoretical_loss": 3.4508215802008952,
      "tokens_seen": 1863450624
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023671505579723852,
      "loss": 2.8075,
      "theoretical_loss": 3.450811320233257,
      "tokens_seen": 1863516160
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023669850576886706,
      "loss": 2.5832,
      "theoretical_loss": 3.4508010607274606,
      "tokens_seen": 1863581696
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023668195574049554,
      "loss": 2.705,
      "theoretical_loss": 3.4507908016834685,
      "tokens_seen": 1863647232
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023666540571212405,
      "loss": 2.7755,
      "theoretical_loss": 3.4507805431012435,
      "tokens_seen": 1863712768
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002366488556837526,
      "loss": 2.7226,
      "theoretical_loss": 3.4507702849807496,
      "tokens_seen": 1863778304
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002366323056553811,
      "loss": 2.6422,
      "theoretical_loss": 3.450760027321948,
      "tokens_seen": 1863843840
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023661575562700964,
      "loss": 2.7468,
      "theoretical_loss": 3.450749770124804,
      "tokens_seen": 1863909376
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023659920559863815,
      "loss": 2.778,
      "theoretical_loss": 3.4507395133892786,
      "tokens_seen": 1863974912
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023658265557026668,
      "loss": 2.8844,
      "theoretical_loss": 3.450729257115336,
      "tokens_seen": 1864040448
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002365661055418952,
      "loss": 2.756,
      "theoretical_loss": 3.4507190013029385,
      "tokens_seen": 1864105984
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023654955551352373,
      "loss": 2.5909,
      "theoretical_loss": 3.45070874595205,
      "tokens_seen": 1864171520
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023653300548515224,
      "loss": 2.5701,
      "theoretical_loss": 3.4506984910626324,
      "tokens_seen": 1864237056
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023651645545678078,
      "loss": 2.6971,
      "theoretical_loss": 3.450688236634649,
      "tokens_seen": 1864302592
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002364999054284093,
      "loss": 2.7031,
      "theoretical_loss": 3.450677982668063,
      "tokens_seen": 1864368128
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023648335540003783,
      "loss": 2.0262,
      "theoretical_loss": 3.450667729162838,
      "tokens_seen": 1864433664
    },
    {
      "epoch": 0.67,
      "objective/train/avg_token_score": 0.009412124752998352,
      "objective/train/avg_weight": 0.98892742395401,
      "objective/train/docs_used": 1053575,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.312844753265381,
      "objective/train/std_weight": 0.0419144444167614,
      "objective/train/theoretical_loss": 3.450657476118936,
      "objective/train/tokens_used": 1884959200,
      "theoretical_loss": 3.450657476118936,
      "tokens_seen": 1864499200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023646680537166634,
      "loss": 2.381,
      "theoretical_loss": 3.450657476118936,
      "tokens_seen": 1864499200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023645025534329488,
      "loss": 2.5781,
      "theoretical_loss": 3.450647223536321,
      "tokens_seen": 1864564736
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002364337053149234,
      "loss": 2.4409,
      "theoretical_loss": 3.450636971414955,
      "tokens_seen": 1864630272
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023641715528655192,
      "loss": 2.958,
      "theoretical_loss": 3.4506267197548017,
      "tokens_seen": 1864695808
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023640060525818043,
      "loss": 2.6498,
      "theoretical_loss": 3.4506164685558245,
      "tokens_seen": 1864761344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023638405522980897,
      "loss": 2.6289,
      "theoretical_loss": 3.4506062178179855,
      "tokens_seen": 1864826880
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023636750520143748,
      "loss": 2.5388,
      "theoretical_loss": 3.450595967541248,
      "tokens_seen": 1864892416
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023635095517306602,
      "loss": 2.6346,
      "theoretical_loss": 3.450585717725576,
      "tokens_seen": 1864957952
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023633440514469453,
      "loss": 2.7233,
      "theoretical_loss": 3.450575468370931,
      "tokens_seen": 1865023488
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023631785511632307,
      "loss": 2.3265,
      "theoretical_loss": 3.4505652194772773,
      "tokens_seen": 1865089024
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023630130508795158,
      "loss": 2.5367,
      "theoretical_loss": 3.4505549710445775,
      "tokens_seen": 1865154560
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023628475505958012,
      "loss": 2.5904,
      "theoretical_loss": 3.450544723072795,
      "tokens_seen": 1865220096
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023626820503120863,
      "loss": 2.5657,
      "theoretical_loss": 3.4505344755618923,
      "tokens_seen": 1865285632
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023625165500283716,
      "loss": 2.5652,
      "theoretical_loss": 3.450524228511833,
      "tokens_seen": 1865351168
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023623510497446567,
      "loss": 2.7243,
      "theoretical_loss": 3.4505139819225796,
      "tokens_seen": 1865416704
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002362185549460942,
      "loss": 2.5047,
      "theoretical_loss": 3.450503735794096,
      "tokens_seen": 1865482240
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023620200491772272,
      "loss": 2.5341,
      "theoretical_loss": 3.4504934901263447,
      "tokens_seen": 1865547776
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002361854548893512,
      "loss": 2.071,
      "theoretical_loss": 3.450483244919289,
      "tokens_seen": 1865613312
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023616890486097974,
      "loss": 2.0133,
      "theoretical_loss": 3.450473000172892,
      "tokens_seen": 1865678848
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023615235483260825,
      "loss": 2.6723,
      "theoretical_loss": 3.450462755887117,
      "tokens_seen": 1865744384
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002361358048042368,
      "loss": 2.3889,
      "theoretical_loss": 3.4504525120619265,
      "tokens_seen": 1865809920
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002361192547758653,
      "loss": 2.7448,
      "theoretical_loss": 3.4504422686972838,
      "tokens_seen": 1865875456
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023610270474749384,
      "loss": 2.1628,
      "theoretical_loss": 3.4504320257931527,
      "tokens_seen": 1865940992
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023608615471912235,
      "loss": 2.3956,
      "theoretical_loss": 3.450421783349496,
      "tokens_seen": 1866006528
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002360696046907509,
      "loss": 2.567,
      "theoretical_loss": 3.450411541366276,
      "tokens_seen": 1866072064
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002360530546623794,
      "loss": 2.7436,
      "theoretical_loss": 3.450401299843457,
      "tokens_seen": 1866137600
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023603650463400794,
      "loss": 2.5786,
      "theoretical_loss": 3.450391058781001,
      "tokens_seen": 1866203136
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023601995460563645,
      "loss": 2.3844,
      "theoretical_loss": 3.450380818178872,
      "tokens_seen": 1866268672
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023600340457726498,
      "loss": 2.7324,
      "theoretical_loss": 3.4503705780370337,
      "tokens_seen": 1866334208
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023598685454889352,
      "loss": 2.7084,
      "theoretical_loss": 3.4503603383554475,
      "tokens_seen": 1866399744
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023597030452052203,
      "loss": 2.7801,
      "theoretical_loss": 3.450350099134078,
      "tokens_seen": 1866465280
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023595375449215057,
      "loss": 2.5571,
      "theoretical_loss": 3.450339860372888,
      "tokens_seen": 1866530816
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023593720446377908,
      "loss": 2.4598,
      "theoretical_loss": 3.4503296220718402,
      "tokens_seen": 1866596352
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023592065443540762,
      "loss": 2.4055,
      "theoretical_loss": 3.4503193842308986,
      "tokens_seen": 1866661888
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023590410440703613,
      "loss": 2.3669,
      "theoretical_loss": 3.4503091468500253,
      "tokens_seen": 1866727424
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023588755437866466,
      "loss": 2.6317,
      "theoretical_loss": 3.4502989099291845,
      "tokens_seen": 1866792960
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023587100435029318,
      "loss": 2.3567,
      "theoretical_loss": 3.450288673468339,
      "tokens_seen": 1866858496
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002358544543219217,
      "loss": 2.6438,
      "theoretical_loss": 3.4502784374674516,
      "tokens_seen": 1866924032
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023583790429355022,
      "loss": 2.6343,
      "theoretical_loss": 3.4502682019264856,
      "tokens_seen": 1866989568
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023582135426517876,
      "loss": 2.7816,
      "theoretical_loss": 3.450257966845405,
      "tokens_seen": 1867055104
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023580480423680727,
      "loss": 2.7606,
      "theoretical_loss": 3.450247732224172,
      "tokens_seen": 1867120640
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002357882542084358,
      "loss": 2.5088,
      "theoretical_loss": 3.4502374980627506,
      "tokens_seen": 1867186176
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023577170418006432,
      "loss": 2.5143,
      "theoretical_loss": 3.4502272643611027,
      "tokens_seen": 1867251712
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023575515415169286,
      "loss": 2.6286,
      "theoretical_loss": 3.4502170311191933,
      "tokens_seen": 1867317248
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023573860412332137,
      "loss": 2.6003,
      "theoretical_loss": 3.4502067983369846,
      "tokens_seen": 1867382784
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002357220540949499,
      "loss": 2.5967,
      "theoretical_loss": 3.45019656601444,
      "tokens_seen": 1867448320
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002357055040665784,
      "loss": 2.609,
      "theoretical_loss": 3.450186334151522,
      "tokens_seen": 1867513856
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002356889540382069,
      "loss": 2.4377,
      "theoretical_loss": 3.4501761027481948,
      "tokens_seen": 1867579392
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023567240400983544,
      "loss": 2.5652,
      "theoretical_loss": 3.4501658718044217,
      "tokens_seen": 1867644928
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023565585398146395,
      "loss": 2.6494,
      "theoretical_loss": 3.4501556413201646,
      "tokens_seen": 1867710464
    },
    {
      "epoch": 0.67,
      "objective/train/avg_token_score": 0.015212046913802624,
      "objective/train/avg_weight": 0.9820915460586548,
      "objective/train/docs_used": 1055519,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.488315105438232,
      "objective/train/std_weight": 0.10470201820135117,
      "objective/train/theoretical_loss": 3.450145411295389,
      "objective/train/tokens_used": 1888236000,
      "theoretical_loss": 3.450145411295389,
      "tokens_seen": 1867776000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023563930395309248,
      "loss": 2.7206,
      "theoretical_loss": 3.450145411295389,
      "tokens_seen": 1867776000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000235622753924721,
      "loss": 2.5496,
      "theoretical_loss": 3.450135181730056,
      "tokens_seen": 1867841536
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023560620389634953,
      "loss": 2.5593,
      "theoretical_loss": 3.4501249526241295,
      "tokens_seen": 1867907072
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023558965386797804,
      "loss": 2.4342,
      "theoretical_loss": 3.450114723977573,
      "tokens_seen": 1867972608
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023557310383960658,
      "loss": 2.3993,
      "theoretical_loss": 3.4501044957903506,
      "tokens_seen": 1868038144
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002355565538112351,
      "loss": 2.6934,
      "theoretical_loss": 3.450094268062424,
      "tokens_seen": 1868103680
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023554000378286363,
      "loss": 2.718,
      "theoretical_loss": 3.4500840407937567,
      "tokens_seen": 1868169216
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023552345375449214,
      "loss": 2.6791,
      "theoretical_loss": 3.450073813984313,
      "tokens_seen": 1868234752
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023550690372612068,
      "loss": 2.3512,
      "theoretical_loss": 3.4500635876340553,
      "tokens_seen": 1868300288
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002354903536977492,
      "loss": 2.5728,
      "theoretical_loss": 3.450053361742947,
      "tokens_seen": 1868365824
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023547380366937772,
      "loss": 2.8308,
      "theoretical_loss": 3.450043136310952,
      "tokens_seen": 1868431360
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023545725364100623,
      "loss": 2.5674,
      "theoretical_loss": 3.450032911338032,
      "tokens_seen": 1868496896
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023544070361263477,
      "loss": 2.6222,
      "theoretical_loss": 3.4500226868241524,
      "tokens_seen": 1868562432
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023542415358426328,
      "loss": 1.8819,
      "theoretical_loss": 3.450012462769275,
      "tokens_seen": 1868627968
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023540760355589182,
      "loss": 2.476,
      "theoretical_loss": 3.450002239173364,
      "tokens_seen": 1868693504
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023539105352752033,
      "loss": 2.6766,
      "theoretical_loss": 3.449992016036382,
      "tokens_seen": 1868759040
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023537450349914887,
      "loss": 2.6893,
      "theoretical_loss": 3.449981793358292,
      "tokens_seen": 1868824576
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023535795347077738,
      "loss": 2.3147,
      "theoretical_loss": 3.4499715711390584,
      "tokens_seen": 1868890112
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023534140344240592,
      "loss": 2.4956,
      "theoretical_loss": 3.4499613493786443,
      "tokens_seen": 1868955648
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023532485341403443,
      "loss": 2.5454,
      "theoretical_loss": 3.4499511280770125,
      "tokens_seen": 1869021184
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023530830338566296,
      "loss": 2.4873,
      "theoretical_loss": 3.4499409072341263,
      "tokens_seen": 1869086720
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023529175335729147,
      "loss": 2.4752,
      "theoretical_loss": 3.4499306868499495,
      "tokens_seen": 1869152256
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023527520332892,
      "loss": 2.5072,
      "theoretical_loss": 3.449920466924445,
      "tokens_seen": 1869217792
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023525865330054852,
      "loss": 2.7221,
      "theoretical_loss": 3.4499102474575762,
      "tokens_seen": 1869283328
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023524210327217706,
      "loss": 2.5003,
      "theoretical_loss": 3.4499000284493073,
      "tokens_seen": 1869348864
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023522555324380557,
      "loss": 2.6045,
      "theoretical_loss": 3.4498898098996,
      "tokens_seen": 1869414400
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023520900321543405,
      "loss": 2.5559,
      "theoretical_loss": 3.4498795918084193,
      "tokens_seen": 1869479936
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002351924531870626,
      "loss": 2.486,
      "theoretical_loss": 3.4498693741757274,
      "tokens_seen": 1869545472
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002351759031586911,
      "loss": 2.655,
      "theoretical_loss": 3.4498591570014883,
      "tokens_seen": 1869611008
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023515935313031964,
      "loss": 2.6036,
      "theoretical_loss": 3.4498489402856647,
      "tokens_seen": 1869676544
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023514280310194815,
      "loss": 2.7189,
      "theoretical_loss": 3.4498387240282207,
      "tokens_seen": 1869742080
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002351262530735767,
      "loss": 2.7967,
      "theoretical_loss": 3.4498285082291194,
      "tokens_seen": 1869807616
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002351097030452052,
      "loss": 2.7191,
      "theoretical_loss": 3.449818292888324,
      "tokens_seen": 1869873152
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023509315301683374,
      "loss": 2.6492,
      "theoretical_loss": 3.4498080780057983,
      "tokens_seen": 1869938688
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023507660298846225,
      "loss": 2.5006,
      "theoretical_loss": 3.449797863581505,
      "tokens_seen": 1870004224
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023506005296009078,
      "loss": 2.576,
      "theoretical_loss": 3.4497876496154083,
      "tokens_seen": 1870069760
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002350435029317193,
      "loss": 2.5233,
      "theoretical_loss": 3.4497774361074707,
      "tokens_seen": 1870135296
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023502695290334783,
      "loss": 2.7585,
      "theoretical_loss": 3.4497672230576564,
      "tokens_seen": 1870200832
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023501040287497634,
      "loss": 2.2181,
      "theoretical_loss": 3.4497570104659285,
      "tokens_seen": 1870266368
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023499385284660488,
      "loss": 2.5146,
      "theoretical_loss": 3.4497467983322503,
      "tokens_seen": 1870331904
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002349773028182334,
      "loss": 2.4337,
      "theoretical_loss": 3.4497365866565852,
      "tokens_seen": 1870397440
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023496075278986193,
      "loss": 2.3316,
      "theoretical_loss": 3.4497263754388965,
      "tokens_seen": 1870462976
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023494420276149044,
      "loss": 2.7511,
      "theoretical_loss": 3.449716164679148,
      "tokens_seen": 1870528512
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023492765273311898,
      "loss": 2.688,
      "theoretical_loss": 3.449705954377303,
      "tokens_seen": 1870594048
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023491110270474749,
      "loss": 2.5872,
      "theoretical_loss": 3.449695744533325,
      "tokens_seen": 1870659584
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023489455267637602,
      "loss": 2.7457,
      "theoretical_loss": 3.449685535147177,
      "tokens_seen": 1870725120
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023487800264800453,
      "loss": 2.6022,
      "theoretical_loss": 3.4496753262188227,
      "tokens_seen": 1870790656
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023486145261963307,
      "loss": 2.5312,
      "theoretical_loss": 3.4496651177482254,
      "tokens_seen": 1870856192
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023484490259126158,
      "loss": 2.7203,
      "theoretical_loss": 3.449654909735349,
      "tokens_seen": 1870921728
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023482835256289012,
      "loss": 2.6013,
      "theoretical_loss": 3.449644702180156,
      "tokens_seen": 1870987264
    },
    {
      "epoch": 0.67,
      "objective/train/avg_token_score": 0.021659405902028084,
      "objective/train/avg_weight": 0.9744968414306641,
      "objective/train/docs_used": 1057572,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.702298164367676,
      "objective/train/std_weight": 0.10491879284381866,
      "objective/train/theoretical_loss": 3.4496344950826114,
      "objective/train/tokens_used": 1891512800,
      "theoretical_loss": 3.4496344950826114,
      "tokens_seen": 1871052800
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023481180253451863,
      "loss": 2.6017,
      "theoretical_loss": 3.4496344950826114,
      "tokens_seen": 1871052800
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023479525250614717,
      "loss": 2.6888,
      "theoretical_loss": 3.4496242884426773,
      "tokens_seen": 1871118336
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023477870247777568,
      "loss": 2.7909,
      "theoretical_loss": 3.449614082260317,
      "tokens_seen": 1871183872
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023476215244940421,
      "loss": 2.5927,
      "theoretical_loss": 3.4496038765354955,
      "tokens_seen": 1871249408
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023474560242103273,
      "loss": 2.5897,
      "theoretical_loss": 3.449593671268175,
      "tokens_seen": 1871314944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023472905239266126,
      "loss": 2.7713,
      "theoretical_loss": 3.449583466458319,
      "tokens_seen": 1871380480
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023471250236428975,
      "loss": 2.2232,
      "theoretical_loss": 3.4495732621058917,
      "tokens_seen": 1871446016
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023469595233591828,
      "loss": 2.4353,
      "theoretical_loss": 3.449563058210856,
      "tokens_seen": 1871511552
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002346794023075468,
      "loss": 2.6492,
      "theoretical_loss": 3.4495528547731755,
      "tokens_seen": 1871577088
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023466285227917533,
      "loss": 2.6021,
      "theoretical_loss": 3.4495426517928136,
      "tokens_seen": 1871642624
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023464630225080384,
      "loss": 2.0055,
      "theoretical_loss": 3.449532449269734,
      "tokens_seen": 1871708160
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023462975222243238,
      "loss": 2.6371,
      "theoretical_loss": 3.4495222472039,
      "tokens_seen": 1871773696
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002346132021940609,
      "loss": 2.5926,
      "theoretical_loss": 3.449512045595275,
      "tokens_seen": 1871839232
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023459665216568943,
      "loss": 2.538,
      "theoretical_loss": 3.449501844443823,
      "tokens_seen": 1871904768
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023458010213731794,
      "loss": 2.5956,
      "theoretical_loss": 3.449491643749507,
      "tokens_seen": 1871970304
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023456355210894648,
      "loss": 2.7791,
      "theoretical_loss": 3.449481443512291,
      "tokens_seen": 1872035840
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023454700208057499,
      "loss": 2.7303,
      "theoretical_loss": 3.449471243732138,
      "tokens_seen": 1872101376
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023453045205220352,
      "loss": 2.7849,
      "theoretical_loss": 3.449461044409012,
      "tokens_seen": 1872166912
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023451390202383203,
      "loss": 2.5381,
      "theoretical_loss": 3.449450845542876,
      "tokens_seen": 1872232448
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023449735199546057,
      "loss": 2.445,
      "theoretical_loss": 3.4494406471336934,
      "tokens_seen": 1872297984
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023448080196708908,
      "loss": 2.3374,
      "theoretical_loss": 3.4494304491814285,
      "tokens_seen": 1872363520
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023446425193871762,
      "loss": 2.5461,
      "theoretical_loss": 3.449420251686045,
      "tokens_seen": 1872429056
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023444770191034613,
      "loss": 2.5894,
      "theoretical_loss": 3.4494100546475055,
      "tokens_seen": 1872494592
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023443115188197467,
      "loss": 2.6012,
      "theoretical_loss": 3.4493998580657736,
      "tokens_seen": 1872560128
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023441460185360318,
      "loss": 2.6555,
      "theoretical_loss": 3.4493896619408138,
      "tokens_seen": 1872625664
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023439805182523172,
      "loss": 2.6204,
      "theoretical_loss": 3.449379466272589,
      "tokens_seen": 1872691200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023438150179686023,
      "loss": 2.5963,
      "theoretical_loss": 3.4493692710610624,
      "tokens_seen": 1872756736
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023436495176848876,
      "loss": 2.4749,
      "theoretical_loss": 3.449359076306198,
      "tokens_seen": 1872822272
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023434840174011727,
      "loss": 2.5109,
      "theoretical_loss": 3.4493488820079596,
      "tokens_seen": 1872887808
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002343318517117458,
      "loss": 2.3114,
      "theoretical_loss": 3.4493386881663106,
      "tokens_seen": 1872953344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023431530168337432,
      "loss": 2.487,
      "theoretical_loss": 3.4493284947812146,
      "tokens_seen": 1873018880
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023429875165500286,
      "loss": 2.4207,
      "theoretical_loss": 3.4493183018526348,
      "tokens_seen": 1873084416
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023428220162663137,
      "loss": 2.5238,
      "theoretical_loss": 3.449308109380535,
      "tokens_seen": 1873149952
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002342656515982599,
      "loss": 2.6196,
      "theoretical_loss": 3.449297917364879,
      "tokens_seen": 1873215488
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023424910156988842,
      "loss": 2.7216,
      "theoretical_loss": 3.44928772580563,
      "tokens_seen": 1873281024
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002342325515415169,
      "loss": 2.4546,
      "theoretical_loss": 3.449277534702752,
      "tokens_seen": 1873346560
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023421600151314544,
      "loss": 2.6942,
      "theoretical_loss": 3.4492673440562083,
      "tokens_seen": 1873412096
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023419945148477395,
      "loss": 2.5803,
      "theoretical_loss": 3.449257153865963,
      "tokens_seen": 1873477632
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002341829014564025,
      "loss": 2.7193,
      "theoretical_loss": 3.449246964131979,
      "tokens_seen": 1873543168
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000234166351428031,
      "loss": 2.7345,
      "theoretical_loss": 3.4492367748542203,
      "tokens_seen": 1873608704
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023414980139965953,
      "loss": 2.8424,
      "theoretical_loss": 3.4492265860326508,
      "tokens_seen": 1873674240
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023413325137128805,
      "loss": 2.6351,
      "theoretical_loss": 3.449216397667233,
      "tokens_seen": 1873739776
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023411670134291658,
      "loss": 2.605,
      "theoretical_loss": 3.449206209757932,
      "tokens_seen": 1873805312
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002341001513145451,
      "loss": 2.4403,
      "theoretical_loss": 3.4491960223047107,
      "tokens_seen": 1873870848
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023408360128617363,
      "loss": 2.5947,
      "theoretical_loss": 3.4491858353075324,
      "tokens_seen": 1873936384
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023406705125780214,
      "loss": 2.6099,
      "theoretical_loss": 3.449175648766362,
      "tokens_seen": 1874001920
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023405050122943068,
      "loss": 2.6852,
      "theoretical_loss": 3.449165462681161,
      "tokens_seen": 1874067456
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002340339512010592,
      "loss": 2.5686,
      "theoretical_loss": 3.4491552770518954,
      "tokens_seen": 1874132992
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023401740117268773,
      "loss": 2.62,
      "theoretical_loss": 3.449145091878527,
      "tokens_seen": 1874198528
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023400085114431624,
      "loss": 2.6964,
      "theoretical_loss": 3.4491349071610204,
      "tokens_seen": 1874264064
    },
    {
      "epoch": 0.67,
      "objective/train/avg_token_score": 0.012178178876638412,
      "objective/train/avg_weight": 0.9856875538825989,
      "objective/train/docs_used": 1059653,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.573986053466797,
      "objective/train/std_weight": 0.07452140748500824,
      "objective/train/theoretical_loss": 3.449124722899339,
      "objective/train/tokens_used": 1894789600,
      "theoretical_loss": 3.449124722899339,
      "tokens_seen": 1874329600
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023398430111594477,
      "loss": 2.4238,
      "theoretical_loss": 3.449124722899339,
      "tokens_seen": 1874329600
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023396775108757329,
      "loss": 2.562,
      "theoretical_loss": 3.449114539093447,
      "tokens_seen": 1874395136
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023395120105920182,
      "loss": 2.7302,
      "theoretical_loss": 3.449104355743307,
      "tokens_seen": 1874460672
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023393465103083033,
      "loss": 2.2857,
      "theoretical_loss": 3.4490941728488838,
      "tokens_seen": 1874526208
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023391810100245887,
      "loss": 2.461,
      "theoretical_loss": 3.44908399041014,
      "tokens_seen": 1874591744
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023390155097408738,
      "loss": 2.5453,
      "theoretical_loss": 3.4490738084270403,
      "tokens_seen": 1874657280
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023388500094571592,
      "loss": 2.7104,
      "theoretical_loss": 3.4490636268995476,
      "tokens_seen": 1874722816
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023386845091734443,
      "loss": 2.4506,
      "theoretical_loss": 3.4490534458276256,
      "tokens_seen": 1874788352
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023385190088897297,
      "loss": 2.8458,
      "theoretical_loss": 3.449043265211239,
      "tokens_seen": 1874853888
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023383535086060148,
      "loss": 2.791,
      "theoretical_loss": 3.4490330850503503,
      "tokens_seen": 1874919424
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023381880083223001,
      "loss": 2.3779,
      "theoretical_loss": 3.4490229053449237,
      "tokens_seen": 1874984960
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023380225080385853,
      "loss": 2.5069,
      "theoretical_loss": 3.449012726094923,
      "tokens_seen": 1875050496
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023378570077548706,
      "loss": 2.6743,
      "theoretical_loss": 3.449002547300312,
      "tokens_seen": 1875116032
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023376915074711557,
      "loss": 2.6136,
      "theoretical_loss": 3.4489923689610538,
      "tokens_seen": 1875181568
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002337526007187441,
      "loss": 2.4764,
      "theoretical_loss": 3.4489821910771123,
      "tokens_seen": 1875247104
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002337360506903726,
      "loss": 2.4803,
      "theoretical_loss": 3.448972013648452,
      "tokens_seen": 1875312640
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002337195006620011,
      "loss": 2.5152,
      "theoretical_loss": 3.4489618366750356,
      "tokens_seen": 1875378176
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023370295063362964,
      "loss": 2.5465,
      "theoretical_loss": 3.448951660156828,
      "tokens_seen": 1875443712
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023368640060525815,
      "loss": 2.5493,
      "theoretical_loss": 3.4489414840937913,
      "tokens_seen": 1875509248
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002336698505768867,
      "loss": 2.3372,
      "theoretical_loss": 3.448931308485891,
      "tokens_seen": 1875574784
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002336533005485152,
      "loss": 2.5274,
      "theoretical_loss": 3.4489211333330894,
      "tokens_seen": 1875640320
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023363675052014374,
      "loss": 2.2372,
      "theoretical_loss": 3.4489109586353512,
      "tokens_seen": 1875705856
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023362020049177225,
      "loss": 2.7025,
      "theoretical_loss": 3.44890078439264,
      "tokens_seen": 1875771392
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023360365046340079,
      "loss": 2.6586,
      "theoretical_loss": 3.448890610604918,
      "tokens_seen": 1875836928
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002335871004350293,
      "loss": 2.6046,
      "theoretical_loss": 3.4488804372721518,
      "tokens_seen": 1875902464
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023357055040665783,
      "loss": 2.6464,
      "theoretical_loss": 3.448870264394303,
      "tokens_seen": 1875968000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023355400037828634,
      "loss": 2.5858,
      "theoretical_loss": 3.448860091971336,
      "tokens_seen": 1876033536
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023353745034991488,
      "loss": 2.4291,
      "theoretical_loss": 3.448849920003215,
      "tokens_seen": 1876099072
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002335209003215434,
      "loss": 2.4684,
      "theoretical_loss": 3.4488397484899034,
      "tokens_seen": 1876164608
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023350435029317193,
      "loss": 2.4939,
      "theoretical_loss": 3.448829577431365,
      "tokens_seen": 1876230144
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023348780026480044,
      "loss": 2.5552,
      "theoretical_loss": 3.448819406827563,
      "tokens_seen": 1876295680
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023347125023642898,
      "loss": 2.5295,
      "theoretical_loss": 3.448809236678462,
      "tokens_seen": 1876361216
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002334547002080575,
      "loss": 2.5659,
      "theoretical_loss": 3.4487990669840256,
      "tokens_seen": 1876426752
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023343815017968603,
      "loss": 2.6865,
      "theoretical_loss": 3.448788897744217,
      "tokens_seen": 1876492288
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023342160015131454,
      "loss": 2.6908,
      "theoretical_loss": 3.4487787289590015,
      "tokens_seen": 1876557824
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023340505012294307,
      "loss": 2.8715,
      "theoretical_loss": 3.4487685606283414,
      "tokens_seen": 1876623360
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023338850009457158,
      "loss": 2.5156,
      "theoretical_loss": 3.4487583927522008,
      "tokens_seen": 1876688896
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023337195006620012,
      "loss": 2.7267,
      "theoretical_loss": 3.4487482253305437,
      "tokens_seen": 1876754432
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023335540003782863,
      "loss": 2.6671,
      "theoretical_loss": 3.4487380583633342,
      "tokens_seen": 1876819968
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023333885000945717,
      "loss": 2.7598,
      "theoretical_loss": 3.4487278918505364,
      "tokens_seen": 1876885504
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023332229998108568,
      "loss": 2.5569,
      "theoretical_loss": 3.448717725792113,
      "tokens_seen": 1876951040
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023330574995271422,
      "loss": 2.5229,
      "theoretical_loss": 3.448707560188028,
      "tokens_seen": 1877016576
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023328919992434273,
      "loss": 2.7077,
      "theoretical_loss": 3.4486973950382467,
      "tokens_seen": 1877082112
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023327264989597127,
      "loss": 2.6136,
      "theoretical_loss": 3.448687230342731,
      "tokens_seen": 1877147648
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023325609986759978,
      "loss": 2.8027,
      "theoretical_loss": 3.448677066101446,
      "tokens_seen": 1877213184
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002332395498392283,
      "loss": 2.1703,
      "theoretical_loss": 3.448666902314355,
      "tokens_seen": 1877278720
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002332229998108568,
      "loss": 2.6714,
      "theoretical_loss": 3.4486567389814224,
      "tokens_seen": 1877344256
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023320644978248533,
      "loss": 2.6473,
      "theoretical_loss": 3.4486465761026115,
      "tokens_seen": 1877409792
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023318989975411385,
      "loss": 2.5228,
      "theoretical_loss": 3.448636413677886,
      "tokens_seen": 1877475328
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023317334972574238,
      "loss": 2.5255,
      "theoretical_loss": 3.4486262517072106,
      "tokens_seen": 1877540864
    },
    {
      "epoch": 0.67,
      "objective/train/avg_token_score": 0.016940295696258545,
      "objective/train/avg_weight": 0.9800722002983093,
      "objective/train/docs_used": 1061454,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.888223171234131,
      "objective/train/std_weight": 0.08463625609874725,
      "objective/train/theoretical_loss": 3.4486160901905487,
      "objective/train/tokens_used": 1898066400,
      "theoretical_loss": 3.4486160901905487,
      "tokens_seen": 1877606400
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002331567996973709,
      "loss": 2.4982,
      "theoretical_loss": 3.4486160901905487,
      "tokens_seen": 1877606400
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023314024966899943,
      "loss": 2.8999,
      "theoretical_loss": 3.448605929127864,
      "tokens_seen": 1877671936
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023312369964062794,
      "loss": 2.4773,
      "theoretical_loss": 3.4485957685191204,
      "tokens_seen": 1877737472
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023310714961225648,
      "loss": 2.6207,
      "theoretical_loss": 3.448585608364282,
      "tokens_seen": 1877803008
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000233090599583885,
      "loss": 2.4565,
      "theoretical_loss": 3.4485754486633127,
      "tokens_seen": 1877868544
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023307404955551353,
      "loss": 2.7271,
      "theoretical_loss": 3.448565289416176,
      "tokens_seen": 1877934080
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023305749952714204,
      "loss": 2.4738,
      "theoretical_loss": 3.448555130622836,
      "tokens_seen": 1877999616
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023304094949877057,
      "loss": 2.6312,
      "theoretical_loss": 3.4485449722832566,
      "tokens_seen": 1878065152
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023302439947039909,
      "loss": 2.8433,
      "theoretical_loss": 3.4485348143974024,
      "tokens_seen": 1878130688
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023300784944202762,
      "loss": 2.5791,
      "theoretical_loss": 3.448524656965236,
      "tokens_seen": 1878196224
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023299129941365613,
      "loss": 2.3077,
      "theoretical_loss": 3.448514499986722,
      "tokens_seen": 1878261760
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023297474938528467,
      "loss": 2.8412,
      "theoretical_loss": 3.448504343461824,
      "tokens_seen": 1878327296
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023295819935691318,
      "loss": 2.6329,
      "theoretical_loss": 3.448494187390507,
      "tokens_seen": 1878392832
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023294164932854172,
      "loss": 2.5255,
      "theoretical_loss": 3.448484031772734,
      "tokens_seen": 1878458368
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023292509930017023,
      "loss": 2.7046,
      "theoretical_loss": 3.4484738766084684,
      "tokens_seen": 1878523904
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023290854927179877,
      "loss": 2.4145,
      "theoretical_loss": 3.4484637218976752,
      "tokens_seen": 1878589440
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023289199924342728,
      "loss": 2.1993,
      "theoretical_loss": 3.4484535676403176,
      "tokens_seen": 1878654976
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023287544921505581,
      "loss": 2.5919,
      "theoretical_loss": 3.44844341383636,
      "tokens_seen": 1878720512
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023285889918668432,
      "loss": 2.5171,
      "theoretical_loss": 3.448433260485766,
      "tokens_seen": 1878786048
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023284234915831286,
      "loss": 2.9178,
      "theoretical_loss": 3.4484231075885,
      "tokens_seen": 1878851584
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023282579912994137,
      "loss": 2.5059,
      "theoretical_loss": 3.4484129551445255,
      "tokens_seen": 1878917120
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002328092491015699,
      "loss": 2.2838,
      "theoretical_loss": 3.4484028031538063,
      "tokens_seen": 1878982656
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023279269907319842,
      "loss": 2.5271,
      "theoretical_loss": 3.4483926516163073,
      "tokens_seen": 1879048192
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023277614904482696,
      "loss": 2.6998,
      "theoretical_loss": 3.4483825005319915,
      "tokens_seen": 1879113728
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023275959901645544,
      "loss": 2.6408,
      "theoretical_loss": 3.448372349900823,
      "tokens_seen": 1879179264
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023274304898808395,
      "loss": 2.4611,
      "theoretical_loss": 3.448362199722766,
      "tokens_seen": 1879244800
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002327264989597125,
      "loss": 2.5048,
      "theoretical_loss": 3.4483520499977844,
      "tokens_seen": 1879310336
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000232709948931341,
      "loss": 2.529,
      "theoretical_loss": 3.4483419007258425,
      "tokens_seen": 1879375872
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023269339890296954,
      "loss": 2.4632,
      "theoretical_loss": 3.4483317519069043,
      "tokens_seen": 1879441408
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023267684887459805,
      "loss": 2.5114,
      "theoretical_loss": 3.4483216035409328,
      "tokens_seen": 1879506944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023266029884622659,
      "loss": 2.5475,
      "theoretical_loss": 3.4483114556278927,
      "tokens_seen": 1879572480
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002326437488178551,
      "loss": 2.7129,
      "theoretical_loss": 3.448301308167748,
      "tokens_seen": 1879638016
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023262719878948363,
      "loss": 2.3339,
      "theoretical_loss": 3.448291161160463,
      "tokens_seen": 1879703552
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023261064876111214,
      "loss": 2.6674,
      "theoretical_loss": 3.4482810146060014,
      "tokens_seen": 1879769088
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023259409873274068,
      "loss": 2.8828,
      "theoretical_loss": 3.4482708685043266,
      "tokens_seen": 1879834624
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002325775487043692,
      "loss": 2.6785,
      "theoretical_loss": 3.4482607228554034,
      "tokens_seen": 1879900160
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023256099867599773,
      "loss": 2.7913,
      "theoretical_loss": 3.448250577659196,
      "tokens_seen": 1879965696
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023254444864762624,
      "loss": 2.386,
      "theoretical_loss": 3.448240432915667,
      "tokens_seen": 1880031232
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023252789861925478,
      "loss": 2.522,
      "theoretical_loss": 3.4482302886247824,
      "tokens_seen": 1880096768
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002325113485908833,
      "loss": 2.4926,
      "theoretical_loss": 3.448220144786505,
      "tokens_seen": 1880162304
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023249479856251183,
      "loss": 2.3584,
      "theoretical_loss": 3.448210001400799,
      "tokens_seen": 1880227840
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023247824853414034,
      "loss": 2.6683,
      "theoretical_loss": 3.4481998584676288,
      "tokens_seen": 1880293376
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023246169850576887,
      "loss": 2.6012,
      "theoretical_loss": 3.448189715986958,
      "tokens_seen": 1880358912
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023244514847739738,
      "loss": 2.652,
      "theoretical_loss": 3.4481795739587504,
      "tokens_seen": 1880424448
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023242859844902592,
      "loss": 2.7041,
      "theoretical_loss": 3.448169432382971,
      "tokens_seen": 1880489984
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023241204842065443,
      "loss": 2.5807,
      "theoretical_loss": 3.448159291259583,
      "tokens_seen": 1880555520
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023239549839228297,
      "loss": 2.7607,
      "theoretical_loss": 3.448149150588551,
      "tokens_seen": 1880621056
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023237894836391148,
      "loss": 2.47,
      "theoretical_loss": 3.4481390103698386,
      "tokens_seen": 1880686592
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023236239833554002,
      "loss": 2.576,
      "theoretical_loss": 3.44812887060341,
      "tokens_seen": 1880752128
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023234584830716853,
      "loss": 2.4843,
      "theoretical_loss": 3.448118731289229,
      "tokens_seen": 1880817664
    },
    {
      "epoch": 0.67,
      "objective/train/avg_token_score": 0.012399973347783089,
      "objective/train/avg_weight": 0.9854126572608948,
      "objective/train/docs_used": 1063292,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.1329681873321533,
      "objective/train/std_weight": 0.034261081367731094,
      "objective/train/theoretical_loss": 3.448108592427261,
      "objective/train/tokens_used": 1901343200,
      "theoretical_loss": 3.448108592427261,
      "tokens_seen": 1880883200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023232929827879707,
      "loss": 2.3376,
      "theoretical_loss": 3.448108592427261,
      "tokens_seen": 1880883200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023231274825042558,
      "loss": 2.6685,
      "theoretical_loss": 3.448098454017469,
      "tokens_seen": 1880948736
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023229619822205411,
      "loss": 2.6422,
      "theoretical_loss": 3.4480883160598164,
      "tokens_seen": 1881014272
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023227964819368262,
      "loss": 2.6889,
      "theoretical_loss": 3.448078178554268,
      "tokens_seen": 1881079808
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002322630981653111,
      "loss": 2.3623,
      "theoretical_loss": 3.4480680415007887,
      "tokens_seen": 1881145344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023224654813693964,
      "loss": 2.4685,
      "theoretical_loss": 3.4480579048993416,
      "tokens_seen": 1881210880
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023222999810856816,
      "loss": 2.6895,
      "theoretical_loss": 3.448047768749891,
      "tokens_seen": 1881276416
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002322134480801967,
      "loss": 2.5117,
      "theoretical_loss": 3.448037633052401,
      "tokens_seen": 1881341952
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002321968980518252,
      "loss": 2.5054,
      "theoretical_loss": 3.4480274978068355,
      "tokens_seen": 1881407488
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023218034802345374,
      "loss": 2.3487,
      "theoretical_loss": 3.448017363013159,
      "tokens_seen": 1881473024
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023216379799508225,
      "loss": 2.6843,
      "theoretical_loss": 3.4480072286713357,
      "tokens_seen": 1881538560
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002321472479667108,
      "loss": 2.423,
      "theoretical_loss": 3.4479970947813294,
      "tokens_seen": 1881604096
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002321306979383393,
      "loss": 2.7973,
      "theoretical_loss": 3.447986961343104,
      "tokens_seen": 1881669632
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023211414790996784,
      "loss": 2.744,
      "theoretical_loss": 3.4479768283566243,
      "tokens_seen": 1881735168
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023209759788159635,
      "loss": 2.7025,
      "theoretical_loss": 3.447966695821854,
      "tokens_seen": 1881800704
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023208104785322488,
      "loss": 2.4646,
      "theoretical_loss": 3.447956563738757,
      "tokens_seen": 1881866240
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002320644978248534,
      "loss": 2.4206,
      "theoretical_loss": 3.4479464321072983,
      "tokens_seen": 1881931776
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023204794779648193,
      "loss": 2.3763,
      "theoretical_loss": 3.4479363009274406,
      "tokens_seen": 1881997312
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023203139776811044,
      "loss": 2.372,
      "theoretical_loss": 3.4479261701991497,
      "tokens_seen": 1882062848
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023201484773973898,
      "loss": 2.8353,
      "theoretical_loss": 3.4479160399223887,
      "tokens_seen": 1882128384
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023199829771136752,
      "loss": 2.7258,
      "theoretical_loss": 3.4479059100971217,
      "tokens_seen": 1882193920
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023198174768299603,
      "loss": 2.5045,
      "theoretical_loss": 3.4478957807233135,
      "tokens_seen": 1882259456
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023196519765462457,
      "loss": 2.3052,
      "theoretical_loss": 3.447885651800928,
      "tokens_seen": 1882324992
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023194864762625308,
      "loss": 2.5105,
      "theoretical_loss": 3.4478755233299294,
      "tokens_seen": 1882390528
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023193209759788161,
      "loss": 2.5045,
      "theoretical_loss": 3.4478653953102816,
      "tokens_seen": 1882456064
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023191554756951012,
      "loss": 2.6168,
      "theoretical_loss": 3.4478552677419487,
      "tokens_seen": 1882521600
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023189899754113866,
      "loss": 2.6573,
      "theoretical_loss": 3.4478451406248953,
      "tokens_seen": 1882587136
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023188244751276717,
      "loss": 2.4584,
      "theoretical_loss": 3.4478350139590854,
      "tokens_seen": 1882652672
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002318658974843957,
      "loss": 2.7653,
      "theoretical_loss": 3.447824887744483,
      "tokens_seen": 1882718208
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023184934745602422,
      "loss": 2.59,
      "theoretical_loss": 3.447814761981053,
      "tokens_seen": 1882783744
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023183279742765276,
      "loss": 2.5507,
      "theoretical_loss": 3.447804636668759,
      "tokens_seen": 1882849280
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023181624739928127,
      "loss": 2.333,
      "theoretical_loss": 3.4477945118075652,
      "tokens_seen": 1882914816
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002317996973709098,
      "loss": 2.7292,
      "theoretical_loss": 3.4477843873974354,
      "tokens_seen": 1882980352
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023178314734253832,
      "loss": 2.4882,
      "theoretical_loss": 3.447774263438335,
      "tokens_seen": 1883045888
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002317665973141668,
      "loss": 2.383,
      "theoretical_loss": 3.447764139930227,
      "tokens_seen": 1883111424
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023175004728579534,
      "loss": 2.5752,
      "theoretical_loss": 3.4477540168730765,
      "tokens_seen": 1883176960
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023173349725742385,
      "loss": 2.5838,
      "theoretical_loss": 3.4477438942668472,
      "tokens_seen": 1883242496
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023171694722905239,
      "loss": 2.7418,
      "theoretical_loss": 3.447733772111503,
      "tokens_seen": 1883308032
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002317003972006809,
      "loss": 2.5718,
      "theoretical_loss": 3.447723650407009,
      "tokens_seen": 1883373568
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023168384717230943,
      "loss": 2.6173,
      "theoretical_loss": 3.447713529153329,
      "tokens_seen": 1883439104
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023166729714393794,
      "loss": 2.7004,
      "theoretical_loss": 3.447703408350427,
      "tokens_seen": 1883504640
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023165074711556648,
      "loss": 2.5324,
      "theoretical_loss": 3.447693287998268,
      "tokens_seen": 1883570176
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000231634197087195,
      "loss": 2.5891,
      "theoretical_loss": 3.447683168096815,
      "tokens_seen": 1883635712
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023161764705882353,
      "loss": 2.5087,
      "theoretical_loss": 3.447673048646034,
      "tokens_seen": 1883701248
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023160109703045204,
      "loss": 2.257,
      "theoretical_loss": 3.4476629296458876,
      "tokens_seen": 1883766784
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023158454700208058,
      "loss": 2.405,
      "theoretical_loss": 3.447652811096341,
      "tokens_seen": 1883832320
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002315679969737091,
      "loss": 2.5602,
      "theoretical_loss": 3.447642692997358,
      "tokens_seen": 1883897856
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023155144694533763,
      "loss": 2.6121,
      "theoretical_loss": 3.4476325753489028,
      "tokens_seen": 1883963392
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023153489691696614,
      "loss": 2.5909,
      "theoretical_loss": 3.4476224581509403,
      "tokens_seen": 1884028928
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023151834688859467,
      "loss": 2.5784,
      "theoretical_loss": 3.447612341403434,
      "tokens_seen": 1884094464
    },
    {
      "epoch": 0.67,
      "objective/train/avg_token_score": 0.013022296130657196,
      "objective/train/avg_weight": 0.9846928715705872,
      "objective/train/docs_used": 1065366,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.598726272583008,
      "objective/train/std_weight": 0.05950305238366127,
      "objective/train/theoretical_loss": 3.4476022251063485,
      "objective/train/tokens_used": 1904620000,
      "theoretical_loss": 3.4476022251063485,
      "tokens_seen": 1884160000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023150179686022318,
      "loss": 2.4846,
      "theoretical_loss": 3.4476022251063485,
      "tokens_seen": 1884160000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023148524683185172,
      "loss": 2.5817,
      "theoretical_loss": 3.447592109259648,
      "tokens_seen": 1884225536
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023146869680348023,
      "loss": 2.7992,
      "theoretical_loss": 3.447581993863297,
      "tokens_seen": 1884291072
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023145214677510877,
      "loss": 2.3144,
      "theoretical_loss": 3.44757187891726,
      "tokens_seen": 1884356608
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023143559674673728,
      "loss": 2.2459,
      "theoretical_loss": 3.4475617644215006,
      "tokens_seen": 1884422144
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023141904671836582,
      "loss": 2.3249,
      "theoretical_loss": 3.447551650375984,
      "tokens_seen": 1884487680
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023140249668999433,
      "loss": 2.5052,
      "theoretical_loss": 3.4475415367806734,
      "tokens_seen": 1884553216
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023138594666162287,
      "loss": 2.2812,
      "theoretical_loss": 3.4475314236355343,
      "tokens_seen": 1884618752
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023136939663325138,
      "loss": 2.6644,
      "theoretical_loss": 3.44752131094053,
      "tokens_seen": 1884684288
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002313528466048799,
      "loss": 2.3306,
      "theoretical_loss": 3.447511198695625,
      "tokens_seen": 1884749824
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023133629657650842,
      "loss": 2.4819,
      "theoretical_loss": 3.4475010869007843,
      "tokens_seen": 1884815360
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023131974654813696,
      "loss": 2.6053,
      "theoretical_loss": 3.4474909755559713,
      "tokens_seen": 1884880896
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023130319651976547,
      "loss": 2.573,
      "theoretical_loss": 3.4474808646611512,
      "tokens_seen": 1884946432
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000231286646491394,
      "loss": 2.4144,
      "theoretical_loss": 3.4474707542162877,
      "tokens_seen": 1885011968
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002312700964630225,
      "loss": 2.5439,
      "theoretical_loss": 3.447460644221345,
      "tokens_seen": 1885077504
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000231253546434651,
      "loss": 2.6599,
      "theoretical_loss": 3.447450534676288,
      "tokens_seen": 1885143040
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023123699640627954,
      "loss": 2.7428,
      "theoretical_loss": 3.447440425581081,
      "tokens_seen": 1885208576
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023122044637790805,
      "loss": 2.6799,
      "theoretical_loss": 3.447430316935688,
      "tokens_seen": 1885274112
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002312038963495366,
      "loss": 2.6161,
      "theoretical_loss": 3.4474202087400734,
      "tokens_seen": 1885339648
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002311873463211651,
      "loss": 2.625,
      "theoretical_loss": 3.447410100994202,
      "tokens_seen": 1885405184
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023117079629279364,
      "loss": 2.5873,
      "theoretical_loss": 3.4473999936980375,
      "tokens_seen": 1885470720
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023115424626442215,
      "loss": 2.5149,
      "theoretical_loss": 3.4473898868515445,
      "tokens_seen": 1885536256
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023113769623605068,
      "loss": 2.6005,
      "theoretical_loss": 3.447379780454688,
      "tokens_seen": 1885601792
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002311211462076792,
      "loss": 2.5257,
      "theoretical_loss": 3.4473696745074314,
      "tokens_seen": 1885667328
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023110459617930773,
      "loss": 2.5752,
      "theoretical_loss": 3.4473595690097394,
      "tokens_seen": 1885732864
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023108804615093624,
      "loss": 2.6178,
      "theoretical_loss": 3.4473494639615767,
      "tokens_seen": 1885798400
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023107149612256478,
      "loss": 2.5051,
      "theoretical_loss": 3.447339359362907,
      "tokens_seen": 1885863936
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002310549460941933,
      "loss": 2.6749,
      "theoretical_loss": 3.447329255213696,
      "tokens_seen": 1885929472
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023103839606582183,
      "loss": 2.5403,
      "theoretical_loss": 3.4473191515139066,
      "tokens_seen": 1885995008
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023102184603745034,
      "loss": 2.505,
      "theoretical_loss": 3.447309048263504,
      "tokens_seen": 1886060544
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023100529600907888,
      "loss": 2.6749,
      "theoretical_loss": 3.4472989454624523,
      "tokens_seen": 1886126080
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002309887459807074,
      "loss": 2.5942,
      "theoretical_loss": 3.447288843110716,
      "tokens_seen": 1886191616
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023097219595233592,
      "loss": 2.5475,
      "theoretical_loss": 3.4472787412082595,
      "tokens_seen": 1886257152
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023095564592396443,
      "loss": 2.6421,
      "theoretical_loss": 3.4472686397550474,
      "tokens_seen": 1886322688
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023093909589559297,
      "loss": 2.4594,
      "theoretical_loss": 3.4472585387510435,
      "tokens_seen": 1886388224
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023092254586722148,
      "loss": 2.6623,
      "theoretical_loss": 3.447248438196213,
      "tokens_seen": 1886453760
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023090599583885002,
      "loss": 2.3791,
      "theoretical_loss": 3.4472383380905205,
      "tokens_seen": 1886519296
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023088944581047853,
      "loss": 2.5802,
      "theoretical_loss": 3.4472282384339294,
      "tokens_seen": 1886584832
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023087289578210707,
      "loss": 2.823,
      "theoretical_loss": 3.4472181392264045,
      "tokens_seen": 1886650368
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023085634575373558,
      "loss": 2.4201,
      "theoretical_loss": 3.4472080404679106,
      "tokens_seen": 1886715904
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023083979572536412,
      "loss": 2.7196,
      "theoretical_loss": 3.447197942158412,
      "tokens_seen": 1886781440
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023082324569699263,
      "loss": 2.4724,
      "theoretical_loss": 3.447187844297873,
      "tokens_seen": 1886846976
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023080669566862116,
      "loss": 2.6963,
      "theoretical_loss": 3.447177746886258,
      "tokens_seen": 1886912512
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023079014564024965,
      "loss": 2.4919,
      "theoretical_loss": 3.4471676499235318,
      "tokens_seen": 1886978048
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023077359561187816,
      "loss": 2.7719,
      "theoretical_loss": 3.4471575534096583,
      "tokens_seen": 1887043584
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002307570455835067,
      "loss": 2.1875,
      "theoretical_loss": 3.4471474573446024,
      "tokens_seen": 1887109120
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023074049555513523,
      "loss": 2.5719,
      "theoretical_loss": 3.4471373617283287,
      "tokens_seen": 1887174656
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023072394552676374,
      "loss": 2.7005,
      "theoretical_loss": 3.447127266560801,
      "tokens_seen": 1887240192
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023070739549839228,
      "loss": 2.7527,
      "theoretical_loss": 3.4471171718419846,
      "tokens_seen": 1887305728
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002306908454700208,
      "loss": 2.552,
      "theoretical_loss": 3.447107077571843,
      "tokens_seen": 1887371264
    },
    {
      "epoch": 0.67,
      "objective/train/avg_token_score": 0.020627381280064583,
      "objective/train/avg_weight": 0.9757266640663147,
      "objective/train/docs_used": 1066806,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.493518352508545,
      "objective/train/std_weight": 0.08975379168987274,
      "objective/train/theoretical_loss": 3.447096983750342,
      "objective/train/tokens_used": 1907896800,
      "theoretical_loss": 3.447096983750342,
      "tokens_seen": 1887436800
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023067429544164933,
      "loss": 2.3929,
      "theoretical_loss": 3.447096983750342,
      "tokens_seen": 1887436800
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023065774541327784,
      "loss": 2.6971,
      "theoretical_loss": 3.447086890377445,
      "tokens_seen": 1887502336
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023064119538490638,
      "loss": 2.5673,
      "theoretical_loss": 3.447076797453117,
      "tokens_seen": 1887567872
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002306246453565349,
      "loss": 2.6051,
      "theoretical_loss": 3.4470667049773214,
      "tokens_seen": 1887633408
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023060809532816343,
      "loss": 2.7078,
      "theoretical_loss": 3.4470566129500244,
      "tokens_seen": 1887698944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023059154529979194,
      "loss": 3.1608,
      "theoretical_loss": 3.4470465213711896,
      "tokens_seen": 1887764480
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023057499527142047,
      "loss": 2.3337,
      "theoretical_loss": 3.4470364302407814,
      "tokens_seen": 1887830016
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023055844524304898,
      "loss": 2.5279,
      "theoretical_loss": 3.447026339558765,
      "tokens_seen": 1887895552
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023054189521467752,
      "loss": 2.6819,
      "theoretical_loss": 3.4470162493251038,
      "tokens_seen": 1887961088
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023052534518630603,
      "loss": 2.6096,
      "theoretical_loss": 3.4470061595397627,
      "tokens_seen": 1888026624
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023050879515793457,
      "loss": 2.7323,
      "theoretical_loss": 3.446996070202707,
      "tokens_seen": 1888092160
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023049224512956308,
      "loss": 2.6098,
      "theoretical_loss": 3.446985981313901,
      "tokens_seen": 1888157696
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023047569510119162,
      "loss": 2.5806,
      "theoretical_loss": 3.4469758928733083,
      "tokens_seen": 1888223232
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023045914507282013,
      "loss": 2.6048,
      "theoretical_loss": 3.446965804880894,
      "tokens_seen": 1888288768
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023044259504444867,
      "loss": 2.4268,
      "theoretical_loss": 3.446955717336623,
      "tokens_seen": 1888354304
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023042604501607718,
      "loss": 2.819,
      "theoretical_loss": 3.4469456302404597,
      "tokens_seen": 1888419840
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002304094949877057,
      "loss": 2.1678,
      "theoretical_loss": 3.4469355435923683,
      "tokens_seen": 1888485376
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023039294495933422,
      "loss": 2.5348,
      "theoretical_loss": 3.4469254573923136,
      "tokens_seen": 1888550912
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023037639493096276,
      "loss": 2.6302,
      "theoretical_loss": 3.4469153716402596,
      "tokens_seen": 1888616448
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023035984490259127,
      "loss": 2.6862,
      "theoretical_loss": 3.4469052863361713,
      "tokens_seen": 1888681984
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002303432948742198,
      "loss": 2.388,
      "theoretical_loss": 3.4468952014800136,
      "tokens_seen": 1888747520
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023032674484584832,
      "loss": 2.5756,
      "theoretical_loss": 3.4468851170717505,
      "tokens_seen": 1888813056
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023031019481747686,
      "loss": 2.5696,
      "theoretical_loss": 3.446875033111347,
      "tokens_seen": 1888878592
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023029364478910534,
      "loss": 2.4697,
      "theoretical_loss": 3.4468649495987678,
      "tokens_seen": 1888944128
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023027709476073385,
      "loss": 2.6152,
      "theoretical_loss": 3.4468548665339767,
      "tokens_seen": 1889009664
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002302605447323624,
      "loss": 2.7005,
      "theoretical_loss": 3.446844783916939,
      "tokens_seen": 1889075200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002302439947039909,
      "loss": 2.3916,
      "theoretical_loss": 3.4468347017476186,
      "tokens_seen": 1889140736
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023022744467561944,
      "loss": 2.3835,
      "theoretical_loss": 3.446824620025981,
      "tokens_seen": 1889206272
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023021089464724795,
      "loss": 2.7235,
      "theoretical_loss": 3.44681453875199,
      "tokens_seen": 1889271808
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023019434461887648,
      "loss": 3.9144,
      "theoretical_loss": 3.4468044579256105,
      "tokens_seen": 1889337344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000230177794590505,
      "loss": 2.6537,
      "theoretical_loss": 3.446794377546807,
      "tokens_seen": 1889402880
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023016124456213353,
      "loss": -0.4327,
      "theoretical_loss": 3.4467842976155447,
      "tokens_seen": 1889468416
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023014469453376204,
      "loss": 2.4594,
      "theoretical_loss": 3.4467742181317873,
      "tokens_seen": 1889533952
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023012814450539058,
      "loss": 2.7425,
      "theoretical_loss": 3.4467641390955,
      "tokens_seen": 1889599488
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002301115944770191,
      "loss": 2.4613,
      "theoretical_loss": 3.446754060506647,
      "tokens_seen": 1889665024
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023009504444864763,
      "loss": 2.5471,
      "theoretical_loss": 3.4467439823651933,
      "tokens_seen": 1889730560
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023007849442027614,
      "loss": 2.9438,
      "theoretical_loss": 3.4467339046711034,
      "tokens_seen": 1889796096
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00023006194439190468,
      "loss": 2.7125,
      "theoretical_loss": 3.446723827424342,
      "tokens_seen": 1889861632
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002300453943635332,
      "loss": 0.164,
      "theoretical_loss": 3.4467137506248733,
      "tokens_seen": 1889927168
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00023002884433516172,
      "loss": 2.7526,
      "theoretical_loss": 3.446703674272663,
      "tokens_seen": 1889992704
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00023001229430679023,
      "loss": 2.6887,
      "theoretical_loss": 3.4466935983676743,
      "tokens_seen": 1890058240
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022999574427841877,
      "loss": 2.5239,
      "theoretical_loss": 3.446683522909873,
      "tokens_seen": 1890123776
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022997919425004728,
      "loss": 2.6657,
      "theoretical_loss": 3.446673447899223,
      "tokens_seen": 1890189312
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022996264422167582,
      "loss": 2.5604,
      "theoretical_loss": 3.4466633733356895,
      "tokens_seen": 1890254848
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022994609419330433,
      "loss": -3.034,
      "theoretical_loss": 3.4466532992192365,
      "tokens_seen": 1890320384
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022992954416493287,
      "loss": 2.7309,
      "theoretical_loss": 3.4466432255498294,
      "tokens_seen": 1890385920
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022991299413656138,
      "loss": 2.6239,
      "theoretical_loss": 3.4466331523274327,
      "tokens_seen": 1890451456
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022989644410818992,
      "loss": 2.6823,
      "theoretical_loss": 3.446623079552011,
      "tokens_seen": 1890516992
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022987989407981843,
      "loss": 4.2611,
      "theoretical_loss": 3.4466130072235286,
      "tokens_seen": 1890582528
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022986334405144696,
      "loss": 2.8941,
      "theoretical_loss": 3.4466029353419505,
      "tokens_seen": 1890648064
    },
    {
      "epoch": 0.68,
      "objective/train/avg_token_score": 0.008219189941883087,
      "objective/train/avg_weight": 0.9903259873390198,
      "objective/train/docs_used": 1068619,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8046422004699707,
      "objective/train/std_weight": 0.055815793573856354,
      "objective/train/theoretical_loss": 3.4465928639072416,
      "objective/train/tokens_used": 1911173600,
      "theoretical_loss": 3.4465928639072416,
      "tokens_seen": 1890713600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022984679402307547,
      "loss": 2.5734,
      "theoretical_loss": 3.4465928639072416,
      "tokens_seen": 1890713600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000229830243994704,
      "loss": 2.6087,
      "theoretical_loss": 3.446582792919366,
      "tokens_seen": 1890779136
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022981369396633252,
      "loss": 2.5474,
      "theoretical_loss": 3.446572722378289,
      "tokens_seen": 1890844672
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000229797143937961,
      "loss": 2.6154,
      "theoretical_loss": 3.4465626522839745,
      "tokens_seen": 1890910208
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022978059390958954,
      "loss": 2.6492,
      "theoretical_loss": 3.4465525826363885,
      "tokens_seen": 1890975744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022976404388121805,
      "loss": 2.3647,
      "theoretical_loss": 3.4465425134354946,
      "tokens_seen": 1891041280
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002297474938528466,
      "loss": 2.5114,
      "theoretical_loss": 3.4465324446812575,
      "tokens_seen": 1891106816
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002297309438244751,
      "loss": 2.5721,
      "theoretical_loss": 3.4465223763736423,
      "tokens_seen": 1891172352
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022971439379610364,
      "loss": 2.601,
      "theoretical_loss": 3.446512308512614,
      "tokens_seen": 1891237888
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022969784376773215,
      "loss": 2.5271,
      "theoretical_loss": 3.446502241098137,
      "tokens_seen": 1891303424
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002296812937393607,
      "loss": 2.6141,
      "theoretical_loss": 3.4464921741301757,
      "tokens_seen": 1891368960
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002296647437109892,
      "loss": 2.9379,
      "theoretical_loss": 3.4464821076086953,
      "tokens_seen": 1891434496
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022964819368261774,
      "loss": 2.7573,
      "theoretical_loss": 3.44647204153366,
      "tokens_seen": 1891500032
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022963164365424625,
      "loss": 2.6195,
      "theoretical_loss": 3.4464619759050352,
      "tokens_seen": 1891565568
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022961509362587478,
      "loss": 2.5956,
      "theoretical_loss": 3.4464519107227853,
      "tokens_seen": 1891631104
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002295985435975033,
      "loss": 2.3362,
      "theoretical_loss": 3.4464418459868753,
      "tokens_seen": 1891696640
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022958199356913183,
      "loss": 2.6578,
      "theoretical_loss": 3.446431781697269,
      "tokens_seen": 1891762176
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022956544354076034,
      "loss": 2.7319,
      "theoretical_loss": 3.4464217178539327,
      "tokens_seen": 1891827712
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022954889351238888,
      "loss": 2.5609,
      "theoretical_loss": 3.4464116544568295,
      "tokens_seen": 1891893248
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002295323434840174,
      "loss": 2.7651,
      "theoretical_loss": 3.446401591505925,
      "tokens_seen": 1891958784
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022951579345564593,
      "loss": 2.4682,
      "theoretical_loss": 3.4463915290011844,
      "tokens_seen": 1892024320
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022949924342727444,
      "loss": 2.5768,
      "theoretical_loss": 3.446381466942572,
      "tokens_seen": 1892089856
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022948269339890298,
      "loss": 2.5194,
      "theoretical_loss": 3.4463714053300523,
      "tokens_seen": 1892155392
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022946614337053149,
      "loss": 2.5254,
      "theoretical_loss": 3.44636134416359,
      "tokens_seen": 1892220928
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022944959334216002,
      "loss": 2.931,
      "theoretical_loss": 3.4463512834431507,
      "tokens_seen": 1892286464
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022943304331378853,
      "loss": 2.4338,
      "theoretical_loss": 3.446341223168698,
      "tokens_seen": 1892352000
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022941649328541707,
      "loss": 2.3533,
      "theoretical_loss": 3.446331163340198,
      "tokens_seen": 1892417536
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022939994325704558,
      "loss": 2.6168,
      "theoretical_loss": 3.446321103957615,
      "tokens_seen": 1892483072
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022938339322867412,
      "loss": 2.6644,
      "theoretical_loss": 3.4463110450209133,
      "tokens_seen": 1892548608
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022936684320030263,
      "loss": 2.6027,
      "theoretical_loss": 3.4463009865300585,
      "tokens_seen": 1892614144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022935029317193117,
      "loss": 2.4082,
      "theoretical_loss": 3.4462909284850145,
      "tokens_seen": 1892679680
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022933374314355968,
      "loss": 2.3361,
      "theoretical_loss": 3.4462808708857464,
      "tokens_seen": 1892745216
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002293171931151882,
      "loss": 2.5705,
      "theoretical_loss": 3.4462708137322196,
      "tokens_seen": 1892810752
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002293006430868167,
      "loss": 2.673,
      "theoretical_loss": 3.446260757024398,
      "tokens_seen": 1892876288
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022928409305844524,
      "loss": 2.4743,
      "theoretical_loss": 3.4462507007622474,
      "tokens_seen": 1892941824
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022926754303007375,
      "loss": 2.406,
      "theoretical_loss": 3.4462406449457315,
      "tokens_seen": 1893007360
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022925099300170228,
      "loss": 2.7795,
      "theoretical_loss": 3.4462305895748164,
      "tokens_seen": 1893072896
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002292344429733308,
      "loss": 2.6035,
      "theoretical_loss": 3.4462205346494654,
      "tokens_seen": 1893138432
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022921789294495933,
      "loss": 2.5213,
      "theoretical_loss": 3.446210480169645,
      "tokens_seen": 1893203968
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022920134291658784,
      "loss": 2.6284,
      "theoretical_loss": 3.446200426135319,
      "tokens_seen": 1893269504
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022918479288821638,
      "loss": 2.7279,
      "theoretical_loss": 3.4461903725464524,
      "tokens_seen": 1893335040
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002291682428598449,
      "loss": 2.6226,
      "theoretical_loss": 3.4461803194030103,
      "tokens_seen": 1893400576
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022915169283147343,
      "loss": 2.6417,
      "theoretical_loss": 3.446170266704957,
      "tokens_seen": 1893466112
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022913514280310194,
      "loss": 2.7275,
      "theoretical_loss": 3.446160214452258,
      "tokens_seen": 1893531648
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022911859277473048,
      "loss": 2.6995,
      "theoretical_loss": 3.4461501626448783,
      "tokens_seen": 1893597184
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022910204274635899,
      "loss": 2.6631,
      "theoretical_loss": 3.446140111282782,
      "tokens_seen": 1893662720
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022908549271798752,
      "loss": 2.7946,
      "theoretical_loss": 3.4461300603659337,
      "tokens_seen": 1893728256
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022906894268961603,
      "loss": 2.5774,
      "theoretical_loss": 3.4461200098942997,
      "tokens_seen": 1893793792
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022905239266124457,
      "loss": 2.505,
      "theoretical_loss": 3.4461099598678437,
      "tokens_seen": 1893859328
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022903584263287308,
      "loss": 2.7357,
      "theoretical_loss": 3.446099910286531,
      "tokens_seen": 1893924864
    },
    {
      "epoch": 0.68,
      "objective/train/avg_token_score": 0.02136770263314247,
      "objective/train/avg_weight": 0.9748478531837463,
      "objective/train/docs_used": 1071183,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.168220043182373,
      "objective/train/std_weight": 0.09433850646018982,
      "objective/train/theoretical_loss": 3.446089861150326,
      "objective/train/tokens_used": 1914450400,
      "theoretical_loss": 3.446089861150326,
      "tokens_seen": 1893990400
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022901929260450162,
      "loss": 2.357,
      "theoretical_loss": 3.446089861150326,
      "tokens_seen": 1893990400
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022900274257613013,
      "loss": 2.5297,
      "theoretical_loss": 3.4460798124591947,
      "tokens_seen": 1894055936
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022898619254775867,
      "loss": 2.6729,
      "theoretical_loss": 3.446069764213101,
      "tokens_seen": 1894121472
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022896964251938718,
      "loss": 2.4948,
      "theoretical_loss": 3.4460597164120097,
      "tokens_seen": 1894187008
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022895309249101572,
      "loss": 2.6672,
      "theoretical_loss": 3.4460496690558866,
      "tokens_seen": 1894252544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022893654246264423,
      "loss": 2.2172,
      "theoretical_loss": 3.4460396221446956,
      "tokens_seen": 1894318080
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022891999243427276,
      "loss": 2.4895,
      "theoretical_loss": 3.4460295756784025,
      "tokens_seen": 1894383616
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022890344240590127,
      "loss": 2.6256,
      "theoretical_loss": 3.446019529656972,
      "tokens_seen": 1894449152
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002288868923775298,
      "loss": 2.2244,
      "theoretical_loss": 3.446009484080368,
      "tokens_seen": 1894514688
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022887034234915832,
      "loss": 2.5572,
      "theoretical_loss": 3.445999438948557,
      "tokens_seen": 1894580224
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022885379232078686,
      "loss": 2.7097,
      "theoretical_loss": 3.4459893942615025,
      "tokens_seen": 1894645760
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022883724229241537,
      "loss": 2.7907,
      "theoretical_loss": 3.4459793500191704,
      "tokens_seen": 1894711296
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022882069226404385,
      "loss": 2.6656,
      "theoretical_loss": 3.4459693062215253,
      "tokens_seen": 1894776832
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002288041422356724,
      "loss": 2.8202,
      "theoretical_loss": 3.445959262868532,
      "tokens_seen": 1894842368
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002287875922073009,
      "loss": 2.8342,
      "theoretical_loss": 3.4459492199601556,
      "tokens_seen": 1894907904
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022877104217892944,
      "loss": 2.3066,
      "theoretical_loss": 3.445939177496361,
      "tokens_seen": 1894973440
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022875449215055795,
      "loss": 2.493,
      "theoretical_loss": 3.445929135477113,
      "tokens_seen": 1895038976
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002287379421221865,
      "loss": 2.3926,
      "theoretical_loss": 3.445919093902377,
      "tokens_seen": 1895104512
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000228721392093815,
      "loss": 2.2468,
      "theoretical_loss": 3.4459090527721177,
      "tokens_seen": 1895170048
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022870484206544354,
      "loss": 2.5166,
      "theoretical_loss": 3.4458990120863,
      "tokens_seen": 1895235584
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022868829203707205,
      "loss": 2.2856,
      "theoretical_loss": 3.445888971844888,
      "tokens_seen": 1895301120
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022867174200870058,
      "loss": 2.5341,
      "theoretical_loss": 3.4458789320478487,
      "tokens_seen": 1895366656
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002286551919803291,
      "loss": 2.3294,
      "theoretical_loss": 3.4458688926951453,
      "tokens_seen": 1895432192
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022863864195195763,
      "loss": 2.6365,
      "theoretical_loss": 3.4458588537867434,
      "tokens_seen": 1895497728
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022862209192358614,
      "loss": 2.8341,
      "theoretical_loss": 3.4458488153226083,
      "tokens_seen": 1895563264
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022860554189521468,
      "loss": 2.4565,
      "theoretical_loss": 3.445838777302704,
      "tokens_seen": 1895628800
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002285889918668432,
      "loss": 2.4472,
      "theoretical_loss": 3.4458287397269967,
      "tokens_seen": 1895694336
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022857244183847173,
      "loss": 2.675,
      "theoretical_loss": 3.4458187025954503,
      "tokens_seen": 1895759872
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022855589181010024,
      "loss": 2.8821,
      "theoretical_loss": 3.4458086659080305,
      "tokens_seen": 1895825408
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022853934178172878,
      "loss": 2.6541,
      "theoretical_loss": 3.4457986296647025,
      "tokens_seen": 1895890944
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022852279175335729,
      "loss": 2.6711,
      "theoretical_loss": 3.4457885938654305,
      "tokens_seen": 1895956480
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022850624172498582,
      "loss": 2.4221,
      "theoretical_loss": 3.44577855851018,
      "tokens_seen": 1896022016
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022848969169661433,
      "loss": 2.6164,
      "theoretical_loss": 3.4457685235989155,
      "tokens_seen": 1896087552
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022847314166824287,
      "loss": 2.5754,
      "theoretical_loss": 3.445758489131603,
      "tokens_seen": 1896153088
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022845659163987138,
      "loss": 2.7598,
      "theoretical_loss": 3.4457484551082063,
      "tokens_seen": 1896218624
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022844004161149992,
      "loss": 2.8202,
      "theoretical_loss": 3.4457384215286915,
      "tokens_seen": 1896284160
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022842349158312843,
      "loss": 2.6281,
      "theoretical_loss": 3.4457283883930234,
      "tokens_seen": 1896349696
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022840694155475697,
      "loss": 2.6685,
      "theoretical_loss": 3.4457183557011666,
      "tokens_seen": 1896415232
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022839039152638548,
      "loss": 2.6156,
      "theoretical_loss": 3.4457083234530863,
      "tokens_seen": 1896480768
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022837384149801401,
      "loss": 2.5977,
      "theoretical_loss": 3.445698291648747,
      "tokens_seen": 1896546304
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022835729146964253,
      "loss": 2.752,
      "theoretical_loss": 3.445688260288115,
      "tokens_seen": 1896611840
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022834074144127106,
      "loss": 2.107,
      "theoretical_loss": 3.445678229371154,
      "tokens_seen": 1896677376
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022832419141289955,
      "loss": 2.5993,
      "theoretical_loss": 3.4456681988978306,
      "tokens_seen": 1896742912
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022830764138452806,
      "loss": 2.4779,
      "theoretical_loss": 3.4456581688681087,
      "tokens_seen": 1896808448
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002282910913561566,
      "loss": 2.7929,
      "theoretical_loss": 3.4456481392819533,
      "tokens_seen": 1896873984
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002282745413277851,
      "loss": 2.5481,
      "theoretical_loss": 3.44563811013933,
      "tokens_seen": 1896939520
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022825799129941364,
      "loss": 2.7222,
      "theoretical_loss": 3.445628081440203,
      "tokens_seen": 1897005056
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022824144127104215,
      "loss": 2.5344,
      "theoretical_loss": 3.445618053184539,
      "tokens_seen": 1897070592
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002282248912426707,
      "loss": 2.5873,
      "theoretical_loss": 3.4456080253723016,
      "tokens_seen": 1897136128
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002282083412142992,
      "loss": 2.7682,
      "theoretical_loss": 3.445597998003456,
      "tokens_seen": 1897201664
    },
    {
      "epoch": 0.68,
      "objective/train/avg_token_score": 0.017528459429740906,
      "objective/train/avg_weight": 0.9794043898582458,
      "objective/train/docs_used": 1073215,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.4678752422332764,
      "objective/train/std_weight": 0.09451679140329361,
      "objective/train/theoretical_loss": 3.445587971077968,
      "objective/train/tokens_used": 1917727200,
      "theoretical_loss": 3.445587971077968,
      "tokens_seen": 1897267200
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022819179118592774,
      "loss": 2.6223,
      "theoretical_loss": 3.445587971077968,
      "tokens_seen": 1897267200
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022817524115755625,
      "loss": 2.4819,
      "theoretical_loss": 3.4455779445958026,
      "tokens_seen": 1897332736
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022815869112918479,
      "loss": 2.7609,
      "theoretical_loss": 3.4455679185569243,
      "tokens_seen": 1897398272
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002281421411008133,
      "loss": 2.6834,
      "theoretical_loss": 3.4455578929612987,
      "tokens_seen": 1897463808
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022812559107244183,
      "loss": 2.7502,
      "theoretical_loss": 3.44554786780889,
      "tokens_seen": 1897529344
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022810904104407034,
      "loss": 2.4205,
      "theoretical_loss": 3.4455378430996646,
      "tokens_seen": 1897594880
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022809249101569888,
      "loss": 2.7723,
      "theoretical_loss": 3.445527818833587,
      "tokens_seen": 1897660416
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002280759409873274,
      "loss": 2.6674,
      "theoretical_loss": 3.4455177950106215,
      "tokens_seen": 1897725952
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022805939095895593,
      "loss": 2.7805,
      "theoretical_loss": 3.445507771630735,
      "tokens_seen": 1897791488
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022804284093058444,
      "loss": 2.6577,
      "theoretical_loss": 3.4454977486938914,
      "tokens_seen": 1897857024
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022802629090221298,
      "loss": 2.1571,
      "theoretical_loss": 3.4454877262000556,
      "tokens_seen": 1897922560
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002280097408738415,
      "loss": 2.525,
      "theoretical_loss": 3.4454777041491935,
      "tokens_seen": 1897988096
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022799319084547003,
      "loss": 2.3771,
      "theoretical_loss": 3.44546768254127,
      "tokens_seen": 1898053632
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022797664081709856,
      "loss": 2.5823,
      "theoretical_loss": 3.44545766137625,
      "tokens_seen": 1898119168
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022796009078872707,
      "loss": 2.9049,
      "theoretical_loss": 3.4454476406540984,
      "tokens_seen": 1898184704
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002279435407603556,
      "loss": 2.4942,
      "theoretical_loss": 3.445437620374781,
      "tokens_seen": 1898250240
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022792699073198412,
      "loss": 2.6303,
      "theoretical_loss": 3.445427600538263,
      "tokens_seen": 1898315776
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022791044070361266,
      "loss": 2.6557,
      "theoretical_loss": 3.4454175811445085,
      "tokens_seen": 1898381312
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022789389067524117,
      "loss": 2.1955,
      "theoretical_loss": 3.445407562193484,
      "tokens_seen": 1898446848
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002278773406468697,
      "loss": 2.7178,
      "theoretical_loss": 3.4453975436851536,
      "tokens_seen": 1898512384
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022786079061849822,
      "loss": 2.5246,
      "theoretical_loss": 3.445387525619483,
      "tokens_seen": 1898577920
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022784424059012676,
      "loss": 2.1063,
      "theoretical_loss": 3.445377507996437,
      "tokens_seen": 1898643456
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022782769056175524,
      "loss": 2.6261,
      "theoretical_loss": 3.445367490815981,
      "tokens_seen": 1898708992
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022781114053338375,
      "loss": 2.4,
      "theoretical_loss": 3.4453574740780804,
      "tokens_seen": 1898774528
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002277945905050123,
      "loss": 2.4975,
      "theoretical_loss": 3.4453474577826997,
      "tokens_seen": 1898840064
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002277780404766408,
      "loss": 2.4651,
      "theoretical_loss": 3.445337441929805,
      "tokens_seen": 1898905600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022776149044826933,
      "loss": 2.5311,
      "theoretical_loss": 3.445327426519361,
      "tokens_seen": 1898971136
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022774494041989785,
      "loss": 2.5231,
      "theoretical_loss": 3.4453174115513323,
      "tokens_seen": 1899036672
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022772839039152638,
      "loss": 2.5886,
      "theoretical_loss": 3.445307397025685,
      "tokens_seen": 1899102208
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002277118403631549,
      "loss": 2.4851,
      "theoretical_loss": 3.445297382942384,
      "tokens_seen": 1899167744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022769529033478343,
      "loss": 2.4396,
      "theoretical_loss": 3.4452873693013943,
      "tokens_seen": 1899233280
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022767874030641194,
      "loss": 2.4552,
      "theoretical_loss": 3.4452773561026815,
      "tokens_seen": 1899298816
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022766219027804048,
      "loss": 2.4777,
      "theoretical_loss": 3.4452673433462104,
      "tokens_seen": 1899364352
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000227645640249669,
      "loss": 2.4697,
      "theoretical_loss": 3.4452573310319465,
      "tokens_seen": 1899429888
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022762909022129753,
      "loss": 2.6437,
      "theoretical_loss": 3.4452473191598547,
      "tokens_seen": 1899495424
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022761254019292604,
      "loss": 2.515,
      "theoretical_loss": 3.4452373077299003,
      "tokens_seen": 1899560960
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022759599016455457,
      "loss": 6.7522,
      "theoretical_loss": 3.4452272967420488,
      "tokens_seen": 1899626496
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022757944013618309,
      "loss": 2.5314,
      "theoretical_loss": 3.4452172861962653,
      "tokens_seen": 1899692032
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022756289010781162,
      "loss": 2.4802,
      "theoretical_loss": 3.445207276092515,
      "tokens_seen": 1899757568
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022754634007944013,
      "loss": 2.517,
      "theoretical_loss": 3.445197266430763,
      "tokens_seen": 1899823104
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022752979005106867,
      "loss": 2.4189,
      "theoretical_loss": 3.4451872572109745,
      "tokens_seen": 1899888640
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022751324002269718,
      "loss": 2.84,
      "theoretical_loss": 3.445177248433115,
      "tokens_seen": 1899954176
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022749668999432572,
      "loss": 2.5396,
      "theoretical_loss": 3.44516724009715,
      "tokens_seen": 1900019712
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022748013996595423,
      "loss": 2.5156,
      "theoretical_loss": 3.445157232203044,
      "tokens_seen": 1900085248
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022746358993758277,
      "loss": 2.6089,
      "theoretical_loss": 3.4451472247507624,
      "tokens_seen": 1900150784
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022744703990921128,
      "loss": 2.5733,
      "theoretical_loss": 3.445137217740271,
      "tokens_seen": 1900216320
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022743048988083981,
      "loss": 2.4351,
      "theoretical_loss": 3.4451272111715348,
      "tokens_seen": 1900281856
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022741393985246833,
      "loss": 2.5085,
      "theoretical_loss": 3.445117205044519,
      "tokens_seen": 1900347392
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022739738982409686,
      "loss": 2.5179,
      "theoretical_loss": 3.445107199359189,
      "tokens_seen": 1900412928
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022738083979572537,
      "loss": 2.4525,
      "theoretical_loss": 3.44509719411551,
      "tokens_seen": 1900478464
    },
    {
      "epoch": 0.68,
      "objective/train/avg_token_score": 0.018475664779543877,
      "objective/train/avg_weight": 0.9782536625862122,
      "objective/train/docs_used": 1074511,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.740579128265381,
      "objective/train/std_weight": 0.08577968925237656,
      "objective/train/theoretical_loss": 3.4450871893134467,
      "objective/train/tokens_used": 1921004000,
      "theoretical_loss": 3.4450871893134467,
      "tokens_seen": 1900544000
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002273642897673539,
      "loss": 2.5564,
      "theoretical_loss": 3.4450871893134467,
      "tokens_seen": 1900544000
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002273477397389824,
      "loss": 2.6725,
      "theoretical_loss": 3.4450771849529653,
      "tokens_seen": 1900609536
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002273311897106109,
      "loss": 2.635,
      "theoretical_loss": 3.4450671810340308,
      "tokens_seen": 1900675072
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022731463968223944,
      "loss": 2.7162,
      "theoretical_loss": 3.4450571775566083,
      "tokens_seen": 1900740608
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022729808965386795,
      "loss": 2.8983,
      "theoretical_loss": 3.445047174520663,
      "tokens_seen": 1900806144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002272815396254965,
      "loss": 2.5431,
      "theoretical_loss": 3.4450371719261605,
      "tokens_seen": 1900871680
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000227264989597125,
      "loss": 2.6604,
      "theoretical_loss": 3.4450271697730663,
      "tokens_seen": 1900937216
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022724843956875354,
      "loss": 2.3604,
      "theoretical_loss": 3.445017168061345,
      "tokens_seen": 1901002752
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022723188954038205,
      "loss": 2.6141,
      "theoretical_loss": 3.4450071667909623,
      "tokens_seen": 1901068288
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022721533951201059,
      "loss": 2.5169,
      "theoretical_loss": 3.4449971659618837,
      "tokens_seen": 1901133824
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002271987894836391,
      "loss": 2.5755,
      "theoretical_loss": 3.4449871655740743,
      "tokens_seen": 1901199360
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022718223945526763,
      "loss": 2.5168,
      "theoretical_loss": 3.444977165627499,
      "tokens_seen": 1901264896
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022716568942689614,
      "loss": 2.5558,
      "theoretical_loss": 3.444967166122124,
      "tokens_seen": 1901330432
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022714913939852468,
      "loss": 2.4433,
      "theoretical_loss": 3.444957167057914,
      "tokens_seen": 1901395968
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002271325893701532,
      "loss": 2.537,
      "theoretical_loss": 3.444947168434835,
      "tokens_seen": 1901461504
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022711603934178173,
      "loss": 2.6052,
      "theoretical_loss": 3.4449371702528513,
      "tokens_seen": 1901527040
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022709948931341024,
      "loss": 2.6802,
      "theoretical_loss": 3.444927172511929,
      "tokens_seen": 1901592576
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022708293928503878,
      "loss": 2.6202,
      "theoretical_loss": 3.4449171752120336,
      "tokens_seen": 1901658112
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002270663892566673,
      "loss": 2.5685,
      "theoretical_loss": 3.4449071783531293,
      "tokens_seen": 1901723648
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022704983922829583,
      "loss": 2.6253,
      "theoretical_loss": 3.4448971819351826,
      "tokens_seen": 1901789184
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022703328919992434,
      "loss": 2.4411,
      "theoretical_loss": 3.4448871859581587,
      "tokens_seen": 1901854720
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022701673917155287,
      "loss": 2.61,
      "theoretical_loss": 3.444877190422023,
      "tokens_seen": 1901920256
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022700018914318138,
      "loss": 2.4625,
      "theoretical_loss": 3.44486719532674,
      "tokens_seen": 1901985792
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022698363911480992,
      "loss": 2.7031,
      "theoretical_loss": 3.444857200672276,
      "tokens_seen": 1902051328
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022696708908643843,
      "loss": 2.7112,
      "theoretical_loss": 3.4448472064585958,
      "tokens_seen": 1902116864
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022695053905806697,
      "loss": 2.3233,
      "theoretical_loss": 3.4448372126856652,
      "tokens_seen": 1902182400
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022693398902969548,
      "loss": 2.7876,
      "theoretical_loss": 3.4448272193534493,
      "tokens_seen": 1902247936
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022691743900132402,
      "loss": 2.5862,
      "theoretical_loss": 3.4448172264619137,
      "tokens_seen": 1902313472
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022690088897295253,
      "loss": 2.642,
      "theoretical_loss": 3.444807234011024,
      "tokens_seen": 1902379008
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022688433894458107,
      "loss": 2.4889,
      "theoretical_loss": 3.4447972420007447,
      "tokens_seen": 1902444544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022686778891620958,
      "loss": 2.425,
      "theoretical_loss": 3.444787250431042,
      "tokens_seen": 1902510080
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022685123888783806,
      "loss": 2.7114,
      "theoretical_loss": 3.444777259301881,
      "tokens_seen": 1902575616
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002268346888594666,
      "loss": 2.6554,
      "theoretical_loss": 3.4447672686132274,
      "tokens_seen": 1902641152
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002268181388310951,
      "loss": 2.3865,
      "theoretical_loss": 3.4447572783650457,
      "tokens_seen": 1902706688
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022680158880272365,
      "loss": 2.4426,
      "theoretical_loss": 3.444747288557303,
      "tokens_seen": 1902772224
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022678503877435216,
      "loss": 2.7074,
      "theoretical_loss": 3.444737299189963,
      "tokens_seen": 1902837760
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002267684887459807,
      "loss": 2.6048,
      "theoretical_loss": 3.444727310262992,
      "tokens_seen": 1902903296
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022675193871760923,
      "loss": 2.448,
      "theoretical_loss": 3.444717321776355,
      "tokens_seen": 1902968832
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022673538868923774,
      "loss": 2.6063,
      "theoretical_loss": 3.444707333730018,
      "tokens_seen": 1903034368
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022671883866086628,
      "loss": 2.5221,
      "theoretical_loss": 3.444697346123946,
      "tokens_seen": 1903099904
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002267022886324948,
      "loss": 2.2068,
      "theoretical_loss": 3.444687358958104,
      "tokens_seen": 1903165440
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022668573860412333,
      "loss": 2.5332,
      "theoretical_loss": 3.4446773722324586,
      "tokens_seen": 1903230976
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022666918857575184,
      "loss": 2.4398,
      "theoretical_loss": 3.4446673859469747,
      "tokens_seen": 1903296512
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022665263854738037,
      "loss": 2.3662,
      "theoretical_loss": 3.444657400101617,
      "tokens_seen": 1903362048
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022663608851900888,
      "loss": 2.217,
      "theoretical_loss": 3.444647414696352,
      "tokens_seen": 1903427584
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022661953849063742,
      "loss": 2.3805,
      "theoretical_loss": 3.444637429731145,
      "tokens_seen": 1903493120
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022660298846226593,
      "loss": 2.5803,
      "theoretical_loss": 3.4446274452059606,
      "tokens_seen": 1903558656
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022658643843389447,
      "loss": 2.707,
      "theoretical_loss": 3.444617461120765,
      "tokens_seen": 1903624192
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022656988840552298,
      "loss": 2.5622,
      "theoretical_loss": 3.444607477475524,
      "tokens_seen": 1903689728
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022655333837715152,
      "loss": 2.7968,
      "theoretical_loss": 3.4445974942702025,
      "tokens_seen": 1903755264
    },
    {
      "epoch": 0.68,
      "objective/train/avg_token_score": 0.009469980373978615,
      "objective/train/avg_weight": 0.9888530373573303,
      "objective/train/docs_used": 1076404,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7698051929473877,
      "objective/train/std_weight": 0.06839723140001297,
      "objective/train/theoretical_loss": 3.4445875115047655,
      "objective/train/tokens_used": 1924280800,
      "theoretical_loss": 3.4445875115047655,
      "tokens_seen": 1903820800
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022653678834878003,
      "loss": 2.6653,
      "theoretical_loss": 3.4445875115047655,
      "tokens_seen": 1903820800
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022652023832040857,
      "loss": 2.5962,
      "theoretical_loss": 3.4445775291791794,
      "tokens_seen": 1903886336
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022650368829203708,
      "loss": 2.6521,
      "theoretical_loss": 3.444567547293409,
      "tokens_seen": 1903951872
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022648713826366561,
      "loss": 2.3084,
      "theoretical_loss": 3.444557565847421,
      "tokens_seen": 1904017408
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022647058823529412,
      "loss": 2.4204,
      "theoretical_loss": 3.444547584841179,
      "tokens_seen": 1904082944
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022645403820692266,
      "loss": 2.1416,
      "theoretical_loss": 3.4445376042746503,
      "tokens_seen": 1904148480
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022643748817855117,
      "loss": 2.4519,
      "theoretical_loss": 3.444527624147799,
      "tokens_seen": 1904214016
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002264209381501797,
      "loss": 2.8705,
      "theoretical_loss": 3.444517644460591,
      "tokens_seen": 1904279552
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022640438812180822,
      "loss": 2.3612,
      "theoretical_loss": 3.4445076652129925,
      "tokens_seen": 1904345088
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022638783809343676,
      "loss": 2.5077,
      "theoretical_loss": 3.4444976864049686,
      "tokens_seen": 1904410624
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022637128806506527,
      "loss": 2.5753,
      "theoretical_loss": 3.444487708036484,
      "tokens_seen": 1904476160
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022635473803669375,
      "loss": 2.7188,
      "theoretical_loss": 3.444477730107505,
      "tokens_seen": 1904541696
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002263381880083223,
      "loss": 2.144,
      "theoretical_loss": 3.4444677526179976,
      "tokens_seen": 1904607232
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002263216379799508,
      "loss": 2.3768,
      "theoretical_loss": 3.444457775567927,
      "tokens_seen": 1904672768
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022630508795157934,
      "loss": 2.7395,
      "theoretical_loss": 3.444447798957258,
      "tokens_seen": 1904738304
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022628853792320785,
      "loss": 2.3219,
      "theoretical_loss": 3.4444378227859564,
      "tokens_seen": 1904803840
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022627198789483639,
      "loss": 2.6195,
      "theoretical_loss": 3.4444278470539884,
      "tokens_seen": 1904869376
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002262554378664649,
      "loss": 2.5282,
      "theoretical_loss": 3.4444178717613188,
      "tokens_seen": 1904934912
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022623888783809343,
      "loss": 2.1198,
      "theoretical_loss": 3.4444078969079133,
      "tokens_seen": 1905000448
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022622233780972194,
      "loss": 2.3888,
      "theoretical_loss": 3.4443979224937378,
      "tokens_seen": 1905065984
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022620578778135048,
      "loss": 2.5523,
      "theoretical_loss": 3.4443879485187576,
      "tokens_seen": 1905131520
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000226189237752979,
      "loss": 2.5007,
      "theoretical_loss": 3.4443779749829386,
      "tokens_seen": 1905197056
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022617268772460753,
      "loss": 2.1649,
      "theoretical_loss": 3.4443680018862457,
      "tokens_seen": 1905262592
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022615613769623604,
      "loss": 2.4378,
      "theoretical_loss": 3.4443580292286446,
      "tokens_seen": 1905328128
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022613958766786458,
      "loss": 2.798,
      "theoretical_loss": 3.4443480570101017,
      "tokens_seen": 1905393664
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002261230376394931,
      "loss": 2.4746,
      "theoretical_loss": 3.444338085230582,
      "tokens_seen": 1905459200
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022610648761112163,
      "loss": 2.2705,
      "theoretical_loss": 3.44432811389005,
      "tokens_seen": 1905524736
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022608993758275014,
      "loss": 2.5904,
      "theoretical_loss": 3.444318142988473,
      "tokens_seen": 1905590272
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022607338755437867,
      "loss": 2.4576,
      "theoretical_loss": 3.444308172525816,
      "tokens_seen": 1905655808
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022605683752600718,
      "loss": 2.5893,
      "theoretical_loss": 3.444298202502044,
      "tokens_seen": 1905721344
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022604028749763572,
      "loss": 2.3959,
      "theoretical_loss": 3.444288232917123,
      "tokens_seen": 1905786880
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022602373746926423,
      "loss": 2.5553,
      "theoretical_loss": 3.444278263771019,
      "tokens_seen": 1905852416
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022600718744089277,
      "loss": 2.4404,
      "theoretical_loss": 3.444268295063697,
      "tokens_seen": 1905917952
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022599063741252128,
      "loss": 2.3356,
      "theoretical_loss": 3.444258326795123,
      "tokens_seen": 1905983488
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022597408738414982,
      "loss": 2.7014,
      "theoretical_loss": 3.444248358965263,
      "tokens_seen": 1906049024
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022595753735577833,
      "loss": 2.6361,
      "theoretical_loss": 3.444238391574081,
      "tokens_seen": 1906114560
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022594098732740687,
      "loss": 2.3567,
      "theoretical_loss": 3.4442284246215444,
      "tokens_seen": 1906180096
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022592443729903538,
      "loss": 2.5138,
      "theoretical_loss": 3.444218458107618,
      "tokens_seen": 1906245632
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002259078872706639,
      "loss": 2.7485,
      "theoretical_loss": 3.444208492032267,
      "tokens_seen": 1906311168
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022589133724229242,
      "loss": -2.6166,
      "theoretical_loss": 3.444198526395458,
      "tokens_seen": 1906376704
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002258747872139209,
      "loss": 2.7052,
      "theoretical_loss": 3.444188561197156,
      "tokens_seen": 1906442240
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022585823718554944,
      "loss": 2.5092,
      "theoretical_loss": 3.4441785964373266,
      "tokens_seen": 1906507776
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022584168715717796,
      "loss": 2.6096,
      "theoretical_loss": 3.444168632115936,
      "tokens_seen": 1906573312
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002258251371288065,
      "loss": -6.936,
      "theoretical_loss": 3.444158668232949,
      "tokens_seen": 1906638848
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000225808587100435,
      "loss": 2.431,
      "theoretical_loss": 3.444148704788332,
      "tokens_seen": 1906704384
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022579203707206354,
      "loss": 5.2728,
      "theoretical_loss": 3.44413874178205,
      "tokens_seen": 1906769920
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022577548704369205,
      "loss": 2.6901,
      "theoretical_loss": 3.4441287792140693,
      "tokens_seen": 1906835456
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002257589370153206,
      "loss": 2.4706,
      "theoretical_loss": 3.4441188170843553,
      "tokens_seen": 1906900992
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002257423869869491,
      "loss": 2.3381,
      "theoretical_loss": 3.444108855392874,
      "tokens_seen": 1906966528
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022572583695857764,
      "loss": 2.6027,
      "theoretical_loss": 3.44409889413959,
      "tokens_seen": 1907032064
    },
    {
      "epoch": 0.68,
      "objective/train/avg_token_score": 0.023262998089194298,
      "objective/train/avg_weight": 0.9726163744926453,
      "objective/train/docs_used": 1078284,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.247208118438721,
      "objective/train/std_weight": 0.12948466837406158,
      "objective/train/theoretical_loss": 3.4440889333244695,
      "objective/train/tokens_used": 1927557600,
      "theoretical_loss": 3.4440889333244695,
      "tokens_seen": 1907097600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022570928693020615,
      "loss": 2.5163,
      "theoretical_loss": 3.4440889333244695,
      "tokens_seen": 1907097600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022569273690183468,
      "loss": 2.5302,
      "theoretical_loss": 3.444078972947479,
      "tokens_seen": 1907163136
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002256761868734632,
      "loss": 2.3701,
      "theoretical_loss": 3.4440690130085825,
      "tokens_seen": 1907228672
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022565963684509173,
      "loss": 2.5512,
      "theoretical_loss": 3.4440590535077478,
      "tokens_seen": 1907294208
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022564308681672024,
      "loss": 2.518,
      "theoretical_loss": 3.4440490944449387,
      "tokens_seen": 1907359744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022562653678834878,
      "loss": 2.4636,
      "theoretical_loss": 3.444039135820122,
      "tokens_seen": 1907425280
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002256099867599773,
      "loss": 2.5083,
      "theoretical_loss": 3.444029177633263,
      "tokens_seen": 1907490816
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022559343673160583,
      "loss": 2.5009,
      "theoretical_loss": 3.4440192198843276,
      "tokens_seen": 1907556352
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022557688670323434,
      "loss": 2.564,
      "theoretical_loss": 3.4440092625732808,
      "tokens_seen": 1907621888
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022556033667486288,
      "loss": 2.4434,
      "theoretical_loss": 3.443999305700089,
      "tokens_seen": 1907687424
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002255437866464914,
      "loss": 2.5957,
      "theoretical_loss": 3.443989349264718,
      "tokens_seen": 1907752960
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022552723661811992,
      "loss": 2.6056,
      "theoretical_loss": 3.4439793932671323,
      "tokens_seen": 1907818496
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022551068658974844,
      "loss": 2.5615,
      "theoretical_loss": 3.4439694377073,
      "tokens_seen": 1907884032
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022549413656137697,
      "loss": 2.3913,
      "theoretical_loss": 3.4439594825851843,
      "tokens_seen": 1907949568
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022547758653300548,
      "loss": 2.7902,
      "theoretical_loss": 3.4439495279007524,
      "tokens_seen": 1908015104
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022546103650463402,
      "loss": 2.4179,
      "theoretical_loss": 3.443939573653969,
      "tokens_seen": 1908080640
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022544448647626253,
      "loss": 2.1667,
      "theoretical_loss": 3.443929619844801,
      "tokens_seen": 1908146176
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022542793644789107,
      "loss": 2.3393,
      "theoretical_loss": 3.4439196664732137,
      "tokens_seen": 1908211712
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022541138641951958,
      "loss": 2.5129,
      "theoretical_loss": 3.4439097135391723,
      "tokens_seen": 1908277248
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022539483639114812,
      "loss": 2.1538,
      "theoretical_loss": 3.443899761042643,
      "tokens_seen": 1908342784
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002253782863627766,
      "loss": 2.4493,
      "theoretical_loss": 3.4438898089835916,
      "tokens_seen": 1908408320
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022536173633440514,
      "loss": 2.5148,
      "theoretical_loss": 3.443879857361984,
      "tokens_seen": 1908473856
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022534518630603365,
      "loss": 2.5767,
      "theoretical_loss": 3.4438699061777855,
      "tokens_seen": 1908539392
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022532863627766219,
      "loss": 2.5003,
      "theoretical_loss": 3.4438599554309617,
      "tokens_seen": 1908604928
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002253120862492907,
      "loss": 2.5056,
      "theoretical_loss": 3.4438500051214787,
      "tokens_seen": 1908670464
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022529553622091923,
      "loss": 2.6787,
      "theoretical_loss": 3.4438400552493027,
      "tokens_seen": 1908736000
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022527898619254774,
      "loss": 2.5901,
      "theoretical_loss": 3.4438301058143987,
      "tokens_seen": 1908801536
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022526243616417628,
      "loss": 2.7166,
      "theoretical_loss": 3.443820156816733,
      "tokens_seen": 1908867072
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002252458861358048,
      "loss": 2.5236,
      "theoretical_loss": 3.443810208256271,
      "tokens_seen": 1908932608
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022522933610743333,
      "loss": 2.5243,
      "theoretical_loss": 3.4438002601329787,
      "tokens_seen": 1908998144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022521278607906184,
      "loss": 2.4356,
      "theoretical_loss": 3.443790312446822,
      "tokens_seen": 1909063680
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022519623605069038,
      "loss": 2.7075,
      "theoretical_loss": 3.4437803651977665,
      "tokens_seen": 1909129216
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002251796860223189,
      "loss": 2.5322,
      "theoretical_loss": 3.4437704183857782,
      "tokens_seen": 1909194752
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022516313599394743,
      "loss": 2.4516,
      "theoretical_loss": 3.443760472010822,
      "tokens_seen": 1909260288
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022514658596557594,
      "loss": 2.227,
      "theoretical_loss": 3.443750526072865,
      "tokens_seen": 1909325824
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022513003593720447,
      "loss": 2.463,
      "theoretical_loss": 3.4437405805718724,
      "tokens_seen": 1909391360
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022511348590883298,
      "loss": 2.6828,
      "theoretical_loss": 3.4437306355078094,
      "tokens_seen": 1909456896
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022509693588046152,
      "loss": 2.5611,
      "theoretical_loss": 3.4437206908806433,
      "tokens_seen": 1909522432
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022508038585209003,
      "loss": 2.7597,
      "theoretical_loss": 3.4437107466903383,
      "tokens_seen": 1909587968
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022506383582371857,
      "loss": 1.5622,
      "theoretical_loss": 3.443700802936861,
      "tokens_seen": 1909653504
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022504728579534708,
      "loss": 2.5606,
      "theoretical_loss": 3.443690859620178,
      "tokens_seen": 1909719040
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022503073576697562,
      "loss": 2.3832,
      "theoretical_loss": 3.443680916740253,
      "tokens_seen": 1909784576
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022501418573860413,
      "loss": 2.3735,
      "theoretical_loss": 3.443670974297054,
      "tokens_seen": 1909850112
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022499763571023267,
      "loss": 2.679,
      "theoretical_loss": 3.443661032290546,
      "tokens_seen": 1909915648
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022498108568186118,
      "loss": 2.4048,
      "theoretical_loss": 3.4436510907206945,
      "tokens_seen": 1909981184
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002249645356534897,
      "loss": 2.4475,
      "theoretical_loss": 3.4436411495874655,
      "tokens_seen": 1910046720
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022494798562511822,
      "loss": 2.4979,
      "theoretical_loss": 3.4436312088908254,
      "tokens_seen": 1910112256
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022493143559674676,
      "loss": 2.3607,
      "theoretical_loss": 3.4436212686307393,
      "tokens_seen": 1910177792
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022491488556837527,
      "loss": 2.4448,
      "theoretical_loss": 3.4436113288071732,
      "tokens_seen": 1910243328
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002248983355400038,
      "loss": 2.5681,
      "theoretical_loss": 3.4436013894200936,
      "tokens_seen": 1910308864
    },
    {
      "epoch": 0.68,
      "objective/train/avg_token_score": 0.015743957832455635,
      "objective/train/avg_weight": 0.9814659357070923,
      "objective/train/docs_used": 1079917,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.942629814147949,
      "objective/train/std_weight": 0.062404271215200424,
      "objective/train/theoretical_loss": 3.443591450469466,
      "objective/train/tokens_used": 1930834400,
      "theoretical_loss": 3.443591450469466,
      "tokens_seen": 1910374400
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002248817855116323,
      "loss": 2.4376,
      "theoretical_loss": 3.443591450469466,
      "tokens_seen": 1910374400
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002248652354832608,
      "loss": 2.6778,
      "theoretical_loss": 3.4435815119552555,
      "tokens_seen": 1910439936
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022484868545488934,
      "loss": 2.2629,
      "theoretical_loss": 3.443571573877429,
      "tokens_seen": 1910505472
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022483213542651785,
      "loss": 2.6209,
      "theoretical_loss": 3.443561636235952,
      "tokens_seen": 1910571008
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002248155853981464,
      "loss": 2.6218,
      "theoretical_loss": 3.4435516990307904,
      "tokens_seen": 1910636544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002247990353697749,
      "loss": 2.5906,
      "theoretical_loss": 3.44354176226191,
      "tokens_seen": 1910702080
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022478248534140344,
      "loss": 2.7927,
      "theoretical_loss": 3.4435318259292766,
      "tokens_seen": 1910767616
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022476593531303195,
      "loss": 2.5172,
      "theoretical_loss": 3.443521890032856,
      "tokens_seen": 1910833152
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022474938528466048,
      "loss": 2.6036,
      "theoretical_loss": 3.4435119545726147,
      "tokens_seen": 1910898688
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000224732835256289,
      "loss": 2.388,
      "theoretical_loss": 3.443502019548518,
      "tokens_seen": 1910964224
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022471628522791753,
      "loss": 2.4212,
      "theoretical_loss": 3.4434920849605324,
      "tokens_seen": 1911029760
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022469973519954604,
      "loss": 2.509,
      "theoretical_loss": 3.443482150808623,
      "tokens_seen": 1911095296
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022468318517117458,
      "loss": 2.5489,
      "theoretical_loss": 3.443472217092756,
      "tokens_seen": 1911160832
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002246666351428031,
      "loss": 2.3635,
      "theoretical_loss": 3.443462283812898,
      "tokens_seen": 1911226368
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022465008511443163,
      "loss": 2.638,
      "theoretical_loss": 3.4434523509690136,
      "tokens_seen": 1911291904
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022463353508606014,
      "loss": 2.5896,
      "theoretical_loss": 3.44344241856107,
      "tokens_seen": 1911357440
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022461698505768868,
      "loss": 2.4699,
      "theoretical_loss": 3.4434324865890322,
      "tokens_seen": 1911422976
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002246004350293172,
      "loss": 2.4707,
      "theoretical_loss": 3.4434225550528668,
      "tokens_seen": 1911488512
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022458388500094572,
      "loss": 2.4449,
      "theoretical_loss": 3.443412623952539,
      "tokens_seen": 1911554048
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022456733497257423,
      "loss": 2.5813,
      "theoretical_loss": 3.4434026932880153,
      "tokens_seen": 1911619584
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022455078494420277,
      "loss": 2.4469,
      "theoretical_loss": 3.4433927630592613,
      "tokens_seen": 1911685120
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022453423491583128,
      "loss": 2.6368,
      "theoretical_loss": 3.4433828332662437,
      "tokens_seen": 1911750656
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022451768488745982,
      "loss": 2.4006,
      "theoretical_loss": 3.4433729039089274,
      "tokens_seen": 1911816192
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022450113485908833,
      "loss": 2.1657,
      "theoretical_loss": 3.443362974987279,
      "tokens_seen": 1911881728
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022448458483071687,
      "loss": 2.54,
      "theoretical_loss": 3.4433530465012643,
      "tokens_seen": 1911947264
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022446803480234538,
      "loss": 2.3743,
      "theoretical_loss": 3.443343118450849,
      "tokens_seen": 1912012800
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022445148477397392,
      "loss": 2.4246,
      "theoretical_loss": 3.4433331908359994,
      "tokens_seen": 1912078336
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022443493474560243,
      "loss": 2.5503,
      "theoretical_loss": 3.4433232636566817,
      "tokens_seen": 1912143872
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022441838471723096,
      "loss": 2.4436,
      "theoretical_loss": 3.443313336912861,
      "tokens_seen": 1912209408
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022440183468885945,
      "loss": 2.5366,
      "theoretical_loss": 3.443303410604504,
      "tokens_seen": 1912274944
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022438528466048796,
      "loss": 2.7372,
      "theoretical_loss": 3.4432934847315764,
      "tokens_seen": 1912340480
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002243687346321165,
      "loss": 2.6192,
      "theoretical_loss": 3.443283559294044,
      "tokens_seen": 1912406016
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000224352184603745,
      "loss": 2.5541,
      "theoretical_loss": 3.4432736342918737,
      "tokens_seen": 1912471552
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022433563457537354,
      "loss": 2.5997,
      "theoretical_loss": 3.44326370972503,
      "tokens_seen": 1912537088
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022431908454700205,
      "loss": 2.4764,
      "theoretical_loss": 3.4432537855934804,
      "tokens_seen": 1912602624
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002243025345186306,
      "loss": 2.5714,
      "theoretical_loss": 3.44324386189719,
      "tokens_seen": 1912668160
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002242859844902591,
      "loss": 2.4665,
      "theoretical_loss": 3.4432339386361246,
      "tokens_seen": 1912733696
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022426943446188764,
      "loss": 2.1679,
      "theoretical_loss": 3.4432240158102507,
      "tokens_seen": 1912799232
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022425288443351615,
      "loss": 2.36,
      "theoretical_loss": 3.443214093419534,
      "tokens_seen": 1912864768
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002242363344051447,
      "loss": 2.4315,
      "theoretical_loss": 3.4432041714639414,
      "tokens_seen": 1912930304
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002242197843767732,
      "loss": 2.8845,
      "theoretical_loss": 3.4431942499434376,
      "tokens_seen": 1912995840
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022420323434840174,
      "loss": 2.5827,
      "theoretical_loss": 3.4431843288579893,
      "tokens_seen": 1913061376
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022418668432003025,
      "loss": -2.684,
      "theoretical_loss": 3.443174408207563,
      "tokens_seen": 1913126912
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022417013429165878,
      "loss": 2.6044,
      "theoretical_loss": 3.4431644879921235,
      "tokens_seen": 1913192448
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002241535842632873,
      "loss": 2.6768,
      "theoretical_loss": 3.4431545682116376,
      "tokens_seen": 1913257984
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022413703423491583,
      "loss": 2.3963,
      "theoretical_loss": 3.4431446488660713,
      "tokens_seen": 1913323520
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022412048420654434,
      "loss": 4.1436,
      "theoretical_loss": 3.44313472995539,
      "tokens_seen": 1913389056
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022410393417817288,
      "loss": 2.9729,
      "theoretical_loss": 3.443124811479561,
      "tokens_seen": 1913454592
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002240873841498014,
      "loss": -6.346,
      "theoretical_loss": 3.44311489343855,
      "tokens_seen": 1913520128
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022407083412142993,
      "loss": 2.7744,
      "theoretical_loss": 3.443104975832322,
      "tokens_seen": 1913585664
    },
    {
      "epoch": 0.68,
      "objective/train/avg_token_score": 0.018198732286691666,
      "objective/train/avg_weight": 0.9785753488540649,
      "objective/train/docs_used": 1081633,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.6086950302124023,
      "objective/train/std_weight": 0.11135104298591614,
      "objective/train/theoretical_loss": 3.4430950586608438,
      "objective/train/tokens_used": 1934111200,
      "theoretical_loss": 3.4430950586608438,
      "tokens_seen": 1913651200
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022405428409305844,
      "loss": 2.3576,
      "theoretical_loss": 3.4430950586608438,
      "tokens_seen": 1913651200
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022403773406468698,
      "loss": 2.7445,
      "theoretical_loss": 3.4430851419240818,
      "tokens_seen": 1913716736
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022402118403631549,
      "loss": 2.6565,
      "theoretical_loss": 3.4430752256220014,
      "tokens_seen": 1913782272
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022400463400794402,
      "loss": 2.5881,
      "theoretical_loss": 3.443065309754569,
      "tokens_seen": 1913847808
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022398808397957256,
      "loss": 2.4315,
      "theoretical_loss": 3.4430553943217506,
      "tokens_seen": 1913913344
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022397153395120107,
      "loss": 2.6923,
      "theoretical_loss": 3.443045479323512,
      "tokens_seen": 1913978880
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002239549839228296,
      "loss": 2.593,
      "theoretical_loss": 3.4430355647598194,
      "tokens_seen": 1914044416
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022393843389445812,
      "loss": 2.6954,
      "theoretical_loss": 3.4430256506306396,
      "tokens_seen": 1914109952
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022392188386608666,
      "loss": 2.6462,
      "theoretical_loss": 3.443015736935938,
      "tokens_seen": 1914175488
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022390533383771514,
      "loss": 2.2718,
      "theoretical_loss": 3.4430058236756804,
      "tokens_seen": 1914241024
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022388878380934365,
      "loss": 2.5679,
      "theoretical_loss": 3.4429959108498336,
      "tokens_seen": 1914306560
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002238722337809722,
      "loss": 2.6137,
      "theoretical_loss": 3.4429859984583633,
      "tokens_seen": 1914372096
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002238556837526007,
      "loss": 2.4339,
      "theoretical_loss": 3.4429760865012353,
      "tokens_seen": 1914437632
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022383913372422924,
      "loss": 2.6035,
      "theoretical_loss": 3.442966174978416,
      "tokens_seen": 1914503168
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022382258369585775,
      "loss": 2.5651,
      "theoretical_loss": 3.4429562638898723,
      "tokens_seen": 1914568704
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022380603366748628,
      "loss": 2.5656,
      "theoretical_loss": 3.442946353235569,
      "tokens_seen": 1914634240
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002237894836391148,
      "loss": 2.3875,
      "theoretical_loss": 3.4429364430154727,
      "tokens_seen": 1914699776
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022377293361074333,
      "loss": 2.5205,
      "theoretical_loss": 3.44292653322955,
      "tokens_seen": 1914765312
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022375638358237184,
      "loss": 2.3562,
      "theoretical_loss": 3.4429166238777658,
      "tokens_seen": 1914830848
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022373983355400038,
      "loss": 2.3868,
      "theoretical_loss": 3.442906714960088,
      "tokens_seen": 1914896384
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002237232835256289,
      "loss": 2.3684,
      "theoretical_loss": 3.4428968064764813,
      "tokens_seen": 1914961920
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022370673349725743,
      "loss": 2.5036,
      "theoretical_loss": 3.442886898426912,
      "tokens_seen": 1915027456
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022369018346888594,
      "loss": 2.5515,
      "theoretical_loss": 3.442876990811347,
      "tokens_seen": 1915092992
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022367363344051448,
      "loss": 2.5269,
      "theoretical_loss": 3.4428670836297517,
      "tokens_seen": 1915158528
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000223657083412143,
      "loss": 2.3881,
      "theoretical_loss": 3.442857176882093,
      "tokens_seen": 1915224064
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022364053338377152,
      "loss": 2.6897,
      "theoretical_loss": 3.442847270568336,
      "tokens_seen": 1915289600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022362398335540003,
      "loss": 2.3725,
      "theoretical_loss": 3.4428373646884474,
      "tokens_seen": 1915355136
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022360743332702857,
      "loss": 2.4972,
      "theoretical_loss": 3.4428274592423938,
      "tokens_seen": 1915420672
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022359088329865708,
      "loss": 2.4293,
      "theoretical_loss": 3.4428175542301407,
      "tokens_seen": 1915486208
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022357433327028562,
      "loss": 2.4952,
      "theoretical_loss": 3.442807649651654,
      "tokens_seen": 1915551744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022355778324191413,
      "loss": 2.6499,
      "theoretical_loss": 3.4427977455069008,
      "tokens_seen": 1915617280
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022354123321354267,
      "loss": 2.5884,
      "theoretical_loss": 3.4427878417958464,
      "tokens_seen": 1915682816
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022352468318517118,
      "loss": 2.592,
      "theoretical_loss": 3.442777938518458,
      "tokens_seen": 1915748352
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022350813315679972,
      "loss": 2.3914,
      "theoretical_loss": 3.4427680356747006,
      "tokens_seen": 1915813888
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022349158312842823,
      "loss": 2.569,
      "theoretical_loss": 3.442758133264541,
      "tokens_seen": 1915879424
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022347503310005676,
      "loss": 2.7163,
      "theoretical_loss": 3.442748231287946,
      "tokens_seen": 1915944960
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022345848307168527,
      "loss": 2.5839,
      "theoretical_loss": 3.4427383297448806,
      "tokens_seen": 1916010496
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002234419330433138,
      "loss": 2.8085,
      "theoretical_loss": 3.4427284286353115,
      "tokens_seen": 1916076032
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022342538301494232,
      "loss": 2.5626,
      "theoretical_loss": 3.442718527959205,
      "tokens_seen": 1916141568
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002234088329865708,
      "loss": 2.7009,
      "theoretical_loss": 3.4427086277165273,
      "tokens_seen": 1916207104
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022339228295819934,
      "loss": 2.3885,
      "theoretical_loss": 3.442698727907244,
      "tokens_seen": 1916272640
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022337573292982785,
      "loss": 2.7676,
      "theoretical_loss": 3.442688828531322,
      "tokens_seen": 1916338176
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002233591829014564,
      "loss": 2.6153,
      "theoretical_loss": 3.4426789295887277,
      "tokens_seen": 1916403712
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002233426328730849,
      "loss": 2.5985,
      "theoretical_loss": 3.4426690310794266,
      "tokens_seen": 1916469248
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022332608284471344,
      "loss": 2.4421,
      "theoretical_loss": 3.442659133003385,
      "tokens_seen": 1916534784
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022330953281634195,
      "loss": 2.2155,
      "theoretical_loss": 3.4426492353605695,
      "tokens_seen": 1916600320
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002232929827879705,
      "loss": 2.4959,
      "theoretical_loss": 3.442639338150946,
      "tokens_seen": 1916665856
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000223276432759599,
      "loss": 2.5689,
      "theoretical_loss": 3.4426294413744816,
      "tokens_seen": 1916731392
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022325988273122754,
      "loss": 2.682,
      "theoretical_loss": 3.4426195450311416,
      "tokens_seen": 1916796928
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022324333270285605,
      "loss": 2.6716,
      "theoretical_loss": 3.442609649120892,
      "tokens_seen": 1916862464
    },
    {
      "epoch": 0.68,
      "objective/train/avg_token_score": 0.01851547509431839,
      "objective/train/avg_weight": 0.9782037138938904,
      "objective/train/docs_used": 1083428,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.916212558746338,
      "objective/train/std_weight": 0.11976942420005798,
      "objective/train/theoretical_loss": 3.4425997536437,
      "objective/train/tokens_used": 1937388000,
      "theoretical_loss": 3.4425997536437,
      "tokens_seen": 1916928000
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022322678267448458,
      "loss": 2.505,
      "theoretical_loss": 3.4425997536437,
      "tokens_seen": 1916928000
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002232102326461131,
      "loss": 2.7154,
      "theoretical_loss": 3.4425898585995314,
      "tokens_seen": 1916993536
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022319368261774163,
      "loss": 2.5905,
      "theoretical_loss": 3.4425799639883525,
      "tokens_seen": 1917059072
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022317713258937014,
      "loss": 2.6392,
      "theoretical_loss": 3.442570069810129,
      "tokens_seen": 1917124608
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022316058256099868,
      "loss": 2.6521,
      "theoretical_loss": 3.442560176064828,
      "tokens_seen": 1917190144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002231440325326272,
      "loss": 2.6119,
      "theoretical_loss": 3.442550282752415,
      "tokens_seen": 1917255680
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022312748250425573,
      "loss": 2.6655,
      "theoretical_loss": 3.4425403898728573,
      "tokens_seen": 1917321216
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022311093247588424,
      "loss": 2.3775,
      "theoretical_loss": 3.44253049742612,
      "tokens_seen": 1917386752
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022309438244751278,
      "loss": 2.7689,
      "theoretical_loss": 3.44252060541217,
      "tokens_seen": 1917452288
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022307783241914129,
      "loss": 2.3856,
      "theoretical_loss": 3.4425107138309734,
      "tokens_seen": 1917517824
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022306128239076982,
      "loss": 2.2928,
      "theoretical_loss": 3.4425008226824967,
      "tokens_seen": 1917583360
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022304473236239833,
      "loss": 2.8671,
      "theoretical_loss": 3.442490931966706,
      "tokens_seen": 1917648896
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022302818233402687,
      "loss": 2.6505,
      "theoretical_loss": 3.4424810416835676,
      "tokens_seen": 1917714432
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022301163230565538,
      "loss": 2.739,
      "theoretical_loss": 3.4424711518330477,
      "tokens_seen": 1917779968
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022299508227728392,
      "loss": 2.6572,
      "theoretical_loss": 3.442461262415113,
      "tokens_seen": 1917845504
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00022297853224891243,
      "loss": 2.5184,
      "theoretical_loss": 3.442451373429729,
      "tokens_seen": 1917911040
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022296198222054097,
      "loss": 2.5152,
      "theoretical_loss": 3.4424414848768627,
      "tokens_seen": 1917976576
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022294543219216948,
      "loss": 3.0361,
      "theoretical_loss": 3.4424315967564807,
      "tokens_seen": 1918042112
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022292888216379802,
      "loss": 2.6922,
      "theoretical_loss": 3.4424217090685483,
      "tokens_seen": 1918107648
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002229123321354265,
      "loss": 2.4616,
      "theoretical_loss": 3.4424118218130326,
      "tokens_seen": 1918173184
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000222895782107055,
      "loss": 2.3494,
      "theoretical_loss": 3.442401934989899,
      "tokens_seen": 1918238720
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022287923207868355,
      "loss": 2.7903,
      "theoretical_loss": 3.442392048599115,
      "tokens_seen": 1918304256
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022286268205031206,
      "loss": 2.6064,
      "theoretical_loss": 3.442382162640646,
      "tokens_seen": 1918369792
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002228461320219406,
      "loss": 2.4935,
      "theoretical_loss": 3.442372277114459,
      "tokens_seen": 1918435328
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002228295819935691,
      "loss": 2.6958,
      "theoretical_loss": 3.44236239202052,
      "tokens_seen": 1918500864
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022281303196519764,
      "loss": 2.7591,
      "theoretical_loss": 3.4423525073587955,
      "tokens_seen": 1918566400
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022279648193682615,
      "loss": 2.5519,
      "theoretical_loss": 3.4423426231292518,
      "tokens_seen": 1918631936
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002227799319084547,
      "loss": 2.7994,
      "theoretical_loss": 3.442332739331855,
      "tokens_seen": 1918697472
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002227633818800832,
      "loss": 2.5032,
      "theoretical_loss": 3.4423228559665713,
      "tokens_seen": 1918763008
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022274683185171174,
      "loss": 2.7313,
      "theoretical_loss": 3.442312973033368,
      "tokens_seen": 1918828544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022273028182334028,
      "loss": 2.5847,
      "theoretical_loss": 3.4423030905322105,
      "tokens_seen": 1918894080
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022271373179496879,
      "loss": 2.4807,
      "theoretical_loss": 3.4422932084630653,
      "tokens_seen": 1918959616
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022269718176659732,
      "loss": 2.4782,
      "theoretical_loss": 3.442283326825899,
      "tokens_seen": 1919025152
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022268063173822583,
      "loss": 2.6246,
      "theoretical_loss": 3.4422734456206783,
      "tokens_seen": 1919090688
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022266408170985437,
      "loss": 2.5423,
      "theoretical_loss": 3.4422635648473685,
      "tokens_seen": 1919156224
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022264753168148288,
      "loss": 2.8115,
      "theoretical_loss": 3.442253684505937,
      "tokens_seen": 1919221760
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022263098165311142,
      "loss": 2.3585,
      "theoretical_loss": 3.4422438045963495,
      "tokens_seen": 1919287296
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022261443162473993,
      "loss": 2.8868,
      "theoretical_loss": 3.4422339251185727,
      "tokens_seen": 1919352832
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022259788159636847,
      "loss": 2.7399,
      "theoretical_loss": 3.4422240460725737,
      "tokens_seen": 1919418368
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022258133156799698,
      "loss": 2.5665,
      "theoretical_loss": 3.4422141674583173,
      "tokens_seen": 1919483904
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022256478153962552,
      "loss": 2.5759,
      "theoretical_loss": 3.4422042892757716,
      "tokens_seen": 1919549440
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022254823151125403,
      "loss": 2.3268,
      "theoretical_loss": 3.4421944115249015,
      "tokens_seen": 1919614976
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022253168148288256,
      "loss": 2.5499,
      "theoretical_loss": 3.442184534205674,
      "tokens_seen": 1919680512
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022251513145451107,
      "loss": 2.5076,
      "theoretical_loss": 3.4421746573180556,
      "tokens_seen": 1919746048
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002224985814261396,
      "loss": 2.4883,
      "theoretical_loss": 3.4421647808620133,
      "tokens_seen": 1919811584
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022248203139776812,
      "loss": 2.5718,
      "theoretical_loss": 3.442154904837512,
      "tokens_seen": 1919877120
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022246548136939666,
      "loss": 2.6446,
      "theoretical_loss": 3.4421450292445197,
      "tokens_seen": 1919942656
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022244893134102517,
      "loss": 2.3586,
      "theoretical_loss": 3.442135154083002,
      "tokens_seen": 1920008192
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022243238131265365,
      "loss": 2.535,
      "theoretical_loss": 3.4421252793529247,
      "tokens_seen": 1920073728
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002224158312842822,
      "loss": 2.4933,
      "theoretical_loss": 3.4421154050542553,
      "tokens_seen": 1920139264
    },
    {
      "epoch": 0.69,
      "objective/train/avg_token_score": 0.013901575468480587,
      "objective/train/avg_weight": 0.9836544990539551,
      "objective/train/docs_used": 1085055,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.798659324645996,
      "objective/train/std_weight": 0.05140009894967079,
      "objective/train/theoretical_loss": 3.4421055311869604,
      "objective/train/tokens_used": 1940664800,
      "theoretical_loss": 3.4421055311869604,
      "tokens_seen": 1920204800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002223992812559107,
      "loss": 2.7846,
      "theoretical_loss": 3.4421055311869604,
      "tokens_seen": 1920204800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022238273122753924,
      "loss": 2.7963,
      "theoretical_loss": 3.442095657751006,
      "tokens_seen": 1920270336
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022236618119916775,
      "loss": 2.6998,
      "theoretical_loss": 3.4420857847463573,
      "tokens_seen": 1920335872
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002223496311707963,
      "loss": 2.6664,
      "theoretical_loss": 3.442075912172983,
      "tokens_seen": 1920401408
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002223330811424248,
      "loss": 2.6547,
      "theoretical_loss": 3.4420660400308476,
      "tokens_seen": 1920466944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022231653111405334,
      "loss": 2.5311,
      "theoretical_loss": 3.442056168319919,
      "tokens_seen": 1920532480
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022229998108568185,
      "loss": 2.8744,
      "theoretical_loss": 3.442046297040163,
      "tokens_seen": 1920598016
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022228343105731038,
      "loss": 2.5692,
      "theoretical_loss": 3.442036426191546,
      "tokens_seen": 1920663552
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002222668810289389,
      "loss": 2.6754,
      "theoretical_loss": 3.442026555774034,
      "tokens_seen": 1920729088
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022225033100056743,
      "loss": 2.6318,
      "theoretical_loss": 3.442016685787595,
      "tokens_seen": 1920794624
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022223378097219594,
      "loss": 2.558,
      "theoretical_loss": 3.442006816232194,
      "tokens_seen": 1920860160
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022221723094382448,
      "loss": 2.5526,
      "theoretical_loss": 3.441996947107798,
      "tokens_seen": 1920925696
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000222200680915453,
      "loss": 2.5783,
      "theoretical_loss": 3.441987078414373,
      "tokens_seen": 1920991232
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022218413088708153,
      "loss": 2.8217,
      "theoretical_loss": 3.4419772101518866,
      "tokens_seen": 1921056768
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022216758085871004,
      "loss": 2.3648,
      "theoretical_loss": 3.4419673423203037,
      "tokens_seen": 1921122304
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022215103083033858,
      "loss": 2.6995,
      "theoretical_loss": 3.4419574749195925,
      "tokens_seen": 1921187840
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022213448080196709,
      "loss": 2.6768,
      "theoretical_loss": 3.441947607949718,
      "tokens_seen": 1921253376
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022211793077359562,
      "loss": 2.5463,
      "theoretical_loss": 3.441937741410648,
      "tokens_seen": 1921318912
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022210138074522413,
      "loss": 2.525,
      "theoretical_loss": 3.441927875302348,
      "tokens_seen": 1921384448
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022208483071685267,
      "loss": 2.3858,
      "theoretical_loss": 3.441918009624785,
      "tokens_seen": 1921449984
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022206828068848118,
      "loss": 2.5751,
      "theoretical_loss": 3.4419081443779254,
      "tokens_seen": 1921515520
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022205173066010972,
      "loss": 2.9469,
      "theoretical_loss": 3.4418982795617357,
      "tokens_seen": 1921581056
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022203518063173823,
      "loss": 2.8655,
      "theoretical_loss": 3.441888415176182,
      "tokens_seen": 1921646592
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022201863060336677,
      "loss": 2.6042,
      "theoretical_loss": 3.4418785512212313,
      "tokens_seen": 1921712128
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022200208057499528,
      "loss": 2.3613,
      "theoretical_loss": 3.44186868769685,
      "tokens_seen": 1921777664
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022198553054662381,
      "loss": 2.7356,
      "theoretical_loss": 3.441858824603005,
      "tokens_seen": 1921843200
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022196898051825233,
      "loss": 2.6449,
      "theoretical_loss": 3.4418489619396624,
      "tokens_seen": 1921908736
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022195243048988086,
      "loss": 2.5526,
      "theoretical_loss": 3.4418390997067885,
      "tokens_seen": 1921974272
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022193588046150935,
      "loss": 2.4574,
      "theoretical_loss": 3.44182923790435,
      "tokens_seen": 1922039808
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022191933043313786,
      "loss": 2.7073,
      "theoretical_loss": 3.4418193765323135,
      "tokens_seen": 1922105344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002219027804047664,
      "loss": 2.6913,
      "theoretical_loss": 3.4418095155906463,
      "tokens_seen": 1922170880
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002218862303763949,
      "loss": 2.3459,
      "theoretical_loss": 3.441799655079314,
      "tokens_seen": 1922236416
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022186968034802344,
      "loss": 2.5613,
      "theoretical_loss": 3.441789794998283,
      "tokens_seen": 1922301952
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022185313031965195,
      "loss": 2.7017,
      "theoretical_loss": 3.4417799353475202,
      "tokens_seen": 1922367488
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002218365802912805,
      "loss": 2.5452,
      "theoretical_loss": 3.4417700761269927,
      "tokens_seen": 1922433024
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000221820030262909,
      "loss": 2.5665,
      "theoretical_loss": 3.4417602173366664,
      "tokens_seen": 1922498560
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022180348023453754,
      "loss": 2.5353,
      "theoretical_loss": 3.441750358976508,
      "tokens_seen": 1922564096
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022178693020616605,
      "loss": 2.5799,
      "theoretical_loss": 3.4417405010464837,
      "tokens_seen": 1922629632
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022177038017779459,
      "loss": 2.6829,
      "theoretical_loss": 3.441730643546561,
      "tokens_seen": 1922695168
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002217538301494231,
      "loss": 2.4354,
      "theoretical_loss": 3.441720786476706,
      "tokens_seen": 1922760704
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022173728012105163,
      "loss": 2.8266,
      "theoretical_loss": 3.441710929836885,
      "tokens_seen": 1922826240
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022172073009268014,
      "loss": 2.4733,
      "theoretical_loss": 3.4417010736270646,
      "tokens_seen": 1922891776
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022170418006430868,
      "loss": 2.715,
      "theoretical_loss": 3.441691217847212,
      "tokens_seen": 1922957312
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002216876300359372,
      "loss": 2.3929,
      "theoretical_loss": 3.4416813624972935,
      "tokens_seen": 1923022848
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022167108000756573,
      "loss": 2.3725,
      "theoretical_loss": 3.441671507577275,
      "tokens_seen": 1923088384
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022165452997919424,
      "loss": 2.4674,
      "theoretical_loss": 3.441661653087124,
      "tokens_seen": 1923153920
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022163797995082278,
      "loss": 2.6134,
      "theoretical_loss": 3.4416517990268067,
      "tokens_seen": 1923219456
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002216214299224513,
      "loss": 2.283,
      "theoretical_loss": 3.44164194539629,
      "tokens_seen": 1923284992
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022160487989407983,
      "loss": 2.625,
      "theoretical_loss": 3.4416320921955403,
      "tokens_seen": 1923350528
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022158832986570834,
      "loss": 2.674,
      "theoretical_loss": 3.4416222394245235,
      "tokens_seen": 1923416064
    },
    {
      "epoch": 0.69,
      "objective/train/avg_token_score": 0.009885657578706741,
      "objective/train/avg_weight": 0.9883856773376465,
      "objective/train/docs_used": 1086767,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.83903169631958,
      "objective/train/std_weight": 0.07036065310239792,
      "objective/train/theoretical_loss": 3.441612387083208,
      "objective/train/tokens_used": 1943941600,
      "theoretical_loss": 3.441612387083208,
      "tokens_seen": 1923481600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022157177983733687,
      "loss": 2.7172,
      "theoretical_loss": 3.441612387083208,
      "tokens_seen": 1923481600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022155522980896538,
      "loss": 2.7285,
      "theoretical_loss": 3.4416025351715587,
      "tokens_seen": 1923547136
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022153867978059392,
      "loss": 2.6456,
      "theoretical_loss": 3.441592683689543,
      "tokens_seen": 1923612672
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022152212975222243,
      "loss": 2.4983,
      "theoretical_loss": 3.4415828326371276,
      "tokens_seen": 1923678208
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022150557972385097,
      "loss": 2.6316,
      "theoretical_loss": 3.4415729820142786,
      "tokens_seen": 1923743744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022148902969547948,
      "loss": 2.5705,
      "theoretical_loss": 3.441563131820963,
      "tokens_seen": 1923809280
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022147247966710802,
      "loss": 2.7302,
      "theoretical_loss": 3.4415532820571473,
      "tokens_seen": 1923874816
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022145592963873653,
      "loss": 2.6859,
      "theoretical_loss": 3.4415434327227983,
      "tokens_seen": 1923940352
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022143937961036504,
      "loss": 2.3315,
      "theoretical_loss": 3.441533583817883,
      "tokens_seen": 1924005888
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022142282958199355,
      "loss": 2.7191,
      "theoretical_loss": 3.4415237353423676,
      "tokens_seen": 1924071424
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002214062795536221,
      "loss": 2.2909,
      "theoretical_loss": 3.4415138872962183,
      "tokens_seen": 1924136960
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002213897295252506,
      "loss": 2.2418,
      "theoretical_loss": 3.4415040396794025,
      "tokens_seen": 1924202496
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022137317949687913,
      "loss": 2.6646,
      "theoretical_loss": 3.441494192491887,
      "tokens_seen": 1924268032
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022135662946850765,
      "loss": 2.3975,
      "theoretical_loss": 3.4414843457336373,
      "tokens_seen": 1924333568
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022134007944013618,
      "loss": 2.9229,
      "theoretical_loss": 3.441474499404621,
      "tokens_seen": 1924399104
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002213235294117647,
      "loss": 2.5767,
      "theoretical_loss": 3.441464653504805,
      "tokens_seen": 1924464640
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022130697938339323,
      "loss": 2.4348,
      "theoretical_loss": 3.4414548080341554,
      "tokens_seen": 1924530176
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022129042935502174,
      "loss": 2.7646,
      "theoretical_loss": 3.441444962992639,
      "tokens_seen": 1924595712
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022127387932665028,
      "loss": 2.8956,
      "theoretical_loss": 3.441435118380223,
      "tokens_seen": 1924661248
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002212573292982788,
      "loss": 2.7689,
      "theoretical_loss": 3.4414252741968734,
      "tokens_seen": 1924726784
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022124077926990733,
      "loss": 2.6277,
      "theoretical_loss": 3.4414154304425573,
      "tokens_seen": 1924792320
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022122422924153584,
      "loss": 2.4247,
      "theoretical_loss": 3.4414055871172406,
      "tokens_seen": 1924857856
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022120767921316437,
      "loss": 2.4147,
      "theoretical_loss": 3.441395744220891,
      "tokens_seen": 1924923392
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022119112918479289,
      "loss": 2.7087,
      "theoretical_loss": 3.441385901753475,
      "tokens_seen": 1924988928
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022117457915642142,
      "loss": 2.6065,
      "theoretical_loss": 3.441376059714959,
      "tokens_seen": 1925054464
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022115802912804993,
      "loss": 2.6835,
      "theoretical_loss": 3.44136621810531,
      "tokens_seen": 1925120000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022114147909967847,
      "loss": 2.8746,
      "theoretical_loss": 3.4413563769244946,
      "tokens_seen": 1925185536
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022112492907130698,
      "loss": 2.9566,
      "theoretical_loss": 3.441346536172479,
      "tokens_seen": 1925251072
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022110837904293552,
      "loss": 2.718,
      "theoretical_loss": 3.441336695849231,
      "tokens_seen": 1925316608
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022109182901456403,
      "loss": 2.565,
      "theoretical_loss": 3.441326855954716,
      "tokens_seen": 1925382144
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022107527898619257,
      "loss": 2.6209,
      "theoretical_loss": 3.4413170164889024,
      "tokens_seen": 1925447680
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022105872895782108,
      "loss": 2.585,
      "theoretical_loss": 3.441307177451755,
      "tokens_seen": 1925513216
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022104217892944961,
      "loss": 2.5682,
      "theoretical_loss": 3.4412973388432424,
      "tokens_seen": 1925578752
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022102562890107813,
      "loss": 2.8118,
      "theoretical_loss": 3.44128750066333,
      "tokens_seen": 1925644288
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022100907887270666,
      "loss": 2.7354,
      "theoretical_loss": 3.4412776629119852,
      "tokens_seen": 1925709824
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022099252884433517,
      "loss": 2.7518,
      "theoretical_loss": 3.441267825589174,
      "tokens_seen": 1925775360
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002209759788159637,
      "loss": 2.7151,
      "theoretical_loss": 3.4412579886948644,
      "tokens_seen": 1925840896
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002209594287875922,
      "loss": 2.4493,
      "theoretical_loss": 3.441248152229022,
      "tokens_seen": 1925906432
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002209428787592207,
      "loss": 2.6093,
      "theoretical_loss": 3.4412383161916145,
      "tokens_seen": 1925971968
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022092632873084924,
      "loss": 2.629,
      "theoretical_loss": 3.4412284805826077,
      "tokens_seen": 1926037504
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022090977870247775,
      "loss": 2.8189,
      "theoretical_loss": 3.441218645401969,
      "tokens_seen": 1926103040
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002208932286741063,
      "loss": -2.5499,
      "theoretical_loss": 3.441208810649665,
      "tokens_seen": 1926168576
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002208766786457348,
      "loss": 2.5018,
      "theoretical_loss": 3.4411989763256625,
      "tokens_seen": 1926234112
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022086012861736334,
      "loss": 2.5816,
      "theoretical_loss": 3.441189142429928,
      "tokens_seen": 1926299648
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022084357858899185,
      "loss": 2.4991,
      "theoretical_loss": 3.441179308962429,
      "tokens_seen": 1926365184
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022082702856062039,
      "loss": 3.0564,
      "theoretical_loss": 3.4411694759231315,
      "tokens_seen": 1926430720
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002208104785322489,
      "loss": 2.5477,
      "theoretical_loss": 3.4411596433120026,
      "tokens_seen": 1926496256
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022079392850387743,
      "loss": 2.7413,
      "theoretical_loss": 3.4411498111290095,
      "tokens_seen": 1926561792
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022077737847550594,
      "loss": 2.5526,
      "theoretical_loss": 3.4411399793741184,
      "tokens_seen": 1926627328
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022076082844713448,
      "loss": 2.6464,
      "theoretical_loss": 3.441130148047296,
      "tokens_seen": 1926692864
    },
    {
      "epoch": 0.69,
      "objective/train/avg_token_score": 0.01561521366238594,
      "objective/train/avg_weight": 0.9816374778747559,
      "objective/train/docs_used": 1088753,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7077689170837402,
      "objective/train/std_weight": 0.08469463139772415,
      "objective/train/theoretical_loss": 3.44112031714851,
      "objective/train/tokens_used": 1947218400,
      "theoretical_loss": 3.44112031714851,
      "tokens_seen": 1926758400
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000220744278418763,
      "loss": 2.6442,
      "theoretical_loss": 3.44112031714851,
      "tokens_seen": 1926758400
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022072772839039153,
      "loss": 2.6288,
      "theoretical_loss": 3.441110486677726,
      "tokens_seen": 1926823936
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022071117836202004,
      "loss": 2.6993,
      "theoretical_loss": 3.4411006566349114,
      "tokens_seen": 1926889472
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022069462833364858,
      "loss": 2.7166,
      "theoretical_loss": 3.441090827020033,
      "tokens_seen": 1926955008
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002206780783052771,
      "loss": 2.7432,
      "theoretical_loss": 3.4410809978330583,
      "tokens_seen": 1927020544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022066152827690563,
      "loss": 2.5814,
      "theoretical_loss": 3.441071169073953,
      "tokens_seen": 1927086080
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022064497824853414,
      "loss": 2.869,
      "theoretical_loss": 3.441061340742684,
      "tokens_seen": 1927151616
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022062842822016267,
      "loss": 2.6449,
      "theoretical_loss": 3.4410515128392194,
      "tokens_seen": 1927217152
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022061187819179118,
      "loss": 2.7374,
      "theoretical_loss": 3.4410416853635244,
      "tokens_seen": 1927282688
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022059532816341972,
      "loss": 2.6583,
      "theoretical_loss": 3.441031858315567,
      "tokens_seen": 1927348224
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022057877813504823,
      "loss": 2.4431,
      "theoretical_loss": 3.4410220316953133,
      "tokens_seen": 1927413760
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022056222810667677,
      "loss": 2.5047,
      "theoretical_loss": 3.441012205502731,
      "tokens_seen": 1927479296
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022054567807830528,
      "loss": 2.6956,
      "theoretical_loss": 3.441002379737786,
      "tokens_seen": 1927544832
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022052912804993382,
      "loss": 2.6022,
      "theoretical_loss": 3.4409925544004456,
      "tokens_seen": 1927610368
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022051257802156233,
      "loss": 2.7668,
      "theoretical_loss": 3.4409827294906767,
      "tokens_seen": 1927675904
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022049602799319087,
      "loss": 2.5535,
      "theoretical_loss": 3.4409729050084463,
      "tokens_seen": 1927741440
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022047947796481938,
      "loss": 2.9286,
      "theoretical_loss": 3.440963080953721,
      "tokens_seen": 1927806976
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022046292793644786,
      "loss": 2.7411,
      "theoretical_loss": 3.440953257326467,
      "tokens_seen": 1927872512
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002204463779080764,
      "loss": 2.5008,
      "theoretical_loss": 3.4409434341266527,
      "tokens_seen": 1927938048
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002204298278797049,
      "loss": 2.6456,
      "theoretical_loss": 3.440933611354244,
      "tokens_seen": 1928003584
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022041327785133345,
      "loss": 2.6779,
      "theoretical_loss": 3.4409237890092075,
      "tokens_seen": 1928069120
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022039672782296196,
      "loss": 2.7112,
      "theoretical_loss": 3.440913967091511,
      "tokens_seen": 1928134656
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002203801777945905,
      "loss": 2.881,
      "theoretical_loss": 3.4409041456011202,
      "tokens_seen": 1928200192
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000220363627766219,
      "loss": 2.7185,
      "theoretical_loss": 3.4408943245380033,
      "tokens_seen": 1928265728
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022034707773784754,
      "loss": 2.6626,
      "theoretical_loss": 3.4408845039021263,
      "tokens_seen": 1928331264
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022033052770947605,
      "loss": 2.6242,
      "theoretical_loss": 3.4408746836934565,
      "tokens_seen": 1928396800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002203139776811046,
      "loss": 2.6903,
      "theoretical_loss": 3.4408648639119606,
      "tokens_seen": 1928462336
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002202974276527331,
      "loss": 2.6378,
      "theoretical_loss": 3.4408550445576056,
      "tokens_seen": 1928527872
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022028087762436164,
      "loss": 2.6804,
      "theoretical_loss": 3.4408452256303583,
      "tokens_seen": 1928593408
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022026432759599015,
      "loss": 2.4164,
      "theoretical_loss": 3.4408354071301854,
      "tokens_seen": 1928658944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022024777756761868,
      "loss": 2.7509,
      "theoretical_loss": 3.4408255890570545,
      "tokens_seen": 1928724480
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002202312275392472,
      "loss": 2.8011,
      "theoretical_loss": 3.440815771410932,
      "tokens_seen": 1928790016
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022021467751087573,
      "loss": 2.6662,
      "theoretical_loss": 3.440805954191785,
      "tokens_seen": 1928855552
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022019812748250424,
      "loss": 3.0068,
      "theoretical_loss": 3.4407961373995803,
      "tokens_seen": 1928921088
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022018157745413278,
      "loss": 2.5816,
      "theoretical_loss": 3.4407863210342846,
      "tokens_seen": 1928986624
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002201650274257613,
      "loss": 2.2293,
      "theoretical_loss": 3.440776505095865,
      "tokens_seen": 1929052160
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022014847739738983,
      "loss": 2.9413,
      "theoretical_loss": 3.440766689584289,
      "tokens_seen": 1929117696
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022013192736901834,
      "loss": 2.6446,
      "theoretical_loss": 3.4407568744995234,
      "tokens_seen": 1929183232
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022011537734064688,
      "loss": 2.9437,
      "theoretical_loss": 3.4407470598415344,
      "tokens_seen": 1929248768
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002200988273122754,
      "loss": 2.2819,
      "theoretical_loss": 3.440737245610289,
      "tokens_seen": 1929314304
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022008227728390392,
      "loss": 2.5383,
      "theoretical_loss": 3.4407274318057555,
      "tokens_seen": 1929379840
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022006572725553244,
      "loss": 2.8515,
      "theoretical_loss": 3.440717618427899,
      "tokens_seen": 1929445376
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022004917722716097,
      "loss": 2.3491,
      "theoretical_loss": 3.4407078054766878,
      "tokens_seen": 1929510912
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022003262719878948,
      "loss": 2.8989,
      "theoretical_loss": 3.440697992952088,
      "tokens_seen": 1929576448
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00022001607717041802,
      "loss": 2.6761,
      "theoretical_loss": 3.4406881808540675,
      "tokens_seen": 1929641984
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021999952714204653,
      "loss": 2.4617,
      "theoretical_loss": 3.440678369182592,
      "tokens_seen": 1929707520
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021998297711367507,
      "loss": 2.7158,
      "theoretical_loss": 3.44066855793763,
      "tokens_seen": 1929773056
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021996642708530355,
      "loss": 2.644,
      "theoretical_loss": 3.440658747119147,
      "tokens_seen": 1929838592
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002199498770569321,
      "loss": 2.7458,
      "theoretical_loss": 3.440648936727111,
      "tokens_seen": 1929904128
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002199333270285606,
      "loss": 2.9172,
      "theoretical_loss": 3.4406391267614884,
      "tokens_seen": 1929969664
    },
    {
      "epoch": 0.69,
      "objective/train/avg_token_score": 0.027505122125148773,
      "objective/train/avg_weight": 0.9676392078399658,
      "objective/train/docs_used": 1090637,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 8.631147384643555,
      "objective/train/std_weight": 0.1652279496192932,
      "objective/train/theoretical_loss": 3.4406293172222466,
      "objective/train/tokens_used": 1950495200,
      "theoretical_loss": 3.4406293172222466,
      "tokens_seen": 1930035200
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021991677700018914,
      "loss": 2.2116,
      "theoretical_loss": 3.4406293172222466,
      "tokens_seen": 1930035200
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021990022697181765,
      "loss": 2.3321,
      "theoretical_loss": 3.4406195081093527,
      "tokens_seen": 1930100736
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021988367694344619,
      "loss": 2.6943,
      "theoretical_loss": 3.4406096994227733,
      "tokens_seen": 1930166272
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002198671269150747,
      "loss": 2.4813,
      "theoretical_loss": 3.440599891162475,
      "tokens_seen": 1930231808
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021985057688670323,
      "loss": 2.7162,
      "theoretical_loss": 3.4405900833284258,
      "tokens_seen": 1930297344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021983402685833174,
      "loss": 2.5071,
      "theoretical_loss": 3.440580275920592,
      "tokens_seen": 1930362880
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021981747682996028,
      "loss": 2.645,
      "theoretical_loss": 3.440570468938941,
      "tokens_seen": 1930428416
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002198009268015888,
      "loss": 2.3405,
      "theoretical_loss": 3.4405606623834393,
      "tokens_seen": 1930493952
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021978437677321733,
      "loss": 2.5516,
      "theoretical_loss": 3.4405508562540543,
      "tokens_seen": 1930559488
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021976782674484584,
      "loss": 2.7841,
      "theoretical_loss": 3.4405410505507534,
      "tokens_seen": 1930625024
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021975127671647438,
      "loss": 2.6733,
      "theoretical_loss": 3.440531245273503,
      "tokens_seen": 1930690560
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002197347266881029,
      "loss": 2.4814,
      "theoretical_loss": 3.4405214404222706,
      "tokens_seen": 1930756096
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021971817665973143,
      "loss": 2.6994,
      "theoretical_loss": 3.4405116359970225,
      "tokens_seen": 1930821632
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021970162663135994,
      "loss": 2.7649,
      "theoretical_loss": 3.4405018319977265,
      "tokens_seen": 1930887168
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021968507660298847,
      "loss": 2.6691,
      "theoretical_loss": 3.440492028424349,
      "tokens_seen": 1930952704
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021966852657461698,
      "loss": 2.5986,
      "theoretical_loss": 3.440482225276858,
      "tokens_seen": 1931018240
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021965197654624552,
      "loss": 2.6281,
      "theoretical_loss": 3.440472422555219,
      "tokens_seen": 1931083776
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021963542651787403,
      "loss": 2.7454,
      "theoretical_loss": 3.4404626202594004,
      "tokens_seen": 1931149312
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021961887648950257,
      "loss": 2.8476,
      "theoretical_loss": 3.4404528183893692,
      "tokens_seen": 1931214848
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021960232646113108,
      "loss": 2.3891,
      "theoretical_loss": 3.4404430169450917,
      "tokens_seen": 1931280384
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021958577643275962,
      "loss": 2.7652,
      "theoretical_loss": 3.4404332159265354,
      "tokens_seen": 1931345920
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021956922640438813,
      "loss": 2.6573,
      "theoretical_loss": 3.440423415333667,
      "tokens_seen": 1931411456
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021955267637601667,
      "loss": 2.6513,
      "theoretical_loss": 3.440413615166454,
      "tokens_seen": 1931476992
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021953612634764518,
      "loss": 2.7292,
      "theoretical_loss": 3.440403815424864,
      "tokens_seen": 1931542528
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002195195763192737,
      "loss": 2.667,
      "theoretical_loss": 3.440394016108863,
      "tokens_seen": 1931608064
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021950302629090222,
      "loss": 2.6475,
      "theoretical_loss": 3.4403842172184183,
      "tokens_seen": 1931673600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002194864762625307,
      "loss": 2.537,
      "theoretical_loss": 3.440374418753497,
      "tokens_seen": 1931739136
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021946992623415924,
      "loss": 2.6916,
      "theoretical_loss": 3.4403646207140666,
      "tokens_seen": 1931804672
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021945337620578776,
      "loss": 2.7281,
      "theoretical_loss": 3.440354823100094,
      "tokens_seen": 1931870208
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002194368261774163,
      "loss": 2.5859,
      "theoretical_loss": 3.4403450259115465,
      "tokens_seen": 1931935744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002194202761490448,
      "loss": 2.7524,
      "theoretical_loss": 3.4403352291483906,
      "tokens_seen": 1932001280
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021940372612067334,
      "loss": 2.4048,
      "theoretical_loss": 3.440325432810594,
      "tokens_seen": 1932066816
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021938717609230185,
      "loss": 2.6498,
      "theoretical_loss": 3.4403156368981236,
      "tokens_seen": 1932132352
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002193706260639304,
      "loss": 2.456,
      "theoretical_loss": 3.4403058414109458,
      "tokens_seen": 1932197888
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002193540760355589,
      "loss": 2.3957,
      "theoretical_loss": 3.4402960463490286,
      "tokens_seen": 1932263424
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021933752600718744,
      "loss": 2.627,
      "theoretical_loss": 3.4402862517123394,
      "tokens_seen": 1932328960
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021932097597881595,
      "loss": 2.6831,
      "theoretical_loss": 3.440276457500844,
      "tokens_seen": 1932394496
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021930442595044448,
      "loss": 2.6501,
      "theoretical_loss": 3.440266663714511,
      "tokens_seen": 1932460032
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000219287875922073,
      "loss": 2.4755,
      "theoretical_loss": 3.440256870353307,
      "tokens_seen": 1932525568
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021927132589370153,
      "loss": 2.7715,
      "theoretical_loss": 3.4402470774171987,
      "tokens_seen": 1932591104
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021925477586533004,
      "loss": 2.5527,
      "theoretical_loss": 3.4402372849061535,
      "tokens_seen": 1932656640
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021923822583695858,
      "loss": 2.4299,
      "theoretical_loss": 3.4402274928201386,
      "tokens_seen": 1932722176
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002192216758085871,
      "loss": 2.1471,
      "theoretical_loss": 3.4402177011591206,
      "tokens_seen": 1932787712
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021920512578021563,
      "loss": 2.6924,
      "theoretical_loss": 3.4402079099230676,
      "tokens_seen": 1932853248
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021918857575184414,
      "loss": 2.5032,
      "theoretical_loss": 3.4401981191119466,
      "tokens_seen": 1932918784
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021917202572347268,
      "loss": 2.6622,
      "theoretical_loss": 3.440188328725724,
      "tokens_seen": 1932984320
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002191554756951012,
      "loss": 2.4947,
      "theoretical_loss": 3.4401785387643677,
      "tokens_seen": 1933049856
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021913892566672972,
      "loss": 2.5998,
      "theoretical_loss": 3.440168749227844,
      "tokens_seen": 1933115392
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021912237563835824,
      "loss": 2.95,
      "theoretical_loss": 3.440158960116121,
      "tokens_seen": 1933180928
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021910582560998677,
      "loss": 2.5121,
      "theoretical_loss": 3.4401491714291654,
      "tokens_seen": 1933246464
    },
    {
      "epoch": 0.69,
      "objective/train/avg_token_score": 0.022286441177129745,
      "objective/train/avg_weight": 0.9737673401832581,
      "objective/train/docs_used": 1092665,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0844321250915527,
      "objective/train/std_weight": 0.10162203013896942,
      "objective/train/theoretical_loss": 3.4401393831669447,
      "objective/train/tokens_used": 1953772000,
      "theoretical_loss": 3.4401393831669447,
      "tokens_seen": 1933312000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021908927558161528,
      "loss": 2.7695,
      "theoretical_loss": 3.4401393831669447,
      "tokens_seen": 1933312000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021907272555324382,
      "loss": 2.6587,
      "theoretical_loss": 3.440129595329425,
      "tokens_seen": 1933377536
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021905617552487233,
      "loss": 2.7835,
      "theoretical_loss": 3.440119807916575,
      "tokens_seen": 1933443072
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021903962549650087,
      "loss": 2.3474,
      "theoretical_loss": 3.440110020928361,
      "tokens_seen": 1933508608
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021902307546812938,
      "loss": 2.5707,
      "theoretical_loss": 3.440100234364751,
      "tokens_seen": 1933574144
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021900652543975792,
      "loss": 2.196,
      "theoretical_loss": 3.4400904482257104,
      "tokens_seen": 1933639680
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002189899754113864,
      "loss": 2.5418,
      "theoretical_loss": 3.4400806625112086,
      "tokens_seen": 1933705216
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002189734253830149,
      "loss": 2.5817,
      "theoretical_loss": 3.440070877221211,
      "tokens_seen": 1933770752
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021895687535464345,
      "loss": 2.6691,
      "theoretical_loss": 3.440061092355686,
      "tokens_seen": 1933836288
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021894032532627196,
      "loss": 2.6432,
      "theoretical_loss": 3.4400513079146,
      "tokens_seen": 1933901824
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002189237752979005,
      "loss": 2.5569,
      "theoretical_loss": 3.440041523897921,
      "tokens_seen": 1933967360
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000218907225269529,
      "loss": 1.4302,
      "theoretical_loss": 3.4400317403056153,
      "tokens_seen": 1934032896
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021889067524115754,
      "loss": 2.5726,
      "theoretical_loss": 3.4400219571376507,
      "tokens_seen": 1934098432
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021887412521278605,
      "loss": 2.594,
      "theoretical_loss": 3.4400121743939946,
      "tokens_seen": 1934163968
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002188575751844146,
      "loss": 2.8153,
      "theoretical_loss": 3.4400023920746134,
      "tokens_seen": 1934229504
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002188410251560431,
      "loss": 2.6755,
      "theoretical_loss": 3.439992610179475,
      "tokens_seen": 1934295040
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021882447512767164,
      "loss": 2.6057,
      "theoretical_loss": 3.4399828287085468,
      "tokens_seen": 1934360576
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021880792509930015,
      "loss": 2.6888,
      "theoretical_loss": 3.4399730476617956,
      "tokens_seen": 1934426112
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002187913750709287,
      "loss": 2.5129,
      "theoretical_loss": 3.4399632670391886,
      "tokens_seen": 1934491648
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002187748250425572,
      "loss": 2.452,
      "theoretical_loss": 3.439953486840693,
      "tokens_seen": 1934557184
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021875827501418574,
      "loss": 2.7178,
      "theoretical_loss": 3.4399437070662766,
      "tokens_seen": 1934622720
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021874172498581427,
      "loss": 2.7125,
      "theoretical_loss": 3.439933927715906,
      "tokens_seen": 1934688256
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021872517495744278,
      "loss": 2.5485,
      "theoretical_loss": 3.439924148789549,
      "tokens_seen": 1934753792
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021870862492907132,
      "loss": 2.5639,
      "theoretical_loss": 3.4399143702871724,
      "tokens_seen": 1934819328
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021869207490069983,
      "loss": 2.5694,
      "theoretical_loss": 3.4399045922087437,
      "tokens_seen": 1934884864
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021867552487232837,
      "loss": 2.624,
      "theoretical_loss": 3.4398948145542296,
      "tokens_seen": 1934950400
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021865897484395688,
      "loss": 1.9966,
      "theoretical_loss": 3.4398850373235987,
      "tokens_seen": 1935015936
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021864242481558542,
      "loss": 2.487,
      "theoretical_loss": 3.439875260516817,
      "tokens_seen": 1935081472
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021862587478721393,
      "loss": 2.4769,
      "theoretical_loss": 3.439865484133852,
      "tokens_seen": 1935147008
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021860932475884247,
      "loss": 2.4833,
      "theoretical_loss": 3.4398557081746715,
      "tokens_seen": 1935212544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021859277473047098,
      "loss": 2.5875,
      "theoretical_loss": 3.439845932639242,
      "tokens_seen": 1935278080
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002185762247020995,
      "loss": 1.9929,
      "theoretical_loss": 3.4398361575275316,
      "tokens_seen": 1935343616
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021855967467372802,
      "loss": 2.5794,
      "theoretical_loss": 3.439826382839507,
      "tokens_seen": 1935409152
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021854312464535656,
      "loss": 2.6721,
      "theoretical_loss": 3.4398166085751356,
      "tokens_seen": 1935474688
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021852657461698507,
      "loss": 2.743,
      "theoretical_loss": 3.4398068347343855,
      "tokens_seen": 1935540224
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002185100245886136,
      "loss": 2.5821,
      "theoretical_loss": 3.439797061317223,
      "tokens_seen": 1935605760
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002184934745602421,
      "loss": 2.4748,
      "theoretical_loss": 3.439787288323615,
      "tokens_seen": 1935671296
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002184769245318706,
      "loss": 2.5703,
      "theoretical_loss": 3.43977751575353,
      "tokens_seen": 1935736832
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021846037450349914,
      "loss": 2.2191,
      "theoretical_loss": 3.439767743606935,
      "tokens_seen": 1935802368
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021844382447512765,
      "loss": 2.7913,
      "theoretical_loss": 3.439757971883797,
      "tokens_seen": 1935867904
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002184272744467562,
      "loss": 2.5454,
      "theoretical_loss": 3.4397482005840834,
      "tokens_seen": 1935933440
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002184107244183847,
      "loss": 2.6379,
      "theoretical_loss": 3.439738429707762,
      "tokens_seen": 1935998976
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021839417439001324,
      "loss": 2.5047,
      "theoretical_loss": 3.439728659254799,
      "tokens_seen": 1936064512
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021837762436164175,
      "loss": 2.8034,
      "theoretical_loss": 3.439718889225163,
      "tokens_seen": 1936130048
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021836107433327028,
      "loss": 2.4436,
      "theoretical_loss": 3.4397091196188203,
      "tokens_seen": 1936195584
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002183445243048988,
      "loss": 2.5808,
      "theoretical_loss": 3.4396993504357387,
      "tokens_seen": 1936261120
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021832797427652733,
      "loss": 2.6582,
      "theoretical_loss": 3.4396895816758857,
      "tokens_seen": 1936326656
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021831142424815584,
      "loss": 2.4846,
      "theoretical_loss": 3.439679813339229,
      "tokens_seen": 1936392192
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021829487421978438,
      "loss": 2.7264,
      "theoretical_loss": 3.439670045425734,
      "tokens_seen": 1936457728
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002182783241914129,
      "loss": 2.7203,
      "theoretical_loss": 3.4396602779353707,
      "tokens_seen": 1936523264
    },
    {
      "epoch": 0.69,
      "objective/train/avg_token_score": 0.01682845875620842,
      "objective/train/avg_weight": 0.9802206754684448,
      "objective/train/docs_used": 1094661,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.968301296234131,
      "objective/train/std_weight": 0.0966724082827568,
      "objective/train/theoretical_loss": 3.4396505108681046,
      "objective/train/tokens_used": 1957048800,
      "theoretical_loss": 3.4396505108681046,
      "tokens_seen": 1936588800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021826177416304143,
      "loss": 2.5986,
      "theoretical_loss": 3.4396505108681046,
      "tokens_seen": 1936588800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021824522413466994,
      "loss": 2.6707,
      "theoretical_loss": 3.4396407442239036,
      "tokens_seen": 1936654336
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021822867410629848,
      "loss": 2.7508,
      "theoretical_loss": 3.4396309780027354,
      "tokens_seen": 1936719872
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000218212124077927,
      "loss": 2.7956,
      "theoretical_loss": 3.4396212122045666,
      "tokens_seen": 1936785408
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021819557404955552,
      "loss": 2.6539,
      "theoretical_loss": 3.4396114468293657,
      "tokens_seen": 1936850944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021817902402118403,
      "loss": 2.8331,
      "theoretical_loss": 3.4396016818770994,
      "tokens_seen": 1936916480
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021816247399281257,
      "loss": 2.6908,
      "theoretical_loss": 3.4395919173477347,
      "tokens_seen": 1936982016
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021814592396444108,
      "loss": 2.7421,
      "theoretical_loss": 3.43958215324124,
      "tokens_seen": 1937047552
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021812937393606962,
      "loss": 2.6256,
      "theoretical_loss": 3.439572389557581,
      "tokens_seen": 1937113088
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021811282390769813,
      "loss": 2.5519,
      "theoretical_loss": 3.439562626296727,
      "tokens_seen": 1937178624
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021809627387932667,
      "loss": 2.8302,
      "theoretical_loss": 3.439552863458644,
      "tokens_seen": 1937244160
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021807972385095518,
      "loss": 2.6192,
      "theoretical_loss": 3.4395431010433,
      "tokens_seen": 1937309696
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021806317382258372,
      "loss": 2.3063,
      "theoretical_loss": 3.439533339050662,
      "tokens_seen": 1937375232
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021804662379421223,
      "loss": 2.5583,
      "theoretical_loss": 3.4395235774806983,
      "tokens_seen": 1937440768
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021803007376584076,
      "loss": 2.8096,
      "theoretical_loss": 3.439513816333376,
      "tokens_seen": 1937506304
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021801352373746927,
      "loss": 2.5595,
      "theoretical_loss": 3.439504055608661,
      "tokens_seen": 1937571840
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021799697370909776,
      "loss": 2.6462,
      "theoretical_loss": 3.439494295306523,
      "tokens_seen": 1937637376
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002179804236807263,
      "loss": 2.6294,
      "theoretical_loss": 3.439484535426928,
      "tokens_seen": 1937702912
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002179638736523548,
      "loss": 2.6367,
      "theoretical_loss": 3.4394747759698436,
      "tokens_seen": 1937768448
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021794732362398334,
      "loss": 2.7019,
      "theoretical_loss": 3.439465016935238,
      "tokens_seen": 1937833984
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021793077359561185,
      "loss": 2.5035,
      "theoretical_loss": 3.4394552583230773,
      "tokens_seen": 1937899520
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002179142235672404,
      "loss": 2.6342,
      "theoretical_loss": 3.4394455001333295,
      "tokens_seen": 1937965056
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002178976735388689,
      "loss": 2.5177,
      "theoretical_loss": 3.439435742365963,
      "tokens_seen": 1938030592
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021788112351049744,
      "loss": 2.678,
      "theoretical_loss": 3.439425985020944,
      "tokens_seen": 1938096128
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021786457348212595,
      "loss": 2.6023,
      "theoretical_loss": 3.43941622809824,
      "tokens_seen": 1938161664
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002178480234537545,
      "loss": 2.5985,
      "theoretical_loss": 3.439406471597819,
      "tokens_seen": 1938227200
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000217831473425383,
      "loss": 2.4062,
      "theoretical_loss": 3.4393967155196483,
      "tokens_seen": 1938292736
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021781492339701154,
      "loss": 2.5723,
      "theoretical_loss": 3.439386959863695,
      "tokens_seen": 1938358272
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021779837336864005,
      "loss": 2.6062,
      "theoretical_loss": 3.439377204629927,
      "tokens_seen": 1938423808
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021778182334026858,
      "loss": 2.5001,
      "theoretical_loss": 3.4393674498183118,
      "tokens_seen": 1938489344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002177652733118971,
      "loss": 2.6002,
      "theoretical_loss": 3.439357695428816,
      "tokens_seen": 1938554880
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021774872328352563,
      "loss": 2.5793,
      "theoretical_loss": 3.4393479414614085,
      "tokens_seen": 1938620416
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021773217325515414,
      "loss": 2.5599,
      "theoretical_loss": 3.4393381879160554,
      "tokens_seen": 1938685952
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021771562322678268,
      "loss": 2.8119,
      "theoretical_loss": 3.439328434792725,
      "tokens_seen": 1938751488
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002176990731984112,
      "loss": 2.4215,
      "theoretical_loss": 3.439318682091385,
      "tokens_seen": 1938817024
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021768252317003973,
      "loss": 2.5669,
      "theoretical_loss": 3.4393089298120016,
      "tokens_seen": 1938882560
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021766597314166824,
      "loss": 2.5558,
      "theoretical_loss": 3.4392991779545437,
      "tokens_seen": 1938948096
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021764942311329678,
      "loss": 2.3866,
      "theoretical_loss": 3.439289426518978,
      "tokens_seen": 1939013632
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021763287308492529,
      "loss": 2.896,
      "theoretical_loss": 3.4392796755052712,
      "tokens_seen": 1939079168
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021761632305655382,
      "loss": 2.4401,
      "theoretical_loss": 3.439269924913393,
      "tokens_seen": 1939144704
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021759977302818233,
      "loss": 2.6238,
      "theoretical_loss": 3.439260174743309,
      "tokens_seen": 1939210240
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021758322299981087,
      "loss": 2.3386,
      "theoretical_loss": 3.439250424994988,
      "tokens_seen": 1939275776
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021756667297143938,
      "loss": 2.4647,
      "theoretical_loss": 3.4392406756683958,
      "tokens_seen": 1939341312
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021755012294306792,
      "loss": 2.6392,
      "theoretical_loss": 3.4392309267635017,
      "tokens_seen": 1939406848
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021753357291469643,
      "loss": 2.6081,
      "theoretical_loss": 3.439221178280272,
      "tokens_seen": 1939472384
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002175170228863249,
      "loss": 2.1415,
      "theoretical_loss": 3.439211430218675,
      "tokens_seen": 1939537920
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021750047285795345,
      "loss": 2.3992,
      "theoretical_loss": 3.439201682578678,
      "tokens_seen": 1939603456
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000217483922829582,
      "loss": 2.4314,
      "theoretical_loss": 3.439191935360249,
      "tokens_seen": 1939668992
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002174673728012105,
      "loss": 2.8491,
      "theoretical_loss": 3.439182188563354,
      "tokens_seen": 1939734528
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021745082277283904,
      "loss": 2.7997,
      "theoretical_loss": 3.4391724421879615,
      "tokens_seen": 1939800064
    },
    {
      "epoch": 0.69,
      "objective/train/avg_token_score": 0.022741537541151047,
      "objective/train/avg_weight": 0.9732460379600525,
      "objective/train/docs_used": 1096594,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.4894332885742188,
      "objective/train/std_weight": 0.11774633079767227,
      "objective/train/theoretical_loss": 3.4391626962340394,
      "objective/train/tokens_used": 1960325600,
      "theoretical_loss": 3.4391626962340394,
      "tokens_seen": 1939865600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021743427274446755,
      "loss": 2.4767,
      "theoretical_loss": 3.4391626962340394,
      "tokens_seen": 1939865600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021741772271609608,
      "loss": 2.6487,
      "theoretical_loss": 3.439152950701555,
      "tokens_seen": 1939931136
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002174011726877246,
      "loss": 2.5205,
      "theoretical_loss": 3.439143205590475,
      "tokens_seen": 1939996672
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021738462265935313,
      "loss": 2.955,
      "theoretical_loss": 3.439133460900768,
      "tokens_seen": 1940062208
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021736807263098164,
      "loss": 2.4832,
      "theoretical_loss": 3.439123716632401,
      "tokens_seen": 1940127744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021735152260261018,
      "loss": 2.5384,
      "theoretical_loss": 3.439113972785342,
      "tokens_seen": 1940193280
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002173349725742387,
      "loss": 2.6817,
      "theoretical_loss": 3.4391042293595584,
      "tokens_seen": 1940258816
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021731842254586723,
      "loss": 2.6889,
      "theoretical_loss": 3.4390944863550175,
      "tokens_seen": 1940324352
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021730187251749574,
      "loss": 2.6413,
      "theoretical_loss": 3.4390847437716863,
      "tokens_seen": 1940389888
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021728532248912428,
      "loss": 2.6814,
      "theoretical_loss": 3.4390750016095337,
      "tokens_seen": 1940455424
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002172687724607528,
      "loss": 2.5835,
      "theoretical_loss": 3.439065259868527,
      "tokens_seen": 1940520960
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021725222243238132,
      "loss": 2.6369,
      "theoretical_loss": 3.4390555185486327,
      "tokens_seen": 1940586496
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021723567240400983,
      "loss": 2.4527,
      "theoretical_loss": 3.4390457776498193,
      "tokens_seen": 1940652032
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021721912237563837,
      "loss": 2.4047,
      "theoretical_loss": 3.4390360371720545,
      "tokens_seen": 1940717568
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021720257234726688,
      "loss": 2.888,
      "theoretical_loss": 3.439026297115305,
      "tokens_seen": 1940783104
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021718602231889542,
      "loss": -5.1767,
      "theoretical_loss": 3.4390165574795395,
      "tokens_seen": 1940848640
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021716947229052393,
      "loss": 2.7059,
      "theoretical_loss": 3.4390068182647244,
      "tokens_seen": 1940914176
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021715292226215247,
      "loss": 2.515,
      "theoretical_loss": 3.4389970794708287,
      "tokens_seen": 1940979712
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021713637223378098,
      "loss": 2.7946,
      "theoretical_loss": 3.4389873410978185,
      "tokens_seen": 1941045248
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021711982220540952,
      "loss": 2.7272,
      "theoretical_loss": 3.4389776031456623,
      "tokens_seen": 1941110784
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021710327217703803,
      "loss": 2.5564,
      "theoretical_loss": 3.4389678656143277,
      "tokens_seen": 1941176320
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021708672214866656,
      "loss": 2.7685,
      "theoretical_loss": 3.4389581285037822,
      "tokens_seen": 1941241856
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021707017212029507,
      "loss": 2.2273,
      "theoretical_loss": 3.4389483918139927,
      "tokens_seen": 1941307392
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002170536220919236,
      "loss": 2.8639,
      "theoretical_loss": 3.438938655544928,
      "tokens_seen": 1941372928
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021703707206355212,
      "loss": 2.8021,
      "theoretical_loss": 3.438928919696555,
      "tokens_seen": 1941438464
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002170205220351806,
      "loss": 2.4819,
      "theoretical_loss": 3.4389191842688414,
      "tokens_seen": 1941504000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021700397200680914,
      "loss": 2.803,
      "theoretical_loss": 3.438909449261755,
      "tokens_seen": 1941569536
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021698742197843765,
      "loss": 2.6402,
      "theoretical_loss": 3.4388997146752636,
      "tokens_seen": 1941635072
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002169708719500662,
      "loss": 2.5236,
      "theoretical_loss": 3.4388899805093343,
      "tokens_seen": 1941700608
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002169543219216947,
      "loss": 2.6519,
      "theoretical_loss": 3.438880246763935,
      "tokens_seen": 1941766144
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021693777189332324,
      "loss": 2.9032,
      "theoretical_loss": 3.438870513439033,
      "tokens_seen": 1941831680
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021692122186495175,
      "loss": 2.8104,
      "theoretical_loss": 3.438860780534597,
      "tokens_seen": 1941897216
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002169046718365803,
      "loss": 2.6671,
      "theoretical_loss": 3.438851048050594,
      "tokens_seen": 1941962752
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002168881218082088,
      "loss": 2.7494,
      "theoretical_loss": 3.4388413159869913,
      "tokens_seen": 1942028288
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021687157177983734,
      "loss": 3.0292,
      "theoretical_loss": 3.438831584343757,
      "tokens_seen": 1942093824
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021685502175146585,
      "loss": 2.8557,
      "theoretical_loss": 3.4388218531208583,
      "tokens_seen": 1942159360
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021683847172309438,
      "loss": 2.8251,
      "theoretical_loss": 3.438812122318263,
      "tokens_seen": 1942224896
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002168219216947229,
      "loss": 2.9076,
      "theoretical_loss": 3.4388023919359396,
      "tokens_seen": 1942290432
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021680537166635143,
      "loss": 2.5774,
      "theoretical_loss": 3.4387926619738547,
      "tokens_seen": 1942355968
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021678882163797994,
      "loss": 2.8374,
      "theoretical_loss": 3.438782932431976,
      "tokens_seen": 1942421504
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021677227160960848,
      "loss": 2.7014,
      "theoretical_loss": 3.438773203310272,
      "tokens_seen": 1942487040
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000216755721581237,
      "loss": 2.5845,
      "theoretical_loss": 3.4387634746087103,
      "tokens_seen": 1942552576
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021673917155286553,
      "loss": 2.7484,
      "theoretical_loss": 3.4387537463272575,
      "tokens_seen": 1942618112
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021672262152449404,
      "loss": 2.6729,
      "theoretical_loss": 3.438744018465882,
      "tokens_seen": 1942683648
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021670607149612258,
      "loss": 2.8624,
      "theoretical_loss": 3.438734291024552,
      "tokens_seen": 1942749184
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021668952146775109,
      "loss": 2.4024,
      "theoretical_loss": 3.438724564003235,
      "tokens_seen": 1942814720
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021667297143937962,
      "loss": 2.531,
      "theoretical_loss": 3.438714837401897,
      "tokens_seen": 1942880256
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021665642141100813,
      "loss": 2.5591,
      "theoretical_loss": 3.4387051112205085,
      "tokens_seen": 1942945792
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021663987138263667,
      "loss": 2.6893,
      "theoretical_loss": 3.4386953854590354,
      "tokens_seen": 1943011328
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021662332135426518,
      "loss": 2.3751,
      "theoretical_loss": 3.4386856601174456,
      "tokens_seen": 1943076864
    },
    {
      "epoch": 0.69,
      "objective/train/avg_token_score": 0.013988016173243523,
      "objective/train/avg_weight": 0.9835357666015625,
      "objective/train/docs_used": 1098399,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.87115478515625,
      "objective/train/std_weight": 0.07979167252779007,
      "objective/train/theoretical_loss": 3.4386759351957066,
      "objective/train/tokens_used": 1963602400,
      "theoretical_loss": 3.4386759351957066,
      "tokens_seen": 1943142400
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021660677132589372,
      "loss": 2.5702,
      "theoretical_loss": 3.4386759351957066,
      "tokens_seen": 1943142400
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021659022129752223,
      "loss": 2.5258,
      "theoretical_loss": 3.4386662106937873,
      "tokens_seen": 1943207936
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021657367126915077,
      "loss": 2.6553,
      "theoretical_loss": 3.438656486611654,
      "tokens_seen": 1943273472
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021655712124077928,
      "loss": 2.6397,
      "theoretical_loss": 3.4386467629492756,
      "tokens_seen": 1943339008
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021654057121240782,
      "loss": 2.5716,
      "theoretical_loss": 3.4386370397066193,
      "tokens_seen": 1943404544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002165240211840363,
      "loss": 2.6215,
      "theoretical_loss": 3.4386273168836525,
      "tokens_seen": 1943470080
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002165074711556648,
      "loss": 2.5761,
      "theoretical_loss": 3.4386175944803434,
      "tokens_seen": 1943535616
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021649092112729335,
      "loss": 2.4128,
      "theoretical_loss": 3.4386078724966596,
      "tokens_seen": 1943601152
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021647437109892186,
      "loss": 2.7588,
      "theoretical_loss": 3.438598150932569,
      "tokens_seen": 1943666688
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002164578210705504,
      "loss": 2.5625,
      "theoretical_loss": 3.4385884297880387,
      "tokens_seen": 1943732224
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002164412710421789,
      "loss": 2.3378,
      "theoretical_loss": 3.4385787090630373,
      "tokens_seen": 1943797760
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021642472101380744,
      "loss": 2.6967,
      "theoretical_loss": 3.438568988757532,
      "tokens_seen": 1943863296
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021640817098543595,
      "loss": 2.8792,
      "theoretical_loss": 3.438559268871491,
      "tokens_seen": 1943928832
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002163916209570645,
      "loss": 2.6877,
      "theoretical_loss": 3.4385495494048817,
      "tokens_seen": 1943994368
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000216375070928693,
      "loss": 2.5121,
      "theoretical_loss": 3.4385398303576715,
      "tokens_seen": 1944059904
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021635852090032154,
      "loss": 2.4904,
      "theoretical_loss": 3.438530111729829,
      "tokens_seen": 1944125440
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021634197087195005,
      "loss": 2.7454,
      "theoretical_loss": 3.4385203935213218,
      "tokens_seen": 1944190976
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021632542084357859,
      "loss": 2.6343,
      "theoretical_loss": 3.4385106757321173,
      "tokens_seen": 1944256512
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002163088708152071,
      "loss": 2.3569,
      "theoretical_loss": 3.438500958362183,
      "tokens_seen": 1944322048
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021629232078683563,
      "loss": 2.3558,
      "theoretical_loss": 3.438491241411487,
      "tokens_seen": 1944387584
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021627577075846414,
      "loss": 2.7372,
      "theoretical_loss": 3.438481524879998,
      "tokens_seen": 1944453120
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021625922073009268,
      "loss": 2.6269,
      "theoretical_loss": 3.438471808767683,
      "tokens_seen": 1944518656
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002162426707017212,
      "loss": 2.4116,
      "theoretical_loss": 3.4384620930745093,
      "tokens_seen": 1944584192
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021622612067334973,
      "loss": 2.735,
      "theoretical_loss": 3.438452377800445,
      "tokens_seen": 1944649728
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021620957064497824,
      "loss": 2.4001,
      "theoretical_loss": 3.4384426629454587,
      "tokens_seen": 1944715264
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021619302061660678,
      "loss": 2.4272,
      "theoretical_loss": 3.438432948509517,
      "tokens_seen": 1944780800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002161764705882353,
      "loss": 2.4058,
      "theoretical_loss": 3.4384232344925882,
      "tokens_seen": 1944846336
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021615992055986383,
      "loss": 2.6324,
      "theoretical_loss": 3.438413520894641,
      "tokens_seen": 1944911872
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021614337053149234,
      "loss": 2.7474,
      "theoretical_loss": 3.438403807715641,
      "tokens_seen": 1944977408
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021612682050312087,
      "loss": 2.6478,
      "theoretical_loss": 3.4383940949555587,
      "tokens_seen": 1945042944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021611027047474938,
      "loss": 2.6683,
      "theoretical_loss": 3.43838438261436,
      "tokens_seen": 1945108480
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021609372044637792,
      "loss": 2.6446,
      "theoretical_loss": 3.4383746706920135,
      "tokens_seen": 1945174016
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021607717041800643,
      "loss": 2.7101,
      "theoretical_loss": 3.438364959188487,
      "tokens_seen": 1945239552
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021606062038963497,
      "loss": 2.6944,
      "theoretical_loss": 3.438355248103748,
      "tokens_seen": 1945305088
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021604407036126345,
      "loss": 2.6992,
      "theoretical_loss": 3.4383455374377645,
      "tokens_seen": 1945370624
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000216027520332892,
      "loss": 2.6989,
      "theoretical_loss": 3.4383358271905045,
      "tokens_seen": 1945436160
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002160109703045205,
      "loss": 2.5328,
      "theoretical_loss": 3.4383261173619353,
      "tokens_seen": 1945501696
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021599442027614904,
      "loss": 3.018,
      "theoretical_loss": 3.438316407952026,
      "tokens_seen": 1945567232
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021597787024777755,
      "loss": 2.848,
      "theoretical_loss": 3.438306698960743,
      "tokens_seen": 1945632768
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002159613202194061,
      "loss": 2.3605,
      "theoretical_loss": 3.4382969903880545,
      "tokens_seen": 1945698304
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002159447701910346,
      "loss": 2.5628,
      "theoretical_loss": 3.4382872822339294,
      "tokens_seen": 1945763840
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021592822016266314,
      "loss": 2.6161,
      "theoretical_loss": 3.4382775744983345,
      "tokens_seen": 1945829376
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021591167013429165,
      "loss": 1.8924,
      "theoretical_loss": 3.4382678671812377,
      "tokens_seen": 1945894912
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00021589512010592018,
      "loss": 2.6292,
      "theoretical_loss": 3.438258160282607,
      "tokens_seen": 1945960448
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002158785700775487,
      "loss": 2.5404,
      "theoretical_loss": 3.4382484538024105,
      "tokens_seen": 1946025984
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021586202004917723,
      "loss": 2.5782,
      "theoretical_loss": 3.438238747740616,
      "tokens_seen": 1946091520
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021584547002080574,
      "loss": 2.4387,
      "theoretical_loss": 3.4382290420971913,
      "tokens_seen": 1946157056
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021582891999243428,
      "loss": 2.6552,
      "theoretical_loss": 3.438219336872104,
      "tokens_seen": 1946222592
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002158123699640628,
      "loss": 2.6141,
      "theoretical_loss": 3.4382096320653224,
      "tokens_seen": 1946288128
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021579581993569133,
      "loss": 2.6154,
      "theoretical_loss": 3.4381999276768145,
      "tokens_seen": 1946353664
    },
    {
      "epoch": 0.7,
      "objective/train/avg_token_score": 0.03261714428663254,
      "objective/train/avg_weight": 0.9616279006004333,
      "objective/train/docs_used": 1099983,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.34322452545166,
      "objective/train/std_weight": 0.159959077835083,
      "objective/train/theoretical_loss": 3.438190223706548,
      "objective/train/tokens_used": 1966879200,
      "theoretical_loss": 3.438190223706548,
      "tokens_seen": 1946419200
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021577926990731984,
      "loss": 2.5027,
      "theoretical_loss": 3.438190223706548,
      "tokens_seen": 1946419200
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021576271987894837,
      "loss": 2.7118,
      "theoretical_loss": 3.43818052015449,
      "tokens_seen": 1946484736
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021574616985057689,
      "loss": 2.2412,
      "theoretical_loss": 3.43817081702061,
      "tokens_seen": 1946550272
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021572961982220542,
      "loss": 2.5731,
      "theoretical_loss": 3.4381611143048745,
      "tokens_seen": 1946615808
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021571306979383393,
      "loss": 2.6726,
      "theoretical_loss": 3.4381514120072523,
      "tokens_seen": 1946681344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021569651976546247,
      "loss": 2.6608,
      "theoretical_loss": 3.4381417101277107,
      "tokens_seen": 1946746880
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021567996973709098,
      "loss": 2.5022,
      "theoretical_loss": 3.438132008666218,
      "tokens_seen": 1946812416
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021566341970871952,
      "loss": 2.7284,
      "theoretical_loss": 3.438122307622742,
      "tokens_seen": 1946877952
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021564686968034803,
      "loss": 2.6463,
      "theoretical_loss": 3.43811260699725,
      "tokens_seen": 1946943488
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021563031965197657,
      "loss": 2.6553,
      "theoretical_loss": 3.4381029067897115,
      "tokens_seen": 1947009024
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021561376962360508,
      "loss": 2.5698,
      "theoretical_loss": 3.438093207000093,
      "tokens_seen": 1947074560
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021559721959523361,
      "loss": 2.5204,
      "theoretical_loss": 3.4380835076283622,
      "tokens_seen": 1947140096
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021558066956686213,
      "loss": 2.4449,
      "theoretical_loss": 3.4380738086744884,
      "tokens_seen": 1947205632
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021556411953849066,
      "loss": 2.4967,
      "theoretical_loss": 3.4380641101384386,
      "tokens_seen": 1947271168
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021554756951011915,
      "loss": 2.6106,
      "theoretical_loss": 3.4380544120201812,
      "tokens_seen": 1947336704
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021553101948174766,
      "loss": 2.3651,
      "theoretical_loss": 3.438044714319684,
      "tokens_seen": 1947402240
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002155144694533762,
      "loss": 2.7035,
      "theoretical_loss": 3.4380350170369143,
      "tokens_seen": 1947467776
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002154979194250047,
      "loss": 2.6263,
      "theoretical_loss": 3.438025320171841,
      "tokens_seen": 1947533312
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021548136939663324,
      "loss": 2.5213,
      "theoretical_loss": 3.438015623724432,
      "tokens_seen": 1947598848
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021546481936826175,
      "loss": 2.4197,
      "theoretical_loss": 3.4380059276946544,
      "tokens_seen": 1947664384
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002154482693398903,
      "loss": 2.6897,
      "theoretical_loss": 3.437996232082477,
      "tokens_seen": 1947729920
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002154317193115188,
      "loss": 2.5188,
      "theoretical_loss": 3.437986536887867,
      "tokens_seen": 1947795456
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021541516928314734,
      "loss": 2.9739,
      "theoretical_loss": 3.437976842110793,
      "tokens_seen": 1947860992
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021539861925477585,
      "loss": 2.9366,
      "theoretical_loss": 3.4379671477512233,
      "tokens_seen": 1947926528
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021538206922640439,
      "loss": 2.8407,
      "theoretical_loss": 3.4379574538091244,
      "tokens_seen": 1947992064
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002153655191980329,
      "loss": 2.5612,
      "theoretical_loss": 3.437947760284466,
      "tokens_seen": 1948057600
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021534896916966143,
      "loss": 2.4343,
      "theoretical_loss": 3.437938067177215,
      "tokens_seen": 1948123136
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021533241914128994,
      "loss": 2.8935,
      "theoretical_loss": 3.43792837448734,
      "tokens_seen": 1948188672
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021531586911291848,
      "loss": 2.6665,
      "theoretical_loss": 3.437918682214808,
      "tokens_seen": 1948254208
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000215299319084547,
      "loss": 2.3055,
      "theoretical_loss": 3.4379089903595883,
      "tokens_seen": 1948319744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021528276905617553,
      "loss": 2.4971,
      "theoretical_loss": 3.437899298921648,
      "tokens_seen": 1948385280
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021526621902780404,
      "loss": 2.6368,
      "theoretical_loss": 3.4378896079009555,
      "tokens_seen": 1948450816
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021524966899943258,
      "loss": 2.6883,
      "theoretical_loss": 3.437879917297479,
      "tokens_seen": 1948516352
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002152331189710611,
      "loss": 2.9136,
      "theoretical_loss": 3.437870227111186,
      "tokens_seen": 1948581888
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021521656894268963,
      "loss": 2.8215,
      "theoretical_loss": 3.4378605373420443,
      "tokens_seen": 1948647424
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021520001891431814,
      "loss": 2.452,
      "theoretical_loss": 3.4378508479900227,
      "tokens_seen": 1948712960
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021518346888594667,
      "loss": 2.5749,
      "theoretical_loss": 3.437841159055089,
      "tokens_seen": 1948778496
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021516691885757518,
      "loss": 2.6492,
      "theoretical_loss": 3.4378314705372106,
      "tokens_seen": 1948844032
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021515036882920372,
      "loss": 2.479,
      "theoretical_loss": 3.437821782436356,
      "tokens_seen": 1948909568
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021513381880083223,
      "loss": 2.589,
      "theoretical_loss": 3.4378120947524935,
      "tokens_seen": 1948975104
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021511726877246077,
      "loss": 2.5355,
      "theoretical_loss": 3.4378024074855906,
      "tokens_seen": 1949040640
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021510071874408928,
      "loss": 2.7466,
      "theoretical_loss": 3.437792720635616,
      "tokens_seen": 1949106176
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021508416871571782,
      "loss": 2.587,
      "theoretical_loss": 3.437783034202537,
      "tokens_seen": 1949171712
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021506761868734633,
      "loss": 2.8195,
      "theoretical_loss": 3.4377733481863215,
      "tokens_seen": 1949237248
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002150510686589748,
      "loss": 2.5755,
      "theoretical_loss": 3.4377636625869386,
      "tokens_seen": 1949302784
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021503451863060335,
      "loss": 2.658,
      "theoretical_loss": 3.4377539774043555,
      "tokens_seen": 1949368320
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021501796860223186,
      "loss": 2.7461,
      "theoretical_loss": 3.4377442926385404,
      "tokens_seen": 1949433856
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002150014185738604,
      "loss": 2.818,
      "theoretical_loss": 3.4377346082894618,
      "tokens_seen": 1949499392
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002149848685454889,
      "loss": 2.7863,
      "theoretical_loss": 3.437724924357087,
      "tokens_seen": 1949564928
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021496831851711745,
      "loss": 2.3749,
      "theoretical_loss": 3.437715240841385,
      "tokens_seen": 1949630464
    },
    {
      "epoch": 0.7,
      "objective/train/avg_token_score": 0.02370428293943405,
      "objective/train/avg_weight": 0.972098171710968,
      "objective/train/docs_used": 1101732,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.819997787475586,
      "objective/train/std_weight": 0.11668632924556732,
      "objective/train/theoretical_loss": 3.437705557742323,
      "objective/train/tokens_used": 1970156000,
      "theoretical_loss": 3.437705557742323,
      "tokens_seen": 1949696000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021495176848874596,
      "loss": 2.6658,
      "theoretical_loss": 3.437705557742323,
      "tokens_seen": 1949696000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002149352184603745,
      "loss": 2.5168,
      "theoretical_loss": 3.4376958750598696,
      "tokens_seen": 1949761536
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000214918668432003,
      "loss": 2.5107,
      "theoretical_loss": 3.437686192793992,
      "tokens_seen": 1949827072
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021490211840363154,
      "loss": 2.6804,
      "theoretical_loss": 3.4376765109446596,
      "tokens_seen": 1949892608
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021488556837526005,
      "loss": 2.5107,
      "theoretical_loss": 3.4376668295118398,
      "tokens_seen": 1949958144
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002148690183468886,
      "loss": 2.5878,
      "theoretical_loss": 3.4376571484955005,
      "tokens_seen": 1950023680
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002148524683185171,
      "loss": 2.6882,
      "theoretical_loss": 3.4376474678956104,
      "tokens_seen": 1950089216
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021483591829014564,
      "loss": 2.6054,
      "theoretical_loss": 3.437637787712137,
      "tokens_seen": 1950154752
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021481936826177415,
      "loss": 2.4623,
      "theoretical_loss": 3.4376281079450486,
      "tokens_seen": 1950220288
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021480281823340269,
      "loss": 2.644,
      "theoretical_loss": 3.437618428594313,
      "tokens_seen": 1950285824
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002147862682050312,
      "loss": 2.7198,
      "theoretical_loss": 3.4376087496598986,
      "tokens_seen": 1950351360
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021476971817665973,
      "loss": 2.402,
      "theoretical_loss": 3.437599071141774,
      "tokens_seen": 1950416896
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021475316814828824,
      "loss": 2.5162,
      "theoretical_loss": 3.4375893930399064,
      "tokens_seen": 1950482432
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021473661811991678,
      "loss": 2.5539,
      "theoretical_loss": 3.4375797153542647,
      "tokens_seen": 1950547968
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021472006809154532,
      "loss": 2.6664,
      "theoretical_loss": 3.437570038084816,
      "tokens_seen": 1950613504
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021470351806317383,
      "loss": 2.6418,
      "theoretical_loss": 3.4375603612315295,
      "tokens_seen": 1950679040
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021468696803480237,
      "loss": 2.6022,
      "theoretical_loss": 3.437550684794373,
      "tokens_seen": 1950744576
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021467041800643088,
      "loss": 2.6363,
      "theoretical_loss": 3.4375410087733145,
      "tokens_seen": 1950810112
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021465386797805941,
      "loss": 2.7308,
      "theoretical_loss": 3.437531333168322,
      "tokens_seen": 1950875648
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021463731794968793,
      "loss": 2.8834,
      "theoretical_loss": 3.4375216579793637,
      "tokens_seen": 1950941184
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021462076792131646,
      "loss": 2.3967,
      "theoretical_loss": 3.4375119832064076,
      "tokens_seen": 1951006720
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021460421789294497,
      "loss": 2.4903,
      "theoretical_loss": 3.4375023088494228,
      "tokens_seen": 1951072256
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002145876678645735,
      "loss": 2.6172,
      "theoretical_loss": 3.4374926349083763,
      "tokens_seen": 1951137792
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000214571117836202,
      "loss": 2.8051,
      "theoretical_loss": 3.4374829613832363,
      "tokens_seen": 1951203328
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002145545678078305,
      "loss": 2.7473,
      "theoretical_loss": 3.4374732882739716,
      "tokens_seen": 1951268864
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021453801777945904,
      "loss": 2.5879,
      "theoretical_loss": 3.43746361558055,
      "tokens_seen": 1951334400
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021452146775108755,
      "loss": 2.3878,
      "theoretical_loss": 3.4374539433029394,
      "tokens_seen": 1951399936
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002145049177227161,
      "loss": 2.5376,
      "theoretical_loss": 3.437444271441109,
      "tokens_seen": 1951465472
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002144883676943446,
      "loss": 2.6628,
      "theoretical_loss": 3.4374345999950253,
      "tokens_seen": 1951531008
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021447181766597314,
      "loss": 2.4394,
      "theoretical_loss": 3.437424928964658,
      "tokens_seen": 1951596544
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021445526763760165,
      "loss": 2.5264,
      "theoretical_loss": 3.437415258349975,
      "tokens_seen": 1951662080
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021443871760923019,
      "loss": 2.749,
      "theoretical_loss": 3.437405588150943,
      "tokens_seen": 1951727616
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002144221675808587,
      "loss": 2.438,
      "theoretical_loss": 3.437395918367532,
      "tokens_seen": 1951793152
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021440561755248723,
      "loss": 2.4916,
      "theoretical_loss": 3.43738624899971,
      "tokens_seen": 1951858688
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021438906752411574,
      "loss": 2.813,
      "theoretical_loss": 3.437376580047444,
      "tokens_seen": 1951924224
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021437251749574428,
      "loss": 2.683,
      "theoretical_loss": 3.437366911510703,
      "tokens_seen": 1951989760
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002143559674673728,
      "loss": 2.5927,
      "theoretical_loss": 3.437357243389455,
      "tokens_seen": 1952055296
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021433941743900133,
      "loss": 2.6611,
      "theoretical_loss": 3.437347575683668,
      "tokens_seen": 1952120832
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021432286741062984,
      "loss": 2.6736,
      "theoretical_loss": 3.437337908393311,
      "tokens_seen": 1952186368
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021430631738225838,
      "loss": 2.3799,
      "theoretical_loss": 3.437328241518352,
      "tokens_seen": 1952251904
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002142897673538869,
      "loss": 2.6583,
      "theoretical_loss": 3.4373185750587583,
      "tokens_seen": 1952317440
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021427321732551543,
      "loss": 2.5397,
      "theoretical_loss": 3.4373089090144986,
      "tokens_seen": 1952382976
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021425666729714394,
      "loss": 2.6673,
      "theoretical_loss": 3.4372992433855414,
      "tokens_seen": 1952448512
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021424011726877247,
      "loss": 2.3209,
      "theoretical_loss": 3.4372895781718547,
      "tokens_seen": 1952514048
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021422356724040098,
      "loss": 2.7308,
      "theoretical_loss": 3.437279913373407,
      "tokens_seen": 1952579584
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021420701721202952,
      "loss": 2.3445,
      "theoretical_loss": 3.437270248990166,
      "tokens_seen": 1952645120
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021419046718365803,
      "loss": 2.4404,
      "theoretical_loss": 3.4372605850221003,
      "tokens_seen": 1952710656
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021417391715528657,
      "loss": 2.5283,
      "theoretical_loss": 3.437250921469178,
      "tokens_seen": 1952776192
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021415736712691508,
      "loss": 2.5847,
      "theoretical_loss": 3.437241258331367,
      "tokens_seen": 1952841728
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021414081709854362,
      "loss": 2.661,
      "theoretical_loss": 3.437231595608636,
      "tokens_seen": 1952907264
    },
    {
      "epoch": 0.7,
      "objective/train/avg_token_score": 0.01880768872797489,
      "objective/train/avg_weight": 0.9778541922569275,
      "objective/train/docs_used": 1103876,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.8133177757263184,
      "objective/train/std_weight": 0.1227811649441719,
      "objective/train/theoretical_loss": 3.4372219333009535,
      "objective/train/tokens_used": 1973432800,
      "theoretical_loss": 3.4372219333009535,
      "tokens_seen": 1952972800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021412426707017213,
      "loss": 2.5098,
      "theoretical_loss": 3.4372219333009535,
      "tokens_seen": 1952972800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021410771704180067,
      "loss": 2.6411,
      "theoretical_loss": 3.437212271408287,
      "tokens_seen": 1953038336
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021409116701342918,
      "loss": 2.8095,
      "theoretical_loss": 3.4372026099306057,
      "tokens_seen": 1953103872
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021407461698505766,
      "loss": 2.6121,
      "theoretical_loss": 3.4371929488678767,
      "tokens_seen": 1953169408
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002140580669566862,
      "loss": 2.6668,
      "theoretical_loss": 3.4371832882200692,
      "tokens_seen": 1953234944
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002140415169283147,
      "loss": 2.7866,
      "theoretical_loss": 3.4371736279871508,
      "tokens_seen": 1953300480
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021402496689994325,
      "loss": 2.3847,
      "theoretical_loss": 3.43716396816909,
      "tokens_seen": 1953366016
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021400841687157176,
      "loss": 2.7593,
      "theoretical_loss": 3.4371543087658556,
      "tokens_seen": 1953431552
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002139918668432003,
      "loss": 2.4583,
      "theoretical_loss": 3.437144649777415,
      "tokens_seen": 1953497088
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002139753168148288,
      "loss": 2.8274,
      "theoretical_loss": 3.4371349912037372,
      "tokens_seen": 1953562624
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021395876678645734,
      "loss": 2.4955,
      "theoretical_loss": 3.43712533304479,
      "tokens_seen": 1953628160
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021394221675808585,
      "loss": 2.5448,
      "theoretical_loss": 3.437115675300542,
      "tokens_seen": 1953693696
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002139256667297144,
      "loss": 2.6564,
      "theoretical_loss": 3.4371060179709607,
      "tokens_seen": 1953759232
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002139091167013429,
      "loss": 2.5632,
      "theoretical_loss": 3.4370963610560157,
      "tokens_seen": 1953824768
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021389256667297144,
      "loss": 2.5876,
      "theoretical_loss": 3.437086704555674,
      "tokens_seen": 1953890304
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021387601664459995,
      "loss": 2.561,
      "theoretical_loss": 3.437077048469905,
      "tokens_seen": 1953955840
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021385946661622848,
      "loss": 2.6834,
      "theoretical_loss": 3.437067392798676,
      "tokens_seen": 1954021376
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000213842916587857,
      "loss": 2.5847,
      "theoretical_loss": 3.4370577375419566,
      "tokens_seen": 1954086912
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021382636655948553,
      "loss": 2.7305,
      "theoretical_loss": 3.4370480826997136,
      "tokens_seen": 1954152448
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021380981653111404,
      "loss": 2.5938,
      "theoretical_loss": 3.437038428271916,
      "tokens_seen": 1954217984
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021379326650274258,
      "loss": 2.6908,
      "theoretical_loss": 3.4370287742585326,
      "tokens_seen": 1954283520
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002137767164743711,
      "loss": 2.6353,
      "theoretical_loss": 3.4370191206595306,
      "tokens_seen": 1954349056
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021376016644599963,
      "loss": 2.7442,
      "theoretical_loss": 3.437009467474879,
      "tokens_seen": 1954414592
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021374361641762814,
      "loss": 2.7009,
      "theoretical_loss": 3.4369998147045466,
      "tokens_seen": 1954480128
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021372706638925668,
      "loss": 2.8479,
      "theoretical_loss": 3.4369901623485006,
      "tokens_seen": 1954545664
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002137105163608852,
      "loss": 2.7015,
      "theoretical_loss": 3.43698051040671,
      "tokens_seen": 1954611200
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021369396633251372,
      "loss": 2.6944,
      "theoretical_loss": 3.4369708588791434,
      "tokens_seen": 1954676736
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021367741630414224,
      "loss": 2.7506,
      "theoretical_loss": 3.4369612077657687,
      "tokens_seen": 1954742272
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021366086627577077,
      "loss": 2.4955,
      "theoretical_loss": 3.4369515570665543,
      "tokens_seen": 1954807808
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021364431624739928,
      "loss": 2.3677,
      "theoretical_loss": 3.4369419067814686,
      "tokens_seen": 1954873344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021362776621902782,
      "loss": 2.7035,
      "theoretical_loss": 3.4369322569104797,
      "tokens_seen": 1954938880
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021361121619065633,
      "loss": 2.4171,
      "theoretical_loss": 3.4369226074535564,
      "tokens_seen": 1955004416
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021359466616228487,
      "loss": 2.781,
      "theoretical_loss": 3.436912958410667,
      "tokens_seen": 1955069952
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021357811613391335,
      "loss": 2.7404,
      "theoretical_loss": 3.436903309781779,
      "tokens_seen": 1955135488
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021356156610554186,
      "loss": 2.8039,
      "theoretical_loss": 3.436893661566862,
      "tokens_seen": 1955201024
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002135450160771704,
      "loss": 2.573,
      "theoretical_loss": 3.4368840137658836,
      "tokens_seen": 1955266560
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002135284660487989,
      "loss": 2.3664,
      "theoretical_loss": 3.436874366378812,
      "tokens_seen": 1955332096
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021351191602042745,
      "loss": 2.6929,
      "theoretical_loss": 3.4368647194056168,
      "tokens_seen": 1955397632
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021349536599205599,
      "loss": 2.5256,
      "theoretical_loss": 3.436855072846265,
      "tokens_seen": 1955463168
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002134788159636845,
      "loss": 2.429,
      "theoretical_loss": 3.4368454267007253,
      "tokens_seen": 1955528704
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021346226593531303,
      "loss": 2.6233,
      "theoretical_loss": 3.4368357809689667,
      "tokens_seen": 1955594240
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021344571590694154,
      "loss": 2.8174,
      "theoretical_loss": 3.436826135650957,
      "tokens_seen": 1955659776
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021342916587857008,
      "loss": 2.6965,
      "theoretical_loss": 3.436816490746665,
      "tokens_seen": 1955725312
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002134126158501986,
      "loss": 2.4248,
      "theoretical_loss": 3.436806846256058,
      "tokens_seen": 1955790848
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021339606582182713,
      "loss": 2.6529,
      "theoretical_loss": 3.436797202179106,
      "tokens_seen": 1955856384
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021337951579345564,
      "loss": 2.5011,
      "theoretical_loss": 3.4367875585157766,
      "tokens_seen": 1955921920
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021336296576508418,
      "loss": 2.7456,
      "theoretical_loss": 3.4367779152660383,
      "tokens_seen": 1955987456
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002133464157367127,
      "loss": 2.8051,
      "theoretical_loss": 3.436768272429859,
      "tokens_seen": 1956052992
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021332986570834123,
      "loss": 2.6629,
      "theoretical_loss": 3.4367586300072075,
      "tokens_seen": 1956118528
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021331331567996974,
      "loss": 2.6822,
      "theoretical_loss": 3.4367489879980524,
      "tokens_seen": 1956184064
    },
    {
      "epoch": 0.7,
      "objective/train/avg_token_score": 0.005943411495536566,
      "objective/train/avg_weight": 0.9930034875869751,
      "objective/train/docs_used": 1105993,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0604453086853027,
      "objective/train/std_weight": 0.039037611335515976,
      "objective/train/theoretical_loss": 3.436739346402362,
      "objective/train/tokens_used": 1976709600,
      "theoretical_loss": 3.436739346402362,
      "tokens_seen": 1956249600
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021329676565159827,
      "loss": 2.6373,
      "theoretical_loss": 3.436739346402362,
      "tokens_seen": 1956249600
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021328021562322678,
      "loss": 2.4423,
      "theoretical_loss": 3.436729705220105,
      "tokens_seen": 1956315136
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021326366559485532,
      "loss": 2.758,
      "theoretical_loss": 3.436720064451249,
      "tokens_seen": 1956380672
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021324711556648383,
      "loss": 2.6888,
      "theoretical_loss": 3.4367104240957635,
      "tokens_seen": 1956446208
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021323056553811237,
      "loss": 2.5037,
      "theoretical_loss": 3.436700784153616,
      "tokens_seen": 1956511744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021321401550974088,
      "loss": 2.7789,
      "theoretical_loss": 3.4366911446247754,
      "tokens_seen": 1956577280
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021319746548136942,
      "loss": 2.7227,
      "theoretical_loss": 3.43668150550921,
      "tokens_seen": 1956642816
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021318091545299793,
      "loss": 2.4863,
      "theoretical_loss": 3.4366718668068885,
      "tokens_seen": 1956708352
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021316436542462647,
      "loss": 2.594,
      "theoretical_loss": 3.436662228517779,
      "tokens_seen": 1956773888
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021314781539625498,
      "loss": 2.6947,
      "theoretical_loss": 3.4366525906418497,
      "tokens_seen": 1956839424
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002131312653678835,
      "loss": 2.5205,
      "theoretical_loss": 3.43664295317907,
      "tokens_seen": 1956904960
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021311471533951202,
      "loss": 2.9181,
      "theoretical_loss": 3.4366333161294076,
      "tokens_seen": 1956970496
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021309816531114056,
      "loss": 2.6665,
      "theoretical_loss": 3.436623679492831,
      "tokens_seen": 1957036032
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021308161528276904,
      "loss": 2.5654,
      "theoretical_loss": 3.436614043269309,
      "tokens_seen": 1957101568
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021306506525439756,
      "loss": 2.2052,
      "theoretical_loss": 3.43660440745881,
      "tokens_seen": 1957167104
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002130485152260261,
      "loss": 2.5003,
      "theoretical_loss": 3.4365947720613015,
      "tokens_seen": 1957232640
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002130319651976546,
      "loss": 2.6766,
      "theoretical_loss": 3.4365851370767535,
      "tokens_seen": 1957298176
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021301541516928314,
      "loss": 2.4885,
      "theoretical_loss": 3.436575502505134,
      "tokens_seen": 1957363712
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021299886514091165,
      "loss": 2.6117,
      "theoretical_loss": 3.4365658683464106,
      "tokens_seen": 1957429248
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002129823151125402,
      "loss": 2.7449,
      "theoretical_loss": 3.4365562346005527,
      "tokens_seen": 1957494784
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002129657650841687,
      "loss": 2.7373,
      "theoretical_loss": 3.436546601267529,
      "tokens_seen": 1957560320
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021294921505579724,
      "loss": 2.7369,
      "theoretical_loss": 3.4365369683473066,
      "tokens_seen": 1957625856
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021293266502742575,
      "loss": 2.8839,
      "theoretical_loss": 3.4365273358398554,
      "tokens_seen": 1957691392
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021291611499905428,
      "loss": 2.4863,
      "theoretical_loss": 3.4365177037451433,
      "tokens_seen": 1957756928
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002128995649706828,
      "loss": 2.5525,
      "theoretical_loss": 3.4365080720631394,
      "tokens_seen": 1957822464
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021288301494231133,
      "loss": 2.5627,
      "theoretical_loss": 3.4364984407938115,
      "tokens_seen": 1957888000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021286646491393984,
      "loss": 2.6234,
      "theoretical_loss": 3.436488809937128,
      "tokens_seen": 1957953536
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021284991488556838,
      "loss": 2.6948,
      "theoretical_loss": 3.4364791794930576,
      "tokens_seen": 1958019072
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002128333648571969,
      "loss": 2.7689,
      "theoretical_loss": 3.4364695494615694,
      "tokens_seen": 1958084608
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021281681482882543,
      "loss": 2.7842,
      "theoretical_loss": 3.4364599198426307,
      "tokens_seen": 1958150144
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021280026480045394,
      "loss": 2.5009,
      "theoretical_loss": 3.4364502906362118,
      "tokens_seen": 1958215680
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021278371477208248,
      "loss": 2.5593,
      "theoretical_loss": 3.4364406618422794,
      "tokens_seen": 1958281216
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000212767164743711,
      "loss": 2.8443,
      "theoretical_loss": 3.436431033460803,
      "tokens_seen": 1958346752
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021275061471533952,
      "loss": 2.4084,
      "theoretical_loss": 3.4364214054917515,
      "tokens_seen": 1958412288
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021273406468696804,
      "loss": 2.5768,
      "theoretical_loss": 3.436411777935092,
      "tokens_seen": 1958477824
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021271751465859657,
      "loss": 2.8348,
      "theoretical_loss": 3.4364021507907943,
      "tokens_seen": 1958543360
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021270096463022508,
      "loss": 2.5737,
      "theoretical_loss": 3.4363925240588267,
      "tokens_seen": 1958608896
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021268441460185362,
      "loss": 2.7286,
      "theoretical_loss": 3.4363828977391577,
      "tokens_seen": 1958674432
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021266786457348213,
      "loss": 2.7164,
      "theoretical_loss": 3.436373271831755,
      "tokens_seen": 1958739968
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021265131454511067,
      "loss": 2.497,
      "theoretical_loss": 3.4363636463365888,
      "tokens_seen": 1958805504
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021263476451673918,
      "loss": 2.6551,
      "theoretical_loss": 3.436354021253626,
      "tokens_seen": 1958871040
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021261821448836772,
      "loss": 2.3693,
      "theoretical_loss": 3.436344396582837,
      "tokens_seen": 1958936576
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002126016644599962,
      "loss": 2.6995,
      "theoretical_loss": 3.436334772324188,
      "tokens_seen": 1959002112
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002125851144316247,
      "loss": 2.6662,
      "theoretical_loss": 3.43632514847765,
      "tokens_seen": 1959067648
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021256856440325325,
      "loss": 2.6955,
      "theoretical_loss": 3.4363155250431894,
      "tokens_seen": 1959133184
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021255201437488176,
      "loss": 2.6753,
      "theoretical_loss": 3.4363059020207762,
      "tokens_seen": 1959198720
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002125354643465103,
      "loss": 2.5318,
      "theoretical_loss": 3.436296279410379,
      "tokens_seen": 1959264256
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002125189143181388,
      "loss": 1.9546,
      "theoretical_loss": 3.436286657211965,
      "tokens_seen": 1959329792
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021250236428976734,
      "loss": 2.709,
      "theoretical_loss": 3.4362770354255043,
      "tokens_seen": 1959395328
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021248581426139585,
      "loss": 2.5795,
      "theoretical_loss": 3.436267414050965,
      "tokens_seen": 1959460864
    },
    {
      "epoch": 0.7,
      "objective/train/avg_token_score": 0.005580605939030647,
      "objective/train/avg_weight": 0.9934310913085938,
      "objective/train/docs_used": 1108290,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9728899002075195,
      "objective/train/std_weight": 0.03248818591237068,
      "objective/train/theoretical_loss": 3.4362577930883154,
      "objective/train/tokens_used": 1979986400,
      "theoretical_loss": 3.4362577930883154,
      "tokens_seen": 1959526400
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002124692642330244,
      "loss": 2.8821,
      "theoretical_loss": 3.4362577930883154,
      "tokens_seen": 1959526400
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002124527142046529,
      "loss": 2.4999,
      "theoretical_loss": 3.436248172537524,
      "tokens_seen": 1959591936
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021243616417628144,
      "loss": 2.6987,
      "theoretical_loss": 3.43623855239856,
      "tokens_seen": 1959657472
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021241961414790995,
      "loss": 2.7354,
      "theoretical_loss": 3.4362289326713915,
      "tokens_seen": 1959723008
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002124030641195385,
      "loss": 2.6119,
      "theoretical_loss": 3.4362193133559877,
      "tokens_seen": 1959788544
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000212386514091167,
      "loss": 2.7191,
      "theoretical_loss": 3.4362096944523164,
      "tokens_seen": 1959854080
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021236996406279554,
      "loss": 2.8386,
      "theoretical_loss": 3.436200075960347,
      "tokens_seen": 1959919616
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021235341403442405,
      "loss": 2.8295,
      "theoretical_loss": 3.4361904578800466,
      "tokens_seen": 1959985152
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021233686400605258,
      "loss": 2.7231,
      "theoretical_loss": 3.436180840211386,
      "tokens_seen": 1960050688
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002123203139776811,
      "loss": 2.6541,
      "theoretical_loss": 3.436171222954332,
      "tokens_seen": 1960116224
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021230376394930963,
      "loss": 2.7289,
      "theoretical_loss": 3.436161606108855,
      "tokens_seen": 1960181760
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021228721392093814,
      "loss": 2.6678,
      "theoretical_loss": 3.436151989674922,
      "tokens_seen": 1960247296
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021227066389256668,
      "loss": 2.515,
      "theoretical_loss": 3.436142373652502,
      "tokens_seen": 1960312832
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002122541138641952,
      "loss": 2.5227,
      "theoretical_loss": 3.4361327580415644,
      "tokens_seen": 1960378368
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021223756383582373,
      "loss": 2.6061,
      "theoretical_loss": 3.436123142842077,
      "tokens_seen": 1960443904
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021222101380745224,
      "loss": 2.7221,
      "theoretical_loss": 3.4361135280540087,
      "tokens_seen": 1960509440
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021220446377908078,
      "loss": 2.5536,
      "theoretical_loss": 3.436103913677328,
      "tokens_seen": 1960574976
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021218791375070929,
      "loss": 2.8247,
      "theoretical_loss": 3.4360942997120043,
      "tokens_seen": 1960640512
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021217136372233782,
      "loss": 2.6685,
      "theoretical_loss": 3.436084686158005,
      "tokens_seen": 1960706048
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021215481369396633,
      "loss": 2.6332,
      "theoretical_loss": 3.4360750730153002,
      "tokens_seen": 1960771584
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021213826366559487,
      "loss": 2.7049,
      "theoretical_loss": 3.4360654602838574,
      "tokens_seen": 1960837120
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021212171363722338,
      "loss": 2.4512,
      "theoretical_loss": 3.4360558479636456,
      "tokens_seen": 1960902656
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002121051636088519,
      "loss": 2.6403,
      "theoretical_loss": 3.4360462360546333,
      "tokens_seen": 1960968192
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002120886135804804,
      "loss": 2.5353,
      "theoretical_loss": 3.43603662455679,
      "tokens_seen": 1961033728
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021207206355210894,
      "loss": 2.5614,
      "theoretical_loss": 3.436027013470083,
      "tokens_seen": 1961099264
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021205551352373745,
      "loss": 2.6489,
      "theoretical_loss": 3.436017402794482,
      "tokens_seen": 1961164800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000212038963495366,
      "loss": 2.7491,
      "theoretical_loss": 3.4360077925299555,
      "tokens_seen": 1961230336
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002120224134669945,
      "loss": 2.3862,
      "theoretical_loss": 3.435998182676472,
      "tokens_seen": 1961295872
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021200586343862304,
      "loss": 2.7714,
      "theoretical_loss": 3.4359885732340008,
      "tokens_seen": 1961361408
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021198931341025155,
      "loss": 2.4872,
      "theoretical_loss": 3.4359789642025094,
      "tokens_seen": 1961426944
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021197276338188008,
      "loss": 2.4138,
      "theoretical_loss": 3.4359693555819675,
      "tokens_seen": 1961492480
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002119562133535086,
      "loss": 2.549,
      "theoretical_loss": 3.435959747372343,
      "tokens_seen": 1961558016
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021193966332513713,
      "loss": 2.5328,
      "theoretical_loss": 3.4359501395736056,
      "tokens_seen": 1961623552
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021192311329676564,
      "loss": 2.5777,
      "theoretical_loss": 3.4359405321857235,
      "tokens_seen": 1961689088
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021190656326839418,
      "loss": 2.4385,
      "theoretical_loss": 3.4359309252086647,
      "tokens_seen": 1961754624
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002118900132400227,
      "loss": 2.7121,
      "theoretical_loss": 3.435921318642399,
      "tokens_seen": 1961820160
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021187346321165123,
      "loss": 2.5974,
      "theoretical_loss": 3.4359117124868948,
      "tokens_seen": 1961885696
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021185691318327974,
      "loss": 2.6946,
      "theoretical_loss": 3.4359021067421205,
      "tokens_seen": 1961951232
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021184036315490828,
      "loss": 2.7344,
      "theoretical_loss": 3.435892501408045,
      "tokens_seen": 1962016768
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002118238131265368,
      "loss": 2.4634,
      "theoretical_loss": 3.435882896484637,
      "tokens_seen": 1962082304
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021180726309816532,
      "loss": 2.3695,
      "theoretical_loss": 3.4358732919718653,
      "tokens_seen": 1962147840
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021179071306979383,
      "loss": 2.5921,
      "theoretical_loss": 3.435863687869699,
      "tokens_seen": 1962213376
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021177416304142237,
      "loss": 2.6662,
      "theoretical_loss": 3.4358540841781062,
      "tokens_seen": 1962278912
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021175761301305088,
      "loss": 2.5043,
      "theoretical_loss": 3.435844480897056,
      "tokens_seen": 1962344448
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021174106298467942,
      "loss": 2.7334,
      "theoretical_loss": 3.435834878026516,
      "tokens_seen": 1962409984
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021172451295630793,
      "loss": 2.7305,
      "theoretical_loss": 3.435825275566457,
      "tokens_seen": 1962475520
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021170796292793647,
      "loss": 2.8148,
      "theoretical_loss": 3.4358156735168466,
      "tokens_seen": 1962541056
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021169141289956498,
      "loss": 2.2694,
      "theoretical_loss": 3.4358060718776535,
      "tokens_seen": 1962606592
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021167486287119352,
      "loss": 2.5433,
      "theoretical_loss": 3.435796470648846,
      "tokens_seen": 1962672128
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021165831284282203,
      "loss": 2.5667,
      "theoretical_loss": 3.4357868698303946,
      "tokens_seen": 1962737664
    },
    {
      "epoch": 0.7,
      "objective/train/avg_token_score": 0.0038467964623123407,
      "objective/train/avg_weight": 0.9954769015312195,
      "objective/train/docs_used": 1110185,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.857368230819702,
      "objective/train/std_weight": 0.03322015702724457,
      "objective/train/theoretical_loss": 3.435777269422266,
      "objective/train/tokens_used": 1983263200,
      "theoretical_loss": 3.435777269422266,
      "tokens_seen": 1962803200
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021164176281445056,
      "loss": 2.8578,
      "theoretical_loss": 3.435777269422266,
      "tokens_seen": 1962803200
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021162521278607907,
      "loss": 2.7211,
      "theoretical_loss": 3.4357676694244303,
      "tokens_seen": 1962868736
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021160866275770756,
      "loss": 2.9623,
      "theoretical_loss": 3.4357580698368557,
      "tokens_seen": 1962934272
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002115921127293361,
      "loss": 2.3815,
      "theoretical_loss": 3.435748470659511,
      "tokens_seen": 1962999808
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002115755627009646,
      "loss": 2.5535,
      "theoretical_loss": 3.4357388718923656,
      "tokens_seen": 1963065344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021155901267259314,
      "loss": 2.6512,
      "theoretical_loss": 3.4357292735353875,
      "tokens_seen": 1963130880
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021154246264422165,
      "loss": 2.6936,
      "theoretical_loss": 3.4357196755885457,
      "tokens_seen": 1963196416
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002115259126158502,
      "loss": 2.7434,
      "theoretical_loss": 3.435710078051809,
      "tokens_seen": 1963261952
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002115093625874787,
      "loss": 2.5091,
      "theoretical_loss": 3.435700480925146,
      "tokens_seen": 1963327488
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021149281255910724,
      "loss": 2.734,
      "theoretical_loss": 3.435690884208526,
      "tokens_seen": 1963393024
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021147626253073575,
      "loss": 2.632,
      "theoretical_loss": 3.4356812879019176,
      "tokens_seen": 1963458560
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002114597125023643,
      "loss": 2.5375,
      "theoretical_loss": 3.4356716920052897,
      "tokens_seen": 1963524096
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002114431624739928,
      "loss": 2.6477,
      "theoretical_loss": 3.4356620965186107,
      "tokens_seen": 1963589632
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021142661244562134,
      "loss": 2.4545,
      "theoretical_loss": 3.435652501441849,
      "tokens_seen": 1963655168
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021141006241724985,
      "loss": 2.6162,
      "theoretical_loss": 3.4356429067749747,
      "tokens_seen": 1963720704
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021139351238887838,
      "loss": 2.542,
      "theoretical_loss": 3.4356333125179557,
      "tokens_seen": 1963786240
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002113769623605069,
      "loss": 2.3954,
      "theoretical_loss": 3.435623718670761,
      "tokens_seen": 1963851776
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021136041233213543,
      "loss": 2.5062,
      "theoretical_loss": 3.43561412523336,
      "tokens_seen": 1963917312
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021134386230376394,
      "loss": 2.387,
      "theoretical_loss": 3.4356045322057205,
      "tokens_seen": 1963982848
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021132731227539248,
      "loss": 2.6546,
      "theoretical_loss": 3.435594939587812,
      "tokens_seen": 1964048384
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000211310762247021,
      "loss": 2.8666,
      "theoretical_loss": 3.435585347379603,
      "tokens_seen": 1964113920
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021129421221864953,
      "loss": 2.6643,
      "theoretical_loss": 3.4355757555810627,
      "tokens_seen": 1964179456
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021127766219027804,
      "loss": 2.5725,
      "theoretical_loss": 3.435566164192159,
      "tokens_seen": 1964244992
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021126111216190658,
      "loss": 2.5241,
      "theoretical_loss": 3.435556573212862,
      "tokens_seen": 1964310528
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021124456213353509,
      "loss": 2.7974,
      "theoretical_loss": 3.4355469826431406,
      "tokens_seen": 1964376064
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021122801210516362,
      "loss": 2.486,
      "theoretical_loss": 3.435537392482962,
      "tokens_seen": 1964441600
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021121146207679213,
      "loss": 2.6555,
      "theoretical_loss": 3.4355278027322966,
      "tokens_seen": 1964507136
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021119491204842067,
      "loss": 2.5796,
      "theoretical_loss": 3.435518213391113,
      "tokens_seen": 1964572672
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021117836202004918,
      "loss": 2.6574,
      "theoretical_loss": 3.4355086244593793,
      "tokens_seen": 1964638208
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021116181199167772,
      "loss": 2.6014,
      "theoretical_loss": 3.435499035937065,
      "tokens_seen": 1964703744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021114526196330623,
      "loss": 2.7136,
      "theoretical_loss": 3.435489447824139,
      "tokens_seen": 1964769280
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002111287119349347,
      "loss": 2.8102,
      "theoretical_loss": 3.43547986012057,
      "tokens_seen": 1964834816
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021111216190656325,
      "loss": 2.7222,
      "theoretical_loss": 3.4354702728263264,
      "tokens_seen": 1964900352
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021109561187819176,
      "loss": 2.7158,
      "theoretical_loss": 3.435460685941378,
      "tokens_seen": 1964965888
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002110790618498203,
      "loss": 2.6204,
      "theoretical_loss": 3.435451099465693,
      "tokens_seen": 1965031424
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002110625118214488,
      "loss": 2.6676,
      "theoretical_loss": 3.4354415133992404,
      "tokens_seen": 1965096960
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021104596179307735,
      "loss": 2.4991,
      "theoretical_loss": 3.4354319277419894,
      "tokens_seen": 1965162496
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021102941176470586,
      "loss": 2.7171,
      "theoretical_loss": 3.4354223424939088,
      "tokens_seen": 1965228032
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002110128617363344,
      "loss": 2.6043,
      "theoretical_loss": 3.435412757654967,
      "tokens_seen": 1965293568
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002109963117079629,
      "loss": 2.7387,
      "theoretical_loss": 3.4354031732251333,
      "tokens_seen": 1965359104
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021097976167959144,
      "loss": 2.6893,
      "theoretical_loss": 3.4353935892043768,
      "tokens_seen": 1965424640
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021096321165121995,
      "loss": 2.6805,
      "theoretical_loss": 3.435384005592666,
      "tokens_seen": 1965490176
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002109466616228485,
      "loss": 2.5253,
      "theoretical_loss": 3.4353744223899696,
      "tokens_seen": 1965555712
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000210930111594477,
      "loss": 2.6281,
      "theoretical_loss": 3.4353648395962573,
      "tokens_seen": 1965621248
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021091356156610554,
      "loss": 2.7155,
      "theoretical_loss": 3.435355257211498,
      "tokens_seen": 1965686784
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021089701153773405,
      "loss": 2.6133,
      "theoretical_loss": 3.4353456752356593,
      "tokens_seen": 1965752320
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002108804615093626,
      "loss": 2.2422,
      "theoretical_loss": 3.4353360936687114,
      "tokens_seen": 1965817856
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002108639114809911,
      "loss": 2.6506,
      "theoretical_loss": 3.4353265125106223,
      "tokens_seen": 1965883392
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021084736145261963,
      "loss": 2.7816,
      "theoretical_loss": 3.4353169317613617,
      "tokens_seen": 1965948928
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021083081142424814,
      "loss": 2.8749,
      "theoretical_loss": 3.4353073514208985,
      "tokens_seen": 1966014464
    },
    {
      "debugging/Self-BLEU-5": 0.4796244782337672,
      "debugging/distinct-1-grams": 0.7874466763773558,
      "debugging/distinct-2-grams": 0.9537592626603371,
      "debugging/entropy-1-grams": 6.225892551400493,
      "debugging/entropy-2-grams": 7.143796200176747,
      "debugging/length": 450.05,
      "debugging/num_segments": 20,
      "epoch": 0.7,
      "objective/train/avg_token_score": 0.028914527967572212,
      "objective/train/avg_weight": 0.9659799933433533,
      "objective/train/docs_used": 1112341,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7207915782928467,
      "objective/train/std_weight": 0.11091215163469315,
      "objective/train/theoretical_loss": 3.4352977714892017,
      "objective/train/tokens_used": 1986540000,
      "theoretical_loss": 3.4352977714892017,
      "tokens_seen": 1966080000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021081426139587668,
      "loss": 2.5824,
      "theoretical_loss": 3.4352977714892017,
      "tokens_seen": 1966080000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002107977113675052,
      "loss": 2.6013,
      "theoretical_loss": 3.4352881919662392,
      "tokens_seen": 1966145536
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021078116133913373,
      "loss": 2.5876,
      "theoretical_loss": 3.435278612851981,
      "tokens_seen": 1966211072
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021076461131076224,
      "loss": 2.6999,
      "theoretical_loss": 3.435269034146396,
      "tokens_seen": 1966276608
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021074806128239078,
      "loss": 2.5714,
      "theoretical_loss": 3.4352594558494527,
      "tokens_seen": 1966342144
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021073151125401932,
      "loss": 2.549,
      "theoretical_loss": 3.4352498779611196,
      "tokens_seen": 1966407680
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021071496122564783,
      "loss": 2.4285,
      "theoretical_loss": 3.435240300481367,
      "tokens_seen": 1966473216
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021069841119727636,
      "loss": 2.5464,
      "theoretical_loss": 3.4352307234101627,
      "tokens_seen": 1966538752
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021068186116890487,
      "loss": 2.4518,
      "theoretical_loss": 3.4352211467474767,
      "tokens_seen": 1966604288
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002106653111405334,
      "loss": 2.6688,
      "theoretical_loss": 3.4352115704932764,
      "tokens_seen": 1966669824
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021064876111216192,
      "loss": 2.5466,
      "theoretical_loss": 3.4352019946475325,
      "tokens_seen": 1966735360
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002106322110837904,
      "loss": 2.7883,
      "theoretical_loss": 3.435192419210213,
      "tokens_seen": 1966800896
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021061566105541894,
      "loss": 2.5221,
      "theoretical_loss": 3.435182844181287,
      "tokens_seen": 1966866432
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021059911102704745,
      "loss": 2.4589,
      "theoretical_loss": 3.435173269560723,
      "tokens_seen": 1966931968
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000210582560998676,
      "loss": 2.3029,
      "theoretical_loss": 3.435163695348491,
      "tokens_seen": 1966997504
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002105660109703045,
      "loss": 2.5479,
      "theoretical_loss": 3.435154121544559,
      "tokens_seen": 1967063040
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021054946094193304,
      "loss": 2.3516,
      "theoretical_loss": 3.4351445481488967,
      "tokens_seen": 1967128576
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021053291091356155,
      "loss": 2.664,
      "theoretical_loss": 3.435134975161473,
      "tokens_seen": 1967194112
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002105163608851901,
      "loss": 2.6118,
      "theoretical_loss": 3.4351254025822566,
      "tokens_seen": 1967259648
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002104998108568186,
      "loss": 2.5562,
      "theoretical_loss": 3.435115830411217,
      "tokens_seen": 1967325184
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021048326082844714,
      "loss": 2.4346,
      "theoretical_loss": 3.4351062586483225,
      "tokens_seen": 1967390720
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021046671080007565,
      "loss": 2.6729,
      "theoretical_loss": 3.4350966872935422,
      "tokens_seen": 1967456256
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021045016077170418,
      "loss": 2.8514,
      "theoretical_loss": 3.4350871163468457,
      "tokens_seen": 1967521792
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002104336107433327,
      "loss": 2.5205,
      "theoretical_loss": 3.435077545808202,
      "tokens_seen": 1967587328
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021041706071496123,
      "loss": 2.3046,
      "theoretical_loss": 3.435067975677579,
      "tokens_seen": 1967652864
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021040051068658974,
      "loss": 2.7144,
      "theoretical_loss": 3.4350584059549467,
      "tokens_seen": 1967718400
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021038396065821828,
      "loss": 2.6239,
      "theoretical_loss": 3.435048836640274,
      "tokens_seen": 1967783936
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002103674106298468,
      "loss": 2.5176,
      "theoretical_loss": 3.4350392677335297,
      "tokens_seen": 1967849472
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021035086060147533,
      "loss": 2.6611,
      "theoretical_loss": 3.435029699234683,
      "tokens_seen": 1967915008
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021033431057310384,
      "loss": 2.7671,
      "theoretical_loss": 3.435020131143703,
      "tokens_seen": 1967980544
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021031776054473238,
      "loss": 2.7922,
      "theoretical_loss": 3.4350105634605583,
      "tokens_seen": 1968046080
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021030121051636089,
      "loss": 2.5753,
      "theoretical_loss": 3.435000996185219,
      "tokens_seen": 1968111616
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021028466048798942,
      "loss": 2.463,
      "theoretical_loss": 3.4349914293176527,
      "tokens_seen": 1968177152
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021026811045961793,
      "loss": 2.6552,
      "theoretical_loss": 3.4349818628578292,
      "tokens_seen": 1968242688
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021025156043124647,
      "loss": 2.7998,
      "theoretical_loss": 3.4349722968057175,
      "tokens_seen": 1968308224
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021023501040287498,
      "loss": 2.416,
      "theoretical_loss": 3.4349627311612867,
      "tokens_seen": 1968373760
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021021846037450352,
      "loss": 2.5221,
      "theoretical_loss": 3.434953165924506,
      "tokens_seen": 1968439296
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021020191034613203,
      "loss": 2.3541,
      "theoretical_loss": 3.434943601095344,
      "tokens_seen": 1968504832
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021018536031776057,
      "loss": 2.5245,
      "theoretical_loss": 3.43493403667377,
      "tokens_seen": 1968570368
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021016881028938908,
      "loss": 2.4413,
      "theoretical_loss": 3.4349244726597528,
      "tokens_seen": 1968635904
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021015226026101762,
      "loss": 2.5099,
      "theoretical_loss": 3.434914909053262,
      "tokens_seen": 1968701440
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002101357102326461,
      "loss": 2.1581,
      "theoretical_loss": 3.434905345854266,
      "tokens_seen": 1968766976
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002101191602042746,
      "loss": 2.8891,
      "theoretical_loss": 3.4348957830627347,
      "tokens_seen": 1968832512
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021010261017590315,
      "loss": 2.5429,
      "theoretical_loss": 3.434886220678637,
      "tokens_seen": 1968898048
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021008606014753166,
      "loss": 2.5335,
      "theoretical_loss": 3.434876658701941,
      "tokens_seen": 1968963584
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002100695101191602,
      "loss": 2.825,
      "theoretical_loss": 3.434867097132617,
      "tokens_seen": 1969029120
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002100529600907887,
      "loss": 2.7102,
      "theoretical_loss": 3.4348575359706333,
      "tokens_seen": 1969094656
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021003641006241724,
      "loss": 2.7673,
      "theoretical_loss": 3.43484797521596,
      "tokens_seen": 1969160192
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00021001986003404575,
      "loss": 2.4906,
      "theoretical_loss": 3.4348384148685644,
      "tokens_seen": 1969225728
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002100033100056743,
      "loss": 2.5048,
      "theoretical_loss": 3.4348288549284174,
      "tokens_seen": 1969291264
    },
    {
      "epoch": 0.7,
      "objective/train/avg_token_score": 0.015260000713169575,
      "objective/train/avg_weight": 0.9820685982704163,
      "objective/train/docs_used": 1114494,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7072296142578125,
      "objective/train/std_weight": 0.08444172143936157,
      "objective/train/theoretical_loss": 3.434819295395487,
      "objective/train/tokens_used": 1989816800,
      "theoretical_loss": 3.434819295395487,
      "tokens_seen": 1969356800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002099867599773028,
      "loss": 2.4065,
      "theoretical_loss": 3.434819295395487,
      "tokens_seen": 1969356800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020997020994893134,
      "loss": 2.7644,
      "theoretical_loss": 3.434809736269743,
      "tokens_seen": 1969422336
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020995365992055985,
      "loss": 2.7804,
      "theoretical_loss": 3.4348001775511543,
      "tokens_seen": 1969487872
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020993710989218839,
      "loss": 2.697,
      "theoretical_loss": 3.4347906192396893,
      "tokens_seen": 1969553408
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002099205598638169,
      "loss": 2.6342,
      "theoretical_loss": 3.4347810613353182,
      "tokens_seen": 1969618944
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020990400983544543,
      "loss": 2.5804,
      "theoretical_loss": 3.434771503838009,
      "tokens_seen": 1969684480
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020988745980707394,
      "loss": 2.626,
      "theoretical_loss": 3.434761946747732,
      "tokens_seen": 1969750016
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020987090977870248,
      "loss": 2.6831,
      "theoretical_loss": 3.434752390064456,
      "tokens_seen": 1969815552
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000209854359750331,
      "loss": 2.4088,
      "theoretical_loss": 3.4347428337881496,
      "tokens_seen": 1969881088
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020983780972195953,
      "loss": 2.4401,
      "theoretical_loss": 3.434733277918782,
      "tokens_seen": 1969946624
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020982125969358804,
      "loss": 2.7038,
      "theoretical_loss": 3.4347237224563223,
      "tokens_seen": 1970012160
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020980470966521658,
      "loss": 2.3162,
      "theoretical_loss": 3.4347141674007404,
      "tokens_seen": 1970077696
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002097881596368451,
      "loss": 2.4435,
      "theoretical_loss": 3.4347046127520047,
      "tokens_seen": 1970143232
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020977160960847363,
      "loss": 2.8345,
      "theoretical_loss": 3.4346950585100853,
      "tokens_seen": 1970208768
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020975505958010214,
      "loss": 2.7212,
      "theoretical_loss": 3.43468550467495,
      "tokens_seen": 1970274304
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020973850955173067,
      "loss": 2.9106,
      "theoretical_loss": 3.434675951246568,
      "tokens_seen": 1970339840
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020972195952335918,
      "loss": 2.5128,
      "theoretical_loss": 3.43466639822491,
      "tokens_seen": 1970405376
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020970540949498772,
      "loss": 2.5226,
      "theoretical_loss": 3.434656845609944,
      "tokens_seen": 1970470912
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020968885946661623,
      "loss": 2.5382,
      "theoretical_loss": 3.434647293401639,
      "tokens_seen": 1970536448
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020967230943824477,
      "loss": 2.5948,
      "theoretical_loss": 3.434637741599965,
      "tokens_seen": 1970601984
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020965575940987325,
      "loss": 2.4127,
      "theoretical_loss": 3.4346281902048905,
      "tokens_seen": 1970667520
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020963920938150176,
      "loss": 2.6431,
      "theoretical_loss": 3.4346186392163847,
      "tokens_seen": 1970733056
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002096226593531303,
      "loss": 2.601,
      "theoretical_loss": 3.434609088634417,
      "tokens_seen": 1970798592
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002096061093247588,
      "loss": 2.5776,
      "theoretical_loss": 3.4345995384589565,
      "tokens_seen": 1970864128
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020958955929638735,
      "loss": 2.4995,
      "theoretical_loss": 3.4345899886899725,
      "tokens_seen": 1970929664
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020957300926801586,
      "loss": 2.6395,
      "theoretical_loss": 3.434580439327434,
      "tokens_seen": 1970995200
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002095564592396444,
      "loss": 2.7429,
      "theoretical_loss": 3.4345708903713104,
      "tokens_seen": 1971060736
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002095399092112729,
      "loss": 2.5317,
      "theoretical_loss": 3.4345613418215706,
      "tokens_seen": 1971126272
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020952335918290145,
      "loss": 2.4056,
      "theoretical_loss": 3.4345517936781835,
      "tokens_seen": 1971191808
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020950680915452996,
      "loss": 2.7027,
      "theoretical_loss": 3.4345422459411195,
      "tokens_seen": 1971257344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002094902591261585,
      "loss": 2.6428,
      "theoretical_loss": 3.434532698610347,
      "tokens_seen": 1971322880
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020947370909778703,
      "loss": 2.3672,
      "theoretical_loss": 3.434523151685835,
      "tokens_seen": 1971388416
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020945715906941554,
      "loss": 2.6131,
      "theoretical_loss": 3.4345136051675524,
      "tokens_seen": 1971453952
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020944060904104408,
      "loss": 2.924,
      "theoretical_loss": 3.43450405905547,
      "tokens_seen": 1971519488
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002094240590126726,
      "loss": 2.298,
      "theoretical_loss": 3.4344945133495552,
      "tokens_seen": 1971585024
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020940750898430113,
      "loss": 2.4779,
      "theoretical_loss": 3.4344849680497784,
      "tokens_seen": 1971650560
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020939095895592964,
      "loss": 2.5971,
      "theoretical_loss": 3.4344754231561083,
      "tokens_seen": 1971716096
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020937440892755817,
      "loss": 2.7012,
      "theoretical_loss": 3.4344658786685143,
      "tokens_seen": 1971781632
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020935785889918669,
      "loss": 2.7613,
      "theoretical_loss": 3.4344563345869656,
      "tokens_seen": 1971847168
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020934130887081522,
      "loss": 2.1638,
      "theoretical_loss": 3.4344467909114313,
      "tokens_seen": 1971912704
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020932475884244373,
      "loss": 2.6353,
      "theoretical_loss": 3.4344372476418807,
      "tokens_seen": 1971978240
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020930820881407227,
      "loss": 2.7667,
      "theoretical_loss": 3.434427704778283,
      "tokens_seen": 1972043776
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020929165878570078,
      "loss": 2.546,
      "theoretical_loss": 3.434418162320608,
      "tokens_seen": 1972109312
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020927510875732932,
      "loss": 2.4962,
      "theoretical_loss": 3.434408620268824,
      "tokens_seen": 1972174848
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020925855872895783,
      "loss": 2.6904,
      "theoretical_loss": 3.434399078622901,
      "tokens_seen": 1972240384
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020924200870058637,
      "loss": 2.5571,
      "theoretical_loss": 3.434389537382808,
      "tokens_seen": 1972305920
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020922545867221488,
      "loss": 2.4978,
      "theoretical_loss": 3.434379996548514,
      "tokens_seen": 1972371456
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020920890864384341,
      "loss": 2.6448,
      "theoretical_loss": 3.4343704561199884,
      "tokens_seen": 1972436992
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020919235861547193,
      "loss": 2.6146,
      "theoretical_loss": 3.4343609160972006,
      "tokens_seen": 1972502528
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020917580858710046,
      "loss": 2.3025,
      "theoretical_loss": 3.4343513764801203,
      "tokens_seen": 1972568064
    },
    {
      "epoch": 0.7,
      "objective/train/avg_token_score": 0.007364606950432062,
      "objective/train/avg_weight": 0.9913301467895508,
      "objective/train/docs_used": 1115816,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7736012935638428,
      "objective/train/std_weight": 0.04948240518569946,
      "objective/train/theoretical_loss": 3.4343418372687156,
      "objective/train/tokens_used": 1993093600,
      "theoretical_loss": 3.4343418372687156,
      "tokens_seen": 1972633600
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020915925855872895,
      "loss": 2.7572,
      "theoretical_loss": 3.4343418372687156,
      "tokens_seen": 1972633600
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020914270853035746,
      "loss": 2.5999,
      "theoretical_loss": 3.434332298462957,
      "tokens_seen": 1972699136
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000209126158501986,
      "loss": 2.4325,
      "theoretical_loss": 3.434322760062813,
      "tokens_seen": 1972764672
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002091096084736145,
      "loss": 2.646,
      "theoretical_loss": 3.434313222068253,
      "tokens_seen": 1972830208
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020909305844524304,
      "loss": 2.025,
      "theoretical_loss": 3.4343036844792465,
      "tokens_seen": 1972895744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020907650841687155,
      "loss": 2.8011,
      "theoretical_loss": 3.4342941472957627,
      "tokens_seen": 1972961280
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002090599583885001,
      "loss": 2.714,
      "theoretical_loss": 3.4342846105177705,
      "tokens_seen": 1973026816
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002090434083601286,
      "loss": 2.5128,
      "theoretical_loss": 3.4342750741452397,
      "tokens_seen": 1973092352
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020902685833175714,
      "loss": 2.6687,
      "theoretical_loss": 3.43426553817814,
      "tokens_seen": 1973157888
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020901030830338565,
      "loss": 2.6755,
      "theoretical_loss": 3.4342560026164395,
      "tokens_seen": 1973223424
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020899375827501419,
      "loss": 2.4679,
      "theoretical_loss": 3.4342464674601083,
      "tokens_seen": 1973288960
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002089772082466427,
      "loss": 2.5806,
      "theoretical_loss": 3.4342369327091156,
      "tokens_seen": 1973354496
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020896065821827123,
      "loss": 2.608,
      "theoretical_loss": 3.434227398363431,
      "tokens_seen": 1973420032
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020894410818989974,
      "loss": 2.4382,
      "theoretical_loss": 3.4342178644230232,
      "tokens_seen": 1973485568
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020892755816152828,
      "loss": 2.6201,
      "theoretical_loss": 3.4342083308878615,
      "tokens_seen": 1973551104
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002089110081331568,
      "loss": 2.5485,
      "theoretical_loss": 3.4341987977579156,
      "tokens_seen": 1973616640
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020889445810478533,
      "loss": 2.9593,
      "theoretical_loss": 3.434189265033155,
      "tokens_seen": 1973682176
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020887790807641384,
      "loss": 2.7438,
      "theoretical_loss": 3.4341797327135484,
      "tokens_seen": 1973747712
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020886135804804238,
      "loss": 2.6368,
      "theoretical_loss": 3.4341702007990658,
      "tokens_seen": 1973813248
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002088448080196709,
      "loss": 2.5089,
      "theoretical_loss": 3.434160669289676,
      "tokens_seen": 1973878784
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00020882825799129943,
      "loss": 2.5291,
      "theoretical_loss": 3.4341511381853485,
      "tokens_seen": 1973944320
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020881170796292794,
      "loss": 2.5458,
      "theoretical_loss": 3.434141607486053,
      "tokens_seen": 1974009856
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020879515793455647,
      "loss": 2.5866,
      "theoretical_loss": 3.4341320771917583,
      "tokens_seen": 1974075392
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020877860790618498,
      "loss": 2.575,
      "theoretical_loss": 3.434122547302434,
      "tokens_seen": 1974140928
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020876205787781352,
      "loss": 2.4401,
      "theoretical_loss": 3.4341130178180492,
      "tokens_seen": 1974206464
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020874550784944203,
      "loss": 2.6041,
      "theoretical_loss": 3.4341034887385735,
      "tokens_seen": 1974272000
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020872895782107057,
      "loss": 2.8005,
      "theoretical_loss": 3.4340939600639766,
      "tokens_seen": 1974337536
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020871240779269908,
      "loss": 2.4715,
      "theoretical_loss": 3.4340844317942274,
      "tokens_seen": 1974403072
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020869585776432762,
      "loss": 2.6344,
      "theoretical_loss": 3.4340749039292957,
      "tokens_seen": 1974468608
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020867930773595613,
      "loss": 2.508,
      "theoretical_loss": 3.4340653764691496,
      "tokens_seen": 1974534144
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002086627577075846,
      "loss": 2.5203,
      "theoretical_loss": 3.43405584941376,
      "tokens_seen": 1974599680
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020864620767921315,
      "loss": 2.6947,
      "theoretical_loss": 3.4340463227630957,
      "tokens_seen": 1974665216
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020862965765084166,
      "loss": 2.5497,
      "theoretical_loss": 3.434036796517126,
      "tokens_seen": 1974730752
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002086131076224702,
      "loss": 2.4954,
      "theoretical_loss": 3.43402727067582,
      "tokens_seen": 1974796288
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002085965575940987,
      "loss": 2.7568,
      "theoretical_loss": 3.4340177452391476,
      "tokens_seen": 1974861824
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020858000756572725,
      "loss": 2.6221,
      "theoretical_loss": 3.4340082202070774,
      "tokens_seen": 1974927360
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020856345753735576,
      "loss": 2.1846,
      "theoretical_loss": 3.43399869557958,
      "tokens_seen": 1974992896
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002085469075089843,
      "loss": 2.5706,
      "theoretical_loss": 3.4339891713566244,
      "tokens_seen": 1975058432
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002085303574806128,
      "loss": 2.6875,
      "theoretical_loss": 3.433979647538179,
      "tokens_seen": 1975123968
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020851380745224134,
      "loss": 2.5517,
      "theoretical_loss": 3.4339701241242144,
      "tokens_seen": 1975189504
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020849725742386985,
      "loss": 2.4244,
      "theoretical_loss": 3.4339606011146993,
      "tokens_seen": 1975255040
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002084807073954984,
      "loss": 2.6633,
      "theoretical_loss": 3.4339510785096037,
      "tokens_seen": 1975320576
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002084641573671269,
      "loss": 2.501,
      "theoretical_loss": 3.4339415563088966,
      "tokens_seen": 1975386112
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020844760733875544,
      "loss": 2.4935,
      "theoretical_loss": 3.433932034512547,
      "tokens_seen": 1975451648
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020843105731038395,
      "loss": 2.4642,
      "theoretical_loss": 3.4339225131205255,
      "tokens_seen": 1975517184
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020841450728201249,
      "loss": 2.6599,
      "theoretical_loss": 3.4339129921328,
      "tokens_seen": 1975582720
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000208397957253641,
      "loss": 2.614,
      "theoretical_loss": 3.4339034715493417,
      "tokens_seen": 1975648256
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020838140722526953,
      "loss": 2.6774,
      "theoretical_loss": 3.433893951370118,
      "tokens_seen": 1975713792
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020836485719689804,
      "loss": 2.7699,
      "theoretical_loss": 3.4338844315951,
      "tokens_seen": 1975779328
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020834830716852658,
      "loss": 2.4927,
      "theoretical_loss": 3.4338749122242564,
      "tokens_seen": 1975844864
    },
    {
      "epoch": 0.71,
      "objective/train/avg_token_score": 0.024974435567855835,
      "objective/train/avg_weight": 0.9706001877784729,
      "objective/train/docs_used": 1117825,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.763598918914795,
      "objective/train/std_weight": 0.12249001115560532,
      "objective/train/theoretical_loss": 3.433865393257557,
      "objective/train/tokens_used": 1996370400,
      "theoretical_loss": 3.433865393257557,
      "tokens_seen": 1975910400
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002083317571401551,
      "loss": 2.4807,
      "theoretical_loss": 3.433865393257557,
      "tokens_seen": 1975910400
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020831520711178363,
      "loss": 2.9054,
      "theoretical_loss": 3.433855874694971,
      "tokens_seen": 1975975936
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020829865708341214,
      "loss": 2.3765,
      "theoretical_loss": 3.433846356536468,
      "tokens_seen": 1976041472
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020828210705504068,
      "loss": 2.6757,
      "theoretical_loss": 3.4338368387820166,
      "tokens_seen": 1976107008
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002082655570266692,
      "loss": 2.455,
      "theoretical_loss": 3.433827321431587,
      "tokens_seen": 1976172544
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020824900699829773,
      "loss": 2.501,
      "theoretical_loss": 3.433817804485149,
      "tokens_seen": 1976238080
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020823245696992624,
      "loss": 2.5443,
      "theoretical_loss": 3.4338082879426715,
      "tokens_seen": 1976303616
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020821590694155477,
      "loss": 2.3094,
      "theoretical_loss": 3.433798771804124,
      "tokens_seen": 1976369152
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020819935691318328,
      "loss": 2.6806,
      "theoretical_loss": 3.433789256069476,
      "tokens_seen": 1976434688
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020818280688481182,
      "loss": 2.7601,
      "theoretical_loss": 3.433779740738697,
      "tokens_seen": 1976500224
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002081662568564403,
      "loss": 2.6639,
      "theoretical_loss": 3.433770225811757,
      "tokens_seen": 1976565760
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020814970682806884,
      "loss": 2.7491,
      "theoretical_loss": 3.433760711288624,
      "tokens_seen": 1976631296
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020813315679969735,
      "loss": 2.5401,
      "theoretical_loss": 3.4337511971692694,
      "tokens_seen": 1976696832
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002081166067713259,
      "loss": 2.4464,
      "theoretical_loss": 3.433741683453661,
      "tokens_seen": 1976762368
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002081000567429544,
      "loss": 2.7256,
      "theoretical_loss": 3.4337321701417696,
      "tokens_seen": 1976827904
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020808350671458294,
      "loss": 2.4161,
      "theoretical_loss": 3.433722657233564,
      "tokens_seen": 1976893440
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020806695668621145,
      "loss": 2.6115,
      "theoretical_loss": 3.4337131447290132,
      "tokens_seen": 1976958976
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020805040665783999,
      "loss": 2.8287,
      "theoretical_loss": 3.4337036326280876,
      "tokens_seen": 1977024512
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002080338566294685,
      "loss": 2.3593,
      "theoretical_loss": 3.433694120930756,
      "tokens_seen": 1977090048
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020801730660109703,
      "loss": 2.2494,
      "theoretical_loss": 3.4336846096369893,
      "tokens_seen": 1977155584
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020800075657272554,
      "loss": 2.5544,
      "theoretical_loss": 3.433675098746755,
      "tokens_seen": 1977221120
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020798420654435408,
      "loss": 2.537,
      "theoretical_loss": 3.4336655882600238,
      "tokens_seen": 1977286656
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002079676565159826,
      "loss": 1.9963,
      "theoretical_loss": 3.433656078176765,
      "tokens_seen": 1977352192
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020795110648761113,
      "loss": 2.5591,
      "theoretical_loss": 3.4336465684969477,
      "tokens_seen": 1977417728
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020793455645923964,
      "loss": 2.499,
      "theoretical_loss": 3.4336370592205423,
      "tokens_seen": 1977483264
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020791800643086818,
      "loss": 2.4111,
      "theoretical_loss": 3.4336275503475178,
      "tokens_seen": 1977548800
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002079014564024967,
      "loss": 2.2999,
      "theoretical_loss": 3.4336180418778435,
      "tokens_seen": 1977614336
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020788490637412523,
      "loss": 2.5626,
      "theoretical_loss": 3.4336085338114897,
      "tokens_seen": 1977679872
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020786835634575374,
      "loss": 2.4636,
      "theoretical_loss": 3.433599026148425,
      "tokens_seen": 1977745408
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020785180631738227,
      "loss": 2.562,
      "theoretical_loss": 3.4335895188886187,
      "tokens_seen": 1977810944
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020783525628901078,
      "loss": 2.7738,
      "theoretical_loss": 3.433580012032042,
      "tokens_seen": 1977876480
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020781870626063932,
      "loss": 2.4647,
      "theoretical_loss": 3.433570505578663,
      "tokens_seen": 1977942016
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020780215623226783,
      "loss": 2.3837,
      "theoretical_loss": 3.433560999528451,
      "tokens_seen": 1978007552
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020778560620389637,
      "loss": 2.3343,
      "theoretical_loss": 3.433551493881377,
      "tokens_seen": 1978073088
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020776905617552488,
      "loss": 2.4261,
      "theoretical_loss": 3.4335419886374092,
      "tokens_seen": 1978138624
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020775250614715342,
      "loss": 2.3936,
      "theoretical_loss": 3.433532483796518,
      "tokens_seen": 1978204160
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020773595611878193,
      "loss": 2.5975,
      "theoretical_loss": 3.433522979358673,
      "tokens_seen": 1978269696
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020771940609041047,
      "loss": 2.6743,
      "theoretical_loss": 3.433513475323843,
      "tokens_seen": 1978335232
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020770285606203898,
      "loss": 2.5891,
      "theoretical_loss": 3.433503971691998,
      "tokens_seen": 1978400768
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020768630603366746,
      "loss": 2.6873,
      "theoretical_loss": 3.433494468463107,
      "tokens_seen": 1978466304
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000207669756005296,
      "loss": 2.4814,
      "theoretical_loss": 3.4334849656371405,
      "tokens_seen": 1978531840
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002076532059769245,
      "loss": 2.6765,
      "theoretical_loss": 3.4334754632140676,
      "tokens_seen": 1978597376
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020763665594855304,
      "loss": 2.2549,
      "theoretical_loss": 3.433465961193858,
      "tokens_seen": 1978662912
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020762010592018156,
      "loss": 2.7518,
      "theoretical_loss": 3.4334564595764814,
      "tokens_seen": 1978728448
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002076035558918101,
      "loss": 2.5151,
      "theoretical_loss": 3.4334469583619067,
      "tokens_seen": 1978793984
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002075870058634386,
      "loss": 2.2976,
      "theoretical_loss": 3.433437457550104,
      "tokens_seen": 1978859520
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020757045583506714,
      "loss": 2.6962,
      "theoretical_loss": 3.4334279571410438,
      "tokens_seen": 1978925056
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020755390580669565,
      "loss": 2.4169,
      "theoretical_loss": 3.4334184571346933,
      "tokens_seen": 1978990592
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002075373557783242,
      "loss": 2.5871,
      "theoretical_loss": 3.4334089575310247,
      "tokens_seen": 1979056128
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002075208057499527,
      "loss": 2.6191,
      "theoretical_loss": 3.433399458330006,
      "tokens_seen": 1979121664
    },
    {
      "epoch": 0.71,
      "objective/train/avg_token_score": 0.019214240834116936,
      "objective/train/avg_weight": 0.9774341583251953,
      "objective/train/docs_used": 1119491,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.620184898376465,
      "objective/train/std_weight": 0.10906728357076645,
      "objective/train/theoretical_loss": 3.433389959531607,
      "objective/train/tokens_used": 1999647200,
      "theoretical_loss": 3.433389959531607,
      "tokens_seen": 1979187200
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020750425572158124,
      "loss": 2.4375,
      "theoretical_loss": 3.433389959531607,
      "tokens_seen": 1979187200
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020748770569320975,
      "loss": 2.5224,
      "theoretical_loss": 3.433380461135798,
      "tokens_seen": 1979252736
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020747115566483828,
      "loss": 2.5784,
      "theoretical_loss": 3.433370963142548,
      "tokens_seen": 1979318272
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002074546056364668,
      "loss": 2.7851,
      "theoretical_loss": 3.433361465551827,
      "tokens_seen": 1979383808
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020743805560809533,
      "loss": 2.5654,
      "theoretical_loss": 3.4333519683636036,
      "tokens_seen": 1979449344
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020742150557972384,
      "loss": 2.5258,
      "theoretical_loss": 3.433342471577849,
      "tokens_seen": 1979514880
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020740495555135238,
      "loss": 2.5597,
      "theoretical_loss": 3.4333329751945314,
      "tokens_seen": 1979580416
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002073884055229809,
      "loss": 2.5137,
      "theoretical_loss": 3.433323479213622,
      "tokens_seen": 1979645952
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020737185549460943,
      "loss": 2.4811,
      "theoretical_loss": 3.4333139836350886,
      "tokens_seen": 1979711488
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020735530546623794,
      "loss": 2.5636,
      "theoretical_loss": 3.433304488458902,
      "tokens_seen": 1979777024
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020733875543786648,
      "loss": 2.2845,
      "theoretical_loss": 3.433294993685031,
      "tokens_seen": 1979842560
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000207322205409495,
      "loss": 2.6381,
      "theoretical_loss": 3.4332854993134463,
      "tokens_seen": 1979908096
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020730565538112352,
      "loss": 2.6836,
      "theoretical_loss": 3.4332760053441174,
      "tokens_seen": 1979973632
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020728910535275204,
      "loss": 2.6526,
      "theoretical_loss": 3.4332665117770125,
      "tokens_seen": 1980039168
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020727255532438057,
      "loss": 2.5446,
      "theoretical_loss": 3.433257018612103,
      "tokens_seen": 1980104704
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020725600529600908,
      "loss": 2.4866,
      "theoretical_loss": 3.433247525849358,
      "tokens_seen": 1980170240
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020723945526763762,
      "loss": 2.5514,
      "theoretical_loss": 3.433238033488747,
      "tokens_seen": 1980235776
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020722290523926613,
      "loss": 2.4273,
      "theoretical_loss": 3.433228541530239,
      "tokens_seen": 1980301312
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020720635521089467,
      "loss": 2.629,
      "theoretical_loss": 3.433219049973805,
      "tokens_seen": 1980366848
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020718980518252315,
      "loss": 2.6582,
      "theoretical_loss": 3.4332095588194136,
      "tokens_seen": 1980432384
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020717325515415166,
      "loss": 2.7741,
      "theoretical_loss": 3.433200068067035,
      "tokens_seen": 1980497920
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002071567051257802,
      "loss": 2.3219,
      "theoretical_loss": 3.4331905777166387,
      "tokens_seen": 1980563456
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002071401550974087,
      "loss": 2.5256,
      "theoretical_loss": 3.4331810877681948,
      "tokens_seen": 1980628992
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020712360506903725,
      "loss": 2.4491,
      "theoretical_loss": 3.433171598221672,
      "tokens_seen": 1980694528
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020710705504066576,
      "loss": 2.3117,
      "theoretical_loss": 3.43316210907704,
      "tokens_seen": 1980760064
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002070905050122943,
      "loss": 2.3998,
      "theoretical_loss": 3.43315262033427,
      "tokens_seen": 1980825600
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002070739549839228,
      "loss": 2.3061,
      "theoretical_loss": 3.4331431319933303,
      "tokens_seen": 1980891136
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020705740495555134,
      "loss": 2.5463,
      "theoretical_loss": 3.4331336440541915,
      "tokens_seen": 1980956672
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020704085492717985,
      "loss": 2.5886,
      "theoretical_loss": 3.433124156516823,
      "tokens_seen": 1981022208
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002070243048988084,
      "loss": 2.6359,
      "theoretical_loss": 3.4331146693811934,
      "tokens_seen": 1981087744
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002070077548704369,
      "loss": 2.371,
      "theoretical_loss": 3.4331051826472736,
      "tokens_seen": 1981153280
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020699120484206544,
      "loss": 2.5801,
      "theoretical_loss": 3.433095696315033,
      "tokens_seen": 1981218816
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020697465481369395,
      "loss": 2.7351,
      "theoretical_loss": 3.4330862103844417,
      "tokens_seen": 1981284352
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002069581047853225,
      "loss": 2.502,
      "theoretical_loss": 3.4330767248554688,
      "tokens_seen": 1981349888
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000206941554756951,
      "loss": 2.57,
      "theoretical_loss": 3.433067239728084,
      "tokens_seen": 1981415424
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020692500472857954,
      "loss": 2.5278,
      "theoretical_loss": 3.4330577550022574,
      "tokens_seen": 1981480960
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020690845470020805,
      "loss": 2.5312,
      "theoretical_loss": 3.4330482706779586,
      "tokens_seen": 1981546496
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020689190467183658,
      "loss": 2.4999,
      "theoretical_loss": 3.4330387867551573,
      "tokens_seen": 1981612032
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002068753546434651,
      "loss": 2.5456,
      "theoretical_loss": 3.4330293032338233,
      "tokens_seen": 1981677568
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020685880461509363,
      "loss": 2.6548,
      "theoretical_loss": 3.433019820113926,
      "tokens_seen": 1981743104
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020684225458672214,
      "loss": 2.623,
      "theoretical_loss": 3.4330103373954355,
      "tokens_seen": 1981808640
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020682570455835068,
      "loss": 2.5655,
      "theoretical_loss": 3.4330008550783213,
      "tokens_seen": 1981874176
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002068091545299792,
      "loss": 2.7337,
      "theoretical_loss": 3.4329913731625536,
      "tokens_seen": 1981939712
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020679260450160773,
      "loss": 2.3732,
      "theoretical_loss": 3.4329818916481014,
      "tokens_seen": 1982005248
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020677605447323624,
      "loss": 2.4941,
      "theoretical_loss": 3.4329724105349353,
      "tokens_seen": 1982070784
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020675950444486478,
      "loss": 2.5873,
      "theoretical_loss": 3.432962929823024,
      "tokens_seen": 1982136320
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002067429544164933,
      "loss": 2.5876,
      "theoretical_loss": 3.432953449512338,
      "tokens_seen": 1982201856
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020672640438812182,
      "loss": 2.6593,
      "theoretical_loss": 3.432943969602847,
      "tokens_seen": 1982267392
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020670985435975036,
      "loss": 2.6425,
      "theoretical_loss": 3.432934490094521,
      "tokens_seen": 1982332928
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020669330433137884,
      "loss": 2.4012,
      "theoretical_loss": 3.4329250109873293,
      "tokens_seen": 1982398464
    },
    {
      "epoch": 0.71,
      "objective/train/avg_token_score": 0.027833102270960808,
      "objective/train/avg_weight": 0.9672377705574036,
      "objective/train/docs_used": 1121511,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.101161479949951,
      "objective/train/std_weight": 0.10513198375701904,
      "objective/train/theoretical_loss": 3.4329155322812417,
      "objective/train/tokens_used": 2002924000,
      "theoretical_loss": 3.4329155322812417,
      "tokens_seen": 1982464000
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020667675430300736,
      "loss": 2.7758,
      "theoretical_loss": 3.4329155322812417,
      "tokens_seen": 1982464000
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002066602042746359,
      "loss": 1.8233,
      "theoretical_loss": 3.432906053976228,
      "tokens_seen": 1982529536
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002066436542462644,
      "loss": 2.6103,
      "theoretical_loss": 3.432896576072258,
      "tokens_seen": 1982595072
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020662710421789294,
      "loss": 2.5502,
      "theoretical_loss": 3.432887098569301,
      "tokens_seen": 1982660608
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020661055418952145,
      "loss": 2.4544,
      "theoretical_loss": 3.4328776214673278,
      "tokens_seen": 1982726144
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020659400416115,
      "loss": 2.5724,
      "theoretical_loss": 3.432868144766308,
      "tokens_seen": 1982791680
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002065774541327785,
      "loss": 2.229,
      "theoretical_loss": 3.432858668466211,
      "tokens_seen": 1982857216
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020656090410440704,
      "loss": 2.3003,
      "theoretical_loss": 3.432849192567006,
      "tokens_seen": 1982922752
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020654435407603555,
      "loss": 2.675,
      "theoretical_loss": 3.4328397170686635,
      "tokens_seen": 1982988288
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020652780404766408,
      "loss": 2.2083,
      "theoretical_loss": 3.4328302419711534,
      "tokens_seen": 1983053824
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002065112540192926,
      "loss": 2.8453,
      "theoretical_loss": 3.4328207672744453,
      "tokens_seen": 1983119360
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020649470399092113,
      "loss": 2.7099,
      "theoretical_loss": 3.432811292978509,
      "tokens_seen": 1983184896
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020647815396254964,
      "loss": 2.4897,
      "theoretical_loss": 3.432801819083315,
      "tokens_seen": 1983250432
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020646160393417818,
      "loss": 2.67,
      "theoretical_loss": 3.4327923455888314,
      "tokens_seen": 1983315968
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002064450539058067,
      "loss": 2.5848,
      "theoretical_loss": 3.4327828724950296,
      "tokens_seen": 1983381504
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020642850387743523,
      "loss": 2.7884,
      "theoretical_loss": 3.432773399801879,
      "tokens_seen": 1983447040
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020641195384906374,
      "loss": 2.4761,
      "theoretical_loss": 3.432763927509349,
      "tokens_seen": 1983512576
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020639540382069228,
      "loss": 2.5885,
      "theoretical_loss": 3.4327544556174097,
      "tokens_seen": 1983578112
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002063788537923208,
      "loss": 2.6244,
      "theoretical_loss": 3.432744984126031,
      "tokens_seen": 1983643648
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020636230376394932,
      "loss": 2.4814,
      "theoretical_loss": 3.4327355130351824,
      "tokens_seen": 1983709184
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020634575373557783,
      "loss": 2.6689,
      "theoretical_loss": 3.4327260423448345,
      "tokens_seen": 1983774720
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020632920370720637,
      "loss": 2.5173,
      "theoretical_loss": 3.4327165720549564,
      "tokens_seen": 1983840256
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020631265367883488,
      "loss": 2.3619,
      "theoretical_loss": 3.4327071021655176,
      "tokens_seen": 1983905792
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020629610365046342,
      "loss": 2.5117,
      "theoretical_loss": 3.432697632676489,
      "tokens_seen": 1983971328
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020627955362209193,
      "loss": 2.6718,
      "theoretical_loss": 3.43268816358784,
      "tokens_seen": 1984036864
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020626300359372047,
      "loss": 2.4844,
      "theoretical_loss": 3.4326786948995407,
      "tokens_seen": 1984102400
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020624645356534898,
      "loss": 2.7342,
      "theoretical_loss": 3.43266922661156,
      "tokens_seen": 1984167936
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020622990353697752,
      "loss": 2.5672,
      "theoretical_loss": 3.4326597587238687,
      "tokens_seen": 1984233472
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000206213353508606,
      "loss": 2.3418,
      "theoretical_loss": 3.432650291236436,
      "tokens_seen": 1984299008
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002061968034802345,
      "loss": 2.1962,
      "theoretical_loss": 3.4326408241492326,
      "tokens_seen": 1984364544
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020618025345186305,
      "loss": 2.3383,
      "theoretical_loss": 3.4326313574622276,
      "tokens_seen": 1984430080
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020616370342349156,
      "loss": 2.6504,
      "theoretical_loss": 3.432621891175391,
      "tokens_seen": 1984495616
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002061471533951201,
      "loss": 2.3034,
      "theoretical_loss": 3.432612425288693,
      "tokens_seen": 1984561152
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002061306033667486,
      "loss": 2.5904,
      "theoretical_loss": 3.4326029598021033,
      "tokens_seen": 1984626688
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020611405333837714,
      "loss": 2.489,
      "theoretical_loss": 3.432593494715592,
      "tokens_seen": 1984692224
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020609750331000565,
      "loss": 2.5283,
      "theoretical_loss": 3.432584030029128,
      "tokens_seen": 1984757760
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002060809532816342,
      "loss": 2.5922,
      "theoretical_loss": 3.4325745657426827,
      "tokens_seen": 1984823296
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002060644032532627,
      "loss": 2.5546,
      "theoretical_loss": 3.432565101856224,
      "tokens_seen": 1984888832
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020604785322489124,
      "loss": 2.7688,
      "theoretical_loss": 3.4325556383697244,
      "tokens_seen": 1984954368
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020603130319651975,
      "loss": 2.7129,
      "theoretical_loss": 3.432546175283152,
      "tokens_seen": 1985019904
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002060147531681483,
      "loss": 2.8313,
      "theoretical_loss": 3.4325367125964767,
      "tokens_seen": 1985085440
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002059982031397768,
      "loss": 2.7693,
      "theoretical_loss": 3.4325272503096693,
      "tokens_seen": 1985150976
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020598165311140534,
      "loss": 2.7035,
      "theoretical_loss": 3.4325177884226985,
      "tokens_seen": 1985216512
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020596510308303385,
      "loss": 2.5663,
      "theoretical_loss": 3.4325083269355354,
      "tokens_seen": 1985282048
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020594855305466238,
      "loss": 2.3211,
      "theoretical_loss": 3.4324988658481494,
      "tokens_seen": 1985347584
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002059320030262909,
      "loss": 2.5424,
      "theoretical_loss": 3.43248940516051,
      "tokens_seen": 1985413120
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020591545299791943,
      "loss": 2.6684,
      "theoretical_loss": 3.432479944872588,
      "tokens_seen": 1985478656
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020589890296954794,
      "loss": 2.5589,
      "theoretical_loss": 3.4324704849843526,
      "tokens_seen": 1985544192
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020588235294117648,
      "loss": 2.3965,
      "theoretical_loss": 3.4324610254957744,
      "tokens_seen": 1985609728
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000205865802912805,
      "loss": 2.7926,
      "theoretical_loss": 3.4324515664068223,
      "tokens_seen": 1985675264
    },
    {
      "epoch": 0.71,
      "objective/train/avg_token_score": 0.009100638329982758,
      "objective/train/avg_weight": 0.9892874956130981,
      "objective/train/docs_used": 1123355,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8225295543670654,
      "objective/train/std_weight": 0.041536394506692886,
      "objective/train/theoretical_loss": 3.432442107717467,
      "objective/train/tokens_used": 2006200800,
      "theoretical_loss": 3.432442107717467,
      "tokens_seen": 1985740800
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020584925288443353,
      "loss": 2.5782,
      "theoretical_loss": 3.432442107717467,
      "tokens_seen": 1985740800
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020583270285606204,
      "loss": 2.367,
      "theoretical_loss": 3.4324326494276782,
      "tokens_seen": 1985806336
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020581615282769058,
      "loss": 2.5441,
      "theoretical_loss": 3.432423191537426,
      "tokens_seen": 1985871872
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020579960279931909,
      "loss": 2.2246,
      "theoretical_loss": 3.4324137340466807,
      "tokens_seen": 1985937408
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020578305277094762,
      "loss": 2.6709,
      "theoretical_loss": 3.432404276955411,
      "tokens_seen": 1986002944
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020576650274257613,
      "loss": 2.6807,
      "theoretical_loss": 3.432394820263588,
      "tokens_seen": 1986068480
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020574995271420467,
      "loss": 2.7524,
      "theoretical_loss": 3.432385363971181,
      "tokens_seen": 1986134016
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020573340268583318,
      "loss": 2.7102,
      "theoretical_loss": 3.4323759080781606,
      "tokens_seen": 1986199552
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020571685265746167,
      "loss": 2.6029,
      "theoretical_loss": 3.432366452584496,
      "tokens_seen": 1986265088
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002057003026290902,
      "loss": 2.4811,
      "theoretical_loss": 3.4323569974901575,
      "tokens_seen": 1986330624
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002056837526007187,
      "loss": 2.678,
      "theoretical_loss": 3.4323475427951156,
      "tokens_seen": 1986396160
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020566720257234725,
      "loss": 2.4638,
      "theoretical_loss": 3.4323380884993395,
      "tokens_seen": 1986461696
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020565065254397576,
      "loss": 2.5784,
      "theoretical_loss": 3.4323286346027997,
      "tokens_seen": 1986527232
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002056341025156043,
      "loss": 2.3844,
      "theoretical_loss": 3.4323191811054654,
      "tokens_seen": 1986592768
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002056175524872328,
      "loss": 2.5493,
      "theoretical_loss": 3.432309728007307,
      "tokens_seen": 1986658304
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020560100245886135,
      "loss": 2.7911,
      "theoretical_loss": 3.432300275308295,
      "tokens_seen": 1986723840
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020558445243048986,
      "loss": 2.5908,
      "theoretical_loss": 3.4322908230083984,
      "tokens_seen": 1986789376
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002055679024021184,
      "loss": 2.6281,
      "theoretical_loss": 3.432281371107588,
      "tokens_seen": 1986854912
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002055513523737469,
      "loss": 2.9702,
      "theoretical_loss": 3.4322719196058338,
      "tokens_seen": 1986920448
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020553480234537544,
      "loss": 2.7559,
      "theoretical_loss": 3.4322624685031053,
      "tokens_seen": 1986985984
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020551825231700395,
      "loss": 2.472,
      "theoretical_loss": 3.4322530177993724,
      "tokens_seen": 1987051520
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002055017022886325,
      "loss": 2.427,
      "theoretical_loss": 3.4322435674946057,
      "tokens_seen": 1987117056
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020548515226026103,
      "loss": 2.517,
      "theoretical_loss": 3.4322341175887745,
      "tokens_seen": 1987182592
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020546860223188954,
      "loss": 2.2464,
      "theoretical_loss": 3.4322246680818496,
      "tokens_seen": 1987248128
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020545205220351808,
      "loss": 2.3804,
      "theoretical_loss": 3.4322152189738,
      "tokens_seen": 1987313664
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002054355021751466,
      "loss": 2.4575,
      "theoretical_loss": 3.4322057702645967,
      "tokens_seen": 1987379200
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020541895214677512,
      "loss": 2.5194,
      "theoretical_loss": 3.4321963219542098,
      "tokens_seen": 1987444736
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020540240211840363,
      "loss": 2.5605,
      "theoretical_loss": 3.4321868740426082,
      "tokens_seen": 1987510272
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020538585209003217,
      "loss": 2.639,
      "theoretical_loss": 3.432177426529763,
      "tokens_seen": 1987575808
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020536930206166068,
      "loss": 2.6204,
      "theoretical_loss": 3.4321679794156434,
      "tokens_seen": 1987641344
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020535275203328922,
      "loss": 2.8425,
      "theoretical_loss": 3.4321585327002198,
      "tokens_seen": 1987706880
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020533620200491773,
      "loss": 2.6182,
      "theoretical_loss": 3.432149086383462,
      "tokens_seen": 1987772416
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020531965197654627,
      "loss": 2.4542,
      "theoretical_loss": 3.4321396404653406,
      "tokens_seen": 1987837952
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020530310194817478,
      "loss": 2.3682,
      "theoretical_loss": 3.4321301949458256,
      "tokens_seen": 1987903488
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020528655191980332,
      "loss": 2.3844,
      "theoretical_loss": 3.432120749824886,
      "tokens_seen": 1987969024
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020527000189143183,
      "loss": 2.5796,
      "theoretical_loss": 3.4321113051024934,
      "tokens_seen": 1988034560
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020525345186306036,
      "loss": 2.5193,
      "theoretical_loss": 3.4321018607786167,
      "tokens_seen": 1988100096
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020523690183468887,
      "loss": 2.4685,
      "theoretical_loss": 3.4320924168532256,
      "tokens_seen": 1988165632
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020522035180631736,
      "loss": 2.7202,
      "theoretical_loss": 3.432082973326292,
      "tokens_seen": 1988231168
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002052038017779459,
      "loss": 2.8611,
      "theoretical_loss": 3.432073530197784,
      "tokens_seen": 1988296704
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002051872517495744,
      "loss": 2.9711,
      "theoretical_loss": 3.4320640874676727,
      "tokens_seen": 1988362240
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020517070172120294,
      "loss": 2.3949,
      "theoretical_loss": 3.4320546451359277,
      "tokens_seen": 1988427776
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020515415169283145,
      "loss": 2.591,
      "theoretical_loss": 3.4320452032025193,
      "tokens_seen": 1988493312
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020513760166446,
      "loss": 2.5532,
      "theoretical_loss": 3.4320357616674175,
      "tokens_seen": 1988558848
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002051210516360885,
      "loss": 2.6298,
      "theoretical_loss": 3.4320263205305928,
      "tokens_seen": 1988624384
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020510450160771704,
      "loss": 2.5158,
      "theoretical_loss": 3.4320168797920143,
      "tokens_seen": 1988689920
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020508795157934555,
      "loss": 2.5435,
      "theoretical_loss": 3.432007439451653,
      "tokens_seen": 1988755456
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002050714015509741,
      "loss": 2.8374,
      "theoretical_loss": 3.4319979995094787,
      "tokens_seen": 1988820992
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002050548515226026,
      "loss": 2.6824,
      "theoretical_loss": 3.431988559965461,
      "tokens_seen": 1988886528
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020503830149423114,
      "loss": 2.5637,
      "theoretical_loss": 3.431979120819571,
      "tokens_seen": 1988952064
    },
    {
      "epoch": 0.71,
      "objective/train/avg_token_score": 0.014597472734749317,
      "objective/train/avg_weight": 0.982821524143219,
      "objective/train/docs_used": 1125335,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9538660049438477,
      "objective/train/std_weight": 0.08197970688343048,
      "objective/train/theoretical_loss": 3.4319696820717778,
      "objective/train/tokens_used": 2009477600,
      "theoretical_loss": 3.4319696820717778,
      "tokens_seen": 1989017600
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020502175146585965,
      "loss": 2.6586,
      "theoretical_loss": 3.4319696820717778,
      "tokens_seen": 1989017600
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020500520143748818,
      "loss": 2.8257,
      "theoretical_loss": 3.431960243722052,
      "tokens_seen": 1989083136
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002049886514091167,
      "loss": 2.4425,
      "theoretical_loss": 3.4319508057703634,
      "tokens_seen": 1989148672
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020497210138074523,
      "loss": 2.7663,
      "theoretical_loss": 3.4319413682166826,
      "tokens_seen": 1989214208
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020495555135237374,
      "loss": 2.58,
      "theoretical_loss": 3.431931931060979,
      "tokens_seen": 1989279744
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020493900132400228,
      "loss": 2.8414,
      "theoretical_loss": 3.4319224943032234,
      "tokens_seen": 1989345280
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002049224512956308,
      "loss": 2.4078,
      "theoretical_loss": 3.431913057943386,
      "tokens_seen": 1989410816
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020490590126725933,
      "loss": 2.6304,
      "theoretical_loss": 3.4319036219814363,
      "tokens_seen": 1989476352
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020488935123888784,
      "loss": 2.7214,
      "theoretical_loss": 3.4318941864173445,
      "tokens_seen": 1989541888
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020487280121051638,
      "loss": 2.7428,
      "theoretical_loss": 3.4318847512510806,
      "tokens_seen": 1989607424
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020485625118214489,
      "loss": 2.7603,
      "theoretical_loss": 3.431875316482615,
      "tokens_seen": 1989672960
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020483970115377342,
      "loss": 2.5755,
      "theoretical_loss": 3.4318658821119183,
      "tokens_seen": 1989738496
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020482315112540193,
      "loss": 2.6178,
      "theoretical_loss": 3.43185644813896,
      "tokens_seen": 1989804032
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020480660109703047,
      "loss": 3.1954,
      "theoretical_loss": 3.43184701456371,
      "tokens_seen": 1989869568
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020479005106865898,
      "loss": 2.5504,
      "theoretical_loss": 3.431837581386139,
      "tokens_seen": 1989935104
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020477350104028752,
      "loss": 2.6683,
      "theoretical_loss": 3.4318281486062174,
      "tokens_seen": 1990000640
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020475695101191603,
      "loss": 2.6489,
      "theoretical_loss": 3.4318187162239147,
      "tokens_seen": 1990066176
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020474040098354457,
      "loss": 2.7254,
      "theoretical_loss": 3.4318092842392005,
      "tokens_seen": 1990131712
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020472385095517305,
      "loss": 2.7592,
      "theoretical_loss": 3.4317998526520466,
      "tokens_seen": 1990197248
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020470730092680156,
      "loss": 2.4257,
      "theoretical_loss": 3.4317904214624217,
      "tokens_seen": 1990262784
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002046907508984301,
      "loss": 2.0353,
      "theoretical_loss": 3.4317809906702967,
      "tokens_seen": 1990328320
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002046742008700586,
      "loss": 2.7871,
      "theoretical_loss": 3.4317715602756413,
      "tokens_seen": 1990393856
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020465765084168715,
      "loss": 2.7717,
      "theoretical_loss": 3.4317621302784262,
      "tokens_seen": 1990459392
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020464110081331566,
      "loss": 2.4067,
      "theoretical_loss": 3.4317527006786217,
      "tokens_seen": 1990524928
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002046245507849442,
      "loss": 2.7083,
      "theoretical_loss": 3.431743271476197,
      "tokens_seen": 1990590464
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002046080007565727,
      "loss": 2.6703,
      "theoretical_loss": 3.4317338426711226,
      "tokens_seen": 1990656000
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020459145072820124,
      "loss": 2.6555,
      "theoretical_loss": 3.4317244142633694,
      "tokens_seen": 1990721536
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020457490069982975,
      "loss": 2.6136,
      "theoretical_loss": 3.4317149862529073,
      "tokens_seen": 1990787072
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002045583506714583,
      "loss": 2.6452,
      "theoretical_loss": 3.4317055586397056,
      "tokens_seen": 1990852608
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002045418006430868,
      "loss": 2.5177,
      "theoretical_loss": 3.4316961314237355,
      "tokens_seen": 1990918144
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020452525061471534,
      "loss": 2.6581,
      "theoretical_loss": 3.431686704604967,
      "tokens_seen": 1990983680
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020450870058634385,
      "loss": 2.7079,
      "theoretical_loss": 3.4316772781833693,
      "tokens_seen": 1991049216
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002044921505579724,
      "loss": 2.7472,
      "theoretical_loss": 3.431667852158914,
      "tokens_seen": 1991114752
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002044756005296009,
      "loss": 2.9096,
      "theoretical_loss": 3.4316584265315706,
      "tokens_seen": 1991180288
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020445905050122943,
      "loss": 2.4343,
      "theoretical_loss": 3.43164900130131,
      "tokens_seen": 1991245824
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020444250047285794,
      "loss": 2.4682,
      "theoretical_loss": 3.4316395764681014,
      "tokens_seen": 1991311360
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020442595044448648,
      "loss": 2.6023,
      "theoretical_loss": 3.4316301520319152,
      "tokens_seen": 1991376896
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000204409400416115,
      "loss": 2.7044,
      "theoretical_loss": 3.4316207279927218,
      "tokens_seen": 1991442432
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020439285038774353,
      "loss": 2.5696,
      "theoretical_loss": 3.4316113043504917,
      "tokens_seen": 1991507968
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020437630035937204,
      "loss": 2.5594,
      "theoretical_loss": 3.4316018811051947,
      "tokens_seen": 1991573504
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020435975033100058,
      "loss": 2.434,
      "theoretical_loss": 3.4315924582568016,
      "tokens_seen": 1991639040
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002043432003026291,
      "loss": 2.733,
      "theoretical_loss": 3.4315830358052817,
      "tokens_seen": 1991704576
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020432665027425763,
      "loss": 2.6819,
      "theoretical_loss": 3.431573613750606,
      "tokens_seen": 1991770112
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020431010024588614,
      "loss": 2.7825,
      "theoretical_loss": 3.431564192092744,
      "tokens_seen": 1991835648
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020429355021751467,
      "loss": 2.5201,
      "theoretical_loss": 3.431554770831667,
      "tokens_seen": 1991901184
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020427700018914318,
      "loss": 2.7404,
      "theoretical_loss": 3.4315453499673443,
      "tokens_seen": 1991966720
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020426045016077172,
      "loss": 2.5899,
      "theoretical_loss": 3.4315359294997467,
      "tokens_seen": 1992032256
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002042439001324002,
      "loss": 2.6476,
      "theoretical_loss": 3.431526509428844,
      "tokens_seen": 1992097792
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020422735010402874,
      "loss": 2.5477,
      "theoretical_loss": 3.431517089754607,
      "tokens_seen": 1992163328
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020421080007565725,
      "loss": 2.6952,
      "theoretical_loss": 3.4315076704770053,
      "tokens_seen": 1992228864
    },
    {
      "epoch": 0.71,
      "objective/train/avg_token_score": 0.01006163191050291,
      "objective/train/avg_weight": 0.9881545901298523,
      "objective/train/docs_used": 1126843,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.616776943206787,
      "objective/train/std_weight": 0.06625470519065857,
      "objective/train/theoretical_loss": 3.4314982515960093,
      "objective/train/tokens_used": 2012754400,
      "theoretical_loss": 3.4314982515960093,
      "tokens_seen": 1992294400
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002041942500472858,
      "loss": 2.6716,
      "theoretical_loss": 3.4314982515960093,
      "tokens_seen": 1992294400
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002041777000189143,
      "loss": 2.3798,
      "theoretical_loss": 3.4314888331115894,
      "tokens_seen": 1992359936
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020416114999054284,
      "loss": 2.7192,
      "theoretical_loss": 3.4314794150237162,
      "tokens_seen": 1992425472
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020414459996217135,
      "loss": 2.4913,
      "theoretical_loss": 3.4314699973323597,
      "tokens_seen": 1992491008
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002041280499337999,
      "loss": 2.562,
      "theoretical_loss": 3.4314605800374895,
      "tokens_seen": 1992556544
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002041114999054284,
      "loss": 2.6803,
      "theoretical_loss": 3.4314511631390774,
      "tokens_seen": 1992622080
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020409494987705694,
      "loss": 2.4495,
      "theoretical_loss": 3.431441746637092,
      "tokens_seen": 1992687616
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020407839984868545,
      "loss": 2.5881,
      "theoretical_loss": 3.4314323305315044,
      "tokens_seen": 1992753152
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020406184982031398,
      "loss": 2.2738,
      "theoretical_loss": 3.431422914822285,
      "tokens_seen": 1992818688
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002040452997919425,
      "loss": 2.5764,
      "theoretical_loss": 3.431413499509404,
      "tokens_seen": 1992884224
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020402874976357103,
      "loss": 2.7986,
      "theoretical_loss": 3.431404084592831,
      "tokens_seen": 1992949760
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020401219973519954,
      "loss": 2.5801,
      "theoretical_loss": 3.4313946700725375,
      "tokens_seen": 1993015296
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020399564970682808,
      "loss": 2.8096,
      "theoretical_loss": 3.4313852559484928,
      "tokens_seen": 1993080832
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002039790996784566,
      "loss": 2.6509,
      "theoretical_loss": 3.4313758422206675,
      "tokens_seen": 1993146368
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020396254965008513,
      "loss": 2.6435,
      "theoretical_loss": 3.4313664288890324,
      "tokens_seen": 1993211904
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020394599962171364,
      "loss": 2.703,
      "theoretical_loss": 3.431357015953557,
      "tokens_seen": 1993277440
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020392944959334218,
      "loss": 2.4862,
      "theoretical_loss": 3.4313476034142116,
      "tokens_seen": 1993342976
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020391289956497069,
      "loss": 2.7815,
      "theoretical_loss": 3.431338191270967,
      "tokens_seen": 1993408512
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020389634953659922,
      "loss": 2.6608,
      "theoretical_loss": 3.4313287795237932,
      "tokens_seen": 1993474048
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020387979950822773,
      "loss": 2.6333,
      "theoretical_loss": 3.4313193681726606,
      "tokens_seen": 1993539584
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020386324947985627,
      "loss": 2.7118,
      "theoretical_loss": 3.4313099572175405,
      "tokens_seen": 1993605120
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020384669945148478,
      "loss": 2.5618,
      "theoretical_loss": 3.4313005466584015,
      "tokens_seen": 1993670656
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020383014942311332,
      "loss": 2.5581,
      "theoretical_loss": 3.431291136495215,
      "tokens_seen": 1993736192
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020381359939474183,
      "loss": 2.5781,
      "theoretical_loss": 3.4312817267279505,
      "tokens_seen": 1993801728
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020379704936637037,
      "loss": 2.844,
      "theoretical_loss": 3.431272317356579,
      "tokens_seen": 1993867264
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020378049933799888,
      "loss": 2.7656,
      "theoretical_loss": 3.431262908381071,
      "tokens_seen": 1993932800
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020376394930962742,
      "loss": 2.7964,
      "theoretical_loss": 3.4312534998013966,
      "tokens_seen": 1993998336
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002037473992812559,
      "loss": 2.6929,
      "theoretical_loss": 3.431244091617526,
      "tokens_seen": 1994063872
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002037308492528844,
      "loss": 2.6298,
      "theoretical_loss": 3.4312346838294294,
      "tokens_seen": 1994129408
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020371429922451295,
      "loss": 2.4086,
      "theoretical_loss": 3.431225276437078,
      "tokens_seen": 1994194944
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020369774919614146,
      "loss": 2.747,
      "theoretical_loss": 3.431215869440441,
      "tokens_seen": 1994260480
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020368119916777,
      "loss": 2.6916,
      "theoretical_loss": 3.431206462839489,
      "tokens_seen": 1994326016
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002036646491393985,
      "loss": 2.6156,
      "theoretical_loss": 3.4311970566341934,
      "tokens_seen": 1994391552
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020364809911102704,
      "loss": 2.6565,
      "theoretical_loss": 3.4311876508245227,
      "tokens_seen": 1994457088
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020363154908265555,
      "loss": 2.6031,
      "theoretical_loss": 3.4311782454104494,
      "tokens_seen": 1994522624
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002036149990542841,
      "loss": 2.7367,
      "theoretical_loss": 3.4311688403919423,
      "tokens_seen": 1994588160
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002035984490259126,
      "loss": 2.6589,
      "theoretical_loss": 3.431159435768972,
      "tokens_seen": 1994653696
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020358189899754114,
      "loss": 2.9367,
      "theoretical_loss": 3.4311500315415095,
      "tokens_seen": 1994719232
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020356534896916965,
      "loss": 2.9284,
      "theoretical_loss": 3.431140627709525,
      "tokens_seen": 1994784768
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020354879894079819,
      "loss": 2.5519,
      "theoretical_loss": 3.4311312242729883,
      "tokens_seen": 1994850304
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002035322489124267,
      "loss": 2.5981,
      "theoretical_loss": 3.43112182123187,
      "tokens_seen": 1994915840
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020351569888405523,
      "loss": 2.6341,
      "theoretical_loss": 3.431112418586141,
      "tokens_seen": 1994981376
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020349914885568374,
      "loss": 2.8679,
      "theoretical_loss": 3.431103016335771,
      "tokens_seen": 1995046912
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020348259882731228,
      "loss": 2.8234,
      "theoretical_loss": 3.4310936144807314,
      "tokens_seen": 1995112448
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002034660487989408,
      "loss": 2.9444,
      "theoretical_loss": 3.4310842130209913,
      "tokens_seen": 1995177984
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020344949877056933,
      "loss": 2.7374,
      "theoretical_loss": 3.431074811956522,
      "tokens_seen": 1995243520
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020343294874219784,
      "loss": 2.7144,
      "theoretical_loss": 3.4310654112872934,
      "tokens_seen": 1995309056
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020341639871382638,
      "loss": 2.7529,
      "theoretical_loss": 3.4310560110132764,
      "tokens_seen": 1995374592
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002033998486854549,
      "loss": 2.8385,
      "theoretical_loss": 3.431046611134441,
      "tokens_seen": 1995440128
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020338329865708343,
      "loss": 2.6319,
      "theoretical_loss": 3.4310372116507573,
      "tokens_seen": 1995505664
    },
    {
      "epoch": 0.71,
      "objective/train/avg_token_score": 0.014820776879787445,
      "objective/train/avg_weight": 0.982559323310852,
      "objective/train/docs_used": 1128158,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5447402000427246,
      "objective/train/std_weight": 0.07266776263713837,
      "objective/train/theoretical_loss": 3.4310278125621965,
      "objective/train/tokens_used": 2016031200,
      "theoretical_loss": 3.4310278125621965,
      "tokens_seen": 1995571200
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020336674862871194,
      "loss": 2.5959,
      "theoretical_loss": 3.4310278125621965,
      "tokens_seen": 1995571200
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020335019860034047,
      "loss": 2.5844,
      "theoretical_loss": 3.431018413868729,
      "tokens_seen": 1995636736
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020333364857196898,
      "loss": 2.6831,
      "theoretical_loss": 3.4310090155703246,
      "tokens_seen": 1995702272
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020331709854359752,
      "loss": 2.7165,
      "theoretical_loss": 3.430999617666954,
      "tokens_seen": 1995767808
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020330054851522603,
      "loss": 2.6127,
      "theoretical_loss": 3.430990220158588,
      "tokens_seen": 1995833344
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020328399848685457,
      "loss": 2.7473,
      "theoretical_loss": 3.4309808230451955,
      "tokens_seen": 1995898880
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020326744845848308,
      "loss": 2.52,
      "theoretical_loss": 3.430971426326749,
      "tokens_seen": 1995964416
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020325089843011156,
      "loss": 2.5678,
      "theoretical_loss": 3.4309620300032178,
      "tokens_seen": 1996029952
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002032343484017401,
      "loss": 2.7591,
      "theoretical_loss": 3.430952634074573,
      "tokens_seen": 1996095488
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002032177983733686,
      "loss": 2.7393,
      "theoretical_loss": 3.4309432385407836,
      "tokens_seen": 1996161024
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020320124834499715,
      "loss": 2.6719,
      "theoretical_loss": 3.4309338434018217,
      "tokens_seen": 1996226560
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020318469831662566,
      "loss": 2.5466,
      "theoretical_loss": 3.430924448657657,
      "tokens_seen": 1996292096
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002031681482882542,
      "loss": 2.5315,
      "theoretical_loss": 3.43091505430826,
      "tokens_seen": 1996357632
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002031515982598827,
      "loss": 2.7867,
      "theoretical_loss": 3.4309056603536012,
      "tokens_seen": 1996423168
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020313504823151125,
      "loss": 2.5065,
      "theoretical_loss": 3.430896266793651,
      "tokens_seen": 1996488704
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020311849820313976,
      "loss": 2.7429,
      "theoretical_loss": 3.43088687362838,
      "tokens_seen": 1996554240
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002031019481747683,
      "loss": 2.7761,
      "theoretical_loss": 3.4308774808577587,
      "tokens_seen": 1996619776
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002030853981463968,
      "loss": 2.8143,
      "theoretical_loss": 3.4308680884817573,
      "tokens_seen": 1996685312
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020306884811802534,
      "loss": 2.6998,
      "theoretical_loss": 3.430858696500346,
      "tokens_seen": 1996750848
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020305229808965385,
      "loss": 2.8518,
      "theoretical_loss": 3.430849304913496,
      "tokens_seen": 1996816384
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002030357480612824,
      "loss": 2.6727,
      "theoretical_loss": 3.4308399137211776,
      "tokens_seen": 1996881920
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002030191980329109,
      "loss": 2.6577,
      "theoretical_loss": 3.430830522923361,
      "tokens_seen": 1996947456
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020300264800453944,
      "loss": 2.5401,
      "theoretical_loss": 3.4308211325200166,
      "tokens_seen": 1997012992
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020298609797616795,
      "loss": 2.6101,
      "theoretical_loss": 3.430811742511115,
      "tokens_seen": 1997078528
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020296954794779649,
      "loss": 2.8214,
      "theoretical_loss": 3.4308023528966274,
      "tokens_seen": 1997144064
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000202952997919425,
      "loss": 2.4803,
      "theoretical_loss": 3.4307929636765233,
      "tokens_seen": 1997209600
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020293644789105353,
      "loss": 2.7804,
      "theoretical_loss": 3.4307835748507736,
      "tokens_seen": 1997275136
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020291989786268204,
      "loss": 2.7979,
      "theoretical_loss": 3.430774186419349,
      "tokens_seen": 1997340672
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020290334783431058,
      "loss": 2.5788,
      "theoretical_loss": 3.4307647983822194,
      "tokens_seen": 1997406208
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002028867978059391,
      "loss": 2.4968,
      "theoretical_loss": 3.430755410739356,
      "tokens_seen": 1997471744
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020287024777756763,
      "loss": 2.4878,
      "theoretical_loss": 3.4307460234907285,
      "tokens_seen": 1997537280
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020285369774919614,
      "loss": 2.5348,
      "theoretical_loss": 3.430736636636308,
      "tokens_seen": 1997602816
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020283714772082468,
      "loss": 2.3697,
      "theoretical_loss": 3.430727250176065,
      "tokens_seen": 1997668352
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002028205976924532,
      "loss": 2.7128,
      "theoretical_loss": 3.4307178641099703,
      "tokens_seen": 1997733888
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020280404766408173,
      "loss": 2.596,
      "theoretical_loss": 3.4307084784379933,
      "tokens_seen": 1997799424
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020278749763571024,
      "loss": 2.6087,
      "theoretical_loss": 3.430699093160106,
      "tokens_seen": 1997864960
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020277094760733875,
      "loss": 2.607,
      "theoretical_loss": 3.4306897082762777,
      "tokens_seen": 1997930496
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020275439757896726,
      "loss": 2.5565,
      "theoretical_loss": 3.4306803237864796,
      "tokens_seen": 1997996032
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002027378475505958,
      "loss": 2.7264,
      "theoretical_loss": 3.4306709396906823,
      "tokens_seen": 1998061568
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002027212975222243,
      "loss": 2.5539,
      "theoretical_loss": 3.4306615559888556,
      "tokens_seen": 1998127104
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020270474749385284,
      "loss": 2.6163,
      "theoretical_loss": 3.4306521726809707,
      "tokens_seen": 1998192640
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020268819746548135,
      "loss": 2.712,
      "theoretical_loss": 3.430642789766998,
      "tokens_seen": 1998258176
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002026716474371099,
      "loss": 2.6789,
      "theoretical_loss": 3.4306334072469085,
      "tokens_seen": 1998323712
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002026550974087384,
      "loss": 2.5322,
      "theoretical_loss": 3.4306240251206717,
      "tokens_seen": 1998389248
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020263854738036694,
      "loss": 2.8088,
      "theoretical_loss": 3.4306146433882585,
      "tokens_seen": 1998454784
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020262199735199545,
      "loss": 2.7156,
      "theoretical_loss": 3.43060526204964,
      "tokens_seen": 1998520320
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020260544732362399,
      "loss": 2.6811,
      "theoretical_loss": 3.4305958811047863,
      "tokens_seen": 1998585856
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002025888972952525,
      "loss": 2.6429,
      "theoretical_loss": 3.4305865005536686,
      "tokens_seen": 1998651392
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020257234726688103,
      "loss": 2.6315,
      "theoretical_loss": 3.430577120396256,
      "tokens_seen": 1998716928
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020255579723850954,
      "loss": 2.7831,
      "theoretical_loss": 3.4305677406325206,
      "tokens_seen": 1998782464
    },
    {
      "epoch": 0.71,
      "objective/train/avg_token_score": 0.004881288856267929,
      "objective/train/avg_weight": 0.994253933429718,
      "objective/train/docs_used": 1130081,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8205318450927734,
      "objective/train/std_weight": 0.04820758476853371,
      "objective/train/theoretical_loss": 3.430558361262433,
      "objective/train/tokens_used": 2019308000,
      "theoretical_loss": 3.430558361262433,
      "tokens_seen": 1998848000
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020253924721013808,
      "loss": 2.7825,
      "theoretical_loss": 3.430558361262433,
      "tokens_seen": 1998848000
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002025226971817666,
      "loss": 2.5154,
      "theoretical_loss": 3.430548982285962,
      "tokens_seen": 1998913536
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020250614715339513,
      "loss": 2.5721,
      "theoretical_loss": 3.43053960370308,
      "tokens_seen": 1998979072
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020248959712502364,
      "loss": 2.7097,
      "theoretical_loss": 3.4305302255137566,
      "tokens_seen": 1999044608
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020247304709665218,
      "loss": 2.4773,
      "theoretical_loss": 3.430520847717963,
      "tokens_seen": 1999110144
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002024564970682807,
      "loss": 2.4495,
      "theoretical_loss": 3.430511470315669,
      "tokens_seen": 1999175680
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020243994703990923,
      "loss": 2.7405,
      "theoretical_loss": 3.4305020933068464,
      "tokens_seen": 1999241216
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020242339701153774,
      "loss": 2.743,
      "theoretical_loss": 3.4304927166914645,
      "tokens_seen": 1999306752
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020240684698316627,
      "loss": 2.7293,
      "theoretical_loss": 3.430483340469495,
      "tokens_seen": 1999372288
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020239029695479478,
      "loss": 2.5407,
      "theoretical_loss": 3.4304739646409077,
      "tokens_seen": 1999437824
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020237374692642332,
      "loss": 2.5902,
      "theoretical_loss": 3.4304645892056733,
      "tokens_seen": 1999503360
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020235719689805183,
      "loss": 2.7636,
      "theoretical_loss": 3.4304552141637625,
      "tokens_seen": 1999568896
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020234064686968037,
      "loss": 2.4441,
      "theoretical_loss": 3.4304458395151465,
      "tokens_seen": 1999634432
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020232409684130888,
      "loss": 2.4851,
      "theoretical_loss": 3.430436465259795,
      "tokens_seen": 1999699968
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020230754681293742,
      "loss": 2.4616,
      "theoretical_loss": 3.4304270913976795,
      "tokens_seen": 1999765504
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020229099678456593,
      "loss": 2.5794,
      "theoretical_loss": 3.430417717928769,
      "tokens_seen": 1999831040
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002022744467561944,
      "loss": 2.8677,
      "theoretical_loss": 3.4304083448530367,
      "tokens_seen": 1999896576
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020225789672782295,
      "loss": 2.7206,
      "theoretical_loss": 3.4303989721704506,
      "tokens_seen": 1999962112
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020224134669945146,
      "loss": 2.5083,
      "theoretical_loss": 3.430389599880983,
      "tokens_seen": 2000027648
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020222479667108,
      "loss": 2.6822,
      "theoretical_loss": 3.430380227984604,
      "tokens_seen": 2000093184
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002022082466427085,
      "loss": 2.5431,
      "theoretical_loss": 3.4303708564812845,
      "tokens_seen": 2000158720
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020219169661433705,
      "loss": 2.578,
      "theoretical_loss": 3.4303614853709945,
      "tokens_seen": 2000224256
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020217514658596556,
      "loss": 2.8337,
      "theoretical_loss": 3.4303521146537053,
      "tokens_seen": 2000289792
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002021585965575941,
      "loss": 2.7192,
      "theoretical_loss": 3.430342744329387,
      "tokens_seen": 2000355328
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002021420465292226,
      "loss": 2.6043,
      "theoretical_loss": 3.430333374398011,
      "tokens_seen": 2000420864
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020212549650085114,
      "loss": 2.406,
      "theoretical_loss": 3.430324004859547,
      "tokens_seen": 2000486400
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020210894647247965,
      "loss": 2.4883,
      "theoretical_loss": 3.430314635713967,
      "tokens_seen": 2000551936
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002020923964441082,
      "loss": 2.424,
      "theoretical_loss": 3.4303052669612395,
      "tokens_seen": 2000617472
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002020758464157367,
      "loss": 2.7292,
      "theoretical_loss": 3.4302958986013374,
      "tokens_seen": 2000683008
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020205929638736524,
      "loss": 2.8994,
      "theoretical_loss": 3.43028653063423,
      "tokens_seen": 2000748544
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020204274635899375,
      "loss": 2.6423,
      "theoretical_loss": 3.4302771630598885,
      "tokens_seen": 2000814080
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020202619633062229,
      "loss": 2.7292,
      "theoretical_loss": 3.4302677958782835,
      "tokens_seen": 2000879616
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002020096463022508,
      "loss": 2.6637,
      "theoretical_loss": 3.4302584290893856,
      "tokens_seen": 2000945152
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020199309627387933,
      "loss": 2.647,
      "theoretical_loss": 3.430249062693165,
      "tokens_seen": 2001010688
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020197654624550784,
      "loss": 2.7803,
      "theoretical_loss": 3.430239696689594,
      "tokens_seen": 2001076224
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020195999621713638,
      "loss": 2.4637,
      "theoretical_loss": 3.4302303310786417,
      "tokens_seen": 2001141760
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002019434461887649,
      "loss": 2.8131,
      "theoretical_loss": 3.4302209658602787,
      "tokens_seen": 2001207296
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020192689616039343,
      "loss": 2.6533,
      "theoretical_loss": 3.4302116010344763,
      "tokens_seen": 2001272832
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020191034613202194,
      "loss": 2.6998,
      "theoretical_loss": 3.4302022366012057,
      "tokens_seen": 2001338368
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020189379610365048,
      "loss": 2.7111,
      "theoretical_loss": 3.4301928725604367,
      "tokens_seen": 2001403904
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000201877246075279,
      "loss": 2.8556,
      "theoretical_loss": 3.4301835089121404,
      "tokens_seen": 2001469440
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020186069604690753,
      "loss": 2.7653,
      "theoretical_loss": 3.430174145656287,
      "tokens_seen": 2001534976
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020184414601853604,
      "loss": 2.3048,
      "theoretical_loss": 3.4301647827928483,
      "tokens_seen": 2001600512
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020182759599016457,
      "loss": 2.5748,
      "theoretical_loss": 3.430155420321794,
      "tokens_seen": 2001666048
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020181104596179308,
      "loss": 2.5906,
      "theoretical_loss": 3.430146058243095,
      "tokens_seen": 2001731584
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020179449593342162,
      "loss": 2.6724,
      "theoretical_loss": 3.430136696556722,
      "tokens_seen": 2001797120
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002017779459050501,
      "loss": 2.565,
      "theoretical_loss": 3.4301273352626462,
      "tokens_seen": 2001862656
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00020176139587667861,
      "loss": 2.745,
      "theoretical_loss": 3.430117974360838,
      "tokens_seen": 2001928192
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020174484584830715,
      "loss": 2.6848,
      "theoretical_loss": 3.4301086138512678,
      "tokens_seen": 2001993728
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020172829581993566,
      "loss": 2.6844,
      "theoretical_loss": 3.430099253733907,
      "tokens_seen": 2002059264
    },
    {
      "epoch": 0.72,
      "objective/train/avg_token_score": 0.007438336499035358,
      "objective/train/avg_weight": 0.9912476539611816,
      "objective/train/docs_used": 1132167,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.829395055770874,
      "objective/train/std_weight": 0.050756681710481644,
      "objective/train/theoretical_loss": 3.4300898940087254,
      "objective/train/tokens_used": 2022584800,
      "theoretical_loss": 3.4300898940087254,
      "tokens_seen": 2002124800
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002017117457915642,
      "loss": 2.7325,
      "theoretical_loss": 3.4300898940087254,
      "tokens_seen": 2002124800
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002016951957631927,
      "loss": 2.6863,
      "theoretical_loss": 3.430080534675695,
      "tokens_seen": 2002190336
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020167864573482125,
      "loss": 2.4842,
      "theoretical_loss": 3.4300711757347853,
      "tokens_seen": 2002255872
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020166209570644976,
      "loss": 2.5416,
      "theoretical_loss": 3.4300618171859676,
      "tokens_seen": 2002321408
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002016455456780783,
      "loss": 2.4414,
      "theoretical_loss": 3.4300524590292127,
      "tokens_seen": 2002386944
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002016289956497068,
      "loss": 2.4864,
      "theoretical_loss": 3.4300431012644914,
      "tokens_seen": 2002452480
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020161244562133534,
      "loss": 2.5755,
      "theoretical_loss": 3.430033743891774,
      "tokens_seen": 2002518016
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020159589559296385,
      "loss": 2.5713,
      "theoretical_loss": 3.4300243869110316,
      "tokens_seen": 2002583552
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002015793455645924,
      "loss": 2.7926,
      "theoretical_loss": 3.430015030322235,
      "tokens_seen": 2002649088
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002015627955362209,
      "loss": 2.748,
      "theoretical_loss": 3.4300056741253546,
      "tokens_seen": 2002714624
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020154624550784944,
      "loss": 2.446,
      "theoretical_loss": 3.4299963183203617,
      "tokens_seen": 2002780160
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020152969547947795,
      "loss": 2.3983,
      "theoretical_loss": 3.4299869629072264,
      "tokens_seen": 2002845696
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002015131454511065,
      "loss": 2.6205,
      "theoretical_loss": 3.4299776078859203,
      "tokens_seen": 2002911232
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020149659542273503,
      "loss": 2.7255,
      "theoretical_loss": 3.4299682532564137,
      "tokens_seen": 2002976768
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020148004539436354,
      "loss": 2.4161,
      "theoretical_loss": 3.4299588990186773,
      "tokens_seen": 2003042304
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020146349536599207,
      "loss": 2.9394,
      "theoretical_loss": 3.429949545172682,
      "tokens_seen": 2003107840
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020144694533762058,
      "loss": 2.5717,
      "theoretical_loss": 3.429940191718398,
      "tokens_seen": 2003173376
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020143039530924912,
      "loss": 2.3974,
      "theoretical_loss": 3.429930838655797,
      "tokens_seen": 2003238912
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020141384528087763,
      "loss": 2.7188,
      "theoretical_loss": 3.4299214859848495,
      "tokens_seen": 2003304448
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020139729525250617,
      "loss": 2.3095,
      "theoretical_loss": 3.429912133705526,
      "tokens_seen": 2003369984
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020138074522413468,
      "loss": 2.6301,
      "theoretical_loss": 3.4299027818177974,
      "tokens_seen": 2003435520
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020136419519576322,
      "loss": 2.5374,
      "theoretical_loss": 3.429893430321635,
      "tokens_seen": 2003501056
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020134764516739173,
      "loss": 2.556,
      "theoretical_loss": 3.4298840792170084,
      "tokens_seen": 2003566592
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020133109513902027,
      "loss": 2.6464,
      "theoretical_loss": 3.42987472850389,
      "tokens_seen": 2003632128
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020131454511064878,
      "loss": 2.5944,
      "theoretical_loss": 3.4298653781822495,
      "tokens_seen": 2003697664
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020129799508227726,
      "loss": 2.7276,
      "theoretical_loss": 3.429856028252058,
      "tokens_seen": 2003763200
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002012814450539058,
      "loss": 2.8416,
      "theoretical_loss": 3.429846678713286,
      "tokens_seen": 2003828736
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002012648950255343,
      "loss": 2.6637,
      "theoretical_loss": 3.4298373295659053,
      "tokens_seen": 2003894272
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020124834499716284,
      "loss": 2.786,
      "theoretical_loss": 3.4298279808098853,
      "tokens_seen": 2003959808
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020123179496879136,
      "loss": 2.6856,
      "theoretical_loss": 3.4298186324451976,
      "tokens_seen": 2004025344
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002012152449404199,
      "loss": 2.6681,
      "theoretical_loss": 3.4298092844718133,
      "tokens_seen": 2004090880
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002011986949120484,
      "loss": 2.6158,
      "theoretical_loss": 3.4297999368897027,
      "tokens_seen": 2004156416
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020118214488367694,
      "loss": 2.1377,
      "theoretical_loss": 3.429790589698837,
      "tokens_seen": 2004221952
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020116559485530545,
      "loss": 2.573,
      "theoretical_loss": 3.4297812428991867,
      "tokens_seen": 2004287488
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000201149044826934,
      "loss": 2.7107,
      "theoretical_loss": 3.429771896490723,
      "tokens_seen": 2004353024
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002011324947985625,
      "loss": 2.8347,
      "theoretical_loss": 3.4297625504734164,
      "tokens_seen": 2004418560
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020111594477019104,
      "loss": 2.7526,
      "theoretical_loss": 3.4297532048472377,
      "tokens_seen": 2004484096
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020109939474181955,
      "loss": 2.5948,
      "theoretical_loss": 3.4297438596121577,
      "tokens_seen": 2004549632
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020108284471344808,
      "loss": 2.6959,
      "theoretical_loss": 3.429734514768148,
      "tokens_seen": 2004615168
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002010662946850766,
      "loss": 2.7718,
      "theoretical_loss": 3.4297251703151783,
      "tokens_seen": 2004680704
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020104974465670513,
      "loss": 2.5935,
      "theoretical_loss": 3.4297158262532204,
      "tokens_seen": 2004746240
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020103319462833364,
      "loss": 2.4922,
      "theoretical_loss": 3.4297064825822448,
      "tokens_seen": 2004811776
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020101664459996218,
      "loss": 2.4367,
      "theoretical_loss": 3.4296971393022226,
      "tokens_seen": 2004877312
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002010000945715907,
      "loss": 2.7071,
      "theoretical_loss": 3.4296877964131243,
      "tokens_seen": 2004942848
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020098354454321923,
      "loss": 2.3483,
      "theoretical_loss": 3.4296784539149208,
      "tokens_seen": 2005008384
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020096699451484774,
      "loss": 2.9703,
      "theoretical_loss": 3.4296691118075833,
      "tokens_seen": 2005073920
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020095044448647628,
      "loss": 2.6348,
      "theoretical_loss": 3.4296597700910825,
      "tokens_seen": 2005139456
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002009338944581048,
      "loss": 2.8865,
      "theoretical_loss": 3.429650428765389,
      "tokens_seen": 2005204992
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020091734442973332,
      "loss": 2.4424,
      "theoretical_loss": 3.4296410878304737,
      "tokens_seen": 2005270528
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020090079440136184,
      "loss": 2.5738,
      "theoretical_loss": 3.429631747286308,
      "tokens_seen": 2005336064
    },
    {
      "epoch": 0.72,
      "objective/train/avg_token_score": 0.022219520062208176,
      "objective/train/avg_weight": 0.973846971988678,
      "objective/train/docs_used": 1133291,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.288991451263428,
      "objective/train/std_weight": 0.11907588690519333,
      "objective/train/theoretical_loss": 3.4296224071328623,
      "objective/train/tokens_used": 2025861600,
      "theoretical_loss": 3.4296224071328623,
      "tokens_seen": 2005401600
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020088424437299037,
      "loss": 2.6142,
      "theoretical_loss": 3.4296224071328623,
      "tokens_seen": 2005401600
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020086769434461888,
      "loss": 2.6621,
      "theoretical_loss": 3.4296130673701075,
      "tokens_seen": 2005467136
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020085114431624742,
      "loss": 2.708,
      "theoretical_loss": 3.429603727998015,
      "tokens_seen": 2005532672
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020083459428787593,
      "loss": 2.5382,
      "theoretical_loss": 3.429594389016555,
      "tokens_seen": 2005598208
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020081804425950447,
      "loss": 2.6544,
      "theoretical_loss": 3.429585050425699,
      "tokens_seen": 2005663744
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020080149423113295,
      "loss": 2.4537,
      "theoretical_loss": 3.429575712225418,
      "tokens_seen": 2005729280
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020078494420276146,
      "loss": 2.4789,
      "theoretical_loss": 3.429566374415682,
      "tokens_seen": 2005794816
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020076839417439,
      "loss": 2.5253,
      "theoretical_loss": 3.4295570369964623,
      "tokens_seen": 2005860352
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002007518441460185,
      "loss": 2.6256,
      "theoretical_loss": 3.4295476999677303,
      "tokens_seen": 2005925888
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020073529411764705,
      "loss": 2.534,
      "theoretical_loss": 3.4295383633294563,
      "tokens_seen": 2005991424
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020071874408927556,
      "loss": 2.6793,
      "theoretical_loss": 3.429529027081612,
      "tokens_seen": 2006056960
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002007021940609041,
      "loss": 2.5608,
      "theoretical_loss": 3.4295196912241677,
      "tokens_seen": 2006122496
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002006856440325326,
      "loss": 2.6055,
      "theoretical_loss": 3.429510355757094,
      "tokens_seen": 2006188032
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020066909400416114,
      "loss": 2.5526,
      "theoretical_loss": 3.4295010206803624,
      "tokens_seen": 2006253568
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020065254397578965,
      "loss": 2.6194,
      "theoretical_loss": 3.429491685993944,
      "tokens_seen": 2006319104
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002006359939474182,
      "loss": 2.5022,
      "theoretical_loss": 3.4294823516978092,
      "tokens_seen": 2006384640
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002006194439190467,
      "loss": 2.6442,
      "theoretical_loss": 3.4294730177919295,
      "tokens_seen": 2006450176
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020060289389067524,
      "loss": 2.3393,
      "theoretical_loss": 3.429463684276275,
      "tokens_seen": 2006515712
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020058634386230375,
      "loss": 2.5966,
      "theoretical_loss": 3.429454351150817,
      "tokens_seen": 2006581248
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002005697938339323,
      "loss": 2.5637,
      "theoretical_loss": 3.429445018415527,
      "tokens_seen": 2006646784
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002005532438055608,
      "loss": 2.7187,
      "theoretical_loss": 3.4294356860703754,
      "tokens_seen": 2006712320
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020053669377718934,
      "loss": 2.5416,
      "theoretical_loss": 3.429426354115334,
      "tokens_seen": 2006777856
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020052014374881785,
      "loss": 2.6452,
      "theoretical_loss": 3.4294170225503717,
      "tokens_seen": 2006843392
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020050359372044638,
      "loss": 2.9292,
      "theoretical_loss": 3.429407691375461,
      "tokens_seen": 2006908928
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002004870436920749,
      "loss": 2.803,
      "theoretical_loss": 3.4293983605905733,
      "tokens_seen": 2006974464
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020047049366370343,
      "loss": 2.5349,
      "theoretical_loss": 3.429389030195679,
      "tokens_seen": 2007040000
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020045394363533194,
      "loss": 2.8632,
      "theoretical_loss": 3.4293797001907484,
      "tokens_seen": 2007105536
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020043739360696048,
      "loss": 2.6071,
      "theoretical_loss": 3.4293703705757537,
      "tokens_seen": 2007171072
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000200420843578589,
      "loss": 2.5262,
      "theoretical_loss": 3.4293610413506643,
      "tokens_seen": 2007236608
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020040429355021753,
      "loss": 2.6067,
      "theoretical_loss": 3.4293517125154525,
      "tokens_seen": 2007302144
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020038774352184604,
      "loss": 2.6799,
      "theoretical_loss": 3.429342384070089,
      "tokens_seen": 2007367680
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020037119349347458,
      "loss": 2.7363,
      "theoretical_loss": 3.4293330560145447,
      "tokens_seen": 2007433216
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020035464346510309,
      "loss": 2.7574,
      "theoretical_loss": 3.4293237283487903,
      "tokens_seen": 2007498752
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020033809343673162,
      "loss": 2.6873,
      "theoretical_loss": 3.4293144010727974,
      "tokens_seen": 2007564288
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020032154340836013,
      "loss": 2.525,
      "theoretical_loss": 3.4293050741865363,
      "tokens_seen": 2007629824
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020030499337998862,
      "loss": 2.5353,
      "theoretical_loss": 3.429295747689978,
      "tokens_seen": 2007695360
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020028844335161716,
      "loss": 2.4948,
      "theoretical_loss": 3.4292864215830945,
      "tokens_seen": 2007760896
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020027189332324567,
      "loss": 2.6861,
      "theoretical_loss": 3.4292770958658556,
      "tokens_seen": 2007826432
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002002553432948742,
      "loss": 2.6674,
      "theoretical_loss": 3.4292677705382335,
      "tokens_seen": 2007891968
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020023879326650274,
      "loss": 2.6646,
      "theoretical_loss": 3.4292584456001975,
      "tokens_seen": 2007957504
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020022224323813125,
      "loss": 2.5688,
      "theoretical_loss": 3.42924912105172,
      "tokens_seen": 2008023040
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002002056932097598,
      "loss": 2.4234,
      "theoretical_loss": 3.429239796892772,
      "tokens_seen": 2008088576
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002001891431813883,
      "loss": 2.5445,
      "theoretical_loss": 3.429230473123324,
      "tokens_seen": 2008154112
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020017259315301684,
      "loss": 2.5317,
      "theoretical_loss": 3.429221149743347,
      "tokens_seen": 2008219648
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020015604312464535,
      "loss": 2.5881,
      "theoretical_loss": 3.4292118267528124,
      "tokens_seen": 2008285184
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020013949309627388,
      "loss": 2.6289,
      "theoretical_loss": 3.429202504151691,
      "tokens_seen": 2008350720
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002001229430679024,
      "loss": 2.5526,
      "theoretical_loss": 3.4291931819399535,
      "tokens_seen": 2008416256
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020010639303953093,
      "loss": 2.5968,
      "theoretical_loss": 3.429183860117572,
      "tokens_seen": 2008481792
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020008984301115944,
      "loss": 2.5547,
      "theoretical_loss": 3.4291745386845163,
      "tokens_seen": 2008547328
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020007329298278798,
      "loss": 2.6841,
      "theoretical_loss": 3.429165217640758,
      "tokens_seen": 2008612864
    },
    {
      "epoch": 0.72,
      "objective/train/avg_token_score": 0.010418794117867947,
      "objective/train/avg_weight": 0.9877451062202454,
      "objective/train/docs_used": 1134873,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0099399089813232,
      "objective/train/std_weight": 0.046785056591033936,
      "objective/train/theoretical_loss": 3.429155896986268,
      "objective/train/tokens_used": 2029138400,
      "theoretical_loss": 3.429155896986268,
      "tokens_seen": 2008678400
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002000567429544165,
      "loss": 2.535,
      "theoretical_loss": 3.429155896986268,
      "tokens_seen": 2008678400
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020004019292604503,
      "loss": 2.4228,
      "theoretical_loss": 3.4291465767210174,
      "tokens_seen": 2008743936
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020002364289767354,
      "loss": 2.9122,
      "theoretical_loss": 3.429137256844977,
      "tokens_seen": 2008809472
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00020000709286930208,
      "loss": 2.2675,
      "theoretical_loss": 3.429127937358119,
      "tokens_seen": 2008875008
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001999905428409306,
      "loss": 2.6196,
      "theoretical_loss": 3.429118618260413,
      "tokens_seen": 2008940544
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019997399281255912,
      "loss": 2.3287,
      "theoretical_loss": 3.429109299551831,
      "tokens_seen": 2009006080
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019995744278418763,
      "loss": 2.6049,
      "theoretical_loss": 3.4290999812323433,
      "tokens_seen": 2009071616
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019994089275581617,
      "loss": 2.5087,
      "theoretical_loss": 3.4290906633019214,
      "tokens_seen": 2009137152
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019992434272744468,
      "loss": 2.6805,
      "theoretical_loss": 3.4290813457605362,
      "tokens_seen": 2009202688
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019990779269907322,
      "loss": 2.3477,
      "theoretical_loss": 3.429072028608159,
      "tokens_seen": 2009268224
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019989124267070173,
      "loss": 2.4531,
      "theoretical_loss": 3.429062711844761,
      "tokens_seen": 2009333760
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019987469264233027,
      "loss": 2.5269,
      "theoretical_loss": 3.429053395470313,
      "tokens_seen": 2009399296
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019985814261395878,
      "loss": 2.7802,
      "theoretical_loss": 3.429044079484786,
      "tokens_seen": 2009464832
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019984159258558732,
      "loss": 2.5727,
      "theoretical_loss": 3.4290347638881515,
      "tokens_seen": 2009530368
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019982504255721583,
      "loss": 2.5338,
      "theoretical_loss": 3.42902544868038,
      "tokens_seen": 2009595904
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001998084925288443,
      "loss": 2.1556,
      "theoretical_loss": 3.4290161338614427,
      "tokens_seen": 2009661440
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019979194250047285,
      "loss": 2.4673,
      "theoretical_loss": 3.4290068194313115,
      "tokens_seen": 2009726976
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019977539247210136,
      "loss": 2.6557,
      "theoretical_loss": 3.428997505389956,
      "tokens_seen": 2009792512
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001997588424437299,
      "loss": 2.6043,
      "theoretical_loss": 3.4289881917373486,
      "tokens_seen": 2009858048
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001997422924153584,
      "loss": 2.4636,
      "theoretical_loss": 3.42897887847346,
      "tokens_seen": 2009923584
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019972574238698694,
      "loss": 2.7026,
      "theoretical_loss": 3.428969565598261,
      "tokens_seen": 2009989120
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019970919235861545,
      "loss": 2.9265,
      "theoretical_loss": 3.428960253111723,
      "tokens_seen": 2010054656
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000199692642330244,
      "loss": 2.6985,
      "theoretical_loss": 3.4289509410138175,
      "tokens_seen": 2010120192
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001996760923018725,
      "loss": 2.3442,
      "theoretical_loss": 3.4289416293045147,
      "tokens_seen": 2010185728
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019965954227350104,
      "loss": 2.464,
      "theoretical_loss": 3.4289323179837865,
      "tokens_seen": 2010251264
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019964299224512955,
      "loss": 2.6683,
      "theoretical_loss": 3.4289230070516035,
      "tokens_seen": 2010316800
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001996264422167581,
      "loss": 2.5216,
      "theoretical_loss": 3.428913696507937,
      "tokens_seen": 2010382336
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001996098921883866,
      "loss": 2.7089,
      "theoretical_loss": 3.428904386352758,
      "tokens_seen": 2010447872
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019959334216001514,
      "loss": 2.5705,
      "theoretical_loss": 3.428895076586038,
      "tokens_seen": 2010513408
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019957679213164365,
      "loss": 2.3864,
      "theoretical_loss": 3.428885767207748,
      "tokens_seen": 2010578944
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019956024210327218,
      "loss": 2.8169,
      "theoretical_loss": 3.428876458217859,
      "tokens_seen": 2010644480
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001995436920749007,
      "loss": 2.6799,
      "theoretical_loss": 3.4288671496163423,
      "tokens_seen": 2010710016
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019952714204652923,
      "loss": 2.6623,
      "theoretical_loss": 3.4288578414031687,
      "tokens_seen": 2010775552
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019951059201815774,
      "loss": 2.6302,
      "theoretical_loss": 3.42884853357831,
      "tokens_seen": 2010841088
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019949404198978628,
      "loss": 2.567,
      "theoretical_loss": 3.4288392261417364,
      "tokens_seen": 2010906624
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001994774919614148,
      "loss": 2.6398,
      "theoretical_loss": 3.4288299190934195,
      "tokens_seen": 2010972160
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019946094193304333,
      "loss": 2.8993,
      "theoretical_loss": 3.4288206124333307,
      "tokens_seen": 2011037696
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019944439190467184,
      "loss": 2.4758,
      "theoretical_loss": 3.4288113061614407,
      "tokens_seen": 2011103232
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019942784187630038,
      "loss": 2.5979,
      "theoretical_loss": 3.4288020002777215,
      "tokens_seen": 2011168768
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019941129184792889,
      "loss": 2.5628,
      "theoretical_loss": 3.4287926947821434,
      "tokens_seen": 2011234304
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019939474181955742,
      "loss": 2.4455,
      "theoretical_loss": 3.4287833896746776,
      "tokens_seen": 2011299840
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019937819179118593,
      "loss": 2.6778,
      "theoretical_loss": 3.428774084955296,
      "tokens_seen": 2011365376
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019936164176281447,
      "loss": 2.6597,
      "theoretical_loss": 3.428764780623969,
      "tokens_seen": 2011430912
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019934509173444298,
      "loss": 2.711,
      "theoretical_loss": 3.428755476680668,
      "tokens_seen": 2011496448
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019932854170607147,
      "loss": 2.5147,
      "theoretical_loss": 3.428746173125364,
      "tokens_seen": 2011561984
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001993119916777,
      "loss": 2.643,
      "theoretical_loss": 3.428736869958029,
      "tokens_seen": 2011627520
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001992954416493285,
      "loss": 2.4935,
      "theoretical_loss": 3.4287275671786333,
      "tokens_seen": 2011693056
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019927889162095705,
      "loss": 2.6414,
      "theoretical_loss": 3.4287182647871486,
      "tokens_seen": 2011758592
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019926234159258556,
      "loss": 2.5879,
      "theoretical_loss": 3.4287089627835456,
      "tokens_seen": 2011824128
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001992457915642141,
      "loss": 2.6803,
      "theoretical_loss": 3.428699661167796,
      "tokens_seen": 2011889664
    },
    {
      "epoch": 0.72,
      "objective/train/avg_token_score": 0.04389012232422829,
      "objective/train/avg_weight": 0.9483633041381836,
      "objective/train/docs_used": 1136813,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.952396869659424,
      "objective/train/std_weight": 0.19600386917591095,
      "objective/train/theoretical_loss": 3.4286903599398704,
      "objective/train/tokens_used": 2032415200,
      "theoretical_loss": 3.4286903599398704,
      "tokens_seen": 2011955200
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001992292415358426,
      "loss": 1.9555,
      "theoretical_loss": 3.4286903599398704,
      "tokens_seen": 2011955200
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019921269150747115,
      "loss": 2.7407,
      "theoretical_loss": 3.4286810590997403,
      "tokens_seen": 2012020736
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019919614147909966,
      "loss": 2.6557,
      "theoretical_loss": 3.4286717586473774,
      "tokens_seen": 2012086272
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001991795914507282,
      "loss": 2.7113,
      "theoretical_loss": 3.4286624585827523,
      "tokens_seen": 2012151808
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001991630414223567,
      "loss": 2.6972,
      "theoretical_loss": 3.428653158905836,
      "tokens_seen": 2012217344
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019914649139398524,
      "loss": 2.3766,
      "theoretical_loss": 3.4286438596166007,
      "tokens_seen": 2012282880
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019912994136561375,
      "loss": 2.3557,
      "theoretical_loss": 3.428634560715017,
      "tokens_seen": 2012348416
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001991133913372423,
      "loss": 2.7104,
      "theoretical_loss": 3.428625262201056,
      "tokens_seen": 2012413952
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001990968413088708,
      "loss": 2.5699,
      "theoretical_loss": 3.4286159640746883,
      "tokens_seen": 2012479488
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019908029128049934,
      "loss": 2.8203,
      "theoretical_loss": 3.4286066663358863,
      "tokens_seen": 2012545024
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019906374125212785,
      "loss": 2.6914,
      "theoretical_loss": 3.428597368984621,
      "tokens_seen": 2012610560
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001990471912237564,
      "loss": 2.6625,
      "theoretical_loss": 3.4285880720208635,
      "tokens_seen": 2012676096
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001990306411953849,
      "loss": 2.5794,
      "theoretical_loss": 3.428578775444585,
      "tokens_seen": 2012741632
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019901409116701343,
      "loss": 2.6241,
      "theoretical_loss": 3.428569479255756,
      "tokens_seen": 2012807168
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019899754113864195,
      "loss": 2.6781,
      "theoretical_loss": 3.4285601834543487,
      "tokens_seen": 2012872704
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019898099111027048,
      "loss": 2.582,
      "theoretical_loss": 3.428550888040334,
      "tokens_seen": 2012938240
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000198964441081899,
      "loss": 2.6532,
      "theoretical_loss": 3.4285415930136836,
      "tokens_seen": 2013003776
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019894789105352753,
      "loss": 2.4485,
      "theoretical_loss": 3.4285322983743676,
      "tokens_seen": 2013069312
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019893134102515604,
      "loss": 2.7494,
      "theoretical_loss": 3.4285230041223587,
      "tokens_seen": 2013134848
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019891479099678458,
      "loss": 2.7782,
      "theoretical_loss": 3.4285137102576275,
      "tokens_seen": 2013200384
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001988982409684131,
      "loss": 2.379,
      "theoretical_loss": 3.428504416780145,
      "tokens_seen": 2013265920
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019888169094004163,
      "loss": 2.4883,
      "theoretical_loss": 3.428495123689882,
      "tokens_seen": 2013331456
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019886514091167014,
      "loss": 2.6753,
      "theoretical_loss": 3.428485830986811,
      "tokens_seen": 2013396992
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019884859088329867,
      "loss": 2.642,
      "theoretical_loss": 3.4284765386709033,
      "tokens_seen": 2013462528
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019883204085492716,
      "loss": 2.6025,
      "theoretical_loss": 3.4284672467421284,
      "tokens_seen": 2013528064
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001988154908265557,
      "loss": 2.6232,
      "theoretical_loss": 3.4284579552004595,
      "tokens_seen": 2013593600
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001987989407981842,
      "loss": 2.2673,
      "theoretical_loss": 3.4284486640458667,
      "tokens_seen": 2013659136
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019878239076981274,
      "loss": 2.6784,
      "theoretical_loss": 3.428439373278322,
      "tokens_seen": 2013724672
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019876584074144125,
      "loss": 2.7047,
      "theoretical_loss": 3.4284300828977967,
      "tokens_seen": 2013790208
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001987492907130698,
      "loss": 2.7305,
      "theoretical_loss": 3.428420792904261,
      "tokens_seen": 2013855744
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001987327406846983,
      "loss": 2.5632,
      "theoretical_loss": 3.428411503297687,
      "tokens_seen": 2013921280
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019871619065632684,
      "loss": 2.5221,
      "theoretical_loss": 3.428402214078046,
      "tokens_seen": 2013986816
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019869964062795535,
      "loss": 2.6716,
      "theoretical_loss": 3.4283929252453094,
      "tokens_seen": 2014052352
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001986830905995839,
      "loss": 2.524,
      "theoretical_loss": 3.4283836367994485,
      "tokens_seen": 2014117888
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001986665405712124,
      "loss": 2.5089,
      "theoretical_loss": 3.428374348740434,
      "tokens_seen": 2014183424
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019864999054284094,
      "loss": 2.7912,
      "theoretical_loss": 3.428365061068238,
      "tokens_seen": 2014248960
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019863344051446945,
      "loss": 2.7364,
      "theoretical_loss": 3.4283557737828314,
      "tokens_seen": 2014314496
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019861689048609798,
      "loss": 2.7563,
      "theoretical_loss": 3.4283464868841853,
      "tokens_seen": 2014380032
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001986003404577265,
      "loss": 2.5252,
      "theoretical_loss": 3.428337200372271,
      "tokens_seen": 2014445568
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019858379042935503,
      "loss": 2.4577,
      "theoretical_loss": 3.4283279142470606,
      "tokens_seen": 2014511104
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019856724040098354,
      "loss": 2.5587,
      "theoretical_loss": 3.4283186285085243,
      "tokens_seen": 2014576640
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019855069037261208,
      "loss": 2.4006,
      "theoretical_loss": 3.4283093431566343,
      "tokens_seen": 2014642176
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001985341403442406,
      "loss": 2.7216,
      "theoretical_loss": 3.428300058191362,
      "tokens_seen": 2014707712
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019851759031586913,
      "loss": 2.9643,
      "theoretical_loss": 3.428290773612678,
      "tokens_seen": 2014773248
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019850104028749764,
      "loss": 2.6309,
      "theoretical_loss": 3.428281489420554,
      "tokens_seen": 2014838784
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019848449025912618,
      "loss": 2.6771,
      "theoretical_loss": 3.4282722056149613,
      "tokens_seen": 2014904320
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019846794023075469,
      "loss": 2.5164,
      "theoretical_loss": 3.428262922195871,
      "tokens_seen": 2014969856
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019845139020238322,
      "loss": 2.4514,
      "theoretical_loss": 3.428253639163255,
      "tokens_seen": 2015035392
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019843484017401173,
      "loss": 2.5317,
      "theoretical_loss": 3.428244356517084,
      "tokens_seen": 2015100928
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019841829014564027,
      "loss": 2.6805,
      "theoretical_loss": 3.42823507425733,
      "tokens_seen": 2015166464
    },
    {
      "epoch": 0.72,
      "objective/train/avg_token_score": 0.014094609767198563,
      "objective/train/avg_weight": 0.9834184646606445,
      "objective/train/docs_used": 1138919,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6727492809295654,
      "objective/train/std_weight": 0.07053889334201813,
      "objective/train/theoretical_loss": 3.428225792383964,
      "objective/train/tokens_used": 2035692000,
      "theoretical_loss": 3.428225792383964,
      "tokens_seen": 2015232000
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019840174011726878,
      "loss": 2.7582,
      "theoretical_loss": 3.428225792383964,
      "tokens_seen": 2015232000
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019838519008889732,
      "loss": 2.6479,
      "theoretical_loss": 3.4282165108969576,
      "tokens_seen": 2015297536
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019836864006052583,
      "loss": 2.617,
      "theoretical_loss": 3.428207229796281,
      "tokens_seen": 2015363072
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019835209003215437,
      "loss": 2.4452,
      "theoretical_loss": 3.4281979490819072,
      "tokens_seen": 2015428608
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019833554000378285,
      "loss": 2.5942,
      "theoretical_loss": 3.4281886687538066,
      "tokens_seen": 2015494144
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019831898997541136,
      "loss": 2.7789,
      "theoretical_loss": 3.428179388811951,
      "tokens_seen": 2015559680
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001983024399470399,
      "loss": 2.7801,
      "theoretical_loss": 3.4281701092563117,
      "tokens_seen": 2015625216
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001982858899186684,
      "loss": 2.6887,
      "theoretical_loss": 3.42816083008686,
      "tokens_seen": 2015690752
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019826933989029695,
      "loss": 2.7697,
      "theoretical_loss": 3.428151551303567,
      "tokens_seen": 2015756288
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019825278986192546,
      "loss": 2.6521,
      "theoretical_loss": 3.428142272906404,
      "tokens_seen": 2015821824
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000198236239833554,
      "loss": 2.5221,
      "theoretical_loss": 3.4281329948953427,
      "tokens_seen": 2015887360
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001982196898051825,
      "loss": 2.8446,
      "theoretical_loss": 3.428123717270355,
      "tokens_seen": 2015952896
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019820313977681104,
      "loss": 2.0468,
      "theoretical_loss": 3.428114440031411,
      "tokens_seen": 2016018432
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019818658974843955,
      "loss": 2.7817,
      "theoretical_loss": 3.4281051631784836,
      "tokens_seen": 2016083968
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001981700397200681,
      "loss": 2.6898,
      "theoretical_loss": 3.428095886711543,
      "tokens_seen": 2016149504
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001981534896916966,
      "loss": 2.6076,
      "theoretical_loss": 3.4280866106305607,
      "tokens_seen": 2016215040
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019813693966332514,
      "loss": 2.5786,
      "theoretical_loss": 3.4280773349355087,
      "tokens_seen": 2016280576
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019812038963495365,
      "loss": 2.4962,
      "theoretical_loss": 3.4280680596263586,
      "tokens_seen": 2016346112
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001981038396065822,
      "loss": 2.6689,
      "theoretical_loss": 3.4280587847030812,
      "tokens_seen": 2016411648
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001980872895782107,
      "loss": 3.1391,
      "theoretical_loss": 3.428049510165647,
      "tokens_seen": 2016477184
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019807073954983923,
      "loss": 2.5835,
      "theoretical_loss": 3.4280402360140294,
      "tokens_seen": 2016542720
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019805418952146774,
      "loss": 2.7381,
      "theoretical_loss": 3.4280309622481986,
      "tokens_seen": 2016608256
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019803763949309628,
      "loss": 2.4964,
      "theoretical_loss": 3.4280216888681263,
      "tokens_seen": 2016673792
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001980210894647248,
      "loss": 2.3993,
      "theoretical_loss": 3.4280124158737837,
      "tokens_seen": 2016739328
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019800453943635333,
      "loss": 2.8098,
      "theoretical_loss": 3.4280031432651423,
      "tokens_seen": 2016804864
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019798798940798184,
      "loss": 2.8264,
      "theoretical_loss": 3.4279938710421742,
      "tokens_seen": 2016870400
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019797143937961038,
      "loss": 2.7875,
      "theoretical_loss": 3.4279845992048497,
      "tokens_seen": 2016935936
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001979548893512389,
      "loss": 2.7871,
      "theoretical_loss": 3.427975327753141,
      "tokens_seen": 2017001472
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019793833932286743,
      "loss": 2.8732,
      "theoretical_loss": 3.427966056687019,
      "tokens_seen": 2017067008
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019792178929449594,
      "loss": 2.6266,
      "theoretical_loss": 3.427956786006456,
      "tokens_seen": 2017132544
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019790523926612447,
      "loss": 2.4486,
      "theoretical_loss": 3.4279475157114225,
      "tokens_seen": 2017198080
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019788868923775298,
      "loss": 2.4651,
      "theoretical_loss": 3.42793824580189,
      "tokens_seen": 2017263616
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019787213920938152,
      "loss": 2.6512,
      "theoretical_loss": 3.427928976277831,
      "tokens_seen": 2017329152
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019785558918101,
      "loss": 2.4842,
      "theoretical_loss": 3.427919707139216,
      "tokens_seen": 2017394688
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019783903915263852,
      "loss": 2.6957,
      "theoretical_loss": 3.4279104383860166,
      "tokens_seen": 2017460224
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019782248912426705,
      "loss": 2.6073,
      "theoretical_loss": 3.4279011700182043,
      "tokens_seen": 2017525760
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019780593909589556,
      "loss": 2.6653,
      "theoretical_loss": 3.427891902035751,
      "tokens_seen": 2017591296
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001977893890675241,
      "loss": 2.636,
      "theoretical_loss": 3.427882634438627,
      "tokens_seen": 2017656832
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001977728390391526,
      "loss": 2.6632,
      "theoretical_loss": 3.4278733672268045,
      "tokens_seen": 2017722368
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019775628901078115,
      "loss": 2.7983,
      "theoretical_loss": 3.4278641004002557,
      "tokens_seen": 2017787904
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019773973898240966,
      "loss": 2.6908,
      "theoretical_loss": 3.4278548339589507,
      "tokens_seen": 2017853440
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001977231889540382,
      "loss": 2.4951,
      "theoretical_loss": 3.427845567902862,
      "tokens_seen": 2017918976
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001977066389256667,
      "loss": 2.5604,
      "theoretical_loss": 3.427836302231961,
      "tokens_seen": 2017984512
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019769008889729525,
      "loss": 2.4282,
      "theoretical_loss": 3.4278270369462183,
      "tokens_seen": 2018050048
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019767353886892376,
      "loss": 2.541,
      "theoretical_loss": 3.4278177720456062,
      "tokens_seen": 2018115584
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001976569888405523,
      "loss": 2.632,
      "theoretical_loss": 3.427808507530096,
      "tokens_seen": 2018181120
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001976404388121808,
      "loss": 2.7741,
      "theoretical_loss": 3.427799243399659,
      "tokens_seen": 2018246656
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019762388878380934,
      "loss": 2.549,
      "theoretical_loss": 3.427789979654267,
      "tokens_seen": 2018312192
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019760733875543785,
      "loss": 2.6722,
      "theoretical_loss": 3.427780716293891,
      "tokens_seen": 2018377728
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001975907887270664,
      "loss": 2.6255,
      "theoretical_loss": 3.427771453318503,
      "tokens_seen": 2018443264
    },
    {
      "epoch": 0.72,
      "objective/train/avg_token_score": 0.01718321070075035,
      "objective/train/avg_weight": 0.9797744154930115,
      "objective/train/docs_used": 1140999,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5398178100585938,
      "objective/train/std_weight": 0.08639667928218842,
      "objective/train/theoretical_loss": 3.4277621907280746,
      "objective/train/tokens_used": 2038968800,
      "theoretical_loss": 3.4277621907280746,
      "tokens_seen": 2018508800
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001975742386986949,
      "loss": 2.6171,
      "theoretical_loss": 3.4277621907280746,
      "tokens_seen": 2018508800
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019755768867032344,
      "loss": 2.6519,
      "theoretical_loss": 3.4277529285225765,
      "tokens_seen": 2018574336
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019754113864195195,
      "loss": 2.4296,
      "theoretical_loss": 3.427743666701981,
      "tokens_seen": 2018639872
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019752458861358049,
      "loss": 2.6699,
      "theoretical_loss": 3.4277344052662597,
      "tokens_seen": 2018705408
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000197508038585209,
      "loss": 2.4624,
      "theoretical_loss": 3.4277251442153833,
      "tokens_seen": 2018770944
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019749148855683753,
      "loss": 2.5254,
      "theoretical_loss": 3.4277158835493236,
      "tokens_seen": 2018836480
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019747493852846607,
      "loss": 2.7103,
      "theoretical_loss": 3.427706623268053,
      "tokens_seen": 2018902016
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019745838850009458,
      "loss": 2.6579,
      "theoretical_loss": 3.4276973633715415,
      "tokens_seen": 2018967552
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019744183847172312,
      "loss": 2.7847,
      "theoretical_loss": 3.4276881038597615,
      "tokens_seen": 2019033088
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019742528844335163,
      "loss": 2.4955,
      "theoretical_loss": 3.427678844732685,
      "tokens_seen": 2019098624
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019740873841498017,
      "loss": 2.5011,
      "theoretical_loss": 3.427669585990283,
      "tokens_seen": 2019164160
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019739218838660868,
      "loss": 2.4235,
      "theoretical_loss": 3.427660327632527,
      "tokens_seen": 2019229696
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019737563835823722,
      "loss": 2.669,
      "theoretical_loss": 3.4276510696593876,
      "tokens_seen": 2019295232
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001973590883298657,
      "loss": 2.7222,
      "theoretical_loss": 3.4276418120708385,
      "tokens_seen": 2019360768
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001973425383014942,
      "loss": 2.6043,
      "theoretical_loss": 3.4276325548668494,
      "tokens_seen": 2019426304
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019732598827312275,
      "loss": 2.7524,
      "theoretical_loss": 3.4276232980473926,
      "tokens_seen": 2019491840
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019730943824475126,
      "loss": 2.4814,
      "theoretical_loss": 3.42761404161244,
      "tokens_seen": 2019557376
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001972928882163798,
      "loss": 2.5194,
      "theoretical_loss": 3.4276047855619622,
      "tokens_seen": 2019622912
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001972763381880083,
      "loss": 2.5427,
      "theoretical_loss": 3.4275955298959313,
      "tokens_seen": 2019688448
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019725978815963684,
      "loss": 2.4312,
      "theoretical_loss": 3.4275862746143186,
      "tokens_seen": 2019753984
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019724323813126535,
      "loss": 2.7288,
      "theoretical_loss": 3.4275770197170963,
      "tokens_seen": 2019819520
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001972266881028939,
      "loss": 2.4832,
      "theoretical_loss": 3.427567765204236,
      "tokens_seen": 2019885056
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001972101380745224,
      "loss": 2.7034,
      "theoretical_loss": 3.4275585110757074,
      "tokens_seen": 2019950592
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019719358804615094,
      "loss": 2.6011,
      "theoretical_loss": 3.4275492573314845,
      "tokens_seen": 2020016128
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019717703801777945,
      "loss": 2.5297,
      "theoretical_loss": 3.4275400039715374,
      "tokens_seen": 2020081664
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019716048798940799,
      "loss": 2.4948,
      "theoretical_loss": 3.427530750995839,
      "tokens_seen": 2020147200
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001971439379610365,
      "loss": 2.6445,
      "theoretical_loss": 3.427521498404359,
      "tokens_seen": 2020212736
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019712738793266503,
      "loss": 2.6005,
      "theoretical_loss": 3.42751224619707,
      "tokens_seen": 2020278272
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019711083790429354,
      "loss": 2.4005,
      "theoretical_loss": 3.427502994373944,
      "tokens_seen": 2020343808
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019709428787592208,
      "loss": 2.5946,
      "theoretical_loss": 3.427493742934952,
      "tokens_seen": 2020409344
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001970777378475506,
      "loss": 2.4391,
      "theoretical_loss": 3.4274844918800658,
      "tokens_seen": 2020474880
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019706118781917913,
      "loss": 2.5536,
      "theoretical_loss": 3.4274752412092573,
      "tokens_seen": 2020540416
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019704463779080764,
      "loss": 1.755,
      "theoretical_loss": 3.427465990922497,
      "tokens_seen": 2020605952
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019702808776243618,
      "loss": 2.6464,
      "theoretical_loss": 3.427456741019758,
      "tokens_seen": 2020671488
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001970115377340647,
      "loss": 2.7495,
      "theoretical_loss": 3.4274474915010105,
      "tokens_seen": 2020737024
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019699498770569323,
      "loss": 2.6426,
      "theoretical_loss": 3.427438242366227,
      "tokens_seen": 2020802560
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019697843767732174,
      "loss": 2.5221,
      "theoretical_loss": 3.427428993615379,
      "tokens_seen": 2020868096
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019696188764895027,
      "loss": 2.3816,
      "theoretical_loss": 3.4274197452484376,
      "tokens_seen": 2020933632
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019694533762057878,
      "loss": 2.6595,
      "theoretical_loss": 3.4274104972653747,
      "tokens_seen": 2020999168
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019692878759220732,
      "loss": 2.7209,
      "theoretical_loss": 3.4274012496661626,
      "tokens_seen": 2021064704
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019691223756383583,
      "loss": 2.3152,
      "theoretical_loss": 3.4273920024507722,
      "tokens_seen": 2021130240
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019689568753546437,
      "loss": 2.479,
      "theoretical_loss": 3.427382755619175,
      "tokens_seen": 2021195776
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019687913750709288,
      "loss": 2.5727,
      "theoretical_loss": 3.427373509171343,
      "tokens_seen": 2021261312
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019686258747872136,
      "loss": 2.6574,
      "theoretical_loss": 3.427364263107248,
      "tokens_seen": 2021326848
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001968460374503499,
      "loss": 2.4459,
      "theoretical_loss": 3.4273550174268603,
      "tokens_seen": 2021392384
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001968294874219784,
      "loss": 2.4737,
      "theoretical_loss": 3.4273457721301535,
      "tokens_seen": 2021457920
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019681293739360695,
      "loss": 2.4254,
      "theoretical_loss": 3.4273365272170984,
      "tokens_seen": 2021523456
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019679638736523546,
      "loss": 2.6082,
      "theoretical_loss": 3.427327282687666,
      "tokens_seen": 2021588992
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000196779837336864,
      "loss": 2.459,
      "theoretical_loss": 3.427318038541829,
      "tokens_seen": 2021654528
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001967632873084925,
      "loss": 2.7457,
      "theoretical_loss": 3.4273087947795586,
      "tokens_seen": 2021720064
    },
    {
      "epoch": 0.72,
      "objective/train/avg_token_score": 0.03967645391821861,
      "objective/train/avg_weight": 0.9533393979072571,
      "objective/train/docs_used": 1143312,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.741224765777588,
      "objective/train/std_weight": 0.175728902220726,
      "objective/train/theoretical_loss": 3.427299551400826,
      "objective/train/tokens_used": 2042245600,
      "theoretical_loss": 3.427299551400826,
      "tokens_seen": 2021785600
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019674673728012105,
      "loss": 2.4015,
      "theoretical_loss": 3.427299551400826,
      "tokens_seen": 2021785600
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019673018725174956,
      "loss": 2.7192,
      "theoretical_loss": 3.4272903084056034,
      "tokens_seen": 2021851136
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001967136372233781,
      "loss": 2.7513,
      "theoretical_loss": 3.4272810657938626,
      "tokens_seen": 2021916672
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001966970871950066,
      "loss": 2.4856,
      "theoretical_loss": 3.427271823565574,
      "tokens_seen": 2021982208
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019668053716663514,
      "loss": 2.6315,
      "theoretical_loss": 3.4272625817207114,
      "tokens_seen": 2022047744
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019666398713826365,
      "loss": 2.4673,
      "theoretical_loss": 3.427253340259245,
      "tokens_seen": 2022113280
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001966474371098922,
      "loss": 2.6444,
      "theoretical_loss": 3.427244099181147,
      "tokens_seen": 2022178816
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001966308870815207,
      "loss": 2.5567,
      "theoretical_loss": 3.427234858486388,
      "tokens_seen": 2022244352
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019661433705314924,
      "loss": 2.6696,
      "theoretical_loss": 3.4272256181749414,
      "tokens_seen": 2022309888
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019659778702477775,
      "loss": 2.5447,
      "theoretical_loss": 3.4272163782467775,
      "tokens_seen": 2022375424
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019658123699640629,
      "loss": 2.6919,
      "theoretical_loss": 3.4272071387018688,
      "tokens_seen": 2022440960
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001965646869680348,
      "loss": 2.6921,
      "theoretical_loss": 3.427197899540187,
      "tokens_seen": 2022506496
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019654813693966333,
      "loss": 2.2968,
      "theoretical_loss": 3.4271886607617024,
      "tokens_seen": 2022572032
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019653158691129184,
      "loss": 2.552,
      "theoretical_loss": 3.4271794223663887,
      "tokens_seen": 2022637568
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019651503688292038,
      "loss": 2.4849,
      "theoretical_loss": 3.427170184354216,
      "tokens_seen": 2022703104
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001964984868545489,
      "loss": 2.5929,
      "theoretical_loss": 3.427160946725157,
      "tokens_seen": 2022768640
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019648193682617743,
      "loss": 2.5733,
      "theoretical_loss": 3.4271517094791832,
      "tokens_seen": 2022834176
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019646538679780594,
      "loss": 2.5815,
      "theoretical_loss": 3.427142472616266,
      "tokens_seen": 2022899712
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019644883676943448,
      "loss": 2.6938,
      "theoretical_loss": 3.4271332361363775,
      "tokens_seen": 2022965248
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000196432286741063,
      "loss": 2.7064,
      "theoretical_loss": 3.4271240000394885,
      "tokens_seen": 2023030784
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019641573671269153,
      "loss": 2.7447,
      "theoretical_loss": 3.427114764325572,
      "tokens_seen": 2023096320
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019639918668432004,
      "loss": 2.6193,
      "theoretical_loss": 3.427105528994599,
      "tokens_seen": 2023161856
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019638263665594852,
      "loss": 2.7974,
      "theoretical_loss": 3.427096294046541,
      "tokens_seen": 2023227392
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019636608662757706,
      "loss": 2.7015,
      "theoretical_loss": 3.4270870594813703,
      "tokens_seen": 2023292928
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019634953659920557,
      "loss": 2.4546,
      "theoretical_loss": 3.427077825299058,
      "tokens_seen": 2023358464
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001963329865708341,
      "loss": 2.5302,
      "theoretical_loss": 3.4270685914995767,
      "tokens_seen": 2023424000
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019631643654246261,
      "loss": 2.5897,
      "theoretical_loss": 3.4270593580828974,
      "tokens_seen": 2023489536
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019629988651409115,
      "loss": 2.3273,
      "theoretical_loss": 3.427050125048992,
      "tokens_seen": 2023555072
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019628333648571966,
      "loss": 2.6416,
      "theoretical_loss": 3.4270408923978324,
      "tokens_seen": 2023620608
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001962667864573482,
      "loss": 2.5864,
      "theoretical_loss": 3.42703166012939,
      "tokens_seen": 2023686144
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019625023642897674,
      "loss": 2.5643,
      "theoretical_loss": 3.427022428243637,
      "tokens_seen": 2023751680
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019623368640060525,
      "loss": 2.7068,
      "theoretical_loss": 3.427013196740545,
      "tokens_seen": 2023817216
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019621713637223379,
      "loss": 2.6985,
      "theoretical_loss": 3.4270039656200852,
      "tokens_seen": 2023882752
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001962005863438623,
      "loss": 2.6336,
      "theoretical_loss": 3.4269947348822303,
      "tokens_seen": 2023948288
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019618403631549083,
      "loss": 2.849,
      "theoretical_loss": 3.426985504526951,
      "tokens_seen": 2024013824
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019616748628711934,
      "loss": 2.492,
      "theoretical_loss": 3.42697627455422,
      "tokens_seen": 2024079360
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019615093625874788,
      "loss": 2.4476,
      "theoretical_loss": 3.426967044964009,
      "tokens_seen": 2024144896
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001961343862303764,
      "loss": 2.4552,
      "theoretical_loss": 3.426957815756289,
      "tokens_seen": 2024210432
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019611783620200493,
      "loss": 2.4763,
      "theoretical_loss": 3.426948586931032,
      "tokens_seen": 2024275968
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019610128617363344,
      "loss": 2.5889,
      "theoretical_loss": 3.4269393584882106,
      "tokens_seen": 2024341504
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019608473614526198,
      "loss": 2.2352,
      "theoretical_loss": 3.4269301304277953,
      "tokens_seen": 2024407040
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001960681861168905,
      "loss": 2.5291,
      "theoretical_loss": 3.4269209027497594,
      "tokens_seen": 2024472576
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019605163608851903,
      "loss": 2.6464,
      "theoretical_loss": 3.4269116754540727,
      "tokens_seen": 2024538112
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019603508606014754,
      "loss": 2.4263,
      "theoretical_loss": 3.4269024485407087,
      "tokens_seen": 2024603648
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019601853603177607,
      "loss": 2.6673,
      "theoretical_loss": 3.426893222009639,
      "tokens_seen": 2024669184
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019600198600340458,
      "loss": 2.59,
      "theoretical_loss": 3.426883995860834,
      "tokens_seen": 2024734720
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019598543597503312,
      "loss": 2.7885,
      "theoretical_loss": 3.4268747700942668,
      "tokens_seen": 2024800256
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019596888594666163,
      "loss": 2.7773,
      "theoretical_loss": 3.426865544709909,
      "tokens_seen": 2024865792
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019595233591829017,
      "loss": 2.5506,
      "theoretical_loss": 3.426856319707732,
      "tokens_seen": 2024931328
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019593578588991868,
      "loss": 2.7152,
      "theoretical_loss": 3.426847095087708,
      "tokens_seen": 2024996864
    },
    {
      "epoch": 0.72,
      "objective/train/avg_token_score": 0.008526789024472237,
      "objective/train/avg_weight": 0.9899924993515015,
      "objective/train/docs_used": 1144755,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.869861364364624,
      "objective/train/std_weight": 0.04428090155124664,
      "objective/train/theoretical_loss": 3.4268378708498086,
      "objective/train/tokens_used": 2045522400,
      "theoretical_loss": 3.4268378708498086,
      "tokens_seen": 2025062400
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019591923586154722,
      "loss": 2.6416,
      "theoretical_loss": 3.4268378708498086,
      "tokens_seen": 2025062400
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019590268583317573,
      "loss": 2.7027,
      "theoretical_loss": 3.426828646994006,
      "tokens_seen": 2025127936
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001958861358048042,
      "loss": 2.7592,
      "theoretical_loss": 3.4268194235202714,
      "tokens_seen": 2025193472
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019586958577643275,
      "loss": 2.3763,
      "theoretical_loss": 3.4268102004285765,
      "tokens_seen": 2025259008
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019585303574806126,
      "loss": 2.6497,
      "theoretical_loss": 3.4268009777188935,
      "tokens_seen": 2025324544
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001958364857196898,
      "loss": 2.477,
      "theoretical_loss": 3.4267917553911946,
      "tokens_seen": 2025390080
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001958199356913183,
      "loss": 2.5994,
      "theoretical_loss": 3.426782533445451,
      "tokens_seen": 2025455616
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019580338566294685,
      "loss": 2.9824,
      "theoretical_loss": 3.426773311881635,
      "tokens_seen": 2025521152
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019578683563457536,
      "loss": 2.5467,
      "theoretical_loss": 3.426764090699718,
      "tokens_seen": 2025586688
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001957702856062039,
      "loss": 2.3107,
      "theoretical_loss": 3.4267548698996717,
      "tokens_seen": 2025652224
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001957537355778324,
      "loss": 2.5232,
      "theoretical_loss": 3.4267456494814685,
      "tokens_seen": 2025717760
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019573718554946094,
      "loss": 2.3791,
      "theoretical_loss": 3.4267364294450795,
      "tokens_seen": 2025783296
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019572063552108945,
      "loss": 2.4259,
      "theoretical_loss": 3.426727209790478,
      "tokens_seen": 2025848832
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000195704085492718,
      "loss": 2.5066,
      "theoretical_loss": 3.426717990517634,
      "tokens_seen": 2025914368
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001956875354643465,
      "loss": 2.4834,
      "theoretical_loss": 3.4267087716265205,
      "tokens_seen": 2025979904
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019567098543597504,
      "loss": 2.605,
      "theoretical_loss": 3.4266995531171087,
      "tokens_seen": 2026045440
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019565443540760355,
      "loss": 2.7041,
      "theoretical_loss": 3.426690334989371,
      "tokens_seen": 2026110976
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019563788537923209,
      "loss": 2.5366,
      "theoretical_loss": 3.426681117243279,
      "tokens_seen": 2026176512
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001956213353508606,
      "loss": 2.6912,
      "theoretical_loss": 3.426671899878804,
      "tokens_seen": 2026242048
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019560478532248913,
      "loss": 2.6178,
      "theoretical_loss": 3.4266626828959197,
      "tokens_seen": 2026307584
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019558823529411764,
      "loss": 2.4706,
      "theoretical_loss": 3.4266534662945958,
      "tokens_seen": 2026373120
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019557168526574618,
      "loss": 2.7665,
      "theoretical_loss": 3.4266442500748053,
      "tokens_seen": 2026438656
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001955551352373747,
      "loss": 2.5588,
      "theoretical_loss": 3.42663503423652,
      "tokens_seen": 2026504192
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019553858520900323,
      "loss": 2.3118,
      "theoretical_loss": 3.426625818779711,
      "tokens_seen": 2026569728
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019552203518063174,
      "loss": 2.7348,
      "theoretical_loss": 3.4266166037043515,
      "tokens_seen": 2026635264
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019550548515226028,
      "loss": 2.6222,
      "theoretical_loss": 3.4266073890104125,
      "tokens_seen": 2026700800
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001954889351238888,
      "loss": 2.8655,
      "theoretical_loss": 3.426598174697866,
      "tokens_seen": 2026766336
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019547238509551732,
      "loss": 2.4744,
      "theoretical_loss": 3.4265889607666837,
      "tokens_seen": 2026831872
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019545583506714584,
      "loss": 2.7628,
      "theoretical_loss": 3.426579747216838,
      "tokens_seen": 2026897408
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019543928503877437,
      "loss": 2.594,
      "theoretical_loss": 3.4265705340483,
      "tokens_seen": 2026962944
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019542273501040288,
      "loss": 2.78,
      "theoretical_loss": 3.426561321261042,
      "tokens_seen": 2027028480
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019540618498203142,
      "loss": 2.7854,
      "theoretical_loss": 3.426552108855037,
      "tokens_seen": 2027094016
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001953896349536599,
      "loss": 2.3964,
      "theoretical_loss": 3.426542896830255,
      "tokens_seen": 2027159552
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019537308492528841,
      "loss": 2.6585,
      "theoretical_loss": 3.4265336851866692,
      "tokens_seen": 2027225088
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019535653489691695,
      "loss": 2.6392,
      "theoretical_loss": 3.4265244739242506,
      "tokens_seen": 2027290624
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019533998486854546,
      "loss": 2.5343,
      "theoretical_loss": 3.4265152630429725,
      "tokens_seen": 2027356160
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000195323434840174,
      "loss": 2.5497,
      "theoretical_loss": 3.426506052542805,
      "tokens_seen": 2027421696
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001953068848118025,
      "loss": 2.6641,
      "theoretical_loss": 3.426496842423721,
      "tokens_seen": 2027487232
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019529033478343105,
      "loss": 2.5875,
      "theoretical_loss": 3.4264876326856926,
      "tokens_seen": 2027552768
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019527378475505956,
      "loss": 2.4241,
      "theoretical_loss": 3.4264784233286916,
      "tokens_seen": 2027618304
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001952572347266881,
      "loss": 2.2877,
      "theoretical_loss": 3.4264692143526894,
      "tokens_seen": 2027683840
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001952406846983166,
      "loss": 2.7297,
      "theoretical_loss": 3.4264600057576584,
      "tokens_seen": 2027749376
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019522413466994514,
      "loss": 2.5794,
      "theoretical_loss": 3.4264507975435707,
      "tokens_seen": 2027814912
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019520758464157365,
      "loss": 2.3084,
      "theoretical_loss": 3.4264415897103975,
      "tokens_seen": 2027880448
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001951910346132022,
      "loss": 2.6979,
      "theoretical_loss": 3.4264323822581115,
      "tokens_seen": 2027945984
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001951744845848307,
      "loss": 2.7526,
      "theoretical_loss": 3.426423175186684,
      "tokens_seen": 2028011520
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019515793455645924,
      "loss": 2.6912,
      "theoretical_loss": 3.4264139684960875,
      "tokens_seen": 2028077056
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019514138452808775,
      "loss": 2.5522,
      "theoretical_loss": 3.4264047621862934,
      "tokens_seen": 2028142592
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001951248344997163,
      "loss": 2.5269,
      "theoretical_loss": 3.4263955562572743,
      "tokens_seen": 2028208128
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001951082844713448,
      "loss": 2.6709,
      "theoretical_loss": 3.4263863507090018,
      "tokens_seen": 2028273664
    },
    {
      "epoch": 0.72,
      "objective/train/avg_token_score": 0.00882747583091259,
      "objective/train/avg_weight": 0.989622950553894,
      "objective/train/docs_used": 1146855,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.813985824584961,
      "objective/train/std_weight": 0.0664902850985527,
      "objective/train/theoretical_loss": 3.426377145541448,
      "objective/train/tokens_used": 2048799200,
      "theoretical_loss": 3.426377145541448,
      "tokens_seen": 2028339200
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019509173444297334,
      "loss": 2.4991,
      "theoretical_loss": 3.426377145541448,
      "tokens_seen": 2028339200
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019507518441460185,
      "loss": 2.6881,
      "theoretical_loss": 3.426367940754584,
      "tokens_seen": 2028404736
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019505863438623038,
      "loss": 2.6801,
      "theoretical_loss": 3.4263587363483827,
      "tokens_seen": 2028470272
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001950420843578589,
      "loss": 2.4947,
      "theoretical_loss": 3.426349532322816,
      "tokens_seen": 2028535808
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019502553432948743,
      "loss": 2.5875,
      "theoretical_loss": 3.4263403286778558,
      "tokens_seen": 2028601344
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019500898430111594,
      "loss": 2.5106,
      "theoretical_loss": 3.4263311254134736,
      "tokens_seen": 2028666880
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019499243427274448,
      "loss": 2.7497,
      "theoretical_loss": 3.4263219225296417,
      "tokens_seen": 2028732416
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000194975884244373,
      "loss": 2.4348,
      "theoretical_loss": 3.4263127200263326,
      "tokens_seen": 2028797952
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019495933421600153,
      "loss": 2.475,
      "theoretical_loss": 3.4263035179035173,
      "tokens_seen": 2028863488
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019494278418763004,
      "loss": 2.4241,
      "theoretical_loss": 3.4262943161611688,
      "tokens_seen": 2028929024
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019492623415925858,
      "loss": 2.6154,
      "theoretical_loss": 3.4262851147992577,
      "tokens_seen": 2028994560
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001949096841308871,
      "loss": 2.3877,
      "theoretical_loss": 3.426275913817757,
      "tokens_seen": 2029060096
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001948931341025156,
      "loss": 2.7521,
      "theoretical_loss": 3.4262667132166387,
      "tokens_seen": 2029125632
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001948765840741441,
      "loss": 2.8061,
      "theoretical_loss": 3.4262575129958748,
      "tokens_seen": 2029191168
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019486003404577264,
      "loss": 2.625,
      "theoretical_loss": 3.426248313155437,
      "tokens_seen": 2029256704
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019484348401740116,
      "loss": 2.476,
      "theoretical_loss": 3.426239113695297,
      "tokens_seen": 2029322240
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001948269339890297,
      "loss": 2.6172,
      "theoretical_loss": 3.426229914615427,
      "tokens_seen": 2029387776
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001948103839606582,
      "loss": 2.5454,
      "theoretical_loss": 3.4262207159158,
      "tokens_seen": 2029453312
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019479383393228674,
      "loss": 2.5834,
      "theoretical_loss": 3.4262115175963865,
      "tokens_seen": 2029518848
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019477728390391525,
      "loss": 2.5939,
      "theoretical_loss": 3.4262023196571594,
      "tokens_seen": 2029584384
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001947607338755438,
      "loss": 2.3833,
      "theoretical_loss": 3.4261931220980903,
      "tokens_seen": 2029649920
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001947441838471723,
      "loss": 2.5481,
      "theoretical_loss": 3.426183924919152,
      "tokens_seen": 2029715456
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019472763381880084,
      "loss": 2.5996,
      "theoretical_loss": 3.4261747281203148,
      "tokens_seen": 2029780992
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019471108379042935,
      "loss": 2.4518,
      "theoretical_loss": 3.4261655317015527,
      "tokens_seen": 2029846528
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00019469453376205788,
      "loss": 2.0985,
      "theoretical_loss": 3.426156335662837,
      "tokens_seen": 2029912064
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001946779837336864,
      "loss": 2.6455,
      "theoretical_loss": 3.426147140004139,
      "tokens_seen": 2029977600
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019466143370531493,
      "loss": 2.5109,
      "theoretical_loss": 3.4261379447254314,
      "tokens_seen": 2030043136
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019464488367694344,
      "loss": 2.6048,
      "theoretical_loss": 3.4261287498266864,
      "tokens_seen": 2030108672
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019462833364857198,
      "loss": 2.4864,
      "theoretical_loss": 3.4261195553078756,
      "tokens_seen": 2030174208
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001946117836202005,
      "loss": 2.5218,
      "theoretical_loss": 3.4261103611689716,
      "tokens_seen": 2030239744
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019459523359182903,
      "loss": 2.5323,
      "theoretical_loss": 3.4261011674099455,
      "tokens_seen": 2030305280
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019457868356345754,
      "loss": 2.6043,
      "theoretical_loss": 3.4260919740307703,
      "tokens_seen": 2030370816
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019456213353508608,
      "loss": 2.7026,
      "theoretical_loss": 3.4260827810314174,
      "tokens_seen": 2030436352
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001945455835067146,
      "loss": 2.565,
      "theoretical_loss": 3.426073588411859,
      "tokens_seen": 2030501888
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019452903347834312,
      "loss": 2.5405,
      "theoretical_loss": 3.426064396172068,
      "tokens_seen": 2030567424
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019451248344997164,
      "loss": 2.34,
      "theoretical_loss": 3.4260552043120147,
      "tokens_seen": 2030632960
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019449593342160017,
      "loss": 2.557,
      "theoretical_loss": 3.4260460128316725,
      "tokens_seen": 2030698496
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019447938339322868,
      "loss": 2.6052,
      "theoretical_loss": 3.4260368217310133,
      "tokens_seen": 2030764032
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019446283336485722,
      "loss": 2.6984,
      "theoretical_loss": 3.426027631010009,
      "tokens_seen": 2030829568
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019444628333648573,
      "loss": 2.49,
      "theoretical_loss": 3.4260184406686314,
      "tokens_seen": 2030895104
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019442973330811427,
      "loss": 2.5431,
      "theoretical_loss": 3.426009250706853,
      "tokens_seen": 2030960640
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019441318327974275,
      "loss": 2.3504,
      "theoretical_loss": 3.4260000611246455,
      "tokens_seen": 2031026176
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019439663325137126,
      "loss": 2.4964,
      "theoretical_loss": 3.4259908719219814,
      "tokens_seen": 2031091712
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001943800832229998,
      "loss": 2.5138,
      "theoretical_loss": 3.4259816830988328,
      "tokens_seen": 2031157248
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001943635331946283,
      "loss": 2.4546,
      "theoretical_loss": 3.425972494655171,
      "tokens_seen": 2031222784
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019434698316625685,
      "loss": 2.6931,
      "theoretical_loss": 3.425963306590969,
      "tokens_seen": 2031288320
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019433043313788536,
      "loss": 2.5681,
      "theoretical_loss": 3.4259541189061977,
      "tokens_seen": 2031353856
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001943138831095139,
      "loss": 2.1733,
      "theoretical_loss": 3.4259449316008306,
      "tokens_seen": 2031419392
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001942973330811424,
      "loss": 2.4963,
      "theoretical_loss": 3.4259357446748395,
      "tokens_seen": 2031484928
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019428078305277094,
      "loss": 2.5775,
      "theoretical_loss": 3.4259265581281957,
      "tokens_seen": 2031550464
    },
    {
      "epoch": 0.73,
      "objective/train/avg_token_score": 0.017058763653039932,
      "objective/train/avg_weight": 0.979920506477356,
      "objective/train/docs_used": 1149039,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7386415004730225,
      "objective/train/std_weight": 0.08685068786144257,
      "objective/train/theoretical_loss": 3.425917371960872,
      "objective/train/tokens_used": 2052076000,
      "theoretical_loss": 3.425917371960872,
      "tokens_seen": 2031616000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019426423302439945,
      "loss": 2.4036,
      "theoretical_loss": 3.425917371960872,
      "tokens_seen": 2031616000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000194247682996028,
      "loss": 2.435,
      "theoretical_loss": 3.4259081861728404,
      "tokens_seen": 2031681536
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001942311329676565,
      "loss": 2.5307,
      "theoretical_loss": 3.4258990007640726,
      "tokens_seen": 2031747072
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019421458293928504,
      "loss": 2.604,
      "theoretical_loss": 3.425889815734541,
      "tokens_seen": 2031812608
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019419803291091355,
      "loss": 2.5709,
      "theoretical_loss": 3.425880631084218,
      "tokens_seen": 2031878144
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001941814828825421,
      "loss": 2.4086,
      "theoretical_loss": 3.4258714468130753,
      "tokens_seen": 2031943680
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001941649328541706,
      "loss": 2.43,
      "theoretical_loss": 3.425862262921085,
      "tokens_seen": 2032009216
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019414838282579914,
      "loss": 2.5216,
      "theoretical_loss": 3.425853079408219,
      "tokens_seen": 2032074752
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019413183279742765,
      "loss": 2.6437,
      "theoretical_loss": 3.4258438962744506,
      "tokens_seen": 2032140288
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019411528276905618,
      "loss": 2.4447,
      "theoretical_loss": 3.4258347135197504,
      "tokens_seen": 2032205824
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001940987327406847,
      "loss": 2.4511,
      "theoretical_loss": 3.425825531144092,
      "tokens_seen": 2032271360
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019408218271231323,
      "loss": 2.6513,
      "theoretical_loss": 3.4258163491474463,
      "tokens_seen": 2032336896
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019406563268394174,
      "loss": 2.5307,
      "theoretical_loss": 3.4258071675297854,
      "tokens_seen": 2032402432
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019404908265557028,
      "loss": 2.3741,
      "theoretical_loss": 3.425797986291083,
      "tokens_seen": 2032467968
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001940325326271988,
      "loss": 2.6175,
      "theoretical_loss": 3.425788805431309,
      "tokens_seen": 2032533504
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019401598259882733,
      "loss": 2.498,
      "theoretical_loss": 3.4257796249504375,
      "tokens_seen": 2032599040
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019399943257045584,
      "loss": 2.4561,
      "theoretical_loss": 3.42577044484844,
      "tokens_seen": 2032664576
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019398288254208438,
      "loss": 2.4981,
      "theoretical_loss": 3.425761265125288,
      "tokens_seen": 2032730112
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019396633251371289,
      "loss": 2.8516,
      "theoretical_loss": 3.4257520857809545,
      "tokens_seen": 2032795648
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019394978248534142,
      "loss": 2.8851,
      "theoretical_loss": 3.4257429068154113,
      "tokens_seen": 2032861184
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019393323245696993,
      "loss": 2.4061,
      "theoretical_loss": 3.4257337282286304,
      "tokens_seen": 2032926720
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019391668242859842,
      "loss": 2.5403,
      "theoretical_loss": 3.4257245500205844,
      "tokens_seen": 2032992256
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019390013240022696,
      "loss": 2.5077,
      "theoretical_loss": 3.4257153721912452,
      "tokens_seen": 2033057792
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019388358237185547,
      "loss": 2.3564,
      "theoretical_loss": 3.4257061947405845,
      "tokens_seen": 2033123328
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000193867032343484,
      "loss": 2.6186,
      "theoretical_loss": 3.4256970176685755,
      "tokens_seen": 2033188864
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001938504823151125,
      "loss": 2.3638,
      "theoretical_loss": 3.42568784097519,
      "tokens_seen": 2033254400
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019383393228674105,
      "loss": 2.4693,
      "theoretical_loss": 3.425678664660399,
      "tokens_seen": 2033319936
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019381738225836956,
      "loss": 2.5723,
      "theoretical_loss": 3.4256694887241768,
      "tokens_seen": 2033385472
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001938008322299981,
      "loss": 2.6375,
      "theoretical_loss": 3.4256603131664933,
      "tokens_seen": 2033451008
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001937842822016266,
      "loss": 2.4024,
      "theoretical_loss": 3.4256511379873227,
      "tokens_seen": 2033516544
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019376773217325515,
      "loss": 2.4537,
      "theoretical_loss": 3.425641963186636,
      "tokens_seen": 2033582080
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019375118214488366,
      "loss": 2.3336,
      "theoretical_loss": 3.425632788764406,
      "tokens_seen": 2033647616
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001937346321165122,
      "loss": 2.3893,
      "theoretical_loss": 3.425623614720604,
      "tokens_seen": 2033713152
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001937180820881407,
      "loss": 2.5021,
      "theoretical_loss": 3.4256144410552034,
      "tokens_seen": 2033778688
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019370153205976924,
      "loss": 2.4515,
      "theoretical_loss": 3.4256052677681756,
      "tokens_seen": 2033844224
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019368498203139775,
      "loss": 2.3877,
      "theoretical_loss": 3.4255960948594932,
      "tokens_seen": 2033909760
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001936684320030263,
      "loss": 2.3305,
      "theoretical_loss": 3.425586922329128,
      "tokens_seen": 2033975296
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001936518819746548,
      "loss": 2.6966,
      "theoretical_loss": 3.4255777501770526,
      "tokens_seen": 2034040832
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019363533194628334,
      "loss": 2.7733,
      "theoretical_loss": 3.4255685784032393,
      "tokens_seen": 2034106368
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019361878191791185,
      "loss": 2.5563,
      "theoretical_loss": 3.4255594070076594,
      "tokens_seen": 2034171904
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001936022318895404,
      "loss": 2.5934,
      "theoretical_loss": 3.4255502359902863,
      "tokens_seen": 2034237440
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001935856818611689,
      "loss": 2.4133,
      "theoretical_loss": 3.4255410653510916,
      "tokens_seen": 2034302976
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019356913183279743,
      "loss": 2.605,
      "theoretical_loss": 3.425531895090047,
      "tokens_seen": 2034368512
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019355258180442595,
      "loss": 2.622,
      "theoretical_loss": 3.425522725207126,
      "tokens_seen": 2034434048
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019353603177605448,
      "loss": 2.4787,
      "theoretical_loss": 3.4255135557022998,
      "tokens_seen": 2034499584
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000193519481747683,
      "loss": 2.6183,
      "theoretical_loss": 3.4255043865755415,
      "tokens_seen": 2034565120
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019350293171931153,
      "loss": 2.5218,
      "theoretical_loss": 3.425495217826822,
      "tokens_seen": 2034630656
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019348638169094007,
      "loss": 2.4547,
      "theoretical_loss": 3.425486049456115,
      "tokens_seen": 2034696192
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019346983166256858,
      "loss": 2.3388,
      "theoretical_loss": 3.425476881463392,
      "tokens_seen": 2034761728
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019345328163419712,
      "loss": 2.4305,
      "theoretical_loss": 3.425467713848625,
      "tokens_seen": 2034827264
    },
    {
      "epoch": 0.73,
      "objective/train/avg_token_score": 0.010520203039050102,
      "objective/train/avg_weight": 0.9876150488853455,
      "objective/train/docs_used": 1151240,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.586012363433838,
      "objective/train/std_weight": 0.06674729287624359,
      "objective/train/theoretical_loss": 3.425458546611787,
      "objective/train/tokens_used": 2055352800,
      "theoretical_loss": 3.425458546611787,
      "tokens_seen": 2034892800
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019343673160582563,
      "loss": 2.6466,
      "theoretical_loss": 3.425458546611787,
      "tokens_seen": 2034892800
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001934201815774541,
      "loss": 2.4585,
      "theoretical_loss": 3.42544937975285,
      "tokens_seen": 2034958336
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019340363154908265,
      "loss": 2.436,
      "theoretical_loss": 3.4254402132717856,
      "tokens_seen": 2035023872
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019338708152071116,
      "loss": 2.7308,
      "theoretical_loss": 3.4254310471685665,
      "tokens_seen": 2035089408
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001933705314923397,
      "loss": 2.1046,
      "theoretical_loss": 3.4254218814431656,
      "tokens_seen": 2035154944
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001933539814639682,
      "loss": 2.5101,
      "theoretical_loss": 3.425412716095554,
      "tokens_seen": 2035220480
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019333743143559674,
      "loss": 2.6331,
      "theoretical_loss": 3.4254035511257053,
      "tokens_seen": 2035286016
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019332088140722525,
      "loss": 2.4361,
      "theoretical_loss": 3.4253943865335903,
      "tokens_seen": 2035351552
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001933043313788538,
      "loss": 2.4891,
      "theoretical_loss": 3.425385222319182,
      "tokens_seen": 2035417088
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001932877813504823,
      "loss": 2.5661,
      "theoretical_loss": 3.425376058482453,
      "tokens_seen": 2035482624
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019327123132211084,
      "loss": 2.4984,
      "theoretical_loss": 3.425366895023375,
      "tokens_seen": 2035548160
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019325468129373935,
      "loss": 2.5217,
      "theoretical_loss": 3.425357731941921,
      "tokens_seen": 2035613696
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001932381312653679,
      "loss": 2.6046,
      "theoretical_loss": 3.425348569238062,
      "tokens_seen": 2035679232
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001932215812369964,
      "loss": 2.7267,
      "theoretical_loss": 3.4253394069117715,
      "tokens_seen": 2035744768
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019320503120862494,
      "loss": 2.3699,
      "theoretical_loss": 3.425330244963021,
      "tokens_seen": 2035810304
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019318848118025345,
      "loss": 2.3687,
      "theoretical_loss": 3.4253210833917835,
      "tokens_seen": 2035875840
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019317193115188198,
      "loss": 2.4749,
      "theoretical_loss": 3.425311922198031,
      "tokens_seen": 2035941376
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001931553811235105,
      "loss": 2.4998,
      "theoretical_loss": 3.4253027613817357,
      "tokens_seen": 2036006912
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019313883109513903,
      "loss": 2.4454,
      "theoretical_loss": 3.4252936009428696,
      "tokens_seen": 2036072448
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019312228106676754,
      "loss": 2.5093,
      "theoretical_loss": 3.4252844408814056,
      "tokens_seen": 2036137984
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019310573103839608,
      "loss": 2.5669,
      "theoretical_loss": 3.4252752811973157,
      "tokens_seen": 2036203520
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001930891810100246,
      "loss": 2.3594,
      "theoretical_loss": 3.4252661218905724,
      "tokens_seen": 2036269056
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019307263098165313,
      "loss": 2.547,
      "theoretical_loss": 3.4252569629611473,
      "tokens_seen": 2036334592
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019305608095328164,
      "loss": 2.2518,
      "theoretical_loss": 3.4252478044090138,
      "tokens_seen": 2036400128
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019303953092491018,
      "loss": 2.3628,
      "theoretical_loss": 3.4252386462341438,
      "tokens_seen": 2036465664
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019302298089653869,
      "loss": 2.523,
      "theoretical_loss": 3.4252294884365093,
      "tokens_seen": 2036531200
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019300643086816722,
      "loss": 2.4594,
      "theoretical_loss": 3.425220331016083,
      "tokens_seen": 2036596736
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019298988083979573,
      "loss": 2.5589,
      "theoretical_loss": 3.4252111739728366,
      "tokens_seen": 2036662272
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019297333081142427,
      "loss": 2.4179,
      "theoretical_loss": 3.4252020173067432,
      "tokens_seen": 2036727808
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019295678078305278,
      "loss": 2.5837,
      "theoretical_loss": 3.4251928610177744,
      "tokens_seen": 2036793344
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019294023075468127,
      "loss": 2.7234,
      "theoretical_loss": 3.4251837051059035,
      "tokens_seen": 2036858880
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001929236807263098,
      "loss": 2.3857,
      "theoretical_loss": 3.425174549571102,
      "tokens_seen": 2036924416
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001929071306979383,
      "loss": 2.3512,
      "theoretical_loss": 3.4251653944133427,
      "tokens_seen": 2036989952
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019289058066956685,
      "loss": 2.585,
      "theoretical_loss": 3.4251562396325976,
      "tokens_seen": 2037055488
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019287403064119536,
      "loss": 2.3285,
      "theoretical_loss": 3.4251470852288395,
      "tokens_seen": 2037121024
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001928574806128239,
      "loss": 2.6128,
      "theoretical_loss": 3.4251379312020402,
      "tokens_seen": 2037186560
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001928409305844524,
      "loss": 2.4683,
      "theoretical_loss": 3.4251287775521724,
      "tokens_seen": 2037252096
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019282438055608095,
      "loss": 2.681,
      "theoretical_loss": 3.4251196242792084,
      "tokens_seen": 2037317632
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019280783052770946,
      "loss": 2.3953,
      "theoretical_loss": 3.42511047138312,
      "tokens_seen": 2037383168
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000192791280499338,
      "loss": 2.5755,
      "theoretical_loss": 3.425101318863881,
      "tokens_seen": 2037448704
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001927747304709665,
      "loss": 2.2292,
      "theoretical_loss": 3.4250921667214627,
      "tokens_seen": 2037514240
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019275818044259504,
      "loss": 2.4343,
      "theoretical_loss": 3.4250830149558373,
      "tokens_seen": 2037579776
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019274163041422355,
      "loss": 2.6409,
      "theoretical_loss": 3.4250738635669773,
      "tokens_seen": 2037645312
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001927250803858521,
      "loss": 2.6211,
      "theoretical_loss": 3.425064712554856,
      "tokens_seen": 2037710848
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001927085303574806,
      "loss": 2.6817,
      "theoretical_loss": 3.425055561919444,
      "tokens_seen": 2037776384
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019269198032910914,
      "loss": 2.7945,
      "theoretical_loss": 3.4250464116607153,
      "tokens_seen": 2037841920
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019267543030073765,
      "loss": 2.1485,
      "theoretical_loss": 3.4250372617786415,
      "tokens_seen": 2037907456
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001926588802723662,
      "loss": 2.6006,
      "theoretical_loss": 3.4250281122731954,
      "tokens_seen": 2037972992
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001926423302439947,
      "loss": 2.6363,
      "theoretical_loss": 3.4250189631443493,
      "tokens_seen": 2038038528
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019262578021562323,
      "loss": 2.5445,
      "theoretical_loss": 3.425009814392075,
      "tokens_seen": 2038104064
    },
    {
      "epoch": 0.73,
      "objective/train/avg_token_score": 0.024454262107610703,
      "objective/train/avg_weight": 0.97124844789505,
      "objective/train/docs_used": 1153257,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6916656494140625,
      "objective/train/std_weight": 0.1040571928024292,
      "objective/train/theoretical_loss": 3.425000666016346,
      "objective/train/tokens_used": 2058629600,
      "theoretical_loss": 3.425000666016346,
      "tokens_seen": 2038169600
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019260923018725175,
      "loss": 2.6149,
      "theoretical_loss": 3.425000666016346,
      "tokens_seen": 2038169600
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019259268015888028,
      "loss": 2.3837,
      "theoretical_loss": 3.4249915180171335,
      "tokens_seen": 2038235136
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001925761301305088,
      "loss": 2.2254,
      "theoretical_loss": 3.4249823703944102,
      "tokens_seen": 2038300672
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019255958010213733,
      "loss": 2.482,
      "theoretical_loss": 3.4249732231481493,
      "tokens_seen": 2038366208
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019254303007376584,
      "loss": 2.3032,
      "theoretical_loss": 3.4249640762783224,
      "tokens_seen": 2038431744
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019252648004539438,
      "loss": 2.3783,
      "theoretical_loss": 3.424954929784902,
      "tokens_seen": 2038497280
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001925099300170229,
      "loss": 2.6793,
      "theoretical_loss": 3.424945783667861,
      "tokens_seen": 2038562816
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019249337998865143,
      "loss": 2.3917,
      "theoretical_loss": 3.4249366379271713,
      "tokens_seen": 2038628352
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019247682996027994,
      "loss": 2.4485,
      "theoretical_loss": 3.4249274925628055,
      "tokens_seen": 2038693888
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019246027993190847,
      "loss": 2.4203,
      "theoretical_loss": 3.424918347574736,
      "tokens_seen": 2038759424
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019244372990353696,
      "loss": 2.196,
      "theoretical_loss": 3.4249092029629353,
      "tokens_seen": 2038824960
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019242717987516547,
      "loss": 2.5867,
      "theoretical_loss": 3.424900058727376,
      "tokens_seen": 2038890496
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000192410629846794,
      "loss": 2.4885,
      "theoretical_loss": 3.4248909148680298,
      "tokens_seen": 2038956032
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019239407981842252,
      "loss": 2.7541,
      "theoretical_loss": 3.42488177138487,
      "tokens_seen": 2039021568
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019237752979005105,
      "loss": 2.4876,
      "theoretical_loss": 3.424872628277868,
      "tokens_seen": 2039087104
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019236097976167956,
      "loss": 2.7458,
      "theoretical_loss": 3.4248634855469975,
      "tokens_seen": 2039152640
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001923444297333081,
      "loss": 2.4469,
      "theoretical_loss": 3.42485434319223,
      "tokens_seen": 2039218176
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001923278797049366,
      "loss": 2.413,
      "theoretical_loss": 3.4248452012135386,
      "tokens_seen": 2039283712
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019231132967656515,
      "loss": 2.6459,
      "theoretical_loss": 3.424836059610895,
      "tokens_seen": 2039349248
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019229477964819366,
      "loss": 2.4917,
      "theoretical_loss": 3.424826918384272,
      "tokens_seen": 2039414784
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001922782296198222,
      "loss": 2.5699,
      "theoretical_loss": 3.424817777533643,
      "tokens_seen": 2039480320
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001922616795914507,
      "loss": 2.414,
      "theoretical_loss": 3.4248086370589785,
      "tokens_seen": 2039545856
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019224512956307925,
      "loss": 2.4403,
      "theoretical_loss": 3.4247994969602527,
      "tokens_seen": 2039611392
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019222857953470778,
      "loss": 2.4608,
      "theoretical_loss": 3.4247903572374367,
      "tokens_seen": 2039676928
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001922120295063363,
      "loss": 2.6049,
      "theoretical_loss": 3.424781217890504,
      "tokens_seen": 2039742464
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019219547947796483,
      "loss": 2.6013,
      "theoretical_loss": 3.4247720789194274,
      "tokens_seen": 2039808000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019217892944959334,
      "loss": 2.5004,
      "theoretical_loss": 3.424762940324178,
      "tokens_seen": 2039873536
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019216237942122188,
      "loss": 2.4761,
      "theoretical_loss": 3.424753802104729,
      "tokens_seen": 2039939072
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001921458293928504,
      "loss": 2.5734,
      "theoretical_loss": 3.4247446642610524,
      "tokens_seen": 2040004608
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019212927936447893,
      "loss": 2.6749,
      "theoretical_loss": 3.4247355267931217,
      "tokens_seen": 2040070144
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019211272933610744,
      "loss": 2.5893,
      "theoretical_loss": 3.4247263897009086,
      "tokens_seen": 2040135680
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019209617930773598,
      "loss": 2.6247,
      "theoretical_loss": 3.4247172529843857,
      "tokens_seen": 2040201216
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019207962927936449,
      "loss": 2.65,
      "theoretical_loss": 3.4247081166435254,
      "tokens_seen": 2040266752
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019206307925099302,
      "loss": 2.7196,
      "theoretical_loss": 3.4246989806783006,
      "tokens_seen": 2040332288
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019204652922262153,
      "loss": 2.4719,
      "theoretical_loss": 3.4246898450886833,
      "tokens_seen": 2040397824
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019202997919425007,
      "loss": 2.4463,
      "theoretical_loss": 3.4246807098746466,
      "tokens_seen": 2040463360
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019201342916587858,
      "loss": 2.4507,
      "theoretical_loss": 3.424671575036162,
      "tokens_seen": 2040528896
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019199687913750712,
      "loss": 2.6688,
      "theoretical_loss": 3.424662440573203,
      "tokens_seen": 2040594432
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019198032910913563,
      "loss": 2.4926,
      "theoretical_loss": 3.4246533064857414,
      "tokens_seen": 2040659968
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019196377908076417,
      "loss": 2.3146,
      "theoretical_loss": 3.4246441727737507,
      "tokens_seen": 2040725504
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019194722905239265,
      "loss": 2.695,
      "theoretical_loss": 3.424635039437202,
      "tokens_seen": 2040791040
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019193067902402116,
      "loss": 2.5359,
      "theoretical_loss": 3.4246259064760687,
      "tokens_seen": 2040856576
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001919141289956497,
      "loss": 2.5383,
      "theoretical_loss": 3.4246167738903233,
      "tokens_seen": 2040922112
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001918975789672782,
      "loss": 2.4416,
      "theoretical_loss": 3.424607641679938,
      "tokens_seen": 2040987648
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019188102893890675,
      "loss": 2.7224,
      "theoretical_loss": 3.424598509844886,
      "tokens_seen": 2041053184
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019186447891053526,
      "loss": 2.5983,
      "theoretical_loss": 3.4245893783851384,
      "tokens_seen": 2041118720
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001918479288821638,
      "loss": 2.5039,
      "theoretical_loss": 3.4245802473006695,
      "tokens_seen": 2041184256
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001918313788537923,
      "loss": 2.3923,
      "theoretical_loss": 3.4245711165914505,
      "tokens_seen": 2041249792
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019181482882542084,
      "loss": 2.5634,
      "theoretical_loss": 3.4245619862574546,
      "tokens_seen": 2041315328
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019179827879704935,
      "loss": 2.5683,
      "theoretical_loss": 3.424552856298654,
      "tokens_seen": 2041380864
    },
    {
      "epoch": 0.73,
      "objective/train/avg_token_score": 0.014046735130250454,
      "objective/train/avg_weight": 0.9834700226783752,
      "objective/train/docs_used": 1155326,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.806427001953125,
      "objective/train/std_weight": 0.0797828733921051,
      "objective/train/theoretical_loss": 3.4245437267150214,
      "objective/train/tokens_used": 2061906400,
      "theoretical_loss": 3.4245437267150214,
      "tokens_seen": 2041446400
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001917817287686779,
      "loss": 2.3591,
      "theoretical_loss": 3.4245437267150214,
      "tokens_seen": 2041446400
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001917651787403064,
      "loss": 2.516,
      "theoretical_loss": 3.4245345975065296,
      "tokens_seen": 2041511936
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019174862871193494,
      "loss": 2.3698,
      "theoretical_loss": 3.4245254686731506,
      "tokens_seen": 2041577472
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019173207868356345,
      "loss": 2.7186,
      "theoretical_loss": 3.424516340214857,
      "tokens_seen": 2041643008
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000191715528655192,
      "loss": 2.2692,
      "theoretical_loss": 3.424507212131622,
      "tokens_seen": 2041708544
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001916989786268205,
      "loss": 2.5227,
      "theoretical_loss": 3.4244980844234174,
      "tokens_seen": 2041774080
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019168242859844903,
      "loss": 2.569,
      "theoretical_loss": 3.424488957090216,
      "tokens_seen": 2041839616
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019166587857007754,
      "loss": 2.6082,
      "theoretical_loss": 3.4244798301319905,
      "tokens_seen": 2041905152
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019164932854170608,
      "loss": 2.5754,
      "theoretical_loss": 3.424470703548714,
      "tokens_seen": 2041970688
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001916327785133346,
      "loss": 2.6621,
      "theoretical_loss": 3.4244615773403577,
      "tokens_seen": 2042036224
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019161622848496313,
      "loss": 2.3013,
      "theoretical_loss": 3.4244524515068946,
      "tokens_seen": 2042101760
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019159967845659164,
      "loss": 2.644,
      "theoretical_loss": 3.4244433260482983,
      "tokens_seen": 2042167296
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019158312842822018,
      "loss": 2.6297,
      "theoretical_loss": 3.4244342009645403,
      "tokens_seen": 2042232832
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001915665783998487,
      "loss": 2.516,
      "theoretical_loss": 3.4244250762555937,
      "tokens_seen": 2042298368
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019155002837147723,
      "loss": 2.4789,
      "theoretical_loss": 3.424415951921431,
      "tokens_seen": 2042363904
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019153347834310574,
      "loss": 2.5423,
      "theoretical_loss": 3.4244068279620246,
      "tokens_seen": 2042429440
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019151692831473427,
      "loss": 2.5263,
      "theoretical_loss": 3.4243977043773466,
      "tokens_seen": 2042494976
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019150037828636278,
      "loss": 2.3489,
      "theoretical_loss": 3.4243885811673707,
      "tokens_seen": 2042560512
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019148382825799132,
      "loss": 2.7078,
      "theoretical_loss": 3.424379458332069,
      "tokens_seen": 2042626048
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001914672782296198,
      "loss": 2.4224,
      "theoretical_loss": 3.424370335871414,
      "tokens_seen": 2042691584
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019145072820124832,
      "loss": 2.5814,
      "theoretical_loss": 3.424361213785378,
      "tokens_seen": 2042757120
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019143417817287685,
      "loss": 2.5321,
      "theoretical_loss": 3.424352092073934,
      "tokens_seen": 2042822656
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019141762814450536,
      "loss": 2.5288,
      "theoretical_loss": 3.424342970737055,
      "tokens_seen": 2042888192
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001914010781161339,
      "loss": 2.329,
      "theoretical_loss": 3.4243338497747127,
      "tokens_seen": 2042953728
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001913845280877624,
      "loss": 2.8377,
      "theoretical_loss": 3.42432472918688,
      "tokens_seen": 2043019264
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019136797805939095,
      "loss": 2.4415,
      "theoretical_loss": 3.4243156089735303,
      "tokens_seen": 2043084800
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019135142803101946,
      "loss": 2.4948,
      "theoretical_loss": 3.424306489134635,
      "tokens_seen": 2043150336
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000191334878002648,
      "loss": 2.6535,
      "theoretical_loss": 3.4242973696701675,
      "tokens_seen": 2043215872
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001913183279742765,
      "loss": 2.8146,
      "theoretical_loss": 3.4242882505801,
      "tokens_seen": 2043281408
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019130177794590505,
      "loss": 2.7781,
      "theoretical_loss": 3.424279131864406,
      "tokens_seen": 2043346944
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019128522791753356,
      "loss": 2.4872,
      "theoretical_loss": 3.4242700135230564,
      "tokens_seen": 2043412480
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001912686778891621,
      "loss": 2.7132,
      "theoretical_loss": 3.424260895556025,
      "tokens_seen": 2043478016
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001912521278607906,
      "loss": 2.3691,
      "theoretical_loss": 3.424251777963285,
      "tokens_seen": 2043543552
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019123557783241914,
      "loss": 2.3939,
      "theoretical_loss": 3.4242426607448078,
      "tokens_seen": 2043609088
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019121902780404765,
      "loss": 2.7442,
      "theoretical_loss": 3.424233543900567,
      "tokens_seen": 2043674624
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001912024777756762,
      "loss": 2.4222,
      "theoretical_loss": 3.4242244274305347,
      "tokens_seen": 2043740160
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001911859277473047,
      "loss": 2.7504,
      "theoretical_loss": 3.4242153113346836,
      "tokens_seen": 2043805696
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019116937771893324,
      "loss": 2.4461,
      "theoretical_loss": 3.424206195612986,
      "tokens_seen": 2043871232
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019115282769056175,
      "loss": 2.5561,
      "theoretical_loss": 3.424197080265415,
      "tokens_seen": 2043936768
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019113627766219029,
      "loss": 2.4221,
      "theoretical_loss": 3.424187965291944,
      "tokens_seen": 2044002304
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001911197276338188,
      "loss": 2.5267,
      "theoretical_loss": 3.4241788506925444,
      "tokens_seen": 2044067840
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019110317760544733,
      "loss": 2.7906,
      "theoretical_loss": 3.4241697364671886,
      "tokens_seen": 2044133376
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019108662757707584,
      "loss": 2.4916,
      "theoretical_loss": 3.424160622615851,
      "tokens_seen": 2044198912
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019107007754870438,
      "loss": 2.7265,
      "theoretical_loss": 3.424151509138502,
      "tokens_seen": 2044264448
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001910535275203329,
      "loss": 2.6065,
      "theoretical_loss": 3.4241423960351165,
      "tokens_seen": 2044329984
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019103697749196143,
      "loss": 2.5027,
      "theoretical_loss": 3.4241332833056664,
      "tokens_seen": 2044395520
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019102042746358994,
      "loss": 2.6195,
      "theoretical_loss": 3.4241241709501233,
      "tokens_seen": 2044461056
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019100387743521848,
      "loss": 2.4586,
      "theoretical_loss": 3.424115058968461,
      "tokens_seen": 2044526592
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000190987327406847,
      "loss": 2.5513,
      "theoretical_loss": 3.424105947360652,
      "tokens_seen": 2044592128
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001909707773784755,
      "loss": 2.4514,
      "theoretical_loss": 3.424096836126669,
      "tokens_seen": 2044657664
    },
    {
      "epoch": 0.73,
      "objective/train/avg_token_score": 0.01049480028450489,
      "objective/train/avg_weight": 0.9876465797424316,
      "objective/train/docs_used": 1157326,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6391329765319824,
      "objective/train/std_weight": 0.06826043128967285,
      "objective/train/theoretical_loss": 3.4240877252664834,
      "objective/train/tokens_used": 2065183200,
      "theoretical_loss": 3.4240877252664834,
      "tokens_seen": 2044723200
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000190954227350104,
      "loss": 2.6052,
      "theoretical_loss": 3.4240877252664834,
      "tokens_seen": 2044723200
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019093767732173255,
      "loss": 2.3609,
      "theoretical_loss": 3.42407861478007,
      "tokens_seen": 2044788736
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019092112729336106,
      "loss": 2.5479,
      "theoretical_loss": 3.4240695046674006,
      "tokens_seen": 2044854272
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001909045772649896,
      "loss": 2.3544,
      "theoretical_loss": 3.424060394928447,
      "tokens_seen": 2044919808
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001908880272366181,
      "loss": 2.6276,
      "theoretical_loss": 3.4240512855631833,
      "tokens_seen": 2044985344
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019087147720824664,
      "loss": 2.576,
      "theoretical_loss": 3.424042176571582,
      "tokens_seen": 2045050880
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019085492717987515,
      "loss": 2.6552,
      "theoretical_loss": 3.424033067953615,
      "tokens_seen": 2045116416
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001908383771515037,
      "loss": 2.5616,
      "theoretical_loss": 3.4240239597092548,
      "tokens_seen": 2045181952
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001908218271231322,
      "loss": 2.5198,
      "theoretical_loss": 3.424014851838475,
      "tokens_seen": 2045247488
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019080527709476074,
      "loss": 2.5725,
      "theoretical_loss": 3.4240057443412484,
      "tokens_seen": 2045313024
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019078872706638925,
      "loss": 2.5365,
      "theoretical_loss": 3.423996637217547,
      "tokens_seen": 2045378560
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019077217703801779,
      "loss": 2.5111,
      "theoretical_loss": 3.4239875304673437,
      "tokens_seen": 2045444096
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001907556270096463,
      "loss": 2.8456,
      "theoretical_loss": 3.4239784240906115,
      "tokens_seen": 2045509632
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019073907698127483,
      "loss": 2.6922,
      "theoretical_loss": 3.423969318087323,
      "tokens_seen": 2045575168
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019072252695290334,
      "loss": 2.3906,
      "theoretical_loss": 3.423960212457451,
      "tokens_seen": 2045640704
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019070597692453188,
      "loss": 2.2935,
      "theoretical_loss": 3.423951107200968,
      "tokens_seen": 2045706240
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001906894268961604,
      "loss": 2.6016,
      "theoretical_loss": 3.423942002317846,
      "tokens_seen": 2045771776
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019067287686778893,
      "loss": 2.3123,
      "theoretical_loss": 3.4239328978080596,
      "tokens_seen": 2045837312
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019065632683941744,
      "loss": 2.6499,
      "theoretical_loss": 3.4239237936715803,
      "tokens_seen": 2045902848
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019063977681104598,
      "loss": 2.6771,
      "theoretical_loss": 3.4239146899083805,
      "tokens_seen": 2045968384
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001906232267826745,
      "loss": 2.5243,
      "theoretical_loss": 3.4239055865184342,
      "tokens_seen": 2046033920
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019060667675430303,
      "loss": 2.4908,
      "theoretical_loss": 3.423896483501713,
      "tokens_seen": 2046099456
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019059012672593154,
      "loss": 2.2831,
      "theoretical_loss": 3.42388738085819,
      "tokens_seen": 2046164992
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019057357669756007,
      "loss": 2.5776,
      "theoretical_loss": 3.423878278587838,
      "tokens_seen": 2046230528
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019055702666918858,
      "loss": 2.6878,
      "theoretical_loss": 3.42386917669063,
      "tokens_seen": 2046296064
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019054047664081712,
      "loss": 2.4753,
      "theoretical_loss": 3.4238600751665382,
      "tokens_seen": 2046361600
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019052392661244563,
      "loss": 2.6372,
      "theoretical_loss": 3.4238509740155356,
      "tokens_seen": 2046427136
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019050737658407417,
      "loss": 2.6052,
      "theoretical_loss": 3.4238418732375955,
      "tokens_seen": 2046492672
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019049082655570268,
      "loss": 2.6542,
      "theoretical_loss": 3.42383277283269,
      "tokens_seen": 2046558208
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019047427652733116,
      "loss": 2.4253,
      "theoretical_loss": 3.4238236728007916,
      "tokens_seen": 2046623744
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001904577264989597,
      "loss": 2.4353,
      "theoretical_loss": 3.4238145731418737,
      "tokens_seen": 2046689280
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001904411764705882,
      "loss": 2.4982,
      "theoretical_loss": 3.4238054738559094,
      "tokens_seen": 2046754816
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019042462644221675,
      "loss": 2.5507,
      "theoretical_loss": 3.4237963749428704,
      "tokens_seen": 2046820352
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019040807641384526,
      "loss": 2.784,
      "theoretical_loss": 3.4237872764027304,
      "tokens_seen": 2046885888
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001903915263854738,
      "loss": 2.592,
      "theoretical_loss": 3.4237781782354615,
      "tokens_seen": 2046951424
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001903749763571023,
      "loss": 2.5133,
      "theoretical_loss": 3.423769080441037,
      "tokens_seen": 2047016960
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019035842632873085,
      "loss": 2.477,
      "theoretical_loss": 3.423759983019429,
      "tokens_seen": 2047082496
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019034187630035936,
      "loss": 2.1351,
      "theoretical_loss": 3.4237508859706116,
      "tokens_seen": 2047148032
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001903253262719879,
      "loss": 2.5214,
      "theoretical_loss": 3.423741789294556,
      "tokens_seen": 2047213568
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001903087762436164,
      "loss": 2.6016,
      "theoretical_loss": 3.423732692991236,
      "tokens_seen": 2047279104
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019029222621524494,
      "loss": 2.5106,
      "theoretical_loss": 3.4237235970606243,
      "tokens_seen": 2047344640
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019027567618687345,
      "loss": 2.6481,
      "theoretical_loss": 3.4237145015026935,
      "tokens_seen": 2047410176
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000190259126158502,
      "loss": 2.6177,
      "theoretical_loss": 3.4237054063174166,
      "tokens_seen": 2047475712
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001902425761301305,
      "loss": 2.5565,
      "theoretical_loss": 3.4236963115047656,
      "tokens_seen": 2047541248
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019022602610175904,
      "loss": 2.4818,
      "theoretical_loss": 3.423687217064715,
      "tokens_seen": 2047606784
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019020947607338755,
      "loss": 2.4675,
      "theoretical_loss": 3.4236781229972353,
      "tokens_seen": 2047672320
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019019292604501609,
      "loss": 2.5213,
      "theoretical_loss": 3.4236690293023018,
      "tokens_seen": 2047737856
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001901763760166446,
      "loss": 2.5466,
      "theoretical_loss": 3.4236599359798854,
      "tokens_seen": 2047803392
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019015982598827313,
      "loss": 2.5857,
      "theoretical_loss": 3.42365084302996,
      "tokens_seen": 2047868928
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019014327595990164,
      "loss": 2.2281,
      "theoretical_loss": 3.4236417504524974,
      "tokens_seen": 2047934464
    },
    {
      "epoch": 0.73,
      "objective/train/avg_token_score": 0.026696883141994476,
      "objective/train/avg_weight": 0.9686170816421509,
      "objective/train/docs_used": 1159281,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.722960948944092,
      "objective/train/std_weight": 0.11805140972137451,
      "objective/train/theoretical_loss": 3.4236326582474716,
      "objective/train/tokens_used": 2068460000,
      "theoretical_loss": 3.4236326582474716,
      "tokens_seen": 2048000000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019012672593153018,
      "loss": 2.5502,
      "theoretical_loss": 3.4236326582474716,
      "tokens_seen": 2048000000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001901101759031587,
      "loss": 2.4328,
      "theoretical_loss": 3.4236235664148547,
      "tokens_seen": 2048065536
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019009362587478723,
      "loss": 2.5968,
      "theoretical_loss": 3.42361447495462,
      "tokens_seen": 2048131072
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019007707584641574,
      "loss": 2.5383,
      "theoretical_loss": 3.4236053838667395,
      "tokens_seen": 2048196608
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019006052581804428,
      "loss": 2.0851,
      "theoretical_loss": 3.4235962931511867,
      "tokens_seen": 2048262144
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001900439757896728,
      "loss": 2.7292,
      "theoretical_loss": 3.4235872028079353,
      "tokens_seen": 2048327680
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019002742576130133,
      "loss": 2.4782,
      "theoretical_loss": 3.423578112836956,
      "tokens_seen": 2048393216
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00019001087573292984,
      "loss": 2.1053,
      "theoretical_loss": 3.4235690232382234,
      "tokens_seen": 2048458752
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018999432570455837,
      "loss": 2.4872,
      "theoretical_loss": 3.4235599340117098,
      "tokens_seen": 2048524288
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018997777567618686,
      "loss": 2.5389,
      "theoretical_loss": 3.4235508451573873,
      "tokens_seen": 2048589824
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018996122564781537,
      "loss": 2.4266,
      "theoretical_loss": 3.42354175667523,
      "tokens_seen": 2048655360
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001899446756194439,
      "loss": 2.7658,
      "theoretical_loss": 3.42353266856521,
      "tokens_seen": 2048720896
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018992812559107241,
      "loss": 2.8413,
      "theoretical_loss": 3.4235235808273012,
      "tokens_seen": 2048786432
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018991157556270095,
      "loss": 2.5379,
      "theoretical_loss": 3.4235144934614747,
      "tokens_seen": 2048851968
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018989502553432946,
      "loss": 2.519,
      "theoretical_loss": 3.423505406467705,
      "tokens_seen": 2048917504
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000189878475505958,
      "loss": 2.3991,
      "theoretical_loss": 3.4234963198459636,
      "tokens_seen": 2048983040
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001898619254775865,
      "loss": 2.5472,
      "theoretical_loss": 3.4234872335962248,
      "tokens_seen": 2049048576
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018984537544921505,
      "loss": 2.241,
      "theoretical_loss": 3.42347814771846,
      "tokens_seen": 2049114112
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018982882542084356,
      "loss": 2.7161,
      "theoretical_loss": 3.423469062212643,
      "tokens_seen": 2049179648
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001898122753924721,
      "loss": 2.6649,
      "theoretical_loss": 3.4234599770787466,
      "tokens_seen": 2049245184
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001897957253641006,
      "loss": 2.7178,
      "theoretical_loss": 3.4234508923167435,
      "tokens_seen": 2049310720
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018977917533572914,
      "loss": 2.5495,
      "theoretical_loss": 3.423441807926607,
      "tokens_seen": 2049376256
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018976262530735765,
      "loss": 2.5295,
      "theoretical_loss": 3.423432723908309,
      "tokens_seen": 2049441792
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001897460752789862,
      "loss": 2.7446,
      "theoretical_loss": 3.423423640261823,
      "tokens_seen": 2049507328
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001897295252506147,
      "loss": 2.3106,
      "theoretical_loss": 3.4234145569871224,
      "tokens_seen": 2049572864
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018971297522224324,
      "loss": 2.7246,
      "theoretical_loss": 3.4234054740841797,
      "tokens_seen": 2049638400
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018969642519387175,
      "loss": 2.5817,
      "theoretical_loss": 3.4233963915529673,
      "tokens_seen": 2049703936
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001896798751655003,
      "loss": 2.6292,
      "theoretical_loss": 3.4233873093934584,
      "tokens_seen": 2049769472
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001896633251371288,
      "loss": 2.319,
      "theoretical_loss": 3.4233782276056264,
      "tokens_seen": 2049835008
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018964677510875734,
      "loss": 2.7194,
      "theoretical_loss": 3.4233691461894438,
      "tokens_seen": 2049900544
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018963022508038585,
      "loss": 2.6266,
      "theoretical_loss": 3.4233600651448834,
      "tokens_seen": 2049966080
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018961367505201438,
      "loss": 2.4649,
      "theoretical_loss": 3.4233509844719183,
      "tokens_seen": 2050031616
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001895971250236429,
      "loss": 2.477,
      "theoretical_loss": 3.4233419041705213,
      "tokens_seen": 2050097152
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018958057499527143,
      "loss": 2.5722,
      "theoretical_loss": 3.4233328242406653,
      "tokens_seen": 2050162688
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018956402496689994,
      "loss": 2.6224,
      "theoretical_loss": 3.4233237446823237,
      "tokens_seen": 2050228224
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018954747493852848,
      "loss": 2.4616,
      "theoretical_loss": 3.4233146654954685,
      "tokens_seen": 2050293760
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000189530924910157,
      "loss": 2.5282,
      "theoretical_loss": 3.4233055866800735,
      "tokens_seen": 2050359296
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018951437488178553,
      "loss": 2.2455,
      "theoretical_loss": 3.4232965082361106,
      "tokens_seen": 2050424832
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000189497824853414,
      "loss": 2.4186,
      "theoretical_loss": 3.4232874301635543,
      "tokens_seen": 2050490368
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018948127482504255,
      "loss": 2.3952,
      "theoretical_loss": 3.4232783524623764,
      "tokens_seen": 2050555904
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018946472479667106,
      "loss": 2.172,
      "theoretical_loss": 3.42326927513255,
      "tokens_seen": 2050621440
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001894481747682996,
      "loss": 2.6685,
      "theoretical_loss": 3.423260198174048,
      "tokens_seen": 2050686976
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001894316247399281,
      "loss": 2.6978,
      "theoretical_loss": 3.4232511215868433,
      "tokens_seen": 2050752512
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018941507471155665,
      "loss": 2.4882,
      "theoretical_loss": 3.42324204537091,
      "tokens_seen": 2050818048
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018939852468318516,
      "loss": 2.5632,
      "theoretical_loss": 3.423232969526219,
      "tokens_seen": 2050883584
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001893819746548137,
      "loss": 2.5927,
      "theoretical_loss": 3.423223894052745,
      "tokens_seen": 2050949120
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001893654246264422,
      "loss": 2.6002,
      "theoretical_loss": 3.4232148189504596,
      "tokens_seen": 2051014656
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018934887459807074,
      "loss": 2.3658,
      "theoretical_loss": 3.423205744219337,
      "tokens_seen": 2051080192
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018933232456969925,
      "loss": 2.586,
      "theoretical_loss": 3.4231966698593492,
      "tokens_seen": 2051145728
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001893157745413278,
      "loss": 2.4839,
      "theoretical_loss": 3.42318759587047,
      "tokens_seen": 2051211264
    },
    {
      "epoch": 0.73,
      "objective/train/avg_token_score": 0.02220361866056919,
      "objective/train/avg_weight": 0.9738765954971313,
      "objective/train/docs_used": 1161569,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6131398677825928,
      "objective/train/std_weight": 0.09282509982585907,
      "objective/train/theoretical_loss": 3.423178522252672,
      "objective/train/tokens_used": 2071736800,
      "theoretical_loss": 3.423178522252672,
      "tokens_seen": 2051276800
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001892992245129563,
      "loss": 2.4934,
      "theoretical_loss": 3.423178522252672,
      "tokens_seen": 2051276800
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018928267448458484,
      "loss": 2.7998,
      "theoretical_loss": 3.4231694490059272,
      "tokens_seen": 2051342336
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018926612445621335,
      "loss": 2.2928,
      "theoretical_loss": 3.4231603761302107,
      "tokens_seen": 2051407872
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018924957442784189,
      "loss": 2.8637,
      "theoretical_loss": 3.4231513036254935,
      "tokens_seen": 2051473408
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001892330243994704,
      "loss": 2.5278,
      "theoretical_loss": 3.4231422314917497,
      "tokens_seen": 2051538944
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018921647437109893,
      "loss": 2.3445,
      "theoretical_loss": 3.423133159728951,
      "tokens_seen": 2051604480
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018919992434272744,
      "loss": 2.408,
      "theoretical_loss": 3.4231240883370724,
      "tokens_seen": 2051670016
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018918337431435598,
      "loss": 2.4612,
      "theoretical_loss": 3.423115017316085,
      "tokens_seen": 2051735552
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001891668242859845,
      "loss": 2.5485,
      "theoretical_loss": 3.423105946665963,
      "tokens_seen": 2051801088
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018915027425761303,
      "loss": 2.4285,
      "theoretical_loss": 3.423096876386679,
      "tokens_seen": 2051866624
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018913372422924154,
      "loss": 2.5133,
      "theoretical_loss": 3.4230878064782058,
      "tokens_seen": 2051932160
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018911717420087008,
      "loss": 2.5056,
      "theoretical_loss": 3.4230787369405165,
      "tokens_seen": 2051997696
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001891006241724986,
      "loss": 2.436,
      "theoretical_loss": 3.4230696677735843,
      "tokens_seen": 2052063232
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018908407414412712,
      "loss": 2.4353,
      "theoretical_loss": 3.423060598977382,
      "tokens_seen": 2052128768
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018906752411575564,
      "loss": 2.6205,
      "theoretical_loss": 3.4230515305518825,
      "tokens_seen": 2052194304
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018905097408738417,
      "loss": 2.7291,
      "theoretical_loss": 3.423042462497059,
      "tokens_seen": 2052259840
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018903442405901268,
      "loss": 2.4751,
      "theoretical_loss": 3.423033394812885,
      "tokens_seen": 2052325376
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018901787403064122,
      "loss": 2.5445,
      "theoretical_loss": 3.4230243274993324,
      "tokens_seen": 2052390912
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001890013240022697,
      "loss": 2.5654,
      "theoretical_loss": 3.423015260556375,
      "tokens_seen": 2052456448
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018898477397389821,
      "loss": 2.4971,
      "theoretical_loss": 3.4230061939839858,
      "tokens_seen": 2052521984
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018896822394552675,
      "loss": 2.4955,
      "theoretical_loss": 3.4229971277821374,
      "tokens_seen": 2052587520
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018895167391715526,
      "loss": 2.7646,
      "theoretical_loss": 3.4229880619508033,
      "tokens_seen": 2052653056
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001889351238887838,
      "loss": 2.5505,
      "theoretical_loss": 3.422978996489956,
      "tokens_seen": 2052718592
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001889185738604123,
      "loss": 2.6666,
      "theoretical_loss": 3.422969931399569,
      "tokens_seen": 2052784128
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018890202383204085,
      "loss": 2.238,
      "theoretical_loss": 3.422960866679616,
      "tokens_seen": 2052849664
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018888547380366936,
      "loss": 2.6029,
      "theoretical_loss": 3.422951802330068,
      "tokens_seen": 2052915200
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001888689237752979,
      "loss": 2.3873,
      "theoretical_loss": 3.4229427383508995,
      "tokens_seen": 2052980736
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001888523737469264,
      "loss": 2.3811,
      "theoretical_loss": 3.422933674742084,
      "tokens_seen": 2053046272
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018883582371855494,
      "loss": 2.0735,
      "theoretical_loss": 3.422924611503593,
      "tokens_seen": 2053111808
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018881927369018345,
      "loss": 2.437,
      "theoretical_loss": 3.422915548635401,
      "tokens_seen": 2053177344
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000188802723661812,
      "loss": 2.5504,
      "theoretical_loss": 3.4229064861374803,
      "tokens_seen": 2053242880
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001887861736334405,
      "loss": 2.459,
      "theoretical_loss": 3.422897424009804,
      "tokens_seen": 2053308416
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018876962360506904,
      "loss": 2.7253,
      "theoretical_loss": 3.422888362252345,
      "tokens_seen": 2053373952
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018875307357669755,
      "loss": 2.6587,
      "theoretical_loss": 3.4228793008650773,
      "tokens_seen": 2053439488
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001887365235483261,
      "loss": 2.4863,
      "theoretical_loss": 3.422870239847973,
      "tokens_seen": 2053505024
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001887199735199546,
      "loss": 2.6249,
      "theoretical_loss": 3.4228611792010053,
      "tokens_seen": 2053570560
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018870342349158314,
      "loss": 2.292,
      "theoretical_loss": 3.4228521189241476,
      "tokens_seen": 2053636096
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018868687346321165,
      "loss": 2.6776,
      "theoretical_loss": 3.4228430590173726,
      "tokens_seen": 2053701632
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018867032343484018,
      "loss": 2.4038,
      "theoretical_loss": 3.4228339994806536,
      "tokens_seen": 2053767168
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001886537734064687,
      "loss": 2.4677,
      "theoretical_loss": 3.422824940313964,
      "tokens_seen": 2053832704
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018863722337809723,
      "loss": 2.7023,
      "theoretical_loss": 3.422815881517276,
      "tokens_seen": 2053898240
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018862067334972574,
      "loss": 2.7378,
      "theoretical_loss": 3.4228068230905633,
      "tokens_seen": 2053963776
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018860412332135428,
      "loss": 2.6197,
      "theoretical_loss": 3.422797765033799,
      "tokens_seen": 2054029312
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001885875732929828,
      "loss": 2.7288,
      "theoretical_loss": 3.4227887073469563,
      "tokens_seen": 2054094848
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018857102326461133,
      "loss": 2.4801,
      "theoretical_loss": 3.422779650030008,
      "tokens_seen": 2054160384
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018855447323623984,
      "loss": 2.0972,
      "theoretical_loss": 3.422770593082927,
      "tokens_seen": 2054225920
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018853792320786838,
      "loss": 2.4426,
      "theoretical_loss": 3.422761536505687,
      "tokens_seen": 2054291456
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018852137317949689,
      "loss": 2.6598,
      "theoretical_loss": 3.4227524802982607,
      "tokens_seen": 2054356992
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018850482315112537,
      "loss": 2.4896,
      "theoretical_loss": 3.422743424460621,
      "tokens_seen": 2054422528
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001884882731227539,
      "loss": 2.5586,
      "theoretical_loss": 3.4227343689927414,
      "tokens_seen": 2054488064
    },
    {
      "epoch": 0.73,
      "objective/train/avg_token_score": 0.017169633880257607,
      "objective/train/avg_weight": 0.9797907471656799,
      "objective/train/docs_used": 1163144,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7262585163116455,
      "objective/train/std_weight": 0.07908392697572708,
      "objective/train/theoretical_loss": 3.4227253138945946,
      "objective/train/tokens_used": 2075013600,
      "theoretical_loss": 3.4227253138945946,
      "tokens_seen": 2054553600
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018847172309438242,
      "loss": 2.2677,
      "theoretical_loss": 3.4227253138945946,
      "tokens_seen": 2054553600
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018845517306601096,
      "loss": 2.5984,
      "theoretical_loss": 3.4227162591661546,
      "tokens_seen": 2054619136
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018843862303763947,
      "loss": 2.5655,
      "theoretical_loss": 3.422707204807393,
      "tokens_seen": 2054684672
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000188422073009268,
      "loss": 2.5589,
      "theoretical_loss": 3.4226981508182845,
      "tokens_seen": 2054750208
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018840552298089651,
      "loss": 2.5663,
      "theoretical_loss": 3.422689097198802,
      "tokens_seen": 2054815744
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018838897295252505,
      "loss": 2.5479,
      "theoretical_loss": 3.422680043948917,
      "tokens_seen": 2054881280
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018837242292415356,
      "loss": 2.4479,
      "theoretical_loss": 3.422670991068605,
      "tokens_seen": 2054946816
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001883558728957821,
      "loss": 2.5154,
      "theoretical_loss": 3.422661938557837,
      "tokens_seen": 2055012352
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001883393228674106,
      "loss": 2.5296,
      "theoretical_loss": 3.4226528864165875,
      "tokens_seen": 2055077888
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018832277283903915,
      "loss": 2.4728,
      "theoretical_loss": 3.4226438346448287,
      "tokens_seen": 2055143424
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018830622281066766,
      "loss": 2.2048,
      "theoretical_loss": 3.422634783242535,
      "tokens_seen": 2055208960
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001882896727822962,
      "loss": 2.5825,
      "theoretical_loss": 3.422625732209678,
      "tokens_seen": 2055274496
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001882731227539247,
      "loss": 2.5949,
      "theoretical_loss": 3.4226166815462316,
      "tokens_seen": 2055340032
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018825657272555324,
      "loss": 2.424,
      "theoretical_loss": 3.42260763125217,
      "tokens_seen": 2055405568
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018824002269718178,
      "loss": 2.5742,
      "theoretical_loss": 3.422598581327464,
      "tokens_seen": 2055471104
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001882234726688103,
      "loss": 2.3808,
      "theoretical_loss": 3.4225895317720885,
      "tokens_seen": 2055536640
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018820692264043883,
      "loss": 2.5982,
      "theoretical_loss": 3.4225804825860164,
      "tokens_seen": 2055602176
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018819037261206734,
      "loss": 2.4409,
      "theoretical_loss": 3.42257143376922,
      "tokens_seen": 2055667712
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018817382258369588,
      "loss": 2.3649,
      "theoretical_loss": 3.422562385321674,
      "tokens_seen": 2055733248
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001881572725553244,
      "loss": 2.4028,
      "theoretical_loss": 3.4225533372433503,
      "tokens_seen": 2055798784
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018814072252695292,
      "loss": 2.7459,
      "theoretical_loss": 3.422544289534222,
      "tokens_seen": 2055864320
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018812417249858144,
      "loss": 2.7557,
      "theoretical_loss": 3.422535242194263,
      "tokens_seen": 2055929856
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018810762247020997,
      "loss": 2.5174,
      "theoretical_loss": 3.4225261952234467,
      "tokens_seen": 2055995392
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018809107244183848,
      "loss": 2.7495,
      "theoretical_loss": 3.422517148621745,
      "tokens_seen": 2056060928
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018807452241346702,
      "loss": 2.2761,
      "theoretical_loss": 3.4225081023891324,
      "tokens_seen": 2056126464
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018805797238509553,
      "loss": 2.6981,
      "theoretical_loss": 3.422499056525581,
      "tokens_seen": 2056192000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018804142235672407,
      "loss": 2.6218,
      "theoretical_loss": 3.4224900110310648,
      "tokens_seen": 2056257536
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018802487232835255,
      "loss": 2.6436,
      "theoretical_loss": 3.422480965905556,
      "tokens_seen": 2056323072
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018800832229998106,
      "loss": 2.5468,
      "theoretical_loss": 3.4224719211490293,
      "tokens_seen": 2056388608
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001879917722716096,
      "loss": 2.4267,
      "theoretical_loss": 3.4224628767614567,
      "tokens_seen": 2056454144
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001879752222432381,
      "loss": 2.7003,
      "theoretical_loss": 3.4224538327428116,
      "tokens_seen": 2056519680
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018795867221486665,
      "loss": 2.5222,
      "theoretical_loss": 3.4224447890930674,
      "tokens_seen": 2056585216
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018794212218649516,
      "loss": 2.3981,
      "theoretical_loss": 3.422435745812197,
      "tokens_seen": 2056650752
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001879255721581237,
      "loss": 2.0736,
      "theoretical_loss": 3.4224267029001743,
      "tokens_seen": 2056716288
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001879090221297522,
      "loss": 2.5148,
      "theoretical_loss": 3.422417660356972,
      "tokens_seen": 2056781824
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018789247210138074,
      "loss": 2.6026,
      "theoretical_loss": 3.422408618182563,
      "tokens_seen": 2056847360
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018787592207300925,
      "loss": 2.5089,
      "theoretical_loss": 3.422399576376921,
      "tokens_seen": 2056912896
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001878593720446378,
      "loss": 2.3871,
      "theoretical_loss": 3.422390534940019,
      "tokens_seen": 2056978432
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001878428220162663,
      "loss": 2.2123,
      "theoretical_loss": 3.42238149387183,
      "tokens_seen": 2057043968
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018782627198789484,
      "loss": 2.5651,
      "theoretical_loss": 3.4223724531723283,
      "tokens_seen": 2057109504
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018780972195952335,
      "loss": 2.7552,
      "theoretical_loss": 3.4223634128414853,
      "tokens_seen": 2057175040
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001877931719311519,
      "loss": 2.7214,
      "theoretical_loss": 3.422354372879276,
      "tokens_seen": 2057240576
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001877766219027804,
      "loss": 2.7853,
      "theoretical_loss": 3.4223453332856724,
      "tokens_seen": 2057306112
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018776007187440894,
      "loss": 2.3731,
      "theoretical_loss": 3.4223362940606483,
      "tokens_seen": 2057371648
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018774352184603745,
      "loss": 2.55,
      "theoretical_loss": 3.422327255204177,
      "tokens_seen": 2057437184
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018772697181766598,
      "loss": 2.7007,
      "theoretical_loss": 3.4223182167162314,
      "tokens_seen": 2057502720
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001877104217892945,
      "loss": 2.6319,
      "theoretical_loss": 3.4223091785967847,
      "tokens_seen": 2057568256
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018769387176092303,
      "loss": 2.6106,
      "theoretical_loss": 3.4223001408458105,
      "tokens_seen": 2057633792
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018767732173255154,
      "loss": 2.6362,
      "theoretical_loss": 3.422291103463282,
      "tokens_seen": 2057699328
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018766077170418008,
      "loss": 2.3267,
      "theoretical_loss": 3.422282066449172,
      "tokens_seen": 2057764864
    },
    {
      "epoch": 0.73,
      "objective/train/avg_token_score": 0.014738215133547783,
      "objective/train/avg_weight": 0.982663631439209,
      "objective/train/docs_used": 1165022,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.9779000282287598,
      "objective/train/std_weight": 0.07284485548734665,
      "objective/train/theoretical_loss": 3.4222730298034545,
      "objective/train/tokens_used": 2078290400,
      "theoretical_loss": 3.4222730298034545,
      "tokens_seen": 2057830400
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001876442216758086,
      "loss": 2.686,
      "theoretical_loss": 3.4222730298034545,
      "tokens_seen": 2057830400
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018762767164743713,
      "loss": 2.5442,
      "theoretical_loss": 3.4222639935261014,
      "tokens_seen": 2057895936
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00018761112161906564,
      "loss": 2.2285,
      "theoretical_loss": 3.4222549576170875,
      "tokens_seen": 2057961472
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018759457159069418,
      "loss": 2.5393,
      "theoretical_loss": 3.422245922076386,
      "tokens_seen": 2058027008
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018757802156232269,
      "loss": 2.3267,
      "theoretical_loss": 3.4222368869039688,
      "tokens_seen": 2058092544
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018756147153395122,
      "loss": 2.5981,
      "theoretical_loss": 3.42222785209981,
      "tokens_seen": 2058158080
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018754492150557973,
      "loss": 2.4243,
      "theoretical_loss": 3.4222188176638833,
      "tokens_seen": 2058223616
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018752837147720822,
      "loss": 2.3394,
      "theoretical_loss": 3.422209783596161,
      "tokens_seen": 2058289152
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018751182144883676,
      "loss": 2.4746,
      "theoretical_loss": 3.422200749896617,
      "tokens_seen": 2058354688
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018749527142046527,
      "loss": 2.5085,
      "theoretical_loss": 3.4221917165652243,
      "tokens_seen": 2058420224
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001874787213920938,
      "loss": 2.4695,
      "theoretical_loss": 3.4221826836019567,
      "tokens_seen": 2058485760
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001874621713637223,
      "loss": 2.5275,
      "theoretical_loss": 3.4221736510067866,
      "tokens_seen": 2058551296
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018744562133535085,
      "loss": 2.5746,
      "theoretical_loss": 3.422164618779688,
      "tokens_seen": 2058616832
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018742907130697936,
      "loss": 2.3757,
      "theoretical_loss": 3.422155586920634,
      "tokens_seen": 2058682368
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001874125212786079,
      "loss": 2.2182,
      "theoretical_loss": 3.4221465554295976,
      "tokens_seen": 2058747904
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001873959712502364,
      "loss": 2.7017,
      "theoretical_loss": 3.422137524306552,
      "tokens_seen": 2058813440
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018737942122186495,
      "loss": 2.4871,
      "theoretical_loss": 3.4221284935514715,
      "tokens_seen": 2058878976
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018736287119349346,
      "loss": 2.813,
      "theoretical_loss": 3.4221194631643286,
      "tokens_seen": 2058944512
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000187346321165122,
      "loss": 2.378,
      "theoretical_loss": 3.4221104331450967,
      "tokens_seen": 2059010048
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001873297711367505,
      "loss": 2.612,
      "theoretical_loss": 3.4221014034937487,
      "tokens_seen": 2059075584
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018731322110837904,
      "loss": 2.4821,
      "theoretical_loss": 3.4220923742102585,
      "tokens_seen": 2059141120
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018729667108000755,
      "loss": 2.3587,
      "theoretical_loss": 3.4220833452945993,
      "tokens_seen": 2059206656
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001872801210516361,
      "loss": 2.5604,
      "theoretical_loss": 3.4220743167467442,
      "tokens_seen": 2059272192
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001872635710232646,
      "loss": 2.4282,
      "theoretical_loss": 3.422065288566667,
      "tokens_seen": 2059337728
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018724702099489314,
      "loss": 2.6119,
      "theoretical_loss": 3.4220562607543403,
      "tokens_seen": 2059403264
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018723047096652165,
      "loss": 2.6482,
      "theoretical_loss": 3.422047233309738,
      "tokens_seen": 2059468800
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001872139209381502,
      "loss": 2.8446,
      "theoretical_loss": 3.422038206232833,
      "tokens_seen": 2059534336
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001871973709097787,
      "loss": 2.6487,
      "theoretical_loss": 3.422029179523599,
      "tokens_seen": 2059599872
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018718082088140723,
      "loss": 2.5829,
      "theoretical_loss": 3.422020153182009,
      "tokens_seen": 2059665408
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018716427085303575,
      "loss": 2.4783,
      "theoretical_loss": 3.4220111272080365,
      "tokens_seen": 2059730944
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018714772082466428,
      "loss": 2.3636,
      "theoretical_loss": 3.4220021016016546,
      "tokens_seen": 2059796480
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001871311707962928,
      "loss": 2.6313,
      "theoretical_loss": 3.4219930763628374,
      "tokens_seen": 2059862016
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018711462076792133,
      "loss": 2.8382,
      "theoretical_loss": 3.421984051491557,
      "tokens_seen": 2059927552
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018709807073954984,
      "loss": 2.7436,
      "theoretical_loss": 3.4219750269877878,
      "tokens_seen": 2059993088
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018708152071117838,
      "loss": 2.2622,
      "theoretical_loss": 3.4219660028515024,
      "tokens_seen": 2060058624
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001870649706828069,
      "loss": 2.6015,
      "theoretical_loss": 3.421956979082675,
      "tokens_seen": 2060124160
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018704842065443543,
      "loss": 2.4635,
      "theoretical_loss": 3.4219479556812784,
      "tokens_seen": 2060189696
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001870318706260639,
      "loss": 2.5946,
      "theoretical_loss": 3.421938932647286,
      "tokens_seen": 2060255232
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018701532059769242,
      "loss": 2.4059,
      "theoretical_loss": 3.421929909980671,
      "tokens_seen": 2060320768
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018699877056932096,
      "loss": 2.5578,
      "theoretical_loss": 3.4219208876814067,
      "tokens_seen": 2060386304
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001869822205409495,
      "loss": 2.3468,
      "theoretical_loss": 3.4219118657494674,
      "tokens_seen": 2060451840
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000186965670512578,
      "loss": 2.5623,
      "theoretical_loss": 3.421902844184825,
      "tokens_seen": 2060517376
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018694912048420654,
      "loss": 2.5212,
      "theoretical_loss": 3.421893822987454,
      "tokens_seen": 2060582912
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018693257045583505,
      "loss": 2.3234,
      "theoretical_loss": 3.421884802157327,
      "tokens_seen": 2060648448
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001869160204274636,
      "loss": 2.3057,
      "theoretical_loss": 3.4218757816944185,
      "tokens_seen": 2060713984
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001868994703990921,
      "loss": 2.6832,
      "theoretical_loss": 3.4218667615987,
      "tokens_seen": 2060779520
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018688292037072064,
      "loss": 2.5782,
      "theoretical_loss": 3.421857741870147,
      "tokens_seen": 2060845056
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018686637034234915,
      "loss": 2.052,
      "theoretical_loss": 3.4218487225087317,
      "tokens_seen": 2060910592
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001868498203139777,
      "loss": 2.4887,
      "theoretical_loss": 3.4218397035144275,
      "tokens_seen": 2060976128
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001868332702856062,
      "loss": 2.687,
      "theoretical_loss": 3.4218306848872073,
      "tokens_seen": 2061041664
    },
    {
      "epoch": 0.74,
      "objective/train/avg_token_score": 0.013521144166588783,
      "objective/train/avg_weight": 0.984105110168457,
      "objective/train/docs_used": 1166735,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.398362636566162,
      "objective/train/std_weight": 0.07943621277809143,
      "objective/train/theoretical_loss": 3.421821666627046,
      "objective/train/tokens_used": 2081567200,
      "theoretical_loss": 3.421821666627046,
      "tokens_seen": 2061107200
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018681672025723474,
      "loss": 2.7449,
      "theoretical_loss": 3.421821666627046,
      "tokens_seen": 2061107200
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018680017022886325,
      "loss": 2.4363,
      "theoretical_loss": 3.4218126487339156,
      "tokens_seen": 2061172736
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018678362020049178,
      "loss": 2.3939,
      "theoretical_loss": 3.4218036312077906,
      "tokens_seen": 2061238272
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001867670701721203,
      "loss": 2.0286,
      "theoretical_loss": 3.4217946140486433,
      "tokens_seen": 2061303808
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018675052014374883,
      "loss": 2.5563,
      "theoretical_loss": 3.4217855972564477,
      "tokens_seen": 2061369344
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018673397011537734,
      "loss": 2.5758,
      "theoretical_loss": 3.4217765808311773,
      "tokens_seen": 2061434880
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018671742008700588,
      "loss": 2.6452,
      "theoretical_loss": 3.4217675647728054,
      "tokens_seen": 2061500416
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001867008700586344,
      "loss": 2.5156,
      "theoretical_loss": 3.4217585490813045,
      "tokens_seen": 2061565952
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018668432003026293,
      "loss": 2.6503,
      "theoretical_loss": 3.4217495337566497,
      "tokens_seen": 2061631488
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018666777000189144,
      "loss": 2.4293,
      "theoretical_loss": 3.4217405187988135,
      "tokens_seen": 2061697024
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018665121997351998,
      "loss": 2.6937,
      "theoretical_loss": 3.421731504207769,
      "tokens_seen": 2061762560
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018663466994514849,
      "loss": 2.6073,
      "theoretical_loss": 3.42172248998349,
      "tokens_seen": 2061828096
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018661811991677702,
      "loss": 2.541,
      "theoretical_loss": 3.42171347612595,
      "tokens_seen": 2061893632
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018660156988840553,
      "loss": 2.6406,
      "theoretical_loss": 3.4217044626351223,
      "tokens_seen": 2061959168
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018658501986003407,
      "loss": 2.6799,
      "theoretical_loss": 3.42169544951098,
      "tokens_seen": 2062024704
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018656846983166258,
      "loss": 2.3712,
      "theoretical_loss": 3.4216864367534976,
      "tokens_seen": 2062090240
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018655191980329107,
      "loss": 2.5774,
      "theoretical_loss": 3.4216774243626475,
      "tokens_seen": 2062155776
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001865353697749196,
      "loss": 2.5101,
      "theoretical_loss": 3.4216684123384034,
      "tokens_seen": 2062221312
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001865188197465481,
      "loss": 2.3026,
      "theoretical_loss": 3.421659400680739,
      "tokens_seen": 2062286848
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018650226971817665,
      "loss": 2.8471,
      "theoretical_loss": 3.421650389389627,
      "tokens_seen": 2062352384
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018648571968980516,
      "loss": 2.6924,
      "theoretical_loss": 3.4216413784650417,
      "tokens_seen": 2062417920
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001864691696614337,
      "loss": 2.6003,
      "theoretical_loss": 3.4216323679069562,
      "tokens_seen": 2062483456
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001864526196330622,
      "loss": 2.6764,
      "theoretical_loss": 3.421623357715344,
      "tokens_seen": 2062548992
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018643606960469075,
      "loss": 2.7059,
      "theoretical_loss": 3.421614347890178,
      "tokens_seen": 2062614528
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018641951957631926,
      "loss": 2.6492,
      "theoretical_loss": 3.4216053384314327,
      "tokens_seen": 2062680064
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001864029695479478,
      "loss": 2.225,
      "theoretical_loss": 3.4215963293390805,
      "tokens_seen": 2062745600
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001863864195195763,
      "loss": 2.5646,
      "theoretical_loss": 3.4215873206130962,
      "tokens_seen": 2062811136
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018636986949120484,
      "loss": 2.671,
      "theoretical_loss": 3.4215783122534518,
      "tokens_seen": 2062876672
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018635331946283335,
      "loss": 2.4598,
      "theoretical_loss": 3.421569304260122,
      "tokens_seen": 2062942208
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001863367694344619,
      "loss": 2.6929,
      "theoretical_loss": 3.421560296633079,
      "tokens_seen": 2063007744
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001863202194060904,
      "loss": 2.6778,
      "theoretical_loss": 3.4215512893722972,
      "tokens_seen": 2063073280
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018630366937771894,
      "loss": 2.2596,
      "theoretical_loss": 3.4215422824777497,
      "tokens_seen": 2063138816
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018628711934934745,
      "loss": 2.5002,
      "theoretical_loss": 3.42153327594941,
      "tokens_seen": 2063204352
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000186270569320976,
      "loss": 2.4897,
      "theoretical_loss": 3.421524269787252,
      "tokens_seen": 2063269888
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001862540192926045,
      "loss": 2.6467,
      "theoretical_loss": 3.4215152639912487,
      "tokens_seen": 2063335424
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018623746926423303,
      "loss": 2.7901,
      "theoretical_loss": 3.421506258561374,
      "tokens_seen": 2063400960
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018622091923586155,
      "loss": 2.5586,
      "theoretical_loss": 3.421497253497601,
      "tokens_seen": 2063466496
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018620436920749008,
      "loss": 2.57,
      "theoretical_loss": 3.4214882487999034,
      "tokens_seen": 2063532032
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001861878191791186,
      "loss": 2.5561,
      "theoretical_loss": 3.4214792444682542,
      "tokens_seen": 2063597568
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018617126915074713,
      "loss": 2.5393,
      "theoretical_loss": 3.421470240502628,
      "tokens_seen": 2063663104
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018615471912237564,
      "loss": 2.6488,
      "theoretical_loss": 3.4214612369029966,
      "tokens_seen": 2063728640
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018613816909400418,
      "loss": 2.6819,
      "theoretical_loss": 3.421452233669335,
      "tokens_seen": 2063794176
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001861216190656327,
      "loss": 2.5017,
      "theoretical_loss": 3.421443230801616,
      "tokens_seen": 2063859712
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018610506903726123,
      "loss": 2.3794,
      "theoretical_loss": 3.421434228299814,
      "tokens_seen": 2063925248
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018608851900888974,
      "loss": 2.5141,
      "theoretical_loss": 3.421425226163901,
      "tokens_seen": 2063990784
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018607196898051827,
      "loss": 2.7991,
      "theoretical_loss": 3.4214162243938517,
      "tokens_seen": 2064056320
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018605541895214676,
      "loss": 2.6375,
      "theoretical_loss": 3.4214072229896395,
      "tokens_seen": 2064121856
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018603886892377527,
      "loss": 2.4138,
      "theoretical_loss": 3.4213982219512373,
      "tokens_seen": 2064187392
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001860223188954038,
      "loss": 2.4946,
      "theoretical_loss": 3.421389221278619,
      "tokens_seen": 2064252928
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018600576886703232,
      "loss": 2.553,
      "theoretical_loss": 3.4213802209717583,
      "tokens_seen": 2064318464
    },
    {
      "epoch": 0.74,
      "objective/train/avg_token_score": 0.009319281205534935,
      "objective/train/avg_weight": 0.9890382289886475,
      "objective/train/docs_used": 1168682,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9240710735321045,
      "objective/train/std_weight": 0.05417429655790329,
      "objective/train/theoretical_loss": 3.4213712210306286,
      "objective/train/tokens_used": 2084844000,
      "theoretical_loss": 3.4213712210306286,
      "tokens_seen": 2064384000
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018598921883866085,
      "loss": 2.5868,
      "theoretical_loss": 3.4213712210306286,
      "tokens_seen": 2064384000
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018597266881028936,
      "loss": 2.2788,
      "theoretical_loss": 3.421362221455203,
      "tokens_seen": 2064449536
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001859561187819179,
      "loss": 2.6376,
      "theoretical_loss": 3.4213532222454557,
      "tokens_seen": 2064515072
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001859395687535464,
      "loss": 2.5886,
      "theoretical_loss": 3.42134422340136,
      "tokens_seen": 2064580608
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018592301872517495,
      "loss": 2.5522,
      "theoretical_loss": 3.421335224922889,
      "tokens_seen": 2064646144
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018590646869680346,
      "loss": 2.5276,
      "theoretical_loss": 3.4213262268100166,
      "tokens_seen": 2064711680
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000185889918668432,
      "loss": 2.3843,
      "theoretical_loss": 3.4213172290627165,
      "tokens_seen": 2064777216
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001858733686400605,
      "loss": 2.2521,
      "theoretical_loss": 3.421308231680962,
      "tokens_seen": 2064842752
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018585681861168905,
      "loss": 2.3508,
      "theoretical_loss": 3.421299234664727,
      "tokens_seen": 2064908288
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018584026858331756,
      "loss": 2.5425,
      "theoretical_loss": 3.421290238013985,
      "tokens_seen": 2064973824
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001858237185549461,
      "loss": 2.5873,
      "theoretical_loss": 3.4212812417287086,
      "tokens_seen": 2065039360
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001858071685265746,
      "loss": 2.647,
      "theoretical_loss": 3.421272245808873,
      "tokens_seen": 2065104896
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018579061849820314,
      "loss": 2.5563,
      "theoretical_loss": 3.42126325025445,
      "tokens_seen": 2065170432
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018577406846983165,
      "loss": 2.6229,
      "theoretical_loss": 3.4212542550654144,
      "tokens_seen": 2065235968
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001857575184414602,
      "loss": 2.6868,
      "theoretical_loss": 3.4212452602417396,
      "tokens_seen": 2065301504
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001857409684130887,
      "loss": 2.3945,
      "theoretical_loss": 3.421236265783399,
      "tokens_seen": 2065367040
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018572441838471724,
      "loss": 2.5879,
      "theoretical_loss": 3.4212272716903653,
      "tokens_seen": 2065432576
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018570786835634575,
      "loss": 2.7442,
      "theoretical_loss": 3.421218277962614,
      "tokens_seen": 2065498112
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018569131832797429,
      "loss": 2.6627,
      "theoretical_loss": 3.421209284600117,
      "tokens_seen": 2065563648
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001856747682996028,
      "loss": 2.6894,
      "theoretical_loss": 3.4212002916028483,
      "tokens_seen": 2065629184
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018565821827123133,
      "loss": 2.4999,
      "theoretical_loss": 3.421191298970782,
      "tokens_seen": 2065694720
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018564166824285984,
      "loss": 2.6784,
      "theoretical_loss": 3.4211823067038907,
      "tokens_seen": 2065760256
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018562511821448838,
      "loss": 2.6212,
      "theoretical_loss": 3.4211733148021493,
      "tokens_seen": 2065825792
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001856085681861169,
      "loss": 2.6343,
      "theoretical_loss": 3.42116432326553,
      "tokens_seen": 2065891328
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018559201815774543,
      "loss": 2.4683,
      "theoretical_loss": 3.421155332094008,
      "tokens_seen": 2065956864
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018557546812937394,
      "loss": 2.6593,
      "theoretical_loss": 3.4211463412875553,
      "tokens_seen": 2066022400
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018555891810100245,
      "loss": 2.5699,
      "theoretical_loss": 3.4211373508461462,
      "tokens_seen": 2066087936
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018554236807263096,
      "loss": 2.5477,
      "theoretical_loss": 3.4211283607697545,
      "tokens_seen": 2066153472
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001855258180442595,
      "loss": 2.5449,
      "theoretical_loss": 3.421119371058354,
      "tokens_seen": 2066219008
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000185509268015888,
      "loss": 2.4159,
      "theoretical_loss": 3.421110381711917,
      "tokens_seen": 2066284544
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018549271798751655,
      "loss": 2.339,
      "theoretical_loss": 3.421101392730418,
      "tokens_seen": 2066350080
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018547616795914506,
      "loss": 2.6242,
      "theoretical_loss": 3.421092404113831,
      "tokens_seen": 2066415616
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001854596179307736,
      "loss": 2.3946,
      "theoretical_loss": 3.421083415862129,
      "tokens_seen": 2066481152
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001854430679024021,
      "loss": 2.6574,
      "theoretical_loss": 3.4210744279752863,
      "tokens_seen": 2066546688
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018542651787403064,
      "loss": 2.5334,
      "theoretical_loss": 3.421065440453276,
      "tokens_seen": 2066612224
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018540996784565915,
      "loss": 2.5591,
      "theoretical_loss": 3.421056453296071,
      "tokens_seen": 2066677760
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001853934178172877,
      "loss": 2.3607,
      "theoretical_loss": 3.4210474665036466,
      "tokens_seen": 2066743296
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001853768677889162,
      "loss": 2.5314,
      "theoretical_loss": 3.421038480075975,
      "tokens_seen": 2066808832
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018536031776054474,
      "loss": 2.4932,
      "theoretical_loss": 3.421029494013031,
      "tokens_seen": 2066874368
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018534376773217325,
      "loss": 2.69,
      "theoretical_loss": 3.421020508314787,
      "tokens_seen": 2066939904
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018532721770380179,
      "loss": 2.6272,
      "theoretical_loss": 3.421011522981217,
      "tokens_seen": 2067005440
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001853106676754303,
      "loss": 2.3983,
      "theoretical_loss": 3.4210025380122953,
      "tokens_seen": 2067070976
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018529411764705883,
      "loss": 2.4426,
      "theoretical_loss": 3.420993553407995,
      "tokens_seen": 2067136512
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018527756761868734,
      "loss": 2.6659,
      "theoretical_loss": 3.42098456916829,
      "tokens_seen": 2067202048
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018526101759031588,
      "loss": 2.6452,
      "theoretical_loss": 3.4209755852931534,
      "tokens_seen": 2067267584
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001852444675619444,
      "loss": 2.4979,
      "theoretical_loss": 3.4209666017825597,
      "tokens_seen": 2067333120
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018522791753357293,
      "loss": 2.5449,
      "theoretical_loss": 3.4209576186364816,
      "tokens_seen": 2067398656
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018521136750520144,
      "loss": 2.6326,
      "theoretical_loss": 3.4209486358548933,
      "tokens_seen": 2067464192
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018519481747682998,
      "loss": 2.7976,
      "theoretical_loss": 3.420939653437769,
      "tokens_seen": 2067529728
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001851782674484585,
      "loss": 2.5217,
      "theoretical_loss": 3.420930671385081,
      "tokens_seen": 2067595264
    },
    {
      "epoch": 0.74,
      "objective/train/avg_token_score": 0.016450561583042145,
      "objective/train/avg_weight": 0.9806646704673767,
      "objective/train/docs_used": 1170870,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.4238853454589844,
      "objective/train/std_weight": 0.09951604157686234,
      "objective/train/theoretical_loss": 3.4209216896968044,
      "objective/train/tokens_used": 2088120800,
      "theoretical_loss": 3.4209216896968044,
      "tokens_seen": 2067660800
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018516171742008703,
      "loss": 2.7094,
      "theoretical_loss": 3.4209216896968044,
      "tokens_seen": 2067660800
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018514516739171554,
      "loss": 2.4633,
      "theoretical_loss": 3.420912708372912,
      "tokens_seen": 2067726336
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018512861736334407,
      "loss": 2.5952,
      "theoretical_loss": 3.4209037274133776,
      "tokens_seen": 2067791872
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018511206733497258,
      "loss": 2.3099,
      "theoretical_loss": 3.420894746818175,
      "tokens_seen": 2067857408
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018509551730660112,
      "loss": 2.5868,
      "theoretical_loss": 3.4208857665872774,
      "tokens_seen": 2067922944
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018507896727822963,
      "loss": 2.539,
      "theoretical_loss": 3.4208767867206595,
      "tokens_seen": 2067988480
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018506241724985812,
      "loss": 2.3976,
      "theoretical_loss": 3.4208678072182943,
      "tokens_seen": 2068054016
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018504586722148665,
      "loss": 2.4497,
      "theoretical_loss": 3.4208588280801555,
      "tokens_seen": 2068119552
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018502931719311516,
      "loss": 2.7234,
      "theoretical_loss": 3.420849849306216,
      "tokens_seen": 2068185088
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001850127671647437,
      "loss": 2.507,
      "theoretical_loss": 3.4208408708964515,
      "tokens_seen": 2068250624
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001849962171363722,
      "loss": 2.2855,
      "theoretical_loss": 3.420831892850834,
      "tokens_seen": 2068316160
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018497966710800075,
      "loss": 2.394,
      "theoretical_loss": 3.420822915169338,
      "tokens_seen": 2068381696
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018496311707962926,
      "loss": 2.5953,
      "theoretical_loss": 3.4208139378519364,
      "tokens_seen": 2068447232
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001849465670512578,
      "loss": 2.5827,
      "theoretical_loss": 3.4208049608986038,
      "tokens_seen": 2068512768
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001849300170228863,
      "loss": 2.3041,
      "theoretical_loss": 3.420795984309313,
      "tokens_seen": 2068578304
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018491346699451485,
      "loss": 2.5104,
      "theoretical_loss": 3.420787008084039,
      "tokens_seen": 2068643840
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018489691696614336,
      "loss": 2.4748,
      "theoretical_loss": 3.4207780322227537,
      "tokens_seen": 2068709376
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001848803669377719,
      "loss": 2.4269,
      "theoretical_loss": 3.4207690567254327,
      "tokens_seen": 2068774912
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001848638169094004,
      "loss": 2.6333,
      "theoretical_loss": 3.4207600815920487,
      "tokens_seen": 2068840448
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018484726688102894,
      "loss": 2.4274,
      "theoretical_loss": 3.4207511068225753,
      "tokens_seen": 2068905984
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018483071685265745,
      "loss": 2.6827,
      "theoretical_loss": 3.4207421324169864,
      "tokens_seen": 2068971520
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000184814166824286,
      "loss": 2.4982,
      "theoretical_loss": 3.4207331583752563,
      "tokens_seen": 2069037056
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001847976167959145,
      "loss": 2.6037,
      "theoretical_loss": 3.4207241846973577,
      "tokens_seen": 2069102592
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018478106676754304,
      "loss": 2.5726,
      "theoretical_loss": 3.420715211383265,
      "tokens_seen": 2069168128
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018476451673917155,
      "loss": 2.369,
      "theoretical_loss": 3.4207062384329516,
      "tokens_seen": 2069233664
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018474796671080009,
      "loss": 2.5642,
      "theoretical_loss": 3.4206972658463917,
      "tokens_seen": 2069299200
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001847314166824286,
      "loss": 2.4933,
      "theoretical_loss": 3.420688293623559,
      "tokens_seen": 2069364736
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018471486665405713,
      "loss": 2.7027,
      "theoretical_loss": 3.4206793217644265,
      "tokens_seen": 2069430272
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018469831662568564,
      "loss": 2.4052,
      "theoretical_loss": 3.420670350268969,
      "tokens_seen": 2069495808
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018468176659731418,
      "loss": 2.5022,
      "theoretical_loss": 3.420661379137159,
      "tokens_seen": 2069561344
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001846652165689427,
      "loss": 2.6359,
      "theoretical_loss": 3.420652408368971,
      "tokens_seen": 2069626880
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018464866654057123,
      "loss": 2.4563,
      "theoretical_loss": 3.4206434379643786,
      "tokens_seen": 2069692416
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018463211651219974,
      "loss": 2.5669,
      "theoretical_loss": 3.420634467923356,
      "tokens_seen": 2069757952
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018461556648382828,
      "loss": 2.4318,
      "theoretical_loss": 3.420625498245876,
      "tokens_seen": 2069823488
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001845990164554568,
      "loss": 2.6858,
      "theoretical_loss": 3.4206165289319133,
      "tokens_seen": 2069889024
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018458246642708527,
      "loss": 2.2194,
      "theoretical_loss": 3.420607559981441,
      "tokens_seen": 2069954560
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001845659163987138,
      "loss": 2.6119,
      "theoretical_loss": 3.420598591394433,
      "tokens_seen": 2070020096
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018454936637034232,
      "loss": 2.6936,
      "theoretical_loss": 3.420589623170864,
      "tokens_seen": 2070085632
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018453281634197086,
      "loss": 2.4405,
      "theoretical_loss": 3.4205806553107063,
      "tokens_seen": 2070151168
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018451626631359937,
      "loss": 2.7386,
      "theoretical_loss": 3.4205716878139345,
      "tokens_seen": 2070216704
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001844997162852279,
      "loss": 2.4325,
      "theoretical_loss": 3.420562720680522,
      "tokens_seen": 2070282240
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018448316625685642,
      "loss": 2.4723,
      "theoretical_loss": 3.420553753910443,
      "tokens_seen": 2070347776
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018446661622848495,
      "loss": 2.4704,
      "theoretical_loss": 3.4205447875036707,
      "tokens_seen": 2070413312
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018445006620011346,
      "loss": 2.4354,
      "theoretical_loss": 3.42053582146018,
      "tokens_seen": 2070478848
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000184433516171742,
      "loss": 2.7287,
      "theoretical_loss": 3.4205268557799435,
      "tokens_seen": 2070544384
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001844169661433705,
      "loss": 2.6667,
      "theoretical_loss": 3.420517890462935,
      "tokens_seen": 2070609920
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018440041611499905,
      "loss": 2.6812,
      "theoretical_loss": 3.420508925509129,
      "tokens_seen": 2070675456
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018438386608662756,
      "loss": 2.4019,
      "theoretical_loss": 3.420499960918499,
      "tokens_seen": 2070740992
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001843673160582561,
      "loss": 2.6576,
      "theoretical_loss": 3.420490996691019,
      "tokens_seen": 2070806528
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001843507660298846,
      "loss": 2.2056,
      "theoretical_loss": 3.420482032826662,
      "tokens_seen": 2070872064
    },
    {
      "epoch": 0.74,
      "objective/train/avg_token_score": 0.030727554112672806,
      "objective/train/avg_weight": 0.9638696908950806,
      "objective/train/docs_used": 1172808,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.482132911682129,
      "objective/train/std_weight": 0.13096143305301666,
      "objective/train/theoretical_loss": 3.4204730693254035,
      "objective/train/tokens_used": 2091397600,
      "theoretical_loss": 3.4204730693254035,
      "tokens_seen": 2070937600
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018433421600151314,
      "loss": 2.5191,
      "theoretical_loss": 3.4204730693254035,
      "tokens_seen": 2070937600
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018431766597314166,
      "loss": 2.4588,
      "theoretical_loss": 3.4204641061872154,
      "tokens_seen": 2071003136
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001843011159447702,
      "loss": 2.4407,
      "theoretical_loss": 3.420455143412072,
      "tokens_seen": 2071068672
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001842845659163987,
      "loss": 2.6263,
      "theoretical_loss": 3.4204461809999476,
      "tokens_seen": 2071134208
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018426801588802724,
      "loss": 2.7924,
      "theoretical_loss": 3.4204372189508163,
      "tokens_seen": 2071199744
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018425146585965575,
      "loss": 2.5894,
      "theoretical_loss": 3.4204282572646507,
      "tokens_seen": 2071265280
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001842349158312843,
      "loss": 2.6468,
      "theoretical_loss": 3.420419295941426,
      "tokens_seen": 2071330816
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018421836580291283,
      "loss": 2.4073,
      "theoretical_loss": 3.4204103349811152,
      "tokens_seen": 2071396352
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018420181577454134,
      "loss": 2.6774,
      "theoretical_loss": 3.420401374383692,
      "tokens_seen": 2071461888
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018418526574616987,
      "loss": 2.6682,
      "theoretical_loss": 3.4203924141491306,
      "tokens_seen": 2071527424
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018416871571779838,
      "loss": 2.4902,
      "theoretical_loss": 3.4203834542774043,
      "tokens_seen": 2071592960
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018415216568942692,
      "loss": 2.6407,
      "theoretical_loss": 3.420374494768488,
      "tokens_seen": 2071658496
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018413561566105543,
      "loss": 2.5899,
      "theoretical_loss": 3.420365535622355,
      "tokens_seen": 2071724032
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018411906563268397,
      "loss": 2.4221,
      "theoretical_loss": 3.4203565768389783,
      "tokens_seen": 2071789568
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018410251560431248,
      "loss": 2.7801,
      "theoretical_loss": 3.420347618418333,
      "tokens_seen": 2071855104
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018408596557594096,
      "loss": 2.5091,
      "theoretical_loss": 3.4203386603603922,
      "tokens_seen": 2071920640
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001840694155475695,
      "loss": 2.7293,
      "theoretical_loss": 3.4203297026651303,
      "tokens_seen": 2071986176
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000184052865519198,
      "loss": 2.8574,
      "theoretical_loss": 3.4203207453325204,
      "tokens_seen": 2072051712
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018403631549082655,
      "loss": 2.6121,
      "theoretical_loss": 3.4203117883625365,
      "tokens_seen": 2072117248
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018401976546245506,
      "loss": 2.8427,
      "theoretical_loss": 3.420302831755153,
      "tokens_seen": 2072182784
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001840032154340836,
      "loss": 2.4913,
      "theoretical_loss": 3.4202938755103434,
      "tokens_seen": 2072248320
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001839866654057121,
      "loss": 2.4991,
      "theoretical_loss": 3.420284919628082,
      "tokens_seen": 2072313856
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018397011537734065,
      "loss": 2.725,
      "theoretical_loss": 3.4202759641083413,
      "tokens_seen": 2072379392
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018395356534896916,
      "loss": 2.5586,
      "theoretical_loss": 3.4202670089510967,
      "tokens_seen": 2072444928
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001839370153205977,
      "loss": 2.6238,
      "theoretical_loss": 3.4202580541563217,
      "tokens_seen": 2072510464
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001839204652922262,
      "loss": 2.5298,
      "theoretical_loss": 3.4202490997239896,
      "tokens_seen": 2072576000
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018390391526385474,
      "loss": 2.541,
      "theoretical_loss": 3.4202401456540747,
      "tokens_seen": 2072641536
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018388736523548325,
      "loss": 2.6504,
      "theoretical_loss": 3.420231191946551,
      "tokens_seen": 2072707072
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001838708152071118,
      "loss": 2.7362,
      "theoretical_loss": 3.420222238601392,
      "tokens_seen": 2072772608
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001838542651787403,
      "loss": 2.7256,
      "theoretical_loss": 3.420213285618572,
      "tokens_seen": 2072838144
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018383771515036884,
      "loss": 2.6848,
      "theoretical_loss": 3.420204332998064,
      "tokens_seen": 2072903680
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018382116512199735,
      "loss": 2.6661,
      "theoretical_loss": 3.420195380739843,
      "tokens_seen": 2072969216
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018380461509362589,
      "loss": 2.2926,
      "theoretical_loss": 3.4201864288438824,
      "tokens_seen": 2073034752
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001837880650652544,
      "loss": 2.4077,
      "theoretical_loss": 3.4201774773101556,
      "tokens_seen": 2073100288
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018377151503688293,
      "loss": 2.4132,
      "theoretical_loss": 3.4201685261386374,
      "tokens_seen": 2073165824
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018375496500851144,
      "loss": 2.6797,
      "theoretical_loss": 3.4201595753293015,
      "tokens_seen": 2073231360
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018373841498013998,
      "loss": 2.5361,
      "theoretical_loss": 3.420150624882121,
      "tokens_seen": 2073296896
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001837218649517685,
      "loss": 2.5915,
      "theoretical_loss": 3.4201416747970708,
      "tokens_seen": 2073362432
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018370531492339703,
      "loss": 2.4994,
      "theoretical_loss": 3.4201327250741245,
      "tokens_seen": 2073427968
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018368876489502554,
      "loss": 2.5223,
      "theoretical_loss": 3.4201237757132557,
      "tokens_seen": 2073493504
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018367221486665408,
      "loss": 2.599,
      "theoretical_loss": 3.420114826714438,
      "tokens_seen": 2073559040
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001836556648382826,
      "loss": 2.5433,
      "theoretical_loss": 3.4201058780776465,
      "tokens_seen": 2073624576
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018363911480991113,
      "loss": 2.7413,
      "theoretical_loss": 3.4200969298028543,
      "tokens_seen": 2073690112
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018362256478153964,
      "loss": 2.5208,
      "theoretical_loss": 3.4200879818900347,
      "tokens_seen": 2073755648
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018360601475316817,
      "loss": 2.7247,
      "theoretical_loss": 3.420079034339163,
      "tokens_seen": 2073821184
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018358946472479666,
      "loss": 2.4477,
      "theoretical_loss": 3.4200700871502123,
      "tokens_seen": 2073886720
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018357291469642517,
      "loss": 2.8223,
      "theoretical_loss": 3.420061140323157,
      "tokens_seen": 2073952256
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001835563646680537,
      "loss": 2.7017,
      "theoretical_loss": 3.4200521938579707,
      "tokens_seen": 2074017792
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018353981463968221,
      "loss": 2.5256,
      "theoretical_loss": 3.420043247754627,
      "tokens_seen": 2074083328
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018352326461131075,
      "loss": 2.5046,
      "theoretical_loss": 3.4200343020131,
      "tokens_seen": 2074148864
    },
    {
      "epoch": 0.74,
      "objective/train/avg_token_score": 0.013477882370352745,
      "objective/train/avg_weight": 0.9841359853744507,
      "objective/train/docs_used": 1174600,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5964999198913574,
      "objective/train/std_weight": 0.08740616589784622,
      "objective/train/theoretical_loss": 3.4200253566333645,
      "objective/train/tokens_used": 2094674400,
      "theoretical_loss": 3.4200253566333645,
      "tokens_seen": 2074214400
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018350671458293926,
      "loss": 2.7158,
      "theoretical_loss": 3.4200253566333645,
      "tokens_seen": 2074214400
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001834901645545678,
      "loss": 2.462,
      "theoretical_loss": 3.4200164116153937,
      "tokens_seen": 2074279936
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001834736145261963,
      "loss": 2.4069,
      "theoretical_loss": 3.420007466959161,
      "tokens_seen": 2074345472
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018345706449782485,
      "loss": 2.7196,
      "theoretical_loss": 3.419998522664641,
      "tokens_seen": 2074411008
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018344051446945336,
      "loss": 2.5577,
      "theoretical_loss": 3.4199895787318084,
      "tokens_seen": 2074476544
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001834239644410819,
      "loss": 2.6492,
      "theoretical_loss": 3.4199806351606354,
      "tokens_seen": 2074542080
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001834074144127104,
      "loss": 2.7851,
      "theoretical_loss": 3.4199716919510976,
      "tokens_seen": 2074607616
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018339086438433894,
      "loss": 2.6649,
      "theoretical_loss": 3.4199627491031674,
      "tokens_seen": 2074673152
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018337431435596745,
      "loss": 2.775,
      "theoretical_loss": 3.41995380661682,
      "tokens_seen": 2074738688
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000183357764327596,
      "loss": 2.684,
      "theoretical_loss": 3.419944864492029,
      "tokens_seen": 2074804224
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001833412142992245,
      "loss": 2.9111,
      "theoretical_loss": 3.4199359227287687,
      "tokens_seen": 2074869760
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018332466427085304,
      "loss": 2.4721,
      "theoretical_loss": 3.4199269813270123,
      "tokens_seen": 2074935296
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018330811424248155,
      "loss": 2.6238,
      "theoretical_loss": 3.419918040286734,
      "tokens_seen": 2075000832
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001832915642141101,
      "loss": 2.6644,
      "theoretical_loss": 3.4199090996079082,
      "tokens_seen": 2075066368
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001832750141857386,
      "loss": 2.4195,
      "theoretical_loss": 3.419900159290509,
      "tokens_seen": 2075131904
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018325846415736714,
      "loss": 2.5478,
      "theoretical_loss": 3.419891219334509,
      "tokens_seen": 2075197440
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018324191412899565,
      "loss": 2.6997,
      "theoretical_loss": 3.419882279739884,
      "tokens_seen": 2075262976
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018322536410062418,
      "loss": 2.6204,
      "theoretical_loss": 3.4198733405066064,
      "tokens_seen": 2075328512
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001832088140722527,
      "loss": 2.6094,
      "theoretical_loss": 3.4198644016346513,
      "tokens_seen": 2075394048
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018319226404388123,
      "loss": 2.5042,
      "theoretical_loss": 3.4198554631239926,
      "tokens_seen": 2075459584
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018317571401550974,
      "loss": 2.5425,
      "theoretical_loss": 3.4198465249746035,
      "tokens_seen": 2075525120
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018315916398713828,
      "loss": 2.6196,
      "theoretical_loss": 3.419837587186459,
      "tokens_seen": 2075590656
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001831426139587668,
      "loss": 2.6504,
      "theoretical_loss": 3.419828649759532,
      "tokens_seen": 2075656192
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018312606393039533,
      "loss": 2.5644,
      "theoretical_loss": 3.419819712693797,
      "tokens_seen": 2075721728
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001831095139020238,
      "loss": 2.4691,
      "theoretical_loss": 3.4198107759892284,
      "tokens_seen": 2075787264
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018309296387365232,
      "loss": 2.6582,
      "theoretical_loss": 3.4198018396457996,
      "tokens_seen": 2075852800
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018307641384528086,
      "loss": 2.5763,
      "theoretical_loss": 3.4197929036634855,
      "tokens_seen": 2075918336
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018305986381690937,
      "loss": 2.396,
      "theoretical_loss": 3.4197839680422586,
      "tokens_seen": 2075983872
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001830433137885379,
      "loss": 2.406,
      "theoretical_loss": 3.419775032782095,
      "tokens_seen": 2076049408
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018302676376016642,
      "loss": 2.5454,
      "theoretical_loss": 3.4197660978829667,
      "tokens_seen": 2076114944
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018301021373179496,
      "loss": 2.2997,
      "theoretical_loss": 3.4197571633448485,
      "tokens_seen": 2076180480
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001829936637034235,
      "loss": 2.5793,
      "theoretical_loss": 3.4197482291677144,
      "tokens_seen": 2076246016
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000182977113675052,
      "loss": 2.679,
      "theoretical_loss": 3.4197392953515386,
      "tokens_seen": 2076311552
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018296056364668054,
      "loss": 2.6798,
      "theoretical_loss": 3.419730361896295,
      "tokens_seen": 2076377088
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018294401361830905,
      "loss": 2.6459,
      "theoretical_loss": 3.4197214288019575,
      "tokens_seen": 2076442624
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001829274635899376,
      "loss": 2.5904,
      "theoretical_loss": 3.4197124960685006,
      "tokens_seen": 2076508160
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001829109135615661,
      "loss": 2.3632,
      "theoretical_loss": 3.419703563695898,
      "tokens_seen": 2076573696
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018289436353319464,
      "loss": 2.5854,
      "theoretical_loss": 3.4196946316841235,
      "tokens_seen": 2076639232
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018287781350482315,
      "loss": 2.5283,
      "theoretical_loss": 3.419685700033151,
      "tokens_seen": 2076704768
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018286126347645168,
      "loss": 2.6331,
      "theoretical_loss": 3.4196767687429555,
      "tokens_seen": 2076770304
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001828447134480802,
      "loss": 2.6541,
      "theoretical_loss": 3.41966783781351,
      "tokens_seen": 2076835840
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018282816341970873,
      "loss": 2.27,
      "theoretical_loss": 3.4196589072447896,
      "tokens_seen": 2076901376
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018281161339133724,
      "loss": 2.6617,
      "theoretical_loss": 3.4196499770367668,
      "tokens_seen": 2076966912
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018279506336296578,
      "loss": 2.8572,
      "theoretical_loss": 3.419641047189417,
      "tokens_seen": 2077032448
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001827785133345943,
      "loss": 2.7926,
      "theoretical_loss": 3.4196321177027142,
      "tokens_seen": 2077097984
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018276196330622283,
      "loss": 2.5914,
      "theoretical_loss": 3.4196231885766317,
      "tokens_seen": 2077163520
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018274541327785134,
      "loss": 2.5963,
      "theoretical_loss": 3.4196142598111443,
      "tokens_seen": 2077229056
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018272886324947988,
      "loss": 2.6621,
      "theoretical_loss": 3.4196053314062254,
      "tokens_seen": 2077294592
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001827123132211084,
      "loss": 2.7221,
      "theoretical_loss": 3.419596403361849,
      "tokens_seen": 2077360128
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018269576319273692,
      "loss": 2.8981,
      "theoretical_loss": 3.4195874756779903,
      "tokens_seen": 2077425664
    },
    {
      "epoch": 0.74,
      "objective/train/avg_token_score": 0.015449879691004753,
      "objective/train/avg_weight": 0.9818115234375,
      "objective/train/docs_used": 1176526,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.672837495803833,
      "objective/train/std_weight": 0.06759706884622574,
      "objective/train/theoretical_loss": 3.4195785483546226,
      "objective/train/tokens_used": 2097951200,
      "theoretical_loss": 3.4195785483546226,
      "tokens_seen": 2077491200
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018267921316436544,
      "loss": 2.4629,
      "theoretical_loss": 3.4195785483546226,
      "tokens_seen": 2077491200
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018266266313599397,
      "loss": 2.7851,
      "theoretical_loss": 3.4195696213917195,
      "tokens_seen": 2077556736
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018264611310762248,
      "loss": 2.5048,
      "theoretical_loss": 3.419560694789256,
      "tokens_seen": 2077622272
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018262956307925102,
      "loss": 2.6729,
      "theoretical_loss": 3.419551768547205,
      "tokens_seen": 2077687808
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001826130130508795,
      "loss": 2.6114,
      "theoretical_loss": 3.419542842665542,
      "tokens_seen": 2077753344
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018259646302250801,
      "loss": 2.7738,
      "theoretical_loss": 3.41953391714424,
      "tokens_seen": 2077818880
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018257991299413655,
      "loss": 4.0316,
      "theoretical_loss": 3.4195249919832738,
      "tokens_seen": 2077884416
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018256336296576506,
      "loss": 2.5715,
      "theoretical_loss": 3.419516067182617,
      "tokens_seen": 2077949952
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001825468129373936,
      "loss": 2.5757,
      "theoretical_loss": 3.419507142742244,
      "tokens_seen": 2078015488
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001825302629090221,
      "loss": 2.6839,
      "theoretical_loss": 3.4194982186621288,
      "tokens_seen": 2078081024
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018251371288065065,
      "loss": 2.5983,
      "theoretical_loss": 3.4194892949422453,
      "tokens_seen": 2078146560
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018249716285227916,
      "loss": 2.6624,
      "theoretical_loss": 3.419480371582568,
      "tokens_seen": 2078212096
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001824806128239077,
      "loss": 2.5622,
      "theoretical_loss": 3.419471448583071,
      "tokens_seen": 2078277632
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001824640627955362,
      "loss": 2.9238,
      "theoretical_loss": 3.419462525943727,
      "tokens_seen": 2078343168
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018244751276716474,
      "loss": 2.6735,
      "theoretical_loss": 3.419453603664512,
      "tokens_seen": 2078408704
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018243096273879325,
      "loss": 2.7244,
      "theoretical_loss": 3.4194446817453996,
      "tokens_seen": 2078474240
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001824144127104218,
      "loss": 2.5721,
      "theoretical_loss": 3.4194357601863636,
      "tokens_seen": 2078539776
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001823978626820503,
      "loss": 2.6761,
      "theoretical_loss": 3.419426838987378,
      "tokens_seen": 2078605312
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018238131265367884,
      "loss": 2.6003,
      "theoretical_loss": 3.419417918148417,
      "tokens_seen": 2078670848
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018236476262530735,
      "loss": 2.5523,
      "theoretical_loss": 3.4194089976694553,
      "tokens_seen": 2078736384
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001823482125969359,
      "loss": 2.522,
      "theoretical_loss": 3.419400077550466,
      "tokens_seen": 2078801920
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001823316625685644,
      "loss": 2.5416,
      "theoretical_loss": 3.4193911577914244,
      "tokens_seen": 2078867456
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018231511254019294,
      "loss": 2.6042,
      "theoretical_loss": 3.419382238392304,
      "tokens_seen": 2078932992
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018229856251182145,
      "loss": 2.5749,
      "theoretical_loss": 3.4193733193530784,
      "tokens_seen": 2078998528
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018228201248344998,
      "loss": 2.5876,
      "theoretical_loss": 3.419364400673723,
      "tokens_seen": 2079064064
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001822654624550785,
      "loss": 2.2271,
      "theoretical_loss": 3.4193554823542107,
      "tokens_seen": 2079129600
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018224891242670703,
      "loss": 2.4565,
      "theoretical_loss": 3.419346564394516,
      "tokens_seen": 2079195136
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018223236239833554,
      "loss": 2.7169,
      "theoretical_loss": 3.419337646794614,
      "tokens_seen": 2079260672
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018221581236996408,
      "loss": 2.5252,
      "theoretical_loss": 3.4193287295544774,
      "tokens_seen": 2079326208
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001821992623415926,
      "loss": 2.6976,
      "theoretical_loss": 3.419319812674081,
      "tokens_seen": 2079391744
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018218271231322113,
      "loss": 2.5897,
      "theoretical_loss": 3.4193108961533993,
      "tokens_seen": 2079457280
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018216616228484964,
      "loss": 2.4798,
      "theoretical_loss": 3.4193019799924063,
      "tokens_seen": 2079522816
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018214961225647818,
      "loss": 2.1101,
      "theoretical_loss": 3.4192930641910753,
      "tokens_seen": 2079588352
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018213306222810669,
      "loss": 2.6327,
      "theoretical_loss": 3.4192841487493815,
      "tokens_seen": 2079653888
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018211651219973517,
      "loss": 2.5648,
      "theoretical_loss": 3.4192752336672987,
      "tokens_seen": 2079719424
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001820999621713637,
      "loss": 2.8049,
      "theoretical_loss": 3.4192663189448007,
      "tokens_seen": 2079784960
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018208341214299222,
      "loss": 2.2931,
      "theoretical_loss": 3.4192574045818622,
      "tokens_seen": 2079850496
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018206686211462076,
      "loss": 2.4754,
      "theoretical_loss": 3.4192484905784575,
      "tokens_seen": 2079916032
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018205031208624927,
      "loss": 2.4378,
      "theoretical_loss": 3.41923957693456,
      "tokens_seen": 2079981568
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001820337620578778,
      "loss": 2.9221,
      "theoretical_loss": 3.4192306636501444,
      "tokens_seen": 2080047104
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018201721202950631,
      "loss": 2.7363,
      "theoretical_loss": 3.4192217507251854,
      "tokens_seen": 2080112640
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018200066200113485,
      "loss": 2.6307,
      "theoretical_loss": 3.419212838159656,
      "tokens_seen": 2080178176
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018198411197276336,
      "loss": 2.4636,
      "theoretical_loss": 3.4192039259535307,
      "tokens_seen": 2080243712
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001819675619443919,
      "loss": 2.4157,
      "theoretical_loss": 3.4191950141067844,
      "tokens_seen": 2080309248
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001819510119160204,
      "loss": 2.5627,
      "theoretical_loss": 3.419186102619391,
      "tokens_seen": 2080374784
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018193446188764895,
      "loss": 2.645,
      "theoretical_loss": 3.419177191491324,
      "tokens_seen": 2080440320
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018191791185927746,
      "loss": 2.5521,
      "theoretical_loss": 3.4191682807225585,
      "tokens_seen": 2080505856
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000181901361830906,
      "loss": 2.3913,
      "theoretical_loss": 3.419159370313068,
      "tokens_seen": 2080571392
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001818848118025345,
      "loss": 2.5696,
      "theoretical_loss": 3.419150460262827,
      "tokens_seen": 2080636928
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018186826177416304,
      "loss": 2.4983,
      "theoretical_loss": 3.41914155057181,
      "tokens_seen": 2080702464
    },
    {
      "epoch": 0.74,
      "objective/train/avg_token_score": 0.01814887672662735,
      "objective/train/avg_weight": 0.9787007570266724,
      "objective/train/docs_used": 1178616,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7187304496765137,
      "objective/train/std_weight": 0.07697233557701111,
      "objective/train/theoretical_loss": 3.419132641239991,
      "objective/train/tokens_used": 2101228000,
      "theoretical_loss": 3.419132641239991,
      "tokens_seen": 2080768000
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018185171174579155,
      "loss": 2.4569,
      "theoretical_loss": 3.419132641239991,
      "tokens_seen": 2080768000
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001818351617174201,
      "loss": 2.6815,
      "theoretical_loss": 3.419123732267344,
      "tokens_seen": 2080833536
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001818186116890486,
      "loss": 2.7227,
      "theoretical_loss": 3.4191148236538433,
      "tokens_seen": 2080899072
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018180206166067714,
      "loss": 2.4244,
      "theoretical_loss": 3.419105915399463,
      "tokens_seen": 2080964608
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018178551163230565,
      "loss": 2.7102,
      "theoretical_loss": 3.4190970075041776,
      "tokens_seen": 2081030144
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001817689616039342,
      "loss": 2.5503,
      "theoretical_loss": 3.4190880999679614,
      "tokens_seen": 2081095680
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001817524115755627,
      "loss": 2.503,
      "theoretical_loss": 3.4190791927907878,
      "tokens_seen": 2081161216
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018173586154719124,
      "loss": 2.6299,
      "theoretical_loss": 3.419070285972632,
      "tokens_seen": 2081226752
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018171931151881975,
      "loss": 2.7483,
      "theoretical_loss": 3.419061379513468,
      "tokens_seen": 2081292288
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018170276149044828,
      "loss": 2.5837,
      "theoretical_loss": 3.4190524734132692,
      "tokens_seen": 2081357824
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001816862114620768,
      "loss": 2.5956,
      "theoretical_loss": 3.419043567672011,
      "tokens_seen": 2081423360
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018166966143370533,
      "loss": 2.6061,
      "theoretical_loss": 3.4190346622896675,
      "tokens_seen": 2081488896
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018165311140533384,
      "loss": 2.2827,
      "theoretical_loss": 3.419025757266212,
      "tokens_seen": 2081554432
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018163656137696238,
      "loss": 2.4201,
      "theoretical_loss": 3.4190168526016196,
      "tokens_seen": 2081619968
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018162001134859086,
      "loss": 2.5606,
      "theoretical_loss": 3.4190079482958637,
      "tokens_seen": 2081685504
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001816034613202194,
      "loss": 2.4389,
      "theoretical_loss": 3.4189990443489195,
      "tokens_seen": 2081751040
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001815869112918479,
      "loss": 2.5768,
      "theoretical_loss": 3.4189901407607612,
      "tokens_seen": 2081816576
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018157036126347645,
      "loss": 2.6328,
      "theoretical_loss": 3.418981237531362,
      "tokens_seen": 2081882112
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018155381123510496,
      "loss": 2.6014,
      "theoretical_loss": 3.4189723346606975,
      "tokens_seen": 2081947648
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001815372612067335,
      "loss": 2.606,
      "theoretical_loss": 3.4189634321487405,
      "tokens_seen": 2082013184
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000181520711178362,
      "loss": 2.5125,
      "theoretical_loss": 3.4189545299954665,
      "tokens_seen": 2082078720
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018150416114999054,
      "loss": 2.6278,
      "theoretical_loss": 3.4189456282008495,
      "tokens_seen": 2082144256
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018148761112161905,
      "loss": 2.5393,
      "theoretical_loss": 3.4189367267648634,
      "tokens_seen": 2082209792
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001814710610932476,
      "loss": 2.6244,
      "theoretical_loss": 3.4189278256874824,
      "tokens_seen": 2082275328
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001814545110648761,
      "loss": 2.636,
      "theoretical_loss": 3.4189189249686813,
      "tokens_seen": 2082340864
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018143796103650464,
      "loss": 2.7393,
      "theoretical_loss": 3.418910024608434,
      "tokens_seen": 2082406400
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018142141100813315,
      "loss": 2.5056,
      "theoretical_loss": 3.4189011246067147,
      "tokens_seen": 2082471936
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001814048609797617,
      "loss": 2.3588,
      "theoretical_loss": 3.418892224963498,
      "tokens_seen": 2082537472
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001813883109513902,
      "loss": 2.5349,
      "theoretical_loss": 3.418883325678758,
      "tokens_seen": 2082603008
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018137176092301874,
      "loss": 2.4015,
      "theoretical_loss": 3.4188744267524687,
      "tokens_seen": 2082668544
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018135521089464725,
      "loss": 2.5804,
      "theoretical_loss": 3.4188655281846048,
      "tokens_seen": 2082734080
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018133866086627578,
      "loss": 2.5527,
      "theoretical_loss": 3.4188566299751404,
      "tokens_seen": 2082799616
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001813221108379043,
      "loss": 2.5911,
      "theoretical_loss": 3.41884773212405,
      "tokens_seen": 2082865152
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018130556080953283,
      "loss": 2.4644,
      "theoretical_loss": 3.418838834631307,
      "tokens_seen": 2082930688
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018128901078116134,
      "loss": 2.5973,
      "theoretical_loss": 3.4188299374968874,
      "tokens_seen": 2082996224
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018127246075278988,
      "loss": 2.4553,
      "theoretical_loss": 3.418821040720764,
      "tokens_seen": 2083061760
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001812559107244184,
      "loss": 2.333,
      "theoretical_loss": 3.418812144302912,
      "tokens_seen": 2083127296
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018123936069604693,
      "loss": 2.7041,
      "theoretical_loss": 3.418803248243305,
      "tokens_seen": 2083192832
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018122281066767544,
      "loss": 2.5386,
      "theoretical_loss": 3.4187943525419175,
      "tokens_seen": 2083258368
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018120626063930398,
      "loss": 2.5096,
      "theoretical_loss": 3.4187854571987244,
      "tokens_seen": 2083323904
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018118971061093249,
      "loss": 2.641,
      "theoretical_loss": 3.418776562213699,
      "tokens_seen": 2083389440
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018117316058256102,
      "loss": 2.7872,
      "theoretical_loss": 3.418767667586817,
      "tokens_seen": 2083454976
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018115661055418953,
      "loss": 2.8038,
      "theoretical_loss": 3.418758773318051,
      "tokens_seen": 2083520512
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018114006052581802,
      "loss": 2.4041,
      "theoretical_loss": 3.4187498794073763,
      "tokens_seen": 2083586048
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018112351049744656,
      "loss": 2.4859,
      "theoretical_loss": 3.4187409858547673,
      "tokens_seen": 2083651584
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018110696046907507,
      "loss": 2.4829,
      "theoretical_loss": 3.418732092660198,
      "tokens_seen": 2083717120
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001810904104407036,
      "loss": 2.4093,
      "theoretical_loss": 3.418723199823643,
      "tokens_seen": 2083782656
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001810738604123321,
      "loss": 2.7381,
      "theoretical_loss": 3.418714307345076,
      "tokens_seen": 2083848192
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018105731038396065,
      "loss": 2.749,
      "theoretical_loss": 3.4187054152244727,
      "tokens_seen": 2083913728
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018104076035558916,
      "loss": 2.6413,
      "theoretical_loss": 3.418696523461806,
      "tokens_seen": 2083979264
    },
    {
      "epoch": 0.74,
      "objective/train/avg_token_score": 0.01385340467095375,
      "objective/train/avg_weight": 0.9837257266044617,
      "objective/train/docs_used": 1180033,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2341468334198,
      "objective/train/std_weight": 0.08297799527645111,
      "objective/train/theoretical_loss": 3.4186876320570505,
      "objective/train/tokens_used": 2104504800,
      "theoretical_loss": 3.4186876320570505,
      "tokens_seen": 2084044800
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001810242103272177,
      "loss": 2.7099,
      "theoretical_loss": 3.4186876320570505,
      "tokens_seen": 2084044800
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001810076602988462,
      "loss": 2.4888,
      "theoretical_loss": 3.4186787410101815,
      "tokens_seen": 2084110336
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018099111027047475,
      "loss": 2.7179,
      "theoretical_loss": 3.418669850321172,
      "tokens_seen": 2084175872
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018097456024210326,
      "loss": 2.6718,
      "theoretical_loss": 3.4186609599899977,
      "tokens_seen": 2084241408
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001809580102137318,
      "loss": 2.5916,
      "theoretical_loss": 3.418652070016632,
      "tokens_seen": 2084306944
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001809414601853603,
      "loss": 2.4465,
      "theoretical_loss": 3.418643180401049,
      "tokens_seen": 2084372480
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018092491015698884,
      "loss": 2.453,
      "theoretical_loss": 3.418634291143224,
      "tokens_seen": 2084438016
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018090836012861735,
      "loss": 2.415,
      "theoretical_loss": 3.418625402243131,
      "tokens_seen": 2084503552
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001808918101002459,
      "loss": 2.6774,
      "theoretical_loss": 3.4186165137007443,
      "tokens_seen": 2084569088
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001808752600718744,
      "loss": 2.2542,
      "theoretical_loss": 3.418607625516038,
      "tokens_seen": 2084634624
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018085871004350294,
      "loss": 2.6071,
      "theoretical_loss": 3.4185987376889866,
      "tokens_seen": 2084700160
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018084216001513145,
      "loss": 2.6609,
      "theoretical_loss": 3.418589850219565,
      "tokens_seen": 2084765696
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018082560998676,
      "loss": 2.7214,
      "theoretical_loss": 3.418580963107746,
      "tokens_seen": 2084831232
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001808090599583885,
      "loss": 2.6639,
      "theoretical_loss": 3.418572076353506,
      "tokens_seen": 2084896768
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018079250993001703,
      "loss": 2.7405,
      "theoretical_loss": 3.4185631899568185,
      "tokens_seen": 2084962304
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018077595990164555,
      "loss": 2.5596,
      "theoretical_loss": 3.4185543039176576,
      "tokens_seen": 2085027840
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018075940987327408,
      "loss": 2.5648,
      "theoretical_loss": 3.4185454182359982,
      "tokens_seen": 2085093376
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001807428598449026,
      "loss": 2.7087,
      "theoretical_loss": 3.418536532911814,
      "tokens_seen": 2085158912
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018072630981653113,
      "loss": 2.4953,
      "theoretical_loss": 3.41852764794508,
      "tokens_seen": 2085224448
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018070975978815964,
      "loss": 2.6359,
      "theoretical_loss": 3.4185187633357703,
      "tokens_seen": 2085289984
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018069320975978818,
      "loss": 2.221,
      "theoretical_loss": 3.4185098790838593,
      "tokens_seen": 2085355520
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001806766597314167,
      "loss": 2.7022,
      "theoretical_loss": 3.418500995189322,
      "tokens_seen": 2085421056
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018066010970304523,
      "loss": 2.663,
      "theoretical_loss": 3.4184921116521316,
      "tokens_seen": 2085486592
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001806435596746737,
      "loss": 2.4932,
      "theoretical_loss": 3.4184832284722635,
      "tokens_seen": 2085552128
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018062700964630222,
      "loss": 2.5822,
      "theoretical_loss": 3.4184743456496918,
      "tokens_seen": 2085617664
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018061045961793076,
      "loss": 2.5431,
      "theoretical_loss": 3.4184654631843907,
      "tokens_seen": 2085683200
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018059390958955927,
      "loss": 2.3375,
      "theoretical_loss": 3.4184565810763345,
      "tokens_seen": 2085748736
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001805773595611878,
      "loss": 2.5593,
      "theoretical_loss": 3.4184476993254984,
      "tokens_seen": 2085814272
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018056080953281632,
      "loss": 2.6549,
      "theoretical_loss": 3.4184388179318557,
      "tokens_seen": 2085879808
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00018054425950444485,
      "loss": 2.6378,
      "theoretical_loss": 3.4184299368953814,
      "tokens_seen": 2085945344
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018052770947607336,
      "loss": 2.6077,
      "theoretical_loss": 3.4184210562160504,
      "tokens_seen": 2086010880
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001805111594477019,
      "loss": 2.5819,
      "theoretical_loss": 3.4184121758938364,
      "tokens_seen": 2086076416
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001804946094193304,
      "loss": 2.5567,
      "theoretical_loss": 3.418403295928714,
      "tokens_seen": 2086141952
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018047805939095895,
      "loss": 1.7238,
      "theoretical_loss": 3.4183944163206577,
      "tokens_seen": 2086207488
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018046150936258746,
      "loss": 2.4418,
      "theoretical_loss": 3.418385537069642,
      "tokens_seen": 2086273024
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000180444959334216,
      "loss": 2.434,
      "theoretical_loss": 3.4183766581756405,
      "tokens_seen": 2086338560
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001804284093058445,
      "loss": 2.86,
      "theoretical_loss": 3.418367779638629,
      "tokens_seen": 2086404096
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018041185927747305,
      "loss": 2.6345,
      "theoretical_loss": 3.418358901458581,
      "tokens_seen": 2086469632
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018039530924910156,
      "loss": 2.7373,
      "theoretical_loss": 3.4183500236354716,
      "tokens_seen": 2086535168
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001803787592207301,
      "loss": 2.6065,
      "theoretical_loss": 3.4183411461692748,
      "tokens_seen": 2086600704
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001803622091923586,
      "loss": 2.6599,
      "theoretical_loss": 3.418332269059965,
      "tokens_seen": 2086666240
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018034565916398714,
      "loss": 2.3836,
      "theoretical_loss": 3.4183233923075163,
      "tokens_seen": 2086731776
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018032910913561565,
      "loss": 2.8004,
      "theoretical_loss": 3.418314515911904,
      "tokens_seen": 2086797312
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001803125591072442,
      "loss": 2.4518,
      "theoretical_loss": 3.418305639873102,
      "tokens_seen": 2086862848
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001802960090788727,
      "loss": 2.3734,
      "theoretical_loss": 3.4182967641910853,
      "tokens_seen": 2086928384
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018027945905050124,
      "loss": 2.8355,
      "theoretical_loss": 3.4182878888658275,
      "tokens_seen": 2086993920
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018026290902212975,
      "loss": 2.3313,
      "theoretical_loss": 3.418279013897304,
      "tokens_seen": 2087059456
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018024635899375829,
      "loss": 2.601,
      "theoretical_loss": 3.418270139285488,
      "tokens_seen": 2087124992
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018022980896538682,
      "loss": 2.6623,
      "theoretical_loss": 3.418261265030355,
      "tokens_seen": 2087190528
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018021325893701533,
      "loss": 2.4904,
      "theoretical_loss": 3.4182523911318796,
      "tokens_seen": 2087256064
    },
    {
      "epoch": 0.75,
      "objective/train/avg_token_score": 0.011564641259610653,
      "objective/train/avg_weight": 0.9863895773887634,
      "objective/train/docs_used": 1182648,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.4164540767669678,
      "objective/train/std_weight": 0.0746951550245285,
      "objective/train/theoretical_loss": 3.418243517590036,
      "objective/train/tokens_used": 2107781600,
      "theoretical_loss": 3.418243517590036,
      "tokens_seen": 2087321600
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018019670890864387,
      "loss": 2.4974,
      "theoretical_loss": 3.418243517590036,
      "tokens_seen": 2087321600
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018018015888027238,
      "loss": 2.1374,
      "theoretical_loss": 3.418234644404798,
      "tokens_seen": 2087387136
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018016360885190092,
      "loss": 2.5044,
      "theoretical_loss": 3.4182257715761404,
      "tokens_seen": 2087452672
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001801470588235294,
      "loss": 2.3918,
      "theoretical_loss": 3.4182168991040385,
      "tokens_seen": 2087518208
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001801305087951579,
      "loss": 2.4922,
      "theoretical_loss": 3.4182080269884656,
      "tokens_seen": 2087583744
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018011395876678645,
      "loss": 2.6878,
      "theoretical_loss": 3.4181991552293973,
      "tokens_seen": 2087649280
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018009740873841496,
      "loss": 2.4501,
      "theoretical_loss": 3.4181902838268075,
      "tokens_seen": 2087714816
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001800808587100435,
      "loss": 2.5442,
      "theoretical_loss": 3.4181814127806702,
      "tokens_seen": 2087780352
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000180064308681672,
      "loss": 2.5087,
      "theoretical_loss": 3.4181725420909608,
      "tokens_seen": 2087845888
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018004775865330055,
      "loss": 2.799,
      "theoretical_loss": 3.4181636717576533,
      "tokens_seen": 2087911424
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00018003120862492906,
      "loss": 2.6346,
      "theoretical_loss": 3.418154801780722,
      "tokens_seen": 2087976960
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001800146585965576,
      "loss": 2.2877,
      "theoretical_loss": 3.418145932160142,
      "tokens_seen": 2088042496
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001799981085681861,
      "loss": 2.579,
      "theoretical_loss": 3.4181370628958874,
      "tokens_seen": 2088108032
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017998155853981464,
      "loss": 2.6082,
      "theoretical_loss": 3.4181281939879327,
      "tokens_seen": 2088173568
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017996500851144315,
      "loss": 2.638,
      "theoretical_loss": 3.418119325436253,
      "tokens_seen": 2088239104
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001799484584830717,
      "loss": 2.585,
      "theoretical_loss": 3.4181104572408216,
      "tokens_seen": 2088304640
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001799319084547002,
      "loss": 2.5332,
      "theoretical_loss": 3.4181015894016142,
      "tokens_seen": 2088370176
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017991535842632874,
      "loss": 2.6228,
      "theoretical_loss": 3.4180927219186046,
      "tokens_seen": 2088435712
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017989880839795725,
      "loss": 2.5083,
      "theoretical_loss": 3.418083854791768,
      "tokens_seen": 2088501248
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001798822583695858,
      "loss": 2.6009,
      "theoretical_loss": 3.4180749880210777,
      "tokens_seen": 2088566784
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001798657083412143,
      "loss": 2.5823,
      "theoretical_loss": 3.4180661216065094,
      "tokens_seen": 2088632320
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017984915831284283,
      "loss": 2.1741,
      "theoretical_loss": 3.418057255548037,
      "tokens_seen": 2088697856
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017983260828447135,
      "loss": 2.5955,
      "theoretical_loss": 3.4180483898456355,
      "tokens_seen": 2088763392
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017981605825609988,
      "loss": 2.6081,
      "theoretical_loss": 3.418039524499279,
      "tokens_seen": 2088828928
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001797995082277284,
      "loss": 2.4851,
      "theoretical_loss": 3.418030659508942,
      "tokens_seen": 2088894464
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017978295819935693,
      "loss": 2.7988,
      "theoretical_loss": 3.4180217948745995,
      "tokens_seen": 2088960000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017976640817098544,
      "loss": 2.4075,
      "theoretical_loss": 3.418012930596226,
      "tokens_seen": 2089025536
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017974985814261398,
      "loss": 2.5024,
      "theoretical_loss": 3.418004066673795,
      "tokens_seen": 2089091072
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001797333081142425,
      "loss": 2.5784,
      "theoretical_loss": 3.4179952031072824,
      "tokens_seen": 2089156608
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017971675808587103,
      "loss": 2.5474,
      "theoretical_loss": 3.417986339896662,
      "tokens_seen": 2089222144
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017970020805749954,
      "loss": 2.5812,
      "theoretical_loss": 3.4179774770419087,
      "tokens_seen": 2089287680
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017968365802912807,
      "loss": 2.7489,
      "theoretical_loss": 3.417968614542997,
      "tokens_seen": 2089353216
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017966710800075656,
      "loss": 2.5273,
      "theoretical_loss": 3.4179597523999012,
      "tokens_seen": 2089418752
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017965055797238507,
      "loss": 2.5202,
      "theoretical_loss": 3.4179508906125955,
      "tokens_seen": 2089484288
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001796340079440136,
      "loss": 2.6594,
      "theoretical_loss": 3.417942029181056,
      "tokens_seen": 2089549824
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017961745791564212,
      "loss": 2.2803,
      "theoretical_loss": 3.417933168105255,
      "tokens_seen": 2089615360
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017960090788727065,
      "loss": 2.4898,
      "theoretical_loss": 3.417924307385169,
      "tokens_seen": 2089680896
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017958435785889916,
      "loss": 2.7465,
      "theoretical_loss": 3.4179154470207718,
      "tokens_seen": 2089746432
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001795678078305277,
      "loss": 2.6597,
      "theoretical_loss": 3.4179065870120375,
      "tokens_seen": 2089811968
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001795512578021562,
      "loss": 2.4807,
      "theoretical_loss": 3.417897727358942,
      "tokens_seen": 2089877504
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017953470777378475,
      "loss": 2.4548,
      "theoretical_loss": 3.4178888680614583,
      "tokens_seen": 2089943040
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017951815774541326,
      "loss": 2.4646,
      "theoretical_loss": 3.4178800091195622,
      "tokens_seen": 2090008576
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001795016077170418,
      "loss": 2.3238,
      "theoretical_loss": 3.4178711505332275,
      "tokens_seen": 2090074112
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001794850576886703,
      "loss": 2.3301,
      "theoretical_loss": 3.417862292302429,
      "tokens_seen": 2090139648
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017946850766029885,
      "loss": 2.3603,
      "theoretical_loss": 3.4178534344271414,
      "tokens_seen": 2090205184
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017945195763192736,
      "loss": 2.8419,
      "theoretical_loss": 3.4178445769073393,
      "tokens_seen": 2090270720
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001794354076035559,
      "loss": 2.3063,
      "theoretical_loss": 3.417835719742998,
      "tokens_seen": 2090336256
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001794188575751844,
      "loss": 2.5545,
      "theoretical_loss": 3.41782686293409,
      "tokens_seen": 2090401792
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017940230754681294,
      "loss": 2.734,
      "theoretical_loss": 3.417818006480592,
      "tokens_seen": 2090467328
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017938575751844145,
      "loss": 2.5961,
      "theoretical_loss": 3.4178091503824777,
      "tokens_seen": 2090532864
    },
    {
      "epoch": 0.75,
      "objective/train/avg_token_score": 0.017300620675086975,
      "objective/train/avg_weight": 0.9796519875526428,
      "objective/train/docs_used": 1184346,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.552057981491089,
      "objective/train/std_weight": 0.10059157013893127,
      "objective/train/theoretical_loss": 3.4178002946397217,
      "objective/train/tokens_used": 2111058400,
      "theoretical_loss": 3.4178002946397217,
      "tokens_seen": 2090598400
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017936920749007,
      "loss": 2.6769,
      "theoretical_loss": 3.4178002946397217,
      "tokens_seen": 2090598400
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001793526574616985,
      "loss": 2.5929,
      "theoretical_loss": 3.417791439252299,
      "tokens_seen": 2090663936
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017933610743332704,
      "loss": 2.6805,
      "theoretical_loss": 3.4177825842201837,
      "tokens_seen": 2090729472
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017931955740495555,
      "loss": 2.4034,
      "theoretical_loss": 3.4177737295433506,
      "tokens_seen": 2090795008
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017930300737658409,
      "loss": 2.6421,
      "theoretical_loss": 3.417764875221774,
      "tokens_seen": 2090860544
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001792864573482126,
      "loss": 2.3532,
      "theoretical_loss": 3.4177560212554297,
      "tokens_seen": 2090926080
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017926990731984113,
      "loss": 2.8603,
      "theoretical_loss": 3.4177471676442908,
      "tokens_seen": 2090991616
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017925335729146964,
      "loss": 2.7082,
      "theoretical_loss": 3.417738314388333,
      "tokens_seen": 2091057152
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017923680726309818,
      "loss": 2.1797,
      "theoretical_loss": 3.4177294614875304,
      "tokens_seen": 2091122688
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001792202572347267,
      "loss": 2.5794,
      "theoretical_loss": 3.4177206089418575,
      "tokens_seen": 2091188224
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017920370720635523,
      "loss": 2.6371,
      "theoretical_loss": 3.4177117567512894,
      "tokens_seen": 2091253760
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017918715717798374,
      "loss": 2.5774,
      "theoretical_loss": 3.4177029049158003,
      "tokens_seen": 2091319296
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017917060714961222,
      "loss": 2.5602,
      "theoretical_loss": 3.4176940534353655,
      "tokens_seen": 2091384832
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017915405712124076,
      "loss": 2.8672,
      "theoretical_loss": 3.417685202309958,
      "tokens_seen": 2091450368
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017913750709286927,
      "loss": 2.7883,
      "theoretical_loss": 3.417676351539555,
      "tokens_seen": 2091515904
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001791209570644978,
      "loss": 2.5145,
      "theoretical_loss": 3.4176675011241286,
      "tokens_seen": 2091581440
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017910440703612632,
      "loss": 2.5479,
      "theoretical_loss": 3.417658651063655,
      "tokens_seen": 2091646976
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017908785700775486,
      "loss": 2.4298,
      "theoretical_loss": 3.4176498013581087,
      "tokens_seen": 2091712512
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017907130697938337,
      "loss": 2.6432,
      "theoretical_loss": 3.4176409520074635,
      "tokens_seen": 2091778048
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001790547569510119,
      "loss": 2.4105,
      "theoretical_loss": 3.417632103011695,
      "tokens_seen": 2091843584
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017903820692264042,
      "loss": 2.4515,
      "theoretical_loss": 3.417623254370777,
      "tokens_seen": 2091909120
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017902165689426895,
      "loss": 2.5863,
      "theoretical_loss": 3.4176144060846854,
      "tokens_seen": 2091974656
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017900510686589746,
      "loss": 2.616,
      "theoretical_loss": 3.4176055581533933,
      "tokens_seen": 2092040192
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000178988556837526,
      "loss": 2.7407,
      "theoretical_loss": 3.417596710576876,
      "tokens_seen": 2092105728
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017897200680915454,
      "loss": 2.5455,
      "theoretical_loss": 3.417587863355109,
      "tokens_seen": 2092171264
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017895545678078305,
      "loss": 2.6109,
      "theoretical_loss": 3.417579016488066,
      "tokens_seen": 2092236800
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017893890675241159,
      "loss": 2.7521,
      "theoretical_loss": 3.417570169975722,
      "tokens_seen": 2092302336
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001789223567240401,
      "loss": 2.5556,
      "theoretical_loss": 3.417561323818051,
      "tokens_seen": 2092367872
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017890580669566863,
      "loss": 2.6224,
      "theoretical_loss": 3.417552478015029,
      "tokens_seen": 2092433408
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017888925666729714,
      "loss": 2.8399,
      "theoretical_loss": 3.4175436325666295,
      "tokens_seen": 2092498944
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017887270663892568,
      "loss": 2.5869,
      "theoretical_loss": 3.4175347874728277,
      "tokens_seen": 2092564480
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001788561566105542,
      "loss": 2.3085,
      "theoretical_loss": 3.417525942733598,
      "tokens_seen": 2092630016
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017883960658218273,
      "loss": 2.524,
      "theoretical_loss": 3.417517098348916,
      "tokens_seen": 2092695552
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017882305655381124,
      "loss": 2.7817,
      "theoretical_loss": 3.4175082543187547,
      "tokens_seen": 2092761088
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017880650652543978,
      "loss": 2.3962,
      "theoretical_loss": 3.4174994106430905,
      "tokens_seen": 2092826624
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001787899564970683,
      "loss": 2.6301,
      "theoretical_loss": 3.417490567321897,
      "tokens_seen": 2092892160
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017877340646869683,
      "loss": 2.5701,
      "theoretical_loss": 3.4174817243551487,
      "tokens_seen": 2092957696
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017875685644032534,
      "loss": 2.2231,
      "theoretical_loss": 3.4174728817428215,
      "tokens_seen": 2093023232
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017874030641195387,
      "loss": 2.5103,
      "theoretical_loss": 3.417464039484889,
      "tokens_seen": 2093088768
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017872375638358238,
      "loss": 2.7689,
      "theoretical_loss": 3.4174551975813268,
      "tokens_seen": 2093154304
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017870720635521092,
      "loss": 2.4117,
      "theoretical_loss": 3.417446356032109,
      "tokens_seen": 2093219840
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017869065632683943,
      "loss": 2.6576,
      "theoretical_loss": 3.4174375148372103,
      "tokens_seen": 2093285376
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017867410629846792,
      "loss": 2.7063,
      "theoretical_loss": 3.4174286739966053,
      "tokens_seen": 2093350912
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017865755627009645,
      "loss": 2.5286,
      "theoretical_loss": 3.4174198335102695,
      "tokens_seen": 2093416448
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017864100624172496,
      "loss": 2.3177,
      "theoretical_loss": 3.417410993378177,
      "tokens_seen": 2093481984
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001786244562133535,
      "loss": 2.4439,
      "theoretical_loss": 3.4174021536003023,
      "tokens_seen": 2093547520
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000178607906184982,
      "loss": 2.6343,
      "theoretical_loss": 3.4173933141766204,
      "tokens_seen": 2093613056
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017859135615661055,
      "loss": 2.8224,
      "theoretical_loss": 3.417384475107106,
      "tokens_seen": 2093678592
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017857480612823906,
      "loss": 2.6332,
      "theoretical_loss": 3.417375636391734,
      "tokens_seen": 2093744128
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001785582560998676,
      "loss": 2.4896,
      "theoretical_loss": 3.417366798030479,
      "tokens_seen": 2093809664
    },
    {
      "epoch": 0.75,
      "objective/train/avg_token_score": 0.012418299913406372,
      "objective/train/avg_weight": 0.9853829145431519,
      "objective/train/docs_used": 1186411,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.620692253112793,
      "objective/train/std_weight": 0.07029148191213608,
      "objective/train/theoretical_loss": 3.4173579600233155,
      "objective/train/tokens_used": 2114335200,
      "theoretical_loss": 3.4173579600233155,
      "tokens_seen": 2093875200
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001785417060714961,
      "loss": 2.5601,
      "theoretical_loss": 3.4173579600233155,
      "tokens_seen": 2093875200
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017852515604312465,
      "loss": 2.649,
      "theoretical_loss": 3.4173491223702186,
      "tokens_seen": 2093940736
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017850860601475316,
      "loss": 2.5371,
      "theoretical_loss": 3.417340285071163,
      "tokens_seen": 2094006272
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001784920559863817,
      "loss": 2.7661,
      "theoretical_loss": 3.4173314481261228,
      "tokens_seen": 2094071808
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001784755059580102,
      "loss": 2.5013,
      "theoretical_loss": 3.4173226115350737,
      "tokens_seen": 2094137344
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017845895592963874,
      "loss": 2.4968,
      "theoretical_loss": 3.41731377529799,
      "tokens_seen": 2094202880
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017844240590126725,
      "loss": 2.4068,
      "theoretical_loss": 3.417304939414846,
      "tokens_seen": 2094268416
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001784258558728958,
      "loss": 2.4237,
      "theoretical_loss": 3.4172961038856173,
      "tokens_seen": 2094333952
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001784093058445243,
      "loss": 2.564,
      "theoretical_loss": 3.417287268710278,
      "tokens_seen": 2094399488
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017839275581615284,
      "loss": 2.028,
      "theoretical_loss": 3.4172784338888036,
      "tokens_seen": 2094465024
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017837620578778135,
      "loss": 2.3751,
      "theoretical_loss": 3.417269599421168,
      "tokens_seen": 2094530560
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017835965575940989,
      "loss": 2.5214,
      "theoretical_loss": 3.417260765307346,
      "tokens_seen": 2094596096
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001783431057310384,
      "loss": 2.3729,
      "theoretical_loss": 3.417251931547313,
      "tokens_seen": 2094661632
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017832655570266693,
      "loss": 2.4328,
      "theoretical_loss": 3.417243098141043,
      "tokens_seen": 2094727168
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017831000567429544,
      "loss": 2.6909,
      "theoretical_loss": 3.4172342650885117,
      "tokens_seen": 2094792704
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017829345564592398,
      "loss": 2.6452,
      "theoretical_loss": 3.417225432389693,
      "tokens_seen": 2094858240
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001782769056175525,
      "loss": 2.7615,
      "theoretical_loss": 3.4172166000445623,
      "tokens_seen": 2094923776
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017826035558918103,
      "loss": 2.6283,
      "theoretical_loss": 3.417207768053094,
      "tokens_seen": 2094989312
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017824380556080954,
      "loss": 2.1574,
      "theoretical_loss": 3.4171989364152626,
      "tokens_seen": 2095054848
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017822725553243808,
      "loss": 2.4965,
      "theoretical_loss": 3.417190105131044,
      "tokens_seen": 2095120384
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001782107055040666,
      "loss": 2.6067,
      "theoretical_loss": 3.4171812742004115,
      "tokens_seen": 2095185920
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017819415547569507,
      "loss": 2.8534,
      "theoretical_loss": 3.417172443623341,
      "tokens_seen": 2095251456
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001781776054473236,
      "loss": 2.5203,
      "theoretical_loss": 3.417163613399807,
      "tokens_seen": 2095316992
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017816105541895212,
      "loss": 2.6061,
      "theoretical_loss": 3.417154783529784,
      "tokens_seen": 2095382528
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017814450539058066,
      "loss": 2.878,
      "theoretical_loss": 3.4171459540132467,
      "tokens_seen": 2095448064
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017812795536220917,
      "loss": 2.5497,
      "theoretical_loss": 3.4171371248501705,
      "tokens_seen": 2095513600
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001781114053338377,
      "loss": 2.6744,
      "theoretical_loss": 3.41712829604053,
      "tokens_seen": 2095579136
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017809485530546622,
      "loss": 2.1622,
      "theoretical_loss": 3.4171194675842997,
      "tokens_seen": 2095644672
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017807830527709475,
      "loss": 2.4647,
      "theoretical_loss": 3.4171106394814545,
      "tokens_seen": 2095710208
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017806175524872326,
      "loss": 2.7201,
      "theoretical_loss": 3.41710181173197,
      "tokens_seen": 2095775744
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001780452052203518,
      "loss": 2.5904,
      "theoretical_loss": 3.4170929843358193,
      "tokens_seen": 2095841280
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001780286551919803,
      "loss": 2.738,
      "theoretical_loss": 3.4170841572929787,
      "tokens_seen": 2095906816
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017801210516360885,
      "loss": 2.5022,
      "theoretical_loss": 3.4170753306034225,
      "tokens_seen": 2095972352
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017799555513523736,
      "loss": 2.6586,
      "theoretical_loss": 3.417066504267125,
      "tokens_seen": 2096037888
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001779790051068659,
      "loss": 2.4348,
      "theoretical_loss": 3.4170576782840625,
      "tokens_seen": 2096103424
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001779624550784944,
      "loss": 2.6198,
      "theoretical_loss": 3.417048852654208,
      "tokens_seen": 2096168960
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017794590505012294,
      "loss": 2.7303,
      "theoretical_loss": 3.4170400273775376,
      "tokens_seen": 2096234496
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017792935502175146,
      "loss": 2.5063,
      "theoretical_loss": 3.4170312024540257,
      "tokens_seen": 2096300032
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017791280499338,
      "loss": 2.5175,
      "theoretical_loss": 3.4170223778836473,
      "tokens_seen": 2096365568
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001778962549650085,
      "loss": 2.4708,
      "theoretical_loss": 3.4170135536663766,
      "tokens_seen": 2096431104
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017787970493663704,
      "loss": 2.649,
      "theoretical_loss": 3.4170047298021893,
      "tokens_seen": 2096496640
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017786315490826555,
      "loss": 2.6823,
      "theoretical_loss": 3.41699590629106,
      "tokens_seen": 2096562176
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001778466048798941,
      "loss": 2.6644,
      "theoretical_loss": 3.4169870831329625,
      "tokens_seen": 2096627712
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001778300548515226,
      "loss": 2.574,
      "theoretical_loss": 3.4169782603278733,
      "tokens_seen": 2096693248
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017781350482315114,
      "loss": 2.5989,
      "theoretical_loss": 3.4169694378757662,
      "tokens_seen": 2096758784
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017779695479477965,
      "loss": 2.4344,
      "theoretical_loss": 3.4169606157766164,
      "tokens_seen": 2096824320
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017778040476640818,
      "loss": 2.8027,
      "theoretical_loss": 3.416951794030399,
      "tokens_seen": 2096889856
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001777638547380367,
      "loss": 2.6415,
      "theoretical_loss": 3.416942972637088,
      "tokens_seen": 2096955392
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017774730470966523,
      "loss": 2.0826,
      "theoretical_loss": 3.4169341515966583,
      "tokens_seen": 2097020928
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017773075468129374,
      "loss": 2.5521,
      "theoretical_loss": 3.416925330909086,
      "tokens_seen": 2097086464
    },
    {
      "epoch": 0.75,
      "objective/train/avg_token_score": 0.011482696048915386,
      "objective/train/avg_weight": 0.9864901900291443,
      "objective/train/docs_used": 1188334,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.851832866668701,
      "objective/train/std_weight": 0.05062809959053993,
      "objective/train/theoretical_loss": 3.416916510574345,
      "objective/train/tokens_used": 2117612000,
      "theoretical_loss": 3.416916510574345,
      "tokens_seen": 2097152000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017771420465292228,
      "loss": 2.551,
      "theoretical_loss": 3.416916510574345,
      "tokens_seen": 2097152000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017769765462455076,
      "loss": 2.5321,
      "theoretical_loss": 3.41690769059241,
      "tokens_seen": 2097217536
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001776811045961793,
      "loss": 2.51,
      "theoretical_loss": 3.416898870963257,
      "tokens_seen": 2097283072
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001776645545678078,
      "loss": 2.5169,
      "theoretical_loss": 3.416890051686859,
      "tokens_seen": 2097348608
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017764800453943635,
      "loss": 2.6249,
      "theoretical_loss": 3.4168812327631928,
      "tokens_seen": 2097414144
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017763145451106486,
      "loss": 2.5383,
      "theoretical_loss": 3.416872414192232,
      "tokens_seen": 2097479680
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001776149044826934,
      "loss": 2.5498,
      "theoretical_loss": 3.416863595973952,
      "tokens_seen": 2097545216
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001775983544543219,
      "loss": 2.443,
      "theoretical_loss": 3.4168547781083274,
      "tokens_seen": 2097610752
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017758180442595045,
      "loss": 2.5748,
      "theoretical_loss": 3.4168459605953334,
      "tokens_seen": 2097676288
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017756525439757896,
      "loss": 2.678,
      "theoretical_loss": 3.416837143434945,
      "tokens_seen": 2097741824
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001775487043692075,
      "loss": 2.2797,
      "theoretical_loss": 3.416828326627136,
      "tokens_seen": 2097807360
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000177532154340836,
      "loss": 2.824,
      "theoretical_loss": 3.416819510171883,
      "tokens_seen": 2097872896
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017751560431246454,
      "loss": 2.5823,
      "theoretical_loss": 3.4168106940691594,
      "tokens_seen": 2097938432
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017749905428409305,
      "loss": 2.4418,
      "theoretical_loss": 3.4168018783189407,
      "tokens_seen": 2098003968
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001774825042557216,
      "loss": 2.3578,
      "theoretical_loss": 3.4167930629212018,
      "tokens_seen": 2098069504
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001774659542273501,
      "loss": 2.5097,
      "theoretical_loss": 3.416784247875918,
      "tokens_seen": 2098135040
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017744940419897864,
      "loss": 2.6196,
      "theoretical_loss": 3.416775433183063,
      "tokens_seen": 2098200576
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017743285417060715,
      "loss": 2.5649,
      "theoretical_loss": 3.4167666188426127,
      "tokens_seen": 2098266112
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017741630414223569,
      "loss": 2.617,
      "theoretical_loss": 3.4167578048545417,
      "tokens_seen": 2098331648
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001773997541138642,
      "loss": 2.6379,
      "theoretical_loss": 3.4167489912188254,
      "tokens_seen": 2098397184
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017738320408549273,
      "loss": 2.5878,
      "theoretical_loss": 3.4167401779354383,
      "tokens_seen": 2098462720
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017736665405712124,
      "loss": 2.2928,
      "theoretical_loss": 3.416731365004355,
      "tokens_seen": 2098528256
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017735010402874978,
      "loss": 2.6364,
      "theoretical_loss": 3.4167225524255507,
      "tokens_seen": 2098593792
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001773335540003783,
      "loss": 2.6743,
      "theoretical_loss": 3.4167137401990004,
      "tokens_seen": 2098659328
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017731700397200683,
      "loss": 2.5841,
      "theoretical_loss": 3.4167049283246786,
      "tokens_seen": 2098724864
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017730045394363534,
      "loss": 2.579,
      "theoretical_loss": 3.4166961168025614,
      "tokens_seen": 2098790400
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017728390391526388,
      "loss": 2.4992,
      "theoretical_loss": 3.4166873056326224,
      "tokens_seen": 2098855936
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001772673538868924,
      "loss": 2.7313,
      "theoretical_loss": 3.416678494814837,
      "tokens_seen": 2098921472
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017725080385852093,
      "loss": 2.3089,
      "theoretical_loss": 3.4166696843491806,
      "tokens_seen": 2098987008
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017723425383014944,
      "loss": 2.5828,
      "theoretical_loss": 3.4166608742356273,
      "tokens_seen": 2099052544
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017721770380177797,
      "loss": 2.6414,
      "theoretical_loss": 3.4166520644741523,
      "tokens_seen": 2099118080
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017720115377340646,
      "loss": 2.4597,
      "theoretical_loss": 3.4166432550647308,
      "tokens_seen": 2099183616
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017718460374503497,
      "loss": 2.4458,
      "theoretical_loss": 3.4166344460073375,
      "tokens_seen": 2099249152
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001771680537166635,
      "loss": 2.7173,
      "theoretical_loss": 3.416625637301948,
      "tokens_seen": 2099314688
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017715150368829201,
      "loss": 2.6974,
      "theoretical_loss": 3.416616828948536,
      "tokens_seen": 2099380224
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017713495365992055,
      "loss": 2.6243,
      "theoretical_loss": 3.4166080209470775,
      "tokens_seen": 2099445760
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017711840363154906,
      "loss": 2.4826,
      "theoretical_loss": 3.416599213297547,
      "tokens_seen": 2099511296
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001771018536031776,
      "loss": 2.6917,
      "theoretical_loss": 3.4165904059999193,
      "tokens_seen": 2099576832
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001770853035748061,
      "loss": 1.9442,
      "theoretical_loss": 3.41658159905417,
      "tokens_seen": 2099642368
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017706875354643465,
      "loss": 2.4836,
      "theoretical_loss": 3.416572792460274,
      "tokens_seen": 2099707904
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017705220351806316,
      "loss": 2.4624,
      "theoretical_loss": 3.4165639862182053,
      "tokens_seen": 2099773440
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001770356534896917,
      "loss": 2.2896,
      "theoretical_loss": 3.4165551803279395,
      "tokens_seen": 2099838976
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001770191034613202,
      "loss": 2.4592,
      "theoretical_loss": 3.416546374789452,
      "tokens_seen": 2099904512
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017700255343294874,
      "loss": 2.4317,
      "theoretical_loss": 3.416537569602717,
      "tokens_seen": 2099970048
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017698600340457725,
      "loss": 2.507,
      "theoretical_loss": 3.41652876476771,
      "tokens_seen": 2100035584
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001769694533762058,
      "loss": 2.5538,
      "theoretical_loss": 3.4165199602844054,
      "tokens_seen": 2100101120
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001769529033478343,
      "loss": 2.3055,
      "theoretical_loss": 3.416511156152779,
      "tokens_seen": 2100166656
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017693635331946284,
      "loss": 2.9406,
      "theoretical_loss": 3.416502352372805,
      "tokens_seen": 2100232192
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017691980329109135,
      "loss": 2.6141,
      "theoretical_loss": 3.416493548944459,
      "tokens_seen": 2100297728
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001769032532627199,
      "loss": 2.2619,
      "theoretical_loss": 3.4164847458677157,
      "tokens_seen": 2100363264
    },
    {
      "epoch": 0.75,
      "objective/train/avg_token_score": 0.036695677787065506,
      "objective/train/avg_weight": 0.9568051695823669,
      "objective/train/docs_used": 1190051,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 8.680294036865234,
      "objective/train/std_weight": 0.15527209639549255,
      "objective/train/theoretical_loss": 3.4164759431425495,
      "objective/train/tokens_used": 2120888800,
      "theoretical_loss": 3.4164759431425495,
      "tokens_seen": 2100428800
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001768867032343484,
      "loss": 2.152,
      "theoretical_loss": 3.4164759431425495,
      "tokens_seen": 2100428800
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017687015320597694,
      "loss": 2.627,
      "theoretical_loss": 3.4164671407689364,
      "tokens_seen": 2100494336
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017685360317760545,
      "loss": 2.6596,
      "theoretical_loss": 3.4164583387468515,
      "tokens_seen": 2100559872
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017683705314923398,
      "loss": 2.6462,
      "theoretical_loss": 3.4164495370762684,
      "tokens_seen": 2100625408
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001768205031208625,
      "loss": 2.6111,
      "theoretical_loss": 3.4164407357571633,
      "tokens_seen": 2100690944
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017680395309249103,
      "loss": 2.5006,
      "theoretical_loss": 3.416431934789511,
      "tokens_seen": 2100756480
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017678740306411954,
      "loss": 2.5273,
      "theoretical_loss": 3.416423134173286,
      "tokens_seen": 2100822016
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017677085303574808,
      "loss": 2.4309,
      "theoretical_loss": 3.416414333908464,
      "tokens_seen": 2100887552
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001767543030073766,
      "loss": 1.8611,
      "theoretical_loss": 3.4164055339950195,
      "tokens_seen": 2100953088
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017673775297900513,
      "loss": 2.2949,
      "theoretical_loss": 3.4163967344329276,
      "tokens_seen": 2101018624
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017672120295063364,
      "loss": 2.6398,
      "theoretical_loss": 3.416387935222164,
      "tokens_seen": 2101084160
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017670465292226212,
      "loss": 2.6664,
      "theoretical_loss": 3.4163791363627025,
      "tokens_seen": 2101149696
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017668810289389066,
      "loss": 2.416,
      "theoretical_loss": 3.4163703378545187,
      "tokens_seen": 2101215232
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017667155286551917,
      "loss": 2.3377,
      "theoretical_loss": 3.4163615396975877,
      "tokens_seen": 2101280768
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001766550028371477,
      "loss": 2.6535,
      "theoretical_loss": 3.4163527418918846,
      "tokens_seen": 2101346304
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017663845280877622,
      "loss": 2.533,
      "theoretical_loss": 3.4163439444373846,
      "tokens_seen": 2101411840
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017662190278040476,
      "loss": 2.3707,
      "theoretical_loss": 3.416335147334062,
      "tokens_seen": 2101477376
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017660535275203327,
      "loss": 2.7455,
      "theoretical_loss": 3.416326350581892,
      "tokens_seen": 2101542912
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001765888027236618,
      "loss": 2.2649,
      "theoretical_loss": 3.416317554180851,
      "tokens_seen": 2101608448
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017657225269529031,
      "loss": 2.3734,
      "theoretical_loss": 3.4163087581309117,
      "tokens_seen": 2101673984
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017655570266691885,
      "loss": 2.5645,
      "theoretical_loss": 3.416299962432051,
      "tokens_seen": 2101739520
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017653915263854736,
      "loss": 2.7727,
      "theoretical_loss": 3.4162911670842435,
      "tokens_seen": 2101805056
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001765226026101759,
      "loss": 2.2022,
      "theoretical_loss": 3.4162823720874633,
      "tokens_seen": 2101870592
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001765060525818044,
      "loss": 2.7439,
      "theoretical_loss": 3.416273577441687,
      "tokens_seen": 2101936128
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017648950255343295,
      "loss": 2.404,
      "theoretical_loss": 3.416264783146888,
      "tokens_seen": 2102001664
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017647295252506146,
      "loss": 2.6565,
      "theoretical_loss": 3.4162559892030426,
      "tokens_seen": 2102067200
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017645640249669,
      "loss": 2.5897,
      "theoretical_loss": 3.4162471956101257,
      "tokens_seen": 2102132736
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001764398524683185,
      "loss": 2.7595,
      "theoretical_loss": 3.416238402368112,
      "tokens_seen": 2102198272
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017642330243994704,
      "loss": 2.7893,
      "theoretical_loss": 3.4162296094769764,
      "tokens_seen": 2102263808
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017640675241157555,
      "loss": 2.545,
      "theoretical_loss": 3.4162208169366943,
      "tokens_seen": 2102329344
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001763902023832041,
      "loss": 2.6977,
      "theoretical_loss": 3.416212024747241,
      "tokens_seen": 2102394880
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001763736523548326,
      "loss": 2.6665,
      "theoretical_loss": 3.416203232908591,
      "tokens_seen": 2102460416
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017635710232646114,
      "loss": 2.601,
      "theoretical_loss": 3.4161944414207195,
      "tokens_seen": 2102525952
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017634055229808965,
      "loss": 2.8736,
      "theoretical_loss": 3.4161856502836017,
      "tokens_seen": 2102591488
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001763240022697182,
      "loss": 2.5408,
      "theoretical_loss": 3.416176859497213,
      "tokens_seen": 2102657024
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001763074522413467,
      "loss": 2.6048,
      "theoretical_loss": 3.416168069061528,
      "tokens_seen": 2102722560
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017629090221297524,
      "loss": 2.5055,
      "theoretical_loss": 3.416159278976522,
      "tokens_seen": 2102788096
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017627435218460375,
      "loss": 2.5438,
      "theoretical_loss": 3.4161504892421695,
      "tokens_seen": 2102853632
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017625780215623228,
      "loss": 2.533,
      "theoretical_loss": 3.4161416998584464,
      "tokens_seen": 2102919168
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017624125212786082,
      "loss": 2.475,
      "theoretical_loss": 3.4161329108253278,
      "tokens_seen": 2102984704
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001762247020994893,
      "loss": 2.5167,
      "theoretical_loss": 3.4161241221427883,
      "tokens_seen": 2103050240
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017620815207111781,
      "loss": 2.8745,
      "theoretical_loss": 3.416115333810803,
      "tokens_seen": 2103115776
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017619160204274635,
      "loss": 2.5248,
      "theoretical_loss": 3.416106545829347,
      "tokens_seen": 2103181312
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017617505201437486,
      "loss": 2.4591,
      "theoretical_loss": 3.4160977581983962,
      "tokens_seen": 2103246848
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001761585019860034,
      "loss": 2.5438,
      "theoretical_loss": 3.4160889709179245,
      "tokens_seen": 2103312384
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001761419519576319,
      "loss": 2.7075,
      "theoretical_loss": 3.4160801839879076,
      "tokens_seen": 2103377920
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017612540192926045,
      "loss": 2.6672,
      "theoretical_loss": 3.4160713974083206,
      "tokens_seen": 2103443456
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017610885190088896,
      "loss": 2.6558,
      "theoretical_loss": 3.4160626111791386,
      "tokens_seen": 2103508992
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001760923018725175,
      "loss": 2.4508,
      "theoretical_loss": 3.4160538253003367,
      "tokens_seen": 2103574528
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000176075751844146,
      "loss": 2.6447,
      "theoretical_loss": 3.41604503977189,
      "tokens_seen": 2103640064
    },
    {
      "epoch": 0.75,
      "objective/train/avg_token_score": 0.02637224830687046,
      "objective/train/avg_weight": 0.9689986705780029,
      "objective/train/docs_used": 1192176,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.784062385559082,
      "objective/train/std_weight": 0.11042698472738266,
      "objective/train/theoretical_loss": 3.4160362545937732,
      "objective/train/tokens_used": 2124165600,
      "theoretical_loss": 3.4160362545937732,
      "tokens_seen": 2103705600
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017605920181577454,
      "loss": 2.6176,
      "theoretical_loss": 3.4160362545937732,
      "tokens_seen": 2103705600
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017604265178740305,
      "loss": 2.7013,
      "theoretical_loss": 3.4160274697659623,
      "tokens_seen": 2103771136
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001760261017590316,
      "loss": 2.5807,
      "theoretical_loss": 3.4160186852884316,
      "tokens_seen": 2103836672
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001760095517306601,
      "loss": 2.7344,
      "theoretical_loss": 3.416009901161157,
      "tokens_seen": 2103902208
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017599300170228864,
      "loss": 2.6279,
      "theoretical_loss": 3.4160011173841127,
      "tokens_seen": 2103967744
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017597645167391715,
      "loss": 2.5844,
      "theoretical_loss": 3.4159923339572744,
      "tokens_seen": 2104033280
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001759599016455457,
      "loss": 2.6275,
      "theoretical_loss": 3.4159835508806173,
      "tokens_seen": 2104098816
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001759433516171742,
      "loss": 2.5213,
      "theoretical_loss": 3.4159747681541166,
      "tokens_seen": 2104164352
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017592680158880274,
      "loss": 2.5841,
      "theoretical_loss": 3.4159659857777465,
      "tokens_seen": 2104229888
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017591025156043125,
      "loss": 2.6865,
      "theoretical_loss": 3.415957203751484,
      "tokens_seen": 2104295424
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017589370153205978,
      "loss": 2.4177,
      "theoretical_loss": 3.415948422075302,
      "tokens_seen": 2104360960
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001758771515036883,
      "loss": 2.5118,
      "theoretical_loss": 3.415939640749177,
      "tokens_seen": 2104426496
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017586060147531683,
      "loss": 2.7315,
      "theoretical_loss": 3.415930859773084,
      "tokens_seen": 2104492032
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017584405144694534,
      "loss": 2.7813,
      "theoretical_loss": 3.415922079146998,
      "tokens_seen": 2104557568
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017582750141857388,
      "loss": 2.5711,
      "theoretical_loss": 3.415913298870894,
      "tokens_seen": 2104623104
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001758109513902024,
      "loss": 2.3731,
      "theoretical_loss": 3.4159045189447474,
      "tokens_seen": 2104688640
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017579440136183093,
      "loss": 2.6716,
      "theoretical_loss": 3.4158957393685334,
      "tokens_seen": 2104754176
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017577785133345944,
      "loss": 2.7617,
      "theoretical_loss": 3.415886960142227,
      "tokens_seen": 2104819712
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017576130130508798,
      "loss": 2.6705,
      "theoretical_loss": 3.4158781812658034,
      "tokens_seen": 2104885248
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017574475127671649,
      "loss": 2.4208,
      "theoretical_loss": 3.415869402739238,
      "tokens_seen": 2104950784
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017572820124834497,
      "loss": 2.4677,
      "theoretical_loss": 3.415860624562505,
      "tokens_seen": 2105016320
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001757116512199735,
      "loss": 2.4002,
      "theoretical_loss": 3.415851846735581,
      "tokens_seen": 2105081856
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017569510119160202,
      "loss": 2.6906,
      "theoretical_loss": 3.4158430692584405,
      "tokens_seen": 2105147392
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017567855116323056,
      "loss": 1.7524,
      "theoretical_loss": 3.4158342921310583,
      "tokens_seen": 2105212928
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017566200113485907,
      "loss": 2.2543,
      "theoretical_loss": 3.4158255153534096,
      "tokens_seen": 2105278464
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001756454511064876,
      "loss": 2.6716,
      "theoretical_loss": 3.4158167389254706,
      "tokens_seen": 2105344000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017562890107811611,
      "loss": 2.6597,
      "theoretical_loss": 3.4158079628472153,
      "tokens_seen": 2105409536
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017561235104974465,
      "loss": 2.7656,
      "theoretical_loss": 3.415799187118619,
      "tokens_seen": 2105475072
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017559580102137316,
      "loss": 2.2641,
      "theoretical_loss": 3.415790411739658,
      "tokens_seen": 2105540608
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001755792509930017,
      "loss": 2.503,
      "theoretical_loss": 3.4157816367103067,
      "tokens_seen": 2105606144
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001755627009646302,
      "loss": 2.4702,
      "theoretical_loss": 3.41577286203054,
      "tokens_seen": 2105671680
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017554615093625875,
      "loss": 2.8605,
      "theoretical_loss": 3.4157640877003335,
      "tokens_seen": 2105737216
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017552960090788726,
      "loss": 2.5789,
      "theoretical_loss": 3.415755313719662,
      "tokens_seen": 2105802752
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001755130508795158,
      "loss": 2.1419,
      "theoretical_loss": 3.4157465400885014,
      "tokens_seen": 2105868288
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001754965008511443,
      "loss": 2.8509,
      "theoretical_loss": 3.415737766806826,
      "tokens_seen": 2105933824
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017547995082277284,
      "loss": 2.0778,
      "theoretical_loss": 3.4157289938746125,
      "tokens_seen": 2105999360
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017546340079440135,
      "loss": 2.5473,
      "theoretical_loss": 3.415720221291834,
      "tokens_seen": 2106064896
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001754468507660299,
      "loss": 2.6136,
      "theoretical_loss": 3.4157114490584677,
      "tokens_seen": 2106130432
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001754303007376584,
      "loss": 2.3419,
      "theoretical_loss": 3.4157026771744876,
      "tokens_seen": 2106195968
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017541375070928694,
      "loss": 2.7628,
      "theoretical_loss": 3.415693905639869,
      "tokens_seen": 2106261504
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017539720068091545,
      "loss": 2.3304,
      "theoretical_loss": 3.4156851344545878,
      "tokens_seen": 2106327040
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000175380650652544,
      "loss": 2.5208,
      "theoretical_loss": 3.4156763636186183,
      "tokens_seen": 2106392576
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001753641006241725,
      "loss": 2.3873,
      "theoretical_loss": 3.4156675931319365,
      "tokens_seen": 2106458112
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017534755059580104,
      "loss": 2.6479,
      "theoretical_loss": 3.4156588229945175,
      "tokens_seen": 2106523648
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017533100056742955,
      "loss": 2.4823,
      "theoretical_loss": 3.415650053206336,
      "tokens_seen": 2106589184
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017531445053905808,
      "loss": 2.5688,
      "theoretical_loss": 3.4156412837673678,
      "tokens_seen": 2106654720
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001752979005106866,
      "loss": 2.6761,
      "theoretical_loss": 3.415632514677588,
      "tokens_seen": 2106720256
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017528135048231513,
      "loss": 2.3506,
      "theoretical_loss": 3.415623745936972,
      "tokens_seen": 2106785792
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017526480045394364,
      "loss": 2.3695,
      "theoretical_loss": 3.415614977545494,
      "tokens_seen": 2106851328
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017524825042557218,
      "loss": 2.5906,
      "theoretical_loss": 3.4156062095031308,
      "tokens_seen": 2106916864
    },
    {
      "epoch": 0.75,
      "objective/train/avg_token_score": 0.025028975680470467,
      "objective/train/avg_weight": 0.9705329537391663,
      "objective/train/docs_used": 1193800,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.344899892807007,
      "objective/train/std_weight": 0.10967812687158585,
      "objective/train/theoretical_loss": 3.4155974418098562,
      "objective/train/tokens_used": 2127442400,
      "theoretical_loss": 3.4155974418098562,
      "tokens_seen": 2106982400
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017523170039720066,
      "loss": 2.6762,
      "theoretical_loss": 3.4155974418098562,
      "tokens_seen": 2106982400
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017521515036882917,
      "loss": 2.3846,
      "theoretical_loss": 3.4155886744656465,
      "tokens_seen": 2107047936
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001751986003404577,
      "loss": 2.6156,
      "theoretical_loss": 3.4155799074704762,
      "tokens_seen": 2107113472
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017518205031208622,
      "loss": 2.5486,
      "theoretical_loss": 3.415571140824321,
      "tokens_seen": 2107179008
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017516550028371476,
      "loss": 2.4048,
      "theoretical_loss": 3.4155623745271564,
      "tokens_seen": 2107244544
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017514895025534327,
      "loss": 2.7867,
      "theoretical_loss": 3.415553608578957,
      "tokens_seen": 2107310080
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001751324002269718,
      "loss": 2.4237,
      "theoretical_loss": 3.4155448429796986,
      "tokens_seen": 2107375616
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017511585019860032,
      "loss": 2.602,
      "theoretical_loss": 3.415536077729356,
      "tokens_seen": 2107441152
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017509930017022885,
      "loss": 2.6497,
      "theoretical_loss": 3.415527312827905,
      "tokens_seen": 2107506688
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017508275014185736,
      "loss": 2.4249,
      "theoretical_loss": 3.4155185482753203,
      "tokens_seen": 2107572224
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001750662001134859,
      "loss": 2.2487,
      "theoretical_loss": 3.4155097840715776,
      "tokens_seen": 2107637760
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001750496500851144,
      "loss": 2.725,
      "theoretical_loss": 3.415501020216652,
      "tokens_seen": 2107703296
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017503310005674295,
      "loss": 2.3656,
      "theoretical_loss": 3.415492256710518,
      "tokens_seen": 2107768832
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017501655002837146,
      "loss": 2.4682,
      "theoretical_loss": 3.4154834935531526,
      "tokens_seen": 2107834368
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000175,
      "loss": 2.5018,
      "theoretical_loss": 3.41547473074453,
      "tokens_seen": 2107899904
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017498344997162854,
      "loss": 2.385,
      "theoretical_loss": 3.415465968284625,
      "tokens_seen": 2107965440
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017496689994325705,
      "loss": 2.6996,
      "theoretical_loss": 3.4154572061734143,
      "tokens_seen": 2108030976
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017495034991488558,
      "loss": 2.5862,
      "theoretical_loss": 3.4154484444108717,
      "tokens_seen": 2108096512
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001749337998865141,
      "loss": 1.9703,
      "theoretical_loss": 3.4154396829969733,
      "tokens_seen": 2108162048
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017491724985814263,
      "loss": 2.709,
      "theoretical_loss": 3.4154309219316943,
      "tokens_seen": 2108227584
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017490069982977114,
      "loss": 2.6587,
      "theoretical_loss": 3.4154221612150097,
      "tokens_seen": 2108293120
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017488414980139968,
      "loss": 2.5534,
      "theoretical_loss": 3.4154134008468953,
      "tokens_seen": 2108358656
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017486759977302816,
      "loss": 2.7632,
      "theoretical_loss": 3.415404640827326,
      "tokens_seen": 2108424192
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001748510497446567,
      "loss": 2.6798,
      "theoretical_loss": 3.4153958811562775,
      "tokens_seen": 2108489728
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001748344997162852,
      "loss": 2.5453,
      "theoretical_loss": 3.4153871218337244,
      "tokens_seen": 2108555264
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017481794968791375,
      "loss": 2.7275,
      "theoretical_loss": 3.4153783628596424,
      "tokens_seen": 2108620800
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017480139965954226,
      "loss": 2.5614,
      "theoretical_loss": 3.415369604234007,
      "tokens_seen": 2108686336
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001747848496311708,
      "loss": 2.4654,
      "theoretical_loss": 3.4153608459567932,
      "tokens_seen": 2108751872
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001747682996027993,
      "loss": 2.6455,
      "theoretical_loss": 3.415352088027977,
      "tokens_seen": 2108817408
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017475174957442784,
      "loss": 2.3195,
      "theoretical_loss": 3.4153433304475325,
      "tokens_seen": 2108882944
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017473519954605635,
      "loss": 2.601,
      "theoretical_loss": 3.415334573215436,
      "tokens_seen": 2108948480
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001747186495176849,
      "loss": 2.7333,
      "theoretical_loss": 3.4153258163316624,
      "tokens_seen": 2109014016
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001747020994893134,
      "loss": 2.5946,
      "theoretical_loss": 3.4153170597961875,
      "tokens_seen": 2109079552
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017468554946094194,
      "loss": 2.5964,
      "theoretical_loss": 3.4153083036089855,
      "tokens_seen": 2109145088
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017466899943257045,
      "loss": 2.6698,
      "theoretical_loss": 3.415299547770033,
      "tokens_seen": 2109210624
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000174652449404199,
      "loss": 2.5831,
      "theoretical_loss": 3.415290792279305,
      "tokens_seen": 2109276160
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001746358993758275,
      "loss": 2.6005,
      "theoretical_loss": 3.415282037136776,
      "tokens_seen": 2109341696
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000174619349347456,
      "loss": 2.5274,
      "theoretical_loss": 3.4152732823424223,
      "tokens_seen": 2109407232
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017460279931908455,
      "loss": 2.6236,
      "theoretical_loss": 3.415264527896219,
      "tokens_seen": 2109472768
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017458624929071306,
      "loss": 2.33,
      "theoretical_loss": 3.4152557737981413,
      "tokens_seen": 2109538304
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001745696992623416,
      "loss": 2.4023,
      "theoretical_loss": 3.4152470200481644,
      "tokens_seen": 2109603840
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001745531492339701,
      "loss": 2.8317,
      "theoretical_loss": 3.4152382666462637,
      "tokens_seen": 2109669376
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017453659920559864,
      "loss": 2.6347,
      "theoretical_loss": 3.4152295135924153,
      "tokens_seen": 2109734912
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017452004917722715,
      "loss": 2.6975,
      "theoretical_loss": 3.4152207608865934,
      "tokens_seen": 2109800448
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001745034991488557,
      "loss": 2.7079,
      "theoretical_loss": 3.4152120085287745,
      "tokens_seen": 2109865984
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001744869491204842,
      "loss": 2.7729,
      "theoretical_loss": 3.4152032565189323,
      "tokens_seen": 2109931520
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017447039909211274,
      "loss": 2.7642,
      "theoretical_loss": 3.415194504857044,
      "tokens_seen": 2109997056
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017445384906374125,
      "loss": 2.5957,
      "theoretical_loss": 3.415185753543084,
      "tokens_seen": 2110062592
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001744372990353698,
      "loss": 2.3702,
      "theoretical_loss": 3.415177002577028,
      "tokens_seen": 2110128128
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001744207490069983,
      "loss": 2.8509,
      "theoretical_loss": 3.4151682519588507,
      "tokens_seen": 2110193664
    },
    {
      "epoch": 0.75,
      "objective/train/avg_token_score": 0.03273293003439903,
      "objective/train/avg_weight": 0.9614900946617126,
      "objective/train/docs_used": 1195692,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.764364719390869,
      "objective/train/std_weight": 0.11532924324274063,
      "objective/train/theoretical_loss": 3.415159501688528,
      "objective/train/tokens_used": 2130719200,
      "theoretical_loss": 3.415159501688528,
      "tokens_seen": 2110259200
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017440419897862683,
      "loss": 2.732,
      "theoretical_loss": 3.415159501688528,
      "tokens_seen": 2110259200
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017438764895025535,
      "loss": 2.6792,
      "theoretical_loss": 3.4151507517660358,
      "tokens_seen": 2110324736
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017437109892188386,
      "loss": 2.6087,
      "theoretical_loss": 3.4151420021913483,
      "tokens_seen": 2110390272
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001743545488935124,
      "loss": 2.5283,
      "theoretical_loss": 3.415133252964442,
      "tokens_seen": 2110455808
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001743379988651409,
      "loss": 2.5609,
      "theoretical_loss": 3.4151245040852913,
      "tokens_seen": 2110521344
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017432144883676944,
      "loss": 2.6787,
      "theoretical_loss": 3.415115755553872,
      "tokens_seen": 2110586880
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017430489880839795,
      "loss": 2.5817,
      "theoretical_loss": 3.4151070073701595,
      "tokens_seen": 2110652416
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001742883487800265,
      "loss": 2.6261,
      "theoretical_loss": 3.4150982595341297,
      "tokens_seen": 2110717952
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000174271798751655,
      "loss": 2.6368,
      "theoretical_loss": 3.415089512045757,
      "tokens_seen": 2110783488
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017425524872328354,
      "loss": 2.648,
      "theoretical_loss": 3.415080764905017,
      "tokens_seen": 2110849024
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017423869869491205,
      "loss": 2.7619,
      "theoretical_loss": 3.415072018111886,
      "tokens_seen": 2110914560
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017422214866654059,
      "loss": 2.4305,
      "theoretical_loss": 3.4150632716663387,
      "tokens_seen": 2110980096
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001742055986381691,
      "loss": 2.7392,
      "theoretical_loss": 3.41505452556835,
      "tokens_seen": 2111045632
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017418904860979763,
      "loss": 2.5228,
      "theoretical_loss": 3.4150457798178966,
      "tokens_seen": 2111111168
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017417249858142614,
      "loss": 2.6319,
      "theoretical_loss": 3.4150370344149525,
      "tokens_seen": 2111176704
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017415594855305468,
      "loss": 2.3375,
      "theoretical_loss": 3.415028289359494,
      "tokens_seen": 2111242240
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017413939852468316,
      "loss": 2.6252,
      "theoretical_loss": 3.4150195446514964,
      "tokens_seen": 2111307776
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001741228484963117,
      "loss": 2.4615,
      "theoretical_loss": 3.415010800290935,
      "tokens_seen": 2111373312
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001741062984679402,
      "loss": 2.591,
      "theoretical_loss": 3.4150020562777854,
      "tokens_seen": 2111438848
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017408974843956875,
      "loss": 2.5953,
      "theoretical_loss": 3.4149933126120224,
      "tokens_seen": 2111504384
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017407319841119726,
      "loss": 2.5274,
      "theoretical_loss": 3.4149845692936216,
      "tokens_seen": 2111569920
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001740566483828258,
      "loss": 2.7433,
      "theoretical_loss": 3.414975826322559,
      "tokens_seen": 2111635456
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001740400983544543,
      "loss": 2.4676,
      "theoretical_loss": 3.4149670836988095,
      "tokens_seen": 2111700992
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017402354832608285,
      "loss": 2.4452,
      "theoretical_loss": 3.4149583414223486,
      "tokens_seen": 2111766528
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017400699829771136,
      "loss": 2.6352,
      "theoretical_loss": 3.414949599493152,
      "tokens_seen": 2111832064
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001739904482693399,
      "loss": 2.531,
      "theoretical_loss": 3.414940857911195,
      "tokens_seen": 2111897600
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001739738982409684,
      "loss": 2.7883,
      "theoretical_loss": 3.4149321166764532,
      "tokens_seen": 2111963136
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017395734821259694,
      "loss": 2.3282,
      "theoretical_loss": 3.4149233757889013,
      "tokens_seen": 2112028672
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017394079818422545,
      "loss": 2.5409,
      "theoretical_loss": 3.4149146352485156,
      "tokens_seen": 2112094208
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000173924248155854,
      "loss": 2.3176,
      "theoretical_loss": 3.414905895055271,
      "tokens_seen": 2112159744
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001739076981274825,
      "loss": 2.8306,
      "theoretical_loss": 3.414897155209143,
      "tokens_seen": 2112225280
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000173891148099111,
      "loss": 2.8059,
      "theoretical_loss": 3.4148884157101076,
      "tokens_seen": 2112290816
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017387459807073955,
      "loss": 2.6878,
      "theoretical_loss": 3.4148796765581397,
      "tokens_seen": 2112356352
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017385804804236806,
      "loss": 2.7333,
      "theoretical_loss": 3.414870937753215,
      "tokens_seen": 2112421888
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001738414980139966,
      "loss": 2.7332,
      "theoretical_loss": 3.4148621992953085,
      "tokens_seen": 2112487424
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001738249479856251,
      "loss": 2.5964,
      "theoretical_loss": 3.4148534611843964,
      "tokens_seen": 2112552960
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017380839795725364,
      "loss": 2.8435,
      "theoretical_loss": 3.414844723420453,
      "tokens_seen": 2112618496
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017379184792888215,
      "loss": 2.0172,
      "theoretical_loss": 3.414835986003455,
      "tokens_seen": 2112684032
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001737752979005107,
      "loss": 2.497,
      "theoretical_loss": 3.4148272489333773,
      "tokens_seen": 2112749568
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001737587478721392,
      "loss": 2.7277,
      "theoretical_loss": 3.4148185122101955,
      "tokens_seen": 2112815104
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017374219784376774,
      "loss": 2.6071,
      "theoretical_loss": 3.414809775833885,
      "tokens_seen": 2112880640
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017372564781539625,
      "loss": 2.2645,
      "theoretical_loss": 3.414801039804421,
      "tokens_seen": 2112946176
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001737090977870248,
      "loss": 2.8712,
      "theoretical_loss": 3.4147923041217796,
      "tokens_seen": 2113011712
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001736925477586533,
      "loss": 2.6881,
      "theoretical_loss": 3.4147835687859356,
      "tokens_seen": 2113077248
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017367599773028184,
      "loss": 2.5148,
      "theoretical_loss": 3.414774833796865,
      "tokens_seen": 2113142784
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017365944770191035,
      "loss": 2.722,
      "theoretical_loss": 3.414766099154543,
      "tokens_seen": 2113208320
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017364289767353886,
      "loss": 2.5077,
      "theoretical_loss": 3.414757364858945,
      "tokens_seen": 2113273856
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001736263476451674,
      "loss": 2.5952,
      "theoretical_loss": 3.414748630910047,
      "tokens_seen": 2113339392
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001736097976167959,
      "loss": 2.3244,
      "theoretical_loss": 3.414739897307824,
      "tokens_seen": 2113404928
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017359324758842444,
      "loss": 2.6372,
      "theoretical_loss": 3.4147311640522515,
      "tokens_seen": 2113470464
    },
    {
      "epoch": 0.75,
      "objective/train/avg_token_score": 0.014581385999917984,
      "objective/train/avg_weight": 0.9828374981880188,
      "objective/train/docs_used": 1197624,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 6.380300998687744,
      "objective/train/std_weight": 0.10778986662626266,
      "objective/train/theoretical_loss": 3.4147224311433053,
      "objective/train/tokens_used": 2133996000,
      "theoretical_loss": 3.4147224311433053,
      "tokens_seen": 2113536000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017357669756005295,
      "loss": 2.2985,
      "theoretical_loss": 3.4147224311433053,
      "tokens_seen": 2113536000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001735601475316815,
      "loss": 2.3946,
      "theoretical_loss": 3.41471369858096,
      "tokens_seen": 2113601536
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017354359750331,
      "loss": 2.7305,
      "theoretical_loss": 3.4147049663651927,
      "tokens_seen": 2113667072
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017352704747493854,
      "loss": 2.6868,
      "theoretical_loss": 3.4146962344959775,
      "tokens_seen": 2113732608
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00017351049744656705,
      "loss": 2.6561,
      "theoretical_loss": 3.4146875029732904,
      "tokens_seen": 2113798144
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001734939474181956,
      "loss": 2.6872,
      "theoretical_loss": 3.4146787717971074,
      "tokens_seen": 2113863680
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001734773973898241,
      "loss": 2.8024,
      "theoretical_loss": 3.414670040967403,
      "tokens_seen": 2113929216
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017346084736145263,
      "loss": 2.4675,
      "theoretical_loss": 3.4146613104841532,
      "tokens_seen": 2113994752
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017344429733308115,
      "loss": 2.3829,
      "theoretical_loss": 3.414652580347334,
      "tokens_seen": 2114060288
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017342774730470968,
      "loss": 2.644,
      "theoretical_loss": 3.41464385055692,
      "tokens_seen": 2114125824
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001734111972763382,
      "loss": 2.7828,
      "theoretical_loss": 3.4146351211128874,
      "tokens_seen": 2114191360
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001733946472479667,
      "loss": 2.4503,
      "theoretical_loss": 3.4146263920152116,
      "tokens_seen": 2114256896
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017337809721959521,
      "loss": 2.5186,
      "theoretical_loss": 3.4146176632638676,
      "tokens_seen": 2114322432
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017336154719122375,
      "loss": 2.5852,
      "theoretical_loss": 3.4146089348588315,
      "tokens_seen": 2114387968
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017334499716285226,
      "loss": 2.4822,
      "theoretical_loss": 3.4146002068000785,
      "tokens_seen": 2114453504
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001733284471344808,
      "loss": 2.6374,
      "theoretical_loss": 3.4145914790875844,
      "tokens_seen": 2114519040
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001733118971061093,
      "loss": 2.5795,
      "theoretical_loss": 3.414582751721325,
      "tokens_seen": 2114584576
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017329534707773785,
      "loss": 2.5692,
      "theoretical_loss": 3.4145740247012752,
      "tokens_seen": 2114650112
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017327879704936636,
      "loss": 2.5761,
      "theoretical_loss": 3.414565298027411,
      "tokens_seen": 2114715648
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001732622470209949,
      "loss": 2.7852,
      "theoretical_loss": 3.4145565716997073,
      "tokens_seen": 2114781184
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001732456969926234,
      "loss": 2.6176,
      "theoretical_loss": 3.41454784571814,
      "tokens_seen": 2114846720
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017322914696425194,
      "loss": 2.1822,
      "theoretical_loss": 3.4145391200826856,
      "tokens_seen": 2114912256
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017321259693588045,
      "loss": 2.4283,
      "theoretical_loss": 3.414530394793318,
      "tokens_seen": 2114977792
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000173196046907509,
      "loss": 2.61,
      "theoretical_loss": 3.4145216698500134,
      "tokens_seen": 2115043328
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001731794968791375,
      "loss": 2.5686,
      "theoretical_loss": 3.414512945252748,
      "tokens_seen": 2115108864
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017316294685076604,
      "loss": 2.5629,
      "theoretical_loss": 3.414504221001496,
      "tokens_seen": 2115174400
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017314639682239455,
      "loss": 2.7751,
      "theoretical_loss": 3.414495497096235,
      "tokens_seen": 2115239936
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017312984679402306,
      "loss": 2.7232,
      "theoretical_loss": 3.4144867735369386,
      "tokens_seen": 2115305472
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001731132967656516,
      "loss": 2.684,
      "theoretical_loss": 3.414478050323583,
      "tokens_seen": 2115371008
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001730967467372801,
      "loss": 2.583,
      "theoretical_loss": 3.4144693274561444,
      "tokens_seen": 2115436544
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017308019670890865,
      "loss": 2.8521,
      "theoretical_loss": 3.414460604934597,
      "tokens_seen": 2115502080
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017306364668053716,
      "loss": 2.8475,
      "theoretical_loss": 3.414451882758918,
      "tokens_seen": 2115567616
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001730470966521657,
      "loss": 2.6393,
      "theoretical_loss": 3.414443160929082,
      "tokens_seen": 2115633152
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001730305466237942,
      "loss": 2.7161,
      "theoretical_loss": 3.4144344394450643,
      "tokens_seen": 2115698688
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017301399659542274,
      "loss": 2.7342,
      "theoretical_loss": 3.414425718306841,
      "tokens_seen": 2115764224
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017299744656705125,
      "loss": 2.5543,
      "theoretical_loss": 3.4144169975143885,
      "tokens_seen": 2115829760
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001729808965386798,
      "loss": 2.4987,
      "theoretical_loss": 3.4144082770676807,
      "tokens_seen": 2115895296
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001729643465103083,
      "loss": 2.5865,
      "theoretical_loss": 3.4143995569666936,
      "tokens_seen": 2115960832
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017294779648193684,
      "loss": 2.3432,
      "theoretical_loss": 3.4143908372114042,
      "tokens_seen": 2116026368
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017293124645356535,
      "loss": 2.871,
      "theoretical_loss": 3.4143821178017864,
      "tokens_seen": 2116091904
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017291469642519389,
      "loss": 2.8121,
      "theoretical_loss": 3.4143733987378164,
      "tokens_seen": 2116157440
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001728981463968224,
      "loss": 2.4126,
      "theoretical_loss": 3.41436468001947,
      "tokens_seen": 2116222976
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001728815963684509,
      "loss": 2.4701,
      "theoretical_loss": 3.4143559616467227,
      "tokens_seen": 2116288512
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017286504634007944,
      "loss": 2.5941,
      "theoretical_loss": 3.41434724361955,
      "tokens_seen": 2116354048
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017284849631170795,
      "loss": 2.585,
      "theoretical_loss": 3.4143385259379277,
      "tokens_seen": 2116419584
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001728319462833365,
      "loss": 2.684,
      "theoretical_loss": 3.4143298086018308,
      "tokens_seen": 2116485120
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000172815396254965,
      "loss": 2.5982,
      "theoretical_loss": 3.4143210916112356,
      "tokens_seen": 2116550656
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017279884622659354,
      "loss": 2.6115,
      "theoretical_loss": 3.4143123749661175,
      "tokens_seen": 2116616192
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017278229619822205,
      "loss": 2.5014,
      "theoretical_loss": 3.414303658666452,
      "tokens_seen": 2116681728
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001727657461698506,
      "loss": 2.2762,
      "theoretical_loss": 3.414294942712215,
      "tokens_seen": 2116747264
    },
    {
      "epoch": 0.76,
      "objective/train/avg_token_score": 0.027111556380987167,
      "objective/train/avg_weight": 0.9681100845336914,
      "objective/train/docs_used": 1198832,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.5071372985839844,
      "objective/train/std_weight": 0.11869955062866211,
      "objective/train/theoretical_loss": 3.414286227103381,
      "objective/train/tokens_used": 2137272800,
      "theoretical_loss": 3.414286227103381,
      "tokens_seen": 2116812800
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001727491961414791,
      "loss": 2.8765,
      "theoretical_loss": 3.414286227103381,
      "tokens_seen": 2116812800
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017273264611310764,
      "loss": 2.6568,
      "theoretical_loss": 3.4142775118399276,
      "tokens_seen": 2116878336
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017271609608473615,
      "loss": 2.6206,
      "theoretical_loss": 3.414268796921829,
      "tokens_seen": 2116943872
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017269954605636468,
      "loss": 2.7775,
      "theoretical_loss": 3.414260082349061,
      "tokens_seen": 2117009408
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001726829960279932,
      "loss": 2.4956,
      "theoretical_loss": 3.414251368121599,
      "tokens_seen": 2117074944
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017266644599962173,
      "loss": 2.4844,
      "theoretical_loss": 3.4142426542394197,
      "tokens_seen": 2117140480
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017264989597125022,
      "loss": 2.4679,
      "theoretical_loss": 3.414233940702498,
      "tokens_seen": 2117206016
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017263334594287875,
      "loss": 2.6452,
      "theoretical_loss": 3.4142252275108094,
      "tokens_seen": 2117271552
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017261679591450726,
      "loss": 2.7079,
      "theoretical_loss": 3.4142165146643295,
      "tokens_seen": 2117337088
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001726002458861358,
      "loss": 2.5869,
      "theoretical_loss": 3.4142078021630344,
      "tokens_seen": 2117402624
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001725836958577643,
      "loss": 2.4175,
      "theoretical_loss": 3.4141990900068997,
      "tokens_seen": 2117468160
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017256714582939285,
      "loss": 2.608,
      "theoretical_loss": 3.4141903781959004,
      "tokens_seen": 2117533696
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017255059580102136,
      "loss": 2.6293,
      "theoretical_loss": 3.4141816667300127,
      "tokens_seen": 2117599232
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001725340457726499,
      "loss": 2.5065,
      "theoretical_loss": 3.414172955609212,
      "tokens_seen": 2117664768
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001725174957442784,
      "loss": 2.3868,
      "theoretical_loss": 3.414164244833475,
      "tokens_seen": 2117730304
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017250094571590694,
      "loss": 2.581,
      "theoretical_loss": 3.4141555344027754,
      "tokens_seen": 2117795840
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017248439568753546,
      "loss": 2.2852,
      "theoretical_loss": 3.41414682431709,
      "tokens_seen": 2117861376
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000172467845659164,
      "loss": 2.5167,
      "theoretical_loss": 3.414138114576395,
      "tokens_seen": 2117926912
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001724512956307925,
      "loss": 2.6535,
      "theoretical_loss": 3.414129405180665,
      "tokens_seen": 2117992448
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017243474560242104,
      "loss": 2.6792,
      "theoretical_loss": 3.414120696129876,
      "tokens_seen": 2118057984
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017241819557404955,
      "loss": 2.3995,
      "theoretical_loss": 3.4141119874240036,
      "tokens_seen": 2118123520
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017240164554567806,
      "loss": 2.505,
      "theoretical_loss": 3.414103279063024,
      "tokens_seen": 2118189056
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001723850955173066,
      "loss": 2.5551,
      "theoretical_loss": 3.4140945710469124,
      "tokens_seen": 2118254592
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001723685454889351,
      "loss": 2.0758,
      "theoretical_loss": 3.414085863375644,
      "tokens_seen": 2118320128
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017235199546056365,
      "loss": 2.5094,
      "theoretical_loss": 3.4140771560491956,
      "tokens_seen": 2118385664
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017233544543219216,
      "loss": 2.6655,
      "theoretical_loss": 3.4140684490675426,
      "tokens_seen": 2118451200
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001723188954038207,
      "loss": 2.3526,
      "theoretical_loss": 3.41405974243066,
      "tokens_seen": 2118516736
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001723023453754492,
      "loss": 2.6313,
      "theoretical_loss": 3.414051036138524,
      "tokens_seen": 2118582272
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017228579534707774,
      "loss": 2.3633,
      "theoretical_loss": 3.41404233019111,
      "tokens_seen": 2118647808
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017226924531870625,
      "loss": 2.7335,
      "theoretical_loss": 3.4140336245883933,
      "tokens_seen": 2118713344
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001722526952903348,
      "loss": 2.7834,
      "theoretical_loss": 3.414024919330351,
      "tokens_seen": 2118778880
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001722361452619633,
      "loss": 2.4653,
      "theoretical_loss": 3.4140162144169577,
      "tokens_seen": 2118844416
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017221959523359184,
      "loss": 2.4703,
      "theoretical_loss": 3.4140075098481892,
      "tokens_seen": 2118909952
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017220304520522035,
      "loss": 2.8778,
      "theoretical_loss": 3.4139988056240216,
      "tokens_seen": 2118975488
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001721864951768489,
      "loss": 2.5481,
      "theoretical_loss": 3.41399010174443,
      "tokens_seen": 2119041024
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001721699451484774,
      "loss": 2.716,
      "theoretical_loss": 3.4139813982093905,
      "tokens_seen": 2119106560
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001721533951201059,
      "loss": 2.5526,
      "theoretical_loss": 3.413972695018879,
      "tokens_seen": 2119172096
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017213684509173445,
      "loss": 2.7807,
      "theoretical_loss": 3.413963992172871,
      "tokens_seen": 2119237632
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017212029506336296,
      "loss": 2.7492,
      "theoretical_loss": 3.4139552896713417,
      "tokens_seen": 2119303168
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001721037450349915,
      "loss": 2.6988,
      "theoretical_loss": 3.4139465875142676,
      "tokens_seen": 2119368704
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017208719500662,
      "loss": 2.8176,
      "theoretical_loss": 3.413937885701624,
      "tokens_seen": 2119434240
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017207064497824854,
      "loss": 2.74,
      "theoretical_loss": 3.4139291842333863,
      "tokens_seen": 2119499776
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017205409494987705,
      "loss": 2.6018,
      "theoretical_loss": 3.413920483109531,
      "tokens_seen": 2119565312
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001720375449215056,
      "loss": 2.5722,
      "theoretical_loss": 3.413911782330034,
      "tokens_seen": 2119630848
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001720209948931341,
      "loss": 2.7745,
      "theoretical_loss": 3.41390308189487,
      "tokens_seen": 2119696384
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017200444486476264,
      "loss": 2.8081,
      "theoretical_loss": 3.413894381804015,
      "tokens_seen": 2119761920
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017198789483639115,
      "loss": 2.7431,
      "theoretical_loss": 3.413885682057445,
      "tokens_seen": 2119827456
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017197134480801969,
      "loss": 2.3897,
      "theoretical_loss": 3.413876982655135,
      "tokens_seen": 2119892992
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001719547947796482,
      "loss": 2.6644,
      "theoretical_loss": 3.4138682835970626,
      "tokens_seen": 2119958528
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017193824475127673,
      "loss": 2.6196,
      "theoretical_loss": 3.413859584883202,
      "tokens_seen": 2120024064
    },
    {
      "epoch": 0.76,
      "objective/train/avg_token_score": 0.018684787675738335,
      "objective/train/avg_weight": 0.9780055284500122,
      "objective/train/docs_used": 1201065,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.78241229057312,
      "objective/train/std_weight": 0.09794425219297409,
      "objective/train/theoretical_loss": 3.413850886513529,
      "objective/train/tokens_used": 2140549600,
      "theoretical_loss": 3.413850886513529,
      "tokens_seen": 2120089600
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017192169472290522,
      "loss": 2.4837,
      "theoretical_loss": 3.413850886513529,
      "tokens_seen": 2120089600
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017190514469453375,
      "loss": 2.4948,
      "theoretical_loss": 3.4138421884880197,
      "tokens_seen": 2120155136
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017188859466616226,
      "loss": 2.5886,
      "theoretical_loss": 3.41383349080665,
      "tokens_seen": 2120220672
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001718720446377908,
      "loss": 2.7972,
      "theoretical_loss": 3.413824793469395,
      "tokens_seen": 2120286208
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001718554946094193,
      "loss": 2.6011,
      "theoretical_loss": 3.4138160964762307,
      "tokens_seen": 2120351744
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017183894458104785,
      "loss": 2.6012,
      "theoretical_loss": 3.4138073998271334,
      "tokens_seen": 2120417280
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017182239455267636,
      "loss": 2.7612,
      "theoretical_loss": 3.4137987035220787,
      "tokens_seen": 2120482816
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001718058445243049,
      "loss": 2.6399,
      "theoretical_loss": 3.4137900075610412,
      "tokens_seen": 2120548352
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001717892944959334,
      "loss": 2.7082,
      "theoretical_loss": 3.413781311943999,
      "tokens_seen": 2120613888
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017177274446756195,
      "loss": 2.5031,
      "theoretical_loss": 3.413772616670925,
      "tokens_seen": 2120679424
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017175619443919046,
      "loss": 2.4751,
      "theoretical_loss": 3.413763921741797,
      "tokens_seen": 2120744960
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000171739644410819,
      "loss": 2.5283,
      "theoretical_loss": 3.4137552271565905,
      "tokens_seen": 2120810496
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001717230943824475,
      "loss": 2.3699,
      "theoretical_loss": 3.4137465329152805,
      "tokens_seen": 2120876032
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017170654435407604,
      "loss": 2.3491,
      "theoretical_loss": 3.4137378390178434,
      "tokens_seen": 2120941568
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017168999432570455,
      "loss": 2.4804,
      "theoretical_loss": 3.413729145464255,
      "tokens_seen": 2121007104
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017167344429733306,
      "loss": 2.6669,
      "theoretical_loss": 3.4137204522544904,
      "tokens_seen": 2121072640
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001716568942689616,
      "loss": 2.4536,
      "theoretical_loss": 3.413711759388526,
      "tokens_seen": 2121138176
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001716403442405901,
      "loss": 2.5706,
      "theoretical_loss": 3.4137030668663373,
      "tokens_seen": 2121203712
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017162379421221865,
      "loss": 2.7046,
      "theoretical_loss": 3.4136943746879007,
      "tokens_seen": 2121269248
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017160724418384716,
      "loss": 2.7772,
      "theoretical_loss": 3.413685682853191,
      "tokens_seen": 2121334784
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001715906941554757,
      "loss": 2.5311,
      "theoretical_loss": 3.413676991362185,
      "tokens_seen": 2121400320
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001715741441271042,
      "loss": 2.8686,
      "theoretical_loss": 3.413668300214858,
      "tokens_seen": 2121465856
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017155759409873274,
      "loss": 2.7358,
      "theoretical_loss": 3.413659609411185,
      "tokens_seen": 2121531392
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017154104407036125,
      "loss": 2.216,
      "theoretical_loss": 3.4136509189511433,
      "tokens_seen": 2121596928
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001715244940419898,
      "loss": 2.595,
      "theoretical_loss": 3.4136422288347075,
      "tokens_seen": 2121662464
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001715079440136183,
      "loss": 2.7216,
      "theoretical_loss": 3.4136335390618537,
      "tokens_seen": 2121728000
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017149139398524684,
      "loss": 2.3302,
      "theoretical_loss": 3.4136248496325585,
      "tokens_seen": 2121793536
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017147484395687535,
      "loss": 2.5532,
      "theoretical_loss": 3.413616160546797,
      "tokens_seen": 2121859072
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001714582939285039,
      "loss": 2.317,
      "theoretical_loss": 3.4136074718045446,
      "tokens_seen": 2121924608
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001714417439001324,
      "loss": 2.598,
      "theoretical_loss": 3.413598783405778,
      "tokens_seen": 2121990144
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001714251938717609,
      "loss": 2.4636,
      "theoretical_loss": 3.4135900953504725,
      "tokens_seen": 2122055680
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017140864384338945,
      "loss": 2.5465,
      "theoretical_loss": 3.4135814076386044,
      "tokens_seen": 2122121216
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017139209381501796,
      "loss": 2.5486,
      "theoretical_loss": 3.4135727202701487,
      "tokens_seen": 2122186752
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001713755437866465,
      "loss": 2.5705,
      "theoretical_loss": 3.4135640332450814,
      "tokens_seen": 2122252288
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000171358993758275,
      "loss": 2.6235,
      "theoretical_loss": 3.4135553465633794,
      "tokens_seen": 2122317824
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017134244372990354,
      "loss": 2.5584,
      "theoretical_loss": 3.413546660225017,
      "tokens_seen": 2122383360
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017132589370153205,
      "loss": 2.5587,
      "theoretical_loss": 3.413537974229971,
      "tokens_seen": 2122448896
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001713093436731606,
      "loss": 2.5911,
      "theoretical_loss": 3.413529288578217,
      "tokens_seen": 2122514432
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001712927936447891,
      "loss": 2.6156,
      "theoretical_loss": 3.4135206032697307,
      "tokens_seen": 2122579968
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017127624361641764,
      "loss": 2.6383,
      "theoretical_loss": 3.413511918304488,
      "tokens_seen": 2122645504
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017125969358804615,
      "loss": 2.4428,
      "theoretical_loss": 3.413503233682465,
      "tokens_seen": 2122711040
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001712431435596747,
      "loss": 2.3014,
      "theoretical_loss": 3.4134945494036373,
      "tokens_seen": 2122776576
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001712265935313032,
      "loss": 2.5003,
      "theoretical_loss": 3.4134858654679805,
      "tokens_seen": 2122842112
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017121004350293173,
      "loss": 2.4754,
      "theoretical_loss": 3.4134771818754714,
      "tokens_seen": 2122907648
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017119349347456025,
      "loss": 2.2976,
      "theoretical_loss": 3.4134684986260844,
      "tokens_seen": 2122973184
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017117694344618876,
      "loss": 2.5105,
      "theoretical_loss": 3.413459815719796,
      "tokens_seen": 2123038720
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017116039341781727,
      "loss": 2.5543,
      "theoretical_loss": 3.413451133156583,
      "tokens_seen": 2123104256
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001711438433894458,
      "loss": 2.5175,
      "theoretical_loss": 3.41344245093642,
      "tokens_seen": 2123169792
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017112729336107431,
      "loss": 2.4898,
      "theoretical_loss": 3.4134337690592833,
      "tokens_seen": 2123235328
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017111074333270285,
      "loss": 2.5563,
      "theoretical_loss": 3.413425087525149,
      "tokens_seen": 2123300864
    },
    {
      "epoch": 0.76,
      "objective/train/avg_token_score": 0.017608921974897385,
      "objective/train/avg_weight": 0.9792799949645996,
      "objective/train/docs_used": 1202332,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6995346546173096,
      "objective/train/std_weight": 0.08404523879289627,
      "objective/train/theoretical_loss": 3.413416406333992,
      "objective/train/tokens_used": 2143826400,
      "theoretical_loss": 3.413416406333992,
      "tokens_seen": 2123366400
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017109419330433136,
      "loss": 2.7369,
      "theoretical_loss": 3.413416406333992,
      "tokens_seen": 2123366400
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001710776432759599,
      "loss": 2.3881,
      "theoretical_loss": 3.413407725485789,
      "tokens_seen": 2123431936
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001710610932475884,
      "loss": 2.732,
      "theoretical_loss": 3.4133990449805163,
      "tokens_seen": 2123497472
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017104454321921695,
      "loss": 2.6037,
      "theoretical_loss": 3.4133903648181487,
      "tokens_seen": 2123563008
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017102799319084546,
      "loss": 2.6176,
      "theoretical_loss": 3.413381684998663,
      "tokens_seen": 2123628544
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000171011443162474,
      "loss": 2.6215,
      "theoretical_loss": 3.4133730055220344,
      "tokens_seen": 2123694080
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017099489313410253,
      "loss": 2.54,
      "theoretical_loss": 3.4133643263882387,
      "tokens_seen": 2123759616
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017097834310573104,
      "loss": 2.7953,
      "theoretical_loss": 3.413355647597253,
      "tokens_seen": 2123825152
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017096179307735958,
      "loss": 2.6732,
      "theoretical_loss": 3.4133469691490514,
      "tokens_seen": 2123890688
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001709452430489881,
      "loss": 2.3776,
      "theoretical_loss": 3.413338291043611,
      "tokens_seen": 2123956224
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001709286930206166,
      "loss": 2.4941,
      "theoretical_loss": 3.4133296132809074,
      "tokens_seen": 2124021760
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001709121429922451,
      "loss": 2.2915,
      "theoretical_loss": 3.4133209358609164,
      "tokens_seen": 2124087296
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017089559296387365,
      "loss": 2.4713,
      "theoretical_loss": 3.413312258783614,
      "tokens_seen": 2124152832
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017087904293550216,
      "loss": 2.5873,
      "theoretical_loss": 3.413303582048976,
      "tokens_seen": 2124218368
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001708624929071307,
      "loss": 2.6301,
      "theoretical_loss": 3.4132949056569784,
      "tokens_seen": 2124283904
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001708459428787592,
      "loss": 2.4929,
      "theoretical_loss": 3.4132862296075968,
      "tokens_seen": 2124349440
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017082939285038775,
      "loss": 2.5805,
      "theoretical_loss": 3.413277553900808,
      "tokens_seen": 2124414976
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017081284282201626,
      "loss": 2.3022,
      "theoretical_loss": 3.413268878536586,
      "tokens_seen": 2124480512
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001707962927936448,
      "loss": 2.6726,
      "theoretical_loss": 3.413260203514909,
      "tokens_seen": 2124546048
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001707797427652733,
      "loss": 2.5449,
      "theoretical_loss": 3.413251528835751,
      "tokens_seen": 2124611584
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017076319273690184,
      "loss": 2.6644,
      "theoretical_loss": 3.4132428544990896,
      "tokens_seen": 2124677120
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017074664270853035,
      "loss": 2.7959,
      "theoretical_loss": 3.4132341805048996,
      "tokens_seen": 2124742656
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001707300926801589,
      "loss": 2.582,
      "theoretical_loss": 3.4132255068531574,
      "tokens_seen": 2124808192
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001707135426517874,
      "loss": 2.5038,
      "theoretical_loss": 3.413216833543838,
      "tokens_seen": 2124873728
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001706969926234159,
      "loss": 2.4106,
      "theoretical_loss": 3.413208160576919,
      "tokens_seen": 2124939264
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017068044259504445,
      "loss": 2.6696,
      "theoretical_loss": 3.413199487952374,
      "tokens_seen": 2125004800
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017066389256667296,
      "loss": 2.5116,
      "theoretical_loss": 3.4131908156701813,
      "tokens_seen": 2125070336
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001706473425383015,
      "loss": 2.807,
      "theoretical_loss": 3.4131821437303156,
      "tokens_seen": 2125135872
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017063079250993,
      "loss": 2.4266,
      "theoretical_loss": 3.4131734721327533,
      "tokens_seen": 2125201408
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017061424248155854,
      "loss": 2.3948,
      "theoretical_loss": 3.41316480087747,
      "tokens_seen": 2125266944
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017059769245318705,
      "loss": 2.6127,
      "theoretical_loss": 3.4131561299644417,
      "tokens_seen": 2125332480
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001705811424248156,
      "loss": 2.8381,
      "theoretical_loss": 3.413147459393644,
      "tokens_seen": 2125398016
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001705645923964441,
      "loss": 2.4719,
      "theoretical_loss": 3.4131387891650533,
      "tokens_seen": 2125463552
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017054804236807264,
      "loss": 2.4501,
      "theoretical_loss": 3.4131301192786454,
      "tokens_seen": 2125529088
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017053149233970115,
      "loss": 2.7524,
      "theoretical_loss": 3.4131214497343967,
      "tokens_seen": 2125594624
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001705149423113297,
      "loss": 2.392,
      "theoretical_loss": 3.4131127805322823,
      "tokens_seen": 2125660160
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001704983922829582,
      "loss": 2.3861,
      "theoretical_loss": 3.4131041116722787,
      "tokens_seen": 2125725696
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017048184225458674,
      "loss": 2.5817,
      "theoretical_loss": 3.413095443154362,
      "tokens_seen": 2125791232
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017046529222621525,
      "loss": 2.9169,
      "theoretical_loss": 3.4130867749785074,
      "tokens_seen": 2125856768
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017044874219784376,
      "loss": 2.5764,
      "theoretical_loss": 3.4130781071446914,
      "tokens_seen": 2125922304
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017043219216947227,
      "loss": 1.95,
      "theoretical_loss": 3.4130694396528902,
      "tokens_seen": 2125987840
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001704156421411008,
      "loss": 2.6326,
      "theoretical_loss": 3.413060772503079,
      "tokens_seen": 2126053376
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017039909211272932,
      "loss": 2.5645,
      "theoretical_loss": 3.4130521056952348,
      "tokens_seen": 2126118912
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017038254208435785,
      "loss": 2.5772,
      "theoretical_loss": 3.413043439229333,
      "tokens_seen": 2126184448
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001703659920559864,
      "loss": 2.5609,
      "theoretical_loss": 3.413034773105349,
      "tokens_seen": 2126249984
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001703494420276149,
      "loss": 2.6349,
      "theoretical_loss": 3.41302610732326,
      "tokens_seen": 2126315520
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017033289199924344,
      "loss": 2.6624,
      "theoretical_loss": 3.4130174418830403,
      "tokens_seen": 2126381056
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017031634197087195,
      "loss": 2.4418,
      "theoretical_loss": 3.4130087767846677,
      "tokens_seen": 2126446592
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001702997919425005,
      "loss": 2.7186,
      "theoretical_loss": 3.4130001120281173,
      "tokens_seen": 2126512128
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000170283241914129,
      "loss": 2.6466,
      "theoretical_loss": 3.412991447613365,
      "tokens_seen": 2126577664
    },
    {
      "epoch": 0.76,
      "objective/train/avg_token_score": 0.01279906090348959,
      "objective/train/avg_weight": 0.9849342703819275,
      "objective/train/docs_used": 1204214,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.401750087738037,
      "objective/train/std_weight": 0.09329646080732346,
      "objective/train/theoretical_loss": 3.412982783540387,
      "objective/train/tokens_used": 2147103200,
      "theoretical_loss": 3.412982783540387,
      "tokens_seen": 2126643200
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017026669188575753,
      "loss": 2.5903,
      "theoretical_loss": 3.412982783540387,
      "tokens_seen": 2126643200
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017025014185738604,
      "loss": 2.6818,
      "theoretical_loss": 3.412974119809159,
      "tokens_seen": 2126708736
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017023359182901458,
      "loss": 2.4562,
      "theoretical_loss": 3.4129654564196574,
      "tokens_seen": 2126774272
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001702170418006431,
      "loss": 2.5971,
      "theoretical_loss": 3.4129567933718583,
      "tokens_seen": 2126839808
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001702004917722716,
      "loss": 2.5341,
      "theoretical_loss": 3.412948130665737,
      "tokens_seen": 2126905344
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017018394174390011,
      "loss": 2.4336,
      "theoretical_loss": 3.4129394683012704,
      "tokens_seen": 2126970880
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017016739171552865,
      "loss": 2.591,
      "theoretical_loss": 3.4129308062784336,
      "tokens_seen": 2127036416
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017015084168715716,
      "loss": 2.447,
      "theoretical_loss": 3.4129221445972027,
      "tokens_seen": 2127101952
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001701342916587857,
      "loss": 2.6108,
      "theoretical_loss": 3.4129134832575545,
      "tokens_seen": 2127167488
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001701177416304142,
      "loss": 2.6312,
      "theoretical_loss": 3.4129048222594642,
      "tokens_seen": 2127233024
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017010119160204275,
      "loss": 2.7727,
      "theoretical_loss": 3.4128961616029083,
      "tokens_seen": 2127298560
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017008464157367126,
      "loss": 2.3485,
      "theoretical_loss": 3.4128875012878623,
      "tokens_seen": 2127364096
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001700680915452998,
      "loss": 2.738,
      "theoretical_loss": 3.4128788413143027,
      "tokens_seen": 2127429632
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001700515415169283,
      "loss": 2.5514,
      "theoretical_loss": 3.412870181682206,
      "tokens_seen": 2127495168
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017003499148855684,
      "loss": 2.4241,
      "theoretical_loss": 3.412861522391547,
      "tokens_seen": 2127560704
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00017001844146018535,
      "loss": 2.6352,
      "theoretical_loss": 3.412852863442302,
      "tokens_seen": 2127626240
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001700018914318139,
      "loss": 2.4781,
      "theoretical_loss": 3.412844204834448,
      "tokens_seen": 2127691776
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001699853414034424,
      "loss": 2.7222,
      "theoretical_loss": 3.4128355465679596,
      "tokens_seen": 2127757312
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016996879137507094,
      "loss": 2.4862,
      "theoretical_loss": 3.412826888642814,
      "tokens_seen": 2127822848
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016995224134669945,
      "loss": 2.5283,
      "theoretical_loss": 3.4128182310589867,
      "tokens_seen": 2127888384
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016993569131832796,
      "loss": 2.4814,
      "theoretical_loss": 3.4128095738164537,
      "tokens_seen": 2127953920
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001699191412899565,
      "loss": 2.423,
      "theoretical_loss": 3.4128009169151916,
      "tokens_seen": 2128019456
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000169902591261585,
      "loss": 2.6969,
      "theoretical_loss": 3.4127922603551752,
      "tokens_seen": 2128084992
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016988604123321355,
      "loss": 2.7243,
      "theoretical_loss": 3.412783604136382,
      "tokens_seen": 2128150528
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016986949120484206,
      "loss": 2.368,
      "theoretical_loss": 3.4127749482587872,
      "tokens_seen": 2128216064
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001698529411764706,
      "loss": 2.4682,
      "theoretical_loss": 3.412766292722367,
      "tokens_seen": 2128281600
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001698363911480991,
      "loss": 2.5651,
      "theoretical_loss": 3.412757637527098,
      "tokens_seen": 2128347136
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016981984111972764,
      "loss": 2.6598,
      "theoretical_loss": 3.412748982672955,
      "tokens_seen": 2128412672
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016980329109135615,
      "loss": 2.5503,
      "theoretical_loss": 3.412740328159915,
      "tokens_seen": 2128478208
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001697867410629847,
      "loss": 2.4051,
      "theoretical_loss": 3.4127316739879543,
      "tokens_seen": 2128543744
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001697701910346132,
      "loss": 2.5286,
      "theoretical_loss": 3.4127230201570478,
      "tokens_seen": 2128609280
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016975364100624174,
      "loss": 2.4868,
      "theoretical_loss": 3.4127143666671724,
      "tokens_seen": 2128674816
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016973709097787025,
      "loss": 2.4038,
      "theoretical_loss": 3.4127057135183043,
      "tokens_seen": 2128740352
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016972054094949879,
      "loss": 2.3791,
      "theoretical_loss": 3.412697060710419,
      "tokens_seen": 2128805888
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001697039909211273,
      "loss": 2.3979,
      "theoretical_loss": 3.4126884082434934,
      "tokens_seen": 2128871424
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001696874408927558,
      "loss": 2.507,
      "theoretical_loss": 3.412679756117502,
      "tokens_seen": 2128936960
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016967089086438434,
      "loss": 2.3434,
      "theoretical_loss": 3.4126711043324227,
      "tokens_seen": 2129002496
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016965434083601285,
      "loss": 2.4856,
      "theoretical_loss": 3.4126624528882306,
      "tokens_seen": 2129068032
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001696377908076414,
      "loss": 2.6644,
      "theoretical_loss": 3.4126538017849017,
      "tokens_seen": 2129133568
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001696212407792699,
      "loss": 2.5737,
      "theoretical_loss": 3.4126451510224127,
      "tokens_seen": 2129199104
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016960469075089844,
      "loss": 2.4116,
      "theoretical_loss": 3.4126365006007386,
      "tokens_seen": 2129264640
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016958814072252695,
      "loss": 2.7425,
      "theoretical_loss": 3.412627850519857,
      "tokens_seen": 2129330176
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001695715906941555,
      "loss": 2.5623,
      "theoretical_loss": 3.412619200779743,
      "tokens_seen": 2129395712
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000169555040665784,
      "loss": 2.6528,
      "theoretical_loss": 3.4126105513803724,
      "tokens_seen": 2129461248
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016953849063741254,
      "loss": 2.5355,
      "theoretical_loss": 3.4126019023217222,
      "tokens_seen": 2129526784
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016952194060904105,
      "loss": 2.5642,
      "theoretical_loss": 3.4125932536037675,
      "tokens_seen": 2129592320
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016950539058066958,
      "loss": 2.6336,
      "theoretical_loss": 3.4125846052264857,
      "tokens_seen": 2129657856
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001694888405522981,
      "loss": 2.3816,
      "theoretical_loss": 3.4125759571898513,
      "tokens_seen": 2129723392
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016947229052392663,
      "loss": 2.7414,
      "theoretical_loss": 3.412567309493842,
      "tokens_seen": 2129788928
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016945574049555512,
      "loss": 2.438,
      "theoretical_loss": 3.4125586621384327,
      "tokens_seen": 2129854464
    },
    {
      "epoch": 0.76,
      "objective/train/avg_token_score": 0.011589955538511276,
      "objective/train/avg_weight": 0.9863681197166443,
      "objective/train/docs_used": 1206282,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.526275634765625,
      "objective/train/std_weight": 0.05875299498438835,
      "objective/train/theoretical_loss": 3.4125500151236,
      "objective/train/tokens_used": 2150380000,
      "theoretical_loss": 3.4125500151236,
      "tokens_seen": 2129920000
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016943919046718365,
      "loss": 2.5398,
      "theoretical_loss": 3.4125500151236,
      "tokens_seen": 2129920000
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016942264043881216,
      "loss": 2.5139,
      "theoretical_loss": 3.41254136844932,
      "tokens_seen": 2129985536
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001694060904104407,
      "loss": 2.3197,
      "theoretical_loss": 3.412532722115569,
      "tokens_seen": 2130051072
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001693895403820692,
      "loss": 2.6101,
      "theoretical_loss": 3.412524076122322,
      "tokens_seen": 2130116608
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016937299035369775,
      "loss": 2.3919,
      "theoretical_loss": 3.412515430469557,
      "tokens_seen": 2130182144
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016935644032532626,
      "loss": 2.6753,
      "theoretical_loss": 3.412506785157248,
      "tokens_seen": 2130247680
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001693398902969548,
      "loss": 2.5178,
      "theoretical_loss": 3.4124981401853733,
      "tokens_seen": 2130313216
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001693233402685833,
      "loss": 2.3337,
      "theoretical_loss": 3.412489495553908,
      "tokens_seen": 2130378752
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016930679024021184,
      "loss": 2.2315,
      "theoretical_loss": 3.4124808512628277,
      "tokens_seen": 2130444288
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016929024021184036,
      "loss": 2.7525,
      "theoretical_loss": 3.412472207312109,
      "tokens_seen": 2130509824
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001692736901834689,
      "loss": 2.6679,
      "theoretical_loss": 3.4124635637017278,
      "tokens_seen": 2130575360
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001692571401550974,
      "loss": 2.6517,
      "theoretical_loss": 3.4124549204316605,
      "tokens_seen": 2130640896
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016924059012672594,
      "loss": 2.3792,
      "theoretical_loss": 3.412446277501884,
      "tokens_seen": 2130706432
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016922404009835445,
      "loss": 2.5448,
      "theoretical_loss": 3.4124376349123726,
      "tokens_seen": 2130771968
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016920749006998296,
      "loss": 2.3097,
      "theoretical_loss": 3.412428992663104,
      "tokens_seen": 2130837504
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001691909400416115,
      "loss": 2.5204,
      "theoretical_loss": 3.4124203507540534,
      "tokens_seen": 2130903040
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016917439001324,
      "loss": 2.7839,
      "theoretical_loss": 3.4124117091851978,
      "tokens_seen": 2130968576
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016915783998486855,
      "loss": 2.4288,
      "theoretical_loss": 3.4124030679565127,
      "tokens_seen": 2131034112
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016914128995649706,
      "loss": 2.5479,
      "theoretical_loss": 3.4123944270679742,
      "tokens_seen": 2131099648
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001691247399281256,
      "loss": 2.5496,
      "theoretical_loss": 3.4123857865195593,
      "tokens_seen": 2131165184
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001691081898997541,
      "loss": 2.5644,
      "theoretical_loss": 3.4123771463112433,
      "tokens_seen": 2131230720
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016909163987138264,
      "loss": 2.644,
      "theoretical_loss": 3.412368506443002,
      "tokens_seen": 2131296256
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016907508984301115,
      "loss": 2.5799,
      "theoretical_loss": 3.412359866914813,
      "tokens_seen": 2131361792
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001690585398146397,
      "loss": 2.5313,
      "theoretical_loss": 3.4123512277266514,
      "tokens_seen": 2131427328
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001690419897862682,
      "loss": 2.1011,
      "theoretical_loss": 3.4123425888784933,
      "tokens_seen": 2131492864
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016902543975789674,
      "loss": 2.6087,
      "theoretical_loss": 3.4123339503703156,
      "tokens_seen": 2131558400
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016900888972952525,
      "loss": 2.4992,
      "theoretical_loss": 3.4123253122020936,
      "tokens_seen": 2131623936
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001689923397011538,
      "loss": 2.6227,
      "theoretical_loss": 3.4123166743738045,
      "tokens_seen": 2131689472
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001689757896727823,
      "loss": 2.7414,
      "theoretical_loss": 3.412308036885423,
      "tokens_seen": 2131755008
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001689592396444108,
      "loss": 2.7393,
      "theoretical_loss": 3.412299399736927,
      "tokens_seen": 2131820544
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016894268961603935,
      "loss": 2.3976,
      "theoretical_loss": 3.4122907629282913,
      "tokens_seen": 2131886080
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016892613958766786,
      "loss": 2.5595,
      "theoretical_loss": 3.412282126459493,
      "tokens_seen": 2131951616
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001689095895592964,
      "loss": 2.6474,
      "theoretical_loss": 3.412273490330507,
      "tokens_seen": 2132017152
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001688930395309249,
      "loss": 2.4439,
      "theoretical_loss": 3.412264854541311,
      "tokens_seen": 2132082688
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016887648950255344,
      "loss": 2.7792,
      "theoretical_loss": 3.412256219091881,
      "tokens_seen": 2132148224
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016885993947418195,
      "loss": 2.5195,
      "theoretical_loss": 3.412247583982192,
      "tokens_seen": 2132213760
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001688433894458105,
      "loss": 2.6405,
      "theoretical_loss": 3.412238949212221,
      "tokens_seen": 2132279296
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000168826839417439,
      "loss": 2.4359,
      "theoretical_loss": 3.4122303147819446,
      "tokens_seen": 2132344832
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016881028938906754,
      "loss": 2.5022,
      "theoretical_loss": 3.412221680691338,
      "tokens_seen": 2132410368
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016879373936069605,
      "loss": 2.6627,
      "theoretical_loss": 3.412213046940378,
      "tokens_seen": 2132475904
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016877718933232459,
      "loss": 2.6898,
      "theoretical_loss": 3.4122044135290412,
      "tokens_seen": 2132541440
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001687606393039531,
      "loss": 2.6929,
      "theoretical_loss": 3.412195780457303,
      "tokens_seen": 2132606976
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016874408927558163,
      "loss": 2.8488,
      "theoretical_loss": 3.41218714772514,
      "tokens_seen": 2132672512
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016872753924721012,
      "loss": 2.4967,
      "theoretical_loss": 3.412178515332528,
      "tokens_seen": 2132738048
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016871098921883865,
      "loss": 2.5183,
      "theoretical_loss": 3.412169883279444,
      "tokens_seen": 2132803584
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016869443919046716,
      "loss": 2.404,
      "theoretical_loss": 3.412161251565864,
      "tokens_seen": 2132869120
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001686778891620957,
      "loss": 2.5992,
      "theoretical_loss": 3.4121526201917636,
      "tokens_seen": 2132934656
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001686613391337242,
      "loss": 2.4987,
      "theoretical_loss": 3.4121439891571193,
      "tokens_seen": 2133000192
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016864478910535275,
      "loss": 2.7005,
      "theoretical_loss": 3.4121353584619074,
      "tokens_seen": 2133065728
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016862823907698126,
      "loss": 2.4816,
      "theoretical_loss": 3.412126728106104,
      "tokens_seen": 2133131264
    },
    {
      "epoch": 0.76,
      "objective/train/avg_token_score": 0.013661704957485199,
      "objective/train/avg_weight": 0.9839208722114563,
      "objective/train/docs_used": 1208010,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9387497901916504,
      "objective/train/std_weight": 0.05751820281147957,
      "objective/train/theoretical_loss": 3.412118098089686,
      "objective/train/tokens_used": 2153656800,
      "theoretical_loss": 3.412118098089686,
      "tokens_seen": 2133196800
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001686116890486098,
      "loss": 2.7134,
      "theoretical_loss": 3.412118098089686,
      "tokens_seen": 2133196800
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001685951390202383,
      "loss": 2.1996,
      "theoretical_loss": 3.4121094684126287,
      "tokens_seen": 2133262336
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016857858899186685,
      "loss": 2.5807,
      "theoretical_loss": 3.412100839074909,
      "tokens_seen": 2133327872
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016856203896349536,
      "loss": 2.5611,
      "theoretical_loss": 3.412092210076503,
      "tokens_seen": 2133393408
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001685454889351239,
      "loss": 2.5268,
      "theoretical_loss": 3.412083581417386,
      "tokens_seen": 2133458944
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001685289389067524,
      "loss": 2.6032,
      "theoretical_loss": 3.412074953097536,
      "tokens_seen": 2133524480
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016851238887838094,
      "loss": 2.3021,
      "theoretical_loss": 3.412066325116928,
      "tokens_seen": 2133590016
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016849583885000945,
      "loss": 2.5607,
      "theoretical_loss": 3.4120576974755386,
      "tokens_seen": 2133655552
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016847928882163796,
      "loss": 2.56,
      "theoretical_loss": 3.4120490701733432,
      "tokens_seen": 2133721088
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001684627387932665,
      "loss": 3.1389,
      "theoretical_loss": 3.4120404432103193,
      "tokens_seen": 2133786624
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000168446188764895,
      "loss": 2.5293,
      "theoretical_loss": 3.4120318165864427,
      "tokens_seen": 2133852160
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016842963873652355,
      "loss": 2.5794,
      "theoretical_loss": 3.41202319030169,
      "tokens_seen": 2133917696
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016841308870815206,
      "loss": 2.7069,
      "theoretical_loss": 3.4120145643560367,
      "tokens_seen": 2133983232
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001683965386797806,
      "loss": 2.6979,
      "theoretical_loss": 3.4120059387494592,
      "tokens_seen": 2134048768
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001683799886514091,
      "loss": 2.8093,
      "theoretical_loss": 3.411997313481934,
      "tokens_seen": 2134114304
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016836343862303764,
      "loss": 2.736,
      "theoretical_loss": 3.411988688553438,
      "tokens_seen": 2134179840
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016834688859466615,
      "loss": 2.7252,
      "theoretical_loss": 3.411980063963946,
      "tokens_seen": 2134245376
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001683303385662947,
      "loss": 2.512,
      "theoretical_loss": 3.4119714397134357,
      "tokens_seen": 2134310912
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001683137885379232,
      "loss": 2.5839,
      "theoretical_loss": 3.4119628158018824,
      "tokens_seen": 2134376448
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016829723850955174,
      "loss": 2.2635,
      "theoretical_loss": 3.411954192229263,
      "tokens_seen": 2134441984
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016828068848118025,
      "loss": 2.8086,
      "theoretical_loss": 3.411945568995553,
      "tokens_seen": 2134507520
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001682641384528088,
      "loss": 2.7814,
      "theoretical_loss": 3.411936946100729,
      "tokens_seen": 2134573056
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001682475884244373,
      "loss": 2.5004,
      "theoretical_loss": 3.4119283235447684,
      "tokens_seen": 2134638592
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001682310383960658,
      "loss": 2.4417,
      "theoretical_loss": 3.411919701327646,
      "tokens_seen": 2134704128
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016821448836769435,
      "loss": 2.4347,
      "theoretical_loss": 3.4119110794493386,
      "tokens_seen": 2134769664
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016819793833932286,
      "loss": 2.7461,
      "theoretical_loss": 3.4119024579098225,
      "tokens_seen": 2134835200
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001681813883109514,
      "loss": 2.5989,
      "theoretical_loss": 3.4118938367090736,
      "tokens_seen": 2134900736
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001681648382825799,
      "loss": 2.3821,
      "theoretical_loss": 3.411885215847069,
      "tokens_seen": 2134966272
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016814828825420844,
      "loss": 2.7216,
      "theoretical_loss": 3.4118765953237844,
      "tokens_seen": 2135031808
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016813173822583695,
      "loss": 2.54,
      "theoretical_loss": 3.4118679751391965,
      "tokens_seen": 2135097344
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001681151881974655,
      "loss": 2.7594,
      "theoretical_loss": 3.411859355293281,
      "tokens_seen": 2135162880
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000168098638169094,
      "loss": 2.7944,
      "theoretical_loss": 3.4118507357860146,
      "tokens_seen": 2135228416
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016808208814072254,
      "loss": 2.5989,
      "theoretical_loss": 3.4118421166173736,
      "tokens_seen": 2135293952
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016806553811235105,
      "loss": 2.3471,
      "theoretical_loss": 3.4118334977873346,
      "tokens_seen": 2135359488
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001680489880839796,
      "loss": 2.5969,
      "theoretical_loss": 3.4118248792958736,
      "tokens_seen": 2135425024
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001680324380556081,
      "loss": 2.6645,
      "theoretical_loss": 3.4118162611429663,
      "tokens_seen": 2135490560
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016801588802723663,
      "loss": 2.5961,
      "theoretical_loss": 3.41180764332859,
      "tokens_seen": 2135556096
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016799933799886515,
      "loss": 2.5761,
      "theoretical_loss": 3.41179902585272,
      "tokens_seen": 2135621632
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016798278797049366,
      "loss": 2.916,
      "theoretical_loss": 3.411790408715334,
      "tokens_seen": 2135687168
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016796623794212217,
      "loss": 2.5842,
      "theoretical_loss": 3.411781791916407,
      "tokens_seen": 2135752704
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001679496879137507,
      "loss": 2.444,
      "theoretical_loss": 3.411773175455916,
      "tokens_seen": 2135818240
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016793313788537921,
      "loss": 2.463,
      "theoretical_loss": 3.4117645593338377,
      "tokens_seen": 2135883776
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016791658785700775,
      "loss": 2.4268,
      "theoretical_loss": 3.411755943550147,
      "tokens_seen": 2135949312
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016790003782863626,
      "loss": 2.6525,
      "theoretical_loss": 3.4117473281048216,
      "tokens_seen": 2136014848
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001678834878002648,
      "loss": 2.5343,
      "theoretical_loss": 3.4117387129978374,
      "tokens_seen": 2136080384
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001678669377718933,
      "loss": 2.5931,
      "theoretical_loss": 3.4117300982291705,
      "tokens_seen": 2136145920
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016785038774352185,
      "loss": 2.476,
      "theoretical_loss": 3.4117214837987975,
      "tokens_seen": 2136211456
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016783383771515036,
      "loss": 2.4832,
      "theoretical_loss": 3.4117128697066947,
      "tokens_seen": 2136276992
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001678172876867789,
      "loss": 2.5619,
      "theoretical_loss": 3.4117042559528388,
      "tokens_seen": 2136342528
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001678007376584074,
      "loss": 2.6233,
      "theoretical_loss": 3.411695642537205,
      "tokens_seen": 2136408064
    },
    {
      "epoch": 0.76,
      "objective/train/avg_token_score": 0.011317876167595387,
      "objective/train/avg_weight": 0.9866769313812256,
      "objective/train/docs_used": 1209970,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.889711380004883,
      "objective/train/std_weight": 0.07975874096155167,
      "objective/train/theoretical_loss": 3.411687029459771,
      "objective/train/tokens_used": 2156933600,
      "theoretical_loss": 3.411687029459771,
      "tokens_seen": 2136473600
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016778418763003594,
      "loss": 2.7127,
      "theoretical_loss": 3.411687029459771,
      "tokens_seen": 2136473600
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016776763760166445,
      "loss": 2.4361,
      "theoretical_loss": 3.4116784167205125,
      "tokens_seen": 2136539136
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000167751087573293,
      "loss": 2.6078,
      "theoretical_loss": 3.4116698043194056,
      "tokens_seen": 2136604672
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001677345375449215,
      "loss": 2.631,
      "theoretical_loss": 3.411661192256427,
      "tokens_seen": 2136670208
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016771798751655,
      "loss": 2.4625,
      "theoretical_loss": 3.4116525805315527,
      "tokens_seen": 2136735744
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016770143748817855,
      "loss": 2.6541,
      "theoretical_loss": 3.4116439691447598,
      "tokens_seen": 2136801280
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016768488745980706,
      "loss": 2.3449,
      "theoretical_loss": 3.4116353580960244,
      "tokens_seen": 2136866816
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001676683374314356,
      "loss": 2.5226,
      "theoretical_loss": 3.411626747385322,
      "tokens_seen": 2136932352
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001676517874030641,
      "loss": 2.6895,
      "theoretical_loss": 3.4116181370126304,
      "tokens_seen": 2136997888
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016763523737469265,
      "loss": 2.2914,
      "theoretical_loss": 3.4116095269779247,
      "tokens_seen": 2137063424
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016761868734632116,
      "loss": 2.4457,
      "theoretical_loss": 3.4116009172811816,
      "tokens_seen": 2137128960
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001676021373179497,
      "loss": 2.5927,
      "theoretical_loss": 3.411592307922378,
      "tokens_seen": 2137194496
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001675855872895782,
      "loss": 2.5954,
      "theoretical_loss": 3.41158369890149,
      "tokens_seen": 2137260032
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016756903726120674,
      "loss": 2.4795,
      "theoretical_loss": 3.411575090218494,
      "tokens_seen": 2137325568
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016755248723283525,
      "loss": 2.6088,
      "theoretical_loss": 3.4115664818733658,
      "tokens_seen": 2137391104
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001675359372044638,
      "loss": 2.5754,
      "theoretical_loss": 3.411557873866083,
      "tokens_seen": 2137456640
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001675193871760923,
      "loss": 2.4071,
      "theoretical_loss": 3.4115492661966202,
      "tokens_seen": 2137522176
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001675028371477208,
      "loss": 2.5833,
      "theoretical_loss": 3.411540658864955,
      "tokens_seen": 2137587712
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016748628711934935,
      "loss": 2.7992,
      "theoretical_loss": 3.411532051871064,
      "tokens_seen": 2137653248
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016746973709097786,
      "loss": 2.6731,
      "theoretical_loss": 3.4115234452149235,
      "tokens_seen": 2137718784
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001674531870626064,
      "loss": 2.5894,
      "theoretical_loss": 3.411514838896509,
      "tokens_seen": 2137784320
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001674366370342349,
      "loss": 2.3122,
      "theoretical_loss": 3.411506232915798,
      "tokens_seen": 2137849856
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016742008700586344,
      "loss": 2.6585,
      "theoretical_loss": 3.411497627272766,
      "tokens_seen": 2137915392
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016740353697749195,
      "loss": 2.3721,
      "theoretical_loss": 3.41148902196739,
      "tokens_seen": 2137980928
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001673869869491205,
      "loss": 2.7222,
      "theoretical_loss": 3.4114804169996464,
      "tokens_seen": 2138046464
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000167370436920749,
      "loss": 2.5644,
      "theoretical_loss": 3.411471812369511,
      "tokens_seen": 2138112000
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016735388689237754,
      "loss": 2.6023,
      "theoretical_loss": 3.4114632080769605,
      "tokens_seen": 2138177536
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016733733686400605,
      "loss": 2.7482,
      "theoretical_loss": 3.4114546041219715,
      "tokens_seen": 2138243072
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001673207868356346,
      "loss": 2.2156,
      "theoretical_loss": 3.4114460005045206,
      "tokens_seen": 2138308608
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001673042368072631,
      "loss": 2.0825,
      "theoretical_loss": 3.4114373972245837,
      "tokens_seen": 2138374144
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016728768677889164,
      "loss": 2.4179,
      "theoretical_loss": 3.4114287942821377,
      "tokens_seen": 2138439680
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016727113675052015,
      "loss": 2.5865,
      "theoretical_loss": 3.4114201916771583,
      "tokens_seen": 2138505216
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016725458672214866,
      "loss": 2.5726,
      "theoretical_loss": 3.4114115894096226,
      "tokens_seen": 2138570752
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016723803669377717,
      "loss": 2.4556,
      "theoretical_loss": 3.411402987479507,
      "tokens_seen": 2138636288
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001672214866654057,
      "loss": 2.4992,
      "theoretical_loss": 3.4113943858867875,
      "tokens_seen": 2138701824
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016720493663703422,
      "loss": 2.4876,
      "theoretical_loss": 3.411385784631441,
      "tokens_seen": 2138767360
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016718838660866275,
      "loss": 2.3448,
      "theoretical_loss": 3.4113771837134435,
      "tokens_seen": 2138832896
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016717183658029126,
      "loss": 2.4025,
      "theoretical_loss": 3.411368583132772,
      "tokens_seen": 2138898432
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001671552865519198,
      "loss": 2.5171,
      "theoretical_loss": 3.411359982889402,
      "tokens_seen": 2138963968
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001671387365235483,
      "loss": 2.4903,
      "theoretical_loss": 3.4113513829833106,
      "tokens_seen": 2139029504
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016712218649517685,
      "loss": 2.5994,
      "theoretical_loss": 3.411342783414474,
      "tokens_seen": 2139095040
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016710563646680536,
      "loss": 2.5608,
      "theoretical_loss": 3.411334184182869,
      "tokens_seen": 2139160576
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001670890864384339,
      "loss": 2.5673,
      "theoretical_loss": 3.411325585288472,
      "tokens_seen": 2139226112
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001670725364100624,
      "loss": 2.431,
      "theoretical_loss": 3.411316986731259,
      "tokens_seen": 2139291648
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016705598638169094,
      "loss": 2.7335,
      "theoretical_loss": 3.411308388511207,
      "tokens_seen": 2139357184
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016703943635331946,
      "loss": 2.7233,
      "theoretical_loss": 3.4112997906282923,
      "tokens_seen": 2139422720
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000167022886324948,
      "loss": 2.588,
      "theoretical_loss": 3.4112911930824907,
      "tokens_seen": 2139488256
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001670063362965765,
      "loss": 2.4491,
      "theoretical_loss": 3.4112825958737796,
      "tokens_seen": 2139553792
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016698978626820501,
      "loss": 2.5162,
      "theoretical_loss": 3.411273999002135,
      "tokens_seen": 2139619328
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016697323623983355,
      "loss": 2.5529,
      "theoretical_loss": 3.4112654024675333,
      "tokens_seen": 2139684864
    },
    {
      "epoch": 0.76,
      "objective/train/avg_token_score": 0.04034283012151718,
      "objective/train/avg_weight": 0.9525216817855835,
      "objective/train/docs_used": 1212088,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9399189949035645,
      "objective/train/std_weight": 0.17058946192264557,
      "objective/train/theoretical_loss": 3.411256806269951,
      "objective/train/tokens_used": 2160210400,
      "theoretical_loss": 3.411256806269951,
      "tokens_seen": 2139750400
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016695668621146206,
      "loss": 2.441,
      "theoretical_loss": 3.411256806269951,
      "tokens_seen": 2139750400
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001669401361830906,
      "loss": 2.6825,
      "theoretical_loss": 3.411248210409364,
      "tokens_seen": 2139815936
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001669235861547191,
      "loss": 2.7071,
      "theoretical_loss": 3.4112396148857504,
      "tokens_seen": 2139881472
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016690703612634765,
      "loss": 2.4773,
      "theoretical_loss": 3.411231019699086,
      "tokens_seen": 2139947008
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016689048609797616,
      "loss": 2.032,
      "theoretical_loss": 3.411222424849346,
      "tokens_seen": 2140012544
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001668739360696047,
      "loss": 2.5647,
      "theoretical_loss": 3.411213830336508,
      "tokens_seen": 2140078080
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001668573860412332,
      "loss": 2.5203,
      "theoretical_loss": 3.4112052361605487,
      "tokens_seen": 2140143616
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016684083601286174,
      "loss": 2.8253,
      "theoretical_loss": 3.4111966423214435,
      "tokens_seen": 2140209152
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016682428598449025,
      "loss": 2.3736,
      "theoretical_loss": 3.41118804881917,
      "tokens_seen": 2140274688
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001668077359561188,
      "loss": 2.5606,
      "theoretical_loss": 3.4111794556537047,
      "tokens_seen": 2140340224
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001667911859277473,
      "loss": 2.317,
      "theoretical_loss": 3.411170862825023,
      "tokens_seen": 2140405760
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016677463589937584,
      "loss": 2.5985,
      "theoretical_loss": 3.4111622703331017,
      "tokens_seen": 2140471296
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016675808587100435,
      "loss": 2.6769,
      "theoretical_loss": 3.4111536781779184,
      "tokens_seen": 2140536832
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016674153584263286,
      "loss": 2.2192,
      "theoretical_loss": 3.4111450863594484,
      "tokens_seen": 2140602368
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001667249858142614,
      "loss": 2.5141,
      "theoretical_loss": 3.4111364948776686,
      "tokens_seen": 2140667904
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001667084357858899,
      "loss": 2.484,
      "theoretical_loss": 3.411127903732556,
      "tokens_seen": 2140733440
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016669188575751845,
      "loss": 2.4894,
      "theoretical_loss": 3.411119312924086,
      "tokens_seen": 2140798976
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016667533572914696,
      "loss": 2.5338,
      "theoretical_loss": 3.4111107224522357,
      "tokens_seen": 2140864512
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001666587857007755,
      "loss": 2.5787,
      "theoretical_loss": 3.411102132316982,
      "tokens_seen": 2140930048
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000166642235672404,
      "loss": 2.5529,
      "theoretical_loss": 3.411093542518301,
      "tokens_seen": 2140995584
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016662568564403254,
      "loss": 2.5478,
      "theoretical_loss": 3.411084953056169,
      "tokens_seen": 2141061120
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016660913561566105,
      "loss": 2.4315,
      "theoretical_loss": 3.411076363930563,
      "tokens_seen": 2141126656
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001665925855872896,
      "loss": 2.4478,
      "theoretical_loss": 3.411067775141459,
      "tokens_seen": 2141192192
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001665760355589181,
      "loss": 2.6635,
      "theoretical_loss": 3.411059186688834,
      "tokens_seen": 2141257728
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016655948553054664,
      "loss": 2.3406,
      "theoretical_loss": 3.411050598572664,
      "tokens_seen": 2141323264
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016654293550217515,
      "loss": 2.492,
      "theoretical_loss": 3.4110420107929262,
      "tokens_seen": 2141388800
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016652638547380369,
      "loss": 2.6887,
      "theoretical_loss": 3.4110334233495965,
      "tokens_seen": 2141454336
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016650983544543217,
      "loss": 2.579,
      "theoretical_loss": 3.4110248362426514,
      "tokens_seen": 2141519872
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001664932854170607,
      "loss": 2.7777,
      "theoretical_loss": 3.4110162494720684,
      "tokens_seen": 2141585408
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016647673538868922,
      "loss": 2.6169,
      "theoretical_loss": 3.4110076630378225,
      "tokens_seen": 2141650944
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016646018536031775,
      "loss": 2.609,
      "theoretical_loss": 3.410999076939892,
      "tokens_seen": 2141716480
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00016644363533194626,
      "loss": 2.5106,
      "theoretical_loss": 3.410990491178252,
      "tokens_seen": 2141782016
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001664270853035748,
      "loss": 2.652,
      "theoretical_loss": 3.4109819057528794,
      "tokens_seen": 2141847552
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001664105352752033,
      "loss": 2.7114,
      "theoretical_loss": 3.4109733206637514,
      "tokens_seen": 2141913088
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016639398524683185,
      "loss": 2.1567,
      "theoretical_loss": 3.4109647359108437,
      "tokens_seen": 2141978624
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016637743521846036,
      "loss": 2.744,
      "theoretical_loss": 3.4109561514941333,
      "tokens_seen": 2142044160
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001663608851900889,
      "loss": 2.2876,
      "theoretical_loss": 3.4109475674135963,
      "tokens_seen": 2142109696
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016634433516171744,
      "loss": 2.3069,
      "theoretical_loss": 3.41093898366921,
      "tokens_seen": 2142175232
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016632778513334595,
      "loss": 2.6742,
      "theoretical_loss": 3.41093040026095,
      "tokens_seen": 2142240768
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016631123510497448,
      "loss": 2.3643,
      "theoretical_loss": 3.410921817188794,
      "tokens_seen": 2142306304
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000166294685076603,
      "loss": 2.5119,
      "theoretical_loss": 3.4109132344527175,
      "tokens_seen": 2142371840
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016627813504823153,
      "loss": 2.4783,
      "theoretical_loss": 3.4109046520526975,
      "tokens_seen": 2142437376
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016626158501986002,
      "loss": 2.6603,
      "theoretical_loss": 3.410896069988711,
      "tokens_seen": 2142502912
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016624503499148855,
      "loss": 2.3197,
      "theoretical_loss": 3.410887488260734,
      "tokens_seen": 2142568448
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016622848496311706,
      "loss": 2.6175,
      "theoretical_loss": 3.410878906868743,
      "tokens_seen": 2142633984
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001662119349347456,
      "loss": 2.717,
      "theoretical_loss": 3.4108703258127147,
      "tokens_seen": 2142699520
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001661953849063741,
      "loss": 2.8585,
      "theoretical_loss": 3.4108617450926255,
      "tokens_seen": 2142765056
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016617883487800265,
      "loss": 2.3651,
      "theoretical_loss": 3.4108531647084526,
      "tokens_seen": 2142830592
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016616228484963116,
      "loss": 2.3353,
      "theoretical_loss": 3.4108445846601727,
      "tokens_seen": 2142896128
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001661457348212597,
      "loss": 2.5847,
      "theoretical_loss": 3.410836004947761,
      "tokens_seen": 2142961664
    },
    {
      "epoch": 0.77,
      "objective/train/avg_token_score": 0.020660296082496643,
      "objective/train/avg_weight": 0.9756791591644287,
      "objective/train/docs_used": 1213995,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.526850938796997,
      "objective/train/std_weight": 0.09951050579547882,
      "objective/train/theoretical_loss": 3.4108274255711954,
      "objective/train/tokens_used": 2163487200,
      "theoretical_loss": 3.4108274255711954,
      "tokens_seen": 2143027200
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001661291847928882,
      "loss": 2.4074,
      "theoretical_loss": 3.4108274255711954,
      "tokens_seen": 2143027200
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016611263476451674,
      "loss": 2.4615,
      "theoretical_loss": 3.410818846530452,
      "tokens_seen": 2143092736
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016609608473614526,
      "loss": 2.7042,
      "theoretical_loss": 3.4108102678255072,
      "tokens_seen": 2143158272
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001660795347077738,
      "loss": 2.6698,
      "theoretical_loss": 3.410801689456338,
      "tokens_seen": 2143223808
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001660629846794023,
      "loss": 2.7055,
      "theoretical_loss": 3.4107931114229206,
      "tokens_seen": 2143289344
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016604643465103084,
      "loss": 2.5631,
      "theoretical_loss": 3.410784533725232,
      "tokens_seen": 2143354880
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016602988462265935,
      "loss": 2.6031,
      "theoretical_loss": 3.4107759563632487,
      "tokens_seen": 2143420416
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016601333459428786,
      "loss": 2.7141,
      "theoretical_loss": 3.4107673793369466,
      "tokens_seen": 2143485952
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001659967845659164,
      "loss": 2.6936,
      "theoretical_loss": 3.4107588026463036,
      "tokens_seen": 2143551488
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001659802345375449,
      "loss": 2.5191,
      "theoretical_loss": 3.410750226291295,
      "tokens_seen": 2143617024
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016596368450917345,
      "loss": 2.6762,
      "theoretical_loss": 3.4107416502718984,
      "tokens_seen": 2143682560
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016594713448080196,
      "loss": 2.2338,
      "theoretical_loss": 3.41073307458809,
      "tokens_seen": 2143748096
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001659305844524305,
      "loss": 2.4669,
      "theoretical_loss": 3.4107244992398464,
      "tokens_seen": 2143813632
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000165914034424059,
      "loss": 2.2809,
      "theoretical_loss": 3.4107159242271443,
      "tokens_seen": 2143879168
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016589748439568754,
      "loss": 2.352,
      "theoretical_loss": 3.41070734954996,
      "tokens_seen": 2143944704
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016588093436731605,
      "loss": 2.5917,
      "theoretical_loss": 3.41069877520827,
      "tokens_seen": 2144010240
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001658643843389446,
      "loss": 2.6552,
      "theoretical_loss": 3.410690201202052,
      "tokens_seen": 2144075776
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001658478343105731,
      "loss": 2.6897,
      "theoretical_loss": 3.4106816275312815,
      "tokens_seen": 2144141312
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016583128428220164,
      "loss": 2.6572,
      "theoretical_loss": 3.410673054195936,
      "tokens_seen": 2144206848
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016581473425383015,
      "loss": 2.3834,
      "theoretical_loss": 3.410664481195991,
      "tokens_seen": 2144272384
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001657981842254587,
      "loss": 2.7229,
      "theoretical_loss": 3.4106559085314236,
      "tokens_seen": 2144337920
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016578163419708717,
      "loss": 2.5685,
      "theoretical_loss": 3.4106473362022114,
      "tokens_seen": 2144403456
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001657650841687157,
      "loss": 2.5324,
      "theoretical_loss": 3.41063876420833,
      "tokens_seen": 2144468992
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016574853414034425,
      "loss": 2.5523,
      "theoretical_loss": 3.410630192549756,
      "tokens_seen": 2144534528
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016573198411197276,
      "loss": 2.5177,
      "theoretical_loss": 3.4106216212264666,
      "tokens_seen": 2144600064
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001657154340836013,
      "loss": 2.0895,
      "theoretical_loss": 3.4106130502384375,
      "tokens_seen": 2144665600
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001656988840552298,
      "loss": 2.448,
      "theoretical_loss": 3.410604479585647,
      "tokens_seen": 2144731136
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016568233402685834,
      "loss": 2.4089,
      "theoretical_loss": 3.4105959092680695,
      "tokens_seen": 2144796672
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016566578399848685,
      "loss": 2.4466,
      "theoretical_loss": 3.410587339285684,
      "tokens_seen": 2144862208
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001656492339701154,
      "loss": 2.5429,
      "theoretical_loss": 3.4105787696384655,
      "tokens_seen": 2144927744
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001656326839417439,
      "loss": 2.5095,
      "theoretical_loss": 3.4105702003263914,
      "tokens_seen": 2144993280
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016561613391337244,
      "loss": 2.5318,
      "theoretical_loss": 3.410561631349438,
      "tokens_seen": 2145058816
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016559958388500095,
      "loss": 2.4925,
      "theoretical_loss": 3.410553062707582,
      "tokens_seen": 2145124352
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016558303385662949,
      "loss": 2.8085,
      "theoretical_loss": 3.4105444944008,
      "tokens_seen": 2145189888
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000165566483828258,
      "loss": 2.5741,
      "theoretical_loss": 3.4105359264290693,
      "tokens_seen": 2145255424
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016554993379988653,
      "loss": 2.7843,
      "theoretical_loss": 3.4105273587923657,
      "tokens_seen": 2145320960
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016553338377151502,
      "loss": 2.6982,
      "theoretical_loss": 3.4105187914906665,
      "tokens_seen": 2145386496
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016551683374314355,
      "loss": 2.6903,
      "theoretical_loss": 3.4105102245239483,
      "tokens_seen": 2145452032
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016550028371477206,
      "loss": 2.5925,
      "theoretical_loss": 3.410501657892187,
      "tokens_seen": 2145517568
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001654837336864006,
      "loss": 2.3539,
      "theoretical_loss": 3.4104930915953604,
      "tokens_seen": 2145583104
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001654671836580291,
      "loss": 2.6672,
      "theoretical_loss": 3.410484525633444,
      "tokens_seen": 2145648640
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016545063362965765,
      "loss": 2.4724,
      "theoretical_loss": 3.4104759600064156,
      "tokens_seen": 2145714176
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016543408360128616,
      "loss": 2.6953,
      "theoretical_loss": 3.410467394714251,
      "tokens_seen": 2145779712
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001654175335729147,
      "loss": 2.6536,
      "theoretical_loss": 3.4104588297569283,
      "tokens_seen": 2145845248
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001654009835445432,
      "loss": 2.7465,
      "theoretical_loss": 3.410450265134422,
      "tokens_seen": 2145910784
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016538443351617175,
      "loss": 2.5337,
      "theoretical_loss": 3.41044170084671,
      "tokens_seen": 2145976320
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016536788348780026,
      "loss": 2.5112,
      "theoretical_loss": 3.4104331368937695,
      "tokens_seen": 2146041856
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001653513334594288,
      "loss": 2.3958,
      "theoretical_loss": 3.4104245732755762,
      "tokens_seen": 2146107392
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001653347834310573,
      "loss": 2.6191,
      "theoretical_loss": 3.4104160099921077,
      "tokens_seen": 2146172928
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016531823340268584,
      "loss": 2.7138,
      "theoretical_loss": 3.41040744704334,
      "tokens_seen": 2146238464
    },
    {
      "epoch": 0.77,
      "objective/train/avg_token_score": 0.015874803066253662,
      "objective/train/avg_weight": 0.9813164472579956,
      "objective/train/docs_used": 1215500,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.618234157562256,
      "objective/train/std_weight": 0.09251457452774048,
      "objective/train/theoretical_loss": 3.4103988844292497,
      "objective/train/tokens_used": 2166764000,
      "theoretical_loss": 3.4103988844292497,
      "tokens_seen": 2146304000
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016530168337431435,
      "loss": 2.4933,
      "theoretical_loss": 3.4103988844292497,
      "tokens_seen": 2146304000
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016528513334594286,
      "loss": 2.1666,
      "theoretical_loss": 3.410390322149814,
      "tokens_seen": 2146369536
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001652685833175714,
      "loss": 1.9729,
      "theoretical_loss": 3.4103817602050093,
      "tokens_seen": 2146435072
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001652520332891999,
      "loss": 2.5048,
      "theoretical_loss": 3.4103731985948125,
      "tokens_seen": 2146500608
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016523548326082845,
      "loss": 2.6867,
      "theoretical_loss": 3.4103646373192005,
      "tokens_seen": 2146566144
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016521893323245696,
      "loss": 2.6111,
      "theoretical_loss": 3.4103560763781493,
      "tokens_seen": 2146631680
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001652023832040855,
      "loss": 2.4367,
      "theoretical_loss": 3.4103475157716363,
      "tokens_seen": 2146697216
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000165185833175714,
      "loss": 2.4948,
      "theoretical_loss": 3.410338955499638,
      "tokens_seen": 2146762752
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016516928314734254,
      "loss": 2.6705,
      "theoretical_loss": 3.4103303955621307,
      "tokens_seen": 2146828288
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016515273311897105,
      "loss": 2.8203,
      "theoretical_loss": 3.410321835959092,
      "tokens_seen": 2146893824
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001651361830905996,
      "loss": 2.4829,
      "theoretical_loss": 3.4103132766904976,
      "tokens_seen": 2146959360
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001651196330622281,
      "loss": 2.5326,
      "theoretical_loss": 3.4103047177563255,
      "tokens_seen": 2147024896
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016510308303385664,
      "loss": 2.6563,
      "theoretical_loss": 3.410296159156551,
      "tokens_seen": 2147090432
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016508653300548515,
      "loss": 2.6958,
      "theoretical_loss": 3.410287600891152,
      "tokens_seen": 2147155968
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001650699829771137,
      "loss": 2.3279,
      "theoretical_loss": 3.410279042960104,
      "tokens_seen": 2147221504
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001650534329487422,
      "loss": 2.7338,
      "theoretical_loss": 3.4102704853633847,
      "tokens_seen": 2147287040
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001650368829203707,
      "loss": 2.6262,
      "theoretical_loss": 3.4102619281009714,
      "tokens_seen": 2147352576
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016502033289199925,
      "loss": 2.5921,
      "theoretical_loss": 3.410253371172839,
      "tokens_seen": 2147418112
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016500378286362776,
      "loss": 2.489,
      "theoretical_loss": 3.4102448145789657,
      "tokens_seen": 2147483648
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001649872328352563,
      "loss": 2.5899,
      "theoretical_loss": 3.410236258319328,
      "tokens_seen": 2147549184
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001649706828068848,
      "loss": 2.6344,
      "theoretical_loss": 3.4102277023939025,
      "tokens_seen": 2147614720
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016495413277851334,
      "loss": 2.6508,
      "theoretical_loss": 3.4102191468026657,
      "tokens_seen": 2147680256
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016493758275014185,
      "loss": 2.548,
      "theoretical_loss": 3.4102105915455945,
      "tokens_seen": 2147745792
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001649210327217704,
      "loss": 2.5186,
      "theoretical_loss": 3.4102020366226657,
      "tokens_seen": 2147811328
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001649044826933989,
      "loss": 2.7718,
      "theoretical_loss": 3.410193482033856,
      "tokens_seen": 2147876864
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016488793266502744,
      "loss": 2.793,
      "theoretical_loss": 3.4101849277791425,
      "tokens_seen": 2147942400
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016487138263665595,
      "loss": 2.7465,
      "theoretical_loss": 3.4101763738585014,
      "tokens_seen": 2148007936
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001648548326082845,
      "loss": 2.6857,
      "theoretical_loss": 3.4101678202719103,
      "tokens_seen": 2148073472
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000164838282579913,
      "loss": 2.4413,
      "theoretical_loss": 3.4101592670193446,
      "tokens_seen": 2148139008
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016482173255154153,
      "loss": 2.4406,
      "theoretical_loss": 3.4101507141007827,
      "tokens_seen": 2148204544
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016480518252317005,
      "loss": 2.3713,
      "theoretical_loss": 3.4101421615162,
      "tokens_seen": 2148270080
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016478863249479856,
      "loss": 2.7479,
      "theoretical_loss": 3.410133609265574,
      "tokens_seen": 2148335616
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016477208246642707,
      "loss": 2.7109,
      "theoretical_loss": 3.410125057348881,
      "tokens_seen": 2148401152
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001647555324380556,
      "loss": 2.4902,
      "theoretical_loss": 3.4101165057660987,
      "tokens_seen": 2148466688
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016473898240968411,
      "loss": 2.5364,
      "theoretical_loss": 3.4101079545172026,
      "tokens_seen": 2148532224
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016472243238131265,
      "loss": 2.5791,
      "theoretical_loss": 3.41009940360217,
      "tokens_seen": 2148597760
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016470588235294116,
      "loss": 2.7904,
      "theoretical_loss": 3.4100908530209786,
      "tokens_seen": 2148663296
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001646893323245697,
      "loss": 2.7246,
      "theoretical_loss": 3.4100823027736036,
      "tokens_seen": 2148728832
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001646727822961982,
      "loss": 2.7479,
      "theoretical_loss": 3.410073752860023,
      "tokens_seen": 2148794368
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016465623226782675,
      "loss": 2.7502,
      "theoretical_loss": 3.4100652032802135,
      "tokens_seen": 2148859904
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016463968223945526,
      "loss": 2.6444,
      "theoretical_loss": 3.410056654034151,
      "tokens_seen": 2148925440
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001646231322110838,
      "loss": 2.6043,
      "theoretical_loss": 3.4100481051218132,
      "tokens_seen": 2148990976
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001646065821827123,
      "loss": 2.7311,
      "theoretical_loss": 3.4100395565431763,
      "tokens_seen": 2149056512
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016459003215434084,
      "loss": 2.4975,
      "theoretical_loss": 3.4100310082982173,
      "tokens_seen": 2149122048
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016457348212596935,
      "loss": 2.3834,
      "theoretical_loss": 3.410022460386913,
      "tokens_seen": 2149187584
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001645569320975979,
      "loss": 2.0978,
      "theoretical_loss": 3.4100139128092404,
      "tokens_seen": 2149253120
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001645403820692264,
      "loss": 2.3004,
      "theoretical_loss": 3.4100053655651763,
      "tokens_seen": 2149318656
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001645238320408549,
      "loss": 2.6075,
      "theoretical_loss": 3.409996818654697,
      "tokens_seen": 2149384192
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016450728201248345,
      "loss": 2.7829,
      "theoretical_loss": 3.4099882720777797,
      "tokens_seen": 2149449728
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016449073198411196,
      "loss": 2.596,
      "theoretical_loss": 3.4099797258344013,
      "tokens_seen": 2149515264
    },
    {
      "epoch": 0.77,
      "objective/train/avg_token_score": 0.016186978667974472,
      "objective/train/avg_weight": 0.9809579849243164,
      "objective/train/docs_used": 1217649,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.635599374771118,
      "objective/train/std_weight": 0.09176909923553467,
      "objective/train/theoretical_loss": 3.4099711799245385,
      "objective/train/tokens_used": 2170040800,
      "theoretical_loss": 3.4099711799245385,
      "tokens_seen": 2149580800
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001644741819557405,
      "loss": 2.4894,
      "theoretical_loss": 3.4099711799245385,
      "tokens_seen": 2149580800
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000164457631927369,
      "loss": 2.7438,
      "theoretical_loss": 3.4099626343481684,
      "tokens_seen": 2149646336
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016444108189899755,
      "loss": 2.518,
      "theoretical_loss": 3.409954089105267,
      "tokens_seen": 2149711872
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016442453187062606,
      "loss": 2.6637,
      "theoretical_loss": 3.4099455441958124,
      "tokens_seen": 2149777408
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001644079818422546,
      "loss": 2.5582,
      "theoretical_loss": 3.40993699961978,
      "tokens_seen": 2149842944
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001643914318138831,
      "loss": 2.6836,
      "theoretical_loss": 3.4099284553771474,
      "tokens_seen": 2149908480
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016437488178551164,
      "loss": 2.7522,
      "theoretical_loss": 3.4099199114678918,
      "tokens_seen": 2149974016
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016435833175714015,
      "loss": 2.6757,
      "theoretical_loss": 3.409911367891989,
      "tokens_seen": 2150039552
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001643417817287687,
      "loss": 2.561,
      "theoretical_loss": 3.4099028246494165,
      "tokens_seen": 2150105088
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001643252317003972,
      "loss": 2.3821,
      "theoretical_loss": 3.409894281740151,
      "tokens_seen": 2150170624
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016430868167202574,
      "loss": 2.645,
      "theoretical_loss": 3.40988573916417,
      "tokens_seen": 2150236160
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016429213164365425,
      "loss": 2.6792,
      "theoretical_loss": 3.4098771969214487,
      "tokens_seen": 2150301696
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016427558161528276,
      "loss": 2.4746,
      "theoretical_loss": 3.409868655011966,
      "tokens_seen": 2150367232
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001642590315869113,
      "loss": 2.7153,
      "theoretical_loss": 3.4098601134356965,
      "tokens_seen": 2150432768
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001642424815585398,
      "loss": 2.7866,
      "theoretical_loss": 3.4098515721926193,
      "tokens_seen": 2150498304
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016422593153016834,
      "loss": 2.5598,
      "theoretical_loss": 3.4098430312827097,
      "tokens_seen": 2150563840
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016420938150179685,
      "loss": 2.5337,
      "theoretical_loss": 3.4098344907059452,
      "tokens_seen": 2150629376
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001641928314734254,
      "loss": 2.7033,
      "theoretical_loss": 3.4098259504623023,
      "tokens_seen": 2150694912
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001641762814450539,
      "loss": 2.7534,
      "theoretical_loss": 3.4098174105517582,
      "tokens_seen": 2150760448
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016415973141668244,
      "loss": 2.2287,
      "theoretical_loss": 3.40980887097429,
      "tokens_seen": 2150825984
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016414318138831095,
      "loss": 2.6709,
      "theoretical_loss": 3.4098003317298735,
      "tokens_seen": 2150891520
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001641266313599395,
      "loss": 2.4891,
      "theoretical_loss": 3.4097917928184867,
      "tokens_seen": 2150957056
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000164110081331568,
      "loss": 2.7473,
      "theoretical_loss": 3.4097832542401063,
      "tokens_seen": 2151022592
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016409353130319654,
      "loss": 2.5524,
      "theoretical_loss": 3.4097747159947085,
      "tokens_seen": 2151088128
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016407698127482505,
      "loss": 2.6078,
      "theoretical_loss": 3.40976617808227,
      "tokens_seen": 2151153664
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016406043124645356,
      "loss": 2.5207,
      "theoretical_loss": 3.409757640502769,
      "tokens_seen": 2151219200
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016404388121808207,
      "loss": 2.7098,
      "theoretical_loss": 3.409749103256181,
      "tokens_seen": 2151284736
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001640273311897106,
      "loss": 2.6766,
      "theoretical_loss": 3.4097405663424842,
      "tokens_seen": 2151350272
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016401078116133912,
      "loss": 2.454,
      "theoretical_loss": 3.4097320297616545,
      "tokens_seen": 2151415808
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016399423113296765,
      "loss": 2.4217,
      "theoretical_loss": 3.4097234935136687,
      "tokens_seen": 2151481344
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016397768110459616,
      "loss": 2.8631,
      "theoretical_loss": 3.4097149575985037,
      "tokens_seen": 2151546880
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001639611310762247,
      "loss": 2.5667,
      "theoretical_loss": 3.4097064220161375,
      "tokens_seen": 2151612416
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001639445810478532,
      "loss": 2.4436,
      "theoretical_loss": 3.409697886766546,
      "tokens_seen": 2151677952
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016392803101948175,
      "loss": 2.6995,
      "theoretical_loss": 3.409689351849706,
      "tokens_seen": 2151743488
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016391148099111026,
      "loss": 2.4266,
      "theoretical_loss": 3.409680817265595,
      "tokens_seen": 2151809024
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001638949309627388,
      "loss": 2.4681,
      "theoretical_loss": 3.4096722830141895,
      "tokens_seen": 2151874560
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001638783809343673,
      "loss": 2.4639,
      "theoretical_loss": 3.409663749095466,
      "tokens_seen": 2151940096
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016386183090599584,
      "loss": 2.5801,
      "theoretical_loss": 3.4096552155094026,
      "tokens_seen": 2152005632
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016384528087762436,
      "loss": 2.4889,
      "theoretical_loss": 3.409646682255975,
      "tokens_seen": 2152071168
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001638287308492529,
      "loss": 2.2793,
      "theoretical_loss": 3.409638149335161,
      "tokens_seen": 2152136704
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001638121808208814,
      "loss": 2.4022,
      "theoretical_loss": 3.4096296167469364,
      "tokens_seen": 2152202240
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016379563079250991,
      "loss": 2.5845,
      "theoretical_loss": 3.4096210844912793,
      "tokens_seen": 2152267776
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016377908076413845,
      "loss": 2.594,
      "theoretical_loss": 3.409612552568166,
      "tokens_seen": 2152333312
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016376253073576696,
      "loss": 2.647,
      "theoretical_loss": 3.4096040209775733,
      "tokens_seen": 2152398848
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001637459807073955,
      "loss": 2.5482,
      "theoretical_loss": 3.4095954897194787,
      "tokens_seen": 2152464384
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000163729430679024,
      "loss": 2.789,
      "theoretical_loss": 3.4095869587938585,
      "tokens_seen": 2152529920
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016371288065065255,
      "loss": 2.5208,
      "theoretical_loss": 3.40957842820069,
      "tokens_seen": 2152595456
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016369633062228106,
      "loss": 2.5173,
      "theoretical_loss": 3.4095698979399502,
      "tokens_seen": 2152660992
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001636797805939096,
      "loss": 2.7322,
      "theoretical_loss": 3.4095613680116155,
      "tokens_seen": 2152726528
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001636632305655381,
      "loss": 2.5577,
      "theoretical_loss": 3.4095528384156633,
      "tokens_seen": 2152792064
    },
    {
      "epoch": 0.77,
      "objective/train/avg_token_score": 0.0202309750020504,
      "objective/train/avg_weight": 0.9761887788772583,
      "objective/train/docs_used": 1219475,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7127838134765625,
      "objective/train/std_weight": 0.0843118205666542,
      "objective/train/theoretical_loss": 3.4095443091520705,
      "objective/train/tokens_used": 2173317600,
      "theoretical_loss": 3.4095443091520705,
      "tokens_seen": 2152857600
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016364668053716664,
      "loss": 2.5276,
      "theoretical_loss": 3.4095443091520705,
      "tokens_seen": 2152857600
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016363013050879515,
      "loss": 2.7085,
      "theoretical_loss": 3.4095357802208133,
      "tokens_seen": 2152923136
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001636135804804237,
      "loss": 2.685,
      "theoretical_loss": 3.4095272516218698,
      "tokens_seen": 2152988672
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001635970304520522,
      "loss": 2.7253,
      "theoretical_loss": 3.4095187233552164,
      "tokens_seen": 2153054208
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016358048042368074,
      "loss": 2.364,
      "theoretical_loss": 3.40951019542083,
      "tokens_seen": 2153119744
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016356393039530925,
      "loss": 2.5594,
      "theoretical_loss": 3.409501667818687,
      "tokens_seen": 2153185280
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016354738036693776,
      "loss": 2.5641,
      "theoretical_loss": 3.409493140548766,
      "tokens_seen": 2153250816
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001635308303385663,
      "loss": 2.6556,
      "theoretical_loss": 3.409484613611042,
      "tokens_seen": 2153316352
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001635142803101948,
      "loss": 2.6301,
      "theoretical_loss": 3.409476087005493,
      "tokens_seen": 2153381888
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016349773028182335,
      "loss": 2.4516,
      "theoretical_loss": 3.4094675607320957,
      "tokens_seen": 2153447424
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016348118025345186,
      "loss": 2.8519,
      "theoretical_loss": 3.4094590347908276,
      "tokens_seen": 2153512960
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001634646302250804,
      "loss": 2.5254,
      "theoretical_loss": 3.4094505091816645,
      "tokens_seen": 2153578496
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001634480801967089,
      "loss": 2.6059,
      "theoretical_loss": 3.4094419839045846,
      "tokens_seen": 2153644032
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016343153016833744,
      "loss": 2.3992,
      "theoretical_loss": 3.409433458959564,
      "tokens_seen": 2153709568
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016341498013996595,
      "loss": 2.3454,
      "theoretical_loss": 3.40942493434658,
      "tokens_seen": 2153775104
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001633984301115945,
      "loss": 2.1845,
      "theoretical_loss": 3.4094164100656092,
      "tokens_seen": 2153840640
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000163381880083223,
      "loss": 2.5146,
      "theoretical_loss": 3.4094078861166297,
      "tokens_seen": 2153906176
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016336533005485154,
      "loss": 2.65,
      "theoretical_loss": 3.4093993624996166,
      "tokens_seen": 2153971712
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016334878002648005,
      "loss": 2.437,
      "theoretical_loss": 3.4093908392145487,
      "tokens_seen": 2154037248
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016333222999810859,
      "loss": 2.5849,
      "theoretical_loss": 3.409382316261402,
      "tokens_seen": 2154102784
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016331567996973707,
      "loss": 2.6438,
      "theoretical_loss": 3.4093737936401536,
      "tokens_seen": 2154168320
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001632991299413656,
      "loss": 2.6526,
      "theoretical_loss": 3.4093652713507803,
      "tokens_seen": 2154233856
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016328257991299412,
      "loss": 2.5735,
      "theoretical_loss": 3.4093567493932597,
      "tokens_seen": 2154299392
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016326602988462265,
      "loss": 2.2973,
      "theoretical_loss": 3.4093482277675684,
      "tokens_seen": 2154364928
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016324947985625116,
      "loss": 2.7937,
      "theoretical_loss": 3.4093397064736832,
      "tokens_seen": 2154430464
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001632329298278797,
      "loss": 2.6257,
      "theoretical_loss": 3.4093311855115815,
      "tokens_seen": 2154496000
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001632163797995082,
      "loss": 2.3943,
      "theoretical_loss": 3.4093226648812403,
      "tokens_seen": 2154561536
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016319982977113675,
      "loss": 2.5481,
      "theoretical_loss": 3.409314144582636,
      "tokens_seen": 2154627072
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016318327974276526,
      "loss": 2.7213,
      "theoretical_loss": 3.4093056246157456,
      "tokens_seen": 2154692608
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001631667297143938,
      "loss": 2.6263,
      "theoretical_loss": 3.4092971049805474,
      "tokens_seen": 2154758144
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001631501796860223,
      "loss": 2.7081,
      "theoretical_loss": 3.4092885856770168,
      "tokens_seen": 2154823680
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016313362965765085,
      "loss": 2.4462,
      "theoretical_loss": 3.4092800667051315,
      "tokens_seen": 2154889216
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016311707962927936,
      "loss": 2.4286,
      "theoretical_loss": 3.4092715480648685,
      "tokens_seen": 2154954752
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001631005296009079,
      "loss": 2.6353,
      "theoretical_loss": 3.4092630297562048,
      "tokens_seen": 2155020288
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001630839795725364,
      "loss": 2.6409,
      "theoretical_loss": 3.4092545117791175,
      "tokens_seen": 2155085824
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016306742954416492,
      "loss": 2.2515,
      "theoretical_loss": 3.4092459941335838,
      "tokens_seen": 2155151360
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016305087951579345,
      "loss": 2.4305,
      "theoretical_loss": 3.4092374768195794,
      "tokens_seen": 2155216896
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016303432948742196,
      "loss": 2.3799,
      "theoretical_loss": 3.4092289598370833,
      "tokens_seen": 2155282432
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001630177794590505,
      "loss": 2.1325,
      "theoretical_loss": 3.4092204431860713,
      "tokens_seen": 2155347968
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000163001229430679,
      "loss": 2.4974,
      "theoretical_loss": 3.4092119268665204,
      "tokens_seen": 2155413504
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016298467940230755,
      "loss": 2.55,
      "theoretical_loss": 3.409203410878408,
      "tokens_seen": 2155479040
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016296812937393606,
      "loss": 2.4683,
      "theoretical_loss": 3.4091948952217113,
      "tokens_seen": 2155544576
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001629515793455646,
      "loss": 2.5133,
      "theoretical_loss": 3.4091863798964064,
      "tokens_seen": 2155610112
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001629350293171931,
      "loss": 2.6657,
      "theoretical_loss": 3.409177864902471,
      "tokens_seen": 2155675648
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016291847928882164,
      "loss": 2.5828,
      "theoretical_loss": 3.4091693502398828,
      "tokens_seen": 2155741184
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016290192926045016,
      "loss": 2.5569,
      "theoretical_loss": 3.409160835908618,
      "tokens_seen": 2155806720
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001628853792320787,
      "loss": 2.255,
      "theoretical_loss": 3.4091523219086532,
      "tokens_seen": 2155872256
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001628688292037072,
      "loss": 2.556,
      "theoretical_loss": 3.4091438082399668,
      "tokens_seen": 2155937792
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016285227917533574,
      "loss": 2.5442,
      "theoretical_loss": 3.4091352949025344,
      "tokens_seen": 2156003328
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016283572914696425,
      "loss": 2.4027,
      "theoretical_loss": 3.4091267818963336,
      "tokens_seen": 2156068864
    },
    {
      "epoch": 0.77,
      "objective/train/avg_token_score": 0.013439245522022247,
      "objective/train/avg_weight": 0.9841875433921814,
      "objective/train/docs_used": 1221633,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.998795986175537,
      "objective/train/std_weight": 0.06728923320770264,
      "objective/train/theoretical_loss": 3.4091182692213415,
      "objective/train/tokens_used": 2176594400,
      "theoretical_loss": 3.4091182692213415,
      "tokens_seen": 2156134400
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016281917911859276,
      "loss": 2.7333,
      "theoretical_loss": 3.4091182692213415,
      "tokens_seen": 2156134400
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001628026290902213,
      "loss": 2.031,
      "theoretical_loss": 3.4091097568775357,
      "tokens_seen": 2156199936
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001627860790618498,
      "loss": 2.5917,
      "theoretical_loss": 3.4091012448648925,
      "tokens_seen": 2156265472
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016276952903347835,
      "loss": 2.7422,
      "theoretical_loss": 3.409092733183389,
      "tokens_seen": 2156331008
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016275297900510686,
      "loss": 2.4792,
      "theoretical_loss": 3.4090842218330026,
      "tokens_seen": 2156396544
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001627364289767354,
      "loss": 2.4166,
      "theoretical_loss": 3.40907571081371,
      "tokens_seen": 2156462080
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001627198789483639,
      "loss": 2.3621,
      "theoretical_loss": 3.409067200125489,
      "tokens_seen": 2156527616
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016270332891999244,
      "loss": 2.5125,
      "theoretical_loss": 3.4090586897683153,
      "tokens_seen": 2156593152
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016268677889162095,
      "loss": 2.3355,
      "theoretical_loss": 3.4090501797421675,
      "tokens_seen": 2156658688
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001626702288632495,
      "loss": 2.1765,
      "theoretical_loss": 3.4090416700470216,
      "tokens_seen": 2156724224
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000162653678834878,
      "loss": 2.5226,
      "theoretical_loss": 3.409033160682855,
      "tokens_seen": 2156789760
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016263712880650654,
      "loss": 2.4134,
      "theoretical_loss": 3.409024651649645,
      "tokens_seen": 2156855296
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016262057877813505,
      "loss": 2.823,
      "theoretical_loss": 3.409016142947368,
      "tokens_seen": 2156920832
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001626040287497636,
      "loss": 2.4958,
      "theoretical_loss": 3.409007634576002,
      "tokens_seen": 2156986368
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016258747872139207,
      "loss": 2.5219,
      "theoretical_loss": 3.408999126535523,
      "tokens_seen": 2157051904
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001625709286930206,
      "loss": 2.4262,
      "theoretical_loss": 3.4089906188259094,
      "tokens_seen": 2157117440
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016255437866464912,
      "loss": 2.3963,
      "theoretical_loss": 3.408982111447137,
      "tokens_seen": 2157182976
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016253782863627766,
      "loss": 2.3679,
      "theoretical_loss": 3.4089736043991836,
      "tokens_seen": 2157248512
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016252127860790617,
      "loss": 2.4129,
      "theoretical_loss": 3.4089650976820263,
      "tokens_seen": 2157314048
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001625047285795347,
      "loss": 2.4865,
      "theoretical_loss": 3.4089565912956417,
      "tokens_seen": 2157379584
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016248817855116321,
      "loss": 2.5623,
      "theoretical_loss": 3.4089480852400076,
      "tokens_seen": 2157445120
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016247162852279175,
      "loss": 2.5136,
      "theoretical_loss": 3.408939579515101,
      "tokens_seen": 2157510656
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016245507849442026,
      "loss": 2.4515,
      "theoretical_loss": 3.408931074120898,
      "tokens_seen": 2157576192
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001624385284660488,
      "loss": 2.4258,
      "theoretical_loss": 3.4089225690573763,
      "tokens_seen": 2157641728
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001624219784376773,
      "loss": 2.5607,
      "theoretical_loss": 3.4089140643245135,
      "tokens_seen": 2157707264
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016240542840930585,
      "loss": 2.7471,
      "theoretical_loss": 3.408905559922286,
      "tokens_seen": 2157772800
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016238887838093436,
      "loss": 2.5397,
      "theoretical_loss": 3.4088970558506717,
      "tokens_seen": 2157838336
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001623723283525629,
      "loss": 2.661,
      "theoretical_loss": 3.408888552109647,
      "tokens_seen": 2157903872
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016235577832419143,
      "loss": 2.3549,
      "theoretical_loss": 3.408880048699189,
      "tokens_seen": 2157969408
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016233922829581992,
      "loss": 2.6297,
      "theoretical_loss": 3.4088715456192755,
      "tokens_seen": 2158034944
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016232267826744845,
      "loss": 2.5145,
      "theoretical_loss": 3.4088630428698825,
      "tokens_seen": 2158100480
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016230612823907696,
      "loss": 2.5189,
      "theoretical_loss": 3.4088545404509882,
      "tokens_seen": 2158166016
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001622895782107055,
      "loss": 2.2396,
      "theoretical_loss": 3.408846038362569,
      "tokens_seen": 2158231552
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000162273028182334,
      "loss": 2.6914,
      "theoretical_loss": 3.4088375366046026,
      "tokens_seen": 2158297088
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016225647815396255,
      "loss": 2.493,
      "theoretical_loss": 3.4088290351770656,
      "tokens_seen": 2158362624
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016223992812559106,
      "loss": 2.4391,
      "theoretical_loss": 3.4088205340799353,
      "tokens_seen": 2158428160
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001622233780972196,
      "loss": 2.4291,
      "theoretical_loss": 3.4088120333131893,
      "tokens_seen": 2158493696
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001622068280688481,
      "loss": 2.8539,
      "theoretical_loss": 3.408803532876804,
      "tokens_seen": 2158559232
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016219027804047665,
      "loss": 2.6625,
      "theoretical_loss": 3.408795032770757,
      "tokens_seen": 2158624768
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016217372801210516,
      "loss": 2.6038,
      "theoretical_loss": 3.408786532995025,
      "tokens_seen": 2158690304
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001621571779837337,
      "loss": 2.3393,
      "theoretical_loss": 3.408778033549585,
      "tokens_seen": 2158755840
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001621406279553622,
      "loss": 2.4722,
      "theoretical_loss": 3.4087695344344153,
      "tokens_seen": 2158821376
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016212407792699074,
      "loss": 2.5245,
      "theoretical_loss": 3.4087610356494924,
      "tokens_seen": 2158886912
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016210752789861925,
      "loss": 2.3375,
      "theoretical_loss": 3.408752537194793,
      "tokens_seen": 2158952448
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016209097787024776,
      "loss": 2.6809,
      "theoretical_loss": 3.408744039070294,
      "tokens_seen": 2159017984
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001620744278418763,
      "loss": 2.5177,
      "theoretical_loss": 3.408735541275974,
      "tokens_seen": 2159083520
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001620578778135048,
      "loss": 2.5766,
      "theoretical_loss": 3.408727043811809,
      "tokens_seen": 2159149056
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016204132778513335,
      "loss": 2.7642,
      "theoretical_loss": 3.4087185466777763,
      "tokens_seen": 2159214592
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016202477775676186,
      "loss": 2.3662,
      "theoretical_loss": 3.4087100498738536,
      "tokens_seen": 2159280128
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001620082277283904,
      "loss": 2.6353,
      "theoretical_loss": 3.408701553400017,
      "tokens_seen": 2159345664
    },
    {
      "epoch": 0.77,
      "objective/train/avg_token_score": 0.014338471926748753,
      "objective/train/avg_weight": 0.9831239581108093,
      "objective/train/docs_used": 1223129,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4873266220092773,
      "objective/train/std_weight": 0.07289232313632965,
      "objective/train/theoretical_loss": 3.4086930572562446,
      "objective/train/tokens_used": 2179871200,
      "theoretical_loss": 3.4086930572562446,
      "tokens_seen": 2159411200
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001619916777000189,
      "loss": 2.1144,
      "theoretical_loss": 3.4086930572562446,
      "tokens_seen": 2159411200
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016197512767164744,
      "loss": 2.4239,
      "theoretical_loss": 3.4086845614425134,
      "tokens_seen": 2159476736
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016195857764327595,
      "loss": 2.4002,
      "theoretical_loss": 3.4086760659588,
      "tokens_seen": 2159542272
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001619420276149045,
      "loss": 2.757,
      "theoretical_loss": 3.408667570805082,
      "tokens_seen": 2159607808
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000161925477586533,
      "loss": 2.5327,
      "theoretical_loss": 3.408659075981337,
      "tokens_seen": 2159673344
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016190892755816154,
      "loss": 2.7424,
      "theoretical_loss": 3.4086505814875414,
      "tokens_seen": 2159738880
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016189237752979005,
      "loss": 2.5871,
      "theoretical_loss": 3.408642087323673,
      "tokens_seen": 2159804416
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001618758275014186,
      "loss": 2.5315,
      "theoretical_loss": 3.4086335934897085,
      "tokens_seen": 2159869952
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001618592774730471,
      "loss": 2.7663,
      "theoretical_loss": 3.4086250999856254,
      "tokens_seen": 2159935488
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001618427274446756,
      "loss": 2.6517,
      "theoretical_loss": 3.4086166068114006,
      "tokens_seen": 2160001024
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016182617741630412,
      "loss": 2.7599,
      "theoretical_loss": 3.408608113967011,
      "tokens_seen": 2160066560
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016180962738793266,
      "loss": 2.4572,
      "theoretical_loss": 3.408599621452435,
      "tokens_seen": 2160132096
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016179307735956117,
      "loss": 2.4472,
      "theoretical_loss": 3.4085911292676485,
      "tokens_seen": 2160197632
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001617765273311897,
      "loss": 2.6259,
      "theoretical_loss": 3.4085826374126293,
      "tokens_seen": 2160263168
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016175997730281822,
      "loss": 2.5759,
      "theoretical_loss": 3.4085741458873544,
      "tokens_seen": 2160328704
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016174342727444675,
      "loss": 2.5698,
      "theoretical_loss": 3.408565654691801,
      "tokens_seen": 2160394240
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001617268772460753,
      "loss": 2.3772,
      "theoretical_loss": 3.408557163825946,
      "tokens_seen": 2160459776
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001617103272177038,
      "loss": 2.7722,
      "theoretical_loss": 3.4085486732897676,
      "tokens_seen": 2160525312
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016169377718933234,
      "loss": 2.4421,
      "theoretical_loss": 3.4085401830832422,
      "tokens_seen": 2160590848
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016167722716096085,
      "loss": 2.5965,
      "theoretical_loss": 3.408531693206347,
      "tokens_seen": 2160656384
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001616606771325894,
      "loss": 2.3957,
      "theoretical_loss": 3.408523203659059,
      "tokens_seen": 2160721920
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001616441271042179,
      "loss": 2.6206,
      "theoretical_loss": 3.4085147144413566,
      "tokens_seen": 2160787456
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016162757707584643,
      "loss": 2.6729,
      "theoretical_loss": 3.4085062255532157,
      "tokens_seen": 2160852992
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016161102704747495,
      "loss": 2.7667,
      "theoretical_loss": 3.408497736994614,
      "tokens_seen": 2160918528
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016159447701910346,
      "loss": 2.5665,
      "theoretical_loss": 3.4084892487655285,
      "tokens_seen": 2160984064
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016157792699073197,
      "loss": 2.4086,
      "theoretical_loss": 3.408480760865937,
      "tokens_seen": 2161049600
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001615613769623605,
      "loss": 2.3361,
      "theoretical_loss": 3.408472273295816,
      "tokens_seen": 2161115136
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016154482693398901,
      "loss": 2.6832,
      "theoretical_loss": 3.408463786055143,
      "tokens_seen": 2161180672
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016152827690561755,
      "loss": 2.557,
      "theoretical_loss": 3.408455299143896,
      "tokens_seen": 2161246208
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016151172687724606,
      "loss": 2.5535,
      "theoretical_loss": 3.408446812562051,
      "tokens_seen": 2161311744
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001614951768488746,
      "loss": 2.7307,
      "theoretical_loss": 3.4084383263095854,
      "tokens_seen": 2161377280
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001614786268205031,
      "loss": 2.6597,
      "theoretical_loss": 3.4084298403864772,
      "tokens_seen": 2161442816
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016146207679213165,
      "loss": 2.5949,
      "theoretical_loss": 3.408421354792703,
      "tokens_seen": 2161508352
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016144552676376016,
      "loss": 2.4917,
      "theoretical_loss": 3.4084128695282403,
      "tokens_seen": 2161573888
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001614289767353887,
      "loss": 2.423,
      "theoretical_loss": 3.408404384593066,
      "tokens_seen": 2161639424
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001614124267070172,
      "loss": 2.4677,
      "theoretical_loss": 3.4083958999871578,
      "tokens_seen": 2161704960
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016139587667864574,
      "loss": 2.24,
      "theoretical_loss": 3.408387415710493,
      "tokens_seen": 2161770496
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016137932665027425,
      "loss": 2.4115,
      "theoretical_loss": 3.408378931763048,
      "tokens_seen": 2161836032
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001613627766219028,
      "loss": 2.2318,
      "theoretical_loss": 3.4083704481448014,
      "tokens_seen": 2161901568
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001613462265935313,
      "loss": 2.684,
      "theoretical_loss": 3.408361964855729,
      "tokens_seen": 2161967104
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001613296765651598,
      "loss": 2.6434,
      "theoretical_loss": 3.408353481895809,
      "tokens_seen": 2162032640
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016131312653678835,
      "loss": 2.4252,
      "theoretical_loss": 3.4083449992650183,
      "tokens_seen": 2162098176
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016129657650841686,
      "loss": 2.5659,
      "theoretical_loss": 3.408336516963334,
      "tokens_seen": 2162163712
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001612800264800454,
      "loss": 2.4745,
      "theoretical_loss": 3.408328034990734,
      "tokens_seen": 2162229248
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001612634764516739,
      "loss": 2.6529,
      "theoretical_loss": 3.4083195533471953,
      "tokens_seen": 2162294784
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016124692642330245,
      "loss": 2.6282,
      "theoretical_loss": 3.4083110720326943,
      "tokens_seen": 2162360320
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016123037639493096,
      "loss": 2.5287,
      "theoretical_loss": 3.4083025910472093,
      "tokens_seen": 2162425856
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001612138263665595,
      "loss": 2.2646,
      "theoretical_loss": 3.4082941103907176,
      "tokens_seen": 2162491392
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000161197276338188,
      "loss": 2.4588,
      "theoretical_loss": 3.4082856300631956,
      "tokens_seen": 2162556928
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016118072630981654,
      "loss": 2.457,
      "theoretical_loss": 3.408277150064621,
      "tokens_seen": 2162622464
    },
    {
      "epoch": 0.77,
      "objective/train/avg_token_score": 0.01191009022295475,
      "objective/train/avg_weight": 0.9859840273857117,
      "objective/train/docs_used": 1224854,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7200255393981934,
      "objective/train/std_weight": 0.05838380753993988,
      "objective/train/theoretical_loss": 3.4082686703949716,
      "objective/train/tokens_used": 2183148000,
      "theoretical_loss": 3.4082686703949716,
      "tokens_seen": 2162688000
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016116417628144505,
      "loss": 2.4861,
      "theoretical_loss": 3.4082686703949716,
      "tokens_seen": 2162688000
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001611476262530736,
      "loss": 2.5331,
      "theoretical_loss": 3.4082601910542243,
      "tokens_seen": 2162753536
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001611310762247021,
      "loss": 2.6575,
      "theoretical_loss": 3.4082517120423557,
      "tokens_seen": 2162819072
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016111452619633064,
      "loss": 2.6798,
      "theoretical_loss": 3.408243233359344,
      "tokens_seen": 2162884608
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016111452619633064,
      "loss": 4.2259,
      "theoretical_loss": 3.408234755005166,
      "tokens_seen": 2162950144
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016109797616795915,
      "loss": 2.4895,
      "theoretical_loss": 3.4082262769797995,
      "tokens_seen": 2163015680
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016108142613958766,
      "loss": 2.6228,
      "theoretical_loss": 3.4082177992832214,
      "tokens_seen": 2163081216
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001610648761112162,
      "loss": 2.3102,
      "theoretical_loss": 3.408209321915409,
      "tokens_seen": 2163146752
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001610483260828447,
      "loss": 2.6785,
      "theoretical_loss": 3.4082008448763395,
      "tokens_seen": 2163212288
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016103177605447324,
      "loss": 2.7179,
      "theoretical_loss": 3.40819236816599,
      "tokens_seen": 2163277824
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016101522602610175,
      "loss": 2.3275,
      "theoretical_loss": 3.408183891784338,
      "tokens_seen": 2163343360
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001609986759977303,
      "loss": 2.613,
      "theoretical_loss": 3.4081754157313617,
      "tokens_seen": 2163408896
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001609821259693588,
      "loss": 2.4304,
      "theoretical_loss": 3.408166940007037,
      "tokens_seen": 2163474432
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016096557594098734,
      "loss": 2.4573,
      "theoretical_loss": 3.408158464611342,
      "tokens_seen": 2163539968
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016094902591261585,
      "loss": 2.5912,
      "theoretical_loss": 3.4081499895442535,
      "tokens_seen": 2163605504
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001609324758842444,
      "loss": 2.6744,
      "theoretical_loss": 3.4081415148057497,
      "tokens_seen": 2163671040
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001609159258558729,
      "loss": 2.7478,
      "theoretical_loss": 3.4081330403958066,
      "tokens_seen": 2163736576
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016089937582750144,
      "loss": 2.7524,
      "theoretical_loss": 3.4081245663144024,
      "tokens_seen": 2163802112
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016088282579912995,
      "loss": 2.8329,
      "theoretical_loss": 3.4081160925615146,
      "tokens_seen": 2163867648
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016086627577075846,
      "loss": 2.5883,
      "theoretical_loss": 3.40810761913712,
      "tokens_seen": 2163933184
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016084972574238697,
      "loss": 2.3796,
      "theoretical_loss": 3.408099146041196,
      "tokens_seen": 2163998720
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001608331757140155,
      "loss": 2.7186,
      "theoretical_loss": 3.40809067327372,
      "tokens_seen": 2164064256
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016081662568564402,
      "loss": 2.6349,
      "theoretical_loss": 3.408082200834669,
      "tokens_seen": 2164129792
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016080007565727255,
      "loss": 2.2573,
      "theoretical_loss": 3.408073728724021,
      "tokens_seen": 2164195328
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016078352562890106,
      "loss": 2.7537,
      "theoretical_loss": 3.408065256941753,
      "tokens_seen": 2164260864
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001607669756005296,
      "loss": 2.7007,
      "theoretical_loss": 3.408056785487842,
      "tokens_seen": 2164326400
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001607504255721581,
      "loss": 2.4742,
      "theoretical_loss": 3.408048314362266,
      "tokens_seen": 2164391936
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016073387554378665,
      "loss": 2.498,
      "theoretical_loss": 3.408039843565002,
      "tokens_seen": 2164457472
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016071732551541516,
      "loss": 2.5705,
      "theoretical_loss": 3.4080313730960268,
      "tokens_seen": 2164523008
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001607007754870437,
      "loss": 2.7397,
      "theoretical_loss": 3.4080229029553184,
      "tokens_seen": 2164588544
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001606842254586722,
      "loss": 2.636,
      "theoretical_loss": 3.408014433142854,
      "tokens_seen": 2164654080
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016066767543030074,
      "loss": 2.4323,
      "theoretical_loss": 3.408005963658611,
      "tokens_seen": 2164719616
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016065112540192926,
      "loss": 2.6666,
      "theoretical_loss": 3.407997494502567,
      "tokens_seen": 2164785152
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001606345753735578,
      "loss": 2.6064,
      "theoretical_loss": 3.4079890256746985,
      "tokens_seen": 2164850688
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001606180253451863,
      "loss": 2.6777,
      "theoretical_loss": 3.4079805571749833,
      "tokens_seen": 2164916224
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016060147531681481,
      "loss": 2.2679,
      "theoretical_loss": 3.4079720890033993,
      "tokens_seen": 2164981760
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016058492528844335,
      "loss": 2.4791,
      "theoretical_loss": 3.4079636211599227,
      "tokens_seen": 2165047296
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016056837526007186,
      "loss": 2.7272,
      "theoretical_loss": 3.407955153644532,
      "tokens_seen": 2165112832
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001605518252317004,
      "loss": 2.4264,
      "theoretical_loss": 3.407946686457204,
      "tokens_seen": 2165178368
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001605352752033289,
      "loss": 2.604,
      "theoretical_loss": 3.407938219597916,
      "tokens_seen": 2165243904
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016051872517495745,
      "loss": 2.7082,
      "theoretical_loss": 3.4079297530666457,
      "tokens_seen": 2165309440
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016050217514658596,
      "loss": 2.4434,
      "theoretical_loss": 3.40792128686337,
      "tokens_seen": 2165374976
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001604856251182145,
      "loss": 2.3446,
      "theoretical_loss": 3.4079128209880665,
      "tokens_seen": 2165440512
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000160469075089843,
      "loss": 2.6307,
      "theoretical_loss": 3.4079043554407127,
      "tokens_seen": 2165506048
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016045252506147154,
      "loss": 2.4965,
      "theoretical_loss": 3.407895890221286,
      "tokens_seen": 2165571584
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016043597503310005,
      "loss": 2.6538,
      "theoretical_loss": 3.4078874253297635,
      "tokens_seen": 2165637120
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001604194250047286,
      "loss": 2.6084,
      "theoretical_loss": 3.407878960766123,
      "tokens_seen": 2165702656
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001604028749763571,
      "loss": 2.4721,
      "theoretical_loss": 3.407870496530341,
      "tokens_seen": 2165768192
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016038632494798564,
      "loss": 2.7583,
      "theoretical_loss": 3.4078620326223956,
      "tokens_seen": 2165833728
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016036977491961415,
      "loss": 2.6045,
      "theoretical_loss": 3.407853569042264,
      "tokens_seen": 2165899264
    },
    {
      "epoch": 0.77,
      "objective/train/avg_token_score": 0.011187296360731125,
      "objective/train/avg_weight": 0.9868378639221191,
      "objective/train/docs_used": 1226563,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.672165870666504,
      "objective/train/std_weight": 0.05638757720589638,
      "objective/train/theoretical_loss": 3.4078451057899244,
      "objective/train/tokens_used": 2186424800,
      "theoretical_loss": 3.4078451057899244,
      "tokens_seen": 2165964800
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016035322489124266,
      "loss": 2.7785,
      "theoretical_loss": 3.4078451057899244,
      "tokens_seen": 2165964800
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001603366748628712,
      "loss": 2.6021,
      "theoretical_loss": 3.4078366428653526,
      "tokens_seen": 2166030336
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001603201248344997,
      "loss": 2.5684,
      "theoretical_loss": 3.4078281802685275,
      "tokens_seen": 2166095872
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016030357480612825,
      "loss": 2.5855,
      "theoretical_loss": 3.407819717999425,
      "tokens_seen": 2166161408
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016028702477775676,
      "loss": 2.7307,
      "theoretical_loss": 3.4078112560580243,
      "tokens_seen": 2166226944
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001602704747493853,
      "loss": 2.5433,
      "theoretical_loss": 3.4078027944443012,
      "tokens_seen": 2166292480
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001602539247210138,
      "loss": 2.7437,
      "theoretical_loss": 3.407794333158234,
      "tokens_seen": 2166358016
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016023737469264234,
      "loss": 2.5891,
      "theoretical_loss": 3.407785872199799,
      "tokens_seen": 2166423552
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016022082466427085,
      "loss": 2.4663,
      "theoretical_loss": 3.407777411568975,
      "tokens_seen": 2166489088
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001602042746358994,
      "loss": 2.5292,
      "theoretical_loss": 3.407768951265739,
      "tokens_seen": 2166554624
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001601877246075279,
      "loss": 2.5571,
      "theoretical_loss": 3.407760491290068,
      "tokens_seen": 2166620160
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016017117457915644,
      "loss": 2.3031,
      "theoretical_loss": 3.4077520316419396,
      "tokens_seen": 2166685696
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016015462455078495,
      "loss": 2.6213,
      "theoretical_loss": 3.4077435723213316,
      "tokens_seen": 2166751232
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016013807452241349,
      "loss": 2.7525,
      "theoretical_loss": 3.4077351133282203,
      "tokens_seen": 2166816768
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016012152449404197,
      "loss": 2.6512,
      "theoretical_loss": 3.407726654662585,
      "tokens_seen": 2166882304
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001601049744656705,
      "loss": 2.653,
      "theoretical_loss": 3.4077181963244008,
      "tokens_seen": 2166947840
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016008842443729902,
      "loss": 2.6261,
      "theoretical_loss": 3.4077097383136468,
      "tokens_seen": 2167013376
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016007187440892755,
      "loss": 2.4234,
      "theoretical_loss": 3.4077012806302998,
      "tokens_seen": 2167078912
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016005532438055606,
      "loss": 2.6848,
      "theoretical_loss": 3.4076928232743375,
      "tokens_seen": 2167144448
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001600387743521846,
      "loss": 2.5812,
      "theoretical_loss": 3.4076843662457375,
      "tokens_seen": 2167209984
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001600222243238131,
      "loss": 2.3942,
      "theoretical_loss": 3.4076759095444764,
      "tokens_seen": 2167275520
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00016000567429544165,
      "loss": 2.5121,
      "theoretical_loss": 3.407667453170532,
      "tokens_seen": 2167341056
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015998912426707016,
      "loss": 2.5787,
      "theoretical_loss": 3.4076589971238826,
      "tokens_seen": 2167406592
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001599725742386987,
      "loss": 2.6408,
      "theoretical_loss": 3.4076505414045046,
      "tokens_seen": 2167472128
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001599560242103272,
      "loss": 2.466,
      "theoretical_loss": 3.4076420860123755,
      "tokens_seen": 2167537664
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015993947418195575,
      "loss": 2.5632,
      "theoretical_loss": 3.4076336309474735,
      "tokens_seen": 2167603200
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015992292415358426,
      "loss": 2.4983,
      "theoretical_loss": 3.407625176209775,
      "tokens_seen": 2167668736
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001599063741252128,
      "loss": 2.7587,
      "theoretical_loss": 3.407616721799258,
      "tokens_seen": 2167734272
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001598898240968413,
      "loss": 2.8894,
      "theoretical_loss": 3.4076082677159003,
      "tokens_seen": 2167799808
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015987327406846982,
      "loss": 2.3156,
      "theoretical_loss": 3.4075998139596786,
      "tokens_seen": 2167865344
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015985672404009835,
      "loss": 2.7632,
      "theoretical_loss": 3.407591360530571,
      "tokens_seen": 2167930880
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015984017401172686,
      "loss": 2.5547,
      "theoretical_loss": 3.4075829074285546,
      "tokens_seen": 2167996416
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001598236239833554,
      "loss": 2.7739,
      "theoretical_loss": 3.4075744546536075,
      "tokens_seen": 2168061952
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001598070739549839,
      "loss": 2.6112,
      "theoretical_loss": 3.407566002205706,
      "tokens_seen": 2168127488
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015979052392661245,
      "loss": 2.6924,
      "theoretical_loss": 3.407557550084828,
      "tokens_seen": 2168193024
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015977397389824096,
      "loss": 2.5492,
      "theoretical_loss": 3.4075490982909518,
      "tokens_seen": 2168258560
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001597574238698695,
      "loss": 2.6609,
      "theoretical_loss": 3.4075406468240534,
      "tokens_seen": 2168324096
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000159740873841498,
      "loss": 2.3764,
      "theoretical_loss": 3.4075321956841114,
      "tokens_seen": 2168389632
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015972432381312654,
      "loss": 2.6718,
      "theoretical_loss": 3.4075237448711033,
      "tokens_seen": 2168455168
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015970777378475506,
      "loss": 2.4996,
      "theoretical_loss": 3.4075152943850058,
      "tokens_seen": 2168520704
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001596912237563836,
      "loss": 2.5652,
      "theoretical_loss": 3.407506844225797,
      "tokens_seen": 2168586240
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001596746737280121,
      "loss": 2.419,
      "theoretical_loss": 3.4074983943934543,
      "tokens_seen": 2168651776
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015965812369964064,
      "loss": 2.7029,
      "theoretical_loss": 3.4074899448879545,
      "tokens_seen": 2168717312
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015964157367126915,
      "loss": 2.4515,
      "theoretical_loss": 3.407481495709276,
      "tokens_seen": 2168782848
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015962502364289766,
      "loss": 2.5409,
      "theoretical_loss": 3.4074730468573957,
      "tokens_seen": 2168848384
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001596084736145262,
      "loss": 2.6805,
      "theoretical_loss": 3.4074645983322913,
      "tokens_seen": 2168913920
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001595919235861547,
      "loss": 2.8002,
      "theoretical_loss": 3.4074561501339407,
      "tokens_seen": 2168979456
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015957537355778325,
      "loss": 2.7136,
      "theoretical_loss": 3.4074477022623206,
      "tokens_seen": 2169044992
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015955882352941176,
      "loss": 2.4772,
      "theoretical_loss": 3.407439254717409,
      "tokens_seen": 2169110528
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001595422735010403,
      "loss": 2.4301,
      "theoretical_loss": 3.407430807499183,
      "tokens_seen": 2169176064
    },
    {
      "epoch": 0.77,
      "objective/train/avg_token_score": 0.019475087523460388,
      "objective/train/avg_weight": 0.9770709276199341,
      "objective/train/docs_used": 1228356,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.906156063079834,
      "objective/train/std_weight": 0.11326521635055542,
      "objective/train/theoretical_loss": 3.4074223606076206,
      "objective/train/tokens_used": 2189701600,
      "theoretical_loss": 3.4074223606076206,
      "tokens_seen": 2169241600
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001595257234726688,
      "loss": 2.5236,
      "theoretical_loss": 3.4074223606076206,
      "tokens_seen": 2169241600
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015950917344429734,
      "loss": 2.3106,
      "theoretical_loss": 3.407413914042699,
      "tokens_seen": 2169307136
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015949262341592585,
      "loss": 2.7531,
      "theoretical_loss": 3.407405467804396,
      "tokens_seen": 2169372672
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001594760733875544,
      "loss": 2.4856,
      "theoretical_loss": 3.407397021892688,
      "tokens_seen": 2169438208
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001594595233591829,
      "loss": 2.3699,
      "theoretical_loss": 3.407388576307554,
      "tokens_seen": 2169503744
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015944297333081144,
      "loss": 2.481,
      "theoretical_loss": 3.407380131048971,
      "tokens_seen": 2169569280
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015942642330243995,
      "loss": 2.8017,
      "theoretical_loss": 3.407371686116916,
      "tokens_seen": 2169634816
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001594098732740685,
      "loss": 2.7326,
      "theoretical_loss": 3.4073632415113675,
      "tokens_seen": 2169700352
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000159393323245697,
      "loss": 2.6747,
      "theoretical_loss": 3.4073547972323017,
      "tokens_seen": 2169765888
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001593767732173255,
      "loss": 2.7741,
      "theoretical_loss": 3.407346353279697,
      "tokens_seen": 2169831424
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015936022318895402,
      "loss": 2.5516,
      "theoretical_loss": 3.4073379096535312,
      "tokens_seen": 2169896960
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00015934367316058256,
      "loss": 2.402,
      "theoretical_loss": 3.4073294663537808,
      "tokens_seen": 2169962496
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015932712313221107,
      "loss": 2.7343,
      "theoretical_loss": 3.4073210233804243,
      "tokens_seen": 2170028032
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001593105731038396,
      "loss": 2.6832,
      "theoretical_loss": 3.4073125807334383,
      "tokens_seen": 2170093568
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015929402307546811,
      "loss": 2.6346,
      "theoretical_loss": 3.407304138412801,
      "tokens_seen": 2170159104
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015927747304709665,
      "loss": 2.4244,
      "theoretical_loss": 3.4072956964184904,
      "tokens_seen": 2170224640
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015926092301872516,
      "loss": 2.4443,
      "theoretical_loss": 3.4072872547504827,
      "tokens_seen": 2170290176
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001592443729903537,
      "loss": 2.7488,
      "theoretical_loss": 3.4072788134087566,
      "tokens_seen": 2170355712
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001592278229619822,
      "loss": 2.7314,
      "theoretical_loss": 3.407270372393289,
      "tokens_seen": 2170421248
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015921127293361075,
      "loss": 2.6995,
      "theoretical_loss": 3.4072619317040576,
      "tokens_seen": 2170486784
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015919472290523926,
      "loss": 2.8284,
      "theoretical_loss": 3.4072534913410397,
      "tokens_seen": 2170552320
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001591781728768678,
      "loss": 2.1698,
      "theoretical_loss": 3.4072450513042134,
      "tokens_seen": 2170617856
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001591616228484963,
      "loss": 2.46,
      "theoretical_loss": 3.407236611593556,
      "tokens_seen": 2170683392
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015914507282012482,
      "loss": 2.4818,
      "theoretical_loss": 3.4072281722090447,
      "tokens_seen": 2170748928
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015912852279175335,
      "loss": 2.7204,
      "theoretical_loss": 3.4072197331506575,
      "tokens_seen": 2170814464
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015911197276338186,
      "loss": 2.6085,
      "theoretical_loss": 3.407211294418372,
      "tokens_seen": 2170880000
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001590954227350104,
      "loss": 2.2438,
      "theoretical_loss": 3.4072028560121654,
      "tokens_seen": 2170945536
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001590788727066389,
      "loss": 2.8867,
      "theoretical_loss": 3.4071944179320153,
      "tokens_seen": 2171011072
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015906232267826745,
      "loss": 2.4383,
      "theoretical_loss": 3.4071859801778994,
      "tokens_seen": 2171076608
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015904577264989596,
      "loss": 2.4926,
      "theoretical_loss": 3.4071775427497952,
      "tokens_seen": 2171142144
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001590292226215245,
      "loss": 2.8168,
      "theoretical_loss": 3.4071691056476805,
      "tokens_seen": 2171207680
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000159012672593153,
      "loss": 2.5529,
      "theoretical_loss": 3.4071606688715326,
      "tokens_seen": 2171273216
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015899612256478155,
      "loss": 2.5096,
      "theoretical_loss": 3.4071522324213293,
      "tokens_seen": 2171338752
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015897957253641006,
      "loss": 2.4502,
      "theoretical_loss": 3.4071437962970474,
      "tokens_seen": 2171404288
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001589630225080386,
      "loss": 2.4613,
      "theoretical_loss": 3.407135360498666,
      "tokens_seen": 2171469824
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001589464724796671,
      "loss": 2.5741,
      "theoretical_loss": 3.4071269250261604,
      "tokens_seen": 2171535360
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015892992245129564,
      "loss": 2.5266,
      "theoretical_loss": 3.407118489879511,
      "tokens_seen": 2171600896
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015891337242292415,
      "loss": 2.6027,
      "theoretical_loss": 3.407110055058693,
      "tokens_seen": 2171666432
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015889682239455266,
      "loss": 2.8591,
      "theoretical_loss": 3.407101620563685,
      "tokens_seen": 2171731968
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001588802723661812,
      "loss": 2.4841,
      "theoretical_loss": 3.4070931863944645,
      "tokens_seen": 2171797504
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001588637223378097,
      "loss": 2.4491,
      "theoretical_loss": 3.4070847525510093,
      "tokens_seen": 2171863040
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015884717230943825,
      "loss": 2.7212,
      "theoretical_loss": 3.4070763190332967,
      "tokens_seen": 2171928576
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015883062228106676,
      "loss": 2.635,
      "theoretical_loss": 3.407067885841304,
      "tokens_seen": 2171994112
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001588140722526953,
      "loss": 2.5409,
      "theoretical_loss": 3.4070594529750093,
      "tokens_seen": 2172059648
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001587975222243238,
      "loss": 2.7517,
      "theoretical_loss": 3.4070510204343902,
      "tokens_seen": 2172125184
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015878097219595234,
      "loss": 2.5808,
      "theoretical_loss": 3.407042588219424,
      "tokens_seen": 2172190720
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015876442216758085,
      "loss": 2.1103,
      "theoretical_loss": 3.407034156330088,
      "tokens_seen": 2172256256
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001587478721392094,
      "loss": 2.3852,
      "theoretical_loss": 3.4070257247663607,
      "tokens_seen": 2172321792
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001587313221108379,
      "loss": 2.6735,
      "theoretical_loss": 3.4070172935282192,
      "tokens_seen": 2172387328
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015871477208246644,
      "loss": 2.4445,
      "theoretical_loss": 3.407008862615641,
      "tokens_seen": 2172452864
    },
    {
      "epoch": 0.78,
      "objective/train/avg_token_score": 0.011331364512443542,
      "objective/train/avg_weight": 0.9867072105407715,
      "objective/train/docs_used": 1230300,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.275857448577881,
      "objective/train/std_weight": 0.07802356779575348,
      "objective/train/theoretical_loss": 3.4070004320286036,
      "objective/train/tokens_used": 2192978400,
      "theoretical_loss": 3.4070004320286036,
      "tokens_seen": 2172518400
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015869822205409495,
      "loss": 2.5216,
      "theoretical_loss": 3.4070004320286036,
      "tokens_seen": 2172518400
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001586816720257235,
      "loss": 2.6189,
      "theoretical_loss": 3.4069920017670854,
      "tokens_seen": 2172583936
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000158665121997352,
      "loss": 2.2594,
      "theoretical_loss": 3.4069835718310633,
      "tokens_seen": 2172649472
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001586485719689805,
      "loss": 2.6972,
      "theoretical_loss": 3.4069751422205146,
      "tokens_seen": 2172715008
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015863202194060902,
      "loss": 2.6046,
      "theoretical_loss": 3.406966712935418,
      "tokens_seen": 2172780544
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015861547191223756,
      "loss": 2.3408,
      "theoretical_loss": 3.40695828397575,
      "tokens_seen": 2172846080
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015859892188386607,
      "loss": 2.6315,
      "theoretical_loss": 3.406949855341489,
      "tokens_seen": 2172911616
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001585823718554946,
      "loss": 2.5978,
      "theoretical_loss": 3.4069414270326126,
      "tokens_seen": 2172977152
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015856582182712312,
      "loss": 2.6083,
      "theoretical_loss": 3.4069329990490975,
      "tokens_seen": 2173042688
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015854927179875165,
      "loss": 2.3593,
      "theoretical_loss": 3.4069245713909226,
      "tokens_seen": 2173108224
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015853272177038016,
      "loss": 2.7874,
      "theoretical_loss": 3.406916144058065,
      "tokens_seen": 2173173760
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001585161717420087,
      "loss": 2.4759,
      "theoretical_loss": 3.4069077170505015,
      "tokens_seen": 2173239296
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001584996217136372,
      "loss": 2.6024,
      "theoretical_loss": 3.4068992903682114,
      "tokens_seen": 2173304832
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015848307168526575,
      "loss": 2.7847,
      "theoretical_loss": 3.406890864011171,
      "tokens_seen": 2173370368
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015846652165689426,
      "loss": 2.4603,
      "theoretical_loss": 3.406882437979358,
      "tokens_seen": 2173435904
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001584499716285228,
      "loss": 2.7441,
      "theoretical_loss": 3.406874012272751,
      "tokens_seen": 2173501440
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001584334216001513,
      "loss": 2.515,
      "theoretical_loss": 3.4068655868913273,
      "tokens_seen": 2173566976
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015841687157177985,
      "loss": 2.5206,
      "theoretical_loss": 3.406857161835064,
      "tokens_seen": 2173632512
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015840032154340836,
      "loss": 2.4496,
      "theoretical_loss": 3.406848737103939,
      "tokens_seen": 2173698048
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015838377151503687,
      "loss": 2.7868,
      "theoretical_loss": 3.4068403126979296,
      "tokens_seen": 2173763584
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001583672214866654,
      "loss": 2.5314,
      "theoretical_loss": 3.406831888617014,
      "tokens_seen": 2173829120
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015835067145829391,
      "loss": 2.5496,
      "theoretical_loss": 3.40682346486117,
      "tokens_seen": 2173894656
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015833412142992245,
      "loss": 2.5779,
      "theoretical_loss": 3.4068150414303755,
      "tokens_seen": 2173960192
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015831757140155096,
      "loss": 2.624,
      "theoretical_loss": 3.406806618324607,
      "tokens_seen": 2174025728
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001583010213731795,
      "loss": 2.5673,
      "theoretical_loss": 3.4067981955438427,
      "tokens_seen": 2174091264
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000158284471344808,
      "loss": 2.2231,
      "theoretical_loss": 3.4067897730880605,
      "tokens_seen": 2174156800
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015826792131643655,
      "loss": 2.6621,
      "theoretical_loss": 3.406781350957238,
      "tokens_seen": 2174222336
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015825137128806506,
      "loss": 2.4814,
      "theoretical_loss": 3.406772929151353,
      "tokens_seen": 2174287872
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001582348212596936,
      "loss": 2.6372,
      "theoretical_loss": 3.4067645076703825,
      "tokens_seen": 2174353408
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001582182712313221,
      "loss": 2.396,
      "theoretical_loss": 3.4067560865143047,
      "tokens_seen": 2174418944
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015820172120295064,
      "loss": 2.6135,
      "theoretical_loss": 3.4067476656830973,
      "tokens_seen": 2174484480
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015818517117457915,
      "loss": 2.8075,
      "theoretical_loss": 3.406739245176738,
      "tokens_seen": 2174550016
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001581686211462077,
      "loss": 2.6362,
      "theoretical_loss": 3.406730824995204,
      "tokens_seen": 2174615552
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001581520711178362,
      "loss": 2.236,
      "theoretical_loss": 3.406722405138474,
      "tokens_seen": 2174681088
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001581355210894647,
      "loss": 2.3812,
      "theoretical_loss": 3.4067139856065247,
      "tokens_seen": 2174746624
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015811897106109325,
      "loss": 2.6139,
      "theoretical_loss": 3.406705566399334,
      "tokens_seen": 2174812160
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015810242103272176,
      "loss": 2.3838,
      "theoretical_loss": 3.40669714751688,
      "tokens_seen": 2174877696
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001580858710043503,
      "loss": 2.7148,
      "theoretical_loss": 3.4066887289591397,
      "tokens_seen": 2174943232
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001580693209759788,
      "loss": 2.572,
      "theoretical_loss": 3.4066803107260912,
      "tokens_seen": 2175008768
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015805277094760735,
      "loss": 2.3538,
      "theoretical_loss": 3.4066718928177124,
      "tokens_seen": 2175074304
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015803622091923586,
      "loss": 2.3182,
      "theoretical_loss": 3.4066634752339806,
      "tokens_seen": 2175139840
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001580196708908644,
      "loss": 2.4842,
      "theoretical_loss": 3.4066550579748736,
      "tokens_seen": 2175205376
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001580031208624929,
      "loss": 2.6566,
      "theoretical_loss": 3.4066466410403695,
      "tokens_seen": 2175270912
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015798657083412144,
      "loss": 2.3716,
      "theoretical_loss": 3.4066382244304454,
      "tokens_seen": 2175336448
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015797002080574995,
      "loss": 2.4097,
      "theoretical_loss": 3.4066298081450794,
      "tokens_seen": 2175401984
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001579534707773785,
      "loss": 2.3939,
      "theoretical_loss": 3.406621392184249,
      "tokens_seen": 2175467520
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000157936920749007,
      "loss": 2.6132,
      "theoretical_loss": 3.4066129765479323,
      "tokens_seen": 2175533056
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015792037072063554,
      "loss": 2.5149,
      "theoretical_loss": 3.406604561236106,
      "tokens_seen": 2175598592
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015790382069226402,
      "loss": 2.7328,
      "theoretical_loss": 3.406596146248749,
      "tokens_seen": 2175664128
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015788727066389256,
      "loss": 2.5,
      "theoretical_loss": 3.4065877315858386,
      "tokens_seen": 2175729664
    },
    {
      "epoch": 0.78,
      "objective/train/avg_token_score": 0.027311909943819046,
      "objective/train/avg_weight": 0.9678521156311035,
      "objective/train/docs_used": 1232394,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.125718116760254,
      "objective/train/std_weight": 0.11969346553087234,
      "objective/train/theoretical_loss": 3.4065793172473526,
      "objective/train/tokens_used": 2196255200,
      "theoretical_loss": 3.4065793172473526,
      "tokens_seen": 2175795200
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015787072063552107,
      "loss": 2.7203,
      "theoretical_loss": 3.4065793172473526,
      "tokens_seen": 2175795200
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001578541706071496,
      "loss": 2.899,
      "theoretical_loss": 3.406570903233268,
      "tokens_seen": 2175860736
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015783762057877812,
      "loss": 2.6595,
      "theoretical_loss": 3.4065624895435636,
      "tokens_seen": 2175926272
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015782107055040665,
      "loss": 2.2558,
      "theoretical_loss": 3.4065540761782165,
      "tokens_seen": 2175991808
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015780452052203517,
      "loss": 2.8041,
      "theoretical_loss": 3.406545663137204,
      "tokens_seen": 2176057344
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001577879704936637,
      "loss": 2.7021,
      "theoretical_loss": 3.4065372504205054,
      "tokens_seen": 2176122880
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001577714204652922,
      "loss": 2.2308,
      "theoretical_loss": 3.406528838028097,
      "tokens_seen": 2176188416
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015775487043692075,
      "loss": 2.2407,
      "theoretical_loss": 3.4065204259599566,
      "tokens_seen": 2176253952
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001577383204085493,
      "loss": 2.4099,
      "theoretical_loss": 3.4065120142160628,
      "tokens_seen": 2176319488
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001577217703801778,
      "loss": 2.5029,
      "theoretical_loss": 3.4065036027963926,
      "tokens_seen": 2176385024
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015770522035180634,
      "loss": 2.6185,
      "theoretical_loss": 3.406495191700924,
      "tokens_seen": 2176450560
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015768867032343485,
      "loss": 2.3434,
      "theoretical_loss": 3.4064867809296344,
      "tokens_seen": 2176516096
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015767212029506338,
      "loss": 2.7007,
      "theoretical_loss": 3.4064783704825023,
      "tokens_seen": 2176581632
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015765557026669187,
      "loss": 2.4847,
      "theoretical_loss": 3.406469960359505,
      "tokens_seen": 2176647168
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001576390202383204,
      "loss": 2.3015,
      "theoretical_loss": 3.4064615505606195,
      "tokens_seen": 2176712704
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015762247020994892,
      "loss": 2.4555,
      "theoretical_loss": 3.4064531410858248,
      "tokens_seen": 2176778240
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015760592018157745,
      "loss": 2.5882,
      "theoretical_loss": 3.4064447319350983,
      "tokens_seen": 2176843776
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015758937015320596,
      "loss": 2.3555,
      "theoretical_loss": 3.4064363231084176,
      "tokens_seen": 2176909312
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001575728201248345,
      "loss": 2.2096,
      "theoretical_loss": 3.4064279146057603,
      "tokens_seen": 2176974848
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000157556270096463,
      "loss": 2.44,
      "theoretical_loss": 3.4064195064271043,
      "tokens_seen": 2177040384
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015753972006809155,
      "loss": 2.5369,
      "theoretical_loss": 3.406411098572428,
      "tokens_seen": 2177105920
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015752317003972006,
      "loss": 2.6583,
      "theoretical_loss": 3.4064026910417073,
      "tokens_seen": 2177171456
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001575066200113486,
      "loss": 2.5804,
      "theoretical_loss": 3.4063942838349224,
      "tokens_seen": 2177236992
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001574900699829771,
      "loss": 2.6326,
      "theoretical_loss": 3.4063858769520494,
      "tokens_seen": 2177302528
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015747351995460564,
      "loss": 2.4022,
      "theoretical_loss": 3.4063774703930667,
      "tokens_seen": 2177368064
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015745696992623416,
      "loss": 2.0835,
      "theoretical_loss": 3.4063690641579516,
      "tokens_seen": 2177433600
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001574404198978627,
      "loss": 2.1405,
      "theoretical_loss": 3.4063606582466823,
      "tokens_seen": 2177499136
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001574238698694912,
      "loss": 2.5285,
      "theoretical_loss": 3.4063522526592367,
      "tokens_seen": 2177564672
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015740731984111971,
      "loss": 2.549,
      "theoretical_loss": 3.4063438473955925,
      "tokens_seen": 2177630208
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015739076981274825,
      "loss": 2.509,
      "theoretical_loss": 3.406335442455727,
      "tokens_seen": 2177695744
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015737421978437676,
      "loss": 2.6108,
      "theoretical_loss": 3.4063270378396187,
      "tokens_seen": 2177761280
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001573576697560053,
      "loss": 2.6519,
      "theoretical_loss": 3.4063186335472446,
      "tokens_seen": 2177826816
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001573411197276338,
      "loss": 2.6795,
      "theoretical_loss": 3.406310229578583,
      "tokens_seen": 2177892352
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015732456969926235,
      "loss": 2.5477,
      "theoretical_loss": 3.4063018259336117,
      "tokens_seen": 2177957888
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015730801967089086,
      "loss": 2.6584,
      "theoretical_loss": 3.4062934226123085,
      "tokens_seen": 2178023424
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001572914696425194,
      "loss": 2.7568,
      "theoretical_loss": 3.406285019614651,
      "tokens_seen": 2178088960
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001572749196141479,
      "loss": 2.5624,
      "theoretical_loss": 3.406276616940617,
      "tokens_seen": 2178154496
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015725836958577644,
      "loss": 2.076,
      "theoretical_loss": 3.4062682145901846,
      "tokens_seen": 2178220032
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015724181955740495,
      "loss": 2.6357,
      "theoretical_loss": 3.406259812563331,
      "tokens_seen": 2178285568
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001572252695290335,
      "loss": 2.6121,
      "theoretical_loss": 3.4062514108600346,
      "tokens_seen": 2178351104
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000157208719500662,
      "loss": 2.6766,
      "theoretical_loss": 3.406243009480273,
      "tokens_seen": 2178416640
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015719216947229054,
      "loss": 2.3022,
      "theoretical_loss": 3.406234608424024,
      "tokens_seen": 2178482176
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015717561944391902,
      "loss": 2.6081,
      "theoretical_loss": 3.4062262076912653,
      "tokens_seen": 2178547712
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015715906941554756,
      "loss": 2.4221,
      "theoretical_loss": 3.4062178072819753,
      "tokens_seen": 2178613248
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015714251938717607,
      "loss": 2.5291,
      "theoretical_loss": 3.406209407196131,
      "tokens_seen": 2178678784
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001571259693588046,
      "loss": 2.5621,
      "theoretical_loss": 3.406201007433711,
      "tokens_seen": 2178744320
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015710941933043315,
      "loss": 2.6911,
      "theoretical_loss": 3.406192607994692,
      "tokens_seen": 2178809856
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015709286930206166,
      "loss": 2.4201,
      "theoretical_loss": 3.4061842088790524,
      "tokens_seen": 2178875392
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001570763192736902,
      "loss": 2.2598,
      "theoretical_loss": 3.4061758100867707,
      "tokens_seen": 2178940928
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001570597692453187,
      "loss": 2.7219,
      "theoretical_loss": 3.4061674116178238,
      "tokens_seen": 2179006464
    },
    {
      "epoch": 0.78,
      "objective/train/avg_token_score": 0.014639534056186676,
      "objective/train/avg_weight": 0.9827972650527954,
      "objective/train/docs_used": 1233537,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.7808754444122314,
      "objective/train/std_weight": 0.09611690044403076,
      "objective/train/theoretical_loss": 3.40615901347219,
      "objective/train/tokens_used": 2199532000,
      "theoretical_loss": 3.40615901347219,
      "tokens_seen": 2179072000
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015704321921694724,
      "loss": 2.643,
      "theoretical_loss": 3.40615901347219,
      "tokens_seen": 2179072000
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015702666918857575,
      "loss": 2.5757,
      "theoretical_loss": 3.4061506156498473,
      "tokens_seen": 2179137536
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001570101191602043,
      "loss": 2.4847,
      "theoretical_loss": 3.406142218150773,
      "tokens_seen": 2179203072
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001569935691318328,
      "loss": 2.5919,
      "theoretical_loss": 3.406133820974945,
      "tokens_seen": 2179268608
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015697701910346134,
      "loss": 2.5948,
      "theoretical_loss": 3.4061254241223415,
      "tokens_seen": 2179334144
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015696046907508985,
      "loss": 2.3649,
      "theoretical_loss": 3.4061170275929404,
      "tokens_seen": 2179399680
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015694391904671839,
      "loss": 2.6943,
      "theoretical_loss": 3.406108631386719,
      "tokens_seen": 2179465216
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015692736901834687,
      "loss": 2.5252,
      "theoretical_loss": 3.4061002355036556,
      "tokens_seen": 2179530752
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001569108189899754,
      "loss": 2.5497,
      "theoretical_loss": 3.406091839943728,
      "tokens_seen": 2179596288
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015689426896160392,
      "loss": 2.6195,
      "theoretical_loss": 3.4060834447069133,
      "tokens_seen": 2179661824
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015687771893323245,
      "loss": 2.5151,
      "theoretical_loss": 3.4060750497931904,
      "tokens_seen": 2179727360
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015686116890486096,
      "loss": 2.6322,
      "theoretical_loss": 3.4060666552025367,
      "tokens_seen": 2179792896
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001568446188764895,
      "loss": 2.6811,
      "theoretical_loss": 3.4060582609349304,
      "tokens_seen": 2179858432
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000156828068848118,
      "loss": 2.7178,
      "theoretical_loss": 3.4060498669903483,
      "tokens_seen": 2179923968
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015681151881974655,
      "loss": 2.6255,
      "theoretical_loss": 3.4060414733687696,
      "tokens_seen": 2179989504
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015679496879137506,
      "loss": 2.653,
      "theoretical_loss": 3.4060330800701717,
      "tokens_seen": 2180055040
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001567784187630036,
      "loss": 2.4085,
      "theoretical_loss": 3.406024687094532,
      "tokens_seen": 2180120576
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001567618687346321,
      "loss": 2.6177,
      "theoretical_loss": 3.4060162944418284,
      "tokens_seen": 2180186112
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015674531870626065,
      "loss": 2.2062,
      "theoretical_loss": 3.40600790211204,
      "tokens_seen": 2180251648
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015672876867788916,
      "loss": 2.6635,
      "theoretical_loss": 3.4059995101051426,
      "tokens_seen": 2180317184
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001567122186495177,
      "loss": 2.6869,
      "theoretical_loss": 3.405991118421116,
      "tokens_seen": 2180382720
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001566956686211462,
      "loss": 2.5354,
      "theoretical_loss": 3.405982727059937,
      "tokens_seen": 2180448256
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015667911859277472,
      "loss": 2.3334,
      "theoretical_loss": 3.405974336021584,
      "tokens_seen": 2180513792
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015666256856440325,
      "loss": 2.8925,
      "theoretical_loss": 3.4059659453060345,
      "tokens_seen": 2180579328
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015664601853603176,
      "loss": 2.5776,
      "theoretical_loss": 3.4059575549132663,
      "tokens_seen": 2180644864
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001566294685076603,
      "loss": 2.6065,
      "theoretical_loss": 3.4059491648432574,
      "tokens_seen": 2180710400
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001566129184792888,
      "loss": 2.5615,
      "theoretical_loss": 3.4059407750959863,
      "tokens_seen": 2180775936
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015659636845091735,
      "loss": 2.568,
      "theoretical_loss": 3.40593238567143,
      "tokens_seen": 2180841472
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015657981842254586,
      "loss": 2.8848,
      "theoretical_loss": 3.405923996569567,
      "tokens_seen": 2180907008
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001565632683941744,
      "loss": 2.5224,
      "theoretical_loss": 3.4059156077903747,
      "tokens_seen": 2180972544
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001565467183658029,
      "loss": 2.3826,
      "theoretical_loss": 3.405907219333831,
      "tokens_seen": 2181038080
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015653016833743144,
      "loss": 2.6088,
      "theoretical_loss": 3.4058988311999148,
      "tokens_seen": 2181103616
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015651361830905996,
      "loss": 2.3075,
      "theoretical_loss": 3.4058904433886026,
      "tokens_seen": 2181169152
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001564970682806885,
      "loss": 2.4549,
      "theoretical_loss": 3.4058820558998733,
      "tokens_seen": 2181234688
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000156480518252317,
      "loss": 2.5608,
      "theoretical_loss": 3.4058736687337046,
      "tokens_seen": 2181300224
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015646396822394554,
      "loss": 2.6521,
      "theoretical_loss": 3.4058652818900734,
      "tokens_seen": 2181365760
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015644741819557405,
      "loss": 2.3578,
      "theoretical_loss": 3.4058568953689594,
      "tokens_seen": 2181431296
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015643086816720256,
      "loss": 2.6001,
      "theoretical_loss": 3.405848509170339,
      "tokens_seen": 2181496832
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001564143181388311,
      "loss": 2.8068,
      "theoretical_loss": 3.405840123294191,
      "tokens_seen": 2181562368
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001563977681104596,
      "loss": 2.6515,
      "theoretical_loss": 3.4058317377404927,
      "tokens_seen": 2181627904
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015638121808208815,
      "loss": 2.3735,
      "theoretical_loss": 3.4058233525092225,
      "tokens_seen": 2181693440
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015636466805371666,
      "loss": 2.0255,
      "theoretical_loss": 3.405814967600358,
      "tokens_seen": 2181758976
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001563481180253452,
      "loss": 2.5989,
      "theoretical_loss": 3.4058065830138773,
      "tokens_seen": 2181824512
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001563315679969737,
      "loss": 2.5748,
      "theoretical_loss": 3.405798198749758,
      "tokens_seen": 2181890048
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015631501796860224,
      "loss": 2.6528,
      "theoretical_loss": 3.405789814807979,
      "tokens_seen": 2181955584
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015629846794023075,
      "loss": 2.3265,
      "theoretical_loss": 3.4057814311885166,
      "tokens_seen": 2182021120
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001562819179118593,
      "loss": 2.727,
      "theoretical_loss": 3.4057730478913504,
      "tokens_seen": 2182086656
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001562653678834878,
      "loss": 2.6595,
      "theoretical_loss": 3.405764664916457,
      "tokens_seen": 2182152192
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015624881785511634,
      "loss": 2.5946,
      "theoretical_loss": 3.405756282263815,
      "tokens_seen": 2182217728
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015623226782674485,
      "loss": 2.5104,
      "theoretical_loss": 3.4057478999334023,
      "tokens_seen": 2182283264
    },
    {
      "epoch": 0.78,
      "objective/train/avg_token_score": 0.009232776239514351,
      "objective/train/avg_weight": 0.9891341924667358,
      "objective/train/docs_used": 1235503,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.843648910522461,
      "objective/train/std_weight": 0.06339240074157715,
      "objective/train/theoretical_loss": 3.405739517925197,
      "objective/train/tokens_used": 2202808800,
      "theoretical_loss": 3.405739517925197,
      "tokens_seen": 2182348800
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001562157177983734,
      "loss": 2.8125,
      "theoretical_loss": 3.405739517925197,
      "tokens_seen": 2182348800
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001561991677700019,
      "loss": 2.5087,
      "theoretical_loss": 3.4057311362391767,
      "tokens_seen": 2182414336
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001561826177416304,
      "loss": 2.4795,
      "theoretical_loss": 3.4057227548753195,
      "tokens_seen": 2182479872
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015616606771325892,
      "loss": 2.6159,
      "theoretical_loss": 3.405714373833603,
      "tokens_seen": 2182545408
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015614951768488746,
      "loss": 2.7856,
      "theoretical_loss": 3.405705993114006,
      "tokens_seen": 2182610944
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015613296765651597,
      "loss": 2.3234,
      "theoretical_loss": 3.4056976127165055,
      "tokens_seen": 2182676480
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001561164176281445,
      "loss": 2.2808,
      "theoretical_loss": 3.40568923264108,
      "tokens_seen": 2182742016
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015609986759977301,
      "loss": 2.5769,
      "theoretical_loss": 3.405680852887707,
      "tokens_seen": 2182807552
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015608331757140155,
      "loss": 2.4611,
      "theoretical_loss": 3.405672473456365,
      "tokens_seen": 2182873088
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015606676754303006,
      "loss": 2.5828,
      "theoretical_loss": 3.4056640943470318,
      "tokens_seen": 2182938624
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001560502175146586,
      "loss": 2.6087,
      "theoretical_loss": 3.405655715559685,
      "tokens_seen": 2183004160
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001560336674862871,
      "loss": 2.4521,
      "theoretical_loss": 3.405647337094303,
      "tokens_seen": 2183069696
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015601711745791565,
      "loss": 2.6451,
      "theoretical_loss": 3.4056389589508633,
      "tokens_seen": 2183135232
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015600056742954416,
      "loss": 2.5588,
      "theoretical_loss": 3.4056305811293446,
      "tokens_seen": 2183200768
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001559840174011727,
      "loss": 2.5998,
      "theoretical_loss": 3.4056222036297243,
      "tokens_seen": 2183266304
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001559674673728012,
      "loss": 2.4809,
      "theoretical_loss": 3.4056138264519804,
      "tokens_seen": 2183331840
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015595091734442972,
      "loss": 2.565,
      "theoretical_loss": 3.405605449596091,
      "tokens_seen": 2183397376
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015593436731605825,
      "loss": 2.6037,
      "theoretical_loss": 3.405597073062034,
      "tokens_seen": 2183462912
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015591781728768676,
      "loss": 2.6202,
      "theoretical_loss": 3.4055886968497875,
      "tokens_seen": 2183528448
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001559012672593153,
      "loss": 2.6666,
      "theoretical_loss": 3.4055803209593294,
      "tokens_seen": 2183593984
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001558847172309438,
      "loss": 2.3821,
      "theoretical_loss": 3.4055719453906375,
      "tokens_seen": 2183659520
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015586816720257235,
      "loss": 2.5134,
      "theoretical_loss": 3.40556357014369,
      "tokens_seen": 2183725056
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015585161717420086,
      "loss": 2.5123,
      "theoretical_loss": 3.405555195218465,
      "tokens_seen": 2183790592
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001558350671458294,
      "loss": 2.7021,
      "theoretical_loss": 3.40554682061494,
      "tokens_seen": 2183856128
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001558185171174579,
      "loss": 2.4945,
      "theoretical_loss": 3.405538446333094,
      "tokens_seen": 2183921664
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015580196708908645,
      "loss": 2.558,
      "theoretical_loss": 3.405530072372904,
      "tokens_seen": 2183987200
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015578541706071496,
      "loss": 2.7351,
      "theoretical_loss": 3.405521698734348,
      "tokens_seen": 2184052736
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001557688670323435,
      "loss": 2.7521,
      "theoretical_loss": 3.405513325417404,
      "tokens_seen": 2184118272
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000155752317003972,
      "loss": 2.4264,
      "theoretical_loss": 3.405504952422051,
      "tokens_seen": 2184183808
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015573576697560054,
      "loss": 2.2815,
      "theoretical_loss": 3.405496579748266,
      "tokens_seen": 2184249344
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015571921694722905,
      "loss": 2.6657,
      "theoretical_loss": 3.4054882073960275,
      "tokens_seen": 2184314880
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015570266691885756,
      "loss": 2.4282,
      "theoretical_loss": 3.405479835365313,
      "tokens_seen": 2184380416
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001556861168904861,
      "loss": 2.5272,
      "theoretical_loss": 3.4054714636561014,
      "tokens_seen": 2184445952
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001556695668621146,
      "loss": 2.2867,
      "theoretical_loss": 3.4054630922683695,
      "tokens_seen": 2184511488
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015565301683374315,
      "loss": 2.7878,
      "theoretical_loss": 3.405454721202096,
      "tokens_seen": 2184577024
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015563646680537166,
      "loss": 2.2643,
      "theoretical_loss": 3.405446350457259,
      "tokens_seen": 2184642560
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001556199167770002,
      "loss": 2.728,
      "theoretical_loss": 3.4054379800338364,
      "tokens_seen": 2184708096
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001556033667486287,
      "loss": 2.4579,
      "theoretical_loss": 3.405429609931806,
      "tokens_seen": 2184773632
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015558681672025724,
      "loss": 2.6677,
      "theoretical_loss": 3.4054212401511466,
      "tokens_seen": 2184839168
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015557026669188575,
      "loss": 2.5924,
      "theoretical_loss": 3.4054128706918347,
      "tokens_seen": 2184904704
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001555537166635143,
      "loss": 2.7066,
      "theoretical_loss": 3.40540450155385,
      "tokens_seen": 2184970240
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001555371666351428,
      "loss": 2.9049,
      "theoretical_loss": 3.4053961327371693,
      "tokens_seen": 2185035776
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015552061660677134,
      "loss": 2.476,
      "theoretical_loss": 3.4053877642417714,
      "tokens_seen": 2185101312
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015550406657839985,
      "loss": 2.4794,
      "theoretical_loss": 3.4053793960676337,
      "tokens_seen": 2185166848
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001554875165500284,
      "loss": 2.3938,
      "theoretical_loss": 3.405371028214735,
      "tokens_seen": 2185232384
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001554709665216569,
      "loss": 2.5993,
      "theoretical_loss": 3.4053626606830525,
      "tokens_seen": 2185297920
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001554544164932854,
      "loss": 2.468,
      "theoretical_loss": 3.405354293472565,
      "tokens_seen": 2185363456
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015543786646491392,
      "loss": 2.4075,
      "theoretical_loss": 3.4053459265832497,
      "tokens_seen": 2185428992
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015542131643654246,
      "loss": 2.6388,
      "theoretical_loss": 3.4053375600150853,
      "tokens_seen": 2185494528
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015540476640817097,
      "loss": 2.5235,
      "theoretical_loss": 3.4053291937680497,
      "tokens_seen": 2185560064
    },
    {
      "epoch": 0.78,
      "objective/train/avg_token_score": 0.010392747819423676,
      "objective/train/avg_weight": 0.9877673387527466,
      "objective/train/docs_used": 1237335,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.773211717605591,
      "objective/train/std_weight": 0.0773540511727333,
      "objective/train/theoretical_loss": 3.4053208278421208,
      "objective/train/tokens_used": 2206085600,
      "theoretical_loss": 3.4053208278421208,
      "tokens_seen": 2185625600
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001553882163797995,
      "loss": 2.6224,
      "theoretical_loss": 3.4053208278421208,
      "tokens_seen": 2185625600
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015537166635142802,
      "loss": 2.4999,
      "theoretical_loss": 3.405312462237277,
      "tokens_seen": 2185691136
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015535511632305655,
      "loss": 2.4542,
      "theoretical_loss": 3.405304096953496,
      "tokens_seen": 2185756672
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015533856629468506,
      "loss": 2.7221,
      "theoretical_loss": 3.4052957319907557,
      "tokens_seen": 2185822208
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001553220162663136,
      "loss": 2.5158,
      "theoretical_loss": 3.4052873673490347,
      "tokens_seen": 2185887744
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001553054662379421,
      "loss": 2.4958,
      "theoretical_loss": 3.4052790030283107,
      "tokens_seen": 2185953280
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015528891620957065,
      "loss": 2.7486,
      "theoretical_loss": 3.405270639028562,
      "tokens_seen": 2186018816
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015527236618119916,
      "loss": 2.7039,
      "theoretical_loss": 3.405262275349766,
      "tokens_seen": 2186084352
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001552558161528277,
      "loss": 2.5174,
      "theoretical_loss": 3.4052539119919016,
      "tokens_seen": 2186149888
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001552392661244562,
      "loss": 2.5444,
      "theoretical_loss": 3.4052455489549462,
      "tokens_seen": 2186215424
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015522271609608475,
      "loss": 2.1297,
      "theoretical_loss": 3.4052371862388786,
      "tokens_seen": 2186280960
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015520616606771326,
      "loss": 2.5446,
      "theoretical_loss": 3.405228823843676,
      "tokens_seen": 2186346496
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015518961603934177,
      "loss": 2.5825,
      "theoretical_loss": 3.4052204617693174,
      "tokens_seen": 2186412032
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001551730660109703,
      "loss": 2.5889,
      "theoretical_loss": 3.4052121000157802,
      "tokens_seen": 2186477568
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015515651598259881,
      "loss": 2.2715,
      "theoretical_loss": 3.4052037385830425,
      "tokens_seen": 2186543104
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015513996595422735,
      "loss": 2.6052,
      "theoretical_loss": 3.405195377471083,
      "tokens_seen": 2186608640
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015512341592585586,
      "loss": 2.7109,
      "theoretical_loss": 3.4051870166798786,
      "tokens_seen": 2186674176
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001551068658974844,
      "loss": 2.4989,
      "theoretical_loss": 3.4051786562094084,
      "tokens_seen": 2186739712
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001550903158691129,
      "loss": 2.4846,
      "theoretical_loss": 3.4051702960596506,
      "tokens_seen": 2186805248
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015507376584074145,
      "loss": 2.6924,
      "theoretical_loss": 3.4051619362305825,
      "tokens_seen": 2186870784
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015505721581236996,
      "loss": 2.635,
      "theoretical_loss": 3.4051535767221823,
      "tokens_seen": 2186936320
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001550406657839985,
      "loss": 2.4113,
      "theoretical_loss": 3.405145217534429,
      "tokens_seen": 2187001856
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000155024115755627,
      "loss": 2.4104,
      "theoretical_loss": 3.4051368586672996,
      "tokens_seen": 2187067392
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015500756572725554,
      "loss": 2.6882,
      "theoretical_loss": 3.4051285001207727,
      "tokens_seen": 2187132928
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015499101569888405,
      "loss": 2.2757,
      "theoretical_loss": 3.4051201418948263,
      "tokens_seen": 2187198464
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001549744656705126,
      "loss": 2.6293,
      "theoretical_loss": 3.4051117839894385,
      "tokens_seen": 2187264000
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001549579156421411,
      "loss": 2.5034,
      "theoretical_loss": 3.4051034264045876,
      "tokens_seen": 2187329536
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001549413656137696,
      "loss": 2.6783,
      "theoretical_loss": 3.4050950691402515,
      "tokens_seen": 2187395072
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015492481558539815,
      "loss": 2.5763,
      "theoretical_loss": 3.4050867121964083,
      "tokens_seen": 2187460608
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015490826555702666,
      "loss": 2.112,
      "theoretical_loss": 3.4050783555730364,
      "tokens_seen": 2187526144
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001548917155286552,
      "loss": 2.4417,
      "theoretical_loss": 3.4050699992701134,
      "tokens_seen": 2187591680
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001548751655002837,
      "loss": 2.2732,
      "theoretical_loss": 3.4050616432876177,
      "tokens_seen": 2187657216
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015485861547191225,
      "loss": 2.1426,
      "theoretical_loss": 3.4050532876255275,
      "tokens_seen": 2187722752
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015484206544354076,
      "loss": 2.6172,
      "theoretical_loss": 3.405044932283821,
      "tokens_seen": 2187788288
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001548255154151693,
      "loss": 2.7663,
      "theoretical_loss": 3.405036577262476,
      "tokens_seen": 2187853824
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001548089653867978,
      "loss": 2.6858,
      "theoretical_loss": 3.4050282225614703,
      "tokens_seen": 2187919360
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015479241535842634,
      "loss": 2.3282,
      "theoretical_loss": 3.405019868180783,
      "tokens_seen": 2187984896
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015477586533005485,
      "loss": 2.5139,
      "theoretical_loss": 3.4050115141203916,
      "tokens_seen": 2188050432
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001547593153016834,
      "loss": 2.2638,
      "theoretical_loss": 3.405003160380274,
      "tokens_seen": 2188115968
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001547427652733119,
      "loss": 2.4205,
      "theoretical_loss": 3.4049948069604095,
      "tokens_seen": 2188181504
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015472621524494044,
      "loss": 2.7071,
      "theoretical_loss": 3.404986453860775,
      "tokens_seen": 2188247040
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015470966521656892,
      "loss": 2.7293,
      "theoretical_loss": 3.4049781010813485,
      "tokens_seen": 2188312576
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015469311518819746,
      "loss": 2.359,
      "theoretical_loss": 3.404969748622109,
      "tokens_seen": 2188378112
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015467656515982597,
      "loss": 2.2352,
      "theoretical_loss": 3.404961396483034,
      "tokens_seen": 2188443648
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001546600151314545,
      "loss": 2.1089,
      "theoretical_loss": 3.4049530446641025,
      "tokens_seen": 2188509184
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015464346510308302,
      "loss": 2.4947,
      "theoretical_loss": 3.404944693165292,
      "tokens_seen": 2188574720
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015462691507471155,
      "loss": 2.3981,
      "theoretical_loss": 3.4049363419865806,
      "tokens_seen": 2188640256
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015461036504634007,
      "loss": 2.0044,
      "theoretical_loss": 3.404927991127946,
      "tokens_seen": 2188705792
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001545938150179686,
      "loss": 2.6345,
      "theoretical_loss": 3.404919640589368,
      "tokens_seen": 2188771328
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001545772649895971,
      "loss": 2.6403,
      "theoretical_loss": 3.4049112903708227,
      "tokens_seen": 2188836864
    },
    {
      "epoch": 0.78,
      "objective/train/avg_token_score": 0.023774731904268265,
      "objective/train/avg_weight": 0.9720308184623718,
      "objective/train/docs_used": 1239239,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 6.240751266479492,
      "objective/train/std_weight": 0.1399824321269989,
      "objective/train/theoretical_loss": 3.4049029404722897,
      "objective/train/tokens_used": 2209362400,
      "theoretical_loss": 3.4049029404722897,
      "tokens_seen": 2188902400
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015456071496122565,
      "loss": 2.4654,
      "theoretical_loss": 3.4049029404722897,
      "tokens_seen": 2188902400
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015454416493285416,
      "loss": 2.5645,
      "theoretical_loss": 3.404894590893747,
      "tokens_seen": 2188967936
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001545276149044827,
      "loss": 2.5851,
      "theoretical_loss": 3.4048862416351717,
      "tokens_seen": 2189033472
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001545110648761112,
      "loss": 2.8441,
      "theoretical_loss": 3.4048778926965433,
      "tokens_seen": 2189099008
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015449451484773975,
      "loss": 2.3769,
      "theoretical_loss": 3.404869544077839,
      "tokens_seen": 2189164544
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015447796481936826,
      "loss": 2.4414,
      "theoretical_loss": 3.404861195779038,
      "tokens_seen": 2189230080
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015446141479099677,
      "loss": 2.5135,
      "theoretical_loss": 3.4048528478001168,
      "tokens_seen": 2189295616
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001544448647626253,
      "loss": 2.3062,
      "theoretical_loss": 3.404844500141055,
      "tokens_seen": 2189361152
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015442831473425382,
      "loss": 2.4053,
      "theoretical_loss": 3.4048361528018303,
      "tokens_seen": 2189426688
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015441176470588235,
      "loss": 2.019,
      "theoretical_loss": 3.404827805782421,
      "tokens_seen": 2189492224
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015439521467751086,
      "loss": 2.5766,
      "theoretical_loss": 3.404819459082805,
      "tokens_seen": 2189557760
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001543786646491394,
      "loss": 2.3635,
      "theoretical_loss": 3.4048111127029603,
      "tokens_seen": 2189623296
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001543621146207679,
      "loss": 2.906,
      "theoretical_loss": 3.404802766642866,
      "tokens_seen": 2189688832
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015434556459239645,
      "loss": 2.4787,
      "theoretical_loss": 3.4047944209025,
      "tokens_seen": 2189754368
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015432901456402496,
      "loss": 2.5205,
      "theoretical_loss": 3.4047860754818395,
      "tokens_seen": 2189819904
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001543124645356535,
      "loss": 2.3944,
      "theoretical_loss": 3.4047777303808635,
      "tokens_seen": 2189885440
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000154295914507282,
      "loss": 2.2009,
      "theoretical_loss": 3.4047693855995504,
      "tokens_seen": 2189950976
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015427936447891054,
      "loss": 2.7252,
      "theoretical_loss": 3.4047610411378777,
      "tokens_seen": 2190016512
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015426281445053906,
      "loss": 2.5798,
      "theoretical_loss": 3.404752696995824,
      "tokens_seen": 2190082048
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001542462644221676,
      "loss": 2.4021,
      "theoretical_loss": 3.404744353173368,
      "tokens_seen": 2190147584
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001542297143937961,
      "loss": 2.6821,
      "theoretical_loss": 3.4047360096704864,
      "tokens_seen": 2190213120
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015421316436542461,
      "loss": 2.5536,
      "theoretical_loss": 3.404727666487159,
      "tokens_seen": 2190278656
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015419661433705315,
      "loss": 2.6186,
      "theoretical_loss": 3.4047193236233633,
      "tokens_seen": 2190344192
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015418006430868166,
      "loss": 2.6471,
      "theoretical_loss": 3.404710981079077,
      "tokens_seen": 2190409728
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001541635142803102,
      "loss": 2.467,
      "theoretical_loss": 3.4047026388542796,
      "tokens_seen": 2190475264
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001541469642519387,
      "loss": 2.3981,
      "theoretical_loss": 3.404694296948948,
      "tokens_seen": 2190540800
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015413041422356725,
      "loss": 2.5945,
      "theoretical_loss": 3.4046859553630613,
      "tokens_seen": 2190606336
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015411386419519576,
      "loss": 2.6994,
      "theoretical_loss": 3.4046776140965975,
      "tokens_seen": 2190671872
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001540973141668243,
      "loss": 2.4536,
      "theoretical_loss": 3.4046692731495343,
      "tokens_seen": 2190737408
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001540807641384528,
      "loss": 3.1563,
      "theoretical_loss": 3.40466093252185,
      "tokens_seen": 2190802944
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015406421411008134,
      "loss": 2.3367,
      "theoretical_loss": 3.404652592213524,
      "tokens_seen": 2190868480
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015404766408170985,
      "loss": 2.3287,
      "theoretical_loss": 3.404644252224533,
      "tokens_seen": 2190934016
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001540311140533384,
      "loss": 2.6621,
      "theoretical_loss": 3.404635912554856,
      "tokens_seen": 2190999552
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001540145640249669,
      "loss": 2.4762,
      "theoretical_loss": 3.404627573204471,
      "tokens_seen": 2191065088
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015399801399659544,
      "loss": 2.609,
      "theoretical_loss": 3.4046192341733565,
      "tokens_seen": 2191130624
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015398146396822392,
      "loss": 2.2929,
      "theoretical_loss": 3.4046108954614906,
      "tokens_seen": 2191196160
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015396491393985246,
      "loss": 2.6631,
      "theoretical_loss": 3.404602557068851,
      "tokens_seen": 2191261696
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015394836391148097,
      "loss": 2.6464,
      "theoretical_loss": 3.404594218995417,
      "tokens_seen": 2191327232
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001539318138831095,
      "loss": 2.3579,
      "theoretical_loss": 3.4045858812411662,
      "tokens_seen": 2191392768
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015391526385473802,
      "loss": 2.6326,
      "theoretical_loss": 3.4045775438060764,
      "tokens_seen": 2191458304
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015389871382636656,
      "loss": 2.4938,
      "theoretical_loss": 3.4045692066901263,
      "tokens_seen": 2191523840
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015388216379799507,
      "loss": 2.5356,
      "theoretical_loss": 3.4045608698932943,
      "tokens_seen": 2191589376
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001538656137696236,
      "loss": 2.6097,
      "theoretical_loss": 3.4045525334155586,
      "tokens_seen": 2191654912
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015384906374125211,
      "loss": 2.0697,
      "theoretical_loss": 3.4045441972568975,
      "tokens_seen": 2191720448
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015383251371288065,
      "loss": 2.3644,
      "theoretical_loss": 3.4045358614172887,
      "tokens_seen": 2191785984
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015381596368450916,
      "loss": 2.5456,
      "theoretical_loss": 3.404527525896711,
      "tokens_seen": 2191851520
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001537994136561377,
      "loss": 2.6029,
      "theoretical_loss": 3.404519190695143,
      "tokens_seen": 2191917056
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001537828636277662,
      "loss": 2.7046,
      "theoretical_loss": 3.4045108558125614,
      "tokens_seen": 2191982592
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015376631359939475,
      "loss": 2.3639,
      "theoretical_loss": 3.404502521248946,
      "tokens_seen": 2192048128
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015374976357102326,
      "loss": 2.5912,
      "theoretical_loss": 3.404494187004275,
      "tokens_seen": 2192113664
    },
    {
      "epoch": 0.78,
      "objective/train/avg_token_score": 0.01629609242081642,
      "objective/train/avg_weight": 0.9808170199394226,
      "objective/train/docs_used": 1240983,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.170804023742676,
      "objective/train/std_weight": 0.09136556833982468,
      "objective/train/theoretical_loss": 3.4044858530785254,
      "objective/train/tokens_used": 2212639200,
      "theoretical_loss": 3.4044858530785254,
      "tokens_seen": 2192179200
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015373321354265177,
      "loss": 2.5606,
      "theoretical_loss": 3.4044858530785254,
      "tokens_seen": 2192179200
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001537166635142803,
      "loss": 2.7368,
      "theoretical_loss": 3.4044775194716768,
      "tokens_seen": 2192244736
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015370011348590882,
      "loss": 2.7674,
      "theoretical_loss": 3.404469186183707,
      "tokens_seen": 2192310272
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015368356345753735,
      "loss": 2.6016,
      "theoretical_loss": 3.404460853214594,
      "tokens_seen": 2192375808
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015366701342916586,
      "loss": 2.6137,
      "theoretical_loss": 3.404452520564316,
      "tokens_seen": 2192441344
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001536504634007944,
      "loss": 2.7592,
      "theoretical_loss": 3.4044441882328518,
      "tokens_seen": 2192506880
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001536339133724229,
      "loss": 2.5484,
      "theoretical_loss": 3.4044358562201795,
      "tokens_seen": 2192572416
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015361736334405145,
      "loss": 2.349,
      "theoretical_loss": 3.404427524526277,
      "tokens_seen": 2192637952
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015360081331567996,
      "loss": 2.7861,
      "theoretical_loss": 3.4044191931511234,
      "tokens_seen": 2192703488
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001535842632873085,
      "loss": 2.5139,
      "theoretical_loss": 3.404410862094696,
      "tokens_seen": 2192769024
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000153567713258937,
      "loss": 2.5369,
      "theoretical_loss": 3.4044025313569737,
      "tokens_seen": 2192834560
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015355116323056555,
      "loss": 2.6146,
      "theoretical_loss": 3.4043942009379347,
      "tokens_seen": 2192900096
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015353461320219406,
      "loss": 2.2245,
      "theoretical_loss": 3.404385870837557,
      "tokens_seen": 2192965632
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001535180631738226,
      "loss": 2.5082,
      "theoretical_loss": 3.404377541055819,
      "tokens_seen": 2193031168
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001535015131454511,
      "loss": 2.5248,
      "theoretical_loss": 3.4043692115926993,
      "tokens_seen": 2193096704
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015348496311707962,
      "loss": 2.5624,
      "theoretical_loss": 3.4043608824481764,
      "tokens_seen": 2193162240
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015346841308870815,
      "loss": 2.6254,
      "theoretical_loss": 3.4043525536222274,
      "tokens_seen": 2193227776
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015345186306033666,
      "loss": 2.6396,
      "theoretical_loss": 3.4043442251148317,
      "tokens_seen": 2193293312
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001534353130319652,
      "loss": 2.6943,
      "theoretical_loss": 3.404335896925967,
      "tokens_seen": 2193358848
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001534187630035937,
      "loss": 2.6679,
      "theoretical_loss": 3.4043275690556123,
      "tokens_seen": 2193424384
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015340221297522225,
      "loss": 2.4394,
      "theoretical_loss": 3.404319241503745,
      "tokens_seen": 2193489920
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015338566294685076,
      "loss": 2.6927,
      "theoretical_loss": 3.404310914270344,
      "tokens_seen": 2193555456
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001533691129184793,
      "loss": 2.6667,
      "theoretical_loss": 3.404302587355388,
      "tokens_seen": 2193620992
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001533525628901078,
      "loss": 2.7913,
      "theoretical_loss": 3.404294260758854,
      "tokens_seen": 2193686528
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015333601286173634,
      "loss": 2.5011,
      "theoretical_loss": 3.4042859344807215,
      "tokens_seen": 2193752064
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015331946283336486,
      "loss": 2.7088,
      "theoretical_loss": 3.4042776085209683,
      "tokens_seen": 2193817600
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001533029128049934,
      "loss": 2.6298,
      "theoretical_loss": 3.404269282879573,
      "tokens_seen": 2193883136
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001532863627766219,
      "loss": 2.8329,
      "theoretical_loss": 3.4042609575565135,
      "tokens_seen": 2193948672
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015326981274825044,
      "loss": 2.5124,
      "theoretical_loss": 3.404252632551769,
      "tokens_seen": 2194014208
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015325326271987895,
      "loss": 2.8477,
      "theoretical_loss": 3.404244307865316,
      "tokens_seen": 2194079744
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015323671269150746,
      "loss": 2.6659,
      "theoretical_loss": 3.4042359834971347,
      "tokens_seen": 2194145280
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015322016266313597,
      "loss": 2.5494,
      "theoretical_loss": 3.4042276594472027,
      "tokens_seen": 2194210816
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001532036126347645,
      "loss": 2.4498,
      "theoretical_loss": 3.4042193357154984,
      "tokens_seen": 2194276352
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015318706260639302,
      "loss": 2.2488,
      "theoretical_loss": 3.404211012302,
      "tokens_seen": 2194341888
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015317051257802156,
      "loss": 2.4443,
      "theoretical_loss": 3.404202689206686,
      "tokens_seen": 2194407424
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015315396254965007,
      "loss": 2.4488,
      "theoretical_loss": 3.4041943664295347,
      "tokens_seen": 2194472960
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001531374125212786,
      "loss": 2.0194,
      "theoretical_loss": 3.4041860439705243,
      "tokens_seen": 2194538496
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015312086249290714,
      "loss": 2.4151,
      "theoretical_loss": 3.404177721829633,
      "tokens_seen": 2194604032
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015310431246453565,
      "loss": 2.2421,
      "theoretical_loss": 3.4041694000068397,
      "tokens_seen": 2194669568
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001530877624361642,
      "loss": 2.685,
      "theoretical_loss": 3.4041610785021224,
      "tokens_seen": 2194735104
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001530712124077927,
      "loss": 2.375,
      "theoretical_loss": 3.4041527573154595,
      "tokens_seen": 2194800640
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015305466237942124,
      "loss": 2.393,
      "theoretical_loss": 3.404144436446829,
      "tokens_seen": 2194866176
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015303811235104975,
      "loss": 2.5239,
      "theoretical_loss": 3.40413611589621,
      "tokens_seen": 2194931712
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001530215623226783,
      "loss": 2.5546,
      "theoretical_loss": 3.40412779566358,
      "tokens_seen": 2194997248
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001530050122943068,
      "loss": 2.7385,
      "theoretical_loss": 3.4041194757489177,
      "tokens_seen": 2195062784
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001529884622659353,
      "loss": 2.7769,
      "theoretical_loss": 3.4041111561522017,
      "tokens_seen": 2195128320
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015297191223756382,
      "loss": 2.2578,
      "theoretical_loss": 3.4041028368734105,
      "tokens_seen": 2195193856
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015295536220919236,
      "loss": 2.5211,
      "theoretical_loss": 3.4040945179125215,
      "tokens_seen": 2195259392
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015293881218082087,
      "loss": 2.5951,
      "theoretical_loss": 3.404086199269514,
      "tokens_seen": 2195324928
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001529222621524494,
      "loss": 2.7838,
      "theoretical_loss": 3.4040778809443664,
      "tokens_seen": 2195390464
    },
    {
      "epoch": 0.78,
      "objective/train/avg_token_score": 0.00980986189097166,
      "objective/train/avg_weight": 0.9884552359580994,
      "objective/train/docs_used": 1242635,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.755491256713867,
      "objective/train/std_weight": 0.059881843626499176,
      "objective/train/theoretical_loss": 3.4040695629370563,
      "objective/train/tokens_used": 2215916000,
      "theoretical_loss": 3.4040695629370563,
      "tokens_seen": 2195456000
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015290571212407791,
      "loss": 2.6569,
      "theoretical_loss": 3.4040695629370563,
      "tokens_seen": 2195456000
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015288916209570645,
      "loss": 2.6919,
      "theoretical_loss": 3.4040612452475627,
      "tokens_seen": 2195521536
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015287261206733496,
      "loss": 2.5686,
      "theoretical_loss": 3.4040529278758638,
      "tokens_seen": 2195587072
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001528560620389635,
      "loss": 2.6909,
      "theoretical_loss": 3.4040446108219378,
      "tokens_seen": 2195652608
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000152839512010592,
      "loss": 2.3625,
      "theoretical_loss": 3.4040362940857634,
      "tokens_seen": 2195718144
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015282296198222055,
      "loss": 2.5818,
      "theoretical_loss": 3.4040279776673183,
      "tokens_seen": 2195783680
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015280641195384906,
      "loss": 2.6242,
      "theoretical_loss": 3.404019661566582,
      "tokens_seen": 2195849216
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001527898619254776,
      "loss": 2.6994,
      "theoretical_loss": 3.4040113457835317,
      "tokens_seen": 2195914752
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001527733118971061,
      "loss": 2.6304,
      "theoretical_loss": 3.4040030303181465,
      "tokens_seen": 2195980288
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015275676186873464,
      "loss": 2.6112,
      "theoretical_loss": 3.403994715170405,
      "tokens_seen": 2196045824
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015274021184036315,
      "loss": 2.6973,
      "theoretical_loss": 3.403986400340285,
      "tokens_seen": 2196111360
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015272366181199166,
      "loss": 2.791,
      "theoretical_loss": 3.403978085827765,
      "tokens_seen": 2196176896
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001527071117836202,
      "loss": 2.4902,
      "theoretical_loss": 3.4039697716328234,
      "tokens_seen": 2196242432
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001526905617552487,
      "loss": 2.4023,
      "theoretical_loss": 3.4039614577554387,
      "tokens_seen": 2196307968
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015267401172687725,
      "loss": 2.8512,
      "theoretical_loss": 3.4039531441955897,
      "tokens_seen": 2196373504
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015265746169850576,
      "loss": 2.66,
      "theoretical_loss": 3.403944830953254,
      "tokens_seen": 2196439040
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001526409116701343,
      "loss": 2.3522,
      "theoretical_loss": 3.4039365180284102,
      "tokens_seen": 2196504576
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001526243616417628,
      "loss": 2.6235,
      "theoretical_loss": 3.403928205421037,
      "tokens_seen": 2196570112
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015260781161339135,
      "loss": 2.6054,
      "theoretical_loss": 3.403919893131113,
      "tokens_seen": 2196635648
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015259126158501986,
      "loss": 2.7072,
      "theoretical_loss": 3.4039115811586163,
      "tokens_seen": 2196701184
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001525747115566484,
      "loss": 2.5405,
      "theoretical_loss": 3.403903269503525,
      "tokens_seen": 2196766720
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001525581615282769,
      "loss": 2.3958,
      "theoretical_loss": 3.403894958165818,
      "tokens_seen": 2196832256
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015254161149990544,
      "loss": 2.3414,
      "theoretical_loss": 3.4038866471454736,
      "tokens_seen": 2196897792
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015252506147153395,
      "loss": 2.1569,
      "theoretical_loss": 3.4038783364424696,
      "tokens_seen": 2196963328
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015250851144316246,
      "loss": 2.4878,
      "theoretical_loss": 3.4038700260567856,
      "tokens_seen": 2197028864
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000152491961414791,
      "loss": 2.6877,
      "theoretical_loss": 3.4038617159883993,
      "tokens_seen": 2197094400
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001524754113864195,
      "loss": 2.7194,
      "theoretical_loss": 3.403853406237289,
      "tokens_seen": 2197159936
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015245886135804805,
      "loss": 2.5618,
      "theoretical_loss": 3.4038450968034333,
      "tokens_seen": 2197225472
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015244231132967656,
      "loss": 2.6501,
      "theoretical_loss": 3.4038367876868105,
      "tokens_seen": 2197291008
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001524257613013051,
      "loss": 2.5137,
      "theoretical_loss": 3.4038284788873994,
      "tokens_seen": 2197356544
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001524092112729336,
      "loss": 2.8259,
      "theoretical_loss": 3.403820170405178,
      "tokens_seen": 2197422080
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015239266124456214,
      "loss": 2.5146,
      "theoretical_loss": 3.403811862240125,
      "tokens_seen": 2197487616
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015237611121619065,
      "loss": 2.402,
      "theoretical_loss": 3.403803554392219,
      "tokens_seen": 2197553152
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001523595611878192,
      "loss": 2.4623,
      "theoretical_loss": 3.403795246861438,
      "tokens_seen": 2197618688
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001523430111594477,
      "loss": 2.7628,
      "theoretical_loss": 3.4037869396477607,
      "tokens_seen": 2197684224
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015232646113107624,
      "loss": 2.4427,
      "theoretical_loss": 3.4037786327511657,
      "tokens_seen": 2197749760
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00015230991110270475,
      "loss": 2.6407,
      "theoretical_loss": 3.4037703261716308,
      "tokens_seen": 2197815296
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001522933610743333,
      "loss": 2.601,
      "theoretical_loss": 3.403762019909135,
      "tokens_seen": 2197880832
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001522768110459618,
      "loss": 2.5304,
      "theoretical_loss": 3.4037537139636567,
      "tokens_seen": 2197946368
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001522602610175903,
      "loss": 2.6545,
      "theoretical_loss": 3.4037454083351744,
      "tokens_seen": 2198011904
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015224371098921882,
      "loss": 2.4546,
      "theoretical_loss": 3.403737103023666,
      "tokens_seen": 2198077440
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015222716096084736,
      "loss": 2.4507,
      "theoretical_loss": 3.403728798029111,
      "tokens_seen": 2198142976
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015221061093247587,
      "loss": 2.7429,
      "theoretical_loss": 3.4037204933514866,
      "tokens_seen": 2198208512
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001521940609041044,
      "loss": 2.5433,
      "theoretical_loss": 3.403712188990772,
      "tokens_seen": 2198274048
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015217751087573292,
      "loss": 2.732,
      "theoretical_loss": 3.403703884946946,
      "tokens_seen": 2198339584
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015216096084736145,
      "loss": 2.6019,
      "theoretical_loss": 3.4036955812199863,
      "tokens_seen": 2198405120
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015214441081898996,
      "loss": 2.5927,
      "theoretical_loss": 3.4036872778098717,
      "tokens_seen": 2198470656
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001521278607906185,
      "loss": 2.3891,
      "theoretical_loss": 3.4036789747165805,
      "tokens_seen": 2198536192
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000152111310762247,
      "loss": 2.5278,
      "theoretical_loss": 3.403670671940091,
      "tokens_seen": 2198601728
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015209476073387555,
      "loss": 2.5324,
      "theoretical_loss": 3.403662369480383,
      "tokens_seen": 2198667264
    },
    {
      "epoch": 0.79,
      "objective/train/avg_token_score": 0.017462341114878654,
      "objective/train/avg_weight": 0.9794541597366333,
      "objective/train/docs_used": 1244165,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5492377281188965,
      "objective/train/std_weight": 0.10757385939359665,
      "objective/train/theoretical_loss": 3.403654067337433,
      "objective/train/tokens_used": 2219192800,
      "theoretical_loss": 3.403654067337433,
      "tokens_seen": 2198732800
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015207821070550406,
      "loss": 2.3854,
      "theoretical_loss": 3.403654067337433,
      "tokens_seen": 2198732800
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001520616606771326,
      "loss": 2.3561,
      "theoretical_loss": 3.4036457655112207,
      "tokens_seen": 2198798336
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001520451106487611,
      "loss": 2.4554,
      "theoretical_loss": 3.403637464001724,
      "tokens_seen": 2198863872
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015202856062038965,
      "loss": 2.6104,
      "theoretical_loss": 3.4036291628089224,
      "tokens_seen": 2198929408
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015201201059201816,
      "loss": 2.5085,
      "theoretical_loss": 3.403620861932793,
      "tokens_seen": 2198994944
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015199546056364667,
      "loss": 2.6006,
      "theoretical_loss": 3.4036125613733152,
      "tokens_seen": 2199060480
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001519789105352752,
      "loss": 2.5868,
      "theoretical_loss": 3.403604261130467,
      "tokens_seen": 2199126016
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015196236050690371,
      "loss": 2.54,
      "theoretical_loss": 3.4035959612042275,
      "tokens_seen": 2199191552
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015194581047853225,
      "loss": 2.117,
      "theoretical_loss": 3.4035876615945746,
      "tokens_seen": 2199257088
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015192926045016076,
      "loss": 2.507,
      "theoretical_loss": 3.4035793623014867,
      "tokens_seen": 2199322624
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001519127104217893,
      "loss": 2.7851,
      "theoretical_loss": 3.403571063324943,
      "tokens_seen": 2199388160
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001518961603934178,
      "loss": 2.6195,
      "theoretical_loss": 3.403562764664921,
      "tokens_seen": 2199453696
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015187961036504635,
      "loss": 2.6194,
      "theoretical_loss": 3.4035544663214004,
      "tokens_seen": 2199519232
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015186306033667486,
      "loss": 2.6579,
      "theoretical_loss": 3.403546168294359,
      "tokens_seen": 2199584768
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001518465103083034,
      "loss": 2.504,
      "theoretical_loss": 3.403537870583775,
      "tokens_seen": 2199650304
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001518299602799319,
      "loss": 2.5976,
      "theoretical_loss": 3.4035295731896276,
      "tokens_seen": 2199715840
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015181341025156044,
      "loss": 2.3293,
      "theoretical_loss": 3.4035212761118947,
      "tokens_seen": 2199781376
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015179686022318895,
      "loss": 2.7046,
      "theoretical_loss": 3.4035129793505554,
      "tokens_seen": 2199846912
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001517803101948175,
      "loss": 2.5087,
      "theoretical_loss": 3.403504682905588,
      "tokens_seen": 2199912448
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000151763760166446,
      "loss": 2.5544,
      "theoretical_loss": 3.40349638677697,
      "tokens_seen": 2199977984
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001517472101380745,
      "loss": 2.696,
      "theoretical_loss": 3.403488090964682,
      "tokens_seen": 2200043520
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015173066010970305,
      "loss": 2.4666,
      "theoretical_loss": 3.4034797954687006,
      "tokens_seen": 2200109056
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015171411008133156,
      "loss": 2.3312,
      "theoretical_loss": 3.403471500289005,
      "tokens_seen": 2200174592
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001516975600529601,
      "loss": 2.5506,
      "theoretical_loss": 3.4034632054255742,
      "tokens_seen": 2200240128
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001516810100245886,
      "loss": 2.4905,
      "theoretical_loss": 3.403454910878386,
      "tokens_seen": 2200305664
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015166445999621715,
      "loss": 2.7348,
      "theoretical_loss": 3.4034466166474195,
      "tokens_seen": 2200371200
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015164790996784566,
      "loss": 2.7052,
      "theoretical_loss": 3.4034383227326526,
      "tokens_seen": 2200436736
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001516313599394742,
      "loss": 2.4694,
      "theoretical_loss": 3.4034300291340647,
      "tokens_seen": 2200502272
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001516148099111027,
      "loss": 2.5051,
      "theoretical_loss": 3.403421735851633,
      "tokens_seen": 2200567808
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015159825988273124,
      "loss": 2.6865,
      "theoretical_loss": 3.4034134428853378,
      "tokens_seen": 2200633344
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015158170985435975,
      "loss": 2.3935,
      "theoretical_loss": 3.403405150235156,
      "tokens_seen": 2200698880
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001515651598259883,
      "loss": 2.5103,
      "theoretical_loss": 3.403396857901067,
      "tokens_seen": 2200764416
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001515486097976168,
      "loss": 2.6077,
      "theoretical_loss": 3.4033885658830494,
      "tokens_seen": 2200829952
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015153205976924534,
      "loss": 2.452,
      "theoretical_loss": 3.403380274181081,
      "tokens_seen": 2200895488
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015151550974087382,
      "loss": 2.6891,
      "theoretical_loss": 3.403371982795141,
      "tokens_seen": 2200961024
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015149895971250236,
      "loss": 2.6364,
      "theoretical_loss": 3.403363691725208,
      "tokens_seen": 2201026560
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015148240968413087,
      "loss": 2.047,
      "theoretical_loss": 3.40335540097126,
      "tokens_seen": 2201092096
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001514658596557594,
      "loss": 2.3799,
      "theoretical_loss": 3.403347110533276,
      "tokens_seen": 2201157632
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015144930962738792,
      "loss": 2.6954,
      "theoretical_loss": 3.4033388204112347,
      "tokens_seen": 2201223168
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015143275959901645,
      "loss": 2.2309,
      "theoretical_loss": 3.403330530605114,
      "tokens_seen": 2201288704
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015141620957064497,
      "loss": 2.459,
      "theoretical_loss": 3.403322241114893,
      "tokens_seen": 2201354240
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001513996595422735,
      "loss": 2.6434,
      "theoretical_loss": 3.40331395194055,
      "tokens_seen": 2201419776
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000151383109513902,
      "loss": 2.6512,
      "theoretical_loss": 3.403305663082064,
      "tokens_seen": 2201485312
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015136655948553055,
      "loss": 2.3188,
      "theoretical_loss": 3.403297374539413,
      "tokens_seen": 2201550848
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015135000945715906,
      "loss": 2.7117,
      "theoretical_loss": 3.4032890863125758,
      "tokens_seen": 2201616384
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001513334594287876,
      "loss": 2.3132,
      "theoretical_loss": 3.403280798401531,
      "tokens_seen": 2201681920
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001513169094004161,
      "loss": 2.3213,
      "theoretical_loss": 3.403272510806257,
      "tokens_seen": 2201747456
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015130035937204465,
      "loss": 2.4557,
      "theoretical_loss": 3.4032642235267327,
      "tokens_seen": 2201812992
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015128380934367316,
      "loss": 2.6011,
      "theoretical_loss": 3.4032559365629362,
      "tokens_seen": 2201878528
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015126725931530167,
      "loss": 2.7352,
      "theoretical_loss": 3.4032476499148467,
      "tokens_seen": 2201944064
    },
    {
      "epoch": 0.79,
      "objective/train/avg_token_score": 0.02762080729007721,
      "objective/train/avg_weight": 0.967502772808075,
      "objective/train/docs_used": 1245946,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.947446346282959,
      "objective/train/std_weight": 0.14260317385196686,
      "objective/train/theoretical_loss": 3.4032393635824416,
      "objective/train/tokens_used": 2222469600,
      "theoretical_loss": 3.4032393635824416,
      "tokens_seen": 2202009600
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001512507092869302,
      "loss": 2.7472,
      "theoretical_loss": 3.4032393635824416,
      "tokens_seen": 2202009600
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015123415925855872,
      "loss": 2.6881,
      "theoretical_loss": 3.4032310775657013,
      "tokens_seen": 2202075136
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015121760923018725,
      "loss": 2.1961,
      "theoretical_loss": 3.4032227918646027,
      "tokens_seen": 2202140672
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015120105920181576,
      "loss": 2.7622,
      "theoretical_loss": 3.4032145064791255,
      "tokens_seen": 2202206208
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001511845091734443,
      "loss": 2.3974,
      "theoretical_loss": 3.4032062214092473,
      "tokens_seen": 2202271744
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001511679591450728,
      "loss": 2.5596,
      "theoretical_loss": 3.403197936654948,
      "tokens_seen": 2202337280
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015115140911670135,
      "loss": 2.2452,
      "theoretical_loss": 3.403189652216205,
      "tokens_seen": 2202402816
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015113485908832986,
      "loss": 2.5298,
      "theoretical_loss": 3.403181368092997,
      "tokens_seen": 2202468352
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001511183090599584,
      "loss": 2.4041,
      "theoretical_loss": 3.4031730842853034,
      "tokens_seen": 2202533888
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001511017590315869,
      "loss": 2.3758,
      "theoretical_loss": 3.403164800793102,
      "tokens_seen": 2202599424
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015108520900321544,
      "loss": 2.5166,
      "theoretical_loss": 3.4031565176163716,
      "tokens_seen": 2202664960
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015106865897484396,
      "loss": 2.4861,
      "theoretical_loss": 3.403148234755091,
      "tokens_seen": 2202730496
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001510521089464725,
      "loss": 2.8013,
      "theoretical_loss": 3.403139952209239,
      "tokens_seen": 2202796032
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000151035558918101,
      "loss": 2.5527,
      "theoretical_loss": 3.4031316699787935,
      "tokens_seen": 2202861568
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015101900888972951,
      "loss": 2.7198,
      "theoretical_loss": 3.403123388063734,
      "tokens_seen": 2202927104
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015100245886135805,
      "loss": 2.5395,
      "theoretical_loss": 3.403115106464038,
      "tokens_seen": 2202992640
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015098590883298656,
      "loss": 2.4508,
      "theoretical_loss": 3.403106825179685,
      "tokens_seen": 2203058176
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001509693588046151,
      "loss": 2.6177,
      "theoretical_loss": 3.403098544210654,
      "tokens_seen": 2203123712
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001509528087762436,
      "loss": 2.2101,
      "theoretical_loss": 3.403090263556922,
      "tokens_seen": 2203189248
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015093625874787215,
      "loss": 1.9725,
      "theoretical_loss": 3.403081983218469,
      "tokens_seen": 2203254784
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015091970871950066,
      "loss": 2.4085,
      "theoretical_loss": 3.403073703195273,
      "tokens_seen": 2203320320
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001509031586911292,
      "loss": 2.4472,
      "theoretical_loss": 3.403065423487313,
      "tokens_seen": 2203385856
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001508866086627577,
      "loss": 2.523,
      "theoretical_loss": 3.4030571440945674,
      "tokens_seen": 2203451392
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015087005863438624,
      "loss": 2.6503,
      "theoretical_loss": 3.4030488650170145,
      "tokens_seen": 2203516928
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015085350860601475,
      "loss": 2.1454,
      "theoretical_loss": 3.4030405862546336,
      "tokens_seen": 2203582464
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001508369585776433,
      "loss": 2.8266,
      "theoretical_loss": 3.403032307807403,
      "tokens_seen": 2203648000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001508204085492718,
      "loss": 2.5651,
      "theoretical_loss": 3.4030240296753016,
      "tokens_seen": 2203713536
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015080385852090034,
      "loss": 2.3423,
      "theoretical_loss": 3.4030157518583075,
      "tokens_seen": 2203779072
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015078730849252882,
      "loss": 2.6769,
      "theoretical_loss": 3.4030074743563996,
      "tokens_seen": 2203844608
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015077075846415736,
      "loss": 2.5111,
      "theoretical_loss": 3.4029991971695566,
      "tokens_seen": 2203910144
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015075420843578587,
      "loss": 2.5904,
      "theoretical_loss": 3.402990920297757,
      "tokens_seen": 2203975680
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001507376584074144,
      "loss": 2.5548,
      "theoretical_loss": 3.40298264374098,
      "tokens_seen": 2204041216
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015072110837904292,
      "loss": 2.7493,
      "theoretical_loss": 3.402974367499203,
      "tokens_seen": 2204106752
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015070455835067146,
      "loss": 2.3763,
      "theoretical_loss": 3.402966091572406,
      "tokens_seen": 2204172288
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015068800832229997,
      "loss": 2.4976,
      "theoretical_loss": 3.4029578159605665,
      "tokens_seen": 2204237824
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001506714582939285,
      "loss": 2.6119,
      "theoretical_loss": 3.4029495406636645,
      "tokens_seen": 2204303360
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015065490826555701,
      "loss": 2.8021,
      "theoretical_loss": 3.4029412656816773,
      "tokens_seen": 2204368896
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015063835823718555,
      "loss": 2.5325,
      "theoretical_loss": 3.4029329910145845,
      "tokens_seen": 2204434432
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015062180820881406,
      "loss": 2.7067,
      "theoretical_loss": 3.4029247166623637,
      "tokens_seen": 2204499968
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001506052581804426,
      "loss": 2.8192,
      "theoretical_loss": 3.402916442624995,
      "tokens_seen": 2204565504
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001505887081520711,
      "loss": 2.1679,
      "theoretical_loss": 3.402908168902456,
      "tokens_seen": 2204631040
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015057215812369965,
      "loss": 2.7015,
      "theoretical_loss": 3.4028998954947256,
      "tokens_seen": 2204696576
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015055560809532816,
      "loss": 2.581,
      "theoretical_loss": 3.4028916224017824,
      "tokens_seen": 2204762112
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015053905806695667,
      "loss": 2.6475,
      "theoretical_loss": 3.4028833496236057,
      "tokens_seen": 2204827648
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001505225080385852,
      "loss": 2.4883,
      "theoretical_loss": 3.4028750771601732,
      "tokens_seen": 2204893184
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015050595801021372,
      "loss": 2.6834,
      "theoretical_loss": 3.402866805011464,
      "tokens_seen": 2204958720
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015048940798184225,
      "loss": 2.6998,
      "theoretical_loss": 3.402858533177457,
      "tokens_seen": 2205024256
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015047285795347076,
      "loss": 1.9326,
      "theoretical_loss": 3.4028502616581306,
      "tokens_seen": 2205089792
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001504563079250993,
      "loss": 2.8612,
      "theoretical_loss": 3.4028419904534637,
      "tokens_seen": 2205155328
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001504397578967278,
      "loss": 2.7752,
      "theoretical_loss": 3.402833719563435,
      "tokens_seen": 2205220864
    },
    {
      "epoch": 0.79,
      "objective/train/avg_token_score": 0.013414936140179634,
      "objective/train/avg_weight": 0.9842434525489807,
      "objective/train/docs_used": 1246233,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8318305015563965,
      "objective/train/std_weight": 0.07480166107416153,
      "objective/train/theoretical_loss": 3.4028254489880227,
      "objective/train/tokens_used": 2225746400,
      "theoretical_loss": 3.4028254489880227,
      "tokens_seen": 2205286400
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015042320786835635,
      "loss": 2.7898,
      "theoretical_loss": 3.4028254489880227,
      "tokens_seen": 2205286400
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015040665783998486,
      "loss": 2.5533,
      "theoretical_loss": 3.4028171787272057,
      "tokens_seen": 2205351936
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001503901078116134,
      "loss": 2.8271,
      "theoretical_loss": 3.402808908780963,
      "tokens_seen": 2205417472
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001503735577832419,
      "loss": 2.798,
      "theoretical_loss": 3.402800639149273,
      "tokens_seen": 2205483008
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015035700775487045,
      "loss": 2.3197,
      "theoretical_loss": 3.4027923698321145,
      "tokens_seen": 2205548544
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015034045772649896,
      "loss": 2.8102,
      "theoretical_loss": 3.402784100829466,
      "tokens_seen": 2205614080
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001503239076981275,
      "loss": 2.4017,
      "theoretical_loss": 3.402775832141307,
      "tokens_seen": 2205679616
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000150307357669756,
      "loss": 2.774,
      "theoretical_loss": 3.4027675637676147,
      "tokens_seen": 2205745152
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015029080764138452,
      "loss": 2.4838,
      "theoretical_loss": 3.402759295708369,
      "tokens_seen": 2205810688
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015027425761301305,
      "loss": 2.7235,
      "theoretical_loss": 3.4027510279635482,
      "tokens_seen": 2205876224
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015025770758464156,
      "loss": 2.8257,
      "theoretical_loss": 3.402742760533131,
      "tokens_seen": 2205941760
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001502411575562701,
      "loss": 2.2963,
      "theoretical_loss": 3.402734493417096,
      "tokens_seen": 2206007296
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001502246075278986,
      "loss": 2.6573,
      "theoretical_loss": 3.402726226615423,
      "tokens_seen": 2206072832
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015020805749952715,
      "loss": 2.3581,
      "theoretical_loss": 3.4027179601280886,
      "tokens_seen": 2206138368
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015019150747115566,
      "loss": 2.7361,
      "theoretical_loss": 3.402709693955073,
      "tokens_seen": 2206203904
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001501749574427842,
      "loss": 2.7497,
      "theoretical_loss": 3.402701428096355,
      "tokens_seen": 2206269440
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001501584074144127,
      "loss": 2.492,
      "theoretical_loss": 3.402693162551912,
      "tokens_seen": 2206334976
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015014185738604124,
      "loss": 2.6772,
      "theoretical_loss": 3.4026848973217243,
      "tokens_seen": 2206400512
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015012530735766976,
      "loss": 2.7111,
      "theoretical_loss": 3.40267663240577,
      "tokens_seen": 2206466048
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001501087573292983,
      "loss": 2.7967,
      "theoretical_loss": 3.4026683678040275,
      "tokens_seen": 2206531584
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001500922073009268,
      "loss": 2.7987,
      "theoretical_loss": 3.402660103516476,
      "tokens_seen": 2206597120
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015007565727255534,
      "loss": 3.1237,
      "theoretical_loss": 3.402651839543094,
      "tokens_seen": 2206662656
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015005910724418385,
      "loss": 2.8729,
      "theoretical_loss": 3.4026435758838596,
      "tokens_seen": 2206728192
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015004255721581236,
      "loss": 2.739,
      "theoretical_loss": 3.4026353125387527,
      "tokens_seen": 2206793728
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00015002600718744087,
      "loss": 2.6781,
      "theoretical_loss": 3.4026270495077515,
      "tokens_seen": 2206859264
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001500094571590694,
      "loss": 2.7271,
      "theoretical_loss": 3.4026187867908346,
      "tokens_seen": 2206924800
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014999290713069792,
      "loss": 2.861,
      "theoretical_loss": 3.4026105243879807,
      "tokens_seen": 2206990336
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014997635710232646,
      "loss": 2.9602,
      "theoretical_loss": 3.402602262299169,
      "tokens_seen": 2207055872
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014995980707395497,
      "loss": 2.7654,
      "theoretical_loss": 3.4025940005243775,
      "tokens_seen": 2207121408
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001499432570455835,
      "loss": 2.7646,
      "theoretical_loss": 3.4025857390635856,
      "tokens_seen": 2207186944
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014992670701721202,
      "loss": 2.9838,
      "theoretical_loss": 3.402577477916772,
      "tokens_seen": 2207252480
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014991015698884055,
      "loss": 2.6283,
      "theoretical_loss": 3.402569217083915,
      "tokens_seen": 2207318016
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014989360696046906,
      "loss": 2.5497,
      "theoretical_loss": 3.4025609565649937,
      "tokens_seen": 2207383552
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001498770569320976,
      "loss": 2.8643,
      "theoretical_loss": 3.4025526963599866,
      "tokens_seen": 2207449088
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001498605069037261,
      "loss": 2.6796,
      "theoretical_loss": 3.402544436468873,
      "tokens_seen": 2207514624
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014984395687535465,
      "loss": 2.7812,
      "theoretical_loss": 3.402536176891631,
      "tokens_seen": 2207580160
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014982740684698316,
      "loss": 2.6947,
      "theoretical_loss": 3.4025279176282397,
      "tokens_seen": 2207645696
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001498108568186117,
      "loss": 2.7534,
      "theoretical_loss": 3.4025196586786777,
      "tokens_seen": 2207711232
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001497943067902402,
      "loss": 2.6397,
      "theoretical_loss": 3.4025114000429237,
      "tokens_seen": 2207776768
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014977775676186872,
      "loss": 2.6847,
      "theoretical_loss": 3.402503141720957,
      "tokens_seen": 2207842304
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014976120673349726,
      "loss": 2.8205,
      "theoretical_loss": 3.4024948837127558,
      "tokens_seen": 2207907840
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014974465670512577,
      "loss": 2.7621,
      "theoretical_loss": 3.4024866260182987,
      "tokens_seen": 2207973376
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001497281066767543,
      "loss": 2.3302,
      "theoretical_loss": 3.4024783686375653,
      "tokens_seen": 2208038912
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014971155664838281,
      "loss": 2.7452,
      "theoretical_loss": 3.4024701115705334,
      "tokens_seen": 2208104448
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014969500662001135,
      "loss": 2.7961,
      "theoretical_loss": 3.4024618548171826,
      "tokens_seen": 2208169984
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014967845659163986,
      "loss": 2.7486,
      "theoretical_loss": 3.402453598377491,
      "tokens_seen": 2208235520
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001496619065632684,
      "loss": 2.75,
      "theoretical_loss": 3.402445342251438,
      "tokens_seen": 2208301056
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001496453565348969,
      "loss": 2.8841,
      "theoretical_loss": 3.4024370864390017,
      "tokens_seen": 2208366592
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014962880650652545,
      "loss": 2.9896,
      "theoretical_loss": 3.4024288309401616,
      "tokens_seen": 2208432128
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014961225647815396,
      "loss": 2.7967,
      "theoretical_loss": 3.402420575754896,
      "tokens_seen": 2208497664
    },
    {
      "epoch": 0.79,
      "objective/train/avg_token_score": 0.009206742979586124,
      "objective/train/avg_weight": 0.9891673922538757,
      "objective/train/docs_used": 1246233,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6186933517456055,
      "objective/train/std_weight": 0.037641193717718124,
      "objective/train/theoretical_loss": 3.4024123208831836,
      "objective/train/tokens_used": 2229023200,
      "theoretical_loss": 3.4024123208831836,
      "tokens_seen": 2208563200
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001495957064497825,
      "loss": 2.5592,
      "theoretical_loss": 3.4024123208831836,
      "tokens_seen": 2208563200
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000149579156421411,
      "loss": 2.6729,
      "theoretical_loss": 3.402404066325004,
      "tokens_seen": 2208628736
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014956260639303954,
      "loss": 2.7733,
      "theoretical_loss": 3.402395812080335,
      "tokens_seen": 2208694272
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014954605636466805,
      "loss": 2.8085,
      "theoretical_loss": 3.4023875581491554,
      "tokens_seen": 2208759808
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014952950633629656,
      "loss": 2.7476,
      "theoretical_loss": 3.4023793045314448,
      "tokens_seen": 2208825344
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001495129563079251,
      "loss": 2.7776,
      "theoretical_loss": 3.402371051227181,
      "tokens_seen": 2208890880
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001494964062795536,
      "loss": 3.1584,
      "theoretical_loss": 3.4023627982363442,
      "tokens_seen": 2208956416
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014947985625118215,
      "loss": 2.6483,
      "theoretical_loss": 3.402354545558912,
      "tokens_seen": 2209021952
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014946330622281066,
      "loss": 2.8336,
      "theoretical_loss": 3.4023462931948636,
      "tokens_seen": 2209087488
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001494467561944392,
      "loss": 2.5448,
      "theoretical_loss": 3.4023380411441777,
      "tokens_seen": 2209153024
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001494302061660677,
      "loss": 2.9501,
      "theoretical_loss": 3.402329789406833,
      "tokens_seen": 2209218560
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014941365613769625,
      "loss": 2.8361,
      "theoretical_loss": 3.402321537982809,
      "tokens_seen": 2209284096
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014939710610932476,
      "loss": 2.99,
      "theoretical_loss": 3.402313286872084,
      "tokens_seen": 2209349632
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001493805560809533,
      "loss": 2.7942,
      "theoretical_loss": 3.4023050360746363,
      "tokens_seen": 2209415168
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001493640060525818,
      "loss": 2.8035,
      "theoretical_loss": 3.4022967855904453,
      "tokens_seen": 2209480704
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014934745602421034,
      "loss": 2.5147,
      "theoretical_loss": 3.4022885354194896,
      "tokens_seen": 2209546240
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014933090599583885,
      "loss": 2.8222,
      "theoretical_loss": 3.4022802855617487,
      "tokens_seen": 2209611776
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014931435596746736,
      "loss": 2.7882,
      "theoretical_loss": 3.4022720360172007,
      "tokens_seen": 2209677312
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014929780593909587,
      "loss": 2.7113,
      "theoretical_loss": 3.402263786785824,
      "tokens_seen": 2209742848
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001492812559107244,
      "loss": 2.8667,
      "theoretical_loss": 3.4022555378675987,
      "tokens_seen": 2209808384
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014926470588235292,
      "loss": 2.9935,
      "theoretical_loss": 3.402247289262503,
      "tokens_seen": 2209873920
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014924815585398146,
      "loss": 2.8532,
      "theoretical_loss": 3.402239040970515,
      "tokens_seen": 2209939456
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014923160582560997,
      "loss": 2.7718,
      "theoretical_loss": 3.4022307929916145,
      "tokens_seen": 2210004992
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001492150557972385,
      "loss": 3.0526,
      "theoretical_loss": 3.40222254532578,
      "tokens_seen": 2210070528
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014919850576886702,
      "loss": 2.7066,
      "theoretical_loss": 3.4022142979729906,
      "tokens_seen": 2210136064
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014918195574049555,
      "loss": 2.6208,
      "theoretical_loss": 3.402206050933225,
      "tokens_seen": 2210201600
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014916540571212407,
      "loss": 2.8357,
      "theoretical_loss": 3.4021978042064616,
      "tokens_seen": 2210267136
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001491488556837526,
      "loss": 2.6351,
      "theoretical_loss": 3.4021895577926795,
      "tokens_seen": 2210332672
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001491323056553811,
      "loss": 2.8317,
      "theoretical_loss": 3.402181311691858,
      "tokens_seen": 2210398208
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014911575562700965,
      "loss": 2.9289,
      "theoretical_loss": 3.402173065903975,
      "tokens_seen": 2210463744
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001490992055986382,
      "loss": 2.9444,
      "theoretical_loss": 3.4021648204290105,
      "tokens_seen": 2210529280
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001490826555702667,
      "loss": 2.8669,
      "theoretical_loss": 3.4021565752669423,
      "tokens_seen": 2210594816
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001490661055418952,
      "loss": 2.5403,
      "theoretical_loss": 3.40214833041775,
      "tokens_seen": 2210660352
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014904955551352372,
      "loss": 2.8407,
      "theoretical_loss": 3.4021400858814124,
      "tokens_seen": 2210725888
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014903300548515226,
      "loss": 2.9286,
      "theoretical_loss": 3.4021318416579076,
      "tokens_seen": 2210791424
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014901645545678077,
      "loss": 2.6406,
      "theoretical_loss": 3.4021235977472153,
      "tokens_seen": 2210856960
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001489999054284093,
      "loss": 2.617,
      "theoretical_loss": 3.4021153541493137,
      "tokens_seen": 2210922496
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014898335540003782,
      "loss": 2.793,
      "theoretical_loss": 3.4021071108641823,
      "tokens_seen": 2210988032
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014896680537166635,
      "loss": 3.0171,
      "theoretical_loss": 3.4020988678918,
      "tokens_seen": 2211053568
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014895025534329486,
      "loss": 3.0001,
      "theoretical_loss": 3.4020906252321446,
      "tokens_seen": 2211119104
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001489337053149234,
      "loss": 2.848,
      "theoretical_loss": 3.4020823828851956,
      "tokens_seen": 2211184640
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001489171552865519,
      "loss": 2.9105,
      "theoretical_loss": 3.4020741408509325,
      "tokens_seen": 2211250176
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014890060525818045,
      "loss": 2.6973,
      "theoretical_loss": 3.4020658991293335,
      "tokens_seen": 2211315712
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014888405522980896,
      "loss": 3.0693,
      "theoretical_loss": 3.402057657720378,
      "tokens_seen": 2211381248
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001488675052014375,
      "loss": 2.7792,
      "theoretical_loss": 3.4020494166240436,
      "tokens_seen": 2211446784
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000148850955173066,
      "loss": 2.4088,
      "theoretical_loss": 3.402041175840311,
      "tokens_seen": 2211512320
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014883440514469455,
      "loss": 2.7645,
      "theoretical_loss": 3.4020329353691574,
      "tokens_seen": 2211577856
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014881785511632306,
      "loss": 2.7609,
      "theoretical_loss": 3.402024695210563,
      "tokens_seen": 2211643392
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014880130508795157,
      "loss": 2.6822,
      "theoretical_loss": 3.4020164553645054,
      "tokens_seen": 2211708928
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001487847550595801,
      "loss": 3.0766,
      "theoretical_loss": 3.4020082158309646,
      "tokens_seen": 2211774464
    },
    {
      "epoch": 0.79,
      "objective/train/avg_token_score": 0.0400233119726181,
      "objective/train/avg_weight": 0.9529027938842773,
      "objective/train/docs_used": 1246233,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.891270875930786,
      "objective/train/std_weight": 0.13878998160362244,
      "objective/train/theoretical_loss": 3.4019999766099187,
      "objective/train/tokens_used": 2232300000,
      "theoretical_loss": 3.4019999766099187,
      "tokens_seen": 2211840000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014876820503120861,
      "loss": 2.9376,
      "theoretical_loss": 3.4019999766099187,
      "tokens_seen": 2211840000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014875165500283715,
      "loss": 2.9727,
      "theoretical_loss": 3.4019917377013473,
      "tokens_seen": 2211905536
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014873510497446566,
      "loss": 2.9258,
      "theoretical_loss": 3.401983499105229,
      "tokens_seen": 2211971072
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001487185549460942,
      "loss": 2.9231,
      "theoretical_loss": 3.4019752608215423,
      "tokens_seen": 2212036608
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001487020049177227,
      "loss": 2.9519,
      "theoretical_loss": 3.401967022850267,
      "tokens_seen": 2212102144
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014868545488935125,
      "loss": 2.7549,
      "theoretical_loss": 3.401958785191381,
      "tokens_seen": 2212167680
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014866890486097976,
      "loss": 2.9076,
      "theoretical_loss": 3.401950547844864,
      "tokens_seen": 2212233216
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001486523548326083,
      "loss": 2.9584,
      "theoretical_loss": 3.4019423108106945,
      "tokens_seen": 2212298752
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001486358048042368,
      "loss": 3.0139,
      "theoretical_loss": 3.4019340740888513,
      "tokens_seen": 2212364288
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014861925477586534,
      "loss": 2.9096,
      "theoretical_loss": 3.401925837679313,
      "tokens_seen": 2212429824
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014860270474749385,
      "loss": 2.8497,
      "theoretical_loss": 3.40191760158206,
      "tokens_seen": 2212495360
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001485861547191224,
      "loss": 2.8582,
      "theoretical_loss": 3.401909365797069,
      "tokens_seen": 2212560896
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014856960469075087,
      "loss": 2.9197,
      "theoretical_loss": 3.4019011303243207,
      "tokens_seen": 2212626432
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001485530546623794,
      "loss": 2.9288,
      "theoretical_loss": 3.4018928951637935,
      "tokens_seen": 2212691968
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014853650463400792,
      "loss": 2.8994,
      "theoretical_loss": 3.401884660315466,
      "tokens_seen": 2212757504
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014851995460563646,
      "loss": 2.9437,
      "theoretical_loss": 3.4018764257793173,
      "tokens_seen": 2212823040
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014850340457726497,
      "loss": 2.854,
      "theoretical_loss": 3.4018681915553266,
      "tokens_seen": 2212888576
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001484868545488935,
      "loss": 2.8126,
      "theoretical_loss": 3.4018599576434725,
      "tokens_seen": 2212954112
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014847030452052205,
      "loss": 2.6698,
      "theoretical_loss": 3.4018517240437336,
      "tokens_seen": 2213019648
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014845375449215056,
      "loss": 2.9043,
      "theoretical_loss": 3.4018434907560895,
      "tokens_seen": 2213085184
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001484372044637791,
      "loss": 2.7799,
      "theoretical_loss": 3.401835257780519,
      "tokens_seen": 2213150720
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001484206544354076,
      "loss": 2.8006,
      "theoretical_loss": 3.401827025117001,
      "tokens_seen": 2213216256
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014840410440703614,
      "loss": 2.7379,
      "theoretical_loss": 3.401818792765514,
      "tokens_seen": 2213281792
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014838755437866465,
      "loss": 2.8893,
      "theoretical_loss": 3.4018105607260374,
      "tokens_seen": 2213347328
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001483710043502932,
      "loss": 2.7921,
      "theoretical_loss": 3.40180232899855,
      "tokens_seen": 2213412864
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001483544543219217,
      "loss": 2.9117,
      "theoretical_loss": 3.4017940975830308,
      "tokens_seen": 2213478400
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014833790429355024,
      "loss": 2.8463,
      "theoretical_loss": 3.4017858664794582,
      "tokens_seen": 2213543936
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014832135426517872,
      "loss": 2.9691,
      "theoretical_loss": 3.401777635687812,
      "tokens_seen": 2213609472
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014830480423680726,
      "loss": 2.7904,
      "theoretical_loss": 3.4017694052080705,
      "tokens_seen": 2213675008
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014828825420843577,
      "loss": 2.9198,
      "theoretical_loss": 3.4017611750402135,
      "tokens_seen": 2213740544
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001482717041800643,
      "loss": 2.9988,
      "theoretical_loss": 3.401752945184219,
      "tokens_seen": 2213806080
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014825515415169282,
      "loss": 2.5195,
      "theoretical_loss": 3.401744715640066,
      "tokens_seen": 2213871616
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014823860412332135,
      "loss": 2.9007,
      "theoretical_loss": 3.401736486407734,
      "tokens_seen": 2213937152
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014822205409494987,
      "loss": 2.792,
      "theoretical_loss": 3.4017282574872016,
      "tokens_seen": 2214002688
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001482055040665784,
      "loss": 2.7483,
      "theoretical_loss": 3.401720028878448,
      "tokens_seen": 2214068224
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001481889540382069,
      "loss": 2.9469,
      "theoretical_loss": 3.4017118005814524,
      "tokens_seen": 2214133760
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014817240400983545,
      "loss": 2.927,
      "theoretical_loss": 3.4017035725961926,
      "tokens_seen": 2214199296
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014815585398146396,
      "loss": 2.9383,
      "theoretical_loss": 3.401695344922649,
      "tokens_seen": 2214264832
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001481393039530925,
      "loss": 2.5808,
      "theoretical_loss": 3.4016871175607997,
      "tokens_seen": 2214330368
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000148122753924721,
      "loss": 2.5356,
      "theoretical_loss": 3.401678890510624,
      "tokens_seen": 2214395904
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014810620389634955,
      "loss": 2.8175,
      "theoretical_loss": 3.4016706637721006,
      "tokens_seen": 2214461440
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014808965386797806,
      "loss": 2.5656,
      "theoretical_loss": 3.4016624373452085,
      "tokens_seen": 2214526976
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014807310383960657,
      "loss": 2.9039,
      "theoretical_loss": 3.401654211229927,
      "tokens_seen": 2214592512
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001480565538112351,
      "loss": 2.647,
      "theoretical_loss": 3.4016459854262346,
      "tokens_seen": 2214658048
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014804000378286362,
      "loss": 2.8832,
      "theoretical_loss": 3.4016377599341108,
      "tokens_seen": 2214723584
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014802345375449215,
      "loss": 2.7106,
      "theoretical_loss": 3.4016295347535346,
      "tokens_seen": 2214789120
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014800690372612066,
      "loss": 2.8623,
      "theoretical_loss": 3.4016213098844843,
      "tokens_seen": 2214854656
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001479903536977492,
      "loss": 3.1278,
      "theoretical_loss": 3.4016130853269395,
      "tokens_seen": 2214920192
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001479738036693777,
      "loss": 3.0314,
      "theoretical_loss": 3.4016048610808785,
      "tokens_seen": 2214985728
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014795725364100625,
      "loss": 3.31,
      "theoretical_loss": 3.401596637146281,
      "tokens_seen": 2215051264
    },
    {
      "epoch": 0.79,
      "objective/train/avg_token_score": 0.015592413023114204,
      "objective/train/avg_weight": 0.9816524386405945,
      "objective/train/docs_used": 1248119,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.0398130416870117,
      "objective/train/std_weight": 0.0802825540304184,
      "objective/train/theoretical_loss": 3.4015884135231262,
      "objective/train/tokens_used": 2235576800,
      "theoretical_loss": 3.4015884135231262,
      "tokens_seen": 2215116800
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014794070361263476,
      "loss": 2.9154,
      "theoretical_loss": 3.4015884135231262,
      "tokens_seen": 2215116800
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001479241535842633,
      "loss": 3.0188,
      "theoretical_loss": 3.4015801902113925,
      "tokens_seen": 2215182336
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001479076035558918,
      "loss": 2.7906,
      "theoretical_loss": 3.401571967211059,
      "tokens_seen": 2215247872
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014789105352752034,
      "loss": 2.9076,
      "theoretical_loss": 3.4015637445221043,
      "tokens_seen": 2215313408
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014787450349914886,
      "loss": 2.8649,
      "theoretical_loss": 3.4015555221445086,
      "tokens_seen": 2215378944
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001478579534707774,
      "loss": 2.8638,
      "theoretical_loss": 3.4015473000782497,
      "tokens_seen": 2215444480
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001478414034424059,
      "loss": 2.8131,
      "theoretical_loss": 3.401539078323307,
      "tokens_seen": 2215510016
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014782485341403441,
      "loss": 3.2655,
      "theoretical_loss": 3.40153085687966,
      "tokens_seen": 2215575552
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014780830338566295,
      "loss": 2.8246,
      "theoretical_loss": 3.4015226357472867,
      "tokens_seen": 2215641088
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014779175335729146,
      "loss": 2.8555,
      "theoretical_loss": 3.401514414926167,
      "tokens_seen": 2215706624
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014777520332892,
      "loss": 2.7722,
      "theoretical_loss": 3.401506194416279,
      "tokens_seen": 2215772160
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001477586533005485,
      "loss": 2.794,
      "theoretical_loss": 3.4014979742176026,
      "tokens_seen": 2215837696
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014774210327217705,
      "loss": 2.7497,
      "theoretical_loss": 3.401489754330117,
      "tokens_seen": 2215903232
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014772555324380556,
      "loss": 2.5649,
      "theoretical_loss": 3.4014815347538,
      "tokens_seen": 2215968768
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001477090032154341,
      "loss": 2.5301,
      "theoretical_loss": 3.401473315488632,
      "tokens_seen": 2216034304
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001476924531870626,
      "loss": 2.8929,
      "theoretical_loss": 3.4014650965345905,
      "tokens_seen": 2216099840
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014767590315869114,
      "loss": 3.1026,
      "theoretical_loss": 3.401456877891656,
      "tokens_seen": 2216165376
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014765935313031965,
      "loss": 2.8601,
      "theoretical_loss": 3.401448659559807,
      "tokens_seen": 2216230912
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001476428031019482,
      "loss": 2.6624,
      "theoretical_loss": 3.401440441539022,
      "tokens_seen": 2216296448
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001476262530735767,
      "loss": 2.7836,
      "theoretical_loss": 3.4014322238292807,
      "tokens_seen": 2216361984
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014760970304520524,
      "loss": 2.7835,
      "theoretical_loss": 3.401424006430562,
      "tokens_seen": 2216427520
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014759315301683372,
      "loss": 2.6853,
      "theoretical_loss": 3.4014157893428445,
      "tokens_seen": 2216493056
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014757660298846226,
      "loss": 2.7186,
      "theoretical_loss": 3.4014075725661077,
      "tokens_seen": 2216558592
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014756005296009077,
      "loss": 2.7564,
      "theoretical_loss": 3.40139935610033,
      "tokens_seen": 2216624128
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001475435029317193,
      "loss": 2.6954,
      "theoretical_loss": 3.401391139945492,
      "tokens_seen": 2216689664
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014752695290334782,
      "loss": 2.6764,
      "theoretical_loss": 3.401382924101571,
      "tokens_seen": 2216755200
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014751040287497636,
      "loss": 2.575,
      "theoretical_loss": 3.4013747085685466,
      "tokens_seen": 2216820736
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014749385284660487,
      "loss": 2.665,
      "theoretical_loss": 3.4013664933463983,
      "tokens_seen": 2216886272
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001474773028182334,
      "loss": 2.72,
      "theoretical_loss": 3.4013582784351044,
      "tokens_seen": 2216951808
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014746075278986191,
      "loss": 2.684,
      "theoretical_loss": 3.401350063834645,
      "tokens_seen": 2217017344
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014744420276149045,
      "loss": 3.01,
      "theoretical_loss": 3.401341849544998,
      "tokens_seen": 2217082880
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014742765273311896,
      "loss": 2.9205,
      "theoretical_loss": 3.401333635566143,
      "tokens_seen": 2217148416
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001474111027047475,
      "loss": 2.8189,
      "theoretical_loss": 3.401325421898059,
      "tokens_seen": 2217213952
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000147394552676376,
      "loss": 2.7629,
      "theoretical_loss": 3.4013172085407253,
      "tokens_seen": 2217279488
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014737800264800455,
      "loss": 2.7014,
      "theoretical_loss": 3.4013089954941202,
      "tokens_seen": 2217345024
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014736145261963306,
      "loss": 2.4646,
      "theoretical_loss": 3.401300782758224,
      "tokens_seen": 2217410560
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014734490259126157,
      "loss": 2.8048,
      "theoretical_loss": 3.4012925703330144,
      "tokens_seen": 2217476096
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001473283525628901,
      "loss": 2.6642,
      "theoretical_loss": 3.4012843582184717,
      "tokens_seen": 2217541632
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014731180253451862,
      "loss": 2.7025,
      "theoretical_loss": 3.401276146414574,
      "tokens_seen": 2217607168
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014729525250614715,
      "loss": 2.7134,
      "theoretical_loss": 3.401267934921301,
      "tokens_seen": 2217672704
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014727870247777566,
      "loss": 2.7049,
      "theoretical_loss": 3.4012597237386313,
      "tokens_seen": 2217738240
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001472621524494042,
      "loss": 2.791,
      "theoretical_loss": 3.401251512866544,
      "tokens_seen": 2217803776
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001472456024210327,
      "loss": 2.9213,
      "theoretical_loss": 3.4012433023050193,
      "tokens_seen": 2217869312
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014722905239266125,
      "loss": 2.7809,
      "theoretical_loss": 3.4012350920540344,
      "tokens_seen": 2217934848
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014721250236428976,
      "loss": 2.826,
      "theoretical_loss": 3.4012268821135696,
      "tokens_seen": 2218000384
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001471959523359183,
      "loss": 2.8853,
      "theoretical_loss": 3.4012186724836035,
      "tokens_seen": 2218065920
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001471794023075468,
      "loss": 2.9743,
      "theoretical_loss": 3.401210463164116,
      "tokens_seen": 2218131456
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014716285227917535,
      "loss": 2.6894,
      "theoretical_loss": 3.401202254155085,
      "tokens_seen": 2218196992
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014714630225080386,
      "loss": 2.7878,
      "theoretical_loss": 3.4011940454564904,
      "tokens_seen": 2218262528
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001471297522224324,
      "loss": 2.7361,
      "theoretical_loss": 3.401185837068311,
      "tokens_seen": 2218328064
    },
    {
      "epoch": 0.79,
      "objective/train/avg_token_score": 0.01746709644794464,
      "objective/train/avg_weight": 0.9794405102729797,
      "objective/train/docs_used": 1249212,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.4928665161132812,
      "objective/train/std_weight": 0.09897966682910919,
      "objective/train/theoretical_loss": 3.401177628990526,
      "objective/train/tokens_used": 2238853600,
      "theoretical_loss": 3.401177628990526,
      "tokens_seen": 2218393600
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001471132021940609,
      "loss": 2.8853,
      "theoretical_loss": 3.401177628990526,
      "tokens_seen": 2218393600
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014709665216568942,
      "loss": 2.8788,
      "theoretical_loss": 3.4011694212231145,
      "tokens_seen": 2218459136
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014708010213731795,
      "loss": 3.0224,
      "theoretical_loss": 3.401161213766055,
      "tokens_seen": 2218524672
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014706355210894646,
      "loss": 2.683,
      "theoretical_loss": 3.401153006619328,
      "tokens_seen": 2218590208
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000147047002080575,
      "loss": 2.9554,
      "theoretical_loss": 3.4011447997829114,
      "tokens_seen": 2218655744
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001470304520522035,
      "loss": 2.8018,
      "theoretical_loss": 3.4011365932567843,
      "tokens_seen": 2218721280
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014701390202383205,
      "loss": 2.7702,
      "theoretical_loss": 3.4011283870409263,
      "tokens_seen": 2218786816
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014699735199546056,
      "loss": 2.6471,
      "theoretical_loss": 3.4011201811353167,
      "tokens_seen": 2218852352
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001469808019670891,
      "loss": 2.6886,
      "theoretical_loss": 3.4011119755399335,
      "tokens_seen": 2218917888
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001469642519387176,
      "loss": 2.7476,
      "theoretical_loss": 3.4011037702547573,
      "tokens_seen": 2218983424
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014694770191034614,
      "loss": 2.7566,
      "theoretical_loss": 3.4010955652797663,
      "tokens_seen": 2219048960
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014693115188197466,
      "loss": 2.8469,
      "theoretical_loss": 3.4010873606149397,
      "tokens_seen": 2219114496
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001469146018536032,
      "loss": 2.6302,
      "theoretical_loss": 3.4010791562602565,
      "tokens_seen": 2219180032
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001468980518252317,
      "loss": 2.6785,
      "theoretical_loss": 3.4010709522156963,
      "tokens_seen": 2219245568
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014688150179686024,
      "loss": 2.9683,
      "theoretical_loss": 3.401062748481238,
      "tokens_seen": 2219311104
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014686495176848875,
      "loss": 3.0625,
      "theoretical_loss": 3.4010545450568603,
      "tokens_seen": 2219376640
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014684840174011726,
      "loss": 2.7042,
      "theoretical_loss": 3.4010463419425427,
      "tokens_seen": 2219442176
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014683185171174577,
      "loss": 2.7931,
      "theoretical_loss": 3.4010381391382647,
      "tokens_seen": 2219507712
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001468153016833743,
      "loss": 2.7793,
      "theoretical_loss": 3.401029936644005,
      "tokens_seen": 2219573248
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014679875165500282,
      "loss": 2.7325,
      "theoretical_loss": 3.4010217344597424,
      "tokens_seen": 2219638784
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014678220162663136,
      "loss": 2.6132,
      "theoretical_loss": 3.4010135325854565,
      "tokens_seen": 2219704320
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014676565159825987,
      "loss": 2.7432,
      "theoretical_loss": 3.4010053310211266,
      "tokens_seen": 2219769856
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001467491015698884,
      "loss": 2.6322,
      "theoretical_loss": 3.4009971297667314,
      "tokens_seen": 2219835392
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014673255154151692,
      "loss": 2.8629,
      "theoretical_loss": 3.40098892882225,
      "tokens_seen": 2219900928
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014671600151314545,
      "loss": 2.5644,
      "theoretical_loss": 3.4009807281876623,
      "tokens_seen": 2219966464
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014669945148477396,
      "loss": 2.7685,
      "theoretical_loss": 3.4009725278629466,
      "tokens_seen": 2220032000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001466829014564025,
      "loss": 2.9602,
      "theoretical_loss": 3.400964327848082,
      "tokens_seen": 2220097536
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000146666351428031,
      "loss": 2.7762,
      "theoretical_loss": 3.4009561281430485,
      "tokens_seen": 2220163072
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014664980139965955,
      "loss": 2.7502,
      "theoretical_loss": 3.4009479287478244,
      "tokens_seen": 2220228608
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014663325137128806,
      "loss": 2.7594,
      "theoretical_loss": 3.4009397296623893,
      "tokens_seen": 2220294144
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001466167013429166,
      "loss": 2.6698,
      "theoretical_loss": 3.400931530886722,
      "tokens_seen": 2220359680
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001466001513145451,
      "loss": 2.7069,
      "theoretical_loss": 3.4009233324208026,
      "tokens_seen": 2220425216
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014658360128617362,
      "loss": 2.8553,
      "theoretical_loss": 3.4009151342646087,
      "tokens_seen": 2220490752
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014656705125780216,
      "loss": 2.7236,
      "theoretical_loss": 3.400906936418121,
      "tokens_seen": 2220556288
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014655050122943067,
      "loss": 2.7532,
      "theoretical_loss": 3.400898738881317,
      "tokens_seen": 2220621824
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001465339512010592,
      "loss": 2.7515,
      "theoretical_loss": 3.400890541654178,
      "tokens_seen": 2220687360
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014651740117268771,
      "loss": 2.57,
      "theoretical_loss": 3.400882344736681,
      "tokens_seen": 2220752896
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014650085114431625,
      "loss": 2.6145,
      "theoretical_loss": 3.400874148128807,
      "tokens_seen": 2220818432
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014648430111594476,
      "loss": 2.6303,
      "theoretical_loss": 3.4008659518305335,
      "tokens_seen": 2220883968
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001464677510875733,
      "loss": 2.7864,
      "theoretical_loss": 3.400857755841841,
      "tokens_seen": 2220949504
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001464512010592018,
      "loss": 2.8894,
      "theoretical_loss": 3.4008495601627082,
      "tokens_seen": 2221015040
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014643465103083035,
      "loss": 2.8387,
      "theoretical_loss": 3.4008413647931137,
      "tokens_seen": 2221080576
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014641810100245886,
      "loss": 2.7137,
      "theoretical_loss": 3.400833169733038,
      "tokens_seen": 2221146112
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001464015509740874,
      "loss": 2.832,
      "theoretical_loss": 3.400824974982459,
      "tokens_seen": 2221211648
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001463850009457159,
      "loss": 2.9044,
      "theoretical_loss": 3.4008167805413563,
      "tokens_seen": 2221277184
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014636845091734444,
      "loss": 2.8055,
      "theoretical_loss": 3.4008085864097093,
      "tokens_seen": 2221342720
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014635190088897295,
      "loss": 2.8361,
      "theoretical_loss": 3.400800392587497,
      "tokens_seen": 2221408256
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014633535086060146,
      "loss": 2.8799,
      "theoretical_loss": 3.4007921990746985,
      "tokens_seen": 2221473792
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014631880083223,
      "loss": 2.8138,
      "theoretical_loss": 3.4007840058712935,
      "tokens_seen": 2221539328
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001463022508038585,
      "loss": 2.7948,
      "theoretical_loss": 3.4007758129772605,
      "tokens_seen": 2221604864
    },
    {
      "epoch": 0.79,
      "objective/train/avg_token_score": 0.013052751310169697,
      "objective/train/avg_weight": 0.984634280204773,
      "objective/train/docs_used": 1251383,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4667701721191406,
      "objective/train/std_weight": 0.06705272197723389,
      "objective/train/theoretical_loss": 3.400767620392579,
      "objective/train/tokens_used": 2242130400,
      "theoretical_loss": 3.400767620392579,
      "tokens_seen": 2221670400
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014628570077548705,
      "loss": 2.6686,
      "theoretical_loss": 3.400767620392579,
      "tokens_seen": 2221670400
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014626915074711556,
      "loss": 2.6246,
      "theoretical_loss": 3.400759428117228,
      "tokens_seen": 2221735936
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001462526007187441,
      "loss": 2.6324,
      "theoretical_loss": 3.4007512361511876,
      "tokens_seen": 2221801472
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001462360506903726,
      "loss": 2.5845,
      "theoretical_loss": 3.400743044494436,
      "tokens_seen": 2221867008
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014621950066200115,
      "loss": 2.4006,
      "theoretical_loss": 3.400734853146952,
      "tokens_seen": 2221932544
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014620295063362966,
      "loss": 2.8512,
      "theoretical_loss": 3.400726662108716,
      "tokens_seen": 2221998080
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001461864006052582,
      "loss": 2.9544,
      "theoretical_loss": 3.4007184713797067,
      "tokens_seen": 2222063616
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001461698505768867,
      "loss": 3.0793,
      "theoretical_loss": 3.4007102809599035,
      "tokens_seen": 2222129152
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014615330054851524,
      "loss": 2.7176,
      "theoretical_loss": 3.4007020908492853,
      "tokens_seen": 2222194688
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014613675052014375,
      "loss": 2.8464,
      "theoretical_loss": 3.4006939010478314,
      "tokens_seen": 2222260224
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001461202004917723,
      "loss": 2.8333,
      "theoretical_loss": 3.400685711555521,
      "tokens_seen": 2222325760
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014610365046340077,
      "loss": 2.7786,
      "theoretical_loss": 3.4006775223723333,
      "tokens_seen": 2222391296
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001460871004350293,
      "loss": 2.6788,
      "theoretical_loss": 3.4006693334982474,
      "tokens_seen": 2222456832
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014607055040665782,
      "loss": 2.5927,
      "theoretical_loss": 3.4006611449332427,
      "tokens_seen": 2222522368
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014605400037828636,
      "loss": 2.7677,
      "theoretical_loss": 3.4006529566772987,
      "tokens_seen": 2222587904
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014603745034991487,
      "loss": 2.5662,
      "theoretical_loss": 3.4006447687303942,
      "tokens_seen": 2222653440
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001460209003215434,
      "loss": 2.9127,
      "theoretical_loss": 3.4006365810925088,
      "tokens_seen": 2222718976
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014600435029317192,
      "loss": 2.7305,
      "theoretical_loss": 3.400628393763621,
      "tokens_seen": 2222784512
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014598780026480045,
      "loss": 2.7985,
      "theoretical_loss": 3.4006202067437106,
      "tokens_seen": 2222850048
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014597125023642897,
      "loss": 2.837,
      "theoretical_loss": 3.400612020032757,
      "tokens_seen": 2222915584
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001459547002080575,
      "loss": 2.8007,
      "theoretical_loss": 3.400603833630739,
      "tokens_seen": 2222981120
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000145938150179686,
      "loss": 2.9037,
      "theoretical_loss": 3.400595647537636,
      "tokens_seen": 2223046656
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014592160015131455,
      "loss": 2.5049,
      "theoretical_loss": 3.4005874617534273,
      "tokens_seen": 2223112192
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014590505012294306,
      "loss": 2.8514,
      "theoretical_loss": 3.400579276278092,
      "tokens_seen": 2223177728
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001458885000945716,
      "loss": 2.8376,
      "theoretical_loss": 3.400571091111609,
      "tokens_seen": 2223243264
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001458719500662001,
      "loss": 2.8825,
      "theoretical_loss": 3.400562906253959,
      "tokens_seen": 2223308800
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014585540003782862,
      "loss": 2.8796,
      "theoretical_loss": 3.4005547217051193,
      "tokens_seen": 2223374336
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014583885000945716,
      "loss": 2.6772,
      "theoretical_loss": 3.4005465374650705,
      "tokens_seen": 2223439872
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014582229998108567,
      "loss": 2.7831,
      "theoretical_loss": 3.400538353533791,
      "tokens_seen": 2223505408
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001458057499527142,
      "loss": 2.7772,
      "theoretical_loss": 3.400530169911261,
      "tokens_seen": 2223570944
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014578919992434272,
      "loss": 2.7436,
      "theoretical_loss": 3.4005219865974587,
      "tokens_seen": 2223636480
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014577264989597125,
      "loss": 2.698,
      "theoretical_loss": 3.4005138035923643,
      "tokens_seen": 2223702016
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014575609986759976,
      "loss": 2.2577,
      "theoretical_loss": 3.4005056208959563,
      "tokens_seen": 2223767552
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001457395498392283,
      "loss": 2.6751,
      "theoretical_loss": 3.4004974385082143,
      "tokens_seen": 2223833088
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001457229998108568,
      "loss": 2.5247,
      "theoretical_loss": 3.400489256429118,
      "tokens_seen": 2223898624
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014570644978248535,
      "loss": 2.817,
      "theoretical_loss": 3.4004810746586456,
      "tokens_seen": 2223964160
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014568989975411386,
      "loss": 2.5114,
      "theoretical_loss": 3.400472893196777,
      "tokens_seen": 2224029696
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001456733497257424,
      "loss": 2.7502,
      "theoretical_loss": 3.4004647120434917,
      "tokens_seen": 2224095232
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001456567996973709,
      "loss": 2.7969,
      "theoretical_loss": 3.4004565311987687,
      "tokens_seen": 2224160768
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014564024966899945,
      "loss": 2.1209,
      "theoretical_loss": 3.400448350662587,
      "tokens_seen": 2224226304
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014562369964062796,
      "loss": 2.6356,
      "theoretical_loss": 3.4004401704349263,
      "tokens_seen": 2224291840
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014560714961225647,
      "loss": 2.6972,
      "theoretical_loss": 3.400431990515766,
      "tokens_seen": 2224357376
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000145590599583885,
      "loss": 2.4675,
      "theoretical_loss": 3.400423810905085,
      "tokens_seen": 2224422912
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014557404955551351,
      "loss": 2.7914,
      "theoretical_loss": 3.400415631602862,
      "tokens_seen": 2224488448
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014555749952714205,
      "loss": 2.568,
      "theoretical_loss": 3.4004074526090777,
      "tokens_seen": 2224553984
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014554094949877056,
      "loss": 2.6512,
      "theoretical_loss": 3.4003992739237106,
      "tokens_seen": 2224619520
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001455243994703991,
      "loss": 2.6849,
      "theoretical_loss": 3.4003910955467394,
      "tokens_seen": 2224685056
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001455078494420276,
      "loss": 2.7682,
      "theoretical_loss": 3.4003829174781446,
      "tokens_seen": 2224750592
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014549129941365615,
      "loss": 2.617,
      "theoretical_loss": 3.400374739717905,
      "tokens_seen": 2224816128
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014547474938528466,
      "loss": 2.6631,
      "theoretical_loss": 3.400366562265999,
      "tokens_seen": 2224881664
    },
    {
      "epoch": 0.79,
      "objective/train/avg_token_score": 0.01728464663028717,
      "objective/train/avg_weight": 0.9796654582023621,
      "objective/train/docs_used": 1252992,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8972644805908203,
      "objective/train/std_weight": 0.08278554677963257,
      "objective/train/theoretical_loss": 3.4003583851224075,
      "objective/train/tokens_used": 2245407200,
      "theoretical_loss": 3.4003583851224075,
      "tokens_seen": 2224947200
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001454581993569132,
      "loss": 2.8243,
      "theoretical_loss": 3.4003583851224075,
      "tokens_seen": 2224947200
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001454416493285417,
      "loss": 2.6805,
      "theoretical_loss": 3.4003502082871084,
      "tokens_seen": 2225012736
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014542509930017024,
      "loss": 2.797,
      "theoretical_loss": 3.400342031760082,
      "tokens_seen": 2225078272
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014540854927179875,
      "loss": 2.4191,
      "theoretical_loss": 3.400333855541307,
      "tokens_seen": 2225143808
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001453919992434273,
      "loss": 2.7164,
      "theoretical_loss": 3.400325679630763,
      "tokens_seen": 2225209344
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014537544921505577,
      "loss": 2.7403,
      "theoretical_loss": 3.400317504028429,
      "tokens_seen": 2225274880
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001453588991866843,
      "loss": 2.7442,
      "theoretical_loss": 3.4003093287342843,
      "tokens_seen": 2225340416
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014534234915831282,
      "loss": 2.0037,
      "theoretical_loss": 3.4003011537483085,
      "tokens_seen": 2225405952
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014532579912994136,
      "loss": 2.773,
      "theoretical_loss": 3.400292979070481,
      "tokens_seen": 2225471488
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014530924910156987,
      "loss": 2.7519,
      "theoretical_loss": 3.4002848047007808,
      "tokens_seen": 2225537024
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001452926990731984,
      "loss": 2.6635,
      "theoretical_loss": 3.4002766306391874,
      "tokens_seen": 2225602560
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014527614904482692,
      "loss": 2.5025,
      "theoretical_loss": 3.40026845688568,
      "tokens_seen": 2225668096
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014525959901645546,
      "loss": 2.8225,
      "theoretical_loss": 3.400260283440238,
      "tokens_seen": 2225733632
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014524304898808397,
      "loss": 2.7036,
      "theoretical_loss": 3.4002521103028407,
      "tokens_seen": 2225799168
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001452264989597125,
      "loss": 2.6971,
      "theoretical_loss": 3.400243937473467,
      "tokens_seen": 2225864704
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00014520994893134101,
      "loss": 2.8281,
      "theoretical_loss": 3.4002357649520967,
      "tokens_seen": 2225930240
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014519339890296955,
      "loss": 2.9884,
      "theoretical_loss": 3.4002275927387093,
      "tokens_seen": 2225995776
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014517684887459806,
      "loss": 2.4491,
      "theoretical_loss": 3.4002194208332837,
      "tokens_seen": 2226061312
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001451602988462266,
      "loss": 2.7592,
      "theoretical_loss": 3.4002112492357996,
      "tokens_seen": 2226126848
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001451437488178551,
      "loss": 2.8448,
      "theoretical_loss": 3.400203077946236,
      "tokens_seen": 2226192384
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014512719878948362,
      "loss": 2.5993,
      "theoretical_loss": 3.4001949069645723,
      "tokens_seen": 2226257920
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014511064876111216,
      "loss": 2.758,
      "theoretical_loss": 3.400186736290788,
      "tokens_seen": 2226323456
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014509409873274067,
      "loss": 2.3569,
      "theoretical_loss": 3.400178565924862,
      "tokens_seen": 2226388992
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001450775487043692,
      "loss": 2.5826,
      "theoretical_loss": 3.400170395866774,
      "tokens_seen": 2226454528
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014506099867599772,
      "loss": 2.8286,
      "theoretical_loss": 3.400162226116504,
      "tokens_seen": 2226520064
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014504444864762625,
      "loss": 2.6845,
      "theoretical_loss": 3.40015405667403,
      "tokens_seen": 2226585600
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014502789861925477,
      "loss": 2.6639,
      "theoretical_loss": 3.4001458875393324,
      "tokens_seen": 2226651136
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001450113485908833,
      "loss": 2.3943,
      "theoretical_loss": 3.40013771871239,
      "tokens_seen": 2226716672
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001449947985625118,
      "loss": 2.8338,
      "theoretical_loss": 3.4001295501931823,
      "tokens_seen": 2226782208
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014497824853414035,
      "loss": 2.781,
      "theoretical_loss": 3.4001213819816885,
      "tokens_seen": 2226847744
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014496169850576886,
      "loss": 2.6311,
      "theoretical_loss": 3.400113214077888,
      "tokens_seen": 2226913280
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001449451484773974,
      "loss": 2.725,
      "theoretical_loss": 3.40010504648176,
      "tokens_seen": 2226978816
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001449285984490259,
      "loss": 2.6198,
      "theoretical_loss": 3.400096879193285,
      "tokens_seen": 2227044352
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014491204842065445,
      "loss": 2.553,
      "theoretical_loss": 3.400088712212441,
      "tokens_seen": 2227109888
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014489549839228296,
      "loss": 2.8833,
      "theoretical_loss": 3.4000805455392076,
      "tokens_seen": 2227175424
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014487894836391147,
      "loss": 2.3793,
      "theoretical_loss": 3.4000723791735643,
      "tokens_seen": 2227240960
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014486239833554,
      "loss": 2.7728,
      "theoretical_loss": 3.400064213115491,
      "tokens_seen": 2227306496
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014484584830716852,
      "loss": 2.7724,
      "theoretical_loss": 3.4000560473649664,
      "tokens_seen": 2227372032
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014482929827879705,
      "loss": 2.7415,
      "theoretical_loss": 3.40004788192197,
      "tokens_seen": 2227437568
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014481274825042556,
      "loss": 2.7589,
      "theoretical_loss": 3.4000397167864813,
      "tokens_seen": 2227503104
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001447961982220541,
      "loss": 2.7268,
      "theoretical_loss": 3.40003155195848,
      "tokens_seen": 2227568640
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001447796481936826,
      "loss": 2.9146,
      "theoretical_loss": 3.400023387437945,
      "tokens_seen": 2227634176
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014476309816531115,
      "loss": 2.9024,
      "theoretical_loss": 3.4000152232248553,
      "tokens_seen": 2227699712
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014474654813693966,
      "loss": 2.8041,
      "theoretical_loss": 3.400007059319191,
      "tokens_seen": 2227765248
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001447299981085682,
      "loss": 2.8745,
      "theoretical_loss": 3.3999988957209313,
      "tokens_seen": 2227830784
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001447134480801967,
      "loss": 2.8056,
      "theoretical_loss": 3.3999907324300556,
      "tokens_seen": 2227896320
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014469689805182524,
      "loss": 2.8367,
      "theoretical_loss": 3.399982569446543,
      "tokens_seen": 2227961856
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014468034802345376,
      "loss": 2.8306,
      "theoretical_loss": 3.399974406770373,
      "tokens_seen": 2228027392
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001446637979950823,
      "loss": 2.5767,
      "theoretical_loss": 3.3999662444015257,
      "tokens_seen": 2228092928
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001446472479667108,
      "loss": 2.7863,
      "theoretical_loss": 3.399958082339979,
      "tokens_seen": 2228158464
    },
    {
      "epoch": 0.8,
      "objective/train/avg_token_score": 0.02657581865787506,
      "objective/train/avg_weight": 0.9687566161155701,
      "objective/train/docs_used": 1254892,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.282907485961914,
      "objective/train/std_weight": 0.12644411623477936,
      "objective/train/theoretical_loss": 3.3999499205857138,
      "objective/train/tokens_used": 2248684000,
      "theoretical_loss": 3.3999499205857138,
      "tokens_seen": 2228224000
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014463069793833931,
      "loss": 2.3238,
      "theoretical_loss": 3.3999499205857138,
      "tokens_seen": 2228224000
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014461414790996782,
      "loss": 2.6709,
      "theoretical_loss": 3.399941759138709,
      "tokens_seen": 2228289536
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014459759788159636,
      "loss": 2.7102,
      "theoretical_loss": 3.399933597998943,
      "tokens_seen": 2228355072
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014458104785322487,
      "loss": 2.6639,
      "theoretical_loss": 3.399925437166397,
      "tokens_seen": 2228420608
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001445644978248534,
      "loss": 2.5767,
      "theoretical_loss": 3.399917276641049,
      "tokens_seen": 2228486144
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014454794779648192,
      "loss": 2.8897,
      "theoretical_loss": 3.399909116422879,
      "tokens_seen": 2228551680
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014453139776811046,
      "loss": 2.6838,
      "theoretical_loss": 3.399900956511866,
      "tokens_seen": 2228617216
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014451484773973897,
      "loss": 2.7761,
      "theoretical_loss": 3.3998927969079897,
      "tokens_seen": 2228682752
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001444982977113675,
      "loss": 2.5389,
      "theoretical_loss": 3.3998846376112297,
      "tokens_seen": 2228748288
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014448174768299604,
      "loss": 2.7131,
      "theoretical_loss": 3.3998764786215654,
      "tokens_seen": 2228813824
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014446519765462455,
      "loss": 2.9202,
      "theoretical_loss": 3.3998683199389754,
      "tokens_seen": 2228879360
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001444486476262531,
      "loss": 2.6509,
      "theoretical_loss": 3.3998601615634403,
      "tokens_seen": 2228944896
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001444320975978816,
      "loss": 2.7905,
      "theoretical_loss": 3.3998520034949387,
      "tokens_seen": 2229010432
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014441554756951014,
      "loss": 2.8189,
      "theoretical_loss": 3.39984384573345,
      "tokens_seen": 2229075968
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014439899754113862,
      "loss": 2.5902,
      "theoretical_loss": 3.3998356882789538,
      "tokens_seen": 2229141504
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014438244751276716,
      "loss": 2.4191,
      "theoretical_loss": 3.39982753113143,
      "tokens_seen": 2229207040
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014436589748439567,
      "loss": 2.7017,
      "theoretical_loss": 3.3998193742908573,
      "tokens_seen": 2229272576
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001443493474560242,
      "loss": 2.6216,
      "theoretical_loss": 3.3998112177572155,
      "tokens_seen": 2229338112
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014433279742765272,
      "loss": 2.6732,
      "theoretical_loss": 3.399803061530484,
      "tokens_seen": 2229403648
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014431624739928126,
      "loss": 2.5031,
      "theoretical_loss": 3.3997949056106425,
      "tokens_seen": 2229469184
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014429969737090977,
      "loss": 2.9006,
      "theoretical_loss": 3.39978674999767,
      "tokens_seen": 2229534720
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001442831473425383,
      "loss": 2.6861,
      "theoretical_loss": 3.399778594691546,
      "tokens_seen": 2229600256
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014426659731416681,
      "loss": 2.4142,
      "theoretical_loss": 3.39977043969225,
      "tokens_seen": 2229665792
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014425004728579535,
      "loss": 2.7304,
      "theoretical_loss": 3.3997622849997615,
      "tokens_seen": 2229731328
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014423349725742386,
      "loss": 2.4855,
      "theoretical_loss": 3.3997541306140597,
      "tokens_seen": 2229796864
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001442169472290524,
      "loss": 2.7768,
      "theoretical_loss": 3.399745976535124,
      "tokens_seen": 2229862400
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001442003972006809,
      "loss": 2.6658,
      "theoretical_loss": 3.3997378227629347,
      "tokens_seen": 2229927936
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014418384717230945,
      "loss": 2.6524,
      "theoretical_loss": 3.39972966929747,
      "tokens_seen": 2229993472
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014416729714393796,
      "loss": 2.7198,
      "theoretical_loss": 3.3997215161387104,
      "tokens_seen": 2230059008
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014415074711556647,
      "loss": 2.7617,
      "theoretical_loss": 3.3997133632866348,
      "tokens_seen": 2230124544
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000144134197087195,
      "loss": 2.8041,
      "theoretical_loss": 3.3997052107412227,
      "tokens_seen": 2230190080
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014411764705882352,
      "loss": 2.4576,
      "theoretical_loss": 3.3996970585024537,
      "tokens_seen": 2230255616
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014410109703045205,
      "loss": 2.6632,
      "theoretical_loss": 3.399688906570307,
      "tokens_seen": 2230321152
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014408454700208056,
      "loss": 2.1143,
      "theoretical_loss": 3.3996807549447627,
      "tokens_seen": 2230386688
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001440679969737091,
      "loss": 2.5459,
      "theoretical_loss": 3.3996726036257994,
      "tokens_seen": 2230452224
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001440514469453376,
      "loss": 2.7068,
      "theoretical_loss": 3.399664452613397,
      "tokens_seen": 2230517760
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014403489691696615,
      "loss": 2.6775,
      "theoretical_loss": 3.3996563019075348,
      "tokens_seen": 2230583296
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014401834688859466,
      "loss": 2.4889,
      "theoretical_loss": 3.3996481515081927,
      "tokens_seen": 2230648832
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001440017968602232,
      "loss": 2.6833,
      "theoretical_loss": 3.3996400014153494,
      "tokens_seen": 2230714368
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001439852468318517,
      "loss": 2.653,
      "theoretical_loss": 3.399631851628985,
      "tokens_seen": 2230779904
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014396869680348025,
      "loss": 2.4718,
      "theoretical_loss": 3.399623702149079,
      "tokens_seen": 2230845440
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014395214677510876,
      "loss": 2.4784,
      "theoretical_loss": 3.3996155529756105,
      "tokens_seen": 2230910976
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001439355967467373,
      "loss": 2.3589,
      "theoretical_loss": 3.399607404108559,
      "tokens_seen": 2230976512
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001439190467183658,
      "loss": 2.4835,
      "theoretical_loss": 3.3995992555479044,
      "tokens_seen": 2231042048
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014390249668999432,
      "loss": 2.6244,
      "theoretical_loss": 3.3995911072936256,
      "tokens_seen": 2231107584
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014388594666162283,
      "loss": 2.7438,
      "theoretical_loss": 3.399582959345703,
      "tokens_seen": 2231173120
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014386939663325136,
      "loss": 2.5735,
      "theoretical_loss": 3.399574811704115,
      "tokens_seen": 2231238656
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001438528466048799,
      "loss": 2.5662,
      "theoretical_loss": 3.3995666643688414,
      "tokens_seen": 2231304192
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001438362965765084,
      "loss": 2.8149,
      "theoretical_loss": 3.399558517339862,
      "tokens_seen": 2231369728
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014381974654813695,
      "loss": 2.6741,
      "theoretical_loss": 3.399550370617156,
      "tokens_seen": 2231435264
    },
    {
      "epoch": 0.8,
      "objective/train/avg_token_score": 0.009761345572769642,
      "objective/train/avg_weight": 0.9885103106498718,
      "objective/train/docs_used": 1257149,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5301570892333984,
      "objective/train/std_weight": 0.060410019010305405,
      "objective/train/theoretical_loss": 3.3995422242007036,
      "objective/train/tokens_used": 2251960800,
      "theoretical_loss": 3.3995422242007036,
      "tokens_seen": 2231500800
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014380319651976546,
      "loss": 2.5572,
      "theoretical_loss": 3.3995422242007036,
      "tokens_seen": 2231500800
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000143786646491394,
      "loss": 2.4341,
      "theoretical_loss": 3.399534078090483,
      "tokens_seen": 2231566336
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001437700964630225,
      "loss": 2.7161,
      "theoretical_loss": 3.399525932286475,
      "tokens_seen": 2231631872
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014375354643465104,
      "loss": 2.5665,
      "theoretical_loss": 3.3995177867886577,
      "tokens_seen": 2231697408
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014373699640627956,
      "loss": 2.5746,
      "theoretical_loss": 3.399509641597012,
      "tokens_seen": 2231762944
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001437204463779081,
      "loss": 2.707,
      "theoretical_loss": 3.399501496711517,
      "tokens_seen": 2231828480
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001437038963495366,
      "loss": 2.6441,
      "theoretical_loss": 3.3994933521321515,
      "tokens_seen": 2231894016
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014368734632116514,
      "loss": 2.7509,
      "theoretical_loss": 3.3994852078588957,
      "tokens_seen": 2231959552
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014367079629279365,
      "loss": 2.82,
      "theoretical_loss": 3.3994770638917293,
      "tokens_seen": 2232025088
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014365424626442216,
      "loss": 2.6032,
      "theoretical_loss": 3.399468920230631,
      "tokens_seen": 2232090624
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014363769623605067,
      "loss": 2.7271,
      "theoretical_loss": 3.3994607768755807,
      "tokens_seen": 2232156160
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001436211462076792,
      "loss": 2.6351,
      "theoretical_loss": 3.399452633826558,
      "tokens_seen": 2232221696
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014360459617930772,
      "loss": 2.7599,
      "theoretical_loss": 3.3994444910835426,
      "tokens_seen": 2232287232
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014358804615093626,
      "loss": 2.6102,
      "theoretical_loss": 3.3994363486465136,
      "tokens_seen": 2232352768
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014357149612256477,
      "loss": 2.4419,
      "theoretical_loss": 3.399428206515451,
      "tokens_seen": 2232418304
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001435549460941933,
      "loss": 2.5314,
      "theoretical_loss": 3.3994200646903336,
      "tokens_seen": 2232483840
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014353839606582182,
      "loss": 2.789,
      "theoretical_loss": 3.3994119231711415,
      "tokens_seen": 2232549376
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014352184603745035,
      "loss": 2.5779,
      "theoretical_loss": 3.3994037819578544,
      "tokens_seen": 2232614912
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014350529600907886,
      "loss": 2.8187,
      "theoretical_loss": 3.399395641050451,
      "tokens_seen": 2232680448
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001434887459807074,
      "loss": 2.8074,
      "theoretical_loss": 3.3993875004489116,
      "tokens_seen": 2232745984
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001434721959523359,
      "loss": 2.4204,
      "theoretical_loss": 3.3993793601532154,
      "tokens_seen": 2232811520
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014345564592396445,
      "loss": 2.5952,
      "theoretical_loss": 3.399371220163342,
      "tokens_seen": 2232877056
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014343909589559296,
      "loss": 2.8357,
      "theoretical_loss": 3.399363080479271,
      "tokens_seen": 2232942592
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001434225458672215,
      "loss": 2.5943,
      "theoretical_loss": 3.399354941100982,
      "tokens_seen": 2233008128
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014340599583885,
      "loss": 2.6113,
      "theoretical_loss": 3.399346802028454,
      "tokens_seen": 2233073664
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014338944581047852,
      "loss": 2.8165,
      "theoretical_loss": 3.3993386632616676,
      "tokens_seen": 2233139200
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014337289578210706,
      "loss": 2.6785,
      "theoretical_loss": 3.399330524800601,
      "tokens_seen": 2233204736
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014335634575373557,
      "loss": 2.7359,
      "theoretical_loss": 3.3993223866452347,
      "tokens_seen": 2233270272
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001433397957253641,
      "loss": 2.7703,
      "theoretical_loss": 3.399314248795548,
      "tokens_seen": 2233335808
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014332324569699261,
      "loss": 2.8086,
      "theoretical_loss": 3.3993061112515206,
      "tokens_seen": 2233401344
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014330669566862115,
      "loss": 2.6661,
      "theoretical_loss": 3.3992979740131317,
      "tokens_seen": 2233466880
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014329014564024966,
      "loss": 2.8007,
      "theoretical_loss": 3.399289837080361,
      "tokens_seen": 2233532416
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001432735956118782,
      "loss": 2.0355,
      "theoretical_loss": 3.399281700453188,
      "tokens_seen": 2233597952
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001432570455835067,
      "loss": 2.546,
      "theoretical_loss": 3.399273564131593,
      "tokens_seen": 2233663488
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014324049555513525,
      "loss": 2.7611,
      "theoretical_loss": 3.3992654281155543,
      "tokens_seen": 2233729024
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014322394552676376,
      "loss": 2.3554,
      "theoretical_loss": 3.399257292405052,
      "tokens_seen": 2233794560
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001432073954983923,
      "loss": 2.7358,
      "theoretical_loss": 3.399249157000066,
      "tokens_seen": 2233860096
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001431908454700208,
      "loss": 2.8544,
      "theoretical_loss": 3.3992410219005755,
      "tokens_seen": 2233925632
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014317429544164934,
      "loss": 2.5013,
      "theoretical_loss": 3.3992328871065602,
      "tokens_seen": 2233991168
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014315774541327785,
      "loss": 2.5799,
      "theoretical_loss": 3.3992247526179997,
      "tokens_seen": 2234056704
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014314119538490636,
      "loss": 2.7791,
      "theoretical_loss": 3.3992166184348735,
      "tokens_seen": 2234122240
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001431246453565349,
      "loss": 2.7632,
      "theoretical_loss": 3.3992084845571613,
      "tokens_seen": 2234187776
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001431080953281634,
      "loss": 2.7185,
      "theoretical_loss": 3.399200350984842,
      "tokens_seen": 2234253312
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014309154529979195,
      "loss": 2.5002,
      "theoretical_loss": 3.3991922177178964,
      "tokens_seen": 2234318848
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014307499527142046,
      "loss": 2.764,
      "theoretical_loss": 3.399184084756303,
      "tokens_seen": 2234384384
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000143058445243049,
      "loss": 2.609,
      "theoretical_loss": 3.3991759521000415,
      "tokens_seen": 2234449920
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001430418952146775,
      "loss": 2.8082,
      "theoretical_loss": 3.399167819749092,
      "tokens_seen": 2234515456
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014302534518630605,
      "loss": 2.6412,
      "theoretical_loss": 3.3991596877034342,
      "tokens_seen": 2234580992
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014300879515793456,
      "loss": 2.4674,
      "theoretical_loss": 3.399151555963047,
      "tokens_seen": 2234646528
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001429922451295631,
      "loss": 2.5694,
      "theoretical_loss": 3.3991434245279106,
      "tokens_seen": 2234712064
    },
    {
      "epoch": 0.8,
      "objective/train/avg_token_score": 0.016377370804548264,
      "objective/train/avg_weight": 0.9807398319244385,
      "objective/train/docs_used": 1259222,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6094956398010254,
      "objective/train/std_weight": 0.0678112655878067,
      "objective/train/theoretical_loss": 3.399135293398004,
      "objective/train/tokens_used": 2255237600,
      "theoretical_loss": 3.399135293398004,
      "tokens_seen": 2234777600
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001429756951011916,
      "loss": 2.6486,
      "theoretical_loss": 3.399135293398004,
      "tokens_seen": 2234777600
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014295914507282014,
      "loss": 2.7478,
      "theoretical_loss": 3.3991271625733073,
      "tokens_seen": 2234843136
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014294259504444865,
      "loss": 2.6466,
      "theoretical_loss": 3.3991190320538,
      "tokens_seen": 2234908672
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001429260450160772,
      "loss": 2.6617,
      "theoretical_loss": 3.399110901839461,
      "tokens_seen": 2234974208
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014290949498770567,
      "loss": 2.6348,
      "theoretical_loss": 3.3991027719302713,
      "tokens_seen": 2235039744
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001428929449593342,
      "loss": 2.8466,
      "theoretical_loss": 3.399094642326209,
      "tokens_seen": 2235105280
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014287639493096272,
      "loss": 2.7811,
      "theoretical_loss": 3.399086513027255,
      "tokens_seen": 2235170816
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014285984490259126,
      "loss": 2.5353,
      "theoretical_loss": 3.399078384033388,
      "tokens_seen": 2235236352
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014284329487421977,
      "loss": 2.5599,
      "theoretical_loss": 3.3990702553445877,
      "tokens_seen": 2235301888
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001428267448458483,
      "loss": 2.697,
      "theoretical_loss": 3.399062126960834,
      "tokens_seen": 2235367424
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014281019481747682,
      "loss": 2.6381,
      "theoretical_loss": 3.3990539988821067,
      "tokens_seen": 2235432960
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014279364478910535,
      "loss": 2.4655,
      "theoretical_loss": 3.399045871108385,
      "tokens_seen": 2235498496
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014277709476073387,
      "loss": 2.5915,
      "theoretical_loss": 3.399037743639649,
      "tokens_seen": 2235564032
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001427605447323624,
      "loss": 2.4112,
      "theoretical_loss": 3.399029616475877,
      "tokens_seen": 2235629568
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001427439947039909,
      "loss": 2.5875,
      "theoretical_loss": 3.3990214896170503,
      "tokens_seen": 2235695104
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014272744467561945,
      "loss": 2.7211,
      "theoretical_loss": 3.3990133630631476,
      "tokens_seen": 2235760640
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014271089464724796,
      "loss": 2.5168,
      "theoretical_loss": 3.399005236814149,
      "tokens_seen": 2235826176
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001426943446188765,
      "loss": 2.7297,
      "theoretical_loss": 3.3989971108700336,
      "tokens_seen": 2235891712
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000142677794590505,
      "loss": 2.3259,
      "theoretical_loss": 3.3989889852307815,
      "tokens_seen": 2235957248
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014266124456213352,
      "loss": 2.6898,
      "theoretical_loss": 3.3989808598963718,
      "tokens_seen": 2236022784
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014264469453376206,
      "loss": 2.926,
      "theoretical_loss": 3.3989727348667844,
      "tokens_seen": 2236088320
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014262814450539057,
      "loss": 2.6035,
      "theoretical_loss": 3.398964610141999,
      "tokens_seen": 2236153856
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001426115944770191,
      "loss": 2.9354,
      "theoretical_loss": 3.3989564857219956,
      "tokens_seen": 2236219392
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014259504444864762,
      "loss": 2.5462,
      "theoretical_loss": 3.398948361606753,
      "tokens_seen": 2236284928
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014257849442027615,
      "loss": 2.5503,
      "theoretical_loss": 3.398940237796251,
      "tokens_seen": 2236350464
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014256194439190466,
      "loss": 2.6687,
      "theoretical_loss": 3.3989321142904703,
      "tokens_seen": 2236416000
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001425453943635332,
      "loss": 2.3959,
      "theoretical_loss": 3.398923991089389,
      "tokens_seen": 2236481536
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001425288443351617,
      "loss": 2.5435,
      "theoretical_loss": 3.3989158681929883,
      "tokens_seen": 2236547072
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014251229430679025,
      "loss": 2.4836,
      "theoretical_loss": 3.3989077456012464,
      "tokens_seen": 2236612608
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014249574427841876,
      "loss": 2.4669,
      "theoretical_loss": 3.3988996233141435,
      "tokens_seen": 2236678144
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001424791942500473,
      "loss": 2.6057,
      "theoretical_loss": 3.3988915013316596,
      "tokens_seen": 2236743680
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001424626442216758,
      "loss": 2.6629,
      "theoretical_loss": 3.398883379653774,
      "tokens_seen": 2236809216
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014244609419330435,
      "loss": 2.6067,
      "theoretical_loss": 3.3988752582804667,
      "tokens_seen": 2236874752
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014242954416493286,
      "loss": 2.5139,
      "theoretical_loss": 3.3988671372117167,
      "tokens_seen": 2236940288
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014241299413656137,
      "loss": 2.7031,
      "theoretical_loss": 3.398859016447504,
      "tokens_seen": 2237005824
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001423964441081899,
      "loss": 2.5226,
      "theoretical_loss": 3.3988508959878088,
      "tokens_seen": 2237071360
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014237989407981841,
      "loss": 2.4677,
      "theoretical_loss": 3.39884277583261,
      "tokens_seen": 2237136896
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014236334405144695,
      "loss": 2.7474,
      "theoretical_loss": 3.3988346559818874,
      "tokens_seen": 2237202432
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014234679402307546,
      "loss": 2.8362,
      "theoretical_loss": 3.3988265364356205,
      "tokens_seen": 2237267968
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000142330243994704,
      "loss": 2.542,
      "theoretical_loss": 3.3988184171937896,
      "tokens_seen": 2237333504
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001423136939663325,
      "loss": 2.3491,
      "theoretical_loss": 3.3988102982563744,
      "tokens_seen": 2237399040
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014229714393796105,
      "loss": 2.6336,
      "theoretical_loss": 3.3988021796233534,
      "tokens_seen": 2237464576
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014228059390958956,
      "loss": 2.7184,
      "theoretical_loss": 3.3987940612947076,
      "tokens_seen": 2237530112
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001422640438812181,
      "loss": 2.8636,
      "theoretical_loss": 3.398785943270416,
      "tokens_seen": 2237595648
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001422474938528466,
      "loss": 2.8845,
      "theoretical_loss": 3.3987778255504586,
      "tokens_seen": 2237661184
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014223094382447514,
      "loss": 2.6965,
      "theoretical_loss": 3.3987697081348145,
      "tokens_seen": 2237726720
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014221439379610365,
      "loss": 2.338,
      "theoretical_loss": 3.3987615910234643,
      "tokens_seen": 2237792256
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001421978437677322,
      "loss": 2.459,
      "theoretical_loss": 3.3987534742163867,
      "tokens_seen": 2237857792
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014218129373936067,
      "loss": 2.555,
      "theoretical_loss": 3.398745357713562,
      "tokens_seen": 2237923328
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001421647437109892,
      "loss": 2.4759,
      "theoretical_loss": 3.39873724151497,
      "tokens_seen": 2237988864
    },
    {
      "epoch": 0.8,
      "objective/train/avg_token_score": 0.01623653620481491,
      "objective/train/avg_weight": 0.9808894991874695,
      "objective/train/docs_used": 1261230,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9781970977783203,
      "objective/train/std_weight": 0.09056559950113297,
      "objective/train/theoretical_loss": 3.39872912562059,
      "objective/train/tokens_used": 2258514400,
      "theoretical_loss": 3.39872912562059,
      "tokens_seen": 2238054400
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014214819368261772,
      "loss": 2.5993,
      "theoretical_loss": 3.39872912562059,
      "tokens_seen": 2238054400
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014213164365424626,
      "loss": 2.4125,
      "theoretical_loss": 3.3987210100304015,
      "tokens_seen": 2238119936
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014211509362587477,
      "loss": 2.5907,
      "theoretical_loss": 3.3987128947443845,
      "tokens_seen": 2238185472
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001420985435975033,
      "loss": 2.776,
      "theoretical_loss": 3.398704779762519,
      "tokens_seen": 2238251008
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014208199356913182,
      "loss": 2.6176,
      "theoretical_loss": 3.398696665084784,
      "tokens_seen": 2238316544
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014206544354076036,
      "loss": 2.2774,
      "theoretical_loss": 3.39868855071116,
      "tokens_seen": 2238382080
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014204889351238887,
      "loss": 2.462,
      "theoretical_loss": 3.398680436641626,
      "tokens_seen": 2238447616
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001420323434840174,
      "loss": 2.6409,
      "theoretical_loss": 3.398672322876162,
      "tokens_seen": 2238513152
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014201579345564591,
      "loss": 2.6904,
      "theoretical_loss": 3.398664209414748,
      "tokens_seen": 2238578688
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014199924342727445,
      "loss": 2.6849,
      "theoretical_loss": 3.398656096257363,
      "tokens_seen": 2238644224
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014198269339890296,
      "loss": 2.7792,
      "theoretical_loss": 3.398647983403987,
      "tokens_seen": 2238709760
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001419661433705315,
      "loss": 2.6602,
      "theoretical_loss": 3.3986398708546,
      "tokens_seen": 2238775296
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014194959334216,
      "loss": 2.5811,
      "theoretical_loss": 3.3986317586091817,
      "tokens_seen": 2238840832
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014193304331378852,
      "loss": 2.5925,
      "theoretical_loss": 3.398623646667712,
      "tokens_seen": 2238906368
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014191649328541706,
      "loss": 2.6334,
      "theoretical_loss": 3.3986155350301694,
      "tokens_seen": 2238971904
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014189994325704557,
      "loss": 2.5861,
      "theoretical_loss": 3.398607423696535,
      "tokens_seen": 2239037440
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001418833932286741,
      "loss": 2.5024,
      "theoretical_loss": 3.3985993126667875,
      "tokens_seen": 2239102976
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014186684320030262,
      "loss": 2.9257,
      "theoretical_loss": 3.3985912019409077,
      "tokens_seen": 2239168512
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014185029317193115,
      "loss": 2.8237,
      "theoretical_loss": 3.3985830915188746,
      "tokens_seen": 2239234048
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014183374314355966,
      "loss": 2.635,
      "theoretical_loss": 3.398574981400668,
      "tokens_seen": 2239299584
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001418171931151882,
      "loss": 2.7689,
      "theoretical_loss": 3.3985668715862674,
      "tokens_seen": 2239365120
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001418006430868167,
      "loss": 2.3046,
      "theoretical_loss": 3.398558762075653,
      "tokens_seen": 2239430656
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014178409305844525,
      "loss": 2.5357,
      "theoretical_loss": 3.3985506528688045,
      "tokens_seen": 2239496192
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014176754303007376,
      "loss": 2.5827,
      "theoretical_loss": 3.3985425439657013,
      "tokens_seen": 2239561728
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001417509930017023,
      "loss": 2.4855,
      "theoretical_loss": 3.398534435366323,
      "tokens_seen": 2239627264
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001417344429733308,
      "loss": 2.7138,
      "theoretical_loss": 3.3985263270706505,
      "tokens_seen": 2239692800
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014171789294495935,
      "loss": 2.4738,
      "theoretical_loss": 3.398518219078662,
      "tokens_seen": 2239758336
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014170134291658786,
      "loss": 2.4477,
      "theoretical_loss": 3.398510111390338,
      "tokens_seen": 2239823872
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014168479288821637,
      "loss": 2.6503,
      "theoretical_loss": 3.3985020040056586,
      "tokens_seen": 2239889408
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001416682428598449,
      "loss": 2.5899,
      "theoretical_loss": 3.3984938969246024,
      "tokens_seen": 2239954944
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014165169283147342,
      "loss": 2.6322,
      "theoretical_loss": 3.3984857901471504,
      "tokens_seen": 2240020480
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014163514280310195,
      "loss": 2.5895,
      "theoretical_loss": 3.3984776836732817,
      "tokens_seen": 2240086016
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014161859277473046,
      "loss": 2.582,
      "theoretical_loss": 3.398469577502976,
      "tokens_seen": 2240151552
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000141602042746359,
      "loss": 2.5991,
      "theoretical_loss": 3.3984614716362134,
      "tokens_seen": 2240217088
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001415854927179875,
      "loss": 2.545,
      "theoretical_loss": 3.398453366072973,
      "tokens_seen": 2240282624
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014156894268961605,
      "loss": 2.7098,
      "theoretical_loss": 3.3984452608132356,
      "tokens_seen": 2240348160
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014155239266124456,
      "loss": 2.4683,
      "theoretical_loss": 3.39843715585698,
      "tokens_seen": 2240413696
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001415358426328731,
      "loss": 2.5996,
      "theoretical_loss": 3.398429051204187,
      "tokens_seen": 2240479232
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001415192926045016,
      "loss": 2.7142,
      "theoretical_loss": 3.398420946854835,
      "tokens_seen": 2240544768
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014150274257613014,
      "loss": 2.8405,
      "theoretical_loss": 3.398412842808905,
      "tokens_seen": 2240610304
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014148619254775866,
      "loss": 2.7285,
      "theoretical_loss": 3.3984047390663754,
      "tokens_seen": 2240675840
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001414696425193872,
      "loss": 2.502,
      "theoretical_loss": 3.3983966356272277,
      "tokens_seen": 2240741376
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001414530924910157,
      "loss": 2.5817,
      "theoretical_loss": 3.39838853249144,
      "tokens_seen": 2240806912
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014143654246264421,
      "loss": 2.7627,
      "theoretical_loss": 3.3983804296589932,
      "tokens_seen": 2240872448
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014141999243427272,
      "loss": 2.6715,
      "theoretical_loss": 3.398372327129867,
      "tokens_seen": 2240937984
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014140344240590126,
      "loss": 2.8583,
      "theoretical_loss": 3.3983642249040407,
      "tokens_seen": 2241003520
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014138689237752977,
      "loss": 2.5338,
      "theoretical_loss": 3.398356122981494,
      "tokens_seen": 2241069056
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001413703423491583,
      "loss": 2.5199,
      "theoretical_loss": 3.398348021362207,
      "tokens_seen": 2241134592
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014135379232078682,
      "loss": 2.6365,
      "theoretical_loss": 3.3983399200461597,
      "tokens_seen": 2241200128
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014133724229241536,
      "loss": 2.6677,
      "theoretical_loss": 3.3983318190333316,
      "tokens_seen": 2241265664
    },
    {
      "epoch": 0.8,
      "objective/train/avg_token_score": 0.01015857420861721,
      "objective/train/avg_weight": 0.9880474805831909,
      "objective/train/docs_used": 1263119,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.115053653717041,
      "objective/train/std_weight": 0.05022725090384483,
      "objective/train/theoretical_loss": 3.3983237183237023,
      "objective/train/tokens_used": 2261791200,
      "theoretical_loss": 3.3983237183237023,
      "tokens_seen": 2241331200
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014132069226404387,
      "loss": 2.5941,
      "theoretical_loss": 3.3983237183237023,
      "tokens_seen": 2241331200
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001413041422356724,
      "loss": 2.4591,
      "theoretical_loss": 3.398315617917252,
      "tokens_seen": 2241396736
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014128759220730092,
      "loss": 2.747,
      "theoretical_loss": 3.3983075178139597,
      "tokens_seen": 2241462272
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014127104217892945,
      "loss": 2.7025,
      "theoretical_loss": 3.3982994180138064,
      "tokens_seen": 2241527808
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014125449215055796,
      "loss": 2.5428,
      "theoretical_loss": 3.3982913185167707,
      "tokens_seen": 2241593344
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001412379421221865,
      "loss": 2.6524,
      "theoretical_loss": 3.3982832193228334,
      "tokens_seen": 2241658880
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000141221392093815,
      "loss": 2.3571,
      "theoretical_loss": 3.3982751204319737,
      "tokens_seen": 2241724416
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014120484206544355,
      "loss": 2.5231,
      "theoretical_loss": 3.3982670218441715,
      "tokens_seen": 2241789952
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014118829203707206,
      "loss": 2.5178,
      "theoretical_loss": 3.398258923559407,
      "tokens_seen": 2241855488
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014117174200870057,
      "loss": 2.7034,
      "theoretical_loss": 3.398250825577659,
      "tokens_seen": 2241921024
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001411551919803291,
      "loss": 2.4901,
      "theoretical_loss": 3.398242727898908,
      "tokens_seen": 2241986560
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014113864195195762,
      "loss": 2.697,
      "theoretical_loss": 3.3982346305231346,
      "tokens_seen": 2242052096
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014112209192358616,
      "loss": 2.6786,
      "theoretical_loss": 3.398226533450317,
      "tokens_seen": 2242117632
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014110554189521467,
      "loss": 2.5743,
      "theoretical_loss": 3.3982184366804358,
      "tokens_seen": 2242183168
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001410889918668432,
      "loss": 2.6418,
      "theoretical_loss": 3.3982103402134713,
      "tokens_seen": 2242248704
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014107244183847171,
      "loss": 2.5934,
      "theoretical_loss": 3.3982022440494024,
      "tokens_seen": 2242314240
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014105589181010025,
      "loss": 2.5618,
      "theoretical_loss": 3.3981941481882094,
      "tokens_seen": 2242379776
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014103934178172876,
      "loss": 2.2768,
      "theoretical_loss": 3.398186052629872,
      "tokens_seen": 2242445312
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001410227917533573,
      "loss": 2.6287,
      "theoretical_loss": 3.39817795737437,
      "tokens_seen": 2242510848
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001410062417249858,
      "loss": 2.5909,
      "theoretical_loss": 3.3981698624216836,
      "tokens_seen": 2242576384
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014098969169661435,
      "loss": 2.4636,
      "theoretical_loss": 3.398161767771792,
      "tokens_seen": 2242641920
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014097314166824286,
      "loss": 2.6226,
      "theoretical_loss": 3.3981536734246753,
      "tokens_seen": 2242707456
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014095659163987137,
      "loss": 2.5924,
      "theoretical_loss": 3.3981455793803135,
      "tokens_seen": 2242772992
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001409400416114999,
      "loss": 2.6834,
      "theoretical_loss": 3.3981374856386863,
      "tokens_seen": 2242838528
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014092349158312842,
      "loss": 2.5402,
      "theoretical_loss": 3.3981293921997735,
      "tokens_seen": 2242904064
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014090694155475695,
      "loss": 2.5539,
      "theoretical_loss": 3.398121299063555,
      "tokens_seen": 2242969600
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014089039152638546,
      "loss": 2.5566,
      "theoretical_loss": 3.3981132062300103,
      "tokens_seen": 2243035136
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000140873841498014,
      "loss": 2.4578,
      "theoretical_loss": 3.3981051136991196,
      "tokens_seen": 2243100672
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001408572914696425,
      "loss": 2.6144,
      "theoretical_loss": 3.398097021470863,
      "tokens_seen": 2243166208
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014084074144127105,
      "loss": 2.6026,
      "theoretical_loss": 3.39808892954522,
      "tokens_seen": 2243231744
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014082419141289956,
      "loss": 2.5785,
      "theoretical_loss": 3.39808083792217,
      "tokens_seen": 2243297280
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001408076413845281,
      "loss": 2.6817,
      "theoretical_loss": 3.3980727466016933,
      "tokens_seen": 2243362816
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001407910913561566,
      "loss": 2.5688,
      "theoretical_loss": 3.39806465558377,
      "tokens_seen": 2243428352
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014077454132778515,
      "loss": 2.7466,
      "theoretical_loss": 3.39805656486838,
      "tokens_seen": 2243493888
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014075799129941366,
      "loss": 2.7387,
      "theoretical_loss": 3.3980484744555026,
      "tokens_seen": 2243559424
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001407414412710422,
      "loss": 2.6749,
      "theoretical_loss": 3.3980403843451175,
      "tokens_seen": 2243624960
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001407248912426707,
      "loss": 2.5881,
      "theoretical_loss": 3.3980322945372055,
      "tokens_seen": 2243690496
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014070834121429922,
      "loss": 2.4145,
      "theoretical_loss": 3.3980242050317457,
      "tokens_seen": 2243756032
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014069179118592773,
      "loss": 2.5942,
      "theoretical_loss": 3.3980161158287183,
      "tokens_seen": 2243821568
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014067524115755626,
      "loss": 2.6405,
      "theoretical_loss": 3.3980080269281028,
      "tokens_seen": 2243887104
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014065869112918477,
      "loss": 2.4604,
      "theoretical_loss": 3.397999938329879,
      "tokens_seen": 2243952640
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001406421411008133,
      "loss": 2.6868,
      "theoretical_loss": 3.3979918500340274,
      "tokens_seen": 2244018176
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014062559107244182,
      "loss": 2.6837,
      "theoretical_loss": 3.3979837620405275,
      "tokens_seen": 2244083712
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014060904104407036,
      "loss": 2.7329,
      "theoretical_loss": 3.397975674349359,
      "tokens_seen": 2244149248
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014059249101569887,
      "loss": 2.4567,
      "theoretical_loss": 3.3979675869605024,
      "tokens_seen": 2244214784
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001405759409873274,
      "loss": 2.2705,
      "theoretical_loss": 3.397959499873937,
      "tokens_seen": 2244280320
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014055939095895592,
      "loss": 2.8526,
      "theoretical_loss": 3.3979514130896424,
      "tokens_seen": 2244345856
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014054284093058446,
      "loss": 2.664,
      "theoretical_loss": 3.3979433266075993,
      "tokens_seen": 2244411392
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014052629090221297,
      "loss": 2.4309,
      "theoretical_loss": 3.3979352404277874,
      "tokens_seen": 2244476928
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001405097408738415,
      "loss": 2.724,
      "theoretical_loss": 3.3979271545501857,
      "tokens_seen": 2244542464
    },
    {
      "epoch": 0.8,
      "objective/train/avg_token_score": 0.013959571719169617,
      "objective/train/avg_weight": 0.9835718274116516,
      "objective/train/docs_used": 1265240,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6149191856384277,
      "objective/train/std_weight": 0.06909426301717758,
      "objective/train/theoretical_loss": 3.397919068974775,
      "objective/train/tokens_used": 2265068000,
      "theoretical_loss": 3.397919068974775,
      "tokens_seen": 2244608000
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014049319084547004,
      "loss": 2.362,
      "theoretical_loss": 3.397919068974775,
      "tokens_seen": 2244608000
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014047664081709855,
      "loss": 2.6429,
      "theoretical_loss": 3.397910983701535,
      "tokens_seen": 2244673536
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014046009078872706,
      "loss": 2.9049,
      "theoretical_loss": 3.3979028987304454,
      "tokens_seen": 2244739072
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014044354076035557,
      "loss": 2.2674,
      "theoretical_loss": 3.397894814061486,
      "tokens_seen": 2244804608
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001404269907319841,
      "loss": 2.4542,
      "theoretical_loss": 3.397886729694637,
      "tokens_seen": 2244870144
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014041044070361262,
      "loss": 2.5105,
      "theoretical_loss": 3.397878645629878,
      "tokens_seen": 2244935680
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014039389067524116,
      "loss": 2.6263,
      "theoretical_loss": 3.3978705618671894,
      "tokens_seen": 2245001216
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014037734064686967,
      "loss": 2.3008,
      "theoretical_loss": 3.3978624784065508,
      "tokens_seen": 2245066752
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001403607906184982,
      "loss": 2.836,
      "theoretical_loss": 3.3978543952479416,
      "tokens_seen": 2245132288
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014034424059012672,
      "loss": 2.5953,
      "theoretical_loss": 3.3978463123913425,
      "tokens_seen": 2245197824
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014032769056175525,
      "loss": 2.661,
      "theoretical_loss": 3.397838229836733,
      "tokens_seen": 2245263360
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014031114053338376,
      "loss": 2.4876,
      "theoretical_loss": 3.397830147584093,
      "tokens_seen": 2245328896
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001402945905050123,
      "loss": 2.5694,
      "theoretical_loss": 3.3978220656334024,
      "tokens_seen": 2245394432
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001402780404766408,
      "loss": 2.3617,
      "theoretical_loss": 3.3978139839846406,
      "tokens_seen": 2245459968
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014026149044826935,
      "loss": 2.5738,
      "theoretical_loss": 3.397805902637789,
      "tokens_seen": 2245525504
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014024494041989786,
      "loss": 2.393,
      "theoretical_loss": 3.397797821592826,
      "tokens_seen": 2245591040
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001402283903915264,
      "loss": 2.3735,
      "theoretical_loss": 3.397789740849732,
      "tokens_seen": 2245656576
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001402118403631549,
      "loss": 2.4787,
      "theoretical_loss": 3.3977816604084876,
      "tokens_seen": 2245722112
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014019529033478342,
      "loss": 2.7366,
      "theoretical_loss": 3.397773580269072,
      "tokens_seen": 2245787648
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014017874030641196,
      "loss": 2.7119,
      "theoretical_loss": 3.397765500431465,
      "tokens_seen": 2245853184
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014016219027804047,
      "loss": 2.3216,
      "theoretical_loss": 3.3977574208956467,
      "tokens_seen": 2245918720
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000140145640249669,
      "loss": 2.5255,
      "theoretical_loss": 3.3977493416615974,
      "tokens_seen": 2245984256
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014012909022129751,
      "loss": 2.6029,
      "theoretical_loss": 3.3977412627292964,
      "tokens_seen": 2246049792
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014011254019292605,
      "loss": 2.736,
      "theoretical_loss": 3.397733184098724,
      "tokens_seen": 2246115328
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014009599016455456,
      "loss": 2.7259,
      "theoretical_loss": 3.39772510576986,
      "tokens_seen": 2246180864
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001400794401361831,
      "loss": 2.6064,
      "theoretical_loss": 3.3977170277426847,
      "tokens_seen": 2246246400
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001400628901078116,
      "loss": 2.4085,
      "theoretical_loss": 3.3977089500171775,
      "tokens_seen": 2246311936
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014004634007944015,
      "loss": 2.5818,
      "theoretical_loss": 3.3977008725933184,
      "tokens_seen": 2246377472
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00014002979005106866,
      "loss": 2.6257,
      "theoretical_loss": 3.3976927954710874,
      "tokens_seen": 2246443008
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001400132400226972,
      "loss": 2.4207,
      "theoretical_loss": 3.397684718650465,
      "tokens_seen": 2246508544
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001399966899943257,
      "loss": 2.4325,
      "theoretical_loss": 3.3976766421314304,
      "tokens_seen": 2246574080
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013998013996595424,
      "loss": 2.3035,
      "theoretical_loss": 3.397668565913964,
      "tokens_seen": 2246639616
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013996358993758273,
      "loss": 2.3224,
      "theoretical_loss": 3.3976604899980454,
      "tokens_seen": 2246705152
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013994703990921126,
      "loss": 2.5529,
      "theoretical_loss": 3.3976524143836544,
      "tokens_seen": 2246770688
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013993048988083977,
      "loss": 2.4307,
      "theoretical_loss": 3.3976443390707716,
      "tokens_seen": 2246836224
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001399139398524683,
      "loss": 2.4445,
      "theoretical_loss": 3.3976362640593765,
      "tokens_seen": 2246901760
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013989738982409682,
      "loss": 2.6122,
      "theoretical_loss": 3.397628189349449,
      "tokens_seen": 2246967296
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013988083979572536,
      "loss": 2.4164,
      "theoretical_loss": 3.3976201149409695,
      "tokens_seen": 2247032832
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001398642897673539,
      "loss": 2.7221,
      "theoretical_loss": 3.3976120408339177,
      "tokens_seen": 2247098368
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001398477397389824,
      "loss": 2.5062,
      "theoretical_loss": 3.3976039670282736,
      "tokens_seen": 2247163904
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013983118971061095,
      "loss": 2.5169,
      "theoretical_loss": 3.3975958935240165,
      "tokens_seen": 2247229440
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013981463968223946,
      "loss": 2.1888,
      "theoretical_loss": 3.397587820321127,
      "tokens_seen": 2247294976
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000139798089653868,
      "loss": 2.6668,
      "theoretical_loss": 3.3975797474195857,
      "tokens_seen": 2247360512
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001397815396254965,
      "loss": 2.5885,
      "theoretical_loss": 3.397571674819371,
      "tokens_seen": 2247426048
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013976498959712504,
      "loss": 2.6197,
      "theoretical_loss": 3.3975636025204645,
      "tokens_seen": 2247491584
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013974843956875355,
      "loss": 2.4887,
      "theoretical_loss": 3.397555530522845,
      "tokens_seen": 2247557120
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001397318895403821,
      "loss": 2.4742,
      "theoretical_loss": 3.3975474588264927,
      "tokens_seen": 2247622656
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013971533951201057,
      "loss": 2.6882,
      "theoretical_loss": 3.397539387431388,
      "tokens_seen": 2247688192
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001396987894836391,
      "loss": 2.597,
      "theoretical_loss": 3.3975313163375107,
      "tokens_seen": 2247753728
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013968223945526762,
      "loss": 2.3631,
      "theoretical_loss": 3.39752324554484,
      "tokens_seen": 2247819264
    },
    {
      "epoch": 0.8,
      "objective/train/avg_token_score": 0.026425013318657875,
      "objective/train/avg_weight": 0.9688868522644043,
      "objective/train/docs_used": 1267114,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 6.771032333374023,
      "objective/train/std_weight": 0.1507129818201065,
      "objective/train/theoretical_loss": 3.3975151750533574,
      "objective/train/tokens_used": 2268344800,
      "theoretical_loss": 3.3975151750533574,
      "tokens_seen": 2247884800
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013966568942689616,
      "loss": 2.4365,
      "theoretical_loss": 3.3975151750533574,
      "tokens_seen": 2247884800
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013964913939852467,
      "loss": 2.5789,
      "theoretical_loss": 3.397507104863042,
      "tokens_seen": 2247950336
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001396325893701532,
      "loss": 2.4715,
      "theoretical_loss": 3.397499034973873,
      "tokens_seen": 2248015872
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013961603934178172,
      "loss": 2.4195,
      "theoretical_loss": 3.397490965385832,
      "tokens_seen": 2248081408
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013959948931341025,
      "loss": 2.6031,
      "theoretical_loss": 3.397482896098898,
      "tokens_seen": 2248146944
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013958293928503877,
      "loss": 2.2915,
      "theoretical_loss": 3.397474827113051,
      "tokens_seen": 2248212480
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001395663892566673,
      "loss": 2.3488,
      "theoretical_loss": 3.397466758428271,
      "tokens_seen": 2248278016
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001395498392282958,
      "loss": 2.5127,
      "theoretical_loss": 3.3974586900445383,
      "tokens_seen": 2248343552
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013953328919992435,
      "loss": 2.1604,
      "theoretical_loss": 3.3974506219618332,
      "tokens_seen": 2248409088
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013951673917155286,
      "loss": 2.3584,
      "theoretical_loss": 3.3974425541801345,
      "tokens_seen": 2248474624
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001395001891431814,
      "loss": 1.9674,
      "theoretical_loss": 3.3974344866994235,
      "tokens_seen": 2248540160
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001394836391148099,
      "loss": 2.6804,
      "theoretical_loss": 3.3974264195196797,
      "tokens_seen": 2248605696
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013946708908643842,
      "loss": 2.5249,
      "theoretical_loss": 3.3974183526408828,
      "tokens_seen": 2248671232
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013945053905806696,
      "loss": 2.9584,
      "theoretical_loss": 3.397410286063013,
      "tokens_seen": 2248736768
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013943398902969547,
      "loss": 2.3491,
      "theoretical_loss": 3.3974022197860503,
      "tokens_seen": 2248802304
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000139417439001324,
      "loss": 2.7055,
      "theoretical_loss": 3.3973941538099752,
      "tokens_seen": 2248867840
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013940088897295252,
      "loss": 2.4513,
      "theoretical_loss": 3.397386088134767,
      "tokens_seen": 2248933376
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013938433894458105,
      "loss": 2.417,
      "theoretical_loss": 3.397378022760406,
      "tokens_seen": 2248998912
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013936778891620956,
      "loss": 2.7216,
      "theoretical_loss": 3.397369957686872,
      "tokens_seen": 2249064448
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001393512388878381,
      "loss": 2.4308,
      "theoretical_loss": 3.3973618929141454,
      "tokens_seen": 2249129984
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001393346888594666,
      "loss": 2.4868,
      "theoretical_loss": 3.397353828442206,
      "tokens_seen": 2249195520
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013931813883109515,
      "loss": 2.345,
      "theoretical_loss": 3.3973457642710336,
      "tokens_seen": 2249261056
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013930158880272366,
      "loss": 2.5435,
      "theoretical_loss": 3.397337700400609,
      "tokens_seen": 2249326592
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001392850387743522,
      "loss": 2.5469,
      "theoretical_loss": 3.397329636830911,
      "tokens_seen": 2249392128
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001392684887459807,
      "loss": 2.7015,
      "theoretical_loss": 3.397321573561921,
      "tokens_seen": 2249457664
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013925193871760925,
      "loss": 2.4766,
      "theoretical_loss": 3.3973135105936176,
      "tokens_seen": 2249523200
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013923538868923776,
      "loss": 2.4181,
      "theoretical_loss": 3.397305447925982,
      "tokens_seen": 2249588736
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013921883866086627,
      "loss": 2.486,
      "theoretical_loss": 3.3972973855589936,
      "tokens_seen": 2249654272
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001392022886324948,
      "loss": 2.5109,
      "theoretical_loss": 3.397289323492633,
      "tokens_seen": 2249719808
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013918573860412331,
      "loss": 2.5528,
      "theoretical_loss": 3.3972812617268797,
      "tokens_seen": 2249785344
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013916918857575185,
      "loss": 2.4958,
      "theoretical_loss": 3.397273200261713,
      "tokens_seen": 2249850880
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013915263854738036,
      "loss": 2.7212,
      "theoretical_loss": 3.397265139097115,
      "tokens_seen": 2249916416
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001391360885190089,
      "loss": 2.6425,
      "theoretical_loss": 3.397257078233064,
      "tokens_seen": 2249981952
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001391195384906374,
      "loss": 2.7002,
      "theoretical_loss": 3.397249017669541,
      "tokens_seen": 2250047488
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013910298846226595,
      "loss": 2.4719,
      "theoretical_loss": 3.397240957406525,
      "tokens_seen": 2250113024
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013908643843389446,
      "loss": 2.482,
      "theoretical_loss": 3.397232897443997,
      "tokens_seen": 2250178560
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000139069888405523,
      "loss": 2.0476,
      "theoretical_loss": 3.397224837781937,
      "tokens_seen": 2250244096
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001390533383771515,
      "loss": 2.5778,
      "theoretical_loss": 3.397216778420325,
      "tokens_seen": 2250309632
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013903678834878004,
      "loss": 2.5282,
      "theoretical_loss": 3.3972087193591403,
      "tokens_seen": 2250375168
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013902023832040855,
      "loss": 2.7754,
      "theoretical_loss": 3.3972006605983633,
      "tokens_seen": 2250440704
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001390036882920371,
      "loss": 2.5992,
      "theoretical_loss": 3.397192602137975,
      "tokens_seen": 2250506240
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013898713826366557,
      "loss": 2.692,
      "theoretical_loss": 3.397184543977954,
      "tokens_seen": 2250571776
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001389705882352941,
      "loss": 2.3758,
      "theoretical_loss": 3.3971764861182816,
      "tokens_seen": 2250637312
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013895403820692262,
      "loss": 2.5897,
      "theoretical_loss": 3.397168428558937,
      "tokens_seen": 2250702848
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013893748817855116,
      "loss": 2.5361,
      "theoretical_loss": 3.3971603712999006,
      "tokens_seen": 2250768384
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013892093815017967,
      "loss": 2.682,
      "theoretical_loss": 3.3971523143411524,
      "tokens_seen": 2250833920
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001389043881218082,
      "loss": 2.6829,
      "theoretical_loss": 3.3971442576826725,
      "tokens_seen": 2250899456
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013888783809343672,
      "loss": 2.3264,
      "theoretical_loss": 3.3971362013244413,
      "tokens_seen": 2250964992
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013887128806506526,
      "loss": 2.5594,
      "theoretical_loss": 3.3971281452664384,
      "tokens_seen": 2251030528
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013885473803669377,
      "loss": 2.6558,
      "theoretical_loss": 3.3971200895086437,
      "tokens_seen": 2251096064
    },
    {
      "epoch": 0.8,
      "objective/train/avg_token_score": 0.017461881041526794,
      "objective/train/avg_weight": 0.9795145988464355,
      "objective/train/docs_used": 1269266,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.9236905574798584,
      "objective/train/std_weight": 0.10659237205982208,
      "objective/train/theoretical_loss": 3.3971120340510375,
      "objective/train/tokens_used": 2271621600,
      "theoretical_loss": 3.3971120340510375,
      "tokens_seen": 2251161600
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001388381880083223,
      "loss": 2.6589,
      "theoretical_loss": 3.3971120340510375,
      "tokens_seen": 2251161600
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013882163797995081,
      "loss": 2.5788,
      "theoretical_loss": 3.3971039788936004,
      "tokens_seen": 2251227136
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013880508795157935,
      "loss": 2.7555,
      "theoretical_loss": 3.3970959240363117,
      "tokens_seen": 2251292672
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013878853792320786,
      "loss": 2.6464,
      "theoretical_loss": 3.397087869479152,
      "tokens_seen": 2251358208
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001387719878948364,
      "loss": 2.2737,
      "theoretical_loss": 3.3970798152221016,
      "tokens_seen": 2251423744
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001387554378664649,
      "loss": 2.5673,
      "theoretical_loss": 3.3970717612651398,
      "tokens_seen": 2251489280
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013873888783809342,
      "loss": 2.2719,
      "theoretical_loss": 3.397063707608247,
      "tokens_seen": 2251554816
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013872233780972196,
      "loss": 2.4775,
      "theoretical_loss": 3.397055654251403,
      "tokens_seen": 2251620352
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013870578778135047,
      "loss": 2.6896,
      "theoretical_loss": 3.397047601194589,
      "tokens_seen": 2251685888
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000138689237752979,
      "loss": 2.6289,
      "theoretical_loss": 3.397039548437784,
      "tokens_seen": 2251751424
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013867268772460752,
      "loss": 2.5677,
      "theoretical_loss": 3.3970314959809684,
      "tokens_seen": 2251816960
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013865613769623605,
      "loss": 2.4847,
      "theoretical_loss": 3.3970234438241222,
      "tokens_seen": 2251882496
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013863958766786456,
      "loss": 2.4584,
      "theoretical_loss": 3.397015391967226,
      "tokens_seen": 2251948032
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001386230376394931,
      "loss": 2.7034,
      "theoretical_loss": 3.3970073404102594,
      "tokens_seen": 2252013568
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001386064876111216,
      "loss": 2.5088,
      "theoretical_loss": 3.3969992891532024,
      "tokens_seen": 2252079104
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013858993758275015,
      "loss": 2.626,
      "theoretical_loss": 3.3969912381960357,
      "tokens_seen": 2252144640
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013857338755437866,
      "loss": 2.1783,
      "theoretical_loss": 3.396983187538739,
      "tokens_seen": 2252210176
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001385568375260072,
      "loss": 2.0426,
      "theoretical_loss": 3.3969751371812915,
      "tokens_seen": 2252275712
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001385402874976357,
      "loss": 2.6172,
      "theoretical_loss": 3.396967087123675,
      "tokens_seen": 2252341248
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013852373746926425,
      "loss": 2.5644,
      "theoretical_loss": 3.396959037365869,
      "tokens_seen": 2252406784
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013850718744089276,
      "loss": 2.4795,
      "theoretical_loss": 3.3969509879078528,
      "tokens_seen": 2252472320
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013849063741252127,
      "loss": 2.7602,
      "theoretical_loss": 3.396942938749608,
      "tokens_seen": 2252537856
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001384740873841498,
      "loss": 2.7358,
      "theoretical_loss": 3.396934889891113,
      "tokens_seen": 2252603392
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013845753735577832,
      "loss": 2.6572,
      "theoretical_loss": 3.3969268413323497,
      "tokens_seen": 2252668928
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013844098732740685,
      "loss": 2.3782,
      "theoretical_loss": 3.3969187930732967,
      "tokens_seen": 2252734464
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013842443729903536,
      "loss": 2.4425,
      "theoretical_loss": 3.396910745113935,
      "tokens_seen": 2252800000
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001384078872706639,
      "loss": 2.587,
      "theoretical_loss": 3.3969026974542444,
      "tokens_seen": 2252865536
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001383913372422924,
      "loss": 2.6344,
      "theoretical_loss": 3.3968946500942048,
      "tokens_seen": 2252931072
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013837478721392095,
      "loss": 2.8352,
      "theoretical_loss": 3.396886603033797,
      "tokens_seen": 2252996608
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013835823718554946,
      "loss": 2.714,
      "theoretical_loss": 3.3968785562730006,
      "tokens_seen": 2253062144
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000138341687157178,
      "loss": 2.7398,
      "theoretical_loss": 3.396870509811796,
      "tokens_seen": 2253127680
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001383251371288065,
      "loss": 2.6556,
      "theoretical_loss": 3.396862463650163,
      "tokens_seen": 2253193216
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013830858710043504,
      "loss": 2.529,
      "theoretical_loss": 3.3968544177880817,
      "tokens_seen": 2253258752
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013829203707206356,
      "loss": 2.5036,
      "theoretical_loss": 3.396846372225533,
      "tokens_seen": 2253324288
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001382754870436921,
      "loss": 2.5708,
      "theoretical_loss": 3.3968383269624964,
      "tokens_seen": 2253389824
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001382589370153206,
      "loss": 2.5479,
      "theoretical_loss": 3.396830281998952,
      "tokens_seen": 2253455360
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013824238698694911,
      "loss": 2.388,
      "theoretical_loss": 3.3968222373348804,
      "tokens_seen": 2253520896
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013822583695857762,
      "loss": 2.4292,
      "theoretical_loss": 3.396814192970261,
      "tokens_seen": 2253586432
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013820928693020616,
      "loss": 2.1523,
      "theoretical_loss": 3.3968061489050747,
      "tokens_seen": 2253651968
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013819273690183467,
      "loss": 2.3725,
      "theoretical_loss": 3.3967981051393012,
      "tokens_seen": 2253717504
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001381761868734632,
      "loss": 2.5578,
      "theoretical_loss": 3.3967900616729207,
      "tokens_seen": 2253783040
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013815963684509172,
      "loss": 2.6612,
      "theoretical_loss": 3.3967820185059137,
      "tokens_seen": 2253848576
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00013814308681672026,
      "loss": 2.3376,
      "theoretical_loss": 3.3967739756382596,
      "tokens_seen": 2253914112
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013812653678834877,
      "loss": 2.8123,
      "theoretical_loss": 3.3967659330699393,
      "tokens_seen": 2253979648
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001381099867599773,
      "loss": 2.5366,
      "theoretical_loss": 3.396757890800933,
      "tokens_seen": 2254045184
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013809343673160582,
      "loss": 2.7257,
      "theoretical_loss": 3.39674984883122,
      "tokens_seen": 2254110720
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013807688670323435,
      "loss": 2.3888,
      "theoretical_loss": 3.3967418071607813,
      "tokens_seen": 2254176256
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013806033667486286,
      "loss": 2.4414,
      "theoretical_loss": 3.396733765789597,
      "tokens_seen": 2254241792
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001380437866464914,
      "loss": 2.1399,
      "theoretical_loss": 3.3967257247176468,
      "tokens_seen": 2254307328
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001380272366181199,
      "loss": 1.8534,
      "theoretical_loss": 3.396717683944911,
      "tokens_seen": 2254372864
    },
    {
      "epoch": 0.81,
      "objective/train/avg_token_score": 0.01351692620664835,
      "objective/train/avg_weight": 0.9840928316116333,
      "objective/train/docs_used": 1270683,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2500736713409424,
      "objective/train/std_weight": 0.07138732075691223,
      "objective/train/theoretical_loss": 3.39670964347137,
      "objective/train/tokens_used": 2274898400,
      "theoretical_loss": 3.39670964347137,
      "tokens_seen": 2254438400
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013801068658974845,
      "loss": 2.4502,
      "theoretical_loss": 3.39670964347137,
      "tokens_seen": 2254438400
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013799413656137696,
      "loss": 2.5938,
      "theoretical_loss": 3.3967016032970037,
      "tokens_seen": 2254503936
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013797758653300547,
      "loss": 2.4271,
      "theoretical_loss": 3.3966935634217927,
      "tokens_seen": 2254569472
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000137961036504634,
      "loss": 2.4643,
      "theoretical_loss": 3.3966855238457168,
      "tokens_seen": 2254635008
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013794448647626252,
      "loss": 2.4357,
      "theoretical_loss": 3.396677484568756,
      "tokens_seen": 2254700544
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013792793644789106,
      "loss": 2.7306,
      "theoretical_loss": 3.396669445590891,
      "tokens_seen": 2254766080
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013791138641951957,
      "loss": 2.796,
      "theoretical_loss": 3.3966614069121017,
      "tokens_seen": 2254831616
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001378948363911481,
      "loss": 2.5106,
      "theoretical_loss": 3.3966533685323683,
      "tokens_seen": 2254897152
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013787828636277661,
      "loss": 2.4807,
      "theoretical_loss": 3.3966453304516713,
      "tokens_seen": 2254962688
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013786173633440515,
      "loss": 2.6378,
      "theoretical_loss": 3.39663729266999,
      "tokens_seen": 2255028224
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013784518630603366,
      "loss": 2.7173,
      "theoretical_loss": 3.3966292551873054,
      "tokens_seen": 2255093760
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001378286362776622,
      "loss": 2.6138,
      "theoretical_loss": 3.3966212180035975,
      "tokens_seen": 2255159296
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001378120862492907,
      "loss": 2.5183,
      "theoretical_loss": 3.3966131811188465,
      "tokens_seen": 2255224832
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013779553622091925,
      "loss": 2.4873,
      "theoretical_loss": 3.3966051445330323,
      "tokens_seen": 2255290368
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013777898619254776,
      "loss": 2.813,
      "theoretical_loss": 3.3965971082461355,
      "tokens_seen": 2255355904
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013776243616417627,
      "loss": 2.4379,
      "theoretical_loss": 3.396589072258136,
      "tokens_seen": 2255421440
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001377458861358048,
      "loss": 2.7026,
      "theoretical_loss": 3.3965810365690143,
      "tokens_seen": 2255486976
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013772933610743332,
      "loss": 2.5207,
      "theoretical_loss": 3.3965730011787505,
      "tokens_seen": 2255552512
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013771278607906185,
      "loss": 2.6209,
      "theoretical_loss": 3.3965649660873245,
      "tokens_seen": 2255618048
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013769623605069036,
      "loss": 2.6866,
      "theoretical_loss": 3.3965569312947164,
      "tokens_seen": 2255683584
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001376796860223189,
      "loss": 2.5032,
      "theoretical_loss": 3.3965488968009074,
      "tokens_seen": 2255749120
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001376631359939474,
      "loss": 2.3538,
      "theoretical_loss": 3.3965408626058764,
      "tokens_seen": 2255814656
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013764658596557595,
      "loss": 2.6442,
      "theoretical_loss": 3.3965328287096046,
      "tokens_seen": 2255880192
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013763003593720446,
      "loss": 2.6637,
      "theoretical_loss": 3.396524795112072,
      "tokens_seen": 2255945728
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000137613485908833,
      "loss": 2.524,
      "theoretical_loss": 3.396516761813259,
      "tokens_seen": 2256011264
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001375969358804615,
      "loss": 2.5517,
      "theoretical_loss": 3.396508728813145,
      "tokens_seen": 2256076800
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013758038585209005,
      "loss": 2.411,
      "theoretical_loss": 3.3965006961117106,
      "tokens_seen": 2256142336
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013756383582371856,
      "loss": 2.2914,
      "theoretical_loss": 3.396492663708936,
      "tokens_seen": 2256207872
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001375472857953471,
      "loss": 1.9003,
      "theoretical_loss": 3.396484631604802,
      "tokens_seen": 2256273408
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001375307357669756,
      "loss": 2.4154,
      "theoretical_loss": 3.3964765997992883,
      "tokens_seen": 2256338944
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013751418573860412,
      "loss": 2.6031,
      "theoretical_loss": 3.396468568292375,
      "tokens_seen": 2256404480
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013749763571023263,
      "loss": 2.663,
      "theoretical_loss": 3.3964605370840424,
      "tokens_seen": 2256470016
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013748108568186116,
      "loss": 2.635,
      "theoretical_loss": 3.3964525061742714,
      "tokens_seen": 2256535552
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013746453565348967,
      "loss": 2.6854,
      "theoretical_loss": 3.396444475563041,
      "tokens_seen": 2256601088
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001374479856251182,
      "loss": 2.1981,
      "theoretical_loss": 3.3964364452503326,
      "tokens_seen": 2256666624
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013743143559674672,
      "loss": 2.5945,
      "theoretical_loss": 3.3964284152361257,
      "tokens_seen": 2256732160
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013741488556837526,
      "loss": 2.7014,
      "theoretical_loss": 3.396420385520401,
      "tokens_seen": 2256797696
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013739833554000377,
      "loss": 2.6545,
      "theoretical_loss": 3.3964123561031383,
      "tokens_seen": 2256863232
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001373817855116323,
      "loss": 2.364,
      "theoretical_loss": 3.3964043269843183,
      "tokens_seen": 2256928768
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013736523548326082,
      "loss": 2.2652,
      "theoretical_loss": 3.396396298163921,
      "tokens_seen": 2256994304
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013734868545488936,
      "loss": 2.1598,
      "theoretical_loss": 3.3963882696419265,
      "tokens_seen": 2257059840
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013733213542651787,
      "loss": 2.5031,
      "theoretical_loss": 3.3963802414183153,
      "tokens_seen": 2257125376
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001373155853981464,
      "loss": 2.6818,
      "theoretical_loss": 3.396372213493067,
      "tokens_seen": 2257190912
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001372990353697749,
      "loss": 1.8926,
      "theoretical_loss": 3.396364185866163,
      "tokens_seen": 2257256448
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013728248534140345,
      "loss": 2.7268,
      "theoretical_loss": 3.396356158537583,
      "tokens_seen": 2257321984
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013726593531303196,
      "loss": 2.3364,
      "theoretical_loss": 3.396348131507307,
      "tokens_seen": 2257387520
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013724938528466047,
      "loss": 2.4809,
      "theoretical_loss": 3.3963401047753154,
      "tokens_seen": 2257453056
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000137232835256289,
      "loss": 2.6202,
      "theoretical_loss": 3.3963320783415885,
      "tokens_seen": 2257518592
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013721628522791752,
      "loss": 2.6815,
      "theoretical_loss": 3.3963240522061064,
      "tokens_seen": 2257584128
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013719973519954606,
      "loss": 2.5285,
      "theoretical_loss": 3.39631602636885,
      "tokens_seen": 2257649664
    },
    {
      "epoch": 0.81,
      "objective/train/avg_token_score": 0.014900019392371178,
      "objective/train/avg_weight": 0.9824824333190918,
      "objective/train/docs_used": 1272388,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.2470414638519287,
      "objective/train/std_weight": 0.07598728686571121,
      "objective/train/theoretical_loss": 3.396308000829799,
      "objective/train/tokens_used": 2278175200,
      "theoretical_loss": 3.396308000829799,
      "tokens_seen": 2257715200
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013718318517117457,
      "loss": 2.5822,
      "theoretical_loss": 3.396308000829799,
      "tokens_seen": 2257715200
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001371666351428031,
      "loss": 2.6489,
      "theoretical_loss": 3.396299975588933,
      "tokens_seen": 2257780736
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013715008511443162,
      "loss": 2.2615,
      "theoretical_loss": 3.396291950646234,
      "tokens_seen": 2257846272
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013713353508606015,
      "loss": 2.5794,
      "theoretical_loss": 3.396283926001681,
      "tokens_seen": 2257911808
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013711698505768866,
      "loss": 2.2554,
      "theoretical_loss": 3.396275901655254,
      "tokens_seen": 2257977344
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001371004350293172,
      "loss": 2.8065,
      "theoretical_loss": 3.3962678776069346,
      "tokens_seen": 2258042880
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001370838850009457,
      "loss": 2.5655,
      "theoretical_loss": 3.396259853856702,
      "tokens_seen": 2258108416
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013706733497257425,
      "loss": 2.7922,
      "theoretical_loss": 3.3962518304045366,
      "tokens_seen": 2258173952
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013705078494420276,
      "loss": 2.4965,
      "theoretical_loss": 3.396243807250419,
      "tokens_seen": 2258239488
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001370342349158313,
      "loss": 1.677,
      "theoretical_loss": 3.3962357843943294,
      "tokens_seen": 2258305024
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001370176848874598,
      "loss": 2.4873,
      "theoretical_loss": 3.396227761836248,
      "tokens_seen": 2258370560
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013700113485908832,
      "loss": 2.3175,
      "theoretical_loss": 3.396219739576155,
      "tokens_seen": 2258436096
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013698458483071686,
      "loss": 2.3625,
      "theoretical_loss": 3.3962117176140305,
      "tokens_seen": 2258501632
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013696803480234537,
      "loss": 2.408,
      "theoretical_loss": 3.396203695949856,
      "tokens_seen": 2258567168
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001369514847739739,
      "loss": 2.5617,
      "theoretical_loss": 3.39619567458361,
      "tokens_seen": 2258632704
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013693493474560241,
      "loss": 2.577,
      "theoretical_loss": 3.396187653515274,
      "tokens_seen": 2258698240
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013691838471723095,
      "loss": 2.4234,
      "theoretical_loss": 3.3961796327448277,
      "tokens_seen": 2258763776
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013690183468885946,
      "loss": 2.4879,
      "theoretical_loss": 3.396171612272252,
      "tokens_seen": 2258829312
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000136885284660488,
      "loss": 2.4635,
      "theoretical_loss": 3.3961635920975266,
      "tokens_seen": 2258894848
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001368687346321165,
      "loss": 2.4755,
      "theoretical_loss": 3.396155572220632,
      "tokens_seen": 2258960384
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013685218460374505,
      "loss": 2.5023,
      "theoretical_loss": 3.3961475526415486,
      "tokens_seen": 2259025920
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013683563457537356,
      "loss": 2.491,
      "theoretical_loss": 3.396139533360256,
      "tokens_seen": 2259091456
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001368190845470021,
      "loss": 2.711,
      "theoretical_loss": 3.3961315143767363,
      "tokens_seen": 2259156992
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001368025345186306,
      "loss": 2.6766,
      "theoretical_loss": 3.396123495690968,
      "tokens_seen": 2259222528
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013678598449025914,
      "loss": 2.6333,
      "theoretical_loss": 3.3961154773029323,
      "tokens_seen": 2259288064
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013676943446188763,
      "loss": 2.5215,
      "theoretical_loss": 3.396107459212609,
      "tokens_seen": 2259353600
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013675288443351616,
      "loss": 2.5436,
      "theoretical_loss": 3.3960994414199783,
      "tokens_seen": 2259419136
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013673633440514467,
      "loss": 2.6088,
      "theoretical_loss": 3.3960914239250215,
      "tokens_seen": 2259484672
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001367197843767732,
      "loss": 2.6,
      "theoretical_loss": 3.396083406727718,
      "tokens_seen": 2259550208
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013670323434840172,
      "loss": 2.6241,
      "theoretical_loss": 3.3960753898280487,
      "tokens_seen": 2259615744
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013668668432003026,
      "loss": 2.3995,
      "theoretical_loss": 3.3960673732259936,
      "tokens_seen": 2259681280
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013667013429165877,
      "loss": 2.5597,
      "theoretical_loss": 3.396059356921533,
      "tokens_seen": 2259746816
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001366535842632873,
      "loss": 2.517,
      "theoretical_loss": 3.396051340914647,
      "tokens_seen": 2259812352
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013663703423491582,
      "loss": 2.5862,
      "theoretical_loss": 3.3960433252053166,
      "tokens_seen": 2259877888
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013662048420654436,
      "loss": 2.3944,
      "theoretical_loss": 3.3960353097935214,
      "tokens_seen": 2259943424
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013660393417817287,
      "loss": 2.5154,
      "theoretical_loss": 3.396027294679242,
      "tokens_seen": 2260008960
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001365873841498014,
      "loss": 2.72,
      "theoretical_loss": 3.3960192798624593,
      "tokens_seen": 2260074496
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013657083412142991,
      "loss": 2.3697,
      "theoretical_loss": 3.3960112653431525,
      "tokens_seen": 2260140032
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013655428409305845,
      "loss": 2.5267,
      "theoretical_loss": 3.396003251121303,
      "tokens_seen": 2260205568
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013653773406468696,
      "loss": 2.6149,
      "theoretical_loss": 3.3959952371968907,
      "tokens_seen": 2260271104
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013652118403631547,
      "loss": 2.3857,
      "theoretical_loss": 3.3959872235698954,
      "tokens_seen": 2260336640
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000136504634007944,
      "loss": 2.5507,
      "theoretical_loss": 3.3959792102402986,
      "tokens_seen": 2260402176
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013648808397957252,
      "loss": 2.5093,
      "theoretical_loss": 3.3959711972080795,
      "tokens_seen": 2260467712
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013647153395120106,
      "loss": 2.6795,
      "theoretical_loss": 3.395963184473219,
      "tokens_seen": 2260533248
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013645498392282957,
      "loss": 2.3307,
      "theoretical_loss": 3.3959551720356975,
      "tokens_seen": 2260598784
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001364384338944581,
      "loss": 2.5591,
      "theoretical_loss": 3.3959471598954956,
      "tokens_seen": 2260664320
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013642188386608662,
      "loss": 2.5303,
      "theoretical_loss": 3.3959391480525927,
      "tokens_seen": 2260729856
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013640533383771515,
      "loss": 2.4004,
      "theoretical_loss": 3.39593113650697,
      "tokens_seen": 2260795392
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013638878380934367,
      "loss": 2.4933,
      "theoretical_loss": 3.3959231252586077,
      "tokens_seen": 2260860928
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001363722337809722,
      "loss": 2.1079,
      "theoretical_loss": 3.3959151143074857,
      "tokens_seen": 2260926464
    },
    {
      "epoch": 0.81,
      "objective/train/avg_token_score": 0.05692027881741524,
      "objective/train/avg_weight": 0.9330305457115173,
      "objective/train/docs_used": 1274638,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.507783889770508,
      "objective/train/std_weight": 0.22566302120685577,
      "objective/train/theoretical_loss": 3.395907103653585,
      "objective/train/tokens_used": 2281452000,
      "theoretical_loss": 3.395907103653585,
      "tokens_seen": 2260992000
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001363556837526007,
      "loss": 2.425,
      "theoretical_loss": 3.395907103653585,
      "tokens_seen": 2260992000
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013633913372422925,
      "loss": 2.4471,
      "theoretical_loss": 3.3958990932968858,
      "tokens_seen": 2261057536
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013632258369585776,
      "loss": 2.5266,
      "theoretical_loss": 3.395891083237368,
      "tokens_seen": 2261123072
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001363060336674863,
      "loss": 2.6632,
      "theoretical_loss": 3.3958830734750123,
      "tokens_seen": 2261188608
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001362894836391148,
      "loss": 2.3657,
      "theoretical_loss": 3.3958750640097994,
      "tokens_seen": 2261254144
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013627293361074332,
      "loss": 2.554,
      "theoretical_loss": 3.395867054841709,
      "tokens_seen": 2261319680
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013625638358237186,
      "loss": 2.7707,
      "theoretical_loss": 3.3958590459707216,
      "tokens_seen": 2261385216
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013623983355400037,
      "loss": 2.4051,
      "theoretical_loss": 3.3958510373968176,
      "tokens_seen": 2261450752
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001362232835256289,
      "loss": 2.3873,
      "theoretical_loss": 3.395843029119978,
      "tokens_seen": 2261516288
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013620673349725742,
      "loss": 2.4719,
      "theoretical_loss": 3.395835021140183,
      "tokens_seen": 2261581824
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013619018346888595,
      "loss": 2.6149,
      "theoretical_loss": 3.395827013457412,
      "tokens_seen": 2261647360
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013617363344051446,
      "loss": 2.9126,
      "theoretical_loss": 3.395819006071646,
      "tokens_seen": 2261712896
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000136157083412143,
      "loss": 2.377,
      "theoretical_loss": 3.3958109989828658,
      "tokens_seen": 2261778432
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001361405333837715,
      "loss": 2.583,
      "theoretical_loss": 3.3958029921910513,
      "tokens_seen": 2261843968
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013612398335540005,
      "loss": 2.5196,
      "theoretical_loss": 3.395794985696183,
      "tokens_seen": 2261909504
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013610743332702856,
      "loss": 2.0658,
      "theoretical_loss": 3.395786979498241,
      "tokens_seen": 2261975040
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001360908832986571,
      "loss": 2.4042,
      "theoretical_loss": 3.3957789735972064,
      "tokens_seen": 2262040576
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001360743332702856,
      "loss": 2.4654,
      "theoretical_loss": 3.3957709679930588,
      "tokens_seen": 2262106112
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013605778324191415,
      "loss": 2.3952,
      "theoretical_loss": 3.395762962685779,
      "tokens_seen": 2262171648
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013604123321354263,
      "loss": 2.3495,
      "theoretical_loss": 3.395754957675347,
      "tokens_seen": 2262237184
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013602468318517117,
      "loss": 2.3845,
      "theoretical_loss": 3.395746952961744,
      "tokens_seen": 2262302720
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013600813315679968,
      "loss": 2.5356,
      "theoretical_loss": 3.39573894854495,
      "tokens_seen": 2262368256
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013599158312842821,
      "loss": 2.3901,
      "theoretical_loss": 3.395730944424945,
      "tokens_seen": 2262433792
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013597503310005672,
      "loss": 2.5202,
      "theoretical_loss": 3.3957229406017095,
      "tokens_seen": 2262499328
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013595848307168526,
      "loss": 2.5806,
      "theoretical_loss": 3.3957149370752244,
      "tokens_seen": 2262564864
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013594193304331377,
      "loss": 2.4171,
      "theoretical_loss": 3.39570693384547,
      "tokens_seen": 2262630400
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001359253830149423,
      "loss": 2.2547,
      "theoretical_loss": 3.3956989309124257,
      "tokens_seen": 2262695936
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013590883298657082,
      "loss": 2.6308,
      "theoretical_loss": 3.3956909282760734,
      "tokens_seen": 2262761472
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013589228295819936,
      "loss": 2.3221,
      "theoretical_loss": 3.3956829259363928,
      "tokens_seen": 2262827008
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013587573292982787,
      "loss": 2.5169,
      "theoretical_loss": 3.395674923893364,
      "tokens_seen": 2262892544
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001358591829014564,
      "loss": 2.4309,
      "theoretical_loss": 3.395666922146968,
      "tokens_seen": 2262958080
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013584263287308494,
      "loss": 2.5081,
      "theoretical_loss": 3.3956589206971852,
      "tokens_seen": 2263023616
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013582608284471345,
      "loss": 2.4682,
      "theoretical_loss": 3.3956509195439954,
      "tokens_seen": 2263089152
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000135809532816342,
      "loss": 2.4376,
      "theoretical_loss": 3.3956429186873796,
      "tokens_seen": 2263154688
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013579298278797047,
      "loss": 2.4858,
      "theoretical_loss": 3.395634918127318,
      "tokens_seen": 2263220224
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000135776432759599,
      "loss": 2.4634,
      "theoretical_loss": 3.3956269178637912,
      "tokens_seen": 2263285760
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013575988273122752,
      "loss": 2.3138,
      "theoretical_loss": 3.395618917896779,
      "tokens_seen": 2263351296
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013574333270285606,
      "loss": 2.4601,
      "theoretical_loss": 3.3956109182262626,
      "tokens_seen": 2263416832
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013572678267448457,
      "loss": 2.5731,
      "theoretical_loss": 3.395602918852222,
      "tokens_seen": 2263482368
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001357102326461131,
      "loss": 2.5124,
      "theoretical_loss": 3.395594919774638,
      "tokens_seen": 2263547904
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013569368261774162,
      "loss": 2.16,
      "theoretical_loss": 3.39558692099349,
      "tokens_seen": 2263613440
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013567713258937016,
      "loss": 2.6963,
      "theoretical_loss": 3.39557892250876,
      "tokens_seen": 2263678976
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013566058256099867,
      "loss": 2.6651,
      "theoretical_loss": 3.3955709243204275,
      "tokens_seen": 2263744512
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001356440325326272,
      "loss": 2.6917,
      "theoretical_loss": 3.3955629264284726,
      "tokens_seen": 2263810048
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013562748250425571,
      "loss": 2.7037,
      "theoretical_loss": 3.3955549288328766,
      "tokens_seen": 2263875584
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013561093247588425,
      "loss": 2.2292,
      "theoretical_loss": 3.3955469315336195,
      "tokens_seen": 2263941120
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013559438244751276,
      "loss": 1.2629,
      "theoretical_loss": 3.395538934530682,
      "tokens_seen": 2264006656
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001355778324191413,
      "loss": 2.6112,
      "theoretical_loss": 3.395530937824044,
      "tokens_seen": 2264072192
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001355612823907698,
      "loss": 2.6236,
      "theoretical_loss": 3.3955229414136863,
      "tokens_seen": 2264137728
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013554473236239832,
      "loss": 2.6274,
      "theoretical_loss": 3.3955149452995896,
      "tokens_seen": 2264203264
    },
    {
      "epoch": 0.81,
      "objective/train/avg_token_score": 0.012770509347319603,
      "objective/train/avg_weight": 0.9849683046340942,
      "objective/train/docs_used": 1276376,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.58665132522583,
      "objective/train/std_weight": 0.044604841619729996,
      "objective/train/theoretical_loss": 3.395506949481734,
      "objective/train/tokens_used": 2284728800,
      "theoretical_loss": 3.395506949481734,
      "tokens_seen": 2264268800
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013552818233402686,
      "loss": 2.5888,
      "theoretical_loss": 3.395506949481734,
      "tokens_seen": 2264268800
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013551163230565537,
      "loss": 2.3305,
      "theoretical_loss": 3.3954989539601,
      "tokens_seen": 2264334336
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001354950822772839,
      "loss": 2.4385,
      "theoretical_loss": 3.395490958734668,
      "tokens_seen": 2264399872
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013547853224891242,
      "loss": 2.486,
      "theoretical_loss": 3.3954829638054185,
      "tokens_seen": 2264465408
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013546198222054095,
      "loss": 2.4144,
      "theoretical_loss": 3.395474969172332,
      "tokens_seen": 2264530944
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013544543219216946,
      "loss": 2.591,
      "theoretical_loss": 3.395466974835389,
      "tokens_seen": 2264596480
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000135428882163798,
      "loss": 2.6899,
      "theoretical_loss": 3.39545898079457,
      "tokens_seen": 2264662016
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001354123321354265,
      "loss": 2.519,
      "theoretical_loss": 3.3954509870498555,
      "tokens_seen": 2264727552
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013539578210705505,
      "loss": 2.6087,
      "theoretical_loss": 3.3954429936012254,
      "tokens_seen": 2264793088
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013537923207868356,
      "loss": 2.772,
      "theoretical_loss": 3.3954350004486606,
      "tokens_seen": 2264858624
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001353626820503121,
      "loss": 2.7118,
      "theoretical_loss": 3.3954270075921418,
      "tokens_seen": 2264924160
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001353461320219406,
      "loss": 2.3772,
      "theoretical_loss": 3.3954190150316492,
      "tokens_seen": 2264989696
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013532958199356915,
      "loss": 2.6525,
      "theoretical_loss": 3.3954110227671634,
      "tokens_seen": 2265055232
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013531303196519766,
      "loss": 2.4378,
      "theoretical_loss": 3.395403030798665,
      "tokens_seen": 2265120768
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013529648193682617,
      "loss": 2.5608,
      "theoretical_loss": 3.395395039126134,
      "tokens_seen": 2265186304
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013527993190845468,
      "loss": 2.7988,
      "theoretical_loss": 3.3953870477495514,
      "tokens_seen": 2265251840
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013526338188008322,
      "loss": 2.4292,
      "theoretical_loss": 3.395379056668897,
      "tokens_seen": 2265317376
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013524683185171175,
      "loss": 2.4685,
      "theoretical_loss": 3.395371065884152,
      "tokens_seen": 2265382912
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013523028182334026,
      "loss": 2.1653,
      "theoretical_loss": 3.395363075395296,
      "tokens_seen": 2265448448
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001352137317949688,
      "loss": 2.346,
      "theoretical_loss": 3.395355085202311,
      "tokens_seen": 2265513984
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001351971817665973,
      "loss": 2.4531,
      "theoretical_loss": 3.395347095305176,
      "tokens_seen": 2265579520
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013518063173822585,
      "loss": 2.5515,
      "theoretical_loss": 3.3953391057038718,
      "tokens_seen": 2265645056
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013516408170985436,
      "loss": 2.2674,
      "theoretical_loss": 3.3953311163983795,
      "tokens_seen": 2265710592
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001351475316814829,
      "loss": 2.3169,
      "theoretical_loss": 3.395323127388679,
      "tokens_seen": 2265776128
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001351309816531114,
      "loss": 2.3447,
      "theoretical_loss": 3.3953151386747513,
      "tokens_seen": 2265841664
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013511443162473994,
      "loss": 2.505,
      "theoretical_loss": 3.3953071502565764,
      "tokens_seen": 2265907200
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013509788159636846,
      "loss": 2.6262,
      "theoretical_loss": 3.395299162134135,
      "tokens_seen": 2265972736
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000135081331567997,
      "loss": 2.7384,
      "theoretical_loss": 3.395291174307408,
      "tokens_seen": 2266038272
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001350647815396255,
      "loss": 2.5468,
      "theoretical_loss": 3.395283186776375,
      "tokens_seen": 2266103808
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013504823151125401,
      "loss": 2.7272,
      "theoretical_loss": 3.3952751995410173,
      "tokens_seen": 2266169344
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013503168148288252,
      "loss": 2.6432,
      "theoretical_loss": 3.395267212601315,
      "tokens_seen": 2266234880
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013501513145451106,
      "loss": 2.3876,
      "theoretical_loss": 3.3952592259572487,
      "tokens_seen": 2266300416
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013499858142613957,
      "loss": 2.6824,
      "theoretical_loss": 3.395251239608799,
      "tokens_seen": 2266365952
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001349820313977681,
      "loss": 2.6403,
      "theoretical_loss": 3.3952432535559463,
      "tokens_seen": 2266431488
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013496548136939662,
      "loss": 2.5047,
      "theoretical_loss": 3.3952352677986712,
      "tokens_seen": 2266497024
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013494893134102516,
      "loss": 2.6211,
      "theoretical_loss": 3.395227282336954,
      "tokens_seen": 2266562560
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013493238131265367,
      "loss": 2.2415,
      "theoretical_loss": 3.3952192971707755,
      "tokens_seen": 2266628096
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001349158312842822,
      "loss": 2.4794,
      "theoretical_loss": 3.395211312300116,
      "tokens_seen": 2266693632
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013489928125591072,
      "loss": 2.6799,
      "theoretical_loss": 3.3952033277249565,
      "tokens_seen": 2266759168
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013488273122753925,
      "loss": 2.756,
      "theoretical_loss": 3.395195343445277,
      "tokens_seen": 2266824704
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013486618119916776,
      "loss": 2.4837,
      "theoretical_loss": 3.3951873594610578,
      "tokens_seen": 2266890240
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001348496311707963,
      "loss": 2.4399,
      "theoretical_loss": 3.3951793757722797,
      "tokens_seen": 2266955776
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001348330811424248,
      "loss": 2.7135,
      "theoretical_loss": 3.395171392378924,
      "tokens_seen": 2267021312
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013481653111405335,
      "loss": 2.4646,
      "theoretical_loss": 3.39516340928097,
      "tokens_seen": 2267086848
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013479998108568186,
      "loss": 2.4927,
      "theoretical_loss": 3.395155426478399,
      "tokens_seen": 2267152384
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013478343105731037,
      "loss": 2.503,
      "theoretical_loss": 3.395147443971191,
      "tokens_seen": 2267217920
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001347668810289389,
      "loss": 2.6134,
      "theoretical_loss": 3.395139461759327,
      "tokens_seen": 2267283456
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013475033100056742,
      "loss": 2.5457,
      "theoretical_loss": 3.395131479842788,
      "tokens_seen": 2267348992
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013473378097219596,
      "loss": 2.5423,
      "theoretical_loss": 3.3951234982215532,
      "tokens_seen": 2267414528
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013471723094382447,
      "loss": 2.4996,
      "theoretical_loss": 3.395115516895604,
      "tokens_seen": 2267480064
    },
    {
      "epoch": 0.81,
      "objective/train/avg_token_score": 0.02744131162762642,
      "objective/train/avg_weight": 0.967720091342926,
      "objective/train/docs_used": 1278407,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.280725479125977,
      "objective/train/std_weight": 0.12944108247756958,
      "objective/train/theoretical_loss": 3.3951075358649208,
      "objective/train/tokens_used": 2288005600,
      "theoretical_loss": 3.3951075358649208,
      "tokens_seen": 2267545600
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000134700680915453,
      "loss": 2.5481,
      "theoretical_loss": 3.3951075358649208,
      "tokens_seen": 2267545600
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013468413088708151,
      "loss": 2.2955,
      "theoretical_loss": 3.3950995551294842,
      "tokens_seen": 2267611136
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013466758085871005,
      "loss": 2.5367,
      "theoretical_loss": 3.3950915746892747,
      "tokens_seen": 2267676672
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013465103083033856,
      "loss": 2.4918,
      "theoretical_loss": 3.395083594544273,
      "tokens_seen": 2267742208
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001346344808019671,
      "loss": 2.3938,
      "theoretical_loss": 3.395075614694459,
      "tokens_seen": 2267807744
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001346179307735956,
      "loss": 2.5865,
      "theoretical_loss": 3.395067635139814,
      "tokens_seen": 2267873280
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013460138074522415,
      "loss": 2.426,
      "theoretical_loss": 3.3950596558803183,
      "tokens_seen": 2267938816
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013458483071685266,
      "loss": 1.837,
      "theoretical_loss": 3.3950516769159527,
      "tokens_seen": 2268004352
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001345682806884812,
      "loss": 2.5173,
      "theoretical_loss": 3.395043698246697,
      "tokens_seen": 2268069888
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001345517306601097,
      "loss": 2.5055,
      "theoretical_loss": 3.3950357198725327,
      "tokens_seen": 2268135424
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013453518063173822,
      "loss": 2.3479,
      "theoretical_loss": 3.39502774179344,
      "tokens_seen": 2268200960
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013451863060336675,
      "loss": 2.619,
      "theoretical_loss": 3.395019764009399,
      "tokens_seen": 2268266496
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013450208057499526,
      "loss": 2.48,
      "theoretical_loss": 3.3950117865203904,
      "tokens_seen": 2268332032
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001344855305466238,
      "loss": 2.6896,
      "theoretical_loss": 3.3950038093263952,
      "tokens_seen": 2268397568
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001344689805182523,
      "loss": 2.507,
      "theoretical_loss": 3.3949958324273943,
      "tokens_seen": 2268463104
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013445243048988085,
      "loss": 2.3914,
      "theoretical_loss": 3.3949878558233673,
      "tokens_seen": 2268528640
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013443588046150936,
      "loss": 2.4535,
      "theoretical_loss": 3.3949798795142954,
      "tokens_seen": 2268594176
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001344193304331379,
      "loss": 2.3217,
      "theoretical_loss": 3.3949719035001586,
      "tokens_seen": 2268659712
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001344027804047664,
      "loss": 2.5192,
      "theoretical_loss": 3.394963927780938,
      "tokens_seen": 2268725248
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013438623037639495,
      "loss": 2.5585,
      "theoretical_loss": 3.3949559523566144,
      "tokens_seen": 2268790784
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013436968034802346,
      "loss": 2.4036,
      "theoretical_loss": 3.394947977227168,
      "tokens_seen": 2268856320
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000134353130319652,
      "loss": 2.6419,
      "theoretical_loss": 3.3949400023925786,
      "tokens_seen": 2268921856
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001343365802912805,
      "loss": 2.6571,
      "theoretical_loss": 3.3949320278528283,
      "tokens_seen": 2268987392
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013432003026290902,
      "loss": 2.5887,
      "theoretical_loss": 3.394924053607897,
      "tokens_seen": 2269052928
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013430348023453753,
      "loss": 2.4342,
      "theoretical_loss": 3.394916079657765,
      "tokens_seen": 2269118464
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013428693020616606,
      "loss": 2.2761,
      "theoretical_loss": 3.394908106002413,
      "tokens_seen": 2269184000
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013427038017779457,
      "loss": 2.6299,
      "theoretical_loss": 3.394900132641822,
      "tokens_seen": 2269249536
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001342538301494231,
      "loss": 2.8226,
      "theoretical_loss": 3.3948921595759725,
      "tokens_seen": 2269315072
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013423728012105162,
      "loss": 2.4301,
      "theoretical_loss": 3.3948841868048447,
      "tokens_seen": 2269380608
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013422073009268016,
      "loss": 2.567,
      "theoretical_loss": 3.3948762143284195,
      "tokens_seen": 2269446144
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013420418006430867,
      "loss": 2.4122,
      "theoretical_loss": 3.394868242146677,
      "tokens_seen": 2269511680
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001341876300359372,
      "loss": 2.4822,
      "theoretical_loss": 3.3948602702595987,
      "tokens_seen": 2269577216
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013417108000756572,
      "loss": 2.6057,
      "theoretical_loss": 3.3948522986671645,
      "tokens_seen": 2269642752
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013415452997919425,
      "loss": 1.4962,
      "theoretical_loss": 3.394844327369355,
      "tokens_seen": 2269708288
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013413797995082277,
      "loss": 2.4432,
      "theoretical_loss": 3.3948363563661514,
      "tokens_seen": 2269773824
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001341214299224513,
      "loss": 2.5954,
      "theoretical_loss": 3.394828385657534,
      "tokens_seen": 2269839360
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001341048798940798,
      "loss": 2.5446,
      "theoretical_loss": 3.394820415243483,
      "tokens_seen": 2269904896
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013408832986570835,
      "loss": 2.3707,
      "theoretical_loss": 3.394812445123979,
      "tokens_seen": 2269970432
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013407177983733686,
      "loss": 2.6293,
      "theoretical_loss": 3.3948044752990034,
      "tokens_seen": 2270035968
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013405522980896537,
      "loss": 2.4711,
      "theoretical_loss": 3.3947965057685363,
      "tokens_seen": 2270101504
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001340386797805939,
      "loss": 2.477,
      "theoretical_loss": 3.3947885365325585,
      "tokens_seen": 2270167040
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013402212975222242,
      "loss": 2.665,
      "theoretical_loss": 3.3947805675910505,
      "tokens_seen": 2270232576
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013400557972385096,
      "loss": 2.2529,
      "theoretical_loss": 3.3947725989439927,
      "tokens_seen": 2270298112
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013398902969547947,
      "loss": 2.813,
      "theoretical_loss": 3.3947646305913657,
      "tokens_seen": 2270363648
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000133972479667108,
      "loss": 2.578,
      "theoretical_loss": 3.394756662533151,
      "tokens_seen": 2270429184
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013395592963873652,
      "loss": 2.7796,
      "theoretical_loss": 3.394748694769328,
      "tokens_seen": 2270494720
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013393937961036505,
      "loss": 2.4813,
      "theoretical_loss": 3.394740727299878,
      "tokens_seen": 2270560256
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013392282958199356,
      "loss": 2.4008,
      "theoretical_loss": 3.394732760124782,
      "tokens_seen": 2270625792
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001339062795536221,
      "loss": 2.2996,
      "theoretical_loss": 3.39472479324402,
      "tokens_seen": 2270691328
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001338897295252506,
      "loss": 2.2526,
      "theoretical_loss": 3.3947168266575725,
      "tokens_seen": 2270756864
    },
    {
      "epoch": 0.81,
      "objective/train/avg_token_score": 0.016306180506944656,
      "objective/train/avg_weight": 0.9808123111724854,
      "objective/train/docs_used": 1280391,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.78997802734375,
      "objective/train/std_weight": 0.07152727246284485,
      "objective/train/theoretical_loss": 3.3947088603654203,
      "objective/train/tokens_used": 2291282400,
      "theoretical_loss": 3.3947088603654203,
      "tokens_seen": 2270822400
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013387317949687915,
      "loss": 2.6381,
      "theoretical_loss": 3.3947088603654203,
      "tokens_seen": 2270822400
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013385662946850766,
      "loss": 2.5007,
      "theoretical_loss": 3.3947008943675447,
      "tokens_seen": 2270887936
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001338400794401362,
      "loss": 2.666,
      "theoretical_loss": 3.3946929286639254,
      "tokens_seen": 2270953472
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001338235294117647,
      "loss": 2.4582,
      "theoretical_loss": 3.3946849632545435,
      "tokens_seen": 2271019008
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013380697938339322,
      "loss": 2.5178,
      "theoretical_loss": 3.3946769981393796,
      "tokens_seen": 2271084544
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013379042935502176,
      "loss": 2.5092,
      "theoretical_loss": 3.394669033318414,
      "tokens_seen": 2271150080
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013377387932665027,
      "loss": 2.6429,
      "theoretical_loss": 3.3946610687916285,
      "tokens_seen": 2271215616
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001337573292982788,
      "loss": 2.4113,
      "theoretical_loss": 3.3946531045590023,
      "tokens_seen": 2271281152
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013374077926990731,
      "loss": 2.8525,
      "theoretical_loss": 3.3946451406205167,
      "tokens_seen": 2271346688
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013372422924153585,
      "loss": 2.697,
      "theoretical_loss": 3.3946371769761527,
      "tokens_seen": 2271412224
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013370767921316436,
      "loss": 2.7531,
      "theoretical_loss": 3.3946292136258904,
      "tokens_seen": 2271477760
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001336911291847929,
      "loss": 2.5385,
      "theoretical_loss": 3.3946212505697106,
      "tokens_seen": 2271543296
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001336745791564214,
      "loss": 2.3643,
      "theoretical_loss": 3.3946132878075934,
      "tokens_seen": 2271608832
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013365802912804995,
      "loss": 2.337,
      "theoretical_loss": 3.394605325339521,
      "tokens_seen": 2271674368
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013364147909967846,
      "loss": 2.34,
      "theoretical_loss": 3.3945973631654724,
      "tokens_seen": 2271739904
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000133624929071307,
      "loss": 2.6422,
      "theoretical_loss": 3.394589401285429,
      "tokens_seen": 2271805440
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001336083790429355,
      "loss": 2.6356,
      "theoretical_loss": 3.3945814396993716,
      "tokens_seen": 2271870976
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013359182901456404,
      "loss": 2.554,
      "theoretical_loss": 3.394573478407281,
      "tokens_seen": 2271936512
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013357527898619253,
      "loss": 2.4739,
      "theoretical_loss": 3.394565517409137,
      "tokens_seen": 2272002048
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013355872895782106,
      "loss": 2.5075,
      "theoretical_loss": 3.394557556704921,
      "tokens_seen": 2272067584
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013354217892944957,
      "loss": 2.5835,
      "theoretical_loss": 3.3945495962946137,
      "tokens_seen": 2272133120
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001335256289010781,
      "loss": 2.5986,
      "theoretical_loss": 3.3945416361781953,
      "tokens_seen": 2272198656
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013350907887270662,
      "loss": 2.5393,
      "theoretical_loss": 3.394533676355647,
      "tokens_seen": 2272264192
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013349252884433516,
      "loss": 2.3636,
      "theoretical_loss": 3.3945257168269487,
      "tokens_seen": 2272329728
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013347597881596367,
      "loss": 2.5951,
      "theoretical_loss": 3.394517757592082,
      "tokens_seen": 2272395264
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001334594287875922,
      "loss": 2.6576,
      "theoretical_loss": 3.3945097986510273,
      "tokens_seen": 2272460800
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013344287875922072,
      "loss": 2.2099,
      "theoretical_loss": 3.3945018400037648,
      "tokens_seen": 2272526336
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013342632873084926,
      "loss": 2.6884,
      "theoretical_loss": 3.3944938816502757,
      "tokens_seen": 2272591872
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013340977870247777,
      "loss": 2.3201,
      "theoretical_loss": 3.3944859235905405,
      "tokens_seen": 2272657408
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001333932286741063,
      "loss": 2.4092,
      "theoretical_loss": 3.3944779658245396,
      "tokens_seen": 2272722944
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013337667864573481,
      "loss": 2.4001,
      "theoretical_loss": 3.394470008352254,
      "tokens_seen": 2272788480
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013336012861736335,
      "loss": 2.451,
      "theoretical_loss": 3.394462051173665,
      "tokens_seen": 2272854016
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013334357858899186,
      "loss": 2.4535,
      "theoretical_loss": 3.3944540942887524,
      "tokens_seen": 2272919552
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013332702856062037,
      "loss": 2.6863,
      "theoretical_loss": 3.394446137697497,
      "tokens_seen": 2272985088
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001333104785322489,
      "loss": 2.698,
      "theoretical_loss": 3.3944381813998796,
      "tokens_seen": 2273050624
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013329392850387742,
      "loss": 2.3214,
      "theoretical_loss": 3.3944302253958814,
      "tokens_seen": 2273116160
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013327737847550596,
      "loss": 2.744,
      "theoretical_loss": 3.394422269685482,
      "tokens_seen": 2273181696
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013326082844713447,
      "loss": 2.7048,
      "theoretical_loss": 3.394414314268663,
      "tokens_seen": 2273247232
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000133244278418763,
      "loss": 2.6355,
      "theoretical_loss": 3.394406359145405,
      "tokens_seen": 2273312768
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013322772839039152,
      "loss": 2.6459,
      "theoretical_loss": 3.3943984043156887,
      "tokens_seen": 2273378304
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013321117836202005,
      "loss": 2.2371,
      "theoretical_loss": 3.3943904497794946,
      "tokens_seen": 2273443840
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013319462833364857,
      "loss": 2.2778,
      "theoretical_loss": 3.3943824955368034,
      "tokens_seen": 2273509376
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001331780783052771,
      "loss": 2.172,
      "theoretical_loss": 3.3943745415875957,
      "tokens_seen": 2273574912
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001331615282769056,
      "loss": 2.3814,
      "theoretical_loss": 3.3943665879318528,
      "tokens_seen": 2273640448
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013314497824853415,
      "loss": 2.6643,
      "theoretical_loss": 3.3943586345695547,
      "tokens_seen": 2273705984
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013312842822016266,
      "loss": 1.9491,
      "theoretical_loss": 3.3943506815006823,
      "tokens_seen": 2273771520
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001331118781917912,
      "loss": 2.5476,
      "theoretical_loss": 3.394342728725217,
      "tokens_seen": 2273837056
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001330953281634197,
      "loss": 2.4368,
      "theoretical_loss": 3.3943347762431384,
      "tokens_seen": 2273902592
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013307877813504822,
      "loss": 2.6041,
      "theoretical_loss": 3.394326824054428,
      "tokens_seen": 2273968128
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013306222810667676,
      "loss": 2.5589,
      "theoretical_loss": 3.3943188721590665,
      "tokens_seen": 2274033664
    },
    {
      "epoch": 0.81,
      "objective/train/avg_token_score": 0.014339342713356018,
      "objective/train/avg_weight": 0.983124315738678,
      "objective/train/docs_used": 1282312,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9759795665740967,
      "objective/train/std_weight": 0.07358884811401367,
      "objective/train/theoretical_loss": 3.394310920557034,
      "objective/train/tokens_used": 2294559200,
      "theoretical_loss": 3.394310920557034,
      "tokens_seen": 2274099200
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013304567807830527,
      "loss": 3.0002,
      "theoretical_loss": 3.394310920557034,
      "tokens_seen": 2274099200
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001330291280499338,
      "loss": 2.7506,
      "theoretical_loss": 3.3943029692483124,
      "tokens_seen": 2274164736
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013301257802156232,
      "loss": 2.6094,
      "theoretical_loss": 3.394295018232881,
      "tokens_seen": 2274230272
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013299602799319085,
      "loss": 2.5951,
      "theoretical_loss": 3.3942870675107217,
      "tokens_seen": 2274295808
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013297947796481936,
      "loss": 1.7351,
      "theoretical_loss": 3.3942791170818145,
      "tokens_seen": 2274361344
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001329629279364479,
      "loss": 2.6241,
      "theoretical_loss": 3.3942711669461407,
      "tokens_seen": 2274426880
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001329463779080764,
      "loss": 2.756,
      "theoretical_loss": 3.3942632171036804,
      "tokens_seen": 2274492416
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013292982787970495,
      "loss": 2.1592,
      "theoretical_loss": 3.394255267554415,
      "tokens_seen": 2274557952
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013291327785133346,
      "loss": 2.5979,
      "theoretical_loss": 3.3942473182983246,
      "tokens_seen": 2274623488
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000132896727822962,
      "loss": 1.9347,
      "theoretical_loss": 3.3942393693353905,
      "tokens_seen": 2274689024
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001328801777945905,
      "loss": 2.3205,
      "theoretical_loss": 3.3942314206655926,
      "tokens_seen": 2274754560
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013286362776621905,
      "loss": 2.4772,
      "theoretical_loss": 3.3942234722889126,
      "tokens_seen": 2274820096
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013284707773784753,
      "loss": 2.829,
      "theoretical_loss": 3.394215524205331,
      "tokens_seen": 2274885632
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013283052770947607,
      "loss": 2.4435,
      "theoretical_loss": 3.3942075764148285,
      "tokens_seen": 2274951168
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013281397768110458,
      "loss": 2.6258,
      "theoretical_loss": 3.3941996289173857,
      "tokens_seen": 2275016704
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013279742765273311,
      "loss": 2.3498,
      "theoretical_loss": 3.3941916817129836,
      "tokens_seen": 2275082240
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013278087762436162,
      "loss": 2.866,
      "theoretical_loss": 3.3941837348016026,
      "tokens_seen": 2275147776
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013276432759599016,
      "loss": 2.708,
      "theoretical_loss": 3.3941757881832233,
      "tokens_seen": 2275213312
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013274777756761867,
      "loss": 3.875,
      "theoretical_loss": 3.3941678418578274,
      "tokens_seen": 2275278848
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001327312275392472,
      "loss": 2.5385,
      "theoretical_loss": 3.394159895825395,
      "tokens_seen": 2275344384
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013271467751087572,
      "loss": 2.7091,
      "theoretical_loss": 3.3941519500859063,
      "tokens_seen": 2275409920
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013269812748250426,
      "loss": 5.1061,
      "theoretical_loss": 3.3941440046393434,
      "tokens_seen": 2275475456
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013268157745413277,
      "loss": 2.4972,
      "theoretical_loss": 3.394136059485686,
      "tokens_seen": 2275540992
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001326650274257613,
      "loss": 2.2685,
      "theoretical_loss": 3.394128114624915,
      "tokens_seen": 2275606528
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013264847739738982,
      "loss": 2.5907,
      "theoretical_loss": 3.394120170057012,
      "tokens_seen": 2275672064
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013263192736901835,
      "loss": 2.205,
      "theoretical_loss": 3.394112225781957,
      "tokens_seen": 2275737600
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013261537734064686,
      "loss": 2.3557,
      "theoretical_loss": 3.3941042817997307,
      "tokens_seen": 2275803136
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013259882731227537,
      "loss": 2.5594,
      "theoretical_loss": 3.394096338110314,
      "tokens_seen": 2275868672
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001325822772839039,
      "loss": 2.7787,
      "theoretical_loss": 3.394088394713688,
      "tokens_seen": 2275934208
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013256572725553242,
      "loss": 2.4685,
      "theoretical_loss": 3.3940804516098333,
      "tokens_seen": 2275999744
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013254917722716096,
      "loss": 2.5098,
      "theoretical_loss": 3.3940725087987307,
      "tokens_seen": 2276065280
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013253262719878947,
      "loss": 2.3362,
      "theoretical_loss": 3.3940645662803606,
      "tokens_seen": 2276130816
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000132516077170418,
      "loss": 2.1148,
      "theoretical_loss": 3.3940566240547043,
      "tokens_seen": 2276196352
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013249952714204652,
      "loss": 2.4772,
      "theoretical_loss": 3.3940486821217424,
      "tokens_seen": 2276261888
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013248297711367506,
      "loss": 2.9696,
      "theoretical_loss": 3.3940407404814557,
      "tokens_seen": 2276327424
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013246642708530357,
      "loss": 2.6793,
      "theoretical_loss": 3.394032799133825,
      "tokens_seen": 2276392960
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001324498770569321,
      "loss": 2.3725,
      "theoretical_loss": 3.394024858078831,
      "tokens_seen": 2276458496
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013243332702856061,
      "loss": 2.4652,
      "theoretical_loss": 3.3940169173164545,
      "tokens_seen": 2276524032
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013241677700018915,
      "loss": 2.3996,
      "theoretical_loss": 3.394008976846676,
      "tokens_seen": 2276589568
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013240022697181766,
      "loss": 2.5819,
      "theoretical_loss": 3.394001036669477,
      "tokens_seen": 2276655104
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001323836769434462,
      "loss": 2.7336,
      "theoretical_loss": 3.393993096784838,
      "tokens_seen": 2276720640
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001323671269150747,
      "loss": 2.3931,
      "theoretical_loss": 3.3939851571927395,
      "tokens_seen": 2276786176
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013235057688670322,
      "loss": 2.4765,
      "theoretical_loss": 3.393977217893163,
      "tokens_seen": 2276851712
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013233402685833176,
      "loss": 2.3204,
      "theoretical_loss": 3.3939692788860887,
      "tokens_seen": 2276917248
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013231747682996027,
      "loss": 2.5696,
      "theoretical_loss": 3.393961340171497,
      "tokens_seen": 2276982784
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001323009268015888,
      "loss": 2.4534,
      "theoretical_loss": 3.39395340174937,
      "tokens_seen": 2277048320
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013228437677321732,
      "loss": 2.5904,
      "theoretical_loss": 3.393945463619687,
      "tokens_seen": 2277113856
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013226782674484585,
      "loss": 2.2894,
      "theoretical_loss": 3.3939375257824302,
      "tokens_seen": 2277179392
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013225127671647436,
      "loss": 2.765,
      "theoretical_loss": 3.3939295882375795,
      "tokens_seen": 2277244928
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001322347266881029,
      "loss": 2.4771,
      "theoretical_loss": 3.3939216509851162,
      "tokens_seen": 2277310464
    },
    {
      "epoch": 0.81,
      "objective/train/avg_token_score": 0.013572119176387787,
      "objective/train/avg_weight": 0.984029233455658,
      "objective/train/docs_used": 1284150,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.3640623092651367,
      "objective/train/std_weight": 0.07640250772237778,
      "objective/train/theoretical_loss": 3.393913714025021,
      "objective/train/tokens_used": 2297836000,
      "theoretical_loss": 3.393913714025021,
      "tokens_seen": 2277376000
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001322181766597314,
      "loss": 2.4667,
      "theoretical_loss": 3.393913714025021,
      "tokens_seen": 2277376000
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013220162663135995,
      "loss": 2.7875,
      "theoretical_loss": 3.393905777357274,
      "tokens_seen": 2277441536
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013218507660298846,
      "loss": 2.5373,
      "theoretical_loss": 3.393897840981857,
      "tokens_seen": 2277507072
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000132168526574617,
      "loss": 2.3714,
      "theoretical_loss": 3.3938899048987508,
      "tokens_seen": 2277572608
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001321519765462455,
      "loss": 2.5321,
      "theoretical_loss": 3.3938819691079356,
      "tokens_seen": 2277638144
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013213542651787405,
      "loss": 1.8745,
      "theoretical_loss": 3.393874033609392,
      "tokens_seen": 2277703680
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013211887648950256,
      "loss": 2.6877,
      "theoretical_loss": 3.393866098403102,
      "tokens_seen": 2277769216
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013210232646113107,
      "loss": 2.2837,
      "theoretical_loss": 3.3938581634890457,
      "tokens_seen": 2277834752
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013208577643275958,
      "loss": 2.5071,
      "theoretical_loss": 3.393850228867204,
      "tokens_seen": 2277900288
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013206922640438812,
      "loss": 1.7721,
      "theoretical_loss": 3.3938422945375577,
      "tokens_seen": 2277965824
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013205267637601663,
      "loss": 2.8128,
      "theoretical_loss": 3.3938343605000876,
      "tokens_seen": 2278031360
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013203612634764516,
      "loss": 2.3981,
      "theoretical_loss": 3.3938264267547744,
      "tokens_seen": 2278096896
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013201957631927367,
      "loss": 2.7257,
      "theoretical_loss": 3.3938184933015996,
      "tokens_seen": 2278162432
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001320030262909022,
      "loss": 2.4492,
      "theoretical_loss": 3.3938105601405435,
      "tokens_seen": 2278227968
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013198647626253072,
      "loss": 2.651,
      "theoretical_loss": 3.3938026272715867,
      "tokens_seen": 2278293504
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013196992623415926,
      "loss": 2.5795,
      "theoretical_loss": 3.3937946946947104,
      "tokens_seen": 2278359040
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013195337620578777,
      "loss": 2.596,
      "theoretical_loss": 3.393786762409896,
      "tokens_seen": 2278424576
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001319368261774163,
      "loss": 2.2776,
      "theoretical_loss": 3.3937788304171232,
      "tokens_seen": 2278490112
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013192027614904482,
      "loss": 2.671,
      "theoretical_loss": 3.3937708987163733,
      "tokens_seen": 2278555648
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013190372612067336,
      "loss": 2.6043,
      "theoretical_loss": 3.393762967307628,
      "tokens_seen": 2278621184
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013188717609230187,
      "loss": 2.6433,
      "theoretical_loss": 3.393755036190867,
      "tokens_seen": 2278686720
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001318706260639304,
      "loss": 2.3789,
      "theoretical_loss": 3.393747105366071,
      "tokens_seen": 2278752256
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013185407603555891,
      "loss": 2.5641,
      "theoretical_loss": 3.393739174833222,
      "tokens_seen": 2278817792
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013183752600718742,
      "loss": 2.5726,
      "theoretical_loss": 3.3937312445923,
      "tokens_seen": 2278883328
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013182097597881596,
      "loss": 2.4284,
      "theoretical_loss": 3.393723314643286,
      "tokens_seen": 2278948864
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013180442595044447,
      "loss": 2.6923,
      "theoretical_loss": 3.393715384986162,
      "tokens_seen": 2279014400
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000131787875922073,
      "loss": 2.4455,
      "theoretical_loss": 3.393707455620907,
      "tokens_seen": 2279079936
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013177132589370152,
      "loss": 2.5865,
      "theoretical_loss": 3.3936995265475027,
      "tokens_seen": 2279145472
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013175477586533006,
      "loss": 2.6113,
      "theoretical_loss": 3.3936915977659305,
      "tokens_seen": 2279211008
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013173822583695857,
      "loss": 2.5039,
      "theoretical_loss": 3.39368366927617,
      "tokens_seen": 2279276544
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001317216758085871,
      "loss": 2.632,
      "theoretical_loss": 3.3936757410782032,
      "tokens_seen": 2279342080
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013170512578021562,
      "loss": 2.6875,
      "theoretical_loss": 3.393667813172011,
      "tokens_seen": 2279407616
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013168857575184415,
      "loss": 2.5265,
      "theoretical_loss": 3.3936598855575735,
      "tokens_seen": 2279473152
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013167202572347266,
      "loss": 2.4369,
      "theoretical_loss": 3.393651958234872,
      "tokens_seen": 2279538688
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001316554756951012,
      "loss": 2.6564,
      "theoretical_loss": 3.393644031203887,
      "tokens_seen": 2279604224
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001316389256667297,
      "loss": 2.6237,
      "theoretical_loss": 3.3936361044646,
      "tokens_seen": 2279669760
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013162237563835825,
      "loss": 2.4785,
      "theoretical_loss": 3.393628178016992,
      "tokens_seen": 2279735296
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013160582560998676,
      "loss": 2.5879,
      "theoretical_loss": 3.3936202518610425,
      "tokens_seen": 2279800832
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013158927558161527,
      "loss": 2.7334,
      "theoretical_loss": 3.393612325996734,
      "tokens_seen": 2279866368
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001315727255532438,
      "loss": 2.5544,
      "theoretical_loss": 3.3936044004240467,
      "tokens_seen": 2279931904
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013155617552487232,
      "loss": 3.1703,
      "theoretical_loss": 3.3935964751429615,
      "tokens_seen": 2279997440
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013153962549650086,
      "loss": 2.4291,
      "theoretical_loss": 3.393588550153459,
      "tokens_seen": 2280062976
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013152307546812937,
      "loss": 2.4991,
      "theoretical_loss": 3.393580625455521,
      "tokens_seen": 2280128512
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001315065254397579,
      "loss": 2.7273,
      "theoretical_loss": 3.3935727010491274,
      "tokens_seen": 2280194048
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013148997541138641,
      "loss": 3.0208,
      "theoretical_loss": 3.3935647769342596,
      "tokens_seen": 2280259584
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013147342538301495,
      "loss": 2.7344,
      "theoretical_loss": 3.3935568531108986,
      "tokens_seen": 2280325120
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013145687535464346,
      "loss": 2.4518,
      "theoretical_loss": 3.393548929579025,
      "tokens_seen": 2280390656
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000131440325326272,
      "loss": 2.358,
      "theoretical_loss": 3.3935410063386193,
      "tokens_seen": 2280456192
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001314237752979005,
      "loss": 2.3992,
      "theoretical_loss": 3.3935330833896638,
      "tokens_seen": 2280521728
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013140722526952905,
      "loss": 2.0534,
      "theoretical_loss": 3.3935251607321377,
      "tokens_seen": 2280587264
    },
    {
      "epoch": 0.81,
      "objective/train/avg_token_score": 0.017764180898666382,
      "objective/train/avg_weight": 0.9790843725204468,
      "objective/train/docs_used": 1285842,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.497668981552124,
      "objective/train/std_weight": 0.09613188356161118,
      "objective/train/theoretical_loss": 3.393517238366023,
      "objective/train/tokens_used": 2301112800,
      "theoretical_loss": 3.393517238366023,
      "tokens_seen": 2280652800
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013139067524115756,
      "loss": 2.5325,
      "theoretical_loss": 3.393517238366023,
      "tokens_seen": 2280652800
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001313741252127861,
      "loss": 2.6069,
      "theoretical_loss": 3.3935093162913,
      "tokens_seen": 2280718336
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013135757518441458,
      "loss": 2.2055,
      "theoretical_loss": 3.3935013945079504,
      "tokens_seen": 2280783872
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013134102515604312,
      "loss": 2.3953,
      "theoretical_loss": 3.3934934730159547,
      "tokens_seen": 2280849408
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013132447512767163,
      "loss": 2.5466,
      "theoretical_loss": 3.3934855518152935,
      "tokens_seen": 2280914944
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013130792509930016,
      "loss": 2.6152,
      "theoretical_loss": 3.393477630905948,
      "tokens_seen": 2280980480
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013129137507092868,
      "loss": 2.3942,
      "theoretical_loss": 3.393469710287899,
      "tokens_seen": 2281046016
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001312748250425572,
      "loss": 2.8286,
      "theoretical_loss": 3.3934617899611275,
      "tokens_seen": 2281111552
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013125827501418572,
      "loss": 2.5902,
      "theoretical_loss": 3.3934538699256143,
      "tokens_seen": 2281177088
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013124172498581426,
      "loss": 2.7759,
      "theoretical_loss": 3.393445950181341,
      "tokens_seen": 2281242624
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001312251749574428,
      "loss": 2.5589,
      "theoretical_loss": 3.393438030728287,
      "tokens_seen": 2281308160
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001312086249290713,
      "loss": 2.7658,
      "theoretical_loss": 3.393430111566435,
      "tokens_seen": 2281373696
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013119207490069985,
      "loss": 2.813,
      "theoretical_loss": 3.3934221926957653,
      "tokens_seen": 2281439232
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013117552487232836,
      "loss": 2.4869,
      "theoretical_loss": 3.393414274116258,
      "tokens_seen": 2281504768
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001311589748439569,
      "loss": 2.2864,
      "theoretical_loss": 3.393406355827895,
      "tokens_seen": 2281570304
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001311424248155854,
      "loss": 2.4317,
      "theoretical_loss": 3.393398437830657,
      "tokens_seen": 2281635840
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013112587478721392,
      "loss": 2.6616,
      "theoretical_loss": 3.3933905201245245,
      "tokens_seen": 2281701376
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013110932475884243,
      "loss": 2.2112,
      "theoretical_loss": 3.393382602709479,
      "tokens_seen": 2281766912
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013109277473047096,
      "loss": 2.3028,
      "theoretical_loss": 3.3933746855855014,
      "tokens_seen": 2281832448
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.00013107622470209947,
      "loss": 2.6886,
      "theoretical_loss": 3.393366768752572,
      "tokens_seen": 2281897984
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.000131059674673728,
      "loss": 2.6195,
      "theoretical_loss": 3.393358852210673,
      "tokens_seen": 2281963520
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013104312464535652,
      "loss": 2.7838,
      "theoretical_loss": 3.3933509359597838,
      "tokens_seen": 2282029056
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013102657461698506,
      "loss": 2.4923,
      "theoretical_loss": 3.393343019999886,
      "tokens_seen": 2282094592
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013101002458861357,
      "loss": 2.6253,
      "theoretical_loss": 3.393335104330961,
      "tokens_seen": 2282160128
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001309934745602421,
      "loss": 2.6788,
      "theoretical_loss": 3.3933271889529895,
      "tokens_seen": 2282225664
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013097692453187062,
      "loss": 2.302,
      "theoretical_loss": 3.393319273865952,
      "tokens_seen": 2282291200
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013096037450349915,
      "loss": 1.9128,
      "theoretical_loss": 3.3933113590698305,
      "tokens_seen": 2282356736
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013094382447512767,
      "loss": 2.4564,
      "theoretical_loss": 3.3933034445646046,
      "tokens_seen": 2282422272
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001309272744467562,
      "loss": 2.547,
      "theoretical_loss": 3.393295530350256,
      "tokens_seen": 2282487808
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001309107244183847,
      "loss": 2.3389,
      "theoretical_loss": 3.3932876164267656,
      "tokens_seen": 2282553344
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013089417439001325,
      "loss": 2.7176,
      "theoretical_loss": 3.3932797027941146,
      "tokens_seen": 2282618880
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013087762436164176,
      "loss": 2.4575,
      "theoretical_loss": 3.3932717894522835,
      "tokens_seen": 2282684416
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013086107433327027,
      "loss": 2.6736,
      "theoretical_loss": 3.393263876401253,
      "tokens_seen": 2282749952
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001308445243048988,
      "loss": 2.6128,
      "theoretical_loss": 3.393255963641005,
      "tokens_seen": 2282815488
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013082797427652732,
      "loss": 2.4028,
      "theoretical_loss": 3.39324805117152,
      "tokens_seen": 2282881024
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013081142424815586,
      "loss": 2.7431,
      "theoretical_loss": 3.393240138992779,
      "tokens_seen": 2282946560
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013079487421978437,
      "loss": 2.5005,
      "theoretical_loss": 3.393232227104763,
      "tokens_seen": 2283012096
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001307783241914129,
      "loss": 2.5305,
      "theoretical_loss": 3.393224315507452,
      "tokens_seen": 2283077632
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013076177416304142,
      "loss": 2.7126,
      "theoretical_loss": 3.393216404200829,
      "tokens_seen": 2283143168
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013074522413466995,
      "loss": 1.4261,
      "theoretical_loss": 3.3932084931848734,
      "tokens_seen": 2283208704
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013072867410629846,
      "loss": 2.5065,
      "theoretical_loss": 3.3932005824595666,
      "tokens_seen": 2283274240
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000130712124077927,
      "loss": 2.5016,
      "theoretical_loss": 3.3931926720248895,
      "tokens_seen": 2283339776
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001306955740495555,
      "loss": 2.3757,
      "theoretical_loss": 3.3931847618808235,
      "tokens_seen": 2283405312
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013067902402118405,
      "loss": 2.4435,
      "theoretical_loss": 3.393176852027349,
      "tokens_seen": 2283470848
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013066247399281256,
      "loss": 2.7632,
      "theoretical_loss": 3.393168942464447,
      "tokens_seen": 2283536384
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001306459239644411,
      "loss": 2.6189,
      "theoretical_loss": 3.3931610331920994,
      "tokens_seen": 2283601920
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013062937393606958,
      "loss": 2.5042,
      "theoretical_loss": 3.3931531242102864,
      "tokens_seen": 2283667456
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013061282390769812,
      "loss": 2.6093,
      "theoretical_loss": 3.393145215518989,
      "tokens_seen": 2283732992
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013059627387932666,
      "loss": 2.7277,
      "theoretical_loss": 3.393137307118188,
      "tokens_seen": 2283798528
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013057972385095517,
      "loss": 2.7348,
      "theoretical_loss": 3.393129399007865,
      "tokens_seen": 2283864064
    },
    {
      "epoch": 0.82,
      "objective/train/avg_token_score": 0.019550718367099762,
      "objective/train/avg_weight": 0.9769809246063232,
      "objective/train/docs_used": 1287535,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.171635627746582,
      "objective/train/std_weight": 0.10565879195928574,
      "objective/train/theoretical_loss": 3.3931214911880008,
      "objective/train/tokens_used": 2304389600,
      "theoretical_loss": 3.3931214911880008,
      "tokens_seen": 2283929600
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001305631738225837,
      "loss": 2.588,
      "theoretical_loss": 3.3931214911880008,
      "tokens_seen": 2283929600
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013054662379421221,
      "loss": 2.6754,
      "theoretical_loss": 3.393113583658576,
      "tokens_seen": 2283995136
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013053007376584075,
      "loss": 2.7839,
      "theoretical_loss": 3.393105676419572,
      "tokens_seen": 2284060672
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013051352373746926,
      "loss": 2.5245,
      "theoretical_loss": 3.3930977694709696,
      "tokens_seen": 2284126208
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001304969737090978,
      "loss": 2.6964,
      "theoretical_loss": 3.3930898628127504,
      "tokens_seen": 2284191744
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001304804236807263,
      "loss": 1.8955,
      "theoretical_loss": 3.3930819564448944,
      "tokens_seen": 2284257280
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013046387365235485,
      "loss": 2.4968,
      "theoretical_loss": 3.3930740503673835,
      "tokens_seen": 2284322816
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013044732362398336,
      "loss": 2.2249,
      "theoretical_loss": 3.3930661445801977,
      "tokens_seen": 2284388352
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001304307735956119,
      "loss": 2.4543,
      "theoretical_loss": 3.3930582390833193,
      "tokens_seen": 2284453888
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001304142235672404,
      "loss": 2.675,
      "theoretical_loss": 3.3930503338767286,
      "tokens_seen": 2284519424
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013039767353886894,
      "loss": 2.4386,
      "theoretical_loss": 3.393042428960406,
      "tokens_seen": 2284584960
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013038112351049743,
      "loss": 2.6205,
      "theoretical_loss": 3.393034524334334,
      "tokens_seen": 2284650496
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013036457348212596,
      "loss": 2.4426,
      "theoretical_loss": 3.3930266199984924,
      "tokens_seen": 2284716032
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013034802345375447,
      "loss": 2.0744,
      "theoretical_loss": 3.3930187159528624,
      "tokens_seen": 2284781568
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000130331473425383,
      "loss": 2.638,
      "theoretical_loss": 3.393010812197425,
      "tokens_seen": 2284847104
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013031492339701152,
      "loss": 2.7186,
      "theoretical_loss": 3.3930029087321625,
      "tokens_seen": 2284912640
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013029837336864006,
      "loss": 2.6169,
      "theoretical_loss": 3.3929950055570544,
      "tokens_seen": 2284978176
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013028182334026857,
      "loss": 2.0807,
      "theoretical_loss": 3.3929871026720817,
      "tokens_seen": 2285043712
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001302652733118971,
      "loss": 2.4115,
      "theoretical_loss": 3.3929792000772263,
      "tokens_seen": 2285109248
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013024872328352562,
      "loss": 2.4325,
      "theoretical_loss": 3.392971297772469,
      "tokens_seen": 2285174784
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013023217325515416,
      "loss": 2.5254,
      "theoretical_loss": 3.3929633957577905,
      "tokens_seen": 2285240320
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013021562322678267,
      "loss": 2.6218,
      "theoretical_loss": 3.3929554940331723,
      "tokens_seen": 2285305856
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001301990731984112,
      "loss": 2.8291,
      "theoretical_loss": 3.392947592598595,
      "tokens_seen": 2285371392
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013018252317003971,
      "loss": 2.5603,
      "theoretical_loss": 3.3929396914540395,
      "tokens_seen": 2285436928
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013016597314166825,
      "loss": 2.4742,
      "theoretical_loss": 3.3929317905994876,
      "tokens_seen": 2285502464
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013014942311329676,
      "loss": 2.4452,
      "theoretical_loss": 3.3929238900349197,
      "tokens_seen": 2285568000
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013013287308492527,
      "loss": 2.4059,
      "theoretical_loss": 3.3929159897603167,
      "tokens_seen": 2285633536
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001301163230565538,
      "loss": 2.6087,
      "theoretical_loss": 3.3929080897756605,
      "tokens_seen": 2285699072
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013009977302818232,
      "loss": 2.0936,
      "theoretical_loss": 3.3929001900809315,
      "tokens_seen": 2285764608
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013008322299981086,
      "loss": 2.7023,
      "theoretical_loss": 3.3928922906761105,
      "tokens_seen": 2285830144
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013006667297143937,
      "loss": 2.7048,
      "theoretical_loss": 3.392884391561179,
      "tokens_seen": 2285895680
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001300501229430679,
      "loss": 2.4931,
      "theoretical_loss": 3.3928764927361184,
      "tokens_seen": 2285961216
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013003357291469642,
      "loss": 2.575,
      "theoretical_loss": 3.392868594200909,
      "tokens_seen": 2286026752
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013001702288632495,
      "loss": 2.75,
      "theoretical_loss": 3.392860695955532,
      "tokens_seen": 2286092288
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00013000047285795347,
      "loss": 2.2913,
      "theoretical_loss": 3.3928527979999688,
      "tokens_seen": 2286157824
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000129983922829582,
      "loss": 2.6913,
      "theoretical_loss": 3.3928449003342003,
      "tokens_seen": 2286223360
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001299673728012105,
      "loss": 2.591,
      "theoretical_loss": 3.3928370029582076,
      "tokens_seen": 2286288896
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012995082277283905,
      "loss": 2.7348,
      "theoretical_loss": 3.3928291058719715,
      "tokens_seen": 2286354432
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012993427274446756,
      "loss": 2.7543,
      "theoretical_loss": 3.3928212090754735,
      "tokens_seen": 2286419968
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001299177227160961,
      "loss": 2.6279,
      "theoretical_loss": 3.3928133125686943,
      "tokens_seen": 2286485504
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001299011726877246,
      "loss": 2.2057,
      "theoretical_loss": 3.3928054163516155,
      "tokens_seen": 2286551040
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012988462265935312,
      "loss": 2.5415,
      "theoretical_loss": 3.392797520424217,
      "tokens_seen": 2286616576
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012986807263098166,
      "loss": 2.5047,
      "theoretical_loss": 3.3927896247864813,
      "tokens_seen": 2286682112
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012985152260261017,
      "loss": 2.5363,
      "theoretical_loss": 3.392781729438388,
      "tokens_seen": 2286747648
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001298349725742387,
      "loss": 2.7029,
      "theoretical_loss": 3.39277383437992,
      "tokens_seen": 2286813184
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012981842254586722,
      "loss": 2.5121,
      "theoretical_loss": 3.3927659396110568,
      "tokens_seen": 2286878720
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012980187251749575,
      "loss": 2.3447,
      "theoretical_loss": 3.39275804513178,
      "tokens_seen": 2286944256
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012978532248912426,
      "loss": 2.5965,
      "theoretical_loss": 3.3927501509420708,
      "tokens_seen": 2287009792
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001297687724607528,
      "loss": 2.1848,
      "theoretical_loss": 3.3927422570419106,
      "tokens_seen": 2287075328
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001297522224323813,
      "loss": 2.5655,
      "theoretical_loss": 3.3927343634312797,
      "tokens_seen": 2287140864
    },
    {
      "epoch": 0.82,
      "objective/train/avg_token_score": 0.011733541265130043,
      "objective/train/avg_weight": 0.9861856698989868,
      "objective/train/docs_used": 1288943,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.644590139389038,
      "objective/train/std_weight": 0.07403979450464249,
      "objective/train/theoretical_loss": 3.3927264701101594,
      "objective/train/tokens_used": 2307666400,
      "theoretical_loss": 3.3927264701101594,
      "tokens_seen": 2287206400
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012973567240400985,
      "loss": 2.4453,
      "theoretical_loss": 3.3927264701101594,
      "tokens_seen": 2287206400
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012971912237563836,
      "loss": 2.8847,
      "theoretical_loss": 3.392718577078531,
      "tokens_seen": 2287271936
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001297025723472669,
      "loss": 2.6621,
      "theoretical_loss": 3.392710684336376,
      "tokens_seen": 2287337472
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001296860223188954,
      "loss": 2.5894,
      "theoretical_loss": 3.3927027918836745,
      "tokens_seen": 2287403008
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012966947229052394,
      "loss": 2.023,
      "theoretical_loss": 3.3926948997204085,
      "tokens_seen": 2287468544
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012965292226215246,
      "loss": 2.7224,
      "theoretical_loss": 3.3926870078465585,
      "tokens_seen": 2287534080
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012963637223378097,
      "loss": 2.5775,
      "theoretical_loss": 3.392679116262106,
      "tokens_seen": 2287599616
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012961982220540948,
      "loss": 2.7697,
      "theoretical_loss": 3.3926712249670317,
      "tokens_seen": 2287665152
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012960327217703801,
      "loss": 2.7767,
      "theoretical_loss": 3.392663333961317,
      "tokens_seen": 2287730688
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012958672214866652,
      "loss": 2.4932,
      "theoretical_loss": 3.392655443244943,
      "tokens_seen": 2287796224
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012957017212029506,
      "loss": 2.6614,
      "theoretical_loss": 3.3926475528178908,
      "tokens_seen": 2287861760
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012955362209192357,
      "loss": 2.5566,
      "theoretical_loss": 3.392639662680141,
      "tokens_seen": 2287927296
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001295370720635521,
      "loss": 2.7817,
      "theoretical_loss": 3.3926317728316757,
      "tokens_seen": 2287992832
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012952052203518062,
      "loss": 2.7416,
      "theoretical_loss": 3.392623883272475,
      "tokens_seen": 2288058368
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012950397200680916,
      "loss": 2.2788,
      "theoretical_loss": 3.3926159940025205,
      "tokens_seen": 2288123904
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012948742197843767,
      "loss": 2.6273,
      "theoretical_loss": 3.3926081050217936,
      "tokens_seen": 2288189440
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001294708719500662,
      "loss": 2.5802,
      "theoretical_loss": 3.392600216330275,
      "tokens_seen": 2288254976
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012945432192169472,
      "loss": 2.7327,
      "theoretical_loss": 3.3925923279279453,
      "tokens_seen": 2288320512
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012943777189332325,
      "loss": 2.5233,
      "theoretical_loss": 3.3925844398147866,
      "tokens_seen": 2288386048
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012942122186495176,
      "loss": 2.4767,
      "theoretical_loss": 3.39257655199078,
      "tokens_seen": 2288451584
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012940467183658027,
      "loss": 2.406,
      "theoretical_loss": 3.3925686644559057,
      "tokens_seen": 2288517120
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001293881218082088,
      "loss": 2.6496,
      "theoretical_loss": 3.3925607772101456,
      "tokens_seen": 2288582656
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012937157177983732,
      "loss": 2.6237,
      "theoretical_loss": 3.39255289025348,
      "tokens_seen": 2288648192
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012935502175146586,
      "loss": 2.4712,
      "theoretical_loss": 3.3925450035858913,
      "tokens_seen": 2288713728
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012933847172309437,
      "loss": 2.6584,
      "theoretical_loss": 3.3925371172073597,
      "tokens_seen": 2288779264
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001293219216947229,
      "loss": 2.3346,
      "theoretical_loss": 3.3925292311178668,
      "tokens_seen": 2288844800
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012930537166635142,
      "loss": 2.4997,
      "theoretical_loss": 3.3925213453173932,
      "tokens_seen": 2288910336
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012928882163797996,
      "loss": 2.6927,
      "theoretical_loss": 3.3925134598059206,
      "tokens_seen": 2288975872
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012927227160960847,
      "loss": 2.0776,
      "theoretical_loss": 3.3925055745834296,
      "tokens_seen": 2289041408
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000129255721581237,
      "loss": 2.6998,
      "theoretical_loss": 3.392497689649902,
      "tokens_seen": 2289106944
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012923917155286551,
      "loss": 2.4819,
      "theoretical_loss": 3.3924898050053183,
      "tokens_seen": 2289172480
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012922262152449405,
      "loss": 2.4592,
      "theoretical_loss": 3.3924819206496597,
      "tokens_seen": 2289238016
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012920607149612256,
      "loss": 2.7082,
      "theoretical_loss": 3.392474036582908,
      "tokens_seen": 2289303552
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001291895214677511,
      "loss": 2.764,
      "theoretical_loss": 3.3924661528050435,
      "tokens_seen": 2289369088
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001291729714393796,
      "loss": 2.4788,
      "theoretical_loss": 3.3924582693160477,
      "tokens_seen": 2289434624
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012915642141100812,
      "loss": 2.6083,
      "theoretical_loss": 3.3924503861159017,
      "tokens_seen": 2289500160
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012913987138263666,
      "loss": 2.5471,
      "theoretical_loss": 3.392442503204587,
      "tokens_seen": 2289565696
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012912332135426517,
      "loss": 2.2496,
      "theoretical_loss": 3.392434620582084,
      "tokens_seen": 2289631232
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001291067713258937,
      "loss": 2.4681,
      "theoretical_loss": 3.392426738248375,
      "tokens_seen": 2289696768
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012909022129752222,
      "loss": 2.7362,
      "theoretical_loss": 3.39241885620344,
      "tokens_seen": 2289762304
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012907367126915075,
      "loss": 2.6216,
      "theoretical_loss": 3.3924109744472606,
      "tokens_seen": 2289827840
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012905712124077926,
      "loss": 2.7385,
      "theoretical_loss": 3.3924030929798183,
      "tokens_seen": 2289893376
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001290405712124078,
      "loss": 2.6276,
      "theoretical_loss": 3.3923952118010936,
      "tokens_seen": 2289958912
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001290240211840363,
      "loss": 2.4631,
      "theoretical_loss": 3.392387330911068,
      "tokens_seen": 2290024448
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012900747115566485,
      "loss": 2.7283,
      "theoretical_loss": 3.392379450309723,
      "tokens_seen": 2290089984
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012899092112729336,
      "loss": 2.6895,
      "theoretical_loss": 3.392371569997039,
      "tokens_seen": 2290155520
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001289743710989219,
      "loss": 2.8538,
      "theoretical_loss": 3.392363689972998,
      "tokens_seen": 2290221056
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001289578210705504,
      "loss": 2.6013,
      "theoretical_loss": 3.3923558102375804,
      "tokens_seen": 2290286592
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012894127104217895,
      "loss": 2.5397,
      "theoretical_loss": 3.392347930790768,
      "tokens_seen": 2290352128
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012892472101380746,
      "loss": 2.7633,
      "theoretical_loss": 3.3923400516325417,
      "tokens_seen": 2290417664
    },
    {
      "epoch": 0.82,
      "objective/train/avg_token_score": 0.016381211578845978,
      "objective/train/avg_weight": 0.9807253479957581,
      "objective/train/docs_used": 1290666,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.275171995162964,
      "objective/train/std_weight": 0.07478269189596176,
      "objective/train/theoretical_loss": 3.392332172762883,
      "objective/train/tokens_used": 2310943200,
      "theoretical_loss": 3.392332172762883,
      "tokens_seen": 2290483200
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012890817098543597,
      "loss": 2.5231,
      "theoretical_loss": 3.392332172762883,
      "tokens_seen": 2290483200
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012889162095706448,
      "loss": 2.3845,
      "theoretical_loss": 3.3923242941817717,
      "tokens_seen": 2290548736
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012887507092869302,
      "loss": 2.562,
      "theoretical_loss": 3.3923164158891908,
      "tokens_seen": 2290614272
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012885852090032153,
      "loss": 2.6107,
      "theoretical_loss": 3.3923085378851208,
      "tokens_seen": 2290679808
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012884197087195006,
      "loss": 2.3251,
      "theoretical_loss": 3.3923006601695427,
      "tokens_seen": 2290745344
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012882542084357857,
      "loss": 2.6018,
      "theoretical_loss": 3.3922927827424374,
      "tokens_seen": 2290810880
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001288088708152071,
      "loss": 2.7145,
      "theoretical_loss": 3.3922849056037867,
      "tokens_seen": 2290876416
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012879232078683562,
      "loss": 2.0791,
      "theoretical_loss": 3.392277028753572,
      "tokens_seen": 2290941952
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012877577075846416,
      "loss": 2.7487,
      "theoretical_loss": 3.3922691521917736,
      "tokens_seen": 2291007488
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012875922073009267,
      "loss": 2.4532,
      "theoretical_loss": 3.392261275918373,
      "tokens_seen": 2291073024
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001287426707017212,
      "loss": 2.5901,
      "theoretical_loss": 3.3922533999333515,
      "tokens_seen": 2291138560
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012872612067334972,
      "loss": 2.4761,
      "theoretical_loss": 3.3922455242366905,
      "tokens_seen": 2291204096
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012870957064497826,
      "loss": 2.5978,
      "theoretical_loss": 3.392237648828371,
      "tokens_seen": 2291269632
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012869302061660677,
      "loss": 2.3963,
      "theoretical_loss": 3.392229773708374,
      "tokens_seen": 2291335168
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001286764705882353,
      "loss": 2.6948,
      "theoretical_loss": 3.3922218988766812,
      "tokens_seen": 2291400704
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012865992055986381,
      "loss": 2.676,
      "theoretical_loss": 3.3922140243332732,
      "tokens_seen": 2291466240
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012864337053149232,
      "loss": 2.5813,
      "theoretical_loss": 3.3922061500781315,
      "tokens_seen": 2291531776
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012862682050312086,
      "loss": 2.3416,
      "theoretical_loss": 3.3921982761112375,
      "tokens_seen": 2291597312
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012861027047474937,
      "loss": 2.6233,
      "theoretical_loss": 3.392190402432572,
      "tokens_seen": 2291662848
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001285937204463779,
      "loss": 2.5509,
      "theoretical_loss": 3.392182529042117,
      "tokens_seen": 2291728384
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012857717041800642,
      "loss": 2.3822,
      "theoretical_loss": 3.3921746559398525,
      "tokens_seen": 2291793920
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012856062038963496,
      "loss": 2.6039,
      "theoretical_loss": 3.3921667831257607,
      "tokens_seen": 2291859456
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012854407036126347,
      "loss": 2.7145,
      "theoretical_loss": 3.392158910599822,
      "tokens_seen": 2291924992
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000128527520332892,
      "loss": 2.5058,
      "theoretical_loss": 3.3921510383620186,
      "tokens_seen": 2291990528
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012851097030452052,
      "loss": 2.3586,
      "theoretical_loss": 3.3921431664123305,
      "tokens_seen": 2292056064
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012849442027614905,
      "loss": 2.4759,
      "theoretical_loss": 3.39213529475074,
      "tokens_seen": 2292121600
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012847787024777756,
      "loss": 2.5447,
      "theoretical_loss": 3.392127423377228,
      "tokens_seen": 2292187136
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001284613202194061,
      "loss": 2.6631,
      "theoretical_loss": 3.3921195522917755,
      "tokens_seen": 2292252672
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001284447701910346,
      "loss": 2.5887,
      "theoretical_loss": 3.392111681494364,
      "tokens_seen": 2292318208
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012842822016266315,
      "loss": 2.5536,
      "theoretical_loss": 3.3921038109849744,
      "tokens_seen": 2292383744
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012841167013429166,
      "loss": 2.5131,
      "theoretical_loss": 3.3920959407635882,
      "tokens_seen": 2292449280
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012839512010592017,
      "loss": 2.7926,
      "theoretical_loss": 3.3920880708301864,
      "tokens_seen": 2292514816
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001283785700775487,
      "loss": 2.5428,
      "theoretical_loss": 3.3920802011847506,
      "tokens_seen": 2292580352
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012836202004917722,
      "loss": 1.8529,
      "theoretical_loss": 3.3920723318272614,
      "tokens_seen": 2292645888
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012834547002080576,
      "loss": 2.472,
      "theoretical_loss": 3.392064462757701,
      "tokens_seen": 2292711424
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012832891999243427,
      "loss": 2.2487,
      "theoretical_loss": 3.39205659397605,
      "tokens_seen": 2292776960
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001283123699640628,
      "loss": 2.4742,
      "theoretical_loss": 3.3920487254822893,
      "tokens_seen": 2292842496
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012829581993569131,
      "loss": 2.6346,
      "theoretical_loss": 3.3920408572764007,
      "tokens_seen": 2292908032
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012827926990731985,
      "loss": 2.6106,
      "theoretical_loss": 3.392032989358365,
      "tokens_seen": 2292973568
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012826271987894836,
      "loss": 2.4629,
      "theoretical_loss": 3.392025121728164,
      "tokens_seen": 2293039104
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001282461698505769,
      "loss": 2.642,
      "theoretical_loss": 3.3920172543857787,
      "tokens_seen": 2293104640
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001282296198222054,
      "loss": 2.5315,
      "theoretical_loss": 3.39200938733119,
      "tokens_seen": 2293170176
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012821306979383395,
      "loss": 2.9154,
      "theoretical_loss": 3.39200152056438,
      "tokens_seen": 2293235712
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012819651976546246,
      "loss": 2.4592,
      "theoretical_loss": 3.3919936540853293,
      "tokens_seen": 2293301248
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000128179969737091,
      "loss": 2.5161,
      "theoretical_loss": 3.3919857878940194,
      "tokens_seen": 2293366784
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012816341970871948,
      "loss": 2.5544,
      "theoretical_loss": 3.391977921990431,
      "tokens_seen": 2293432320
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012814686968034802,
      "loss": 2.3175,
      "theoretical_loss": 3.3919700563745456,
      "tokens_seen": 2293497856
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012813031965197653,
      "loss": 2.4492,
      "theoretical_loss": 3.3919621910463453,
      "tokens_seen": 2293563392
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012811376962360506,
      "loss": 2.1967,
      "theoretical_loss": 3.39195432600581,
      "tokens_seen": 2293628928
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012809721959523358,
      "loss": 2.7121,
      "theoretical_loss": 3.391946461252922,
      "tokens_seen": 2293694464
    },
    {
      "debugging/Self-BLEU-5": 0.515096219338814,
      "debugging/distinct-1-grams": 0.7923118197529954,
      "debugging/distinct-2-grams": 0.9542743672238616,
      "debugging/entropy-1-grams": 6.147062465574793,
      "debugging/entropy-2-grams": 7.107820582543805,
      "debugging/length": 504.1666666666667,
      "debugging/num_segments": 18,
      "epoch": 0.82,
      "objective/train/avg_token_score": 0.012544393539428711,
      "objective/train/avg_weight": 0.9852434396743774,
      "objective/train/docs_used": 1292636,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.770737886428833,
      "objective/train/std_weight": 0.05353361740708351,
      "objective/train/theoretical_loss": 3.3919385967876625,
      "objective/train/tokens_used": 2314220000,
      "theoretical_loss": 3.3919385967876625,
      "tokens_seen": 2293760000
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001280806695668621,
      "loss": 2.5196,
      "theoretical_loss": 3.3919385967876625,
      "tokens_seen": 2293760000
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012806411953849062,
      "loss": 2.5693,
      "theoretical_loss": 3.391930732610012,
      "tokens_seen": 2293825536
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012804756951011916,
      "loss": 2.5067,
      "theoretical_loss": 3.3919228687199525,
      "tokens_seen": 2293891072
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012803101948174767,
      "loss": 2.6134,
      "theoretical_loss": 3.3919150051174647,
      "tokens_seen": 2293956608
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001280144694533762,
      "loss": 2.7212,
      "theoretical_loss": 3.3919071418025304,
      "tokens_seen": 2294022144
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012799791942500472,
      "loss": 2.3673,
      "theoretical_loss": 3.3918992787751305,
      "tokens_seen": 2294087680
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012798136939663326,
      "loss": 2.4399,
      "theoretical_loss": 3.391891416035247,
      "tokens_seen": 2294153216
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012796481936826177,
      "loss": 2.7548,
      "theoretical_loss": 3.39188355358286,
      "tokens_seen": 2294218752
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001279482693398903,
      "loss": 2.6361,
      "theoretical_loss": 3.3918756914179515,
      "tokens_seen": 2294284288
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012793171931151882,
      "loss": 2.5957,
      "theoretical_loss": 3.3918678295405025,
      "tokens_seen": 2294349824
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012791516928314733,
      "loss": 2.5192,
      "theoretical_loss": 3.3918599679504946,
      "tokens_seen": 2294415360
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012789861925477586,
      "loss": 2.4542,
      "theoretical_loss": 3.391852106647909,
      "tokens_seen": 2294480896
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012788206922640437,
      "loss": 2.5258,
      "theoretical_loss": 3.3918442456327265,
      "tokens_seen": 2294546432
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001278655191980329,
      "loss": 2.5031,
      "theoretical_loss": 3.391836384904929,
      "tokens_seen": 2294611968
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012784896916966142,
      "loss": 2.4647,
      "theoretical_loss": 3.3918285244644975,
      "tokens_seen": 2294677504
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012783241914128996,
      "loss": 3.0967,
      "theoretical_loss": 3.3918206643114135,
      "tokens_seen": 2294743040
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012781586911291847,
      "loss": 2.5392,
      "theoretical_loss": 3.391812804445658,
      "tokens_seen": 2294808576
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000127799319084547,
      "loss": 2.7662,
      "theoretical_loss": 3.3918049448672125,
      "tokens_seen": 2294874112
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012778276905617552,
      "loss": 2.7123,
      "theoretical_loss": 3.391797085576058,
      "tokens_seen": 2294939648
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012776621902780405,
      "loss": 1.9585,
      "theoretical_loss": 3.3917892265721763,
      "tokens_seen": 2295005184
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012774966899943257,
      "loss": 2.6181,
      "theoretical_loss": 3.391781367855548,
      "tokens_seen": 2295070720
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001277331189710611,
      "loss": 2.7524,
      "theoretical_loss": 3.391773509426155,
      "tokens_seen": 2295136256
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001277165689426896,
      "loss": 2.9867,
      "theoretical_loss": 3.3917656512839787,
      "tokens_seen": 2295201792
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012770001891431815,
      "loss": 2.7232,
      "theoretical_loss": 3.391757793429,
      "tokens_seen": 2295267328
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012768346888594666,
      "loss": 2.7662,
      "theoretical_loss": 3.3917499358612,
      "tokens_seen": 2295332864
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012766691885757517,
      "loss": 2.658,
      "theoretical_loss": 3.391742078580561,
      "tokens_seen": 2295398400
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001276503688292037,
      "loss": 2.4436,
      "theoretical_loss": 3.391734221587063,
      "tokens_seen": 2295463936
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012763381880083222,
      "loss": 2.4641,
      "theoretical_loss": 3.391726364880688,
      "tokens_seen": 2295529472
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012761726877246076,
      "loss": 2.6788,
      "theoretical_loss": 3.3917185084614174,
      "tokens_seen": 2295595008
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012760071874408927,
      "loss": 2.5179,
      "theoretical_loss": 3.391710652329232,
      "tokens_seen": 2295660544
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001275841687157178,
      "loss": 2.8604,
      "theoretical_loss": 3.391702796484114,
      "tokens_seen": 2295726080
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012756761868734632,
      "loss": 2.665,
      "theoretical_loss": 3.391694940926044,
      "tokens_seen": 2295791616
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012755106865897485,
      "loss": 2.7052,
      "theoretical_loss": 3.3916870856550037,
      "tokens_seen": 2295857152
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012753451863060336,
      "loss": 2.6256,
      "theoretical_loss": 3.3916792306709738,
      "tokens_seen": 2295922688
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001275179686022319,
      "loss": 2.6098,
      "theoretical_loss": 3.391671375973936,
      "tokens_seen": 2295988224
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001275014185738604,
      "loss": 2.6843,
      "theoretical_loss": 3.3916635215638724,
      "tokens_seen": 2296053760
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012748486854548895,
      "loss": 2.4612,
      "theoretical_loss": 3.391655667440763,
      "tokens_seen": 2296119296
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012746831851711746,
      "loss": 2.6191,
      "theoretical_loss": 3.39164781360459,
      "tokens_seen": 2296184832
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000127451768488746,
      "loss": 2.702,
      "theoretical_loss": 3.391639960055334,
      "tokens_seen": 2296250368
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012743521846037448,
      "loss": 2.4651,
      "theoretical_loss": 3.391632106792977,
      "tokens_seen": 2296315904
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012741866843200302,
      "loss": 2.6426,
      "theoretical_loss": 3.3916242538175005,
      "tokens_seen": 2296381440
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012740211840363153,
      "loss": 2.5095,
      "theoretical_loss": 3.3916164011288847,
      "tokens_seen": 2296446976
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012738556837526007,
      "loss": 2.7853,
      "theoretical_loss": 3.3916085487271124,
      "tokens_seen": 2296512512
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012736901834688858,
      "loss": 2.8854,
      "theoretical_loss": 3.391600696612164,
      "tokens_seen": 2296578048
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012735246831851711,
      "loss": 2.752,
      "theoretical_loss": 3.3915928447840207,
      "tokens_seen": 2296643584
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012733591829014562,
      "loss": 2.511,
      "theoretical_loss": 3.3915849932426645,
      "tokens_seen": 2296709120
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012731936826177416,
      "loss": 2.5878,
      "theoretical_loss": 3.3915771419880762,
      "tokens_seen": 2296774656
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012730281823340267,
      "loss": 2.6202,
      "theoretical_loss": 3.3915692910202377,
      "tokens_seen": 2296840192
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001272862682050312,
      "loss": 2.3862,
      "theoretical_loss": 3.39156144033913,
      "tokens_seen": 2296905728
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012726971817665972,
      "loss": 2.5127,
      "theoretical_loss": 3.3915535899447344,
      "tokens_seen": 2296971264
    },
    {
      "epoch": 0.82,
      "objective/train/avg_token_score": 0.01656987890601158,
      "objective/train/avg_weight": 0.9805304408073425,
      "objective/train/docs_used": 1294107,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.504160165786743,
      "objective/train/std_weight": 0.08566085994243622,
      "objective/train/theoretical_loss": 3.3915457398370323,
      "objective/train/tokens_used": 2317496800,
      "theoretical_loss": 3.3915457398370323,
      "tokens_seen": 2297036800
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012725316814828826,
      "loss": 2.4147,
      "theoretical_loss": 3.3915457398370323,
      "tokens_seen": 2297036800
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001272366181199168,
      "loss": 2.8233,
      "theoretical_loss": 3.391537890016005,
      "tokens_seen": 2297102336
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001272200680915453,
      "loss": 2.8211,
      "theoretical_loss": 3.3915300404816344,
      "tokens_seen": 2297167872
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012720351806317384,
      "loss": 2.6049,
      "theoretical_loss": 3.391522191233901,
      "tokens_seen": 2297233408
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012718696803480233,
      "loss": 2.5016,
      "theoretical_loss": 3.391514342272786,
      "tokens_seen": 2297298944
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012717041800643086,
      "loss": 2.6544,
      "theoretical_loss": 3.391506493598272,
      "tokens_seen": 2297364480
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012715386797805937,
      "loss": 2.4966,
      "theoretical_loss": 3.3914986452103397,
      "tokens_seen": 2297430016
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001271373179496879,
      "loss": 2.8262,
      "theoretical_loss": 3.39149079710897,
      "tokens_seen": 2297495552
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012712076792131642,
      "loss": 2.5005,
      "theoretical_loss": 3.391482949294145,
      "tokens_seen": 2297561088
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012710421789294496,
      "loss": 2.7037,
      "theoretical_loss": 3.3914751017658458,
      "tokens_seen": 2297626624
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012708766786457347,
      "loss": 2.4953,
      "theoretical_loss": 3.3914672545240534,
      "tokens_seen": 2297692160
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000127071117836202,
      "loss": 2.5966,
      "theoretical_loss": 3.3914594075687496,
      "tokens_seen": 2297757696
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012705456780783052,
      "loss": 2.8113,
      "theoretical_loss": 3.391451560899916,
      "tokens_seen": 2297823232
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012703801777945906,
      "loss": 2.7194,
      "theoretical_loss": 3.391443714517533,
      "tokens_seen": 2297888768
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012702146775108757,
      "loss": 2.6725,
      "theoretical_loss": 3.391435868421583,
      "tokens_seen": 2297954304
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001270049177227161,
      "loss": 2.4506,
      "theoretical_loss": 3.391428022612047,
      "tokens_seen": 2298019840
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012698836769434461,
      "loss": 2.1753,
      "theoretical_loss": 3.391420177088906,
      "tokens_seen": 2298085376
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012697181766597315,
      "loss": 2.7171,
      "theoretical_loss": 3.391412331852142,
      "tokens_seen": 2298150912
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012695526763760166,
      "loss": 2.5332,
      "theoretical_loss": 3.391404486901736,
      "tokens_seen": 2298216448
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012693871760923017,
      "loss": 2.7865,
      "theoretical_loss": 3.3913966422376696,
      "tokens_seen": 2298281984
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001269221675808587,
      "loss": 2.2699,
      "theoretical_loss": 3.391388797859924,
      "tokens_seen": 2298347520
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012690561755248722,
      "loss": 2.425,
      "theoretical_loss": 3.3913809537684805,
      "tokens_seen": 2298413056
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012688906752411576,
      "loss": 2.5848,
      "theoretical_loss": 3.391373109963321,
      "tokens_seen": 2298478592
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012687251749574427,
      "loss": 2.7007,
      "theoretical_loss": 3.391365266444426,
      "tokens_seen": 2298544128
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001268559674673728,
      "loss": 2.6704,
      "theoretical_loss": 3.3913574232117782,
      "tokens_seen": 2298609664
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012683941743900132,
      "loss": 2.5687,
      "theoretical_loss": 3.3913495802653575,
      "tokens_seen": 2298675200
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012682286741062985,
      "loss": 2.5728,
      "theoretical_loss": 3.3913417376051465,
      "tokens_seen": 2298740736
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012680631738225837,
      "loss": 2.7991,
      "theoretical_loss": 3.391333895231126,
      "tokens_seen": 2298806272
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001267897673538869,
      "loss": 2.6108,
      "theoretical_loss": 3.3913260531432776,
      "tokens_seen": 2298871808
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001267732173255154,
      "loss": 2.5691,
      "theoretical_loss": 3.3913182113415825,
      "tokens_seen": 2298937344
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012675666729714395,
      "loss": 1.8473,
      "theoretical_loss": 3.391310369826022,
      "tokens_seen": 2299002880
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012674011726877246,
      "loss": 2.7462,
      "theoretical_loss": 3.3913025285965777,
      "tokens_seen": 2299068416
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000126723567240401,
      "loss": 2.718,
      "theoretical_loss": 3.3912946876532315,
      "tokens_seen": 2299133952
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001267070172120295,
      "loss": 2.5761,
      "theoretical_loss": 3.391286846995964,
      "tokens_seen": 2299199488
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012669046718365802,
      "loss": 2.7943,
      "theoretical_loss": 3.3912790066247567,
      "tokens_seen": 2299265024
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012667391715528653,
      "loss": 2.8136,
      "theoretical_loss": 3.3912711665395916,
      "tokens_seen": 2299330560
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012665736712691507,
      "loss": 2.5702,
      "theoretical_loss": 3.3912633267404497,
      "tokens_seen": 2299396096
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012664081709854358,
      "loss": 2.7688,
      "theoretical_loss": 3.391255487227313,
      "tokens_seen": 2299461632
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012662426707017212,
      "loss": 2.6006,
      "theoretical_loss": 3.3912476480001614,
      "tokens_seen": 2299527168
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012660771704180065,
      "loss": 2.8165,
      "theoretical_loss": 3.3912398090589777,
      "tokens_seen": 2299592704
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012659116701342916,
      "loss": 2.8177,
      "theoretical_loss": 3.391231970403743,
      "tokens_seen": 2299658240
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001265746169850577,
      "loss": 2.6449,
      "theoretical_loss": 3.3912241320344387,
      "tokens_seen": 2299723776
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001265580669566862,
      "loss": 2.7482,
      "theoretical_loss": 3.3912162939510457,
      "tokens_seen": 2299789312
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012654151692831475,
      "loss": 2.731,
      "theoretical_loss": 3.3912084561535467,
      "tokens_seen": 2299854848
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012652496689994326,
      "loss": 2.8239,
      "theoretical_loss": 3.3912006186419217,
      "tokens_seen": 2299920384
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001265084168715718,
      "loss": 2.8579,
      "theoretical_loss": 3.391192781416153,
      "tokens_seen": 2299985920
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001264918668432003,
      "loss": 2.4701,
      "theoretical_loss": 3.3911849444762217,
      "tokens_seen": 2300051456
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012647531681482884,
      "loss": 2.6704,
      "theoretical_loss": 3.3911771078221093,
      "tokens_seen": 2300116992
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012645876678645736,
      "loss": 2.531,
      "theoretical_loss": 3.391169271453797,
      "tokens_seen": 2300182528
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012644221675808587,
      "loss": 2.7019,
      "theoretical_loss": 3.3911614353712665,
      "tokens_seen": 2300248064
    },
    {
      "epoch": 0.82,
      "objective/train/avg_token_score": 0.020872579887509346,
      "objective/train/avg_weight": 0.9754710793495178,
      "objective/train/docs_used": 1294624,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.863074779510498,
      "objective/train/std_weight": 0.08568324893712997,
      "objective/train/theoretical_loss": 3.3911535995744995,
      "objective/train/tokens_used": 2320773600,
      "theoretical_loss": 3.3911535995744995,
      "tokens_seen": 2300313600
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012642566672971438,
      "loss": 2.8555,
      "theoretical_loss": 3.3911535995744995,
      "tokens_seen": 2300313600
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012640911670134291,
      "loss": 2.6894,
      "theoretical_loss": 3.3911457640634772,
      "tokens_seen": 2300379136
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012639256667297142,
      "loss": 2.8404,
      "theoretical_loss": 3.3911379288381807,
      "tokens_seen": 2300444672
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012637601664459996,
      "loss": 2.6141,
      "theoretical_loss": 3.391130093898592,
      "tokens_seen": 2300510208
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012635946661622847,
      "loss": 2.6521,
      "theoretical_loss": 3.391122259244692,
      "tokens_seen": 2300575744
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000126342916587857,
      "loss": 2.8349,
      "theoretical_loss": 3.3911144248764624,
      "tokens_seen": 2300641280
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012632636655948552,
      "loss": 2.5932,
      "theoretical_loss": 3.3911065907938847,
      "tokens_seen": 2300706816
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012630981653111406,
      "loss": 2.87,
      "theoretical_loss": 3.3910987569969406,
      "tokens_seen": 2300772352
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012629326650274257,
      "loss": 2.6557,
      "theoretical_loss": 3.391090923485611,
      "tokens_seen": 2300837888
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001262767164743711,
      "loss": 2.9719,
      "theoretical_loss": 3.391083090259878,
      "tokens_seen": 2300903424
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012626016644599962,
      "loss": 2.7321,
      "theoretical_loss": 3.391075257319722,
      "tokens_seen": 2300968960
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012624361641762815,
      "loss": 2.6971,
      "theoretical_loss": 3.391067424665126,
      "tokens_seen": 2301034496
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012622706638925666,
      "loss": 2.7179,
      "theoretical_loss": 3.3910595922960693,
      "tokens_seen": 2301100032
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012621051636088517,
      "loss": 2.6173,
      "theoretical_loss": 3.3910517602125356,
      "tokens_seen": 2301165568
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001261939663325137,
      "loss": 2.9877,
      "theoretical_loss": 3.391043928414505,
      "tokens_seen": 2301231104
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012617741630414222,
      "loss": 2.7899,
      "theoretical_loss": 3.3910360969019595,
      "tokens_seen": 2301296640
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012616086627577076,
      "loss": 2.7972,
      "theoretical_loss": 3.3910282656748807,
      "tokens_seen": 2301362176
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012614431624739927,
      "loss": 2.431,
      "theoretical_loss": 3.391020434733249,
      "tokens_seen": 2301427712
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001261277662190278,
      "loss": 2.8214,
      "theoretical_loss": 3.3910126040770474,
      "tokens_seen": 2301493248
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012611121619065632,
      "loss": 2.6628,
      "theoretical_loss": 3.3910047737062565,
      "tokens_seen": 2301558784
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012609466616228486,
      "loss": 2.5538,
      "theoretical_loss": 3.390996943620858,
      "tokens_seen": 2301624320
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012607811613391337,
      "loss": 2.7702,
      "theoretical_loss": 3.3909891138208326,
      "tokens_seen": 2301689856
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001260615661055419,
      "loss": 2.8275,
      "theoretical_loss": 3.390981284306163,
      "tokens_seen": 2301755392
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012604501607717041,
      "loss": 2.8119,
      "theoretical_loss": 3.3909734550768302,
      "tokens_seen": 2301820928
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012602846604879895,
      "loss": 2.7603,
      "theoretical_loss": 3.3909656261328154,
      "tokens_seen": 2301886464
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012601191602042746,
      "loss": 2.6897,
      "theoretical_loss": 3.3909577974741003,
      "tokens_seen": 2301952000
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000125995365992056,
      "loss": 2.5803,
      "theoretical_loss": 3.390949969100666,
      "tokens_seen": 2302017536
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001259788159636845,
      "loss": 2.8434,
      "theoretical_loss": 3.390942141012495,
      "tokens_seen": 2302083072
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012596226593531302,
      "loss": 2.846,
      "theoretical_loss": 3.3909343132095673,
      "tokens_seen": 2302148608
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012594571590694156,
      "loss": 2.4212,
      "theoretical_loss": 3.3909264856918657,
      "tokens_seen": 2302214144
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012592916587857007,
      "loss": 2.6794,
      "theoretical_loss": 3.3909186584593716,
      "tokens_seen": 2302279680
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001259126158501986,
      "loss": 2.5008,
      "theoretical_loss": 3.3909108315120653,
      "tokens_seen": 2302345216
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012589606582182712,
      "loss": 2.7198,
      "theoretical_loss": 3.3909030048499296,
      "tokens_seen": 2302410752
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012587951579345565,
      "loss": 2.9819,
      "theoretical_loss": 3.3908951784729453,
      "tokens_seen": 2302476288
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012586296576508416,
      "loss": 2.8286,
      "theoretical_loss": 3.3908873523810943,
      "tokens_seen": 2302541824
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001258464157367127,
      "loss": 2.7451,
      "theoretical_loss": 3.3908795265743574,
      "tokens_seen": 2302607360
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001258298657083412,
      "loss": 2.7812,
      "theoretical_loss": 3.390871701052717,
      "tokens_seen": 2302672896
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012581331567996975,
      "loss": 2.6114,
      "theoretical_loss": 3.3908638758161542,
      "tokens_seen": 2302738432
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012579676565159826,
      "loss": 2.3017,
      "theoretical_loss": 3.3908560508646497,
      "tokens_seen": 2302803968
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001257802156232268,
      "loss": 2.8557,
      "theoretical_loss": 3.3908482261981865,
      "tokens_seen": 2302869504
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001257636655948553,
      "loss": 2.0856,
      "theoretical_loss": 3.390840401816745,
      "tokens_seen": 2302935040
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012574711556648385,
      "loss": 2.6206,
      "theoretical_loss": 3.3908325777203077,
      "tokens_seen": 2303000576
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012573056553811236,
      "loss": 2.214,
      "theoretical_loss": 3.3908247539088547,
      "tokens_seen": 2303066112
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012571401550974087,
      "loss": 2.7316,
      "theoretical_loss": 3.3908169303823685,
      "tokens_seen": 2303131648
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012569746548136938,
      "loss": 2.7134,
      "theoretical_loss": 3.390809107140831,
      "tokens_seen": 2303197184
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012568091545299792,
      "loss": 2.6791,
      "theoretical_loss": 3.3908012841842226,
      "tokens_seen": 2303262720
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012566436542462643,
      "loss": 2.7279,
      "theoretical_loss": 3.3907934615125255,
      "tokens_seen": 2303328256
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012564781539625496,
      "loss": 2.9094,
      "theoretical_loss": 3.390785639125721,
      "tokens_seen": 2303393792
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012563126536788347,
      "loss": 2.7081,
      "theoretical_loss": 3.3907778170237908,
      "tokens_seen": 2303459328
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000125614715339512,
      "loss": 2.8401,
      "theoretical_loss": 3.390769995206716,
      "tokens_seen": 2303524864
    },
    {
      "epoch": 0.82,
      "objective/train/avg_token_score": 0.040719255805015564,
      "objective/train/avg_weight": 0.9521021246910095,
      "objective/train/docs_used": 1296468,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6210556030273438,
      "objective/train/std_weight": 0.13445089757442474,
      "objective/train/theoretical_loss": 3.390762173674479,
      "objective/train/tokens_used": 2324050400,
      "theoretical_loss": 3.390762173674479,
      "tokens_seen": 2303590400
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012559816531114052,
      "loss": 2.5773,
      "theoretical_loss": 3.390762173674479,
      "tokens_seen": 2303590400
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012558161528276906,
      "loss": 2.8504,
      "theoretical_loss": 3.3907543524270602,
      "tokens_seen": 2303655936
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012556506525439757,
      "loss": 2.9076,
      "theoretical_loss": 3.3907465314644414,
      "tokens_seen": 2303721472
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001255485152260261,
      "loss": 2.7755,
      "theoretical_loss": 3.3907387107866054,
      "tokens_seen": 2303787008
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012553196519765462,
      "loss": 2.7953,
      "theoretical_loss": 3.390730890393532,
      "tokens_seen": 2303852544
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012551541516928316,
      "loss": 2.6248,
      "theoretical_loss": 3.3907230702852034,
      "tokens_seen": 2303918080
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012549886514091167,
      "loss": 2.6043,
      "theoretical_loss": 3.3907152504616014,
      "tokens_seen": 2303983616
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001254823151125402,
      "loss": 2.6151,
      "theoretical_loss": 3.390707430922707,
      "tokens_seen": 2304049152
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012546576508416871,
      "loss": 2.4847,
      "theoretical_loss": 3.390699611668503,
      "tokens_seen": 2304114688
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012544921505579722,
      "loss": 2.5918,
      "theoretical_loss": 3.3906917926989695,
      "tokens_seen": 2304180224
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012543266502742576,
      "loss": 2.651,
      "theoretical_loss": 3.390683974014088,
      "tokens_seen": 2304245760
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012541611499905427,
      "loss": 2.9238,
      "theoretical_loss": 3.3906761556138414,
      "tokens_seen": 2304311296
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001253995649706828,
      "loss": 2.8583,
      "theoretical_loss": 3.39066833749821,
      "tokens_seen": 2304376832
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012538301494231132,
      "loss": 2.8828,
      "theoretical_loss": 3.390660519667176,
      "tokens_seen": 2304442368
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012536646491393986,
      "loss": 2.9966,
      "theoretical_loss": 3.3906527021207205,
      "tokens_seen": 2304507904
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012534991488556837,
      "loss": 2.9505,
      "theoretical_loss": 3.3906448848588253,
      "tokens_seen": 2304573440
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001253333648571969,
      "loss": 1.9141,
      "theoretical_loss": 3.390637067881472,
      "tokens_seen": 2304638976
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012531681482882542,
      "loss": 2.4732,
      "theoretical_loss": 3.390629251188642,
      "tokens_seen": 2304704512
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012530026480045395,
      "loss": 2.6044,
      "theoretical_loss": 3.390621434780317,
      "tokens_seen": 2304770048
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012528371477208246,
      "loss": 2.7488,
      "theoretical_loss": 3.3906136186564786,
      "tokens_seen": 2304835584
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000125267164743711,
      "loss": 2.7929,
      "theoretical_loss": 3.3906058028171078,
      "tokens_seen": 2304901120
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001252506147153395,
      "loss": 2.4875,
      "theoretical_loss": 3.390597987262187,
      "tokens_seen": 2304966656
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012523406468696805,
      "loss": 2.6959,
      "theoretical_loss": 3.3905901719916973,
      "tokens_seen": 2305032192
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012521751465859656,
      "loss": 2.7261,
      "theoretical_loss": 3.3905823570056204,
      "tokens_seen": 2305097728
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012520096463022507,
      "loss": 2.6911,
      "theoretical_loss": 3.3905745423039377,
      "tokens_seen": 2305163264
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001251844146018536,
      "loss": 2.8505,
      "theoretical_loss": 3.3905667278866307,
      "tokens_seen": 2305228800
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012516786457348212,
      "loss": 2.6072,
      "theoretical_loss": 3.390558913753681,
      "tokens_seen": 2305294336
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012515131454511066,
      "loss": 2.7819,
      "theoretical_loss": 3.3905510999050708,
      "tokens_seen": 2305359872
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012513476451673917,
      "loss": 2.4507,
      "theoretical_loss": 3.3905432863407805,
      "tokens_seen": 2305425408
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001251182144883677,
      "loss": 2.7227,
      "theoretical_loss": 3.390535473060793,
      "tokens_seen": 2305490944
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012510166445999621,
      "loss": 2.6754,
      "theoretical_loss": 3.390527660065089,
      "tokens_seen": 2305556480
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012508511443162475,
      "loss": 1.5224,
      "theoretical_loss": 3.3905198473536498,
      "tokens_seen": 2305622016
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012506856440325326,
      "loss": 2.7426,
      "theoretical_loss": 3.3905120349264584,
      "tokens_seen": 2305687552
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001250520143748818,
      "loss": 2.6556,
      "theoretical_loss": 3.3905042227834947,
      "tokens_seen": 2305753088
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001250354643465103,
      "loss": 2.8245,
      "theoretical_loss": 3.390496410924741,
      "tokens_seen": 2305818624
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012501891431813885,
      "loss": 2.4229,
      "theoretical_loss": 3.390488599350179,
      "tokens_seen": 2305884160
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012500236428976736,
      "loss": 2.8017,
      "theoretical_loss": 3.3904807880597905,
      "tokens_seen": 2305949696
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001249858142613959,
      "loss": 2.4657,
      "theoretical_loss": 3.390472977053556,
      "tokens_seen": 2306015232
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012496926423302438,
      "loss": 2.9132,
      "theoretical_loss": 3.3904651663314587,
      "tokens_seen": 2306080768
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012495271420465292,
      "loss": 2.6252,
      "theoretical_loss": 3.390457355893479,
      "tokens_seen": 2306146304
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012493616417628143,
      "loss": 2.9636,
      "theoretical_loss": 3.390449545739599,
      "tokens_seen": 2306211840
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012491961414790996,
      "loss": 2.8729,
      "theoretical_loss": 3.3904417358698002,
      "tokens_seen": 2306277376
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012490306411953848,
      "loss": 2.8719,
      "theoretical_loss": 3.390433926284064,
      "tokens_seen": 2306342912
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000124886514091167,
      "loss": 2.6012,
      "theoretical_loss": 3.390426116982372,
      "tokens_seen": 2306408448
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012486996406279552,
      "loss": 2.7056,
      "theoretical_loss": 3.390418307964706,
      "tokens_seen": 2306473984
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012485341403442406,
      "loss": 2.6527,
      "theoretical_loss": 3.3904104992310478,
      "tokens_seen": 2306539520
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012483686400605257,
      "loss": 2.7204,
      "theoretical_loss": 3.3904026907813787,
      "tokens_seen": 2306605056
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001248203139776811,
      "loss": 2.7965,
      "theoretical_loss": 3.39039488261568,
      "tokens_seen": 2306670592
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012480376394930962,
      "loss": 2.626,
      "theoretical_loss": 3.3903870747339333,
      "tokens_seen": 2306736128
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012478721392093816,
      "loss": 2.7493,
      "theoretical_loss": 3.3903792671361215,
      "tokens_seen": 2306801664
    },
    {
      "epoch": 0.82,
      "objective/train/avg_token_score": 0.01601943001151085,
      "objective/train/avg_weight": 0.9811593294143677,
      "objective/train/docs_used": 1298252,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1092255115509033,
      "objective/train/std_weight": 0.09072182327508926,
      "objective/train/theoretical_loss": 3.3903714598222243,
      "objective/train/tokens_used": 2327327200,
      "theoretical_loss": 3.3903714598222243,
      "tokens_seen": 2306867200
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012477066389256667,
      "loss": 2.4309,
      "theoretical_loss": 3.3903714598222243,
      "tokens_seen": 2306867200
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001247541138641952,
      "loss": 2.452,
      "theoretical_loss": 3.390363652792225,
      "tokens_seen": 2306932736
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012473756383582372,
      "loss": 2.6646,
      "theoretical_loss": 3.390355846046104,
      "tokens_seen": 2306998272
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012472101380745223,
      "loss": 2.7364,
      "theoretical_loss": 3.3903480395838437,
      "tokens_seen": 2307063808
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012470446377908076,
      "loss": 2.6282,
      "theoretical_loss": 3.3903402334054253,
      "tokens_seen": 2307129344
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012468791375070927,
      "loss": 2.6391,
      "theoretical_loss": 3.39033242751083,
      "tokens_seen": 2307194880
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001246713637223378,
      "loss": 2.6305,
      "theoretical_loss": 3.390324621900041,
      "tokens_seen": 2307260416
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012465481369396632,
      "loss": 2.7392,
      "theoretical_loss": 3.390316816573038,
      "tokens_seen": 2307325952
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012463826366559486,
      "loss": 2.5583,
      "theoretical_loss": 3.390309011529804,
      "tokens_seen": 2307391488
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012462171363722337,
      "loss": 2.63,
      "theoretical_loss": 3.39030120677032,
      "tokens_seen": 2307457024
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001246051636088519,
      "loss": 2.5879,
      "theoretical_loss": 3.390293402294567,
      "tokens_seen": 2307522560
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012458861358048042,
      "loss": 2.7935,
      "theoretical_loss": 3.3902855981025284,
      "tokens_seen": 2307588096
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012457206355210895,
      "loss": 2.6716,
      "theoretical_loss": 3.3902777941941844,
      "tokens_seen": 2307653632
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012455551352373747,
      "loss": 2.8282,
      "theoretical_loss": 3.390269990569517,
      "tokens_seen": 2307719168
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000124538963495366,
      "loss": 2.9512,
      "theoretical_loss": 3.3902621872285077,
      "tokens_seen": 2307784704
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001245224134669945,
      "loss": 2.6219,
      "theoretical_loss": 3.3902543841711386,
      "tokens_seen": 2307850240
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012450586343862305,
      "loss": 2.493,
      "theoretical_loss": 3.3902465813973905,
      "tokens_seen": 2307915776
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012448931341025156,
      "loss": 2.7566,
      "theoretical_loss": 3.390238778907246,
      "tokens_seen": 2307981312
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012447276338188007,
      "loss": 2.5401,
      "theoretical_loss": 3.3902309767006864,
      "tokens_seen": 2308046848
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001244562133535086,
      "loss": 2.6205,
      "theoretical_loss": 3.390223174777693,
      "tokens_seen": 2308112384
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012443966332513712,
      "loss": 2.7578,
      "theoretical_loss": 3.390215373138248,
      "tokens_seen": 2308177920
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012442311329676566,
      "loss": 2.8045,
      "theoretical_loss": 3.390207571782332,
      "tokens_seen": 2308243456
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012440656326839417,
      "loss": 2.1307,
      "theoretical_loss": 3.390199770709928,
      "tokens_seen": 2308308992
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001243900132400227,
      "loss": 2.8166,
      "theoretical_loss": 3.390191969921017,
      "tokens_seen": 2308374528
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012437346321165122,
      "loss": 2.3473,
      "theoretical_loss": 3.3901841694155803,
      "tokens_seen": 2308440064
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012435691318327975,
      "loss": 2.2276,
      "theoretical_loss": 3.3901763691936,
      "tokens_seen": 2308505600
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012434036315490826,
      "loss": 2.6129,
      "theoretical_loss": 3.390168569255058,
      "tokens_seen": 2308571136
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001243238131265368,
      "loss": 2.8158,
      "theoretical_loss": 3.3901607695999356,
      "tokens_seen": 2308636672
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001243072630981653,
      "loss": 2.676,
      "theoretical_loss": 3.390152970228214,
      "tokens_seen": 2308702208
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012429071306979385,
      "loss": 2.6612,
      "theoretical_loss": 3.390145171139876,
      "tokens_seen": 2308767744
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012427416304142236,
      "loss": 2.5975,
      "theoretical_loss": 3.390137372334902,
      "tokens_seen": 2308833280
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001242576130130509,
      "loss": 2.7498,
      "theoretical_loss": 3.3901295738132746,
      "tokens_seen": 2308898816
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012424106298467938,
      "loss": 2.4302,
      "theoretical_loss": 3.390121775574975,
      "tokens_seen": 2308964352
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012422451295630792,
      "loss": 2.812,
      "theoretical_loss": 3.3901139776199853,
      "tokens_seen": 2309029888
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012420796292793643,
      "loss": 2.8407,
      "theoretical_loss": 3.390106179948287,
      "tokens_seen": 2309095424
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012419141289956497,
      "loss": 2.778,
      "theoretical_loss": 3.390098382559861,
      "tokens_seen": 2309160960
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012417486287119348,
      "loss": 1.8446,
      "theoretical_loss": 3.39009058545469,
      "tokens_seen": 2309226496
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012415831284282201,
      "loss": 2.7056,
      "theoretical_loss": 3.390082788632755,
      "tokens_seen": 2309292032
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012414176281445052,
      "loss": 2.402,
      "theoretical_loss": 3.3900749920940383,
      "tokens_seen": 2309357568
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012412521278607906,
      "loss": 2.6384,
      "theoretical_loss": 3.3900671958385207,
      "tokens_seen": 2309423104
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012410866275770757,
      "loss": 2.6265,
      "theoretical_loss": 3.390059399866185,
      "tokens_seen": 2309488640
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001240921127293361,
      "loss": 2.5871,
      "theoretical_loss": 3.390051604177012,
      "tokens_seen": 2309554176
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012407556270096462,
      "loss": 2.5454,
      "theoretical_loss": 3.3900438087709834,
      "tokens_seen": 2309619712
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012405901267259316,
      "loss": 2.7571,
      "theoretical_loss": 3.3900360136480816,
      "tokens_seen": 2309685248
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012404246264422167,
      "loss": 2.8223,
      "theoretical_loss": 3.3900282188082875,
      "tokens_seen": 2309750784
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0001240259126158502,
      "loss": 2.4399,
      "theoretical_loss": 3.3900204242515835,
      "tokens_seen": 2309816320
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012400936258747872,
      "loss": 2.5092,
      "theoretical_loss": 3.3900126299779503,
      "tokens_seen": 2309881856
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.00012399281255910723,
      "loss": 2.6998,
      "theoretical_loss": 3.3900048359873707,
      "tokens_seen": 2309947392
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012397626253073576,
      "loss": 2.6353,
      "theoretical_loss": 3.3899970422798256,
      "tokens_seen": 2310012928
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012395971250236427,
      "loss": 2.6479,
      "theoretical_loss": 3.389989248855297,
      "tokens_seen": 2310078464
    },
    {
      "epoch": 0.83,
      "objective/train/avg_token_score": 0.030694492161273956,
      "objective/train/avg_weight": 0.9638602137565613,
      "objective/train/docs_used": 1299979,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.324735641479492,
      "objective/train/std_weight": 0.14230284094810486,
      "objective/train/theoretical_loss": 3.389981455713767,
      "objective/train/tokens_used": 2330604000,
      "theoretical_loss": 3.389981455713767,
      "tokens_seen": 2310144000
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001239431624739928,
      "loss": 1.9581,
      "theoretical_loss": 3.389981455713767,
      "tokens_seen": 2310144000
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012392661244562132,
      "loss": 2.6665,
      "theoretical_loss": 3.3899736628552164,
      "tokens_seen": 2310209536
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012391006241724986,
      "loss": 2.6976,
      "theoretical_loss": 3.389965870279627,
      "tokens_seen": 2310275072
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012389351238887837,
      "loss": 2.4995,
      "theoretical_loss": 3.389958077986982,
      "tokens_seen": 2310340608
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001238769623605069,
      "loss": 2.5298,
      "theoretical_loss": 3.3899502859772612,
      "tokens_seen": 2310406144
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012386041233213542,
      "loss": 2.8182,
      "theoretical_loss": 3.389942494250447,
      "tokens_seen": 2310471680
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012384386230376396,
      "loss": 2.6601,
      "theoretical_loss": 3.389934702806521,
      "tokens_seen": 2310537216
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012382731227539247,
      "loss": 2.634,
      "theoretical_loss": 3.389926911645466,
      "tokens_seen": 2310602752
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.000123810762247021,
      "loss": 1.5726,
      "theoretical_loss": 3.389919120767262,
      "tokens_seen": 2310668288
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012379421221864951,
      "loss": 2.5905,
      "theoretical_loss": 3.3899113301718913,
      "tokens_seen": 2310733824
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012377766219027805,
      "loss": 2.7179,
      "theoretical_loss": 3.3899035398593362,
      "tokens_seen": 2310799360
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012376111216190656,
      "loss": 2.426,
      "theoretical_loss": 3.389895749829578,
      "tokens_seen": 2310864896
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012374456213353507,
      "loss": 2.7254,
      "theoretical_loss": 3.389887960082598,
      "tokens_seen": 2310930432
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001237280121051636,
      "loss": 2.8189,
      "theoretical_loss": 3.3898801706183788,
      "tokens_seen": 2310995968
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012371146207679212,
      "loss": 2.8178,
      "theoretical_loss": 3.3898723814369016,
      "tokens_seen": 2311061504
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012369491204842066,
      "loss": 2.8606,
      "theoretical_loss": 3.389864592538148,
      "tokens_seen": 2311127040
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012367836202004917,
      "loss": 2.8446,
      "theoretical_loss": 3.3898568039221004,
      "tokens_seen": 2311192576
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001236618119916777,
      "loss": 2.9624,
      "theoretical_loss": 3.3898490155887395,
      "tokens_seen": 2311258112
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012364526196330622,
      "loss": 2.635,
      "theoretical_loss": 3.3898412275380476,
      "tokens_seen": 2311323648
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012362871193493475,
      "loss": 2.4606,
      "theoretical_loss": 3.3898334397700065,
      "tokens_seen": 2311389184
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012361216190656327,
      "loss": 2.6532,
      "theoretical_loss": 3.389825652284598,
      "tokens_seen": 2311454720
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001235956118781918,
      "loss": 2.6916,
      "theoretical_loss": 3.3898178650818034,
      "tokens_seen": 2311520256
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001235790618498203,
      "loss": 2.3961,
      "theoretical_loss": 3.389810078161605,
      "tokens_seen": 2311585792
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012356251182144885,
      "loss": 2.6985,
      "theoretical_loss": 3.3898022915239836,
      "tokens_seen": 2311651328
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012354596179307736,
      "loss": 2.445,
      "theoretical_loss": 3.389794505168922,
      "tokens_seen": 2311716864
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001235294117647059,
      "loss": 2.6765,
      "theoretical_loss": 3.389786719096401,
      "tokens_seen": 2311782400
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001235128617363344,
      "loss": 2.7168,
      "theoretical_loss": 3.389778933306403,
      "tokens_seen": 2311847936
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012349631170796292,
      "loss": 2.5977,
      "theoretical_loss": 3.3897711477989096,
      "tokens_seen": 2311913472
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012347976167959143,
      "loss": 2.558,
      "theoretical_loss": 3.3897633625739028,
      "tokens_seen": 2311979008
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012346321165121997,
      "loss": 2.8759,
      "theoretical_loss": 3.3897555776313637,
      "tokens_seen": 2312044544
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012344666162284848,
      "loss": 2.7711,
      "theoretical_loss": 3.3897477929712743,
      "tokens_seen": 2312110080
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012343011159447702,
      "loss": 2.7526,
      "theoretical_loss": 3.3897400085936167,
      "tokens_seen": 2312175616
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012341356156610553,
      "loss": 2.6702,
      "theoretical_loss": 3.3897322244983723,
      "tokens_seen": 2312241152
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012339701153773406,
      "loss": 2.338,
      "theoretical_loss": 3.389724440685523,
      "tokens_seen": 2312306688
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012338046150936257,
      "loss": 2.7935,
      "theoretical_loss": 3.3897166571550503,
      "tokens_seen": 2312372224
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001233639114809911,
      "loss": 2.8135,
      "theoretical_loss": 3.389708873906936,
      "tokens_seen": 2312437760
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012334736145261962,
      "loss": 2.6247,
      "theoretical_loss": 3.3897010909411622,
      "tokens_seen": 2312503296
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012333081142424816,
      "loss": 2.5938,
      "theoretical_loss": 3.3896933082577103,
      "tokens_seen": 2312568832
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012331426139587667,
      "loss": 2.5937,
      "theoretical_loss": 3.3896855258565624,
      "tokens_seen": 2312634368
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001232977113675052,
      "loss": 2.6965,
      "theoretical_loss": 3.3896777437376997,
      "tokens_seen": 2312699904
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012328116133913372,
      "loss": 2.7475,
      "theoretical_loss": 3.3896699619011046,
      "tokens_seen": 2312765440
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012326461131076226,
      "loss": 2.771,
      "theoretical_loss": 3.389662180346759,
      "tokens_seen": 2312830976
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012324806128239077,
      "loss": 2.6621,
      "theoretical_loss": 3.3896543990746437,
      "tokens_seen": 2312896512
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012323151125401928,
      "loss": 2.7662,
      "theoretical_loss": 3.3896466180847407,
      "tokens_seen": 2312962048
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012321496122564781,
      "loss": 2.5316,
      "theoretical_loss": 3.3896388373770328,
      "tokens_seen": 2313027584
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012319841119727632,
      "loss": 2.5102,
      "theoretical_loss": 3.389631056951501,
      "tokens_seen": 2313093120
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012318186116890486,
      "loss": 2.5193,
      "theoretical_loss": 3.389623276808127,
      "tokens_seen": 2313158656
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012316531114053337,
      "loss": 2.5607,
      "theoretical_loss": 3.389615496946892,
      "tokens_seen": 2313224192
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001231487611121619,
      "loss": 2.586,
      "theoretical_loss": 3.389607717367779,
      "tokens_seen": 2313289728
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012313221108379042,
      "loss": 2.7669,
      "theoretical_loss": 3.3895999380707695,
      "tokens_seen": 2313355264
    },
    {
      "epoch": 0.83,
      "objective/train/avg_token_score": 0.01694384217262268,
      "objective/train/avg_weight": 0.9801348447799683,
      "objective/train/docs_used": 1302096,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.54665470123291,
      "objective/train/std_weight": 0.1007770523428917,
      "objective/train/theoretical_loss": 3.389592159055845,
      "objective/train/tokens_used": 2333880800,
      "theoretical_loss": 3.389592159055845,
      "tokens_seen": 2313420800
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012311566105541896,
      "loss": 2.7789,
      "theoretical_loss": 3.389592159055845,
      "tokens_seen": 2313420800
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012309911102704747,
      "loss": 2.5385,
      "theoretical_loss": 3.389584380322987,
      "tokens_seen": 2313486336
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.000123082560998676,
      "loss": 2.8191,
      "theoretical_loss": 3.389576601872178,
      "tokens_seen": 2313551872
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012306601097030452,
      "loss": 2.1948,
      "theoretical_loss": 3.389568823703399,
      "tokens_seen": 2313617408
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012304946094193305,
      "loss": 2.6575,
      "theoretical_loss": 3.3895610458166323,
      "tokens_seen": 2313682944
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012303291091356156,
      "loss": 2.6397,
      "theoretical_loss": 3.3895532682118596,
      "tokens_seen": 2313748480
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001230163608851901,
      "loss": 2.5777,
      "theoretical_loss": 3.3895454908890628,
      "tokens_seen": 2313814016
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001229998108568186,
      "loss": 2.7244,
      "theoretical_loss": 3.389537713848223,
      "tokens_seen": 2313879552
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012298326082844712,
      "loss": 2.6869,
      "theoretical_loss": 3.389529937089323,
      "tokens_seen": 2313945088
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012296671080007566,
      "loss": 2.5846,
      "theoretical_loss": 3.3895221606123442,
      "tokens_seen": 2314010624
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012295016077170417,
      "loss": 2.5441,
      "theoretical_loss": 3.389514384417268,
      "tokens_seen": 2314076160
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001229336107433327,
      "loss": 2.4402,
      "theoretical_loss": 3.389506608504077,
      "tokens_seen": 2314141696
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012291706071496122,
      "loss": 2.6617,
      "theoretical_loss": 3.3894988328727518,
      "tokens_seen": 2314207232
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012290051068658976,
      "loss": 2.9668,
      "theoretical_loss": 3.3894910575232755,
      "tokens_seen": 2314272768
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012288396065821827,
      "loss": 2.5806,
      "theoretical_loss": 3.389483282455629,
      "tokens_seen": 2314338304
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001228674106298468,
      "loss": 2.4737,
      "theoretical_loss": 3.389475507669794,
      "tokens_seen": 2314403840
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012285086060147531,
      "loss": 2.9548,
      "theoretical_loss": 3.3894677331657537,
      "tokens_seen": 2314469376
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012283431057310385,
      "loss": 2.6137,
      "theoretical_loss": 3.3894599589434886,
      "tokens_seen": 2314534912
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012281776054473236,
      "loss": 2.5453,
      "theoretical_loss": 3.3894521850029804,
      "tokens_seen": 2314600448
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001228012105163609,
      "loss": 2.5872,
      "theoretical_loss": 3.3894444113442117,
      "tokens_seen": 2314665984
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001227846604879894,
      "loss": 2.6217,
      "theoretical_loss": 3.389436637967164,
      "tokens_seen": 2314731520
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012276811045961792,
      "loss": 2.68,
      "theoretical_loss": 3.389428864871819,
      "tokens_seen": 2314797056
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012275156043124643,
      "loss": 2.49,
      "theoretical_loss": 3.389421092058159,
      "tokens_seen": 2314862592
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012273501040287497,
      "loss": 2.4869,
      "theoretical_loss": 3.389413319526165,
      "tokens_seen": 2314928128
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012271846037450348,
      "loss": 2.6481,
      "theoretical_loss": 3.389405547275819,
      "tokens_seen": 2314993664
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012270191034613202,
      "loss": 2.7345,
      "theoretical_loss": 3.3893977753071036,
      "tokens_seen": 2315059200
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012268536031776053,
      "loss": 2.7006,
      "theoretical_loss": 3.3893900036199995,
      "tokens_seen": 2315124736
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012266881028938906,
      "loss": 2.569,
      "theoretical_loss": 3.3893822322144898,
      "tokens_seen": 2315190272
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012265226026101758,
      "loss": 2.5735,
      "theoretical_loss": 3.389374461090555,
      "tokens_seen": 2315255808
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001226357102326461,
      "loss": 2.5784,
      "theoretical_loss": 3.3893666902481776,
      "tokens_seen": 2315321344
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012261916020427465,
      "loss": 2.7676,
      "theoretical_loss": 3.38935891968734,
      "tokens_seen": 2315386880
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012260261017590316,
      "loss": 2.5296,
      "theoretical_loss": 3.389351149408023,
      "tokens_seen": 2315452416
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001225860601475317,
      "loss": 2.4919,
      "theoretical_loss": 3.389343379410209,
      "tokens_seen": 2315517952
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001225695101191602,
      "loss": 2.4699,
      "theoretical_loss": 3.3893356096938794,
      "tokens_seen": 2315583488
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012255296009078875,
      "loss": 2.684,
      "theoretical_loss": 3.389327840259017,
      "tokens_seen": 2315649024
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012253641006241726,
      "loss": 2.6616,
      "theoretical_loss": 3.389320071105602,
      "tokens_seen": 2315714560
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012251986003404577,
      "loss": 2.7114,
      "theoretical_loss": 3.3893123022336176,
      "tokens_seen": 2315780096
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012250331000567428,
      "loss": 2.601,
      "theoretical_loss": 3.3893045336430454,
      "tokens_seen": 2315845632
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012248675997730282,
      "loss": 2.5111,
      "theoretical_loss": 3.389296765333867,
      "tokens_seen": 2315911168
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012247020994893133,
      "loss": 2.7381,
      "theoretical_loss": 3.3892889973060645,
      "tokens_seen": 2315976704
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012245365992055986,
      "loss": 2.581,
      "theoretical_loss": 3.3892812295596193,
      "tokens_seen": 2316042240
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012243710989218837,
      "loss": 2.6282,
      "theoretical_loss": 3.3892734620945135,
      "tokens_seen": 2316107776
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001224205598638169,
      "loss": 2.4521,
      "theoretical_loss": 3.3892656949107294,
      "tokens_seen": 2316173312
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012240400983544542,
      "loss": 2.6156,
      "theoretical_loss": 3.3892579280082478,
      "tokens_seen": 2316238848
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012238745980707396,
      "loss": 2.4322,
      "theoretical_loss": 3.389250161387052,
      "tokens_seen": 2316304384
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012237090977870247,
      "loss": 2.6297,
      "theoretical_loss": 3.3892423950471224,
      "tokens_seen": 2316369920
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.000122354359750331,
      "loss": 2.5807,
      "theoretical_loss": 3.389234628988442,
      "tokens_seen": 2316435456
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012233780972195952,
      "loss": 2.8788,
      "theoretical_loss": 3.3892268632109914,
      "tokens_seen": 2316500992
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012232125969358806,
      "loss": 2.4544,
      "theoretical_loss": 3.389219097714754,
      "tokens_seen": 2316566528
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012230470966521657,
      "loss": 2.8134,
      "theoretical_loss": 3.38921133249971,
      "tokens_seen": 2316632064
    },
    {
      "epoch": 0.83,
      "objective/train/avg_token_score": 0.020373407751321793,
      "objective/train/avg_weight": 0.9760119915008545,
      "objective/train/docs_used": 1304213,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.201322555541992,
      "objective/train/std_weight": 0.12593387067317963,
      "objective/train/theoretical_loss": 3.389203567565843,
      "objective/train/tokens_used": 2337157600,
      "theoretical_loss": 3.389203567565843,
      "tokens_seen": 2316697600
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001222881596368451,
      "loss": 2.4768,
      "theoretical_loss": 3.389203567565843,
      "tokens_seen": 2316697600
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012227160960847361,
      "loss": 2.5196,
      "theoretical_loss": 3.3891958029131333,
      "tokens_seen": 2316763136
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012225505958010212,
      "loss": 2.4154,
      "theoretical_loss": 3.389188038541564,
      "tokens_seen": 2316828672
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012223850955173066,
      "loss": 2.5958,
      "theoretical_loss": 3.3891802744511166,
      "tokens_seen": 2316894208
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012222195952335917,
      "loss": 2.1624,
      "theoretical_loss": 3.3891725106417723,
      "tokens_seen": 2316959744
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001222054094949877,
      "loss": 2.5569,
      "theoretical_loss": 3.3891647471135133,
      "tokens_seen": 2317025280
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012218885946661622,
      "loss": 2.5533,
      "theoretical_loss": 3.3891569838663225,
      "tokens_seen": 2317090816
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012217230943824476,
      "loss": 2.181,
      "theoretical_loss": 3.38914922090018,
      "tokens_seen": 2317156352
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012215575940987327,
      "loss": 2.54,
      "theoretical_loss": 3.389141458215069,
      "tokens_seen": 2317221888
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001221392093815018,
      "loss": 2.4729,
      "theoretical_loss": 3.3891336958109712,
      "tokens_seen": 2317287424
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012212265935313032,
      "loss": 2.5008,
      "theoretical_loss": 3.3891259336878683,
      "tokens_seen": 2317352960
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012210610932475885,
      "loss": 2.045,
      "theoretical_loss": 3.389118171845742,
      "tokens_seen": 2317418496
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012208955929638736,
      "loss": 2.4852,
      "theoretical_loss": 3.389110410284574,
      "tokens_seen": 2317484032
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001220730092680159,
      "loss": 2.5112,
      "theoretical_loss": 3.3891026490043465,
      "tokens_seen": 2317549568
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012205645923964441,
      "loss": 2.6026,
      "theoretical_loss": 3.389094888005042,
      "tokens_seen": 2317615104
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012203990921127294,
      "loss": 2.4687,
      "theoretical_loss": 3.3890871272866416,
      "tokens_seen": 2317680640
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012202335918290145,
      "loss": 2.6246,
      "theoretical_loss": 3.389079366849127,
      "tokens_seen": 2317746176
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012200680915452997,
      "loss": 2.4672,
      "theoretical_loss": 3.3890716066924806,
      "tokens_seen": 2317811712
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001219902591261585,
      "loss": 1.8052,
      "theoretical_loss": 3.3890638468166845,
      "tokens_seen": 2317877248
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012197370909778702,
      "loss": 2.5433,
      "theoretical_loss": 3.38905608722172,
      "tokens_seen": 2317942784
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012195715906941554,
      "loss": 2.5426,
      "theoretical_loss": 3.3890483279075694,
      "tokens_seen": 2318008320
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012194060904104407,
      "loss": 2.5519,
      "theoretical_loss": 3.3890405688742145,
      "tokens_seen": 2318073856
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012192405901267259,
      "loss": 2.6072,
      "theoretical_loss": 3.3890328101216367,
      "tokens_seen": 2318139392
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012190750898430111,
      "loss": 1.895,
      "theoretical_loss": 3.389025051649819,
      "tokens_seen": 2318204928
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012189095895592964,
      "loss": 2.4786,
      "theoretical_loss": 3.389017293458742,
      "tokens_seen": 2318270464
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012187440892755816,
      "loss": 2.2579,
      "theoretical_loss": 3.389009535548389,
      "tokens_seen": 2318336000
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012185785889918669,
      "loss": 2.6854,
      "theoretical_loss": 3.3890017779187405,
      "tokens_seen": 2318401536
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012184130887081521,
      "loss": 2.6291,
      "theoretical_loss": 3.3889940205697795,
      "tokens_seen": 2318467072
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012182475884244373,
      "loss": 2.7401,
      "theoretical_loss": 3.3889862635014874,
      "tokens_seen": 2318532608
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012180820881407226,
      "loss": 2.649,
      "theoretical_loss": 3.3889785067138463,
      "tokens_seen": 2318598144
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012179165878570078,
      "loss": 2.7094,
      "theoretical_loss": 3.3889707502068376,
      "tokens_seen": 2318663680
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012177510875732929,
      "loss": 2.3569,
      "theoretical_loss": 3.388962993980444,
      "tokens_seen": 2318729216
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012175855872895782,
      "loss": 2.6352,
      "theoretical_loss": 3.3889552380346473,
      "tokens_seen": 2318794752
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012174200870058634,
      "loss": 2.5344,
      "theoretical_loss": 3.3889474823694288,
      "tokens_seen": 2318860288
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012172545867221486,
      "loss": 2.7324,
      "theoretical_loss": 3.3889397269847708,
      "tokens_seen": 2318925824
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012170890864384339,
      "loss": 2.4468,
      "theoretical_loss": 3.3889319718806554,
      "tokens_seen": 2318991360
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012169235861547191,
      "loss": 2.7994,
      "theoretical_loss": 3.388924217057064,
      "tokens_seen": 2319056896
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012167580858710044,
      "loss": 2.4845,
      "theoretical_loss": 3.388916462513979,
      "tokens_seen": 2319122432
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012165925855872896,
      "loss": 1.7916,
      "theoretical_loss": 3.3889087082513827,
      "tokens_seen": 2319187968
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012164270853035748,
      "loss": 2.574,
      "theoretical_loss": 3.388900954269256,
      "tokens_seen": 2319253504
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012162615850198601,
      "loss": 2.7509,
      "theoretical_loss": 3.388893200567581,
      "tokens_seen": 2319319040
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012160960847361453,
      "loss": 2.337,
      "theoretical_loss": 3.3888854471463405,
      "tokens_seen": 2319384576
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012159305844524306,
      "loss": 2.376,
      "theoretical_loss": 3.388877694005516,
      "tokens_seen": 2319450112
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012157650841687158,
      "loss": 2.6156,
      "theoretical_loss": 3.388869941145089,
      "tokens_seen": 2319515648
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001215599583885001,
      "loss": 2.6857,
      "theoretical_loss": 3.3888621885650423,
      "tokens_seen": 2319581184
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012154340836012863,
      "loss": 2.7514,
      "theoretical_loss": 3.388854436265357,
      "tokens_seen": 2319646720
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012152685833175713,
      "loss": 2.3615,
      "theoretical_loss": 3.3888466842460154,
      "tokens_seen": 2319712256
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012151030830338565,
      "loss": 2.407,
      "theoretical_loss": 3.3888389325069994,
      "tokens_seen": 2319777792
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012149375827501417,
      "loss": 2.5247,
      "theoretical_loss": 3.3888311810482907,
      "tokens_seen": 2319843328
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001214772082466427,
      "loss": 2.4051,
      "theoretical_loss": 3.388823429869872,
      "tokens_seen": 2319908864
    },
    {
      "epoch": 0.83,
      "objective/train/avg_token_score": 0.016623402014374733,
      "objective/train/avg_weight": 0.98043292760849,
      "objective/train/docs_used": 1306171,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1827986240386963,
      "objective/train/std_weight": 0.08088142424821854,
      "objective/train/theoretical_loss": 3.3888156789717243,
      "objective/train/tokens_used": 2340434400,
      "theoretical_loss": 3.3888156789717243,
      "tokens_seen": 2319974400
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012146065821827122,
      "loss": 2.6863,
      "theoretical_loss": 3.3888156789717243,
      "tokens_seen": 2319974400
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012144410818989975,
      "loss": 2.5098,
      "theoretical_loss": 3.38880792835383,
      "tokens_seen": 2320039936
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012142755816152827,
      "loss": 2.6869,
      "theoretical_loss": 3.3888001780161714,
      "tokens_seen": 2320105472
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001214110081331568,
      "loss": 2.8335,
      "theoretical_loss": 3.3887924279587303,
      "tokens_seen": 2320171008
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012139445810478532,
      "loss": 2.6528,
      "theoretical_loss": 3.388784678181488,
      "tokens_seen": 2320236544
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012137790807641384,
      "loss": 2.5088,
      "theoretical_loss": 3.388776928684427,
      "tokens_seen": 2320302080
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012136135804804237,
      "loss": 2.6731,
      "theoretical_loss": 3.388769179467529,
      "tokens_seen": 2320367616
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012134480801967089,
      "loss": 2.6545,
      "theoretical_loss": 3.3887614305307765,
      "tokens_seen": 2320433152
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012132825799129941,
      "loss": 2.5687,
      "theoretical_loss": 3.388753681874151,
      "tokens_seen": 2320498688
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012131170796292794,
      "loss": 2.6877,
      "theoretical_loss": 3.3887459334976344,
      "tokens_seen": 2320564224
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012129515793455645,
      "loss": 2.3137,
      "theoretical_loss": 3.3887381854012086,
      "tokens_seen": 2320629760
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012127860790618497,
      "loss": 2.6639,
      "theoretical_loss": 3.388730437584856,
      "tokens_seen": 2320695296
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001212620578778135,
      "loss": 2.6878,
      "theoretical_loss": 3.3887226900485583,
      "tokens_seen": 2320760832
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012124550784944202,
      "loss": 2.5553,
      "theoretical_loss": 3.388714942792298,
      "tokens_seen": 2320826368
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012122895782107054,
      "loss": 2.5996,
      "theoretical_loss": 3.388707195816056,
      "tokens_seen": 2320891904
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012121240779269907,
      "loss": 2.8444,
      "theoretical_loss": 3.388699449119815,
      "tokens_seen": 2320957440
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012119585776432759,
      "loss": 2.7613,
      "theoretical_loss": 3.3886917027035572,
      "tokens_seen": 2321022976
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012117930773595612,
      "loss": 2.5334,
      "theoretical_loss": 3.3886839565672635,
      "tokens_seen": 2321088512
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012116275770758464,
      "loss": 2.8668,
      "theoretical_loss": 3.3886762107109174,
      "tokens_seen": 2321154048
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012114620767921316,
      "loss": 2.6091,
      "theoretical_loss": 3.3886684651344994,
      "tokens_seen": 2321219584
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012112965765084169,
      "loss": 2.9003,
      "theoretical_loss": 3.3886607198379926,
      "tokens_seen": 2321285120
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012111310762247021,
      "loss": 2.2321,
      "theoretical_loss": 3.388652974821378,
      "tokens_seen": 2321350656
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012109655759409874,
      "loss": 2.702,
      "theoretical_loss": 3.3886452300846384,
      "tokens_seen": 2321416192
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012108000756572726,
      "loss": 2.3681,
      "theoretical_loss": 3.3886374856277555,
      "tokens_seen": 2321481728
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012106345753735578,
      "loss": 2.468,
      "theoretical_loss": 3.3886297414507114,
      "tokens_seen": 2321547264
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001210469075089843,
      "loss": 2.5706,
      "theoretical_loss": 3.388621997553488,
      "tokens_seen": 2321612800
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012103035748061282,
      "loss": 2.7426,
      "theoretical_loss": 3.388614253936067,
      "tokens_seen": 2321678336
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012101380745224134,
      "loss": 2.6348,
      "theoretical_loss": 3.3886065105984304,
      "tokens_seen": 2321743872
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012099725742386987,
      "loss": 2.3377,
      "theoretical_loss": 3.3885987675405613,
      "tokens_seen": 2321809408
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012098070739549839,
      "loss": 2.8882,
      "theoretical_loss": 3.38859102476244,
      "tokens_seen": 2321874944
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012096415736712691,
      "loss": 2.606,
      "theoretical_loss": 3.38858328226405,
      "tokens_seen": 2321940480
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012094760733875544,
      "loss": 2.5972,
      "theoretical_loss": 3.3885755400453723,
      "tokens_seen": 2322006016
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012093105731038396,
      "loss": 2.3234,
      "theoretical_loss": 3.388567798106389,
      "tokens_seen": 2322071552
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012091450728201249,
      "loss": 2.4881,
      "theoretical_loss": 3.388560056447083,
      "tokens_seen": 2322137088
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012089795725364101,
      "loss": 2.6344,
      "theoretical_loss": 3.3885523150674355,
      "tokens_seen": 2322202624
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012088140722526953,
      "loss": 2.6613,
      "theoretical_loss": 3.3885445739674283,
      "tokens_seen": 2322268160
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012086485719689806,
      "loss": 2.2776,
      "theoretical_loss": 3.388536833147044,
      "tokens_seen": 2322333696
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012084830716852658,
      "loss": 2.7418,
      "theoretical_loss": 3.388529092606264,
      "tokens_seen": 2322399232
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001208317571401551,
      "loss": 2.4236,
      "theoretical_loss": 3.3885213523450712,
      "tokens_seen": 2322464768
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012081520711178363,
      "loss": 2.559,
      "theoretical_loss": 3.3885136123634467,
      "tokens_seen": 2322530304
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012079865708341213,
      "loss": 2.4977,
      "theoretical_loss": 3.3885058726613733,
      "tokens_seen": 2322595840
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012078210705504065,
      "loss": 2.6349,
      "theoretical_loss": 3.388498133238832,
      "tokens_seen": 2322661376
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012076555702666917,
      "loss": 2.6298,
      "theoretical_loss": 3.388490394095806,
      "tokens_seen": 2322726912
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001207490069982977,
      "loss": 1.9696,
      "theoretical_loss": 3.3884826552322767,
      "tokens_seen": 2322792448
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012073245696992622,
      "loss": 2.516,
      "theoretical_loss": 3.388474916648226,
      "tokens_seen": 2322857984
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012071590694155475,
      "loss": 2.6771,
      "theoretical_loss": 3.388467178343636,
      "tokens_seen": 2322923520
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012069935691318327,
      "loss": 2.7456,
      "theoretical_loss": 3.3884594403184893,
      "tokens_seen": 2322989056
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001206828068848118,
      "loss": 2.6703,
      "theoretical_loss": 3.3884517025727674,
      "tokens_seen": 2323054592
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012066625685644032,
      "loss": 2.7658,
      "theoretical_loss": 3.388443965106452,
      "tokens_seen": 2323120128
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012064970682806884,
      "loss": 2.7046,
      "theoretical_loss": 3.3884362279195255,
      "tokens_seen": 2323185664
    },
    {
      "epoch": 0.83,
      "objective/train/avg_token_score": 0.02712240070104599,
      "objective/train/avg_weight": 0.96806800365448,
      "objective/train/docs_used": 1308132,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.2700605392456055,
      "objective/train/std_weight": 0.15027031302452087,
      "objective/train/theoretical_loss": 3.3884284910119704,
      "objective/train/tokens_used": 2343711200,
      "theoretical_loss": 3.3884284910119704,
      "tokens_seen": 2323251200
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012063315679969737,
      "loss": 2.3048,
      "theoretical_loss": 3.3884284910119704,
      "tokens_seen": 2323251200
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001206166067713259,
      "loss": 2.5899,
      "theoretical_loss": 3.388420754383768,
      "tokens_seen": 2323316736
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012060005674295443,
      "loss": 2.6347,
      "theoretical_loss": 3.3884130180349006,
      "tokens_seen": 2323382272
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012058350671458295,
      "loss": 2.6994,
      "theoretical_loss": 3.3884052819653503,
      "tokens_seen": 2323447808
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012056695668621148,
      "loss": 2.6304,
      "theoretical_loss": 3.388397546175099,
      "tokens_seen": 2323513344
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012055040665783997,
      "loss": 2.7705,
      "theoretical_loss": 3.388389810664129,
      "tokens_seen": 2323578880
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001205338566294685,
      "loss": 2.6389,
      "theoretical_loss": 3.3883820754324216,
      "tokens_seen": 2323644416
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012051730660109702,
      "loss": 2.4544,
      "theoretical_loss": 3.3883743404799596,
      "tokens_seen": 2323709952
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012050075657272555,
      "loss": 2.6448,
      "theoretical_loss": 3.388366605806725,
      "tokens_seen": 2323775488
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012048420654435407,
      "loss": 2.323,
      "theoretical_loss": 3.3883588714127,
      "tokens_seen": 2323841024
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012046765651598259,
      "loss": 2.7043,
      "theoretical_loss": 3.388351137297866,
      "tokens_seen": 2323906560
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012045110648761112,
      "loss": 2.5449,
      "theoretical_loss": 3.3883434034622057,
      "tokens_seen": 2323972096
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012043455645923964,
      "loss": 2.6631,
      "theoretical_loss": 3.3883356699057003,
      "tokens_seen": 2324037632
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012041800643086817,
      "loss": 2.5649,
      "theoretical_loss": 3.3883279366283325,
      "tokens_seen": 2324103168
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012040145640249669,
      "loss": 2.4333,
      "theoretical_loss": 3.3883202036300846,
      "tokens_seen": 2324168704
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012038490637412521,
      "loss": 2.7429,
      "theoretical_loss": 3.3883124709109382,
      "tokens_seen": 2324234240
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012036835634575374,
      "loss": 2.554,
      "theoretical_loss": 3.3883047384708753,
      "tokens_seen": 2324299776
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012035180631738226,
      "loss": 2.5311,
      "theoretical_loss": 3.3882970063098785,
      "tokens_seen": 2324365312
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012033525628901079,
      "loss": 2.7473,
      "theoretical_loss": 3.3882892744279287,
      "tokens_seen": 2324430848
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012031870626063931,
      "loss": 2.7906,
      "theoretical_loss": 3.3882815428250095,
      "tokens_seen": 2324496384
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012030215623226782,
      "loss": 2.6239,
      "theoretical_loss": 3.3882738115011017,
      "tokens_seen": 2324561920
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012028560620389634,
      "loss": 2.7922,
      "theoretical_loss": 3.388266080456188,
      "tokens_seen": 2324627456
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012026905617552487,
      "loss": 2.4487,
      "theoretical_loss": 3.388258349690251,
      "tokens_seen": 2324692992
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012025250614715339,
      "loss": 2.637,
      "theoretical_loss": 3.388250619203271,
      "tokens_seen": 2324758528
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012023595611878192,
      "loss": 2.5008,
      "theoretical_loss": 3.388242888995232,
      "tokens_seen": 2324824064
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012021940609041044,
      "loss": 2.7234,
      "theoretical_loss": 3.388235159066115,
      "tokens_seen": 2324889600
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012020285606203896,
      "loss": 2.6627,
      "theoretical_loss": 3.388227429415902,
      "tokens_seen": 2324955136
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012018630603366749,
      "loss": 2.4324,
      "theoretical_loss": 3.3882197000445755,
      "tokens_seen": 2325020672
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012016975600529601,
      "loss": 2.6071,
      "theoretical_loss": 3.388211970952118,
      "tokens_seen": 2325086208
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012015320597692454,
      "loss": 2.2007,
      "theoretical_loss": 3.3882042421385106,
      "tokens_seen": 2325151744
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012013665594855306,
      "loss": 2.5792,
      "theoretical_loss": 3.388196513603736,
      "tokens_seen": 2325217280
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012012010592018158,
      "loss": 2.3359,
      "theoretical_loss": 3.388188785347776,
      "tokens_seen": 2325282816
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012010355589181011,
      "loss": 2.509,
      "theoretical_loss": 3.3881810573706126,
      "tokens_seen": 2325348352
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012008700586343863,
      "loss": 2.5652,
      "theoretical_loss": 3.3881733296722283,
      "tokens_seen": 2325413888
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012007045583506716,
      "loss": 2.587,
      "theoretical_loss": 3.388165602252605,
      "tokens_seen": 2325479424
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012005390580669565,
      "loss": 2.6688,
      "theoretical_loss": 3.388157875111725,
      "tokens_seen": 2325544960
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012003735577832418,
      "loss": 2.6928,
      "theoretical_loss": 3.38815014824957,
      "tokens_seen": 2325610496
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001200208057499527,
      "loss": 2.6577,
      "theoretical_loss": 3.3881424216661222,
      "tokens_seen": 2325676032
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00012000425572158122,
      "loss": 2.4645,
      "theoretical_loss": 3.3881346953613636,
      "tokens_seen": 2325741568
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011998770569320976,
      "loss": 2.3556,
      "theoretical_loss": 3.3881269693352762,
      "tokens_seen": 2325807104
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011997115566483829,
      "loss": 2.5863,
      "theoretical_loss": 3.388119243587843,
      "tokens_seen": 2325872640
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011995460563646681,
      "loss": 2.555,
      "theoretical_loss": 3.388111518119045,
      "tokens_seen": 2325938176
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011993805560809533,
      "loss": 2.5878,
      "theoretical_loss": 3.388103792928865,
      "tokens_seen": 2326003712
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011992150557972386,
      "loss": 2.5294,
      "theoretical_loss": 3.3880960680172842,
      "tokens_seen": 2326069248
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011990495555135238,
      "loss": 2.6519,
      "theoretical_loss": 3.388088343384286,
      "tokens_seen": 2326134784
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001198884055229809,
      "loss": 2.6405,
      "theoretical_loss": 3.3880806190298514,
      "tokens_seen": 2326200320
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011987185549460943,
      "loss": 2.8641,
      "theoretical_loss": 3.388072894953963,
      "tokens_seen": 2326265856
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011985530546623795,
      "loss": 2.7489,
      "theoretical_loss": 3.388065171156603,
      "tokens_seen": 2326331392
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011983875543786648,
      "loss": 2.7379,
      "theoretical_loss": 3.3880574476377534,
      "tokens_seen": 2326396928
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.000119822205409495,
      "loss": 2.44,
      "theoretical_loss": 3.3880497243973964,
      "tokens_seen": 2326462464
    },
    {
      "epoch": 0.83,
      "objective/train/avg_token_score": 0.014057767577469349,
      "objective/train/avg_weight": 0.9834492802619934,
      "objective/train/docs_used": 1310381,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.871222496032715,
      "objective/train/std_weight": 0.06661440432071686,
      "objective/train/theoretical_loss": 3.3880420014355135,
      "objective/train/tokens_used": 2346988000,
      "theoretical_loss": 3.3880420014355135,
      "tokens_seen": 2326528000
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001198056553811235,
      "loss": 2.6522,
      "theoretical_loss": 3.3880420014355135,
      "tokens_seen": 2326528000
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011978910535275202,
      "loss": 2.6748,
      "theoretical_loss": 3.388034278752088,
      "tokens_seen": 2326593536
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011977255532438055,
      "loss": 2.6923,
      "theoretical_loss": 3.388026556347101,
      "tokens_seen": 2326659072
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011975600529600907,
      "loss": 2.5527,
      "theoretical_loss": 3.3880188342205346,
      "tokens_seen": 2326724608
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001197394552676376,
      "loss": 2.6579,
      "theoretical_loss": 3.3880111123723715,
      "tokens_seen": 2326790144
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011972290523926612,
      "loss": 2.4142,
      "theoretical_loss": 3.3880033908025937,
      "tokens_seen": 2326855680
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011970635521089464,
      "loss": 2.6528,
      "theoretical_loss": 3.387995669511183,
      "tokens_seen": 2326921216
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011968980518252317,
      "loss": 2.4675,
      "theoretical_loss": 3.3879879484981217,
      "tokens_seen": 2326986752
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011967325515415169,
      "loss": 2.57,
      "theoretical_loss": 3.3879802277633924,
      "tokens_seen": 2327052288
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011965670512578021,
      "loss": 2.3383,
      "theoretical_loss": 3.3879725073069764,
      "tokens_seen": 2327117824
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011964015509740874,
      "loss": 2.5354,
      "theoretical_loss": 3.3879647871288565,
      "tokens_seen": 2327183360
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011962360506903726,
      "loss": 2.5043,
      "theoretical_loss": 3.3879570672290145,
      "tokens_seen": 2327248896
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011960705504066579,
      "loss": 2.5269,
      "theoretical_loss": 3.387949347607432,
      "tokens_seen": 2327314432
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011959050501229431,
      "loss": 2.8421,
      "theoretical_loss": 3.3879416282640924,
      "tokens_seen": 2327379968
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011957395498392282,
      "loss": 2.637,
      "theoretical_loss": 3.387933909198977,
      "tokens_seen": 2327445504
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011955740495555134,
      "loss": 2.746,
      "theoretical_loss": 3.3879261904120677,
      "tokens_seen": 2327511040
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011954085492717987,
      "loss": 2.7118,
      "theoretical_loss": 3.3879184719033475,
      "tokens_seen": 2327576576
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011952430489880839,
      "loss": 2.703,
      "theoretical_loss": 3.3879107536727977,
      "tokens_seen": 2327642112
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011950775487043692,
      "loss": 2.5496,
      "theoretical_loss": 3.3879030357204014,
      "tokens_seen": 2327707648
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011949120484206544,
      "loss": 2.7703,
      "theoretical_loss": 3.3878953180461395,
      "tokens_seen": 2327773184
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011947465481369396,
      "loss": 2.418,
      "theoretical_loss": 3.387887600649995,
      "tokens_seen": 2327838720
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011945810478532249,
      "loss": 2.5578,
      "theoretical_loss": 3.38787988353195,
      "tokens_seen": 2327904256
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011944155475695101,
      "loss": 2.3902,
      "theoretical_loss": 3.387872166691986,
      "tokens_seen": 2327969792
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011942500472857954,
      "loss": 2.5559,
      "theoretical_loss": 3.3878644501300865,
      "tokens_seen": 2328035328
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011940845470020806,
      "loss": 2.3683,
      "theoretical_loss": 3.3878567338462324,
      "tokens_seen": 2328100864
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011939190467183658,
      "loss": 2.606,
      "theoretical_loss": 3.387849017840406,
      "tokens_seen": 2328166400
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011937535464346511,
      "loss": 2.368,
      "theoretical_loss": 3.3878413021125904,
      "tokens_seen": 2328231936
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011935880461509363,
      "loss": 2.7205,
      "theoretical_loss": 3.3878335866627665,
      "tokens_seen": 2328297472
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011934225458672216,
      "loss": 2.7233,
      "theoretical_loss": 3.387825871490917,
      "tokens_seen": 2328363008
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011932570455835067,
      "loss": 2.5034,
      "theoretical_loss": 3.3878181565970245,
      "tokens_seen": 2328428544
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011930915452997919,
      "loss": 2.6117,
      "theoretical_loss": 3.387810441981071,
      "tokens_seen": 2328494080
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011929260450160772,
      "loss": 2.3975,
      "theoretical_loss": 3.3878027276430376,
      "tokens_seen": 2328559616
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011927605447323624,
      "loss": 2.5033,
      "theoretical_loss": 3.387795013582908,
      "tokens_seen": 2328625152
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011925950444486476,
      "loss": 2.6724,
      "theoretical_loss": 3.387787299800663,
      "tokens_seen": 2328690688
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011924295441649329,
      "loss": 2.3713,
      "theoretical_loss": 3.387779586296286,
      "tokens_seen": 2328756224
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011922640438812181,
      "loss": 1.0608,
      "theoretical_loss": 3.387771873069758,
      "tokens_seen": 2328821760
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011920985435975034,
      "loss": 2.8049,
      "theoretical_loss": 3.3877641601210624,
      "tokens_seen": 2328887296
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011919330433137886,
      "loss": 2.7276,
      "theoretical_loss": 3.3877564474501805,
      "tokens_seen": 2328952832
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011917675430300738,
      "loss": 2.5673,
      "theoretical_loss": 3.3877487350570945,
      "tokens_seen": 2329018368
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011916020427463591,
      "loss": 2.5527,
      "theoretical_loss": 3.3877410229417873,
      "tokens_seen": 2329083904
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011914365424626443,
      "loss": 2.5921,
      "theoretical_loss": 3.3877333111042405,
      "tokens_seen": 2329149440
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011912710421789296,
      "loss": 2.3593,
      "theoretical_loss": 3.387725599544436,
      "tokens_seen": 2329214976
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011911055418952148,
      "loss": 2.6214,
      "theoretical_loss": 3.3877178882623564,
      "tokens_seen": 2329280512
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011909400416115,
      "loss": 2.7069,
      "theoretical_loss": 3.387710177257984,
      "tokens_seen": 2329346048
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001190774541327785,
      "loss": 2.5313,
      "theoretical_loss": 3.3877024665313007,
      "tokens_seen": 2329411584
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011906090410440702,
      "loss": 2.783,
      "theoretical_loss": 3.387694756082289,
      "tokens_seen": 2329477120
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011904435407603555,
      "loss": 2.5761,
      "theoretical_loss": 3.3876870459109307,
      "tokens_seen": 2329542656
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011902780404766407,
      "loss": 2.5572,
      "theoretical_loss": 3.387679336017208,
      "tokens_seen": 2329608192
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001190112540192926,
      "loss": 2.5862,
      "theoretical_loss": 3.387671626401104,
      "tokens_seen": 2329673728
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011899470399092112,
      "loss": 2.8277,
      "theoretical_loss": 3.3876639170626,
      "tokens_seen": 2329739264
    },
    {
      "epoch": 0.83,
      "objective/train/avg_token_score": 0.011159960180521011,
      "objective/train/avg_weight": 0.9868701696395874,
      "objective/train/docs_used": 1312090,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.620255947113037,
      "objective/train/std_weight": 0.06689024716615677,
      "objective/train/theoretical_loss": 3.387656208001678,
      "objective/train/tokens_used": 2350264800,
      "theoretical_loss": 3.387656208001678,
      "tokens_seen": 2329804800
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011897815396254964,
      "loss": 2.6862,
      "theoretical_loss": 3.387656208001678,
      "tokens_seen": 2329804800
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011896160393417817,
      "loss": 2.7338,
      "theoretical_loss": 3.3876484992183205,
      "tokens_seen": 2329870336
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011894505390580669,
      "loss": 2.7014,
      "theoretical_loss": 3.38764079071251,
      "tokens_seen": 2329935872
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011892850387743522,
      "loss": 2.3664,
      "theoretical_loss": 3.3876330824842285,
      "tokens_seen": 2330001408
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011891195384906374,
      "loss": 2.3951,
      "theoretical_loss": 3.387625374533458,
      "tokens_seen": 2330066944
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011889540382069226,
      "loss": 2.6606,
      "theoretical_loss": 3.387617666860181,
      "tokens_seen": 2330132480
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011887885379232079,
      "loss": 2.5059,
      "theoretical_loss": 3.3876099594643794,
      "tokens_seen": 2330198016
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011886230376394931,
      "loss": 2.4837,
      "theoretical_loss": 3.387602252346036,
      "tokens_seen": 2330263552
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011884575373557784,
      "loss": 2.5008,
      "theoretical_loss": 3.3875945455051317,
      "tokens_seen": 2330329088
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011882920370720635,
      "loss": 2.4014,
      "theoretical_loss": 3.38758683894165,
      "tokens_seen": 2330394624
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011881265367883487,
      "loss": 2.4815,
      "theoretical_loss": 3.3875791326555733,
      "tokens_seen": 2330460160
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001187961036504634,
      "loss": 2.1669,
      "theoretical_loss": 3.3875714266468826,
      "tokens_seen": 2330525696
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011877955362209192,
      "loss": 2.6619,
      "theoretical_loss": 3.387563720915561,
      "tokens_seen": 2330591232
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011876300359372044,
      "loss": 2.4325,
      "theoretical_loss": 3.3875560154615907,
      "tokens_seen": 2330656768
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011874645356534897,
      "loss": 2.7111,
      "theoretical_loss": 3.3875483102849535,
      "tokens_seen": 2330722304
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011872990353697749,
      "loss": 2.7847,
      "theoretical_loss": 3.387540605385632,
      "tokens_seen": 2330787840
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011871335350860601,
      "loss": 2.8034,
      "theoretical_loss": 3.387532900763608,
      "tokens_seen": 2330853376
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011869680348023454,
      "loss": 2.5665,
      "theoretical_loss": 3.3875251964188635,
      "tokens_seen": 2330918912
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011868025345186306,
      "loss": 2.6001,
      "theoretical_loss": 3.3875174923513818,
      "tokens_seen": 2330984448
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011866370342349159,
      "loss": 2.7118,
      "theoretical_loss": 3.387509788561144,
      "tokens_seen": 2331049984
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011864715339512011,
      "loss": 2.5744,
      "theoretical_loss": 3.3875020850481334,
      "tokens_seen": 2331115520
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011863060336674863,
      "loss": 2.5201,
      "theoretical_loss": 3.3874943818123313,
      "tokens_seen": 2331181056
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011861405333837716,
      "loss": 2.7117,
      "theoretical_loss": 3.387486678853721,
      "tokens_seen": 2331246592
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011859750331000568,
      "loss": 2.6531,
      "theoretical_loss": 3.3874789761722828,
      "tokens_seen": 2331312128
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011858095328163419,
      "loss": 2.6305,
      "theoretical_loss": 3.387471273768001,
      "tokens_seen": 2331377664
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011856440325326272,
      "loss": 2.5402,
      "theoretical_loss": 3.3874635716408568,
      "tokens_seen": 2331443200
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011854785322489124,
      "loss": 2.3529,
      "theoretical_loss": 3.3874558697908324,
      "tokens_seen": 2331508736
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011853130319651976,
      "loss": 2.4331,
      "theoretical_loss": 3.3874481682179107,
      "tokens_seen": 2331574272
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011851475316814829,
      "loss": 2.5574,
      "theoretical_loss": 3.3874404669220732,
      "tokens_seen": 2331639808
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011849820313977681,
      "loss": 2.4306,
      "theoretical_loss": 3.3874327659033026,
      "tokens_seen": 2331705344
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011848165311140534,
      "loss": 2.8029,
      "theoretical_loss": 3.3874250651615814,
      "tokens_seen": 2331770880
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011846510308303386,
      "loss": 2.4205,
      "theoretical_loss": 3.3874173646968906,
      "tokens_seen": 2331836416
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011844855305466238,
      "loss": 2.5781,
      "theoretical_loss": 3.3874096645092138,
      "tokens_seen": 2331901952
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011843200302629091,
      "loss": 2.3591,
      "theoretical_loss": 3.3874019645985327,
      "tokens_seen": 2331967488
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011841545299791943,
      "loss": 2.7461,
      "theoretical_loss": 3.3873942649648296,
      "tokens_seen": 2332033024
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011839890296954796,
      "loss": 2.5642,
      "theoretical_loss": 3.387386565608087,
      "tokens_seen": 2332098560
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011838235294117648,
      "loss": 2.2517,
      "theoretical_loss": 3.3873788665282865,
      "tokens_seen": 2332164096
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.000118365802912805,
      "loss": 2.7152,
      "theoretical_loss": 3.3873711677254112,
      "tokens_seen": 2332229632
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011834925288443353,
      "loss": 2.6168,
      "theoretical_loss": 3.3873634691994425,
      "tokens_seen": 2332295168
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011833270285606203,
      "loss": 2.9651,
      "theoretical_loss": 3.3873557709503634,
      "tokens_seen": 2332360704
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011831615282769055,
      "loss": 2.609,
      "theoretical_loss": 3.3873480729781558,
      "tokens_seen": 2332426240
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011829960279931907,
      "loss": 2.4167,
      "theoretical_loss": 3.387340375282802,
      "tokens_seen": 2332491776
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001182830527709476,
      "loss": 2.7201,
      "theoretical_loss": 3.3873326778642845,
      "tokens_seen": 2332557312
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011826650274257612,
      "loss": 2.507,
      "theoretical_loss": 3.387324980722585,
      "tokens_seen": 2332622848
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011824995271420465,
      "loss": 2.4107,
      "theoretical_loss": 3.387317283857686,
      "tokens_seen": 2332688384
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011823340268583317,
      "loss": 2.6698,
      "theoretical_loss": 3.38730958726957,
      "tokens_seen": 2332753920
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001182168526574617,
      "loss": 2.4191,
      "theoretical_loss": 3.387301890958219,
      "tokens_seen": 2332819456
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011820030262909022,
      "loss": 2.6874,
      "theoretical_loss": 3.3872941949236157,
      "tokens_seen": 2332884992
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011818375260071874,
      "loss": 2.4707,
      "theoretical_loss": 3.387286499165742,
      "tokens_seen": 2332950528
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011816720257234727,
      "loss": 2.5834,
      "theoretical_loss": 3.3872788036845805,
      "tokens_seen": 2333016064
    },
    {
      "epoch": 0.83,
      "objective/train/avg_token_score": 0.008725408464670181,
      "objective/train/avg_weight": 0.9897310733795166,
      "objective/train/docs_used": 1313833,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.898077964782715,
      "objective/train/std_weight": 0.045301154255867004,
      "objective/train/theoretical_loss": 3.387271108480113,
      "objective/train/tokens_used": 2353541600,
      "theoretical_loss": 3.387271108480113,
      "tokens_seen": 2333081600
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011815065254397579,
      "loss": 2.734,
      "theoretical_loss": 3.387271108480113,
      "tokens_seen": 2333081600
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011813410251560431,
      "loss": 2.6252,
      "theoretical_loss": 3.3872634135523225,
      "tokens_seen": 2333147136
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011811755248723284,
      "loss": 2.5341,
      "theoretical_loss": 3.3872557189011903,
      "tokens_seen": 2333212672
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011810100245886136,
      "loss": 2.6463,
      "theoretical_loss": 3.3872480245266994,
      "tokens_seen": 2333278208
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011808445243048987,
      "loss": 2.6105,
      "theoretical_loss": 3.3872403304288317,
      "tokens_seen": 2333343744
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001180679024021184,
      "loss": 2.1503,
      "theoretical_loss": 3.38723263660757,
      "tokens_seen": 2333409280
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011805135237374692,
      "loss": 2.7542,
      "theoretical_loss": 3.387224943062896,
      "tokens_seen": 2333474816
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011803480234537544,
      "loss": 2.3792,
      "theoretical_loss": 3.3872172497947926,
      "tokens_seen": 2333540352
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011801825231700397,
      "loss": 2.6879,
      "theoretical_loss": 3.387209556803241,
      "tokens_seen": 2333605888
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011800170228863249,
      "loss": 2.6174,
      "theoretical_loss": 3.3872018640882247,
      "tokens_seen": 2333671424
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011798515226026102,
      "loss": 2.6112,
      "theoretical_loss": 3.3871941716497256,
      "tokens_seen": 2333736960
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011796860223188954,
      "loss": 2.3867,
      "theoretical_loss": 3.3871864794877258,
      "tokens_seen": 2333802496
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011795205220351806,
      "loss": 2.5574,
      "theoretical_loss": 3.387178787602208,
      "tokens_seen": 2333868032
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011793550217514659,
      "loss": 2.5639,
      "theoretical_loss": 3.3871710959931534,
      "tokens_seen": 2333933568
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011791895214677511,
      "loss": 2.5168,
      "theoretical_loss": 3.387163404660546,
      "tokens_seen": 2333999104
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011790240211840364,
      "loss": 2.6726,
      "theoretical_loss": 3.3871557136043666,
      "tokens_seen": 2334064640
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011788585209003216,
      "loss": 2.4135,
      "theoretical_loss": 3.3871480228245985,
      "tokens_seen": 2334130176
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011786930206166068,
      "loss": 2.3717,
      "theoretical_loss": 3.3871403323212235,
      "tokens_seen": 2334195712
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001178527520332892,
      "loss": 2.6222,
      "theoretical_loss": 3.3871326420942243,
      "tokens_seen": 2334261248
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011783620200491772,
      "loss": 2.7171,
      "theoretical_loss": 3.3871249521435827,
      "tokens_seen": 2334326784
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011781965197654624,
      "loss": 2.462,
      "theoretical_loss": 3.387117262469281,
      "tokens_seen": 2334392320
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011780310194817477,
      "loss": 2.5827,
      "theoretical_loss": 3.387109573071302,
      "tokens_seen": 2334457856
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011778655191980329,
      "loss": 2.3797,
      "theoretical_loss": 3.387101883949628,
      "tokens_seen": 2334523392
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011777000189143181,
      "loss": 2.6783,
      "theoretical_loss": 3.387094195104241,
      "tokens_seen": 2334588928
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011775345186306034,
      "loss": 2.2484,
      "theoretical_loss": 3.387086506535123,
      "tokens_seen": 2334654464
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011773690183468886,
      "loss": 2.4581,
      "theoretical_loss": 3.3870788182422573,
      "tokens_seen": 2334720000
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011772035180631739,
      "loss": 2.6113,
      "theoretical_loss": 3.3870711302256256,
      "tokens_seen": 2334785536
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011770380177794591,
      "loss": 2.6236,
      "theoretical_loss": 3.3870634424852097,
      "tokens_seen": 2334851072
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011768725174957443,
      "loss": 2.5804,
      "theoretical_loss": 3.387055755020993,
      "tokens_seen": 2334916608
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011767070172120296,
      "loss": 2.0798,
      "theoretical_loss": 3.387048067832957,
      "tokens_seen": 2334982144
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011765415169283148,
      "loss": 2.4711,
      "theoretical_loss": 3.3870403809210847,
      "tokens_seen": 2335047680
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011763760166446,
      "loss": 2.3697,
      "theoretical_loss": 3.3870326942853577,
      "tokens_seen": 2335113216
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011762105163608853,
      "loss": 2.5064,
      "theoretical_loss": 3.387025007925759,
      "tokens_seen": 2335178752
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011760450160771703,
      "loss": 2.7722,
      "theoretical_loss": 3.387017321842271,
      "tokens_seen": 2335244288
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011758795157934555,
      "loss": 2.5936,
      "theoretical_loss": 3.387009636034875,
      "tokens_seen": 2335309824
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011757140155097407,
      "loss": 2.5318,
      "theoretical_loss": 3.3870019505035542,
      "tokens_seen": 2335375360
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001175548515226026,
      "loss": 2.5008,
      "theoretical_loss": 3.386994265248291,
      "tokens_seen": 2335440896
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011753830149423112,
      "loss": 2.6145,
      "theoretical_loss": 3.386986580269067,
      "tokens_seen": 2335506432
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011752175146585965,
      "loss": 2.6827,
      "theoretical_loss": 3.3869788955658655,
      "tokens_seen": 2335571968
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011750520143748817,
      "loss": 2.5181,
      "theoretical_loss": 3.3869712111386683,
      "tokens_seen": 2335637504
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001174886514091167,
      "loss": 2.7091,
      "theoretical_loss": 3.3869635269874574,
      "tokens_seen": 2335703040
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011747210138074522,
      "loss": 2.5422,
      "theoretical_loss": 3.386955843112216,
      "tokens_seen": 2335768576
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011745555135237374,
      "loss": 2.4173,
      "theoretical_loss": 3.386948159512926,
      "tokens_seen": 2335834112
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011743900132400227,
      "loss": 2.3683,
      "theoretical_loss": 3.386940476189569,
      "tokens_seen": 2335899648
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011742245129563079,
      "loss": 2.3597,
      "theoretical_loss": 3.3869327931421287,
      "tokens_seen": 2335965184
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011740590126725931,
      "loss": 2.4968,
      "theoretical_loss": 3.3869251103705866,
      "tokens_seen": 2336030720
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011738935123888784,
      "loss": 2.8372,
      "theoretical_loss": 3.3869174278749257,
      "tokens_seen": 2336096256
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011737280121051636,
      "loss": 2.7197,
      "theoretical_loss": 3.386909745655128,
      "tokens_seen": 2336161792
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011735625118214487,
      "loss": 2.5599,
      "theoretical_loss": 3.386902063711175,
      "tokens_seen": 2336227328
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0001173397011537734,
      "loss": 2.4913,
      "theoretical_loss": 3.386894382043051,
      "tokens_seen": 2336292864
    },
    {
      "epoch": 0.83,
      "objective/train/avg_token_score": 0.024882851168513298,
      "objective/train/avg_weight": 0.9707625508308411,
      "objective/train/docs_used": 1316188,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.060212135314941,
      "objective/train/std_weight": 0.12923628091812134,
      "objective/train/theoretical_loss": 3.386886700650736,
      "objective/train/tokens_used": 2356818400,
      "theoretical_loss": 3.386886700650736,
      "tokens_seen": 2336358400
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011732315112540192,
      "loss": 2.4628,
      "theoretical_loss": 3.386886700650736,
      "tokens_seen": 2336358400
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011730660109703045,
      "loss": 2.3757,
      "theoretical_loss": 3.3868790195342147,
      "tokens_seen": 2336423936
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011729005106865897,
      "loss": 2.4674,
      "theoretical_loss": 3.3868713386934677,
      "tokens_seen": 2336489472
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011727350104028749,
      "loss": 1.7968,
      "theoretical_loss": 3.386863658128478,
      "tokens_seen": 2336555008
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011725695101191602,
      "loss": 2.6876,
      "theoretical_loss": 3.386855977839228,
      "tokens_seen": 2336620544
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011724040098354454,
      "loss": 2.6312,
      "theoretical_loss": 3.3868482978257006,
      "tokens_seen": 2336686080
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011722385095517307,
      "loss": 2.4039,
      "theoretical_loss": 3.386840618087877,
      "tokens_seen": 2336751616
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011720730092680159,
      "loss": 2.7339,
      "theoretical_loss": 3.3868329386257403,
      "tokens_seen": 2336817152
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011719075089843011,
      "loss": 2.4309,
      "theoretical_loss": 3.386825259439273,
      "tokens_seen": 2336882688
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011717420087005864,
      "loss": 2.4141,
      "theoretical_loss": 3.386817580528457,
      "tokens_seen": 2336948224
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011715765084168716,
      "loss": 2.6,
      "theoretical_loss": 3.386809901893275,
      "tokens_seen": 2337013760
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011714110081331569,
      "loss": 2.5977,
      "theoretical_loss": 3.386802223533709,
      "tokens_seen": 2337079296
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011712455078494421,
      "loss": 2.8031,
      "theoretical_loss": 3.3867945454497423,
      "tokens_seen": 2337144832
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011710800075657272,
      "loss": 2.3258,
      "theoretical_loss": 3.386786867641356,
      "tokens_seen": 2337210368
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011709145072820124,
      "loss": 2.8176,
      "theoretical_loss": 3.3867791901085336,
      "tokens_seen": 2337275904
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011707490069982977,
      "loss": 2.2092,
      "theoretical_loss": 3.386771512851257,
      "tokens_seen": 2337341440
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011705835067145829,
      "loss": 1.855,
      "theoretical_loss": 3.386763835869508,
      "tokens_seen": 2337406976
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011704180064308682,
      "loss": 2.5664,
      "theoretical_loss": 3.38675615916327,
      "tokens_seen": 2337472512
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011702525061471534,
      "loss": 2.3177,
      "theoretical_loss": 3.386748482732525,
      "tokens_seen": 2337538048
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011700870058634386,
      "loss": 2.7177,
      "theoretical_loss": 3.3867408065772553,
      "tokens_seen": 2337603584
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011699215055797239,
      "loss": 2.4505,
      "theoretical_loss": 3.3867331306974435,
      "tokens_seen": 2337669120
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011697560052960091,
      "loss": 2.287,
      "theoretical_loss": 3.3867254550930714,
      "tokens_seen": 2337734656
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011695905050122944,
      "loss": 2.729,
      "theoretical_loss": 3.386717779764122,
      "tokens_seen": 2337800192
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011694250047285796,
      "loss": 2.5388,
      "theoretical_loss": 3.386710104710578,
      "tokens_seen": 2337865728
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00011692595044448648,
      "loss": 2.4435,
      "theoretical_loss": 3.386702429932421,
      "tokens_seen": 2337931264
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011690940041611501,
      "loss": 2.5473,
      "theoretical_loss": 3.386694755429634,
      "tokens_seen": 2337996800
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011689285038774353,
      "loss": 2.3895,
      "theoretical_loss": 3.3866870812021985,
      "tokens_seen": 2338062336
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011687630035937206,
      "loss": 2.0879,
      "theoretical_loss": 3.3866794072500976,
      "tokens_seen": 2338127872
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011685975033100055,
      "loss": 2.4967,
      "theoretical_loss": 3.3866717335733143,
      "tokens_seen": 2338193408
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011684320030262908,
      "loss": 2.491,
      "theoretical_loss": 3.38666406017183,
      "tokens_seen": 2338258944
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001168266502742576,
      "loss": 2.7286,
      "theoretical_loss": 3.3866563870456274,
      "tokens_seen": 2338324480
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011681010024588612,
      "loss": 2.6065,
      "theoretical_loss": 3.386648714194689,
      "tokens_seen": 2338390016
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011679355021751465,
      "loss": 2.2418,
      "theoretical_loss": 3.386641041618997,
      "tokens_seen": 2338455552
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011677700018914317,
      "loss": 2.5843,
      "theoretical_loss": 3.386633369318534,
      "tokens_seen": 2338521088
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001167604501607717,
      "loss": 2.4853,
      "theoretical_loss": 3.3866256972932827,
      "tokens_seen": 2338586624
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011674390013240022,
      "loss": 2.5596,
      "theoretical_loss": 3.386618025543225,
      "tokens_seen": 2338652160
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011672735010402874,
      "loss": 2.4072,
      "theoretical_loss": 3.3866103540683437,
      "tokens_seen": 2338717696
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011671080007565727,
      "loss": 2.667,
      "theoretical_loss": 3.386602682868621,
      "tokens_seen": 2338783232
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011669425004728579,
      "loss": 2.4978,
      "theoretical_loss": 3.3865950119440393,
      "tokens_seen": 2338848768
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011667770001891432,
      "loss": 2.3953,
      "theoretical_loss": 3.3865873412945815,
      "tokens_seen": 2338914304
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011666114999054284,
      "loss": 2.8163,
      "theoretical_loss": 3.386579670920229,
      "tokens_seen": 2338979840
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011664459996217136,
      "loss": 2.6252,
      "theoretical_loss": 3.386572000820965,
      "tokens_seen": 2339045376
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011662804993379989,
      "loss": 2.4142,
      "theoretical_loss": 3.386564330996772,
      "tokens_seen": 2339110912
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001166114999054284,
      "loss": 2.7621,
      "theoretical_loss": 3.3865566614476323,
      "tokens_seen": 2339176448
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011659494987705692,
      "loss": 2.4467,
      "theoretical_loss": 3.3865489921735277,
      "tokens_seen": 2339241984
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011657839984868545,
      "loss": 2.8228,
      "theoretical_loss": 3.3865413231744417,
      "tokens_seen": 2339307520
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011656184982031397,
      "loss": 2.5822,
      "theoretical_loss": 3.386533654450356,
      "tokens_seen": 2339373056
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001165452997919425,
      "loss": 2.7091,
      "theoretical_loss": 3.386525986001253,
      "tokens_seen": 2339438592
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011652874976357102,
      "loss": 2.7877,
      "theoretical_loss": 3.3865183178271154,
      "tokens_seen": 2339504128
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011651219973519954,
      "loss": 2.0063,
      "theoretical_loss": 3.386510649927926,
      "tokens_seen": 2339569664
    },
    {
      "epoch": 0.84,
      "objective/train/avg_token_score": 0.018511462956666946,
      "objective/train/avg_weight": 0.9782274961471558,
      "objective/train/docs_used": 1317929,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6221070289611816,
      "objective/train/std_weight": 0.07913627475500107,
      "objective/train/theoretical_loss": 3.386502982303667,
      "objective/train/tokens_used": 2360095200,
      "theoretical_loss": 3.386502982303667,
      "tokens_seen": 2339635200
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011649564970682807,
      "loss": 2.6357,
      "theoretical_loss": 3.386502982303667,
      "tokens_seen": 2339635200
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011647909967845659,
      "loss": 2.4875,
      "theoretical_loss": 3.38649531495432,
      "tokens_seen": 2339700736
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011646254965008511,
      "loss": 2.7253,
      "theoretical_loss": 3.3864876478798687,
      "tokens_seen": 2339766272
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011644599962171364,
      "loss": 2.4895,
      "theoretical_loss": 3.386479981080295,
      "tokens_seen": 2339831808
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011642944959334216,
      "loss": 2.4813,
      "theoretical_loss": 3.3864723145555806,
      "tokens_seen": 2339897344
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011641289956497069,
      "loss": 2.3772,
      "theoretical_loss": 3.3864646483057093,
      "tokens_seen": 2339962880
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011639634953659921,
      "loss": 2.3001,
      "theoretical_loss": 3.386456982330663,
      "tokens_seen": 2340028416
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011637979950822772,
      "loss": 2.5732,
      "theoretical_loss": 3.3864493166304235,
      "tokens_seen": 2340093952
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011636324947985624,
      "loss": 2.7129,
      "theoretical_loss": 3.3864416512049744,
      "tokens_seen": 2340159488
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011634669945148477,
      "loss": 2.4394,
      "theoretical_loss": 3.386433986054297,
      "tokens_seen": 2340225024
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011633014942311329,
      "loss": 2.6247,
      "theoretical_loss": 3.386426321178375,
      "tokens_seen": 2340290560
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011631359939474182,
      "loss": 2.6934,
      "theoretical_loss": 3.38641865657719,
      "tokens_seen": 2340356096
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011629704936637034,
      "loss": 2.1476,
      "theoretical_loss": 3.3864109922507244,
      "tokens_seen": 2340421632
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011628049933799886,
      "loss": 2.437,
      "theoretical_loss": 3.386403328198961,
      "tokens_seen": 2340487168
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011626394930962739,
      "loss": 2.404,
      "theoretical_loss": 3.3863956644218822,
      "tokens_seen": 2340552704
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011624739928125591,
      "loss": 2.5701,
      "theoretical_loss": 3.386388000919471,
      "tokens_seen": 2340618240
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011623084925288444,
      "loss": 2.4022,
      "theoretical_loss": 3.3863803376917083,
      "tokens_seen": 2340683776
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011621429922451296,
      "loss": 2.4818,
      "theoretical_loss": 3.3863726747385785,
      "tokens_seen": 2340749312
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011619774919614148,
      "loss": 2.399,
      "theoretical_loss": 3.386365012060063,
      "tokens_seen": 2340814848
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011618119916777001,
      "loss": 2.6558,
      "theoretical_loss": 3.386357349656144,
      "tokens_seen": 2340880384
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011616464913939853,
      "loss": 2.4667,
      "theoretical_loss": 3.3863496875268044,
      "tokens_seen": 2340945920
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011614809911102706,
      "loss": 2.3516,
      "theoretical_loss": 3.386342025672027,
      "tokens_seen": 2341011456
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011613154908265555,
      "loss": 2.6392,
      "theoretical_loss": 3.386334364091794,
      "tokens_seen": 2341076992
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011611499905428408,
      "loss": 2.1273,
      "theoretical_loss": 3.386326702786087,
      "tokens_seen": 2341142528
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001160984490259126,
      "loss": 1.9061,
      "theoretical_loss": 3.38631904175489,
      "tokens_seen": 2341208064
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011608189899754113,
      "loss": 2.4679,
      "theoretical_loss": 3.386311380998185,
      "tokens_seen": 2341273600
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011606534896916965,
      "loss": 2.5858,
      "theoretical_loss": 3.386303720515954,
      "tokens_seen": 2341339136
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011604879894079817,
      "loss": 2.3556,
      "theoretical_loss": 3.3862960603081795,
      "tokens_seen": 2341404672
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001160322489124267,
      "loss": 2.322,
      "theoretical_loss": 3.3862884003748444,
      "tokens_seen": 2341470208
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011601569888405522,
      "loss": 2.6791,
      "theoretical_loss": 3.386280740715931,
      "tokens_seen": 2341535744
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011599914885568376,
      "loss": 2.4329,
      "theoretical_loss": 3.3862730813314217,
      "tokens_seen": 2341601280
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011598259882731228,
      "loss": 2.2814,
      "theoretical_loss": 3.3862654222212996,
      "tokens_seen": 2341666816
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011596604879894081,
      "loss": 2.2658,
      "theoretical_loss": 3.386257763385546,
      "tokens_seen": 2341732352
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011594949877056933,
      "loss": 2.5829,
      "theoretical_loss": 3.3862501048241445,
      "tokens_seen": 2341797888
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011593294874219786,
      "loss": 2.8312,
      "theoretical_loss": 3.386242446537077,
      "tokens_seen": 2341863424
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011591639871382638,
      "loss": 2.5102,
      "theoretical_loss": 3.386234788524326,
      "tokens_seen": 2341928960
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001158998486854549,
      "loss": 2.6753,
      "theoretical_loss": 3.3862271307858745,
      "tokens_seen": 2341994496
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001158832986570834,
      "loss": 2.5575,
      "theoretical_loss": 3.3862194733217046,
      "tokens_seen": 2342060032
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011586674862871192,
      "loss": 2.4375,
      "theoretical_loss": 3.386211816131799,
      "tokens_seen": 2342125568
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011585019860034045,
      "loss": 2.5729,
      "theoretical_loss": 3.3862041592161396,
      "tokens_seen": 2342191104
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011583364857196897,
      "loss": 2.7157,
      "theoretical_loss": 3.38619650257471,
      "tokens_seen": 2342256640
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001158170985435975,
      "loss": 2.4897,
      "theoretical_loss": 3.3861888462074914,
      "tokens_seen": 2342322176
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011580054851522602,
      "loss": 2.3261,
      "theoretical_loss": 3.3861811901144674,
      "tokens_seen": 2342387712
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011578399848685454,
      "loss": 2.456,
      "theoretical_loss": 3.38617353429562,
      "tokens_seen": 2342453248
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011576744845848307,
      "loss": 2.7295,
      "theoretical_loss": 3.3861658787509317,
      "tokens_seen": 2342518784
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011575089843011159,
      "loss": 2.4451,
      "theoretical_loss": 3.3861582234803853,
      "tokens_seen": 2342584320
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011573434840174012,
      "loss": 2.6513,
      "theoretical_loss": 3.3861505684839632,
      "tokens_seen": 2342649856
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011571779837336864,
      "loss": 2.6351,
      "theoretical_loss": 3.386142913761648,
      "tokens_seen": 2342715392
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011570124834499716,
      "loss": 2.1883,
      "theoretical_loss": 3.3861352593134217,
      "tokens_seen": 2342780928
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011568469831662569,
      "loss": 2.8393,
      "theoretical_loss": 3.3861276051392672,
      "tokens_seen": 2342846464
    },
    {
      "epoch": 0.84,
      "objective/train/avg_token_score": 0.007082349620759487,
      "objective/train/avg_weight": 0.9917154312133789,
      "objective/train/docs_used": 1319369,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9728035926818848,
      "objective/train/std_weight": 0.05484147369861603,
      "objective/train/theoretical_loss": 3.3861199512391673,
      "objective/train/tokens_used": 2363372000,
      "theoretical_loss": 3.3861199512391673,
      "tokens_seen": 2342912000
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011566814828825421,
      "loss": 2.6218,
      "theoretical_loss": 3.3861199512391673,
      "tokens_seen": 2342912000
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011565159825988274,
      "loss": 2.5363,
      "theoretical_loss": 3.386112297613104,
      "tokens_seen": 2342977536
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011563504823151125,
      "loss": 3.046,
      "theoretical_loss": 3.38610464426106,
      "tokens_seen": 2343043072
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011561849820313977,
      "loss": 2.5717,
      "theoretical_loss": 3.386096991183018,
      "tokens_seen": 2343108608
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001156019481747683,
      "loss": 2.9625,
      "theoretical_loss": 3.3860893383789605,
      "tokens_seen": 2343174144
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011558539814639682,
      "loss": 2.6584,
      "theoretical_loss": 3.38608168584887,
      "tokens_seen": 2343239680
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011556884811802534,
      "loss": 2.7337,
      "theoretical_loss": 3.386074033592729,
      "tokens_seen": 2343305216
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011555229808965387,
      "loss": 2.6706,
      "theoretical_loss": 3.3860663816105196,
      "tokens_seen": 2343370752
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011553574806128239,
      "loss": 2.7399,
      "theoretical_loss": 3.386058729902225,
      "tokens_seen": 2343436288
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011551919803291091,
      "loss": 2.7002,
      "theoretical_loss": 3.3860510784678275,
      "tokens_seen": 2343501824
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011550264800453944,
      "loss": 2.8351,
      "theoretical_loss": 3.3860434273073095,
      "tokens_seen": 2343567360
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011548609797616796,
      "loss": 2.9229,
      "theoretical_loss": 3.3860357764206537,
      "tokens_seen": 2343632896
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011546954794779649,
      "loss": 2.6857,
      "theoretical_loss": 3.3860281258078424,
      "tokens_seen": 2343698432
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011545299791942501,
      "loss": 2.741,
      "theoretical_loss": 3.3860204754688583,
      "tokens_seen": 2343763968
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011543644789105353,
      "loss": 2.3839,
      "theoretical_loss": 3.3860128254036845,
      "tokens_seen": 2343829504
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011541989786268206,
      "loss": 2.6756,
      "theoretical_loss": 3.3860051756123024,
      "tokens_seen": 2343895040
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011540334783431058,
      "loss": 2.8125,
      "theoretical_loss": 3.3859975260946955,
      "tokens_seen": 2343960576
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011538679780593908,
      "loss": 2.4578,
      "theoretical_loss": 3.385989876850846,
      "tokens_seen": 2344026112
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011537024777756762,
      "loss": 2.4182,
      "theoretical_loss": 3.3859822278807363,
      "tokens_seen": 2344091648
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011535369774919614,
      "loss": 2.618,
      "theoretical_loss": 3.3859745791843494,
      "tokens_seen": 2344157184
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011533714772082466,
      "loss": 2.584,
      "theoretical_loss": 3.385966930761667,
      "tokens_seen": 2344222720
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011532059769245319,
      "loss": 2.6578,
      "theoretical_loss": 3.3859592826126725,
      "tokens_seen": 2344288256
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011530404766408171,
      "loss": 2.6889,
      "theoretical_loss": 3.3859516347373484,
      "tokens_seen": 2344353792
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011528749763571024,
      "loss": 2.835,
      "theoretical_loss": 3.3859439871356765,
      "tokens_seen": 2344419328
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011527094760733876,
      "loss": 2.6959,
      "theoretical_loss": 3.3859363398076407,
      "tokens_seen": 2344484864
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011525439757896728,
      "loss": 2.1043,
      "theoretical_loss": 3.3859286927532217,
      "tokens_seen": 2344550400
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011523784755059581,
      "loss": 2.5681,
      "theoretical_loss": 3.385921045972404,
      "tokens_seen": 2344615936
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011522129752222433,
      "loss": 2.5712,
      "theoretical_loss": 3.385913399465169,
      "tokens_seen": 2344681472
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011520474749385286,
      "loss": 2.524,
      "theoretical_loss": 3.3859057532314996,
      "tokens_seen": 2344747008
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011518819746548138,
      "loss": 2.4456,
      "theoretical_loss": 3.385898107271378,
      "tokens_seen": 2344812544
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001151716474371099,
      "loss": 2.5761,
      "theoretical_loss": 3.385890461584787,
      "tokens_seen": 2344878080
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011515509740873843,
      "loss": 2.4662,
      "theoretical_loss": 3.3858828161717094,
      "tokens_seen": 2344943616
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011513854738036693,
      "loss": 2.4594,
      "theoretical_loss": 3.385875171032128,
      "tokens_seen": 2345009152
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011512199735199545,
      "loss": 2.504,
      "theoretical_loss": 3.3858675261660247,
      "tokens_seen": 2345074688
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011510544732362397,
      "loss": 2.3488,
      "theoretical_loss": 3.3858598815733822,
      "tokens_seen": 2345140224
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001150888972952525,
      "loss": 1.9532,
      "theoretical_loss": 3.3858522372541833,
      "tokens_seen": 2345205760
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011507234726688102,
      "loss": 2.6577,
      "theoretical_loss": 3.385844593208411,
      "tokens_seen": 2345271296
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011505579723850955,
      "loss": 2.8072,
      "theoretical_loss": 3.3858369494360465,
      "tokens_seen": 2345336832
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011503924721013807,
      "loss": 2.6223,
      "theoretical_loss": 3.385829305937074,
      "tokens_seen": 2345402368
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001150226971817666,
      "loss": 2.7536,
      "theoretical_loss": 3.385821662711475,
      "tokens_seen": 2345467904
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011500614715339512,
      "loss": 2.6629,
      "theoretical_loss": 3.385814019759233,
      "tokens_seen": 2345533440
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011498959712502364,
      "loss": 2.5738,
      "theoretical_loss": 3.3858063770803293,
      "tokens_seen": 2345598976
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011497304709665217,
      "loss": 2.6741,
      "theoretical_loss": 3.3857987346747476,
      "tokens_seen": 2345664512
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011495649706828069,
      "loss": 2.6613,
      "theoretical_loss": 3.38579109254247,
      "tokens_seen": 2345730048
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011493994703990921,
      "loss": 2.6149,
      "theoretical_loss": 3.3857834506834794,
      "tokens_seen": 2345795584
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011492339701153774,
      "loss": 2.444,
      "theoretical_loss": 3.385775809097758,
      "tokens_seen": 2345861120
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011490684698316626,
      "loss": 2.6332,
      "theoretical_loss": 3.3857681677852884,
      "tokens_seen": 2345926656
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011489029695479477,
      "loss": 2.4937,
      "theoretical_loss": 3.3857605267460538,
      "tokens_seen": 2345992192
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001148737469264233,
      "loss": 2.5742,
      "theoretical_loss": 3.385752885980036,
      "tokens_seen": 2346057728
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011485719689805182,
      "loss": 2.4897,
      "theoretical_loss": 3.385745245487218,
      "tokens_seen": 2346123264
    },
    {
      "epoch": 0.84,
      "objective/train/avg_token_score": 0.016279932111501694,
      "objective/train/avg_weight": 0.980861246585846,
      "objective/train/docs_used": 1321381,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6676526069641113,
      "objective/train/std_weight": 0.10505659878253937,
      "objective/train/theoretical_loss": 3.385737605267583,
      "objective/train/tokens_used": 2366648800,
      "theoretical_loss": 3.385737605267583,
      "tokens_seen": 2346188800
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011484064686968034,
      "loss": 2.4944,
      "theoretical_loss": 3.385737605267583,
      "tokens_seen": 2346188800
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011482409684130887,
      "loss": 2.5958,
      "theoretical_loss": 3.3857299653211124,
      "tokens_seen": 2346254336
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011480754681293739,
      "loss": 2.5499,
      "theoretical_loss": 3.3857223256477895,
      "tokens_seen": 2346319872
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011479099678456592,
      "loss": 2.609,
      "theoretical_loss": 3.3857146862475966,
      "tokens_seen": 2346385408
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011477444675619444,
      "loss": 2.6809,
      "theoretical_loss": 3.385707047120517,
      "tokens_seen": 2346450944
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011475789672782296,
      "loss": 2.7711,
      "theoretical_loss": 3.3856994082665324,
      "tokens_seen": 2346516480
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011474134669945149,
      "loss": 2.6825,
      "theoretical_loss": 3.385691769685626,
      "tokens_seen": 2346582016
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011472479667108001,
      "loss": 2.4828,
      "theoretical_loss": 3.38568413137778,
      "tokens_seen": 2346647552
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011470824664270854,
      "loss": 2.8464,
      "theoretical_loss": 3.3856764933429773,
      "tokens_seen": 2346713088
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011469169661433706,
      "loss": 2.5897,
      "theoretical_loss": 3.3856688555812005,
      "tokens_seen": 2346778624
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011467514658596558,
      "loss": 2.6267,
      "theoretical_loss": 3.3856612180924324,
      "tokens_seen": 2346844160
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001146585965575941,
      "loss": 2.1742,
      "theoretical_loss": 3.385653580876655,
      "tokens_seen": 2346909696
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011464204652922262,
      "loss": 2.6001,
      "theoretical_loss": 3.3856459439338513,
      "tokens_seen": 2346975232
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011462549650085114,
      "loss": 2.3619,
      "theoretical_loss": 3.3856383072640046,
      "tokens_seen": 2347040768
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011460894647247967,
      "loss": 3.0401,
      "theoretical_loss": 3.385630670867096,
      "tokens_seen": 2347106304
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011459239644410819,
      "loss": 2.8764,
      "theoretical_loss": 3.3856230347431095,
      "tokens_seen": 2347171840
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011457584641573671,
      "loss": 2.9611,
      "theoretical_loss": 3.385615398892027,
      "tokens_seen": 2347237376
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011455929638736524,
      "loss": 2.5856,
      "theoretical_loss": 3.3856077633138315,
      "tokens_seen": 2347302912
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011454274635899376,
      "loss": 2.8338,
      "theoretical_loss": 3.3856001280085053,
      "tokens_seen": 2347368448
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011452619633062229,
      "loss": 2.5345,
      "theoretical_loss": 3.3855924929760315,
      "tokens_seen": 2347433984
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011450964630225081,
      "loss": 2.7005,
      "theoretical_loss": 3.385584858216392,
      "tokens_seen": 2347499520
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011449309627387933,
      "loss": 1.9525,
      "theoretical_loss": 3.38557722372957,
      "tokens_seen": 2347565056
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011447654624550786,
      "loss": 2.4633,
      "theoretical_loss": 3.3855695895155478,
      "tokens_seen": 2347630592
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011445999621713638,
      "loss": 2.475,
      "theoretical_loss": 3.3855619555743086,
      "tokens_seen": 2347696128
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001144434461887649,
      "loss": 2.6344,
      "theoretical_loss": 3.3855543219058344,
      "tokens_seen": 2347761664
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011442689616039343,
      "loss": 2.6925,
      "theoretical_loss": 3.385546688510108,
      "tokens_seen": 2347827200
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011441034613202193,
      "loss": 2.4713,
      "theoretical_loss": 3.3855390553871128,
      "tokens_seen": 2347892736
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011439379610365045,
      "loss": 2.8122,
      "theoretical_loss": 3.3855314225368303,
      "tokens_seen": 2347958272
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011437724607527897,
      "loss": 2.5073,
      "theoretical_loss": 3.385523789959244,
      "tokens_seen": 2348023808
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001143606960469075,
      "loss": 2.3924,
      "theoretical_loss": 3.385516157654336,
      "tokens_seen": 2348089344
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011434414601853602,
      "loss": 2.6977,
      "theoretical_loss": 3.3855085256220887,
      "tokens_seen": 2348154880
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011432759599016455,
      "loss": 2.4033,
      "theoretical_loss": 3.3855008938624858,
      "tokens_seen": 2348220416
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011431104596179307,
      "loss": 2.393,
      "theoretical_loss": 3.3854932623755087,
      "tokens_seen": 2348285952
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001142944959334216,
      "loss": 2.4646,
      "theoretical_loss": 3.385485631161141,
      "tokens_seen": 2348351488
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011427794590505012,
      "loss": 2.5862,
      "theoretical_loss": 3.3854780002193654,
      "tokens_seen": 2348417024
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011426139587667864,
      "loss": 2.6771,
      "theoretical_loss": 3.3854703695501636,
      "tokens_seen": 2348482560
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011424484584830717,
      "loss": 2.7773,
      "theoretical_loss": 3.3854627391535193,
      "tokens_seen": 2348548096
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011422829581993569,
      "loss": 2.6951,
      "theoretical_loss": 3.3854551090294143,
      "tokens_seen": 2348613632
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011421174579156421,
      "loss": 2.5594,
      "theoretical_loss": 3.385447479177832,
      "tokens_seen": 2348679168
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011419519576319274,
      "loss": 2.5814,
      "theoretical_loss": 3.3854398495987548,
      "tokens_seen": 2348744704
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011417864573482126,
      "loss": 2.4912,
      "theoretical_loss": 3.385432220292165,
      "tokens_seen": 2348810240
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011416209570644977,
      "loss": 2.7365,
      "theoretical_loss": 3.3854245912580456,
      "tokens_seen": 2348875776
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001141455456780783,
      "loss": 2.4976,
      "theoretical_loss": 3.3854169624963797,
      "tokens_seen": 2348941312
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011412899564970682,
      "loss": 2.5751,
      "theoretical_loss": 3.385409334007149,
      "tokens_seen": 2349006848
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011411244562133535,
      "loss": 2.5642,
      "theoretical_loss": 3.385401705790337,
      "tokens_seen": 2349072384
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011409589559296387,
      "loss": 2.0887,
      "theoretical_loss": 3.385394077845926,
      "tokens_seen": 2349137920
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011407934556459239,
      "loss": 2.4371,
      "theoretical_loss": 3.3853864501738986,
      "tokens_seen": 2349203456
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011406279553622092,
      "loss": 2.5538,
      "theoretical_loss": 3.3853788227742374,
      "tokens_seen": 2349268992
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011404624550784944,
      "loss": 2.4394,
      "theoretical_loss": 3.385371195646926,
      "tokens_seen": 2349334528
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011402969547947797,
      "loss": 2.5626,
      "theoretical_loss": 3.3853635687919454,
      "tokens_seen": 2349400064
    },
    {
      "epoch": 0.84,
      "objective/train/avg_token_score": 0.01526274811476469,
      "objective/train/avg_weight": 0.9820370674133301,
      "objective/train/docs_used": 1323513,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.7493014335632324,
      "objective/train/std_weight": 0.10053753852844238,
      "objective/train/theoretical_loss": 3.38535594220928,
      "objective/train/tokens_used": 2369925600,
      "theoretical_loss": 3.38535594220928,
      "tokens_seen": 2349465600
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011401314545110649,
      "loss": 2.3751,
      "theoretical_loss": 3.38535594220928,
      "tokens_seen": 2349465600
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011399659542273501,
      "loss": 2.1892,
      "theoretical_loss": 3.3853483158989115,
      "tokens_seen": 2349531136
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011398004539436354,
      "loss": 2.5694,
      "theoretical_loss": 3.3853406898608225,
      "tokens_seen": 2349596672
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011396349536599206,
      "loss": 2.5351,
      "theoretical_loss": 3.385333064094996,
      "tokens_seen": 2349662208
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011394694533762059,
      "loss": 2.4342,
      "theoretical_loss": 3.385325438601415,
      "tokens_seen": 2349727744
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011393039530924911,
      "loss": 2.4798,
      "theoretical_loss": 3.3853178133800617,
      "tokens_seen": 2349793280
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011391384528087762,
      "loss": 2.5492,
      "theoretical_loss": 3.385310188430919,
      "tokens_seen": 2349858816
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011389729525250614,
      "loss": 2.5408,
      "theoretical_loss": 3.3853025637539695,
      "tokens_seen": 2349924352
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011388074522413467,
      "loss": 2.7034,
      "theoretical_loss": 3.385294939349196,
      "tokens_seen": 2349989888
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011386419519576319,
      "loss": 2.6187,
      "theoretical_loss": 3.385287315216581,
      "tokens_seen": 2350055424
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011384764516739172,
      "loss": 2.3702,
      "theoretical_loss": 3.385279691356107,
      "tokens_seen": 2350120960
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011383109513902024,
      "loss": 2.566,
      "theoretical_loss": 3.3852720677677577,
      "tokens_seen": 2350186496
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011381454511064876,
      "loss": 2.3697,
      "theoretical_loss": 3.385264444451515,
      "tokens_seen": 2350252032
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011379799508227729,
      "loss": 2.7755,
      "theoretical_loss": 3.3852568214073617,
      "tokens_seen": 2350317568
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011378144505390581,
      "loss": 2.5744,
      "theoretical_loss": 3.38524919863528,
      "tokens_seen": 2350383104
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011376489502553434,
      "loss": 2.6788,
      "theoretical_loss": 3.3852415761352534,
      "tokens_seen": 2350448640
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011374834499716286,
      "loss": 2.6198,
      "theoretical_loss": 3.3852339539072647,
      "tokens_seen": 2350514176
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011373179496879138,
      "loss": 2.4882,
      "theoretical_loss": 3.3852263319512956,
      "tokens_seen": 2350579712
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011371524494041991,
      "loss": 2.6438,
      "theoretical_loss": 3.38521871026733,
      "tokens_seen": 2350645248
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011369869491204843,
      "loss": 2.5506,
      "theoretical_loss": 3.3852110888553497,
      "tokens_seen": 2350710784
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011368214488367696,
      "loss": 2.4982,
      "theoretical_loss": 3.385203467715338,
      "tokens_seen": 2350776320
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011366559485530545,
      "loss": 2.6232,
      "theoretical_loss": 3.385195846847277,
      "tokens_seen": 2350841856
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011364904482693398,
      "loss": 2.8233,
      "theoretical_loss": 3.38518822625115,
      "tokens_seen": 2350907392
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001136324947985625,
      "loss": 2.5106,
      "theoretical_loss": 3.38518060592694,
      "tokens_seen": 2350972928
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011361594477019102,
      "loss": 2.3554,
      "theoretical_loss": 3.3851729858746284,
      "tokens_seen": 2351038464
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011359939474181955,
      "loss": 2.5123,
      "theoretical_loss": 3.385165366094199,
      "tokens_seen": 2351104000
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011358284471344807,
      "loss": 2.1479,
      "theoretical_loss": 3.3851577465856346,
      "tokens_seen": 2351169536
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001135662946850766,
      "loss": 2.7299,
      "theoretical_loss": 3.3851501273489175,
      "tokens_seen": 2351235072
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011354974465670512,
      "loss": 2.7205,
      "theoretical_loss": 3.38514250838403,
      "tokens_seen": 2351300608
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011353319462833364,
      "loss": 2.5063,
      "theoretical_loss": 3.3851348896909563,
      "tokens_seen": 2351366144
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011351664459996217,
      "loss": 2.6454,
      "theoretical_loss": 3.3851272712696776,
      "tokens_seen": 2351431680
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011350009457159069,
      "loss": 2.5135,
      "theoretical_loss": 3.385119653120177,
      "tokens_seen": 2351497216
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011350009457159069,
      "loss": 9.6417,
      "theoretical_loss": 3.3851120352424378,
      "tokens_seen": 2351562752
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011348354454321922,
      "loss": 2.5347,
      "theoretical_loss": 3.385104417636442,
      "tokens_seen": 2351628288
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011346699451484774,
      "loss": 2.4093,
      "theoretical_loss": 3.385096800302173,
      "tokens_seen": 2351693824
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011345044448647626,
      "loss": 2.624,
      "theoretical_loss": 3.385089183239613,
      "tokens_seen": 2351759360
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011343389445810479,
      "loss": 2.838,
      "theoretical_loss": 3.3850815664487452,
      "tokens_seen": 2351824896
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001134173444297333,
      "loss": 2.5145,
      "theoretical_loss": 3.385073949929552,
      "tokens_seen": 2351890432
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011340079440136182,
      "loss": 2.59,
      "theoretical_loss": 3.3850663336820164,
      "tokens_seen": 2351955968
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011338424437299035,
      "loss": 2.2927,
      "theoretical_loss": 3.3850587177061207,
      "tokens_seen": 2352021504
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011336769434461887,
      "loss": 2.794,
      "theoretical_loss": 3.385051102001848,
      "tokens_seen": 2352087040
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001133511443162474,
      "loss": 2.6576,
      "theoretical_loss": 3.385043486569181,
      "tokens_seen": 2352152576
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011333459428787592,
      "loss": 2.6523,
      "theoretical_loss": 3.3850358714081024,
      "tokens_seen": 2352218112
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011331804425950444,
      "loss": 2.507,
      "theoretical_loss": 3.385028256518595,
      "tokens_seen": 2352283648
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011330149423113297,
      "loss": 2.5712,
      "theoretical_loss": 3.385020641900642,
      "tokens_seen": 2352349184
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011328494420276149,
      "loss": 2.5146,
      "theoretical_loss": 3.385013027554225,
      "tokens_seen": 2352414720
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011326839417439001,
      "loss": 2.6981,
      "theoretical_loss": 3.3850054134793277,
      "tokens_seen": 2352480256
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011325184414601854,
      "loss": 2.4696,
      "theoretical_loss": 3.3849977996759324,
      "tokens_seen": 2352545792
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011323529411764706,
      "loss": 2.3921,
      "theoretical_loss": 3.384990186144022,
      "tokens_seen": 2352611328
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011321874408927559,
      "loss": 2.4124,
      "theoretical_loss": 3.3849825728835796,
      "tokens_seen": 2352676864
    },
    {
      "epoch": 0.84,
      "objective/train/avg_token_score": 0.014310961589217186,
      "objective/train/avg_weight": 0.9831714630126953,
      "objective/train/docs_used": 1325482,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.225241184234619,
      "objective/train/std_weight": 0.09955517202615738,
      "objective/train/theoretical_loss": 3.3849749598945875,
      "objective/train/tokens_used": 2373202400,
      "theoretical_loss": 3.3849749598945875,
      "tokens_seen": 2352742400
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011320219406090411,
      "loss": 2.6601,
      "theoretical_loss": 3.3849749598945875,
      "tokens_seen": 2352742400
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011318564403253263,
      "loss": 2.4845,
      "theoretical_loss": 3.3849673471770285,
      "tokens_seen": 2352807936
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011316909400416114,
      "loss": 2.4477,
      "theoretical_loss": 3.3849597347308853,
      "tokens_seen": 2352873472
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011315254397578967,
      "loss": 2.4587,
      "theoretical_loss": 3.3849521225561414,
      "tokens_seen": 2352939008
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011313599394741819,
      "loss": 2.5548,
      "theoretical_loss": 3.3849445106527787,
      "tokens_seen": 2353004544
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011311944391904672,
      "loss": 2.4407,
      "theoretical_loss": 3.38493689902078,
      "tokens_seen": 2353070080
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011310289389067524,
      "loss": 2.4837,
      "theoretical_loss": 3.3849292876601287,
      "tokens_seen": 2353135616
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011308634386230376,
      "loss": 2.2193,
      "theoretical_loss": 3.384921676570807,
      "tokens_seen": 2353201152
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011306979383393229,
      "loss": 2.4644,
      "theoretical_loss": 3.384914065752798,
      "tokens_seen": 2353266688
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011305324380556081,
      "loss": 2.8254,
      "theoretical_loss": 3.3849064552060844,
      "tokens_seen": 2353332224
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011303669377718934,
      "loss": 2.3593,
      "theoretical_loss": 3.3848988449306487,
      "tokens_seen": 2353397760
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011302014374881786,
      "loss": 2.6065,
      "theoretical_loss": 3.3848912349264744,
      "tokens_seen": 2353463296
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011300359372044638,
      "loss": 2.4261,
      "theoretical_loss": 3.384883625193543,
      "tokens_seen": 2353528832
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011298704369207491,
      "loss": 2.4045,
      "theoretical_loss": 3.3848760157318387,
      "tokens_seen": 2353594368
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011297049366370343,
      "loss": 2.7862,
      "theoretical_loss": 3.3848684065413437,
      "tokens_seen": 2353659904
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011295394363533196,
      "loss": 2.2884,
      "theoretical_loss": 3.38486079762204,
      "tokens_seen": 2353725440
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011293739360696045,
      "loss": 2.1905,
      "theoretical_loss": 3.3848531889739117,
      "tokens_seen": 2353790976
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011292084357858898,
      "loss": 2.228,
      "theoretical_loss": 3.3848455805969406,
      "tokens_seen": 2353856512
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001129042935502175,
      "loss": 2.7079,
      "theoretical_loss": 3.3848379724911104,
      "tokens_seen": 2353922048
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011288774352184603,
      "loss": 2.5866,
      "theoretical_loss": 3.384830364656403,
      "tokens_seen": 2353987584
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011287119349347455,
      "loss": 2.7692,
      "theoretical_loss": 3.3848227570928016,
      "tokens_seen": 2354053120
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011285464346510307,
      "loss": 2.7061,
      "theoretical_loss": 3.384815149800289,
      "tokens_seen": 2354118656
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001128380934367316,
      "loss": 2.5126,
      "theoretical_loss": 3.384807542778848,
      "tokens_seen": 2354184192
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011282154340836012,
      "loss": 2.5181,
      "theoretical_loss": 3.384799936028461,
      "tokens_seen": 2354249728
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011280499337998865,
      "loss": 2.6774,
      "theoretical_loss": 3.3847923295491116,
      "tokens_seen": 2354315264
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011278844335161717,
      "loss": 2.4711,
      "theoretical_loss": 3.384784723340782,
      "tokens_seen": 2354380800
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001127718933232457,
      "loss": 2.7644,
      "theoretical_loss": 3.384777117403455,
      "tokens_seen": 2354446336
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011275534329487422,
      "loss": 2.5849,
      "theoretical_loss": 3.3847695117371135,
      "tokens_seen": 2354511872
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011273879326650274,
      "loss": 2.7566,
      "theoretical_loss": 3.384761906341741,
      "tokens_seen": 2354577408
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011272224323813127,
      "loss": 2.4866,
      "theoretical_loss": 3.3847543012173187,
      "tokens_seen": 2354642944
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011270569320975979,
      "loss": 2.5809,
      "theoretical_loss": 3.384746696363831,
      "tokens_seen": 2354708480
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001126891431813883,
      "loss": 2.5863,
      "theoretical_loss": 3.3847390917812596,
      "tokens_seen": 2354774016
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011267259315301682,
      "loss": 2.6072,
      "theoretical_loss": 3.3847314874695877,
      "tokens_seen": 2354839552
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011265604312464535,
      "loss": 2.6493,
      "theoretical_loss": 3.3847238834287987,
      "tokens_seen": 2354905088
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011263949309627387,
      "loss": 2.4502,
      "theoretical_loss": 3.384716279658875,
      "tokens_seen": 2354970624
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001126229430679024,
      "loss": 2.5032,
      "theoretical_loss": 3.3847086761597986,
      "tokens_seen": 2355036160
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011260639303953092,
      "loss": 2.4068,
      "theoretical_loss": 3.3847010729315534,
      "tokens_seen": 2355101696
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011258984301115944,
      "loss": 2.6777,
      "theoretical_loss": 3.384693469974122,
      "tokens_seen": 2355167232
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011257329298278797,
      "loss": 2.4962,
      "theoretical_loss": 3.3846858672874864,
      "tokens_seen": 2355232768
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011255674295441649,
      "loss": 2.6649,
      "theoretical_loss": 3.384678264871631,
      "tokens_seen": 2355298304
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011254019292604502,
      "loss": 2.4729,
      "theoretical_loss": 3.384670662726537,
      "tokens_seen": 2355363840
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011252364289767354,
      "loss": 2.4598,
      "theoretical_loss": 3.3846630608521884,
      "tokens_seen": 2355429376
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011250709286930206,
      "loss": 2.4917,
      "theoretical_loss": 3.384655459248567,
      "tokens_seen": 2355494912
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011249054284093059,
      "loss": 2.6368,
      "theoretical_loss": 3.3846478579156565,
      "tokens_seen": 2355560448
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011247399281255911,
      "loss": 2.661,
      "theoretical_loss": 3.3846402568534395,
      "tokens_seen": 2355625984
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011245744278418764,
      "loss": 2.5631,
      "theoretical_loss": 3.3846326560618984,
      "tokens_seen": 2355691520
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011244089275581615,
      "loss": 2.5586,
      "theoretical_loss": 3.384625055541017,
      "tokens_seen": 2355757056
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011242434272744467,
      "loss": 2.4872,
      "theoretical_loss": 3.3846174552907766,
      "tokens_seen": 2355822592
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001124077926990732,
      "loss": 2.7241,
      "theoretical_loss": 3.3846098553111617,
      "tokens_seen": 2355888128
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011239124267070172,
      "loss": 2.3041,
      "theoretical_loss": 3.3846022556021538,
      "tokens_seen": 2355953664
    },
    {
      "epoch": 0.84,
      "objective/train/avg_token_score": 0.013297040946781635,
      "objective/train/avg_weight": 0.9843533039093018,
      "objective/train/docs_used": 1326516,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.984947681427002,
      "objective/train/std_weight": 0.08192041516304016,
      "objective/train/theoretical_loss": 3.3845946561637366,
      "objective/train/tokens_used": 2376479200,
      "theoretical_loss": 3.3845946561637366,
      "tokens_seen": 2356019200
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011237469264233024,
      "loss": 2.4548,
      "theoretical_loss": 3.3845946561637366,
      "tokens_seen": 2356019200
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011235814261395877,
      "loss": 2.5387,
      "theoretical_loss": 3.3845870569958927,
      "tokens_seen": 2356084736
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011234159258558729,
      "loss": 2.5212,
      "theoretical_loss": 3.384579458098605,
      "tokens_seen": 2356150272
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011232504255721581,
      "loss": 2.6952,
      "theoretical_loss": 3.3845718594718557,
      "tokens_seen": 2356215808
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011230849252884434,
      "loss": 2.674,
      "theoretical_loss": 3.3845642611156284,
      "tokens_seen": 2356281344
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011229194250047286,
      "loss": 2.6412,
      "theoretical_loss": 3.3845566630299055,
      "tokens_seen": 2356346880
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011227539247210139,
      "loss": 2.581,
      "theoretical_loss": 3.384549065214671,
      "tokens_seen": 2356412416
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011225884244372991,
      "loss": 2.6085,
      "theoretical_loss": 3.384541467669906,
      "tokens_seen": 2356477952
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011224229241535843,
      "loss": 2.2192,
      "theoretical_loss": 3.384533870395594,
      "tokens_seen": 2356543488
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011222574238698696,
      "loss": 2.2567,
      "theoretical_loss": 3.3845262733917183,
      "tokens_seen": 2356609024
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011220919235861548,
      "loss": 2.5495,
      "theoretical_loss": 3.384518676658262,
      "tokens_seen": 2356674560
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011219264233024398,
      "loss": 2.7431,
      "theoretical_loss": 3.3845110801952063,
      "tokens_seen": 2356740096
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001121760923018725,
      "loss": 2.5779,
      "theoretical_loss": 3.384503484002536,
      "tokens_seen": 2356805632
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011215954227350103,
      "loss": 2.0686,
      "theoretical_loss": 3.3844958880802327,
      "tokens_seen": 2356871168
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011214299224512955,
      "loss": 2.7434,
      "theoretical_loss": 3.38448829242828,
      "tokens_seen": 2356936704
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011212644221675808,
      "loss": 2.5938,
      "theoretical_loss": 3.3844806970466603,
      "tokens_seen": 2357002240
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001121098921883866,
      "loss": 2.745,
      "theoretical_loss": 3.384473101935357,
      "tokens_seen": 2357067776
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011209334216001512,
      "loss": 2.551,
      "theoretical_loss": 3.384465507094352,
      "tokens_seen": 2357133312
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011207679213164365,
      "loss": 2.6026,
      "theoretical_loss": 3.384457912523629,
      "tokens_seen": 2357198848
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011206024210327217,
      "loss": 2.7563,
      "theoretical_loss": 3.3844503182231707,
      "tokens_seen": 2357264384
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001120436920749007,
      "loss": 2.4489,
      "theoretical_loss": 3.3844427241929598,
      "tokens_seen": 2357329920
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011202714204652922,
      "loss": 2.6778,
      "theoretical_loss": 3.384435130432979,
      "tokens_seen": 2357395456
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011201059201815774,
      "loss": 2.83,
      "theoretical_loss": 3.3844275369432113,
      "tokens_seen": 2357460992
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011199404198978628,
      "loss": 2.7581,
      "theoretical_loss": 3.3844199437236404,
      "tokens_seen": 2357526528
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001119774919614148,
      "loss": 2.7064,
      "theoretical_loss": 3.384412350774248,
      "tokens_seen": 2357592064
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011196094193304333,
      "loss": 2.6843,
      "theoretical_loss": 3.384404758095018,
      "tokens_seen": 2357657600
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011194439190467183,
      "loss": 2.6647,
      "theoretical_loss": 3.3843971656859315,
      "tokens_seen": 2357723136
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011192784187630035,
      "loss": 2.628,
      "theoretical_loss": 3.384389573546974,
      "tokens_seen": 2357788672
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011191129184792887,
      "loss": 2.4572,
      "theoretical_loss": 3.384381981678126,
      "tokens_seen": 2357854208
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001118947418195574,
      "loss": 2.7568,
      "theoretical_loss": 3.3843743900793717,
      "tokens_seen": 2357919744
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011187819179118592,
      "loss": 2.8025,
      "theoretical_loss": 3.384366798750694,
      "tokens_seen": 2357985280
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011186164176281445,
      "loss": 2.5973,
      "theoretical_loss": 3.384359207692075,
      "tokens_seen": 2358050816
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011184509173444297,
      "loss": 2.5354,
      "theoretical_loss": 3.384351616903498,
      "tokens_seen": 2358116352
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001118285417060715,
      "loss": 1.8902,
      "theoretical_loss": 3.3843440263849462,
      "tokens_seen": 2358181888
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011181199167770002,
      "loss": 2.6729,
      "theoretical_loss": 3.3843364361364023,
      "tokens_seen": 2358247424
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011179544164932854,
      "loss": 2.511,
      "theoretical_loss": 3.3843288461578487,
      "tokens_seen": 2358312960
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011177889162095707,
      "loss": 2.8758,
      "theoretical_loss": 3.384321256449269,
      "tokens_seen": 2358378496
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011176234159258559,
      "loss": 2.6626,
      "theoretical_loss": 3.3843136670106455,
      "tokens_seen": 2358444032
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011174579156421411,
      "loss": 2.757,
      "theoretical_loss": 3.384306077841962,
      "tokens_seen": 2358509568
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011172924153584264,
      "loss": 2.6154,
      "theoretical_loss": 3.3842984889432,
      "tokens_seen": 2358575104
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011171269150747116,
      "loss": 2.2877,
      "theoretical_loss": 3.3842909003143435,
      "tokens_seen": 2358640640
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011169614147909967,
      "loss": 2.6189,
      "theoretical_loss": 3.3842833119553752,
      "tokens_seen": 2358706176
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001116795914507282,
      "loss": 2.5357,
      "theoretical_loss": 3.3842757238662777,
      "tokens_seen": 2358771712
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011166304142235672,
      "loss": 3.5453,
      "theoretical_loss": 3.384268136047034,
      "tokens_seen": 2358837248
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011164649139398524,
      "loss": 2.5065,
      "theoretical_loss": 3.3842605484976276,
      "tokens_seen": 2358902784
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011162994136561377,
      "loss": 2.7283,
      "theoretical_loss": 3.3842529612180408,
      "tokens_seen": 2358968320
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011161339133724229,
      "loss": 2.5353,
      "theoretical_loss": 3.3842453742082563,
      "tokens_seen": 2359033856
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011159684130887082,
      "loss": 2.6794,
      "theoretical_loss": 3.3842377874682574,
      "tokens_seen": 2359099392
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011158029128049934,
      "loss": 3.017,
      "theoretical_loss": 3.3842302009980267,
      "tokens_seen": 2359164928
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011156374125212786,
      "loss": 2.6094,
      "theoretical_loss": 3.384222614797548,
      "tokens_seen": 2359230464
    },
    {
      "epoch": 0.84,
      "objective/train/avg_token_score": 0.027562158182263374,
      "objective/train/avg_weight": 0.9676087498664856,
      "objective/train/docs_used": 1327582,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.1266069412231445,
      "objective/train/std_weight": 0.12888839840888977,
      "objective/train/theoretical_loss": 3.384215028866803,
      "objective/train/tokens_used": 2379756000,
      "theoretical_loss": 3.384215028866803,
      "tokens_seen": 2359296000
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011154719122375639,
      "loss": 2.6358,
      "theoretical_loss": 3.384215028866803,
      "tokens_seen": 2359296000
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011153064119538491,
      "loss": 2.7736,
      "theoretical_loss": 3.3842074432057756,
      "tokens_seen": 2359361536
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011151409116701344,
      "loss": 2.6446,
      "theoretical_loss": 3.384199857814448,
      "tokens_seen": 2359427072
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011149754113864196,
      "loss": 2.6653,
      "theoretical_loss": 3.3841922726928035,
      "tokens_seen": 2359492608
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011148099111027048,
      "loss": 2.5538,
      "theoretical_loss": 3.384184687840825,
      "tokens_seen": 2359558144
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011146444108189901,
      "loss": 2.6027,
      "theoretical_loss": 3.3841771032584953,
      "tokens_seen": 2359623680
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001114478910535275,
      "loss": 2.7229,
      "theoretical_loss": 3.3841695189457974,
      "tokens_seen": 2359689216
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011143134102515603,
      "loss": 2.6157,
      "theoretical_loss": 3.384161934902714,
      "tokens_seen": 2359754752
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011141479099678455,
      "loss": 2.7283,
      "theoretical_loss": 3.384154351129229,
      "tokens_seen": 2359820288
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011139824096841308,
      "loss": 2.4969,
      "theoretical_loss": 3.384146767625324,
      "tokens_seen": 2359885824
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001113816909400416,
      "loss": 2.7396,
      "theoretical_loss": 3.3841391843909827,
      "tokens_seen": 2359951360
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011136514091167014,
      "loss": 2.7063,
      "theoretical_loss": 3.384131601426188,
      "tokens_seen": 2360016896
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011134859088329866,
      "loss": 2.6578,
      "theoretical_loss": 3.384124018730922,
      "tokens_seen": 2360082432
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011133204085492719,
      "loss": 2.5686,
      "theoretical_loss": 3.3841164363051686,
      "tokens_seen": 2360147968
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011131549082655571,
      "loss": 2.5309,
      "theoretical_loss": 3.384108854148911,
      "tokens_seen": 2360213504
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011129894079818423,
      "loss": 2.7075,
      "theoretical_loss": 3.3841012722621313,
      "tokens_seen": 2360279040
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011128239076981276,
      "loss": 2.6787,
      "theoretical_loss": 3.3840936906448125,
      "tokens_seen": 2360344576
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011126584074144128,
      "loss": 2.7278,
      "theoretical_loss": 3.3840861092969377,
      "tokens_seen": 2360410112
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001112492907130698,
      "loss": 2.4527,
      "theoretical_loss": 3.38407852821849,
      "tokens_seen": 2360475648
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011123274068469833,
      "loss": 2.7292,
      "theoretical_loss": 3.3840709474094526,
      "tokens_seen": 2360541184
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011121619065632683,
      "loss": 2.6335,
      "theoretical_loss": 3.3840633668698077,
      "tokens_seen": 2360606720
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011119964062795535,
      "loss": 2.5133,
      "theoretical_loss": 3.384055786599539,
      "tokens_seen": 2360672256
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011118309059958387,
      "loss": 2.5962,
      "theoretical_loss": 3.384048206598629,
      "tokens_seen": 2360737792
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001111665405712124,
      "loss": 2.6626,
      "theoretical_loss": 3.3840406268670606,
      "tokens_seen": 2360803328
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011114999054284092,
      "loss": 2.8434,
      "theoretical_loss": 3.384033047404817,
      "tokens_seen": 2360868864
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011113344051446945,
      "loss": 2.5486,
      "theoretical_loss": 3.384025468211881,
      "tokens_seen": 2360934400
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011111689048609797,
      "loss": 2.5246,
      "theoretical_loss": 3.384017889288236,
      "tokens_seen": 2360999936
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001111003404577265,
      "loss": 2.6825,
      "theoretical_loss": 3.384010310633864,
      "tokens_seen": 2361065472
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011108379042935502,
      "loss": 2.686,
      "theoretical_loss": 3.384002732248749,
      "tokens_seen": 2361131008
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011106724040098354,
      "loss": 2.557,
      "theoretical_loss": 3.383995154132873,
      "tokens_seen": 2361196544
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011105069037261207,
      "loss": 2.8552,
      "theoretical_loss": 3.38398757628622,
      "tokens_seen": 2361262080
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011103414034424059,
      "loss": 2.7518,
      "theoretical_loss": 3.383979998708772,
      "tokens_seen": 2361327616
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011101759031586911,
      "loss": 2.3094,
      "theoretical_loss": 3.3839724214005127,
      "tokens_seen": 2361393152
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011100104028749764,
      "loss": 2.6705,
      "theoretical_loss": 3.3839648443614245,
      "tokens_seen": 2361458688
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011098449025912616,
      "loss": 2.6133,
      "theoretical_loss": 3.383957267591491,
      "tokens_seen": 2361524224
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011096794023075467,
      "loss": 2.8413,
      "theoretical_loss": 3.3839496910906943,
      "tokens_seen": 2361589760
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001109513902023832,
      "loss": 2.4957,
      "theoretical_loss": 3.3839421148590185,
      "tokens_seen": 2361655296
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011093484017401172,
      "loss": 2.8568,
      "theoretical_loss": 3.3839345388964457,
      "tokens_seen": 2361720832
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011091829014564025,
      "loss": 2.4745,
      "theoretical_loss": 3.383926963202959,
      "tokens_seen": 2361786368
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011090174011726877,
      "loss": 2.3976,
      "theoretical_loss": 3.3839193877785414,
      "tokens_seen": 2361851904
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011088519008889729,
      "loss": 2.3604,
      "theoretical_loss": 3.3839118126231758,
      "tokens_seen": 2361917440
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011086864006052582,
      "loss": 2.5179,
      "theoretical_loss": 3.3839042377368456,
      "tokens_seen": 2361982976
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011085209003215434,
      "loss": 2.7214,
      "theoretical_loss": 3.3838966631195335,
      "tokens_seen": 2362048512
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011083554000378287,
      "loss": 2.3876,
      "theoretical_loss": 3.3838890887712223,
      "tokens_seen": 2362114048
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011081898997541139,
      "loss": 2.8632,
      "theoretical_loss": 3.383881514691896,
      "tokens_seen": 2362179584
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011080243994703991,
      "loss": 2.9728,
      "theoretical_loss": 3.3838739408815357,
      "tokens_seen": 2362245120
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011078588991866844,
      "loss": 2.6943,
      "theoretical_loss": 3.383866367340126,
      "tokens_seen": 2362310656
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011076933989029696,
      "loss": 2.404,
      "theoretical_loss": 3.3838587940676494,
      "tokens_seen": 2362376192
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011075278986192549,
      "loss": 2.6416,
      "theoretical_loss": 3.3838512210640888,
      "tokens_seen": 2362441728
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011073623983355401,
      "loss": 2.543,
      "theoretical_loss": 3.383843648329427,
      "tokens_seen": 2362507264
    },
    {
      "epoch": 0.84,
      "objective/train/avg_token_score": 0.007959910668432713,
      "objective/train/avg_weight": 0.9906386733055115,
      "objective/train/docs_used": 1329627,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.674154758453369,
      "objective/train/std_weight": 0.04970913007855415,
      "objective/train/theoretical_loss": 3.3838360758636474,
      "objective/train/tokens_used": 2383032800,
      "theoretical_loss": 3.3838360758636474,
      "tokens_seen": 2362572800
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011071968980518252,
      "loss": 2.7509,
      "theoretical_loss": 3.3838360758636474,
      "tokens_seen": 2362572800
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011070313977681104,
      "loss": 2.5362,
      "theoretical_loss": 3.383828503666733,
      "tokens_seen": 2362638336
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011068658974843957,
      "loss": 2.8199,
      "theoretical_loss": 3.3838209317386667,
      "tokens_seen": 2362703872
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011067003972006809,
      "loss": 2.6355,
      "theoretical_loss": 3.383813360079431,
      "tokens_seen": 2362769408
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011065348969169662,
      "loss": 2.4796,
      "theoretical_loss": 3.3838057886890094,
      "tokens_seen": 2362834944
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011063693966332514,
      "loss": 2.5958,
      "theoretical_loss": 3.383798217567385,
      "tokens_seen": 2362900480
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011062038963495366,
      "loss": 2.6922,
      "theoretical_loss": 3.3837906467145404,
      "tokens_seen": 2362966016
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011060383960658219,
      "loss": 2.597,
      "theoretical_loss": 3.383783076130459,
      "tokens_seen": 2363031552
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011058728957821071,
      "loss": 2.482,
      "theoretical_loss": 3.383775505815124,
      "tokens_seen": 2363097088
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011057073954983924,
      "loss": 2.6187,
      "theoretical_loss": 3.3837679357685175,
      "tokens_seen": 2363162624
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011055418952146776,
      "loss": 2.6146,
      "theoretical_loss": 3.3837603659906232,
      "tokens_seen": 2363228160
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011053763949309628,
      "loss": 2.8351,
      "theoretical_loss": 3.383752796481424,
      "tokens_seen": 2363293696
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011052108946472481,
      "loss": 2.7139,
      "theoretical_loss": 3.3837452272409028,
      "tokens_seen": 2363359232
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011050453943635333,
      "loss": 2.6452,
      "theoretical_loss": 3.3837376582690424,
      "tokens_seen": 2363424768
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011048798940798186,
      "loss": 2.4493,
      "theoretical_loss": 3.3837300895658267,
      "tokens_seen": 2363490304
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011047143937961035,
      "loss": 2.6881,
      "theoretical_loss": 3.3837225211312374,
      "tokens_seen": 2363555840
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011045488935123888,
      "loss": 2.5294,
      "theoretical_loss": 3.3837149529652586,
      "tokens_seen": 2363621376
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001104383393228674,
      "loss": 2.6748,
      "theoretical_loss": 3.383707385067873,
      "tokens_seen": 2363686912
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011042178929449592,
      "loss": 2.716,
      "theoretical_loss": 3.383699817439064,
      "tokens_seen": 2363752448
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011040523926612445,
      "loss": 2.7262,
      "theoretical_loss": 3.3836922500788136,
      "tokens_seen": 2363817984
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011038868923775297,
      "loss": 2.4945,
      "theoretical_loss": 3.3836846829871057,
      "tokens_seen": 2363883520
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001103721392093815,
      "loss": 2.4207,
      "theoretical_loss": 3.3836771161639225,
      "tokens_seen": 2363949056
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011035558918101002,
      "loss": 2.654,
      "theoretical_loss": 3.383669549609248,
      "tokens_seen": 2364014592
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011033903915263854,
      "loss": 2.2834,
      "theoretical_loss": 3.383661983323065,
      "tokens_seen": 2364080128
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011032248912426707,
      "loss": 2.4673,
      "theoretical_loss": 3.383654417305356,
      "tokens_seen": 2364145664
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011030593909589559,
      "loss": 2.4241,
      "theoretical_loss": 3.3836468515561045,
      "tokens_seen": 2364211200
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011028938906752412,
      "loss": 2.6959,
      "theoretical_loss": 3.3836392860752933,
      "tokens_seen": 2364276736
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011027283903915264,
      "loss": 2.3165,
      "theoretical_loss": 3.3836317208629056,
      "tokens_seen": 2364342272
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011025628901078116,
      "loss": 2.6108,
      "theoretical_loss": 3.383624155918924,
      "tokens_seen": 2364407808
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011023973898240969,
      "loss": 2.5023,
      "theoretical_loss": 3.3836165912433325,
      "tokens_seen": 2364473344
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001102231889540382,
      "loss": 2.724,
      "theoretical_loss": 3.3836090268361136,
      "tokens_seen": 2364538880
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011020663892566672,
      "loss": 2.5997,
      "theoretical_loss": 3.38360146269725,
      "tokens_seen": 2364604416
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011019008889729525,
      "loss": 2.5849,
      "theoretical_loss": 3.383593898826725,
      "tokens_seen": 2364669952
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011017353886892377,
      "loss": 2.5504,
      "theoretical_loss": 3.3835863352245217,
      "tokens_seen": 2364735488
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0001101569888405523,
      "loss": 2.4288,
      "theoretical_loss": 3.383578771890623,
      "tokens_seen": 2364801024
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011014043881218082,
      "loss": 2.7097,
      "theoretical_loss": 3.3835712088250123,
      "tokens_seen": 2364866560
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011012388878380934,
      "loss": 2.2732,
      "theoretical_loss": 3.383563646027672,
      "tokens_seen": 2364932096
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011010733875543787,
      "loss": 2.616,
      "theoretical_loss": 3.383556083498586,
      "tokens_seen": 2364997632
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011009078872706639,
      "loss": 2.3817,
      "theoretical_loss": 3.3835485212377368,
      "tokens_seen": 2365063168
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011007423869869491,
      "loss": 2.5025,
      "theoretical_loss": 3.3835409592451073,
      "tokens_seen": 2365128704
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011005768867032344,
      "loss": 2.7735,
      "theoretical_loss": 3.383533397520681,
      "tokens_seen": 2365194240
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011004113864195196,
      "loss": 2.7878,
      "theoretical_loss": 3.383525836064441,
      "tokens_seen": 2365259776
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011002458861358049,
      "loss": 2.4221,
      "theoretical_loss": 3.38351827487637,
      "tokens_seen": 2365325312
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00011000803858520901,
      "loss": 2.6491,
      "theoretical_loss": 3.383510713956451,
      "tokens_seen": 2365390848
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00010999148855683753,
      "loss": 2.619,
      "theoretical_loss": 3.3835031533046678,
      "tokens_seen": 2365456384
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00010997493852846604,
      "loss": 2.3643,
      "theoretical_loss": 3.3834955929210024,
      "tokens_seen": 2365521920
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00010995838850009457,
      "loss": 2.3713,
      "theoretical_loss": 3.3834880328054386,
      "tokens_seen": 2365587456
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00010994183847172309,
      "loss": 2.7543,
      "theoretical_loss": 3.383480472957959,
      "tokens_seen": 2365652992
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00010992528844335162,
      "loss": 2.296,
      "theoretical_loss": 3.3834729133785473,
      "tokens_seen": 2365718528
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00010990873841498014,
      "loss": 2.3884,
      "theoretical_loss": 3.383465354067186,
      "tokens_seen": 2365784064
    },
    {
      "epoch": 0.84,
      "objective/train/avg_token_score": 0.022031933069229126,
      "objective/train/avg_weight": 0.9740912318229675,
      "objective/train/docs_used": 1331397,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 7.130276679992676,
      "objective/train/std_weight": 0.13993236422538757,
      "objective/train/theoretical_loss": 3.3834577950238582,
      "objective/train/tokens_used": 2386309600,
      "theoretical_loss": 3.3834577950238582,
      "tokens_seen": 2365849600
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00010989218838660866,
      "loss": 2.1537,
      "theoretical_loss": 3.3834577950238582,
      "tokens_seen": 2365849600
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.00010987563835823719,
      "loss": 2.5702,
      "theoretical_loss": 3.3834502362485472,
      "tokens_seen": 2365915136
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010985908832986571,
      "loss": 2.5938,
      "theoretical_loss": 3.383442677741236,
      "tokens_seen": 2365980672
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010984253830149424,
      "loss": 2.481,
      "theoretical_loss": 3.383435119501908,
      "tokens_seen": 2366046208
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010982598827312276,
      "loss": 2.6169,
      "theoretical_loss": 3.383427561530546,
      "tokens_seen": 2366111744
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010980943824475128,
      "loss": 2.5813,
      "theoretical_loss": 3.3834200038271325,
      "tokens_seen": 2366177280
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010979288821637981,
      "loss": 2.6644,
      "theoretical_loss": 3.3834124463916515,
      "tokens_seen": 2366242816
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010977633818800833,
      "loss": 2.7425,
      "theoretical_loss": 3.383404889224085,
      "tokens_seen": 2366308352
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010975978815963686,
      "loss": 2.5024,
      "theoretical_loss": 3.3833973323244178,
      "tokens_seen": 2366373888
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010974323813126535,
      "loss": 2.7578,
      "theoretical_loss": 3.383389775692631,
      "tokens_seen": 2366439424
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010972668810289388,
      "loss": 2.4235,
      "theoretical_loss": 3.3833822193287095,
      "tokens_seen": 2366504960
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001097101380745224,
      "loss": 2.6259,
      "theoretical_loss": 3.383374663232635,
      "tokens_seen": 2366570496
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010969358804615093,
      "loss": 2.6485,
      "theoretical_loss": 3.3833671074043914,
      "tokens_seen": 2366636032
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010967703801777945,
      "loss": 2.3751,
      "theoretical_loss": 3.383359551843961,
      "tokens_seen": 2366701568
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010966048798940797,
      "loss": 2.623,
      "theoretical_loss": 3.3833519965513283,
      "tokens_seen": 2366767104
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001096439379610365,
      "loss": 2.5969,
      "theoretical_loss": 3.383344441526475,
      "tokens_seen": 2366832640
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010962738793266502,
      "loss": 2.3195,
      "theoretical_loss": 3.3833368867693845,
      "tokens_seen": 2366898176
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010961083790429355,
      "loss": 2.5851,
      "theoretical_loss": 3.38332933228004,
      "tokens_seen": 2366963712
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010959428787592207,
      "loss": 2.7159,
      "theoretical_loss": 3.3833217780584253,
      "tokens_seen": 2367029248
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001095777378475506,
      "loss": 2.8997,
      "theoretical_loss": 3.3833142241045224,
      "tokens_seen": 2367094784
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010956118781917912,
      "loss": 2.4745,
      "theoretical_loss": 3.3833066704183152,
      "tokens_seen": 2367160320
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010954463779080764,
      "loss": 2.4443,
      "theoretical_loss": 3.3832991169997864,
      "tokens_seen": 2367225856
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010952808776243617,
      "loss": 2.6445,
      "theoretical_loss": 3.383291563848919,
      "tokens_seen": 2367291392
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010951153773406469,
      "loss": 2.4528,
      "theoretical_loss": 3.3832840109656965,
      "tokens_seen": 2367356928
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001094949877056932,
      "loss": 2.6373,
      "theoretical_loss": 3.383276458350102,
      "tokens_seen": 2367422464
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010947843767732172,
      "loss": 2.7527,
      "theoretical_loss": 3.383268906002118,
      "tokens_seen": 2367488000
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010946188764895025,
      "loss": 2.6365,
      "theoretical_loss": 3.383261353921728,
      "tokens_seen": 2367553536
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010944533762057877,
      "loss": 2.6114,
      "theoretical_loss": 3.383253802108915,
      "tokens_seen": 2367619072
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001094287875922073,
      "loss": 2.7454,
      "theoretical_loss": 3.383246250563663,
      "tokens_seen": 2367684608
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010941223756383582,
      "loss": 2.3676,
      "theoretical_loss": 3.383238699285954,
      "tokens_seen": 2367750144
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010939568753546434,
      "loss": 2.4637,
      "theoretical_loss": 3.383231148275771,
      "tokens_seen": 2367815680
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010937913750709287,
      "loss": 2.5393,
      "theoretical_loss": 3.383223597533098,
      "tokens_seen": 2367881216
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010936258747872139,
      "loss": 2.7059,
      "theoretical_loss": 3.383216047057918,
      "tokens_seen": 2367946752
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010934603745034992,
      "loss": 2.6927,
      "theoretical_loss": 3.383208496850213,
      "tokens_seen": 2368012288
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010932948742197844,
      "loss": 2.4858,
      "theoretical_loss": 3.3832009469099678,
      "tokens_seen": 2368077824
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010931293739360696,
      "loss": 2.7831,
      "theoretical_loss": 3.3831933972371644,
      "tokens_seen": 2368143360
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010929638736523549,
      "loss": 2.5989,
      "theoretical_loss": 3.383185847831786,
      "tokens_seen": 2368208896
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010927983733686401,
      "loss": 2.6373,
      "theoretical_loss": 3.3831782986938164,
      "tokens_seen": 2368274432
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010926328730849254,
      "loss": 2.7088,
      "theoretical_loss": 3.383170749823238,
      "tokens_seen": 2368339968
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010924673728012105,
      "loss": 2.7292,
      "theoretical_loss": 3.383163201220034,
      "tokens_seen": 2368405504
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010923018725174957,
      "loss": 2.5801,
      "theoretical_loss": 3.3831556528841884,
      "tokens_seen": 2368471040
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001092136372233781,
      "loss": 2.6045,
      "theoretical_loss": 3.383148104815683,
      "tokens_seen": 2368536576
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010919708719500662,
      "loss": 2.5339,
      "theoretical_loss": 3.383140557014502,
      "tokens_seen": 2368602112
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010918053716663514,
      "loss": 2.782,
      "theoretical_loss": 3.383133009480628,
      "tokens_seen": 2368667648
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010916398713826367,
      "loss": 2.317,
      "theoretical_loss": 3.383125462214044,
      "tokens_seen": 2368733184
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010914743710989219,
      "loss": 2.073,
      "theoretical_loss": 3.3831179152147337,
      "tokens_seen": 2368798720
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010913088708152071,
      "loss": 2.4993,
      "theoretical_loss": 3.3831103684826798,
      "tokens_seen": 2368864256
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010911433705314924,
      "loss": 2.2997,
      "theoretical_loss": 3.383102822017866,
      "tokens_seen": 2368929792
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010909778702477776,
      "loss": 2.4174,
      "theoretical_loss": 3.3830952758202746,
      "tokens_seen": 2368995328
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010908123699640629,
      "loss": 2.6833,
      "theoretical_loss": 3.383087729889889,
      "tokens_seen": 2369060864
    },
    {
      "epoch": 0.85,
      "objective/train/avg_token_score": 0.008084027096629143,
      "objective/train/avg_weight": 0.9905106425285339,
      "objective/train/docs_used": 1333210,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8225440979003906,
      "objective/train/std_weight": 0.04812023416161537,
      "objective/train/theoretical_loss": 3.383080184226693,
      "objective/train/tokens_used": 2389586400,
      "theoretical_loss": 3.383080184226693,
      "tokens_seen": 2369126400
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010906468696803481,
      "loss": 2.5414,
      "theoretical_loss": 3.383080184226693,
      "tokens_seen": 2369126400
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010904813693966333,
      "loss": 2.4667,
      "theoretical_loss": 3.383072638830669,
      "tokens_seen": 2369191936
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010903158691129186,
      "loss": 2.7949,
      "theoretical_loss": 3.3830650937018008,
      "tokens_seen": 2369257472
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010901503688292038,
      "loss": 2.7453,
      "theoretical_loss": 3.3830575488400707,
      "tokens_seen": 2369323008
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010899848685454888,
      "loss": 2.6535,
      "theoretical_loss": 3.3830500042454625,
      "tokens_seen": 2369388544
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001089819368261774,
      "loss": 2.5168,
      "theoretical_loss": 3.3830424599179594,
      "tokens_seen": 2369454080
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010896538679780593,
      "loss": 2.4687,
      "theoretical_loss": 3.383034915857544,
      "tokens_seen": 2369519616
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010894883676943445,
      "loss": 2.4038,
      "theoretical_loss": 3.3830273720642,
      "tokens_seen": 2369585152
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010893228674106298,
      "loss": 2.7038,
      "theoretical_loss": 3.3830198285379103,
      "tokens_seen": 2369650688
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001089157367126915,
      "loss": 2.6318,
      "theoretical_loss": 3.3830122852786584,
      "tokens_seen": 2369716224
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010889918668432002,
      "loss": 2.5879,
      "theoretical_loss": 3.3830047422864267,
      "tokens_seen": 2369781760
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010888263665594855,
      "loss": 2.7113,
      "theoretical_loss": 3.382997199561199,
      "tokens_seen": 2369847296
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010886608662757707,
      "loss": 2.6591,
      "theoretical_loss": 3.382989657102959,
      "tokens_seen": 2369912832
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001088495365992056,
      "loss": 2.6047,
      "theoretical_loss": 3.3829821149116883,
      "tokens_seen": 2369978368
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010883298657083412,
      "loss": 2.6803,
      "theoretical_loss": 3.3829745729873713,
      "tokens_seen": 2370043904
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010881643654246264,
      "loss": 2.5495,
      "theoretical_loss": 3.382967031329991,
      "tokens_seen": 2370109440
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010879988651409117,
      "loss": 2.4342,
      "theoretical_loss": 3.38295948993953,
      "tokens_seen": 2370174976
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010878333648571969,
      "loss": 2.6644,
      "theoretical_loss": 3.382951948815972,
      "tokens_seen": 2370240512
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010876678645734821,
      "loss": 2.6961,
      "theoretical_loss": 3.3829444079593003,
      "tokens_seen": 2370306048
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010875023642897673,
      "loss": 2.595,
      "theoretical_loss": 3.3829368673694975,
      "tokens_seen": 2370371584
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010873368640060525,
      "loss": 2.1186,
      "theoretical_loss": 3.382929327046547,
      "tokens_seen": 2370437120
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010871713637223377,
      "loss": 2.6142,
      "theoretical_loss": 3.3829217869904324,
      "tokens_seen": 2370502656
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001087005863438623,
      "loss": 2.4936,
      "theoretical_loss": 3.3829142472011364,
      "tokens_seen": 2370568192
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010868403631549082,
      "loss": 2.5642,
      "theoretical_loss": 3.382906707678642,
      "tokens_seen": 2370633728
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010866748628711935,
      "loss": 2.55,
      "theoretical_loss": 3.382899168422933,
      "tokens_seen": 2370699264
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010865093625874787,
      "loss": 2.9565,
      "theoretical_loss": 3.382891629433993,
      "tokens_seen": 2370764800
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001086343862303764,
      "loss": 2.8772,
      "theoretical_loss": 3.382884090711803,
      "tokens_seen": 2370830336
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010861783620200492,
      "loss": 2.67,
      "theoretical_loss": 3.3828765522563486,
      "tokens_seen": 2370895872
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010860128617363344,
      "loss": 2.7547,
      "theoretical_loss": 3.3828690140676123,
      "tokens_seen": 2370961408
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010858473614526197,
      "loss": 2.6418,
      "theoretical_loss": 3.382861476145577,
      "tokens_seen": 2371026944
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010856818611689049,
      "loss": 2.816,
      "theoretical_loss": 3.3828539384902254,
      "tokens_seen": 2371092480
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010855163608851901,
      "loss": 2.3991,
      "theoretical_loss": 3.3828464011015416,
      "tokens_seen": 2371158016
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010853508606014754,
      "loss": 2.5679,
      "theoretical_loss": 3.3828388639795084,
      "tokens_seen": 2371223552
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010851853603177606,
      "loss": 2.5829,
      "theoretical_loss": 3.382831327124109,
      "tokens_seen": 2371289088
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010850198600340457,
      "loss": 2.7803,
      "theoretical_loss": 3.3828237905353267,
      "tokens_seen": 2371354624
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001084854359750331,
      "loss": 2.4908,
      "theoretical_loss": 3.382816254213145,
      "tokens_seen": 2371420160
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010846888594666162,
      "loss": 2.143,
      "theoretical_loss": 3.3828087181575457,
      "tokens_seen": 2371485696
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010845233591829014,
      "loss": 2.6943,
      "theoretical_loss": 3.382801182368514,
      "tokens_seen": 2371551232
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010843578588991867,
      "loss": 2.0199,
      "theoretical_loss": 3.382793646846032,
      "tokens_seen": 2371616768
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010841923586154719,
      "loss": 2.5975,
      "theoretical_loss": 3.3827861115900824,
      "tokens_seen": 2371682304
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010840268583317572,
      "loss": 2.6606,
      "theoretical_loss": 3.3827785766006495,
      "tokens_seen": 2371747840
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010838613580480424,
      "loss": 2.659,
      "theoretical_loss": 3.3827710418777164,
      "tokens_seen": 2371813376
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010836958577643276,
      "loss": 2.6325,
      "theoretical_loss": 3.3827635074212656,
      "tokens_seen": 2371878912
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010835303574806129,
      "loss": 2.5419,
      "theoretical_loss": 3.382755973231281,
      "tokens_seen": 2371944448
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010833648571968981,
      "loss": 2.5733,
      "theoretical_loss": 3.382748439307745,
      "tokens_seen": 2372009984
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010831993569131834,
      "loss": 2.764,
      "theoretical_loss": 3.3827409056506417,
      "tokens_seen": 2372075520
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010830338566294686,
      "loss": 2.5644,
      "theoretical_loss": 3.382733372259954,
      "tokens_seen": 2372141056
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010828683563457538,
      "loss": 2.6383,
      "theoretical_loss": 3.3827258391356647,
      "tokens_seen": 2372206592
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010827028560620391,
      "loss": 2.5512,
      "theoretical_loss": 3.3827183062777575,
      "tokens_seen": 2372272128
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001082537355778324,
      "loss": 2.6025,
      "theoretical_loss": 3.382710773686216,
      "tokens_seen": 2372337664
    },
    {
      "epoch": 0.85,
      "objective/train/avg_token_score": 0.023113328963518143,
      "objective/train/avg_weight": 0.9727855920791626,
      "objective/train/docs_used": 1334412,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.065147876739502,
      "objective/train/std_weight": 0.1082872748374939,
      "objective/train/theoretical_loss": 3.3827032413610225,
      "objective/train/tokens_used": 2392863200,
      "theoretical_loss": 3.3827032413610225,
      "tokens_seen": 2372403200
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010823718554946093,
      "loss": 2.5334,
      "theoretical_loss": 3.3827032413610225,
      "tokens_seen": 2372403200
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010822063552108945,
      "loss": 2.741,
      "theoretical_loss": 3.3826957093021606,
      "tokens_seen": 2372468736
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010820408549271798,
      "loss": 2.0093,
      "theoretical_loss": 3.3826881775096136,
      "tokens_seen": 2372534272
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001081875354643465,
      "loss": 2.4979,
      "theoretical_loss": 3.3826806459833647,
      "tokens_seen": 2372599808
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010817098543597502,
      "loss": 2.5204,
      "theoretical_loss": 3.3826731147233975,
      "tokens_seen": 2372665344
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010815443540760355,
      "loss": 2.6492,
      "theoretical_loss": 3.3826655837296946,
      "tokens_seen": 2372730880
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010813788537923207,
      "loss": 2.5514,
      "theoretical_loss": 3.382658053002239,
      "tokens_seen": 2372796416
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001081213353508606,
      "loss": 2.5908,
      "theoretical_loss": 3.3826505225410153,
      "tokens_seen": 2372861952
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010810478532248912,
      "loss": 2.5368,
      "theoretical_loss": 3.3826429923460055,
      "tokens_seen": 2372927488
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010808823529411764,
      "loss": 2.4763,
      "theoretical_loss": 3.382635462417193,
      "tokens_seen": 2372993024
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010807168526574617,
      "loss": 1.9684,
      "theoretical_loss": 3.382627932754562,
      "tokens_seen": 2373058560
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010805513523737469,
      "loss": 2.6667,
      "theoretical_loss": 3.3826204033580938,
      "tokens_seen": 2373124096
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010803858520900322,
      "loss": 2.7641,
      "theoretical_loss": 3.3826128742277737,
      "tokens_seen": 2373189632
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010802203518063173,
      "loss": 2.6446,
      "theoretical_loss": 3.382605345363584,
      "tokens_seen": 2373255168
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010800548515226025,
      "loss": 2.4697,
      "theoretical_loss": 3.3825978167655077,
      "tokens_seen": 2373320704
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010798893512388877,
      "loss": 2.6759,
      "theoretical_loss": 3.382590288433528,
      "tokens_seen": 2373386240
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001079723850955173,
      "loss": 2.303,
      "theoretical_loss": 3.3825827603676295,
      "tokens_seen": 2373451776
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010795583506714582,
      "loss": 2.7958,
      "theoretical_loss": 3.3825752325677936,
      "tokens_seen": 2373517312
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010793928503877435,
      "loss": 2.4924,
      "theoretical_loss": 3.382567705034005,
      "tokens_seen": 2373582848
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010792273501040287,
      "loss": 2.6421,
      "theoretical_loss": 3.382560177766246,
      "tokens_seen": 2373648384
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001079061849820314,
      "loss": 2.3325,
      "theoretical_loss": 3.3825526507645005,
      "tokens_seen": 2373713920
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010788963495365992,
      "loss": 2.6625,
      "theoretical_loss": 3.382545124028751,
      "tokens_seen": 2373779456
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010787308492528844,
      "loss": 2.5572,
      "theoretical_loss": 3.3825375975589815,
      "tokens_seen": 2373844992
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010785653489691697,
      "loss": 2.3527,
      "theoretical_loss": 3.382530071355175,
      "tokens_seen": 2373910528
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010783998486854549,
      "loss": 2.6781,
      "theoretical_loss": 3.3825225454173147,
      "tokens_seen": 2373976064
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010782343484017401,
      "loss": 2.5803,
      "theoretical_loss": 3.382515019745384,
      "tokens_seen": 2374041600
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010780688481180254,
      "loss": 2.3532,
      "theoretical_loss": 3.3825074943393663,
      "tokens_seen": 2374107136
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010779033478343106,
      "loss": 2.5621,
      "theoretical_loss": 3.382499969199244,
      "tokens_seen": 2374172672
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010777378475505957,
      "loss": 2.5826,
      "theoretical_loss": 3.3824924443250013,
      "tokens_seen": 2374238208
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001077572347266881,
      "loss": 2.7899,
      "theoretical_loss": 3.3824849197166214,
      "tokens_seen": 2374303744
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010774068469831662,
      "loss": 2.5524,
      "theoretical_loss": 3.382477395374087,
      "tokens_seen": 2374369280
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010772413466994515,
      "loss": 2.4711,
      "theoretical_loss": 3.3824698712973817,
      "tokens_seen": 2374434816
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010770758464157367,
      "loss": 2.5653,
      "theoretical_loss": 3.3824623474864888,
      "tokens_seen": 2374500352
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010769103461320219,
      "loss": 2.7969,
      "theoretical_loss": 3.3824548239413916,
      "tokens_seen": 2374565888
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010767448458483072,
      "loss": 2.8945,
      "theoretical_loss": 3.3824473006620734,
      "tokens_seen": 2374631424
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010765793455645924,
      "loss": 2.5686,
      "theoretical_loss": 3.3824397776485173,
      "tokens_seen": 2374696960
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010764138452808777,
      "loss": 2.488,
      "theoretical_loss": 3.382432254900707,
      "tokens_seen": 2374762496
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010762483449971629,
      "loss": 2.6677,
      "theoretical_loss": 3.382424732418625,
      "tokens_seen": 2374828032
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010760828447134481,
      "loss": 2.4783,
      "theoretical_loss": 3.382417210202255,
      "tokens_seen": 2374893568
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010759173444297334,
      "loss": 2.5599,
      "theoretical_loss": 3.3824096882515806,
      "tokens_seen": 2374959104
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010757518441460186,
      "loss": 2.6077,
      "theoretical_loss": 3.382402166566585,
      "tokens_seen": 2375024640
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010755863438623038,
      "loss": 2.144,
      "theoretical_loss": 3.382394645147251,
      "tokens_seen": 2375090176
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010754208435785891,
      "loss": 2.8009,
      "theoretical_loss": 3.382387123993562,
      "tokens_seen": 2375155712
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001075255343294874,
      "loss": 2.6334,
      "theoretical_loss": 3.382379603105502,
      "tokens_seen": 2375221248
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010750898430111593,
      "loss": 2.4912,
      "theoretical_loss": 3.382372082483053,
      "tokens_seen": 2375286784
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010749243427274445,
      "loss": 2.6979,
      "theoretical_loss": 3.3823645621261997,
      "tokens_seen": 2375352320
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010747588424437298,
      "loss": 2.4191,
      "theoretical_loss": 3.3823570420349243,
      "tokens_seen": 2375417856
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001074593342160015,
      "loss": 2.605,
      "theoretical_loss": 3.3823495222092106,
      "tokens_seen": 2375483392
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010744278418763003,
      "loss": 2.4309,
      "theoretical_loss": 3.3823420026490423,
      "tokens_seen": 2375548928
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010742623415925855,
      "loss": 2.5665,
      "theoretical_loss": 3.3823344833544016,
      "tokens_seen": 2375614464
    },
    {
      "epoch": 0.85,
      "objective/train/avg_token_score": 0.020016659051179886,
      "objective/train/avg_weight": 0.9764650464057922,
      "objective/train/docs_used": 1336383,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.358576774597168,
      "objective/train/std_weight": 0.09200353175401688,
      "objective/train/theoretical_loss": 3.382326964325273,
      "objective/train/tokens_used": 2396140000,
      "theoretical_loss": 3.382326964325273,
      "tokens_seen": 2375680000
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010740968413088707,
      "loss": 2.4468,
      "theoretical_loss": 3.382326964325273,
      "tokens_seen": 2375680000
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001073931341025156,
      "loss": 2.6536,
      "theoretical_loss": 3.3823194455616385,
      "tokens_seen": 2375745536
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010737658407414412,
      "loss": 2.3946,
      "theoretical_loss": 3.3823119270634825,
      "tokens_seen": 2375811072
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010736003404577266,
      "loss": 2.5655,
      "theoretical_loss": 3.3823044088307883,
      "tokens_seen": 2375876608
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010734348401740118,
      "loss": 2.5014,
      "theoretical_loss": 3.3822968908635382,
      "tokens_seen": 2375942144
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010732693398902971,
      "loss": 2.5056,
      "theoretical_loss": 3.3822893731617167,
      "tokens_seen": 2376007680
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010731038396065823,
      "loss": 2.4536,
      "theoretical_loss": 3.382281855725306,
      "tokens_seen": 2376073216
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010729383393228676,
      "loss": 2.6294,
      "theoretical_loss": 3.38227433855429,
      "tokens_seen": 2376138752
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010727728390391525,
      "loss": 2.5178,
      "theoretical_loss": 3.3822668216486527,
      "tokens_seen": 2376204288
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010726073387554378,
      "loss": 1.9793,
      "theoretical_loss": 3.3822593050083762,
      "tokens_seen": 2376269824
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001072441838471723,
      "loss": 2.5122,
      "theoretical_loss": 3.382251788633444,
      "tokens_seen": 2376335360
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010722763381880082,
      "loss": 2.3459,
      "theoretical_loss": 3.38224427252384,
      "tokens_seen": 2376400896
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010721108379042935,
      "loss": 2.7979,
      "theoretical_loss": 3.3822367566795473,
      "tokens_seen": 2376466432
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010719453376205787,
      "loss": 2.6613,
      "theoretical_loss": 3.382229241100549,
      "tokens_seen": 2376531968
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001071779837336864,
      "loss": 2.6306,
      "theoretical_loss": 3.382221725786829,
      "tokens_seen": 2376597504
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010716143370531492,
      "loss": 2.4801,
      "theoretical_loss": 3.3822142107383697,
      "tokens_seen": 2376663040
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010714488367694344,
      "loss": 2.6065,
      "theoretical_loss": 3.382206695955155,
      "tokens_seen": 2376728576
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010712833364857197,
      "loss": 2.6498,
      "theoretical_loss": 3.3821991814371684,
      "tokens_seen": 2376794112
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010711178362020049,
      "loss": 2.4768,
      "theoretical_loss": 3.3821916671843923,
      "tokens_seen": 2376859648
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010709523359182902,
      "loss": 2.4801,
      "theoretical_loss": 3.3821841531968113,
      "tokens_seen": 2376925184
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010707868356345754,
      "loss": 2.6852,
      "theoretical_loss": 3.382176639474408,
      "tokens_seen": 2376990720
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010706213353508606,
      "loss": 2.5997,
      "theoretical_loss": 3.382169126017166,
      "tokens_seen": 2377056256
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010704558350671459,
      "loss": 2.5905,
      "theoretical_loss": 3.382161612825068,
      "tokens_seen": 2377121792
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001070290334783431,
      "loss": 2.5761,
      "theoretical_loss": 3.3821540998980986,
      "tokens_seen": 2377187328
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010701248344997162,
      "loss": 2.7132,
      "theoretical_loss": 3.3821465872362397,
      "tokens_seen": 2377252864
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010699593342160015,
      "loss": 2.7367,
      "theoretical_loss": 3.3821390748394755,
      "tokens_seen": 2377318400
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010697938339322867,
      "loss": 2.5172,
      "theoretical_loss": 3.382131562707789,
      "tokens_seen": 2377383936
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001069628333648572,
      "loss": 2.6498,
      "theoretical_loss": 3.382124050841164,
      "tokens_seen": 2377449472
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010694628333648572,
      "loss": 2.4979,
      "theoretical_loss": 3.382116539239583,
      "tokens_seen": 2377515008
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010692973330811424,
      "loss": 2.706,
      "theoretical_loss": 3.38210902790303,
      "tokens_seen": 2377580544
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010691318327974277,
      "loss": 2.4758,
      "theoretical_loss": 3.3821015168314883,
      "tokens_seen": 2377646080
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010689663325137129,
      "loss": 2.3955,
      "theoretical_loss": 3.3820940060249414,
      "tokens_seen": 2377711616
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010688008322299981,
      "loss": 2.4977,
      "theoretical_loss": 3.382086495483372,
      "tokens_seen": 2377777152
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010686353319462834,
      "loss": 2.6375,
      "theoretical_loss": 3.382078985206764,
      "tokens_seen": 2377842688
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010684698316625686,
      "loss": 2.7721,
      "theoretical_loss": 3.3820714751951004,
      "tokens_seen": 2377908224
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010683043313788539,
      "loss": 2.5136,
      "theoretical_loss": 3.3820639654483653,
      "tokens_seen": 2377973760
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010681388310951391,
      "loss": 2.5078,
      "theoretical_loss": 3.382056455966541,
      "tokens_seen": 2378039296
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010679733308114243,
      "loss": 2.6234,
      "theoretical_loss": 3.3820489467496113,
      "tokens_seen": 2378104832
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010678078305277093,
      "loss": 2.5286,
      "theoretical_loss": 3.38204143779756,
      "tokens_seen": 2378170368
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010676423302439946,
      "loss": 2.8243,
      "theoretical_loss": 3.38203392911037,
      "tokens_seen": 2378235904
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010674768299602799,
      "loss": 2.4569,
      "theoretical_loss": 3.382026420688024,
      "tokens_seen": 2378301440
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010673113296765652,
      "loss": 2.4484,
      "theoretical_loss": 3.3820189125305067,
      "tokens_seen": 2378366976
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010671458293928504,
      "loss": 2.5182,
      "theoretical_loss": 3.382011404637801,
      "tokens_seen": 2378432512
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010669803291091356,
      "loss": 2.5108,
      "theoretical_loss": 3.38200389700989,
      "tokens_seen": 2378498048
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010668148288254209,
      "loss": 2.5864,
      "theoretical_loss": 3.381996389646757,
      "tokens_seen": 2378563584
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010666493285417061,
      "loss": 2.5491,
      "theoretical_loss": 3.3819888825483853,
      "tokens_seen": 2378629120
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010664838282579914,
      "loss": 2.6686,
      "theoretical_loss": 3.381981375714759,
      "tokens_seen": 2378694656
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010663183279742766,
      "loss": 2.4037,
      "theoretical_loss": 3.3819738691458605,
      "tokens_seen": 2378760192
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010661528276905618,
      "loss": 2.633,
      "theoretical_loss": 3.3819663628416743,
      "tokens_seen": 2378825728
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010659873274068471,
      "loss": 2.4618,
      "theoretical_loss": 3.3819588568021826,
      "tokens_seen": 2378891264
    },
    {
      "epoch": 0.85,
      "objective/train/avg_token_score": 0.025618497282266617,
      "objective/train/avg_weight": 0.9698943495750427,
      "objective/train/docs_used": 1338132,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.719021797180176,
      "objective/train/std_weight": 0.11872956156730652,
      "objective/train/theoretical_loss": 3.3819513510273698,
      "objective/train/tokens_used": 2399416800,
      "theoretical_loss": 3.3819513510273698,
      "tokens_seen": 2378956800
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010658218271231323,
      "loss": 2.4768,
      "theoretical_loss": 3.3819513510273698,
      "tokens_seen": 2378956800
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010656563268394176,
      "loss": 2.706,
      "theoretical_loss": 3.381943845517218,
      "tokens_seen": 2379022336
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010654908265557028,
      "loss": 2.7546,
      "theoretical_loss": 3.3819363402717117,
      "tokens_seen": 2379087872
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010653253262719878,
      "loss": 2.078,
      "theoretical_loss": 3.381928835290834,
      "tokens_seen": 2379153408
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001065159825988273,
      "loss": 2.7693,
      "theoretical_loss": 3.3819213305745683,
      "tokens_seen": 2379218944
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010649943257045583,
      "loss": 2.6559,
      "theoretical_loss": 3.381913826122898,
      "tokens_seen": 2379284480
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010648288254208435,
      "loss": 2.3398,
      "theoretical_loss": 3.3819063219358063,
      "tokens_seen": 2379350016
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010646633251371287,
      "loss": 2.665,
      "theoretical_loss": 3.3818988180132763,
      "tokens_seen": 2379415552
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001064497824853414,
      "loss": 2.7718,
      "theoretical_loss": 3.381891314355292,
      "tokens_seen": 2379481088
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010643323245696992,
      "loss": 2.5679,
      "theoretical_loss": 3.3818838109618365,
      "tokens_seen": 2379546624
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010641668242859845,
      "loss": 2.4358,
      "theoretical_loss": 3.3818763078328935,
      "tokens_seen": 2379612160
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010640013240022697,
      "loss": 2.4239,
      "theoretical_loss": 3.3818688049684456,
      "tokens_seen": 2379677696
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001063835823718555,
      "loss": 2.9343,
      "theoretical_loss": 3.381861302368477,
      "tokens_seen": 2379743232
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010636703234348402,
      "loss": 2.4368,
      "theoretical_loss": 3.381853800032971,
      "tokens_seen": 2379808768
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010635048231511254,
      "loss": 2.4105,
      "theoretical_loss": 3.3818462979619106,
      "tokens_seen": 2379874304
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010633393228674107,
      "loss": 2.6876,
      "theoretical_loss": 3.381838796155279,
      "tokens_seen": 2379939840
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010631738225836959,
      "loss": 2.5039,
      "theoretical_loss": 3.3818312946130606,
      "tokens_seen": 2380005376
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001063008322299981,
      "loss": 2.7211,
      "theoretical_loss": 3.381823793335238,
      "tokens_seen": 2380070912
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010628428220162662,
      "loss": 2.6867,
      "theoretical_loss": 3.381816292321795,
      "tokens_seen": 2380136448
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010626773217325515,
      "loss": 2.3763,
      "theoretical_loss": 3.3818087915727144,
      "tokens_seen": 2380201984
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010625118214488367,
      "loss": 2.5672,
      "theoretical_loss": 3.3818012910879798,
      "tokens_seen": 2380267520
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001062346321165122,
      "loss": 2.7521,
      "theoretical_loss": 3.3817937908675755,
      "tokens_seen": 2380333056
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010621808208814072,
      "loss": 2.6214,
      "theoretical_loss": 3.3817862909114837,
      "tokens_seen": 2380398592
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010620153205976924,
      "loss": 2.5158,
      "theoretical_loss": 3.3817787912196886,
      "tokens_seen": 2380464128
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010618498203139777,
      "loss": 2.4407,
      "theoretical_loss": 3.381771291792173,
      "tokens_seen": 2380529664
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010616843200302629,
      "loss": 2.4688,
      "theoretical_loss": 3.381763792628921,
      "tokens_seen": 2380595200
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010615188197465482,
      "loss": 2.5171,
      "theoretical_loss": 3.3817562937299153,
      "tokens_seen": 2380660736
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010613533194628334,
      "loss": 2.4224,
      "theoretical_loss": 3.38174879509514,
      "tokens_seen": 2380726272
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010611878191791186,
      "loss": 2.6183,
      "theoretical_loss": 3.381741296724578,
      "tokens_seen": 2380791808
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010610223188954039,
      "loss": 2.7061,
      "theoretical_loss": 3.381733798618213,
      "tokens_seen": 2380857344
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010608568186116891,
      "loss": 2.7091,
      "theoretical_loss": 3.3817263007760285,
      "tokens_seen": 2380922880
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010606913183279744,
      "loss": 2.1201,
      "theoretical_loss": 3.3817188031980074,
      "tokens_seen": 2380988416
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010605258180442595,
      "loss": 2.5236,
      "theoretical_loss": 3.3817113058841337,
      "tokens_seen": 2381053952
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010603603177605447,
      "loss": 2.6149,
      "theoretical_loss": 3.3817038088343905,
      "tokens_seen": 2381119488
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.000106019481747683,
      "loss": 2.473,
      "theoretical_loss": 3.3816963120487613,
      "tokens_seen": 2381185024
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010600293171931152,
      "loss": 2.7585,
      "theoretical_loss": 3.3816888155272298,
      "tokens_seen": 2381250560
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010598638169094004,
      "loss": 2.766,
      "theoretical_loss": 3.381681319269779,
      "tokens_seen": 2381316096
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010596983166256857,
      "loss": 2.4945,
      "theoretical_loss": 3.381673823276392,
      "tokens_seen": 2381381632
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010595328163419709,
      "loss": 2.45,
      "theoretical_loss": 3.3816663275470535,
      "tokens_seen": 2381447168
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010593673160582561,
      "loss": 2.7232,
      "theoretical_loss": 3.3816588320817456,
      "tokens_seen": 2381512704
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010592018157745414,
      "loss": 2.0731,
      "theoretical_loss": 3.381651336880453,
      "tokens_seen": 2381578240
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010590363154908266,
      "loss": 2.4334,
      "theoretical_loss": 3.381643841943158,
      "tokens_seen": 2381643776
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010588708152071119,
      "loss": 2.0477,
      "theoretical_loss": 3.3816363472698443,
      "tokens_seen": 2381709312
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010587053149233971,
      "loss": 2.4791,
      "theoretical_loss": 3.3816288528604956,
      "tokens_seen": 2381774848
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010585398146396823,
      "loss": 2.87,
      "theoretical_loss": 3.3816213587150954,
      "tokens_seen": 2381840384
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010583743143559676,
      "loss": 2.6384,
      "theoretical_loss": 3.3816138648336267,
      "tokens_seen": 2381905920
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010582088140722528,
      "loss": 2.4883,
      "theoretical_loss": 3.3816063712160735,
      "tokens_seen": 2381971456
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010580433137885378,
      "loss": 2.4976,
      "theoretical_loss": 3.381598877862419,
      "tokens_seen": 2382036992
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001057877813504823,
      "loss": 2.5547,
      "theoretical_loss": 3.381591384772647,
      "tokens_seen": 2382102528
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010577123132211083,
      "loss": 2.6048,
      "theoretical_loss": 3.38158389194674,
      "tokens_seen": 2382168064
    },
    {
      "epoch": 0.85,
      "objective/train/avg_token_score": 0.01599414274096489,
      "objective/train/avg_weight": 0.981173574924469,
      "objective/train/docs_used": 1340136,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.921898126602173,
      "objective/train/std_weight": 0.09938832372426987,
      "objective/train/theoretical_loss": 3.381576399384682,
      "objective/train/tokens_used": 2402693600,
      "theoretical_loss": 3.381576399384682,
      "tokens_seen": 2382233600
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010575468129373935,
      "loss": 2.6117,
      "theoretical_loss": 3.381576399384682,
      "tokens_seen": 2382233600
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010573813126536787,
      "loss": 2.7259,
      "theoretical_loss": 3.3815689070864567,
      "tokens_seen": 2382299136
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001057215812369964,
      "loss": 2.7556,
      "theoretical_loss": 3.3815614150520474,
      "tokens_seen": 2382364672
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010570503120862492,
      "loss": 3.0815,
      "theoretical_loss": 3.3815539232814373,
      "tokens_seen": 2382430208
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010568848118025345,
      "loss": 2.7404,
      "theoretical_loss": 3.3815464317746105,
      "tokens_seen": 2382495744
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010567193115188197,
      "loss": 2.6791,
      "theoretical_loss": 3.3815389405315495,
      "tokens_seen": 2382561280
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001056553811235105,
      "loss": 2.447,
      "theoretical_loss": 3.3815314495522384,
      "tokens_seen": 2382626816
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010563883109513902,
      "loss": 2.7539,
      "theoretical_loss": 3.3815239588366603,
      "tokens_seen": 2382692352
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010562228106676754,
      "loss": 2.4978,
      "theoretical_loss": 3.381516468384799,
      "tokens_seen": 2382757888
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010560573103839607,
      "loss": 2.5568,
      "theoretical_loss": 3.381508978196638,
      "tokens_seen": 2382823424
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010558918101002459,
      "loss": 2.5031,
      "theoretical_loss": 3.3815014882721606,
      "tokens_seen": 2382888960
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010557263098165311,
      "loss": 2.1837,
      "theoretical_loss": 3.38149399861135,
      "tokens_seen": 2382954496
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010555608095328163,
      "loss": 2.6612,
      "theoretical_loss": 3.38148650921419,
      "tokens_seen": 2383020032
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010553953092491015,
      "loss": 2.8158,
      "theoretical_loss": 3.3814790200806644,
      "tokens_seen": 2383085568
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010552298089653867,
      "loss": 2.4521,
      "theoretical_loss": 3.381471531210756,
      "tokens_seen": 2383151104
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001055064308681672,
      "loss": 2.7569,
      "theoretical_loss": 3.3814640426044487,
      "tokens_seen": 2383216640
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010548988083979572,
      "loss": 2.3185,
      "theoretical_loss": 3.3814565542617254,
      "tokens_seen": 2383282176
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010547333081142425,
      "loss": 2.6614,
      "theoretical_loss": 3.3814490661825705,
      "tokens_seen": 2383347712
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010545678078305277,
      "loss": 2.4549,
      "theoretical_loss": 3.3814415783669665,
      "tokens_seen": 2383413248
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001054402307546813,
      "loss": 2.4649,
      "theoretical_loss": 3.3814340908148974,
      "tokens_seen": 2383478784
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010542368072630982,
      "loss": 2.7146,
      "theoretical_loss": 3.381426603526347,
      "tokens_seen": 2383544320
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010540713069793834,
      "loss": 2.3546,
      "theoretical_loss": 3.381419116501298,
      "tokens_seen": 2383609856
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010539058066956687,
      "loss": 2.749,
      "theoretical_loss": 3.381411629739735,
      "tokens_seen": 2383675392
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010537403064119539,
      "loss": 2.5573,
      "theoretical_loss": 3.3814041432416397,
      "tokens_seen": 2383740928
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010535748061282391,
      "loss": 2.7734,
      "theoretical_loss": 3.3813966570069973,
      "tokens_seen": 2383806464
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010534093058445244,
      "loss": 2.8825,
      "theoretical_loss": 3.381389171035791,
      "tokens_seen": 2383872000
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010532438055608096,
      "loss": 2.6645,
      "theoretical_loss": 3.3813816853280034,
      "tokens_seen": 2383937536
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010530783052770947,
      "loss": 2.5811,
      "theoretical_loss": 3.381374199883618,
      "tokens_seen": 2384003072
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.000105291280499338,
      "loss": 2.5499,
      "theoretical_loss": 3.3813667147026196,
      "tokens_seen": 2384068608
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010527473047096652,
      "loss": 2.5318,
      "theoretical_loss": 3.381359229784991,
      "tokens_seen": 2384134144
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010525818044259504,
      "loss": 2.4851,
      "theoretical_loss": 3.3813517451307153,
      "tokens_seen": 2384199680
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010524163041422357,
      "loss": 2.5434,
      "theoretical_loss": 3.3813442607397763,
      "tokens_seen": 2384265216
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010522508038585209,
      "loss": 2.5284,
      "theoretical_loss": 3.381336776612158,
      "tokens_seen": 2384330752
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010520853035748062,
      "loss": 2.6026,
      "theoretical_loss": 3.3813292927478424,
      "tokens_seen": 2384396288
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010519198032910914,
      "loss": 2.7151,
      "theoretical_loss": 3.3813218091468147,
      "tokens_seen": 2384461824
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010517543030073766,
      "loss": 2.7948,
      "theoretical_loss": 3.3813143258090577,
      "tokens_seen": 2384527360
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010515888027236619,
      "loss": 2.6041,
      "theoretical_loss": 3.3813068427345545,
      "tokens_seen": 2384592896
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010514233024399471,
      "loss": 2.6179,
      "theoretical_loss": 3.3812993599232892,
      "tokens_seen": 2384658432
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010512578021562324,
      "loss": 2.552,
      "theoretical_loss": 3.381291877375245,
      "tokens_seen": 2384723968
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010510923018725176,
      "loss": 2.5002,
      "theoretical_loss": 3.3812843950904057,
      "tokens_seen": 2384789504
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010509268015888028,
      "loss": 2.497,
      "theoretical_loss": 3.3812769130687546,
      "tokens_seen": 2384855040
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010507613013050881,
      "loss": 2.648,
      "theoretical_loss": 3.381269431310275,
      "tokens_seen": 2384920576
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001050595801021373,
      "loss": 2.3741,
      "theoretical_loss": 3.3812619498149505,
      "tokens_seen": 2384986112
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010504303007376583,
      "loss": 2.6396,
      "theoretical_loss": 3.3812544685827652,
      "tokens_seen": 2385051648
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010502648004539435,
      "loss": 2.5042,
      "theoretical_loss": 3.381246987613702,
      "tokens_seen": 2385117184
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010500993001702288,
      "loss": 2.536,
      "theoretical_loss": 3.3812395069077446,
      "tokens_seen": 2385182720
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001049933799886514,
      "loss": 2.5276,
      "theoretical_loss": 3.3812320264648767,
      "tokens_seen": 2385248256
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010497682996027992,
      "loss": 2.5256,
      "theoretical_loss": 3.381224546285081,
      "tokens_seen": 2385313792
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010496027993190845,
      "loss": 2.4932,
      "theoretical_loss": 3.3812170663683423,
      "tokens_seen": 2385379328
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010494372990353697,
      "loss": 2.5673,
      "theoretical_loss": 3.3812095867146432,
      "tokens_seen": 2385444864
    },
    {
      "epoch": 0.85,
      "objective/train/avg_token_score": 0.018505292013287544,
      "objective/train/avg_weight": 0.9782348275184631,
      "objective/train/docs_used": 1341857,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.719496250152588,
      "objective/train/std_weight": 0.11356586217880249,
      "objective/train/theoretical_loss": 3.3812021073239675,
      "objective/train/tokens_used": 2405970400,
      "theoretical_loss": 3.3812021073239675,
      "tokens_seen": 2385510400
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001049271798751655,
      "loss": 2.3314,
      "theoretical_loss": 3.3812021073239675,
      "tokens_seen": 2385510400
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010491062984679402,
      "loss": 2.6257,
      "theoretical_loss": 3.381194628196299,
      "tokens_seen": 2385575936
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010489407981842254,
      "loss": 2.5747,
      "theoretical_loss": 3.3811871493316206,
      "tokens_seen": 2385641472
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010487752979005107,
      "loss": 2.525,
      "theoretical_loss": 3.3811796707299164,
      "tokens_seen": 2385707008
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010486097976167959,
      "loss": 2.6757,
      "theoretical_loss": 3.3811721923911695,
      "tokens_seen": 2385772544
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010484442973330812,
      "loss": 2.621,
      "theoretical_loss": 3.381164714315364,
      "tokens_seen": 2385838080
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010482787970493663,
      "loss": 1.5902,
      "theoretical_loss": 3.3811572365024825,
      "tokens_seen": 2385903616
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010481132967656515,
      "loss": 2.7497,
      "theoretical_loss": 3.3811497589525095,
      "tokens_seen": 2385969152
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010479477964819367,
      "loss": 2.8797,
      "theoretical_loss": 3.381142281665428,
      "tokens_seen": 2386034688
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001047782296198222,
      "loss": 2.7584,
      "theoretical_loss": 3.3811348046412215,
      "tokens_seen": 2386100224
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010476167959145072,
      "loss": 2.6816,
      "theoretical_loss": 3.3811273278798737,
      "tokens_seen": 2386165760
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010474512956307925,
      "loss": 2.2567,
      "theoretical_loss": 3.3811198513813685,
      "tokens_seen": 2386231296
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010472857953470777,
      "loss": 2.6183,
      "theoretical_loss": 3.3811123751456886,
      "tokens_seen": 2386296832
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001047120295063363,
      "loss": 2.5565,
      "theoretical_loss": 3.3811048991728185,
      "tokens_seen": 2386362368
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010469547947796482,
      "loss": 2.7844,
      "theoretical_loss": 3.3810974234627413,
      "tokens_seen": 2386427904
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010467892944959334,
      "loss": 2.8123,
      "theoretical_loss": 3.38108994801544,
      "tokens_seen": 2386493440
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010466237942122187,
      "loss": 2.6009,
      "theoretical_loss": 3.3810824728308995,
      "tokens_seen": 2386558976
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010464582939285039,
      "loss": 2.321,
      "theoretical_loss": 3.381074997909102,
      "tokens_seen": 2386624512
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010462927936447891,
      "loss": 2.6239,
      "theoretical_loss": 3.3810675232500316,
      "tokens_seen": 2386690048
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010461272933610744,
      "loss": 2.7685,
      "theoretical_loss": 3.3810600488536715,
      "tokens_seen": 2386755584
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010459617930773596,
      "loss": 2.5387,
      "theoretical_loss": 3.381052574720006,
      "tokens_seen": 2386821120
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010457962927936447,
      "loss": 2.6193,
      "theoretical_loss": 3.3810451008490183,
      "tokens_seen": 2386886656
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.000104563079250993,
      "loss": 2.5753,
      "theoretical_loss": 3.3810376272406915,
      "tokens_seen": 2386952192
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010454652922262152,
      "loss": 2.5664,
      "theoretical_loss": 3.38103015389501,
      "tokens_seen": 2387017728
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010452997919425005,
      "loss": 2.526,
      "theoretical_loss": 3.381022680811957,
      "tokens_seen": 2387083264
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010451342916587857,
      "loss": 2.7341,
      "theoretical_loss": 3.3810152079915152,
      "tokens_seen": 2387148800
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010449687913750709,
      "loss": 2.5229,
      "theoretical_loss": 3.3810077354336694,
      "tokens_seen": 2387214336
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010448032910913562,
      "loss": 2.8089,
      "theoretical_loss": 3.381000263138403,
      "tokens_seen": 2387279872
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010446377908076414,
      "loss": 2.5887,
      "theoretical_loss": 3.3809927911056987,
      "tokens_seen": 2387345408
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010444722905239267,
      "loss": 2.6135,
      "theoretical_loss": 3.380985319335541,
      "tokens_seen": 2387410944
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010443067902402119,
      "loss": 2.1109,
      "theoretical_loss": 3.380977847827913,
      "tokens_seen": 2387476480
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010441412899564971,
      "loss": 2.5134,
      "theoretical_loss": 3.3809703765827983,
      "tokens_seen": 2387542016
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010439757896727824,
      "loss": 2.7356,
      "theoretical_loss": 3.3809629056001804,
      "tokens_seen": 2387607552
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010438102893890676,
      "loss": 2.8956,
      "theoretical_loss": 3.3809554348800432,
      "tokens_seen": 2387673088
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010436447891053528,
      "loss": 2.6528,
      "theoretical_loss": 3.38094796442237,
      "tokens_seen": 2387738624
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010434792888216381,
      "loss": 2.2767,
      "theoretical_loss": 3.3809404942271444,
      "tokens_seen": 2387804160
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001043313788537923,
      "loss": 2.5838,
      "theoretical_loss": 3.38093302429435,
      "tokens_seen": 2387869696
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010431482882542083,
      "loss": 2.568,
      "theoretical_loss": 3.3809255546239707,
      "tokens_seen": 2387935232
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010429827879704935,
      "loss": 2.6795,
      "theoretical_loss": 3.3809180852159892,
      "tokens_seen": 2388000768
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010428172876867788,
      "loss": 2.43,
      "theoretical_loss": 3.3809106160703903,
      "tokens_seen": 2388066304
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001042651787403064,
      "loss": 2.6736,
      "theoretical_loss": 3.3809031471871567,
      "tokens_seen": 2388131840
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010424862871193493,
      "loss": 2.7188,
      "theoretical_loss": 3.3808956785662723,
      "tokens_seen": 2388197376
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010423207868356345,
      "loss": 2.7226,
      "theoretical_loss": 3.3808882102077207,
      "tokens_seen": 2388262912
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010421552865519197,
      "loss": 2.5801,
      "theoretical_loss": 3.3808807421114855,
      "tokens_seen": 2388328448
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001041989786268205,
      "loss": 2.457,
      "theoretical_loss": 3.38087327427755,
      "tokens_seen": 2388393984
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010418242859844902,
      "loss": 2.5127,
      "theoretical_loss": 3.3808658067058976,
      "tokens_seen": 2388459520
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010416587857007755,
      "loss": 2.5496,
      "theoretical_loss": 3.380858339396513,
      "tokens_seen": 2388525056
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010414932854170607,
      "loss": 2.5202,
      "theoretical_loss": 3.3808508723493786,
      "tokens_seen": 2388590592
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001041327785133346,
      "loss": 2.6014,
      "theoretical_loss": 3.3808434055644785,
      "tokens_seen": 2388656128
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010411622848496312,
      "loss": 2.6493,
      "theoretical_loss": 3.3808359390417966,
      "tokens_seen": 2388721664
    },
    {
      "epoch": 0.85,
      "objective/train/avg_token_score": 0.03957618400454521,
      "objective/train/avg_weight": 0.953440248966217,
      "objective/train/docs_used": 1343487,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.443852424621582,
      "objective/train/std_weight": 0.16509301960468292,
      "objective/train/theoretical_loss": 3.380828472781316,
      "objective/train/tokens_used": 2409247200,
      "theoretical_loss": 3.380828472781316,
      "tokens_seen": 2388787200
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010409967845659164,
      "loss": 2.4303,
      "theoretical_loss": 3.380828472781316,
      "tokens_seen": 2388787200
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010408312842822015,
      "loss": 2.7081,
      "theoretical_loss": 3.3808210067830204,
      "tokens_seen": 2388852736
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010406657839984868,
      "loss": 2.4042,
      "theoretical_loss": 3.3808135410468934,
      "tokens_seen": 2388918272
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001040500283714772,
      "loss": 2.5939,
      "theoretical_loss": 3.380806075572919,
      "tokens_seen": 2388983808
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010403347834310572,
      "loss": 2.2062,
      "theoretical_loss": 3.3807986103610803,
      "tokens_seen": 2389049344
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010401692831473425,
      "loss": 2.6619,
      "theoretical_loss": 3.3807911454113606,
      "tokens_seen": 2389114880
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010400037828636277,
      "loss": 2.5473,
      "theoretical_loss": 3.3807836807237446,
      "tokens_seen": 2389180416
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001039838282579913,
      "loss": 2.3802,
      "theoretical_loss": 3.3807762162982153,
      "tokens_seen": 2389245952
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010396727822961982,
      "loss": 2.5856,
      "theoretical_loss": 3.380768752134756,
      "tokens_seen": 2389311488
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010395072820124834,
      "loss": 2.3405,
      "theoretical_loss": 3.3807612882333506,
      "tokens_seen": 2389377024
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010393417817287687,
      "loss": 2.5906,
      "theoretical_loss": 3.3807538245939828,
      "tokens_seen": 2389442560
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010391762814450539,
      "loss": 2.9428,
      "theoretical_loss": 3.3807463612166364,
      "tokens_seen": 2389508096
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010390107811613392,
      "loss": 2.4457,
      "theoretical_loss": 3.3807388981012942,
      "tokens_seen": 2389573632
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010388452808776244,
      "loss": 2.5868,
      "theoretical_loss": 3.3807314352479407,
      "tokens_seen": 2389639168
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010386797805939096,
      "loss": 2.6989,
      "theoretical_loss": 3.3807239726565594,
      "tokens_seen": 2389704704
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010385142803101949,
      "loss": 2.4222,
      "theoretical_loss": 3.3807165103271335,
      "tokens_seen": 2389770240
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.000103834878002648,
      "loss": 2.6937,
      "theoretical_loss": 3.380709048259647,
      "tokens_seen": 2389835776
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010381832797427652,
      "loss": 2.412,
      "theoretical_loss": 3.380701586454083,
      "tokens_seen": 2389901312
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010380177794590505,
      "loss": 2.0693,
      "theoretical_loss": 3.3806941249104256,
      "tokens_seen": 2389966848
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010378522791753357,
      "loss": 3.0241,
      "theoretical_loss": 3.3806866636286586,
      "tokens_seen": 2390032384
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001037686778891621,
      "loss": 2.4404,
      "theoretical_loss": 3.380679202608765,
      "tokens_seen": 2390097920
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010375212786079062,
      "loss": 2.7141,
      "theoretical_loss": 3.380671741850729,
      "tokens_seen": 2390163456
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010373557783241914,
      "loss": 2.6822,
      "theoretical_loss": 3.380664281354534,
      "tokens_seen": 2390228992
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010371902780404767,
      "loss": 2.8497,
      "theoretical_loss": 3.380656821120163,
      "tokens_seen": 2390294528
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010370247777567619,
      "loss": 2.6594,
      "theoretical_loss": 3.3806493611476007,
      "tokens_seen": 2390360064
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010368592774730471,
      "loss": 2.8681,
      "theoretical_loss": 3.3806419014368303,
      "tokens_seen": 2390425600
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010366937771893324,
      "loss": 2.6003,
      "theoretical_loss": 3.3806344419878354,
      "tokens_seen": 2390491136
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010365282769056176,
      "loss": 2.6401,
      "theoretical_loss": 3.3806269828005995,
      "tokens_seen": 2390556672
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010363627766219029,
      "loss": 2.7076,
      "theoretical_loss": 3.3806195238751067,
      "tokens_seen": 2390622208
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010361972763381881,
      "loss": 2.4776,
      "theoretical_loss": 3.38061206521134,
      "tokens_seen": 2390687744
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010360317760544733,
      "loss": 2.7699,
      "theoretical_loss": 3.3806046068092837,
      "tokens_seen": 2390753280
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010358662757707583,
      "loss": 2.5517,
      "theoretical_loss": 3.380597148668921,
      "tokens_seen": 2390818816
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010357007754870436,
      "loss": 2.7798,
      "theoretical_loss": 3.380589690790236,
      "tokens_seen": 2390884352
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010355352752033288,
      "loss": 2.1137,
      "theoretical_loss": 3.380582233173212,
      "tokens_seen": 2390949888
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001035369774919614,
      "loss": 2.1344,
      "theoretical_loss": 3.380574775817832,
      "tokens_seen": 2391015424
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010352042746358993,
      "loss": 2.8195,
      "theoretical_loss": 3.3805673187240806,
      "tokens_seen": 2391080960
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010350387743521845,
      "loss": 2.7189,
      "theoretical_loss": 3.3805598618919412,
      "tokens_seen": 2391146496
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010348732740684698,
      "loss": 2.6563,
      "theoretical_loss": 3.3805524053213976,
      "tokens_seen": 2391212032
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001034707773784755,
      "loss": 2.4509,
      "theoretical_loss": 3.3805449490124335,
      "tokens_seen": 2391277568
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010345422735010402,
      "loss": 2.7374,
      "theoretical_loss": 3.380537492965032,
      "tokens_seen": 2391343104
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010343767732173255,
      "loss": 2.4824,
      "theoretical_loss": 3.3805300371791764,
      "tokens_seen": 2391408640
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010342112729336107,
      "loss": 2.5899,
      "theoretical_loss": 3.3805225816548523,
      "tokens_seen": 2391474176
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001034045772649896,
      "loss": 2.5979,
      "theoretical_loss": 3.380515126392041,
      "tokens_seen": 2391539712
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010338802723661812,
      "loss": 2.5658,
      "theoretical_loss": 3.380507671390728,
      "tokens_seen": 2391605248
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010337147720824666,
      "loss": 2.6262,
      "theoretical_loss": 3.380500216650896,
      "tokens_seen": 2391670784
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010335492717987518,
      "loss": 2.6221,
      "theoretical_loss": 3.380492762172529,
      "tokens_seen": 2391736320
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010333837715150368,
      "loss": 2.4791,
      "theoretical_loss": 3.3804853079556105,
      "tokens_seen": 2391801856
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001033218271231322,
      "loss": 2.6883,
      "theoretical_loss": 3.380477854000124,
      "tokens_seen": 2391867392
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010330527709476073,
      "loss": 2.5586,
      "theoretical_loss": 3.3804704003060535,
      "tokens_seen": 2391932928
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010328872706638925,
      "loss": 2.8172,
      "theoretical_loss": 3.380462946873383,
      "tokens_seen": 2391998464
    },
    {
      "epoch": 0.85,
      "objective/train/avg_token_score": 0.020660866051912308,
      "objective/train/avg_weight": 0.9756754636764526,
      "objective/train/docs_used": 1344994,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.851940155029297,
      "objective/train/std_weight": 0.10401567071676254,
      "objective/train/theoretical_loss": 3.380455493702095,
      "objective/train/tokens_used": 2412524000,
      "theoretical_loss": 3.380455493702095,
      "tokens_seen": 2392064000
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010327217703801777,
      "loss": 2.5254,
      "theoretical_loss": 3.380455493702095,
      "tokens_seen": 2392064000
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001032556270096463,
      "loss": 2.499,
      "theoretical_loss": 3.3804480407921744,
      "tokens_seen": 2392129536
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010323907698127482,
      "loss": 2.731,
      "theoretical_loss": 3.380440588143604,
      "tokens_seen": 2392195072
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010322252695290335,
      "loss": 2.4128,
      "theoretical_loss": 3.380433135756368,
      "tokens_seen": 2392260608
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010320597692453187,
      "loss": 2.7503,
      "theoretical_loss": 3.3804256836304503,
      "tokens_seen": 2392326144
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001031894268961604,
      "loss": 2.5345,
      "theoretical_loss": 3.380418231765834,
      "tokens_seen": 2392391680
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010317287686778892,
      "loss": 2.597,
      "theoretical_loss": 3.380410780162503,
      "tokens_seen": 2392457216
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010315632683941744,
      "loss": 2.5848,
      "theoretical_loss": 3.3804033288204405,
      "tokens_seen": 2392522752
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010313977681104597,
      "loss": 2.83,
      "theoretical_loss": 3.3803958777396312,
      "tokens_seen": 2392588288
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010312322678267449,
      "loss": 2.5573,
      "theoretical_loss": 3.380388426920058,
      "tokens_seen": 2392653824
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.000103106676754303,
      "loss": 2.5495,
      "theoretical_loss": 3.380380976361705,
      "tokens_seen": 2392719360
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010309012672593152,
      "loss": 2.606,
      "theoretical_loss": 3.3803735260645555,
      "tokens_seen": 2392784896
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010307357669756005,
      "loss": 2.7102,
      "theoretical_loss": 3.3803660760285936,
      "tokens_seen": 2392850432
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010305702666918857,
      "loss": 2.2742,
      "theoretical_loss": 3.3803586262538023,
      "tokens_seen": 2392915968
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001030404766408171,
      "loss": 2.5304,
      "theoretical_loss": 3.3803511767401666,
      "tokens_seen": 2392981504
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010302392661244562,
      "loss": 2.7249,
      "theoretical_loss": 3.3803437274876686,
      "tokens_seen": 2393047040
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010300737658407414,
      "loss": 1.8608,
      "theoretical_loss": 3.3803362784962934,
      "tokens_seen": 2393112576
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010299082655570267,
      "loss": 2.5374,
      "theoretical_loss": 3.380328829766024,
      "tokens_seen": 2393178112
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010297427652733119,
      "loss": 2.784,
      "theoretical_loss": 3.380321381296844,
      "tokens_seen": 2393243648
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010295772649895972,
      "loss": 2.5829,
      "theoretical_loss": 3.380313933088737,
      "tokens_seen": 2393309184
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010294117647058824,
      "loss": 2.6467,
      "theoretical_loss": 3.380306485141687,
      "tokens_seen": 2393374720
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010292462644221676,
      "loss": 2.6251,
      "theoretical_loss": 3.380299037455678,
      "tokens_seen": 2393440256
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010290807641384529,
      "loss": 2.6387,
      "theoretical_loss": 3.3802915900306933,
      "tokens_seen": 2393505792
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010289152638547381,
      "loss": 2.5762,
      "theoretical_loss": 3.3802841428667163,
      "tokens_seen": 2393571328
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010287497635710234,
      "loss": 2.5793,
      "theoretical_loss": 3.3802766959637314,
      "tokens_seen": 2393636864
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010285842632873083,
      "loss": 2.3174,
      "theoretical_loss": 3.380269249321722,
      "tokens_seen": 2393702400
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010284187630035936,
      "loss": 2.6909,
      "theoretical_loss": 3.3802618029406717,
      "tokens_seen": 2393767936
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010282532627198788,
      "loss": 2.5249,
      "theoretical_loss": 3.3802543568205645,
      "tokens_seen": 2393833472
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0001028087762436164,
      "loss": 2.777,
      "theoretical_loss": 3.3802469109613837,
      "tokens_seen": 2393899008
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00010279222621524493,
      "loss": 2.6704,
      "theoretical_loss": 3.3802394653631134,
      "tokens_seen": 2393964544
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010277567618687345,
      "loss": 2.5982,
      "theoretical_loss": 3.380232020025737,
      "tokens_seen": 2394030080
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010275912615850198,
      "loss": 2.3466,
      "theoretical_loss": 3.3802245749492386,
      "tokens_seen": 2394095616
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010274257613013051,
      "loss": 2.6248,
      "theoretical_loss": 3.3802171301336017,
      "tokens_seen": 2394161152
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010272602610175904,
      "loss": 2.6322,
      "theoretical_loss": 3.38020968557881,
      "tokens_seen": 2394226688
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010270947607338756,
      "loss": 2.5841,
      "theoretical_loss": 3.380202241284847,
      "tokens_seen": 2394292224
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010269292604501609,
      "loss": 2.429,
      "theoretical_loss": 3.3801947972516966,
      "tokens_seen": 2394357760
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010267637601664461,
      "loss": 2.9016,
      "theoretical_loss": 3.3801873534793425,
      "tokens_seen": 2394423296
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010265982598827313,
      "loss": 2.0156,
      "theoretical_loss": 3.3801799099677687,
      "tokens_seen": 2394488832
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010264327595990166,
      "loss": 2.2781,
      "theoretical_loss": 3.380172466716959,
      "tokens_seen": 2394554368
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010262672593153018,
      "loss": 2.548,
      "theoretical_loss": 3.3801650237268968,
      "tokens_seen": 2394619904
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010261017590315868,
      "loss": 2.3517,
      "theoretical_loss": 3.3801575809975652,
      "tokens_seen": 2394685440
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001025936258747872,
      "loss": 2.7072,
      "theoretical_loss": 3.380150138528949,
      "tokens_seen": 2394750976
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010257707584641573,
      "loss": 2.5749,
      "theoretical_loss": 3.380142696321032,
      "tokens_seen": 2394816512
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010256052581804425,
      "loss": 2.8845,
      "theoretical_loss": 3.3801352543737972,
      "tokens_seen": 2394882048
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010254397578967277,
      "loss": 2.3891,
      "theoretical_loss": 3.3801278126872285,
      "tokens_seen": 2394947584
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001025274257613013,
      "loss": 2.0687,
      "theoretical_loss": 3.3801203712613095,
      "tokens_seen": 2395013120
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010251087573292982,
      "loss": 2.6767,
      "theoretical_loss": 3.3801129300960246,
      "tokens_seen": 2395078656
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010249432570455835,
      "loss": 2.621,
      "theoretical_loss": 3.380105489191357,
      "tokens_seen": 2395144192
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010247777567618687,
      "loss": 2.4612,
      "theoretical_loss": 3.380098048547291,
      "tokens_seen": 2395209728
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001024612256478154,
      "loss": 2.649,
      "theoretical_loss": 3.380090608163809,
      "tokens_seen": 2395275264
    },
    {
      "epoch": 0.86,
      "objective/train/avg_token_score": 0.01223299652338028,
      "objective/train/avg_weight": 0.9856345057487488,
      "objective/train/docs_used": 1347049,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.093480110168457,
      "objective/train/std_weight": 0.08330517262220383,
      "objective/train/theoretical_loss": 3.3800831680408963,
      "objective/train/tokens_used": 2415800800,
      "theoretical_loss": 3.3800831680408963,
      "tokens_seen": 2395340800
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010244467561944392,
      "loss": 3.7986,
      "theoretical_loss": 3.3800831680408963,
      "tokens_seen": 2395340800
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010242812559107244,
      "loss": 2.6256,
      "theoretical_loss": 3.380075728178536,
      "tokens_seen": 2395406336
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010241157556270097,
      "loss": 2.6275,
      "theoretical_loss": 3.380068288576712,
      "tokens_seen": 2395471872
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010239502553432949,
      "loss": 2.6344,
      "theoretical_loss": 3.3800608492354076,
      "tokens_seen": 2395537408
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010237847550595801,
      "loss": 2.4758,
      "theoretical_loss": 3.3800534101546065,
      "tokens_seen": 2395602944
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010236192547758653,
      "loss": 2.4763,
      "theoretical_loss": 3.3800459713342934,
      "tokens_seen": 2395668480
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010234537544921505,
      "loss": 2.4418,
      "theoretical_loss": 3.380038532774451,
      "tokens_seen": 2395734016
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010232882542084357,
      "loss": 2.9005,
      "theoretical_loss": 3.380031094475064,
      "tokens_seen": 2395799552
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001023122753924721,
      "loss": 2.8081,
      "theoretical_loss": 3.3800236564361157,
      "tokens_seen": 2395865088
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010229572536410062,
      "loss": 2.6837,
      "theoretical_loss": 3.3800162186575893,
      "tokens_seen": 2395930624
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010227917533572915,
      "loss": 2.3247,
      "theoretical_loss": 3.3800087811394697,
      "tokens_seen": 2395996160
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010226262530735767,
      "loss": 2.6519,
      "theoretical_loss": 3.3800013438817396,
      "tokens_seen": 2396061696
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001022460752789862,
      "loss": 2.5064,
      "theoretical_loss": 3.3799939068843834,
      "tokens_seen": 2396127232
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010222952525061472,
      "loss": 2.6806,
      "theoretical_loss": 3.3799864701473847,
      "tokens_seen": 2396192768
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010221297522224324,
      "loss": 2.2953,
      "theoretical_loss": 3.379979033670727,
      "tokens_seen": 2396258304
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010219642519387177,
      "loss": 2.6229,
      "theoretical_loss": 3.3799715974543947,
      "tokens_seen": 2396323840
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010217987516550029,
      "loss": 2.5194,
      "theoretical_loss": 3.379964161498371,
      "tokens_seen": 2396389376
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010216332513712881,
      "loss": 2.6411,
      "theoretical_loss": 3.3799567258026397,
      "tokens_seen": 2396454912
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010214677510875734,
      "loss": 2.5578,
      "theoretical_loss": 3.3799492903671844,
      "tokens_seen": 2396520448
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010213022508038586,
      "loss": 2.491,
      "theoretical_loss": 3.3799418551919898,
      "tokens_seen": 2396585984
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010211367505201437,
      "loss": 2.5888,
      "theoretical_loss": 3.379934420277039,
      "tokens_seen": 2396651520
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001020971250236429,
      "loss": 2.6401,
      "theoretical_loss": 3.3799269856223155,
      "tokens_seen": 2396717056
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010208057499527142,
      "loss": 2.449,
      "theoretical_loss": 3.379919551227804,
      "tokens_seen": 2396782592
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010206402496689994,
      "loss": 2.3962,
      "theoretical_loss": 3.379912117093487,
      "tokens_seen": 2396848128
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010204747493852847,
      "loss": 2.3597,
      "theoretical_loss": 3.3799046832193493,
      "tokens_seen": 2396913664
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010203092491015699,
      "loss": 2.5305,
      "theoretical_loss": 3.3798972496053743,
      "tokens_seen": 2396979200
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010201437488178552,
      "loss": 2.6932,
      "theoretical_loss": 3.379889816251546,
      "tokens_seen": 2397044736
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010199782485341404,
      "loss": 2.4953,
      "theoretical_loss": 3.3798823831578475,
      "tokens_seen": 2397110272
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010198127482504256,
      "loss": 2.6995,
      "theoretical_loss": 3.3798749503242638,
      "tokens_seen": 2397175808
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010196472479667109,
      "loss": 2.5061,
      "theoretical_loss": 3.3798675177507773,
      "tokens_seen": 2397241344
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010194817476829961,
      "loss": 2.4321,
      "theoretical_loss": 3.379860085437373,
      "tokens_seen": 2397306880
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010193162473992814,
      "loss": 2.5607,
      "theoretical_loss": 3.3798526533840336,
      "tokens_seen": 2397372416
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010191507471155666,
      "loss": 2.6489,
      "theoretical_loss": 3.379845221590744,
      "tokens_seen": 2397437952
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010189852468318518,
      "loss": 2.6658,
      "theoretical_loss": 3.379837790057487,
      "tokens_seen": 2397503488
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010188197465481371,
      "loss": 2.7577,
      "theoretical_loss": 3.3798303587842473,
      "tokens_seen": 2397569024
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001018654246264422,
      "loss": 2.5877,
      "theoretical_loss": 3.3798229277710083,
      "tokens_seen": 2397634560
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010184887459807073,
      "loss": 2.6502,
      "theoretical_loss": 3.3798154970177534,
      "tokens_seen": 2397700096
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010183232456969925,
      "loss": 2.4872,
      "theoretical_loss": 3.3798080665244665,
      "tokens_seen": 2397765632
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010181577454132778,
      "loss": 2.5632,
      "theoretical_loss": 3.379800636291132,
      "tokens_seen": 2397831168
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001017992245129563,
      "loss": 0.2276,
      "theoretical_loss": 3.379793206317734,
      "tokens_seen": 2397896704
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010178267448458482,
      "loss": 2.7372,
      "theoretical_loss": 3.3797857766042547,
      "tokens_seen": 2397962240
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010176612445621335,
      "loss": 2.4948,
      "theoretical_loss": 3.379778347150679,
      "tokens_seen": 2398027776
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010174957442784187,
      "loss": 2.5311,
      "theoretical_loss": 3.3797709179569906,
      "tokens_seen": 2398093312
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001017330243994704,
      "loss": 2.2708,
      "theoretical_loss": 3.379763489023173,
      "tokens_seen": 2398158848
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010171647437109892,
      "loss": 2.66,
      "theoretical_loss": 3.3797560603492105,
      "tokens_seen": 2398224384
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010169992434272744,
      "loss": 2.6311,
      "theoretical_loss": 3.3797486319350867,
      "tokens_seen": 2398289920
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010168337431435597,
      "loss": 2.4438,
      "theoretical_loss": 3.3797412037807857,
      "tokens_seen": 2398355456
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010166682428598449,
      "loss": 2.1187,
      "theoretical_loss": 3.3797337758862906,
      "tokens_seen": 2398420992
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010165027425761302,
      "loss": 2.6891,
      "theoretical_loss": 3.3797263482515856,
      "tokens_seen": 2398486528
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010163372422924154,
      "loss": 2.5288,
      "theoretical_loss": 3.3797189208766545,
      "tokens_seen": 2398552064
    },
    {
      "epoch": 0.86,
      "objective/train/avg_token_score": 0.005777479149401188,
      "objective/train/avg_weight": 0.9932004809379578,
      "objective/train/docs_used": 1348115,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.917660713195801,
      "objective/train/std_weight": 0.029734211042523384,
      "objective/train/theoretical_loss": 3.3797114937614814,
      "objective/train/tokens_used": 2419077600,
      "theoretical_loss": 3.3797114937614814,
      "tokens_seen": 2398617600
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010161717420087005,
      "loss": 2.6583,
      "theoretical_loss": 3.3797114937614814,
      "tokens_seen": 2398617600
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010160062417249857,
      "loss": 2.5306,
      "theoretical_loss": 3.37970406690605,
      "tokens_seen": 2398683136
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001015840741441271,
      "loss": 2.6059,
      "theoretical_loss": 3.3796966403103434,
      "tokens_seen": 2398748672
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010156752411575562,
      "loss": 2.8656,
      "theoretical_loss": 3.3796892139743466,
      "tokens_seen": 2398814208
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010155097408738415,
      "loss": 2.4702,
      "theoretical_loss": 3.3796817878980425,
      "tokens_seen": 2398879744
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010153442405901267,
      "loss": 2.7528,
      "theoretical_loss": 3.379674362081415,
      "tokens_seen": 2398945280
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001015178740306412,
      "loss": 2.4554,
      "theoretical_loss": 3.3796669365244485,
      "tokens_seen": 2399010816
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010150132400226972,
      "loss": 2.6226,
      "theoretical_loss": 3.3796595112271266,
      "tokens_seen": 2399076352
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010148477397389824,
      "loss": 2.5632,
      "theoretical_loss": 3.379652086189433,
      "tokens_seen": 2399141888
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010146822394552677,
      "loss": 2.5603,
      "theoretical_loss": 3.3796446614113513,
      "tokens_seen": 2399207424
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010145167391715529,
      "loss": 2.7207,
      "theoretical_loss": 3.379637236892866,
      "tokens_seen": 2399272960
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010143512388878381,
      "loss": 2.5848,
      "theoretical_loss": 3.37962981263396,
      "tokens_seen": 2399338496
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010141857386041234,
      "loss": 2.4594,
      "theoretical_loss": 3.379622388634618,
      "tokens_seen": 2399404032
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010140202383204086,
      "loss": 2.7836,
      "theoretical_loss": 3.3796149648948237,
      "tokens_seen": 2399469568
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010138547380366937,
      "loss": 2.4232,
      "theoretical_loss": 3.3796075414145603,
      "tokens_seen": 2399535104
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001013689237752979,
      "loss": 2.6174,
      "theoretical_loss": 3.3796001181938125,
      "tokens_seen": 2399600640
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010135237374692642,
      "loss": 2.7024,
      "theoretical_loss": 3.3795926952325637,
      "tokens_seen": 2399666176
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010133582371855495,
      "loss": 2.8133,
      "theoretical_loss": 3.3795852725307975,
      "tokens_seen": 2399731712
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010131927369018347,
      "loss": 2.246,
      "theoretical_loss": 3.3795778500884976,
      "tokens_seen": 2399797248
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010130272366181199,
      "loss": 2.4942,
      "theoretical_loss": 3.379570427905649,
      "tokens_seen": 2399862784
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010128617363344052,
      "loss": 3.0347,
      "theoretical_loss": 3.3795630059822344,
      "tokens_seen": 2399928320
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010126962360506904,
      "loss": 2.7942,
      "theoretical_loss": 3.379555584318238,
      "tokens_seen": 2399993856
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010125307357669756,
      "loss": 2.2459,
      "theoretical_loss": 3.379548162913644,
      "tokens_seen": 2400059392
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010123652354832609,
      "loss": 2.4926,
      "theoretical_loss": 3.379540741768436,
      "tokens_seen": 2400124928
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010121997351995461,
      "loss": 2.5957,
      "theoretical_loss": 3.3795333208825973,
      "tokens_seen": 2400190464
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010120342349158314,
      "loss": 2.812,
      "theoretical_loss": 3.3795259002561124,
      "tokens_seen": 2400256000
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010118687346321166,
      "loss": 2.4869,
      "theoretical_loss": 3.379518479888965,
      "tokens_seen": 2400321536
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010117032343484018,
      "loss": 2.4733,
      "theoretical_loss": 3.379511059781139,
      "tokens_seen": 2400387072
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010115377340646871,
      "loss": 2.8289,
      "theoretical_loss": 3.3795036399326186,
      "tokens_seen": 2400452608
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001011372233780972,
      "loss": 2.8812,
      "theoretical_loss": 3.3794962203433867,
      "tokens_seen": 2400518144
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010112067334972573,
      "loss": 2.6476,
      "theoretical_loss": 3.3794888010134283,
      "tokens_seen": 2400583680
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010110412332135425,
      "loss": 2.7641,
      "theoretical_loss": 3.379481381942726,
      "tokens_seen": 2400649216
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010108757329298278,
      "loss": 2.6262,
      "theoretical_loss": 3.379473963131265,
      "tokens_seen": 2400714752
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001010710232646113,
      "loss": 2.5764,
      "theoretical_loss": 3.3794665445790284,
      "tokens_seen": 2400780288
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010105447323623983,
      "loss": 2.4369,
      "theoretical_loss": 3.379459126286,
      "tokens_seen": 2400845824
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010103792320786835,
      "loss": 2.7131,
      "theoretical_loss": 3.379451708252164,
      "tokens_seen": 2400911360
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010102137317949687,
      "loss": 2.3455,
      "theoretical_loss": 3.3794442904775037,
      "tokens_seen": 2400976896
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001010048231511254,
      "loss": 2.4131,
      "theoretical_loss": 3.3794368729620037,
      "tokens_seen": 2401042432
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010098827312275392,
      "loss": 2.6127,
      "theoretical_loss": 3.3794294557056475,
      "tokens_seen": 2401107968
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010097172309438245,
      "loss": 2.4965,
      "theoretical_loss": 3.3794220387084195,
      "tokens_seen": 2401173504
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010095517306601097,
      "loss": 2.6252,
      "theoretical_loss": 3.3794146219703025,
      "tokens_seen": 2401239040
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001009386230376395,
      "loss": 2.558,
      "theoretical_loss": 3.3794072054912814,
      "tokens_seen": 2401304576
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010092207300926802,
      "loss": 2.8044,
      "theoretical_loss": 3.379399789271339,
      "tokens_seen": 2401370112
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010090552298089654,
      "loss": 2.4863,
      "theoretical_loss": 3.379392373310461,
      "tokens_seen": 2401435648
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010088897295252505,
      "loss": 2.5772,
      "theoretical_loss": 3.3793849576086297,
      "tokens_seen": 2401501184
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010087242292415358,
      "loss": 2.7781,
      "theoretical_loss": 3.379377542165829,
      "tokens_seen": 2401566720
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001008558728957821,
      "loss": 2.5198,
      "theoretical_loss": 3.3793701269820433,
      "tokens_seen": 2401632256
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010083932286741062,
      "loss": 1.8071,
      "theoretical_loss": 3.3793627120572567,
      "tokens_seen": 2401697792
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010082277283903915,
      "loss": 2.6711,
      "theoretical_loss": 3.3793552973914522,
      "tokens_seen": 2401763328
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010080622281066767,
      "loss": 2.2897,
      "theoretical_loss": 3.379347882984615,
      "tokens_seen": 2401828864
    },
    {
      "epoch": 0.86,
      "objective/train/avg_token_score": 0.004511049948632717,
      "objective/train/avg_weight": 0.9946910738945007,
      "objective/train/docs_used": 1349979,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1285452842712402,
      "objective/train/std_weight": 0.034413762390613556,
      "objective/train/theoretical_loss": 3.3793404688367277,
      "objective/train/tokens_used": 2422354400,
      "theoretical_loss": 3.3793404688367277,
      "tokens_seen": 2401894400
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001007896727822962,
      "loss": 2.83,
      "theoretical_loss": 3.3793404688367277,
      "tokens_seen": 2401894400
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010077312275392472,
      "loss": 2.6866,
      "theoretical_loss": 3.379333054947775,
      "tokens_seen": 2401959936
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010075657272555324,
      "loss": 2.8652,
      "theoretical_loss": 3.3793256413177404,
      "tokens_seen": 2402025472
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010074002269718177,
      "loss": 2.6384,
      "theoretical_loss": 3.3793182279466083,
      "tokens_seen": 2402091008
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010072347266881029,
      "loss": 2.4587,
      "theoretical_loss": 3.379310814834362,
      "tokens_seen": 2402156544
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010070692264043882,
      "loss": 2.6074,
      "theoretical_loss": 3.3793034019809856,
      "tokens_seen": 2402222080
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010069037261206734,
      "loss": 2.8617,
      "theoretical_loss": 3.3792959893864625,
      "tokens_seen": 2402287616
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010067382258369586,
      "loss": 2.5342,
      "theoretical_loss": 3.3792885770507777,
      "tokens_seen": 2402353152
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010065727255532439,
      "loss": 2.6877,
      "theoretical_loss": 3.3792811649739143,
      "tokens_seen": 2402418688
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001006407225269529,
      "loss": 2.4562,
      "theoretical_loss": 3.3792737531558568,
      "tokens_seen": 2402484224
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010062417249858142,
      "loss": 2.4043,
      "theoretical_loss": 3.3792663415965882,
      "tokens_seen": 2402549760
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010060762247020995,
      "loss": 2.562,
      "theoretical_loss": 3.379258930296093,
      "tokens_seen": 2402615296
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010059107244183847,
      "loss": 2.6551,
      "theoretical_loss": 3.3792515192543555,
      "tokens_seen": 2402680832
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.000100574522413467,
      "loss": 2.3084,
      "theoretical_loss": 3.379244108471359,
      "tokens_seen": 2402746368
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010055797238509552,
      "loss": 2.4116,
      "theoretical_loss": 3.3792366979470874,
      "tokens_seen": 2402811904
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010054142235672404,
      "loss": 2.6095,
      "theoretical_loss": 3.3792292876815244,
      "tokens_seen": 2402877440
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010052487232835257,
      "loss": 2.576,
      "theoretical_loss": 3.379221877674655,
      "tokens_seen": 2402942976
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010050832229998109,
      "loss": 2.6299,
      "theoretical_loss": 3.3792144679264617,
      "tokens_seen": 2403008512
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010049177227160961,
      "loss": 2.3588,
      "theoretical_loss": 3.3792070584369296,
      "tokens_seen": 2403074048
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010047522224323814,
      "loss": 2.6711,
      "theoretical_loss": 3.3791996492060417,
      "tokens_seen": 2403139584
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010045867221486666,
      "loss": 2.5831,
      "theoretical_loss": 3.379192240233783,
      "tokens_seen": 2403205120
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010044212218649519,
      "loss": 2.5174,
      "theoretical_loss": 3.3791848315201363,
      "tokens_seen": 2403270656
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010042557215812371,
      "loss": 2.7486,
      "theoretical_loss": 3.379177423065086,
      "tokens_seen": 2403336192
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010040902212975223,
      "loss": 2.6413,
      "theoretical_loss": 3.379170014868616,
      "tokens_seen": 2403401728
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010039247210138073,
      "loss": 2.5224,
      "theoretical_loss": 3.37916260693071,
      "tokens_seen": 2403467264
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010037592207300926,
      "loss": 2.5341,
      "theoretical_loss": 3.3791551992513527,
      "tokens_seen": 2403532800
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010035937204463778,
      "loss": 2.5777,
      "theoretical_loss": 3.3791477918305275,
      "tokens_seen": 2403598336
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001003428220162663,
      "loss": 2.5593,
      "theoretical_loss": 3.3791403846682178,
      "tokens_seen": 2403663872
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010032627198789483,
      "loss": 2.3952,
      "theoretical_loss": 3.3791329777644084,
      "tokens_seen": 2403729408
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010030972195952335,
      "loss": 2.5244,
      "theoretical_loss": 3.379125571119083,
      "tokens_seen": 2403794944
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010029317193115188,
      "loss": 2.6587,
      "theoretical_loss": 3.379118164732225,
      "tokens_seen": 2403860480
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001002766219027804,
      "loss": 2.5182,
      "theoretical_loss": 3.3791107586038187,
      "tokens_seen": 2403926016
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010026007187440892,
      "loss": 2.4658,
      "theoretical_loss": 3.379103352733848,
      "tokens_seen": 2403991552
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010024352184603745,
      "loss": 2.597,
      "theoretical_loss": 3.3790959471222974,
      "tokens_seen": 2404057088
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010022697181766597,
      "loss": 2.5933,
      "theoretical_loss": 3.3790885417691503,
      "tokens_seen": 2404122624
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001002104217892945,
      "loss": 2.616,
      "theoretical_loss": 3.3790811366743903,
      "tokens_seen": 2404188160
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010019387176092302,
      "loss": 2.4374,
      "theoretical_loss": 3.379073731838002,
      "tokens_seen": 2404253696
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010017732173255154,
      "loss": 2.5668,
      "theoretical_loss": 3.3790663272599692,
      "tokens_seen": 2404319232
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010016077170418007,
      "loss": 2.7477,
      "theoretical_loss": 3.3790589229402754,
      "tokens_seen": 2404384768
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010014422167580858,
      "loss": 2.7648,
      "theoretical_loss": 3.379051518878905,
      "tokens_seen": 2404450304
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001001276716474371,
      "loss": 2.5386,
      "theoretical_loss": 3.3790441150758417,
      "tokens_seen": 2404515840
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010011112161906563,
      "loss": 2.5586,
      "theoretical_loss": 3.3790367115310698,
      "tokens_seen": 2404581376
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010009457159069415,
      "loss": 2.4812,
      "theoretical_loss": 3.379029308244573,
      "tokens_seen": 2404646912
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010007802156232267,
      "loss": 2.5384,
      "theoretical_loss": 3.379021905216335,
      "tokens_seen": 2404712448
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001000614715339512,
      "loss": 2.692,
      "theoretical_loss": 3.3790145024463403,
      "tokens_seen": 2404777984
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010004492150557972,
      "loss": 2.5792,
      "theoretical_loss": 3.3790070999345723,
      "tokens_seen": 2404843520
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010002837147720825,
      "loss": 2.1345,
      "theoretical_loss": 3.3789996976810155,
      "tokens_seen": 2404909056
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.00010001182144883677,
      "loss": 2.4284,
      "theoretical_loss": 3.378992295685653,
      "tokens_seen": 2404974592
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.99952714204653e-05,
      "loss": 2.6218,
      "theoretical_loss": 3.37898489394847,
      "tokens_seen": 2405040128
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.997872139209382e-05,
      "loss": 2.7462,
      "theoretical_loss": 3.3789774924694496,
      "tokens_seen": 2405105664
    },
    {
      "epoch": 0.86,
      "objective/train/avg_token_score": 0.03191853314638138,
      "objective/train/avg_weight": 0.9624383449554443,
      "objective/train/docs_used": 1352287,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.236175060272217,
      "objective/train/std_weight": 0.15201294422149658,
      "objective/train/theoretical_loss": 3.378970091248576,
      "objective/train/tokens_used": 2425631200,
      "theoretical_loss": 3.378970091248576,
      "tokens_seen": 2405171200
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.996217136372234e-05,
      "loss": 2.5647,
      "theoretical_loss": 3.378970091248576,
      "tokens_seen": 2405171200
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.994562133535087e-05,
      "loss": 2.4692,
      "theoretical_loss": 3.378962690285833,
      "tokens_seen": 2405236736
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.992907130697939e-05,
      "loss": 2.6608,
      "theoretical_loss": 3.3789552895812047,
      "tokens_seen": 2405302272
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.991252127860791e-05,
      "loss": 2.3223,
      "theoretical_loss": 3.378947889134675,
      "tokens_seen": 2405367808
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.989597125023642e-05,
      "loss": 2.64,
      "theoretical_loss": 3.378940488946228,
      "tokens_seen": 2405433344
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.987942122186495e-05,
      "loss": 2.5781,
      "theoretical_loss": 3.3789330890158475,
      "tokens_seen": 2405498880
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.986287119349347e-05,
      "loss": 2.6402,
      "theoretical_loss": 3.378925689343518,
      "tokens_seen": 2405564416
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.9846321165122e-05,
      "loss": 2.6125,
      "theoretical_loss": 3.3789182899292225,
      "tokens_seen": 2405629952
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.982977113675052e-05,
      "loss": 2.5371,
      "theoretical_loss": 3.3789108907729455,
      "tokens_seen": 2405695488
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.981322110837904e-05,
      "loss": 2.8113,
      "theoretical_loss": 3.378903491874671,
      "tokens_seen": 2405761024
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.979667108000757e-05,
      "loss": 2.5909,
      "theoretical_loss": 3.3788960932343834,
      "tokens_seen": 2405826560
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.978012105163609e-05,
      "loss": 2.4046,
      "theoretical_loss": 3.378888694852066,
      "tokens_seen": 2405892096
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.976357102326462e-05,
      "loss": 2.6129,
      "theoretical_loss": 3.3788812967277027,
      "tokens_seen": 2405957632
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.974702099489314e-05,
      "loss": 2.5555,
      "theoretical_loss": 3.3788738988612783,
      "tokens_seen": 2406023168
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.973047096652166e-05,
      "loss": 2.5022,
      "theoretical_loss": 3.3788665012527757,
      "tokens_seen": 2406088704
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.971392093815019e-05,
      "loss": 2.3121,
      "theoretical_loss": 3.37885910390218,
      "tokens_seen": 2406154240
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.969737090977871e-05,
      "loss": 2.4534,
      "theoretical_loss": 3.378851706809474,
      "tokens_seen": 2406219776
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.968082088140724e-05,
      "loss": 2.5478,
      "theoretical_loss": 3.3788443099746432,
      "tokens_seen": 2406285312
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.966427085303573e-05,
      "loss": 2.3763,
      "theoretical_loss": 3.37883691339767,
      "tokens_seen": 2406350848
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.964772082466426e-05,
      "loss": 2.6252,
      "theoretical_loss": 3.3788295170785396,
      "tokens_seen": 2406416384
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.963117079629278e-05,
      "loss": 2.664,
      "theoretical_loss": 3.378822121017235,
      "tokens_seen": 2406481920
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.96146207679213e-05,
      "loss": 2.4382,
      "theoretical_loss": 3.378814725213741,
      "tokens_seen": 2406547456
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.959807073954983e-05,
      "loss": 2.6314,
      "theoretical_loss": 3.378807329668041,
      "tokens_seen": 2406612992
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.958152071117835e-05,
      "loss": 2.8613,
      "theoretical_loss": 3.3787999343801194,
      "tokens_seen": 2406678528
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.956497068280688e-05,
      "loss": 2.6803,
      "theoretical_loss": 3.37879253934996,
      "tokens_seen": 2406744064
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.95484206544354e-05,
      "loss": 2.5277,
      "theoretical_loss": 3.378785144577547,
      "tokens_seen": 2406809600
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.953187062606392e-05,
      "loss": 2.6028,
      "theoretical_loss": 3.3787777500628646,
      "tokens_seen": 2406875136
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.951532059769245e-05,
      "loss": 2.6511,
      "theoretical_loss": 3.378770355805896,
      "tokens_seen": 2406940672
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.949877056932097e-05,
      "loss": 2.6762,
      "theoretical_loss": 3.3787629618066255,
      "tokens_seen": 2407006208
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.94822205409495e-05,
      "loss": 2.3525,
      "theoretical_loss": 3.3787555680650376,
      "tokens_seen": 2407071744
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.946567051257802e-05,
      "loss": 2.7037,
      "theoretical_loss": 3.3787481745811156,
      "tokens_seen": 2407137280
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.944912048420654e-05,
      "loss": 2.6262,
      "theoretical_loss": 3.3787407813548445,
      "tokens_seen": 2407202816
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.943257045583507e-05,
      "loss": 2.5253,
      "theoretical_loss": 3.378733388386207,
      "tokens_seen": 2407268352
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.941602042746358e-05,
      "loss": 2.3702,
      "theoretical_loss": 3.3787259956751883,
      "tokens_seen": 2407333888
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.93994703990921e-05,
      "loss": 2.4247,
      "theoretical_loss": 3.3787186032217718,
      "tokens_seen": 2407399424
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.938292037072063e-05,
      "loss": 2.1543,
      "theoretical_loss": 3.378711211025941,
      "tokens_seen": 2407464960
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.936637034234915e-05,
      "loss": 2.243,
      "theoretical_loss": 3.3787038190876815,
      "tokens_seen": 2407530496
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.934982031397767e-05,
      "loss": 2.5057,
      "theoretical_loss": 3.3786964274069757,
      "tokens_seen": 2407596032
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.93332702856062e-05,
      "loss": 2.8957,
      "theoretical_loss": 3.3786890359838084,
      "tokens_seen": 2407661568
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.931672025723472e-05,
      "loss": 2.4635,
      "theoretical_loss": 3.3786816448181636,
      "tokens_seen": 2407727104
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.930017022886325e-05,
      "loss": 2.5871,
      "theoretical_loss": 3.378674253910025,
      "tokens_seen": 2407792640
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.928362020049177e-05,
      "loss": 2.7478,
      "theoretical_loss": 3.378666863259377,
      "tokens_seen": 2407858176
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.92670701721203e-05,
      "loss": 2.5123,
      "theoretical_loss": 3.3786594728662034,
      "tokens_seen": 2407923712
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.925052014374882e-05,
      "loss": 2.8649,
      "theoretical_loss": 3.378652082730488,
      "tokens_seen": 2407989248
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.923397011537734e-05,
      "loss": 2.1613,
      "theoretical_loss": 3.378644692852215,
      "tokens_seen": 2408054784
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.921742008700587e-05,
      "loss": 2.7055,
      "theoretical_loss": 3.3786373032313692,
      "tokens_seen": 2408120320
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.920087005863439e-05,
      "loss": 2.4613,
      "theoretical_loss": 3.3786299138679334,
      "tokens_seen": 2408185856
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.918432003026291e-05,
      "loss": 2.2398,
      "theoretical_loss": 3.378622524761892,
      "tokens_seen": 2408251392
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.916777000189143e-05,
      "loss": 2.3653,
      "theoretical_loss": 3.3786151359132295,
      "tokens_seen": 2408316928
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.915121997351995e-05,
      "loss": 2.2915,
      "theoretical_loss": 3.37860774732193,
      "tokens_seen": 2408382464
    },
    {
      "epoch": 0.86,
      "objective/train/avg_token_score": 0.027598315849900246,
      "objective/train/avg_weight": 0.9675138592720032,
      "objective/train/docs_used": 1354179,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.905780792236328,
      "objective/train/std_weight": 0.10537509620189667,
      "objective/train/theoretical_loss": 3.3786003589879763,
      "objective/train/tokens_used": 2428908000,
      "theoretical_loss": 3.3786003589879763,
      "tokens_seen": 2408448000
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.913466994514847e-05,
      "loss": 2.5277,
      "theoretical_loss": 3.3786003589879763,
      "tokens_seen": 2408448000
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.9118119916777e-05,
      "loss": 2.4551,
      "theoretical_loss": 3.378592970911354,
      "tokens_seen": 2408513536
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.910156988840552e-05,
      "loss": 2.547,
      "theoretical_loss": 3.378585583092046,
      "tokens_seen": 2408579072
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.908501986003405e-05,
      "loss": 2.2265,
      "theoretical_loss": 3.3785781955300367,
      "tokens_seen": 2408644608
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.906846983166257e-05,
      "loss": 2.5841,
      "theoretical_loss": 3.3785708082253105,
      "tokens_seen": 2408710144
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.90519198032911e-05,
      "loss": 2.5504,
      "theoretical_loss": 3.378563421177851,
      "tokens_seen": 2408775680
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.903536977491962e-05,
      "loss": 2.4736,
      "theoretical_loss": 3.378556034387642,
      "tokens_seen": 2408841216
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.901881974654814e-05,
      "loss": 2.4811,
      "theoretical_loss": 3.378548647854669,
      "tokens_seen": 2408906752
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.900226971817667e-05,
      "loss": 2.4254,
      "theoretical_loss": 3.378541261578914,
      "tokens_seen": 2408972288
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.898571968980519e-05,
      "loss": 2.6258,
      "theoretical_loss": 3.378533875560362,
      "tokens_seen": 2409037824
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.896916966143371e-05,
      "loss": 2.5942,
      "theoretical_loss": 3.3785264897989977,
      "tokens_seen": 2409103360
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.895261963306224e-05,
      "loss": 2.5646,
      "theoretical_loss": 3.378519104294804,
      "tokens_seen": 2409168896
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.893606960469076e-05,
      "loss": 2.1013,
      "theoretical_loss": 3.378511719047766,
      "tokens_seen": 2409234432
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.891951957631926e-05,
      "loss": 2.3918,
      "theoretical_loss": 3.3785043340578667,
      "tokens_seen": 2409299968
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.890296954794778e-05,
      "loss": 2.3449,
      "theoretical_loss": 3.3784969493250907,
      "tokens_seen": 2409365504
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.88864195195763e-05,
      "loss": 2.6508,
      "theoretical_loss": 3.378489564849422,
      "tokens_seen": 2409431040
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.886986949120483e-05,
      "loss": 2.5886,
      "theoretical_loss": 3.378482180630845,
      "tokens_seen": 2409496576
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.885331946283335e-05,
      "loss": 2.649,
      "theoretical_loss": 3.378474796669343,
      "tokens_seen": 2409562112
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.883676943446188e-05,
      "loss": 2.6628,
      "theoretical_loss": 3.378467412964901,
      "tokens_seen": 2409627648
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.88202194060904e-05,
      "loss": 2.4439,
      "theoretical_loss": 3.3784600295175022,
      "tokens_seen": 2409693184
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.880366937771893e-05,
      "loss": 2.5766,
      "theoretical_loss": 3.3784526463271307,
      "tokens_seen": 2409758720
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.878711934934745e-05,
      "loss": 2.4588,
      "theoretical_loss": 3.3784452633937714,
      "tokens_seen": 2409824256
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.877056932097597e-05,
      "loss": 2.5438,
      "theoretical_loss": 3.378437880717408,
      "tokens_seen": 2409889792
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.87540192926045e-05,
      "loss": 2.7301,
      "theoretical_loss": 3.378430498298024,
      "tokens_seen": 2409955328
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.873746926423304e-05,
      "loss": 2.6814,
      "theoretical_loss": 3.378423116135604,
      "tokens_seen": 2410020864
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.872091923586156e-05,
      "loss": 2.4194,
      "theoretical_loss": 3.3784157342301318,
      "tokens_seen": 2410086400
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.870436920749008e-05,
      "loss": 2.4985,
      "theoretical_loss": 3.378408352581592,
      "tokens_seen": 2410151936
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.868781917911861e-05,
      "loss": 2.2714,
      "theoretical_loss": 3.3784009711899676,
      "tokens_seen": 2410217472
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.86712691507471e-05,
      "loss": 2.4795,
      "theoretical_loss": 3.3783935900552438,
      "tokens_seen": 2410283008
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.865471912237563e-05,
      "loss": 2.7175,
      "theoretical_loss": 3.378386209177404,
      "tokens_seen": 2410348544
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.863816909400415e-05,
      "loss": 2.1945,
      "theoretical_loss": 3.378378828556433,
      "tokens_seen": 2410414080
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.862161906563268e-05,
      "loss": 2.5911,
      "theoretical_loss": 3.378371448192314,
      "tokens_seen": 2410479616
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.86050690372612e-05,
      "loss": 2.4913,
      "theoretical_loss": 3.3783640680850313,
      "tokens_seen": 2410545152
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.858851900888972e-05,
      "loss": 2.5018,
      "theoretical_loss": 3.378356688234569,
      "tokens_seen": 2410610688
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.857196898051825e-05,
      "loss": 2.7642,
      "theoretical_loss": 3.378349308640912,
      "tokens_seen": 2410676224
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.855541895214677e-05,
      "loss": 2.5318,
      "theoretical_loss": 3.3783419293040433,
      "tokens_seen": 2410741760
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.85388689237753e-05,
      "loss": 2.6404,
      "theoretical_loss": 3.378334550223947,
      "tokens_seen": 2410807296
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.852231889540382e-05,
      "loss": 2.4969,
      "theoretical_loss": 3.378327171400608,
      "tokens_seen": 2410872832
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.850576886703234e-05,
      "loss": 2.4148,
      "theoretical_loss": 3.3783197928340103,
      "tokens_seen": 2410938368
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.848921883866087e-05,
      "loss": 2.0994,
      "theoretical_loss": 3.378312414524137,
      "tokens_seen": 2411003904
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.847266881028939e-05,
      "loss": 2.7862,
      "theoretical_loss": 3.378305036470973,
      "tokens_seen": 2411069440
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.845611878191792e-05,
      "loss": 2.4818,
      "theoretical_loss": 3.378297658674503,
      "tokens_seen": 2411134976
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.843956875354644e-05,
      "loss": 2.4222,
      "theoretical_loss": 3.3782902811347095,
      "tokens_seen": 2411200512
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.842301872517495e-05,
      "loss": 2.6385,
      "theoretical_loss": 3.3782829038515776,
      "tokens_seen": 2411266048
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.840646869680347e-05,
      "loss": 2.4633,
      "theoretical_loss": 3.378275526825091,
      "tokens_seen": 2411331584
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.8389918668432e-05,
      "loss": 2.6097,
      "theoretical_loss": 3.3782681500552343,
      "tokens_seen": 2411397120
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.837336864006052e-05,
      "loss": 2.6431,
      "theoretical_loss": 3.378260773541991,
      "tokens_seen": 2411462656
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.835681861168905e-05,
      "loss": 2.1456,
      "theoretical_loss": 3.378253397285346,
      "tokens_seen": 2411528192
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.834026858331757e-05,
      "loss": 2.0939,
      "theoretical_loss": 3.3782460212852827,
      "tokens_seen": 2411593728
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.83237185549461e-05,
      "loss": 1.9671,
      "theoretical_loss": 3.378238645541785,
      "tokens_seen": 2411659264
    },
    {
      "epoch": 0.86,
      "objective/train/avg_token_score": 0.018498850986361504,
      "objective/train/avg_weight": 0.9782499074935913,
      "objective/train/docs_used": 1355984,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9745779037475586,
      "objective/train/std_weight": 0.09847763925790787,
      "objective/train/theoretical_loss": 3.3782312700548376,
      "objective/train/tokens_used": 2432184800,
      "theoretical_loss": 3.3782312700548376,
      "tokens_seen": 2411724800
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.830716852657462e-05,
      "loss": 2.4757,
      "theoretical_loss": 3.3782312700548376,
      "tokens_seen": 2411724800
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.829061849820314e-05,
      "loss": 2.4905,
      "theoretical_loss": 3.3782238948244245,
      "tokens_seen": 2411790336
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.827406846983167e-05,
      "loss": 2.4982,
      "theoretical_loss": 3.37821651985053,
      "tokens_seen": 2411855872
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.825751844146019e-05,
      "loss": 2.486,
      "theoretical_loss": 3.3782091451331375,
      "tokens_seen": 2411921408
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.824096841308871e-05,
      "loss": 2.3832,
      "theoretical_loss": 3.3782017706722316,
      "tokens_seen": 2411986944
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.822441838471724e-05,
      "loss": 2.6873,
      "theoretical_loss": 3.3781943964677965,
      "tokens_seen": 2412052480
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.820786835634576e-05,
      "loss": 2.7601,
      "theoretical_loss": 3.3781870225198163,
      "tokens_seen": 2412118016
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.819131832797426e-05,
      "loss": 2.7693,
      "theoretical_loss": 3.378179648828275,
      "tokens_seen": 2412183552
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.817476829960278e-05,
      "loss": 2.67,
      "theoretical_loss": 3.3781722753931565,
      "tokens_seen": 2412249088
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.815821827123131e-05,
      "loss": 2.4942,
      "theoretical_loss": 3.378164902214445,
      "tokens_seen": 2412314624
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.814166824285983e-05,
      "loss": 2.7303,
      "theoretical_loss": 3.3781575292921247,
      "tokens_seen": 2412380160
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.812511821448837e-05,
      "loss": 2.3855,
      "theoretical_loss": 3.37815015662618,
      "tokens_seen": 2412445696
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.810856818611689e-05,
      "loss": 2.6779,
      "theoretical_loss": 3.3781427842165948,
      "tokens_seen": 2412511232
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.809201815774542e-05,
      "loss": 2.571,
      "theoretical_loss": 3.3781354120633527,
      "tokens_seen": 2412576768
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.807546812937394e-05,
      "loss": 2.8138,
      "theoretical_loss": 3.378128040166439,
      "tokens_seen": 2412642304
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.805891810100246e-05,
      "loss": 2.5677,
      "theoretical_loss": 3.3781206685258365,
      "tokens_seen": 2412707840
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.804236807263099e-05,
      "loss": 2.5636,
      "theoretical_loss": 3.3781132971415304,
      "tokens_seen": 2412773376
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.802581804425951e-05,
      "loss": 2.1961,
      "theoretical_loss": 3.3781059260135042,
      "tokens_seen": 2412838912
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.800926801588804e-05,
      "loss": 2.5206,
      "theoretical_loss": 3.3780985551417424,
      "tokens_seen": 2412904448
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.799271798751656e-05,
      "loss": 2.6102,
      "theoretical_loss": 3.3780911845262285,
      "tokens_seen": 2412969984
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.797616795914508e-05,
      "loss": 2.4783,
      "theoretical_loss": 3.378083814166948,
      "tokens_seen": 2413035520
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.795961793077361e-05,
      "loss": 2.4524,
      "theoretical_loss": 3.378076444063883,
      "tokens_seen": 2413101056
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.79430679024021e-05,
      "loss": 2.5883,
      "theoretical_loss": 3.378069074217019,
      "tokens_seen": 2413166592
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.792651787403063e-05,
      "loss": 2.5301,
      "theoretical_loss": 3.3780617046263406,
      "tokens_seen": 2413232128
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.790996784565915e-05,
      "loss": 2.5516,
      "theoretical_loss": 3.378054335291831,
      "tokens_seen": 2413297664
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.789341781728768e-05,
      "loss": 2.6263,
      "theoretical_loss": 3.378046966213474,
      "tokens_seen": 2413363200
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.78768677889162e-05,
      "loss": 2.4496,
      "theoretical_loss": 3.378039597391255,
      "tokens_seen": 2413428736
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.786031776054473e-05,
      "loss": 2.612,
      "theoretical_loss": 3.378032228825157,
      "tokens_seen": 2413494272
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.784376773217325e-05,
      "loss": 2.6802,
      "theoretical_loss": 3.378024860515165,
      "tokens_seen": 2413559808
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.782721770380177e-05,
      "loss": 2.5217,
      "theoretical_loss": 3.3780174924612623,
      "tokens_seen": 2413625344
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.78106676754303e-05,
      "loss": 2.531,
      "theoretical_loss": 3.3780101246634335,
      "tokens_seen": 2413690880
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.779411764705882e-05,
      "loss": 2.4863,
      "theoretical_loss": 3.3780027571216626,
      "tokens_seen": 2413756416
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.777756761868735e-05,
      "loss": 2.4984,
      "theoretical_loss": 3.3779953898359345,
      "tokens_seen": 2413821952
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.776101759031587e-05,
      "loss": 2.5335,
      "theoretical_loss": 3.3779880228062322,
      "tokens_seen": 2413887488
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.77444675619444e-05,
      "loss": 2.5778,
      "theoretical_loss": 3.3779806560325407,
      "tokens_seen": 2413953024
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.772791753357292e-05,
      "loss": 2.5117,
      "theoretical_loss": 3.3779732895148435,
      "tokens_seen": 2414018560
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.771136750520144e-05,
      "loss": 2.2852,
      "theoretical_loss": 3.3779659232531256,
      "tokens_seen": 2414084096
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.769481747682995e-05,
      "loss": 2.4637,
      "theoretical_loss": 3.3779585572473705,
      "tokens_seen": 2414149632
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.767826744845848e-05,
      "loss": 2.4136,
      "theoretical_loss": 3.377951191497562,
      "tokens_seen": 2414215168
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.7661717420087e-05,
      "loss": 2.4841,
      "theoretical_loss": 3.377943826003685,
      "tokens_seen": 2414280704
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.764516739171552e-05,
      "loss": 2.5391,
      "theoretical_loss": 3.377936460765724,
      "tokens_seen": 2414346240
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.762861736334405e-05,
      "loss": 2.4157,
      "theoretical_loss": 3.377929095783662,
      "tokens_seen": 2414411776
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.761206733497257e-05,
      "loss": 2.5016,
      "theoretical_loss": 3.3779217310574836,
      "tokens_seen": 2414477312
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.75955173066011e-05,
      "loss": 2.4303,
      "theoretical_loss": 3.3779143665871736,
      "tokens_seen": 2414542848
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.757896727822962e-05,
      "loss": 2.427,
      "theoretical_loss": 3.3779070023727154,
      "tokens_seen": 2414608384
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.756241724985814e-05,
      "loss": 2.3841,
      "theoretical_loss": 3.377899638414094,
      "tokens_seen": 2414673920
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.754586722148667e-05,
      "loss": 2.7239,
      "theoretical_loss": 3.377892274711292,
      "tokens_seen": 2414739456
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.752931719311519e-05,
      "loss": 2.4544,
      "theoretical_loss": 3.3778849112642955,
      "tokens_seen": 2414804992
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.751276716474372e-05,
      "loss": 2.4549,
      "theoretical_loss": 3.377877548073087,
      "tokens_seen": 2414870528
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.749621713637224e-05,
      "loss": 2.4778,
      "theoretical_loss": 3.377870185137652,
      "tokens_seen": 2414936064
    },
    {
      "epoch": 0.86,
      "objective/train/avg_token_score": 0.01474373135715723,
      "objective/train/avg_weight": 0.9826673269271851,
      "objective/train/docs_used": 1358162,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.3250296115875244,
      "objective/train/std_weight": 0.08232724666595459,
      "objective/train/theoretical_loss": 3.377862822457974,
      "objective/train/tokens_used": 2435461600,
      "theoretical_loss": 3.377862822457974,
      "tokens_seen": 2415001600
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.747966710800076e-05,
      "loss": 2.6154,
      "theoretical_loss": 3.377862822457974,
      "tokens_seen": 2415001600
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.746311707962929e-05,
      "loss": 2.665,
      "theoretical_loss": 3.3778554600340374,
      "tokens_seen": 2415067136
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.74465670512578e-05,
      "loss": 2.6001,
      "theoretical_loss": 3.377848097865826,
      "tokens_seen": 2415132672
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.743001702288632e-05,
      "loss": 2.5772,
      "theoretical_loss": 3.3778407359533245,
      "tokens_seen": 2415198208
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.741346699451485e-05,
      "loss": 2.4727,
      "theoretical_loss": 3.3778333742965163,
      "tokens_seen": 2415263744
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.739691696614337e-05,
      "loss": 2.3243,
      "theoretical_loss": 3.3778260128953868,
      "tokens_seen": 2415329280
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.73803669377719e-05,
      "loss": 2.6949,
      "theoretical_loss": 3.377818651749919,
      "tokens_seen": 2415394816
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.736381690940042e-05,
      "loss": 2.5036,
      "theoretical_loss": 3.3778112908600977,
      "tokens_seen": 2415460352
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.734726688102894e-05,
      "loss": 2.7435,
      "theoretical_loss": 3.377803930225907,
      "tokens_seen": 2415525888
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.733071685265747e-05,
      "loss": 2.5109,
      "theoretical_loss": 3.377796569847331,
      "tokens_seen": 2415591424
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.731416682428599e-05,
      "loss": 2.283,
      "theoretical_loss": 3.377789209724354,
      "tokens_seen": 2415656960
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.729761679591451e-05,
      "loss": 2.5207,
      "theoretical_loss": 3.37778184985696,
      "tokens_seen": 2415722496
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.728106676754304e-05,
      "loss": 2.5995,
      "theoretical_loss": 3.3777744902451334,
      "tokens_seen": 2415788032
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.726451673917156e-05,
      "loss": 2.1523,
      "theoretical_loss": 3.3777671308888584,
      "tokens_seen": 2415853568
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.724796671080009e-05,
      "loss": 2.3105,
      "theoretical_loss": 3.377759771788119,
      "tokens_seen": 2415919104
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.723141668242861e-05,
      "loss": 2.744,
      "theoretical_loss": 3.3777524129428995,
      "tokens_seen": 2415984640
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.721486665405713e-05,
      "loss": 2.5552,
      "theoretical_loss": 3.3777450543531837,
      "tokens_seen": 2416050176
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.719831662568563e-05,
      "loss": 2.4794,
      "theoretical_loss": 3.377737696018957,
      "tokens_seen": 2416115712
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.718176659731416e-05,
      "loss": 2.6487,
      "theoretical_loss": 3.377730337940202,
      "tokens_seen": 2416181248
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.716521656894268e-05,
      "loss": 2.4692,
      "theoretical_loss": 3.377722980116904,
      "tokens_seen": 2416246784
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.71486665405712e-05,
      "loss": 2.4964,
      "theoretical_loss": 3.3777156225490472,
      "tokens_seen": 2416312320
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.713211651219973e-05,
      "loss": 2.3014,
      "theoretical_loss": 3.377708265236615,
      "tokens_seen": 2416377856
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.711556648382825e-05,
      "loss": 2.4885,
      "theoretical_loss": 3.3777009081795923,
      "tokens_seen": 2416443392
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.709901645545678e-05,
      "loss": 2.5927,
      "theoretical_loss": 3.377693551377963,
      "tokens_seen": 2416508928
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.70824664270853e-05,
      "loss": 2.3402,
      "theoretical_loss": 3.377686194831712,
      "tokens_seen": 2416574464
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.706591639871382e-05,
      "loss": 2.5645,
      "theoretical_loss": 3.377678838540822,
      "tokens_seen": 2416640000
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.704936637034235e-05,
      "loss": 2.6986,
      "theoretical_loss": 3.3776714825052787,
      "tokens_seen": 2416705536
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.703281634197087e-05,
      "loss": 2.604,
      "theoretical_loss": 3.3776641267250658,
      "tokens_seen": 2416771072
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.70162663135994e-05,
      "loss": 2.6193,
      "theoretical_loss": 3.3776567712001673,
      "tokens_seen": 2416836608
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.699971628522792e-05,
      "loss": 2.5473,
      "theoretical_loss": 3.3776494159305677,
      "tokens_seen": 2416902144
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.698316625685644e-05,
      "loss": 2.571,
      "theoretical_loss": 3.377642060916251,
      "tokens_seen": 2416967680
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.696661622848497e-05,
      "loss": 2.5836,
      "theoretical_loss": 3.3776347061572016,
      "tokens_seen": 2417033216
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.695006620011348e-05,
      "loss": 2.5664,
      "theoretical_loss": 3.3776273516534037,
      "tokens_seen": 2417098752
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.6933516171742e-05,
      "loss": 2.5056,
      "theoretical_loss": 3.377619997404841,
      "tokens_seen": 2417164288
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.691696614337053e-05,
      "loss": 2.7269,
      "theoretical_loss": 3.377612643411499,
      "tokens_seen": 2417229824
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.690041611499905e-05,
      "loss": 2.7266,
      "theoretical_loss": 3.3776052896733604,
      "tokens_seen": 2417295360
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.688386608662757e-05,
      "loss": 2.7664,
      "theoretical_loss": 3.3775979361904103,
      "tokens_seen": 2417360896
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.68673160582561e-05,
      "loss": 2.4212,
      "theoretical_loss": 3.377590582962633,
      "tokens_seen": 2417426432
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.685076602988462e-05,
      "loss": 2.4823,
      "theoretical_loss": 3.3775832299900124,
      "tokens_seen": 2417491968
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.683421600151315e-05,
      "loss": 2.4445,
      "theoretical_loss": 3.3775758772725326,
      "tokens_seen": 2417557504
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.681766597314167e-05,
      "loss": 2.6572,
      "theoretical_loss": 3.377568524810178,
      "tokens_seen": 2417623040
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.68011159447702e-05,
      "loss": 2.4529,
      "theoretical_loss": 3.377561172602933,
      "tokens_seen": 2417688576
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.678456591639872e-05,
      "loss": 2.5909,
      "theoretical_loss": 3.377553820650782,
      "tokens_seen": 2417754112
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.676801588802724e-05,
      "loss": 2.5043,
      "theoretical_loss": 3.3775464689537085,
      "tokens_seen": 2417819648
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.675146585965577e-05,
      "loss": 2.4904,
      "theoretical_loss": 3.3775391175116973,
      "tokens_seen": 2417885184
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.673491583128429e-05,
      "loss": 2.6597,
      "theoretical_loss": 3.3775317663247324,
      "tokens_seen": 2417950720
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.671836580291281e-05,
      "loss": 2.3628,
      "theoretical_loss": 3.3775244153927986,
      "tokens_seen": 2418016256
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.670181577454132e-05,
      "loss": 2.5334,
      "theoretical_loss": 3.3775170647158794,
      "tokens_seen": 2418081792
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.668526574616985e-05,
      "loss": 2.5378,
      "theoretical_loss": 3.3775097142939594,
      "tokens_seen": 2418147328
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.666871571779837e-05,
      "loss": 2.661,
      "theoretical_loss": 3.3775023641270225,
      "tokens_seen": 2418212864
    },
    {
      "epoch": 0.86,
      "objective/train/avg_token_score": 0.0057706646621227264,
      "objective/train/avg_weight": 0.9932088255882263,
      "objective/train/docs_used": 1359820,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4630846977233887,
      "objective/train/std_weight": 0.026669219136238098,
      "objective/train/theoretical_loss": 3.3774950142150537,
      "objective/train/tokens_used": 2438738400,
      "theoretical_loss": 3.3774950142150537,
      "tokens_seen": 2418278400
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.66521656894269e-05,
      "loss": 2.116,
      "theoretical_loss": 3.3774950142150537,
      "tokens_seen": 2418278400
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.663561566105542e-05,
      "loss": 2.7108,
      "theoretical_loss": 3.3774876645580365,
      "tokens_seen": 2418343936
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.661906563268394e-05,
      "loss": 2.4972,
      "theoretical_loss": 3.3774803151559554,
      "tokens_seen": 2418409472
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.660251560431247e-05,
      "loss": 2.538,
      "theoretical_loss": 3.3774729660087948,
      "tokens_seen": 2418475008
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.658596557594099e-05,
      "loss": 2.3615,
      "theoretical_loss": 3.3774656171165387,
      "tokens_seen": 2418540544
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.656941554756952e-05,
      "loss": 2.7618,
      "theoretical_loss": 3.3774582684791716,
      "tokens_seen": 2418606080
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.655286551919804e-05,
      "loss": 2.7746,
      "theoretical_loss": 3.3774509200966776,
      "tokens_seen": 2418671616
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.653631549082656e-05,
      "loss": 2.6191,
      "theoretical_loss": 3.377443571969041,
      "tokens_seen": 2418737152
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.651976546245509e-05,
      "loss": 2.6451,
      "theoretical_loss": 3.377436224096246,
      "tokens_seen": 2418802688
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.650321543408361e-05,
      "loss": 2.4339,
      "theoretical_loss": 3.3774288764782767,
      "tokens_seen": 2418868224
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.648666540571214e-05,
      "loss": 2.6366,
      "theoretical_loss": 3.377421529115118,
      "tokens_seen": 2418933760
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.647011537734063e-05,
      "loss": 2.5624,
      "theoretical_loss": 3.377414182006753,
      "tokens_seen": 2418999296
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.645356534896916e-05,
      "loss": 2.5141,
      "theoretical_loss": 3.377406835153167,
      "tokens_seen": 2419064832
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.643701532059768e-05,
      "loss": 2.6548,
      "theoretical_loss": 3.3773994885543446,
      "tokens_seen": 2419130368
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.64204652922262e-05,
      "loss": 2.4887,
      "theoretical_loss": 3.377392142210269,
      "tokens_seen": 2419195904
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.640391526385473e-05,
      "loss": 2.7149,
      "theoretical_loss": 3.3773847961209245,
      "tokens_seen": 2419261440
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.638736523548325e-05,
      "loss": 2.6435,
      "theoretical_loss": 3.377377450286296,
      "tokens_seen": 2419326976
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.637081520711178e-05,
      "loss": 2.3773,
      "theoretical_loss": 3.377370104706367,
      "tokens_seen": 2419392512
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.63542651787403e-05,
      "loss": 2.2528,
      "theoretical_loss": 3.377362759381123,
      "tokens_seen": 2419458048
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.633771515036882e-05,
      "loss": 2.3677,
      "theoretical_loss": 3.3773554143105473,
      "tokens_seen": 2419523584
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.632116512199735e-05,
      "loss": 2.9123,
      "theoretical_loss": 3.3773480694946243,
      "tokens_seen": 2419589120
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.630461509362587e-05,
      "loss": 2.3039,
      "theoretical_loss": 3.3773407249333385,
      "tokens_seen": 2419654656
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.62880650652544e-05,
      "loss": 2.7504,
      "theoretical_loss": 3.3773333806266743,
      "tokens_seen": 2419720192
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.627151503688292e-05,
      "loss": 2.4545,
      "theoretical_loss": 3.3773260365746154,
      "tokens_seen": 2419785728
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.625496500851144e-05,
      "loss": 2.4116,
      "theoretical_loss": 3.3773186927771466,
      "tokens_seen": 2419851264
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.623841498013997e-05,
      "loss": 2.5675,
      "theoretical_loss": 3.377311349234252,
      "tokens_seen": 2419916800
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.622186495176848e-05,
      "loss": 2.5635,
      "theoretical_loss": 3.3773040059459154,
      "tokens_seen": 2419982336
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.6205314923397e-05,
      "loss": 2.3029,
      "theoretical_loss": 3.377296662912122,
      "tokens_seen": 2420047872
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.618876489502553e-05,
      "loss": 2.7336,
      "theoretical_loss": 3.3772893201328555,
      "tokens_seen": 2420113408
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.617221486665405e-05,
      "loss": 2.5986,
      "theoretical_loss": 3.3772819776081007,
      "tokens_seen": 2420178944
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.615566483828257e-05,
      "loss": 2.7074,
      "theoretical_loss": 3.3772746353378413,
      "tokens_seen": 2420244480
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.61391148099111e-05,
      "loss": 2.6815,
      "theoretical_loss": 3.377267293322062,
      "tokens_seen": 2420310016
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.612256478153962e-05,
      "loss": 2.6156,
      "theoretical_loss": 3.3772599515607467,
      "tokens_seen": 2420375552
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.610601475316815e-05,
      "loss": 2.3131,
      "theoretical_loss": 3.3772526100538798,
      "tokens_seen": 2420441088
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.608946472479667e-05,
      "loss": 2.5396,
      "theoretical_loss": 3.3772452688014454,
      "tokens_seen": 2420506624
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.60729146964252e-05,
      "loss": 2.0351,
      "theoretical_loss": 3.377237927803429,
      "tokens_seen": 2420572160
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.605636466805372e-05,
      "loss": 2.6751,
      "theoretical_loss": 3.377230587059813,
      "tokens_seen": 2420637696
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.603981463968224e-05,
      "loss": 2.7192,
      "theoretical_loss": 3.3772232465705834,
      "tokens_seen": 2420703232
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.602326461131077e-05,
      "loss": 2.568,
      "theoretical_loss": 3.3772159063357234,
      "tokens_seen": 2420768768
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.600671458293929e-05,
      "loss": 2.4168,
      "theoretical_loss": 3.3772085663552174,
      "tokens_seen": 2420834304
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.599016455456781e-05,
      "loss": 2.4544,
      "theoretical_loss": 3.37720122662905,
      "tokens_seen": 2420899840
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.597361452619633e-05,
      "loss": 2.4938,
      "theoretical_loss": 3.377193887157206,
      "tokens_seen": 2420965376
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.595706449782485e-05,
      "loss": 2.4644,
      "theoretical_loss": 3.3771865479396688,
      "tokens_seen": 2421030912
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.594051446945337e-05,
      "loss": 2.7424,
      "theoretical_loss": 3.3771792089764237,
      "tokens_seen": 2421096448
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.59239644410819e-05,
      "loss": 2.5764,
      "theoretical_loss": 3.3771718702674534,
      "tokens_seen": 2421161984
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.590741441271042e-05,
      "loss": 2.5604,
      "theoretical_loss": 3.3771645318127437,
      "tokens_seen": 2421227520
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.589086438433895e-05,
      "loss": 2.6491,
      "theoretical_loss": 3.3771571936122786,
      "tokens_seen": 2421293056
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.587431435596747e-05,
      "loss": 2.6118,
      "theoretical_loss": 3.3771498556660418,
      "tokens_seen": 2421358592
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.5857764327596e-05,
      "loss": 2.4926,
      "theoretical_loss": 3.3771425179740184,
      "tokens_seen": 2421424128
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.584121429922452e-05,
      "loss": 2.2841,
      "theoretical_loss": 3.3771351805361918,
      "tokens_seen": 2421489664
    },
    {
      "epoch": 0.86,
      "objective/train/avg_token_score": 0.007263542152941227,
      "objective/train/avg_weight": 0.9914515018463135,
      "objective/train/docs_used": 1361438,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.549990177154541,
      "objective/train/std_weight": 0.05167626589536667,
      "objective/train/theoretical_loss": 3.3771278433525476,
      "objective/train/tokens_used": 2442015200,
      "theoretical_loss": 3.3771278433525476,
      "tokens_seen": 2421555200
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.582466427085304e-05,
      "loss": 2.4595,
      "theoretical_loss": 3.3771278433525476,
      "tokens_seen": 2421555200
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.580811424248157e-05,
      "loss": 2.381,
      "theoretical_loss": 3.3771205064230685,
      "tokens_seen": 2421620736
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.579156421411009e-05,
      "loss": 2.7732,
      "theoretical_loss": 3.3771131697477403,
      "tokens_seen": 2421686272
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.577501418573861e-05,
      "loss": 2.6108,
      "theoretical_loss": 3.3771058333265467,
      "tokens_seen": 2421751808
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.575846415736714e-05,
      "loss": 2.5914,
      "theoretical_loss": 3.3770984971594715,
      "tokens_seen": 2421817344
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.574191412899566e-05,
      "loss": 2.4451,
      "theoretical_loss": 3.3770911612464998,
      "tokens_seen": 2421882880
    },
    {
      "epoch": 0.86,
      "learning_rate": 9.572536410062416e-05,
      "loss": 2.4801,
      "theoretical_loss": 3.377083825587616,
      "tokens_seen": 2421948416
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.570881407225268e-05,
      "loss": 2.4907,
      "theoretical_loss": 3.377076490182804,
      "tokens_seen": 2422013952
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.56922640438812e-05,
      "loss": 2.2964,
      "theoretical_loss": 3.3770691550320477,
      "tokens_seen": 2422079488
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.567571401550973e-05,
      "loss": 2.467,
      "theoretical_loss": 3.3770618201353324,
      "tokens_seen": 2422145024
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.565916398713825e-05,
      "loss": 2.5741,
      "theoretical_loss": 3.377054485492642,
      "tokens_seen": 2422210560
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.564261395876678e-05,
      "loss": 2.3627,
      "theoretical_loss": 3.3770471511039606,
      "tokens_seen": 2422276096
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.56260639303953e-05,
      "loss": 2.693,
      "theoretical_loss": 3.3770398169692726,
      "tokens_seen": 2422341632
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.560951390202383e-05,
      "loss": 2.7642,
      "theoretical_loss": 3.3770324830885627,
      "tokens_seen": 2422407168
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.559296387365235e-05,
      "loss": 2.4844,
      "theoretical_loss": 3.377025149461815,
      "tokens_seen": 2422472704
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.557641384528087e-05,
      "loss": 2.7899,
      "theoretical_loss": 3.3770178160890136,
      "tokens_seen": 2422538240
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.55598638169094e-05,
      "loss": 2.6789,
      "theoretical_loss": 3.3770104829701437,
      "tokens_seen": 2422603776
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.554331378853792e-05,
      "loss": 2.6899,
      "theoretical_loss": 3.3770031501051885,
      "tokens_seen": 2422669312
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.552676376016645e-05,
      "loss": 2.3533,
      "theoretical_loss": 3.376995817494133,
      "tokens_seen": 2422734848
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.551021373179497e-05,
      "loss": 2.4268,
      "theoretical_loss": 3.376988485136961,
      "tokens_seen": 2422800384
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.54936637034235e-05,
      "loss": 2.5631,
      "theoretical_loss": 3.3769811530336575,
      "tokens_seen": 2422865920
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.5477113675052e-05,
      "loss": 2.7622,
      "theoretical_loss": 3.376973821184207,
      "tokens_seen": 2422931456
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.546056364668053e-05,
      "loss": 2.58,
      "theoretical_loss": 3.3769664895885927,
      "tokens_seen": 2422996992
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.544401361830905e-05,
      "loss": 2.6669,
      "theoretical_loss": 3.3769591582468004,
      "tokens_seen": 2423062528
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.542746358993758e-05,
      "loss": 2.5282,
      "theoretical_loss": 3.376951827158813,
      "tokens_seen": 2423128064
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.54109135615661e-05,
      "loss": 2.6681,
      "theoretical_loss": 3.376944496324616,
      "tokens_seen": 2423193600
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.539436353319462e-05,
      "loss": 3.2586,
      "theoretical_loss": 3.3769371657441933,
      "tokens_seen": 2423259136
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.537781350482315e-05,
      "loss": 2.8243,
      "theoretical_loss": 3.376929835417529,
      "tokens_seen": 2423324672
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.536126347645167e-05,
      "loss": 2.7109,
      "theoretical_loss": 3.376922505344608,
      "tokens_seen": 2423390208
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.53447134480802e-05,
      "loss": 2.6718,
      "theoretical_loss": 3.376915175525414,
      "tokens_seen": 2423455744
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.532816341970872e-05,
      "loss": 2.5182,
      "theoretical_loss": 3.3769078459599324,
      "tokens_seen": 2423521280
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.531161339133724e-05,
      "loss": 2.412,
      "theoretical_loss": 3.3769005166481465,
      "tokens_seen": 2423586816
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.529506336296577e-05,
      "loss": 2.5132,
      "theoretical_loss": 3.376893187590041,
      "tokens_seen": 2423652352
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.527851333459429e-05,
      "loss": 2.5668,
      "theoretical_loss": 3.3768858587856005,
      "tokens_seen": 2423717888
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.526196330622282e-05,
      "loss": 2.7059,
      "theoretical_loss": 3.376878530234809,
      "tokens_seen": 2423783424
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.524541327785134e-05,
      "loss": 2.579,
      "theoretical_loss": 3.376871201937651,
      "tokens_seen": 2423848960
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.522886324947985e-05,
      "loss": 2.3111,
      "theoretical_loss": 3.376863873894111,
      "tokens_seen": 2423914496
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.521231322110837e-05,
      "loss": 2.2585,
      "theoretical_loss": 3.3768565461041735,
      "tokens_seen": 2423980032
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.51957631927369e-05,
      "loss": 2.3772,
      "theoretical_loss": 3.376849218567822,
      "tokens_seen": 2424045568
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.517921316436542e-05,
      "loss": 2.6714,
      "theoretical_loss": 3.376841891285042,
      "tokens_seen": 2424111104
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.516266313599395e-05,
      "loss": 2.6082,
      "theoretical_loss": 3.3768345642558173,
      "tokens_seen": 2424176640
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.514611310762247e-05,
      "loss": 2.3872,
      "theoretical_loss": 3.3768272374801325,
      "tokens_seen": 2424242176
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.5129563079251e-05,
      "loss": 2.4861,
      "theoretical_loss": 3.3768199109579715,
      "tokens_seen": 2424307712
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.511301305087952e-05,
      "loss": 2.7875,
      "theoretical_loss": 3.376812584689319,
      "tokens_seen": 2424373248
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.509646302250804e-05,
      "loss": 2.6096,
      "theoretical_loss": 3.3768052586741595,
      "tokens_seen": 2424438784
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.507991299413657e-05,
      "loss": 2.4459,
      "theoretical_loss": 3.3767979329124773,
      "tokens_seen": 2424504320
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.506336296576509e-05,
      "loss": 2.6097,
      "theoretical_loss": 3.3767906074042564,
      "tokens_seen": 2424569856
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.504681293739361e-05,
      "loss": 2.2458,
      "theoretical_loss": 3.376783282149482,
      "tokens_seen": 2424635392
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.503026290902214e-05,
      "loss": 2.6129,
      "theoretical_loss": 3.376775957148138,
      "tokens_seen": 2424700928
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.501371288065066e-05,
      "loss": 2.6259,
      "theoretical_loss": 3.376768632400208,
      "tokens_seen": 2424766464
    },
    {
      "epoch": 0.87,
      "objective/train/avg_token_score": 0.03175375983119011,
      "objective/train/avg_weight": 0.9626287221908569,
      "objective/train/docs_used": 1363257,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.8509960174560547,
      "objective/train/std_weight": 0.12791074812412262,
      "objective/train/theoretical_loss": 3.376761307905678,
      "objective/train/tokens_used": 2445292000,
      "theoretical_loss": 3.376761307905678,
      "tokens_seen": 2424832000
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.499716285227919e-05,
      "loss": 2.4697,
      "theoretical_loss": 3.376761307905678,
      "tokens_seen": 2424832000
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.498061282390768e-05,
      "loss": 2.7478,
      "theoretical_loss": 3.376753983664531,
      "tokens_seen": 2424897536
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.496406279553621e-05,
      "loss": 2.6541,
      "theoretical_loss": 3.376746659676752,
      "tokens_seen": 2424963072
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.494751276716473e-05,
      "loss": 2.5977,
      "theoretical_loss": 3.3767393359423257,
      "tokens_seen": 2425028608
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.493096273879326e-05,
      "loss": 2.4659,
      "theoretical_loss": 3.3767320124612357,
      "tokens_seen": 2425094144
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.491441271042178e-05,
      "loss": 2.6901,
      "theoretical_loss": 3.376724689233467,
      "tokens_seen": 2425159680
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.48978626820503e-05,
      "loss": 2.6065,
      "theoretical_loss": 3.3767173662590038,
      "tokens_seen": 2425225216
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.488131265367883e-05,
      "loss": 2.5273,
      "theoretical_loss": 3.3767100435378303,
      "tokens_seen": 2425290752
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.486476262530735e-05,
      "loss": 2.7322,
      "theoretical_loss": 3.3767027210699307,
      "tokens_seen": 2425356288
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.484821259693588e-05,
      "loss": 2.553,
      "theoretical_loss": 3.3766953988552904,
      "tokens_seen": 2425421824
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.48316625685644e-05,
      "loss": 2.5256,
      "theoretical_loss": 3.376688076893893,
      "tokens_seen": 2425487360
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.481511254019292e-05,
      "loss": 2.3898,
      "theoretical_loss": 3.376680755185723,
      "tokens_seen": 2425552896
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.479856251182145e-05,
      "loss": 2.4764,
      "theoretical_loss": 3.376673433730765,
      "tokens_seen": 2425618432
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.478201248344997e-05,
      "loss": 2.7988,
      "theoretical_loss": 3.376666112529003,
      "tokens_seen": 2425683968
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.47654624550785e-05,
      "loss": 2.7277,
      "theoretical_loss": 3.376658791580422,
      "tokens_seen": 2425749504
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.4748912426707e-05,
      "loss": 2.0144,
      "theoretical_loss": 3.376651470885006,
      "tokens_seen": 2425815040
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.473236239833553e-05,
      "loss": 2.5616,
      "theoretical_loss": 3.3766441504427394,
      "tokens_seen": 2425880576
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.471581236996405e-05,
      "loss": 2.3098,
      "theoretical_loss": 3.3766368302536067,
      "tokens_seen": 2425946112
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.469926234159258e-05,
      "loss": 2.6169,
      "theoretical_loss": 3.376629510317592,
      "tokens_seen": 2426011648
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.46827123132211e-05,
      "loss": 2.758,
      "theoretical_loss": 3.3766221906346807,
      "tokens_seen": 2426077184
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.466616228484963e-05,
      "loss": 2.5319,
      "theoretical_loss": 3.3766148712048563,
      "tokens_seen": 2426142720
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.464961225647815e-05,
      "loss": 2.5242,
      "theoretical_loss": 3.376607552028103,
      "tokens_seen": 2426208256
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.463306222810667e-05,
      "loss": 2.2282,
      "theoretical_loss": 3.376600233104406,
      "tokens_seen": 2426273792
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.46165121997352e-05,
      "loss": 2.5861,
      "theoretical_loss": 3.376592914433749,
      "tokens_seen": 2426339328
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.459996217136372e-05,
      "loss": 2.663,
      "theoretical_loss": 3.3765855960161173,
      "tokens_seen": 2426404864
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.458341214299225e-05,
      "loss": 2.7208,
      "theoretical_loss": 3.3765782778514946,
      "tokens_seen": 2426470400
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.456686211462077e-05,
      "loss": 2.8778,
      "theoretical_loss": 3.3765709599398654,
      "tokens_seen": 2426535936
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.45503120862493e-05,
      "loss": 2.7846,
      "theoretical_loss": 3.3765636422812144,
      "tokens_seen": 2426601472
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.453376205787782e-05,
      "loss": 2.61,
      "theoretical_loss": 3.376556324875526,
      "tokens_seen": 2426667008
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.451721202950634e-05,
      "loss": 2.3752,
      "theoretical_loss": 3.376549007722784,
      "tokens_seen": 2426732544
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.450066200113485e-05,
      "loss": 2.5264,
      "theoretical_loss": 3.3765416908229735,
      "tokens_seen": 2426798080
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.448411197276338e-05,
      "loss": 2.6573,
      "theoretical_loss": 3.376534374176079,
      "tokens_seen": 2426863616
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.44675619443919e-05,
      "loss": 2.3686,
      "theoretical_loss": 3.3765270577820847,
      "tokens_seen": 2426929152
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.445101191602042e-05,
      "loss": 2.4017,
      "theoretical_loss": 3.3765197416409745,
      "tokens_seen": 2426994688
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.443446188764895e-05,
      "loss": 2.5038,
      "theoretical_loss": 3.376512425752734,
      "tokens_seen": 2427060224
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.441791185927747e-05,
      "loss": 2.4718,
      "theoretical_loss": 3.376505110117346,
      "tokens_seen": 2427125760
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.4401361830906e-05,
      "loss": 2.7867,
      "theoretical_loss": 3.3764977947347967,
      "tokens_seen": 2427191296
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.438481180253452e-05,
      "loss": 2.5495,
      "theoretical_loss": 3.37649047960507,
      "tokens_seen": 2427256832
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.436826177416304e-05,
      "loss": 2.4572,
      "theoretical_loss": 3.3764831647281492,
      "tokens_seen": 2427322368
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.435171174579157e-05,
      "loss": 2.3766,
      "theoretical_loss": 3.37647585010402,
      "tokens_seen": 2427387904
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.433516171742009e-05,
      "loss": 2.5,
      "theoretical_loss": 3.3764685357326667,
      "tokens_seen": 2427453440
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.431861168904862e-05,
      "loss": 2.4541,
      "theoretical_loss": 3.3764612216140732,
      "tokens_seen": 2427518976
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.430206166067714e-05,
      "loss": 2.861,
      "theoretical_loss": 3.376453907748224,
      "tokens_seen": 2427584512
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.428551163230566e-05,
      "loss": 2.6101,
      "theoretical_loss": 3.3764465941351043,
      "tokens_seen": 2427650048
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.426896160393419e-05,
      "loss": 2.3875,
      "theoretical_loss": 3.3764392807746977,
      "tokens_seen": 2427715584
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.425241157556268e-05,
      "loss": 2.2846,
      "theoretical_loss": 3.376431967666989,
      "tokens_seen": 2427781120
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.423586154719121e-05,
      "loss": 2.7128,
      "theoretical_loss": 3.3764246548119625,
      "tokens_seen": 2427846656
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.421931151881973e-05,
      "loss": 2.3921,
      "theoretical_loss": 3.376417342209603,
      "tokens_seen": 2427912192
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.420276149044826e-05,
      "loss": 2.5725,
      "theoretical_loss": 3.3764100298598945,
      "tokens_seen": 2427977728
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.418621146207678e-05,
      "loss": 2.5544,
      "theoretical_loss": 3.3764027177628217,
      "tokens_seen": 2428043264
    },
    {
      "epoch": 0.87,
      "objective/train/avg_token_score": 0.008435784839093685,
      "objective/train/avg_weight": 0.9900768399238586,
      "objective/train/docs_used": 1365224,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5721864700317383,
      "objective/train/std_weight": 0.0378696583211422,
      "objective/train/theoretical_loss": 3.376395405918369,
      "objective/train/tokens_used": 2448568800,
      "theoretical_loss": 3.376395405918369,
      "tokens_seen": 2428108800
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.41696614337053e-05,
      "loss": 2.4495,
      "theoretical_loss": 3.376395405918369,
      "tokens_seen": 2428108800
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.415311140533383e-05,
      "loss": 2.5414,
      "theoretical_loss": 3.376388094326521,
      "tokens_seen": 2428174336
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.413656137696235e-05,
      "loss": 2.5747,
      "theoretical_loss": 3.376380782987262,
      "tokens_seen": 2428239872
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.412001134859089e-05,
      "loss": 2.5385,
      "theoretical_loss": 3.3763734719005765,
      "tokens_seen": 2428305408
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.410346132021941e-05,
      "loss": 2.4862,
      "theoretical_loss": 3.376366161066449,
      "tokens_seen": 2428370944
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.408691129184794e-05,
      "loss": 2.8222,
      "theoretical_loss": 3.3763588504848636,
      "tokens_seen": 2428436480
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.407036126347646e-05,
      "loss": 2.3135,
      "theoretical_loss": 3.376351540155805,
      "tokens_seen": 2428502016
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.405381123510499e-05,
      "loss": 2.101,
      "theoretical_loss": 3.3763442300792583,
      "tokens_seen": 2428567552
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.403726120673351e-05,
      "loss": 2.5612,
      "theoretical_loss": 3.376336920255207,
      "tokens_seen": 2428633088
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.402071117836203e-05,
      "loss": 2.524,
      "theoretical_loss": 3.376329610683636,
      "tokens_seen": 2428698624
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.400416114999053e-05,
      "loss": 2.3145,
      "theoretical_loss": 3.37632230136453,
      "tokens_seen": 2428764160
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.398761112161906e-05,
      "loss": 2.5162,
      "theoretical_loss": 3.3763149922978726,
      "tokens_seen": 2428829696
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.397106109324758e-05,
      "loss": 2.7613,
      "theoretical_loss": 3.376307683483649,
      "tokens_seen": 2428895232
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.39545110648761e-05,
      "loss": 2.4826,
      "theoretical_loss": 3.376300374921844,
      "tokens_seen": 2428960768
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.393796103650463e-05,
      "loss": 2.4716,
      "theoretical_loss": 3.376293066612441,
      "tokens_seen": 2429026304
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.392141100813315e-05,
      "loss": 2.4699,
      "theoretical_loss": 3.3762857585554253,
      "tokens_seen": 2429091840
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.390486097976168e-05,
      "loss": 2.7432,
      "theoretical_loss": 3.3762784507507813,
      "tokens_seen": 2429157376
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.38883109513902e-05,
      "loss": 2.4829,
      "theoretical_loss": 3.376271143198493,
      "tokens_seen": 2429222912
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.387176092301872e-05,
      "loss": 2.6718,
      "theoretical_loss": 3.3762638358985457,
      "tokens_seen": 2429288448
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.385521089464725e-05,
      "loss": 2.5621,
      "theoretical_loss": 3.376256528850923,
      "tokens_seen": 2429353984
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.383866086627577e-05,
      "loss": 2.4775,
      "theoretical_loss": 3.37624922205561,
      "tokens_seen": 2429419520
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.38221108379043e-05,
      "loss": 2.4888,
      "theoretical_loss": 3.3762419155125905,
      "tokens_seen": 2429485056
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.380556080953282e-05,
      "loss": 2.4698,
      "theoretical_loss": 3.37623460922185,
      "tokens_seen": 2429550592
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.378901078116134e-05,
      "loss": 2.6739,
      "theoretical_loss": 3.3762273031833723,
      "tokens_seen": 2429616128
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.377246075278987e-05,
      "loss": 2.6782,
      "theoretical_loss": 3.3762199973971416,
      "tokens_seen": 2429681664
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.375591072441838e-05,
      "loss": 2.6508,
      "theoretical_loss": 3.3762126918631434,
      "tokens_seen": 2429747200
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.37393606960469e-05,
      "loss": 2.2971,
      "theoretical_loss": 3.376205386581361,
      "tokens_seen": 2429812736
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.372281066767543e-05,
      "loss": 2.6393,
      "theoretical_loss": 3.3761980815517796,
      "tokens_seen": 2429878272
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.370626063930395e-05,
      "loss": 2.5817,
      "theoretical_loss": 3.3761907767743837,
      "tokens_seen": 2429943808
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.368971061093247e-05,
      "loss": 2.333,
      "theoretical_loss": 3.3761834722491577,
      "tokens_seen": 2430009344
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.3673160582561e-05,
      "loss": 2.7914,
      "theoretical_loss": 3.376176167976086,
      "tokens_seen": 2430074880
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.365661055418952e-05,
      "loss": 2.564,
      "theoretical_loss": 3.376168863955153,
      "tokens_seen": 2430140416
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.364006052581805e-05,
      "loss": 2.5927,
      "theoretical_loss": 3.3761615601863433,
      "tokens_seen": 2430205952
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.362351049744657e-05,
      "loss": 2.5112,
      "theoretical_loss": 3.3761542566696416,
      "tokens_seen": 2430271488
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.36069604690751e-05,
      "loss": 2.6799,
      "theoretical_loss": 3.376146953405032,
      "tokens_seen": 2430337024
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.359041044070362e-05,
      "loss": 2.3858,
      "theoretical_loss": 3.3761396503924996,
      "tokens_seen": 2430402560
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.357386041233214e-05,
      "loss": 2.5766,
      "theoretical_loss": 3.376132347632028,
      "tokens_seen": 2430468096
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.355731038396067e-05,
      "loss": 2.5646,
      "theoretical_loss": 3.3761250451236027,
      "tokens_seen": 2430533632
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.354076035558919e-05,
      "loss": 2.4785,
      "theoretical_loss": 3.376117742867208,
      "tokens_seen": 2430599168
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.352421032721771e-05,
      "loss": 2.6435,
      "theoretical_loss": 3.3761104408628277,
      "tokens_seen": 2430664704
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.350766029884621e-05,
      "loss": 2.7526,
      "theoretical_loss": 3.376103139110447,
      "tokens_seen": 2430730240
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.349111027047475e-05,
      "loss": 2.326,
      "theoretical_loss": 3.37609583761005,
      "tokens_seen": 2430795776
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.347456024210327e-05,
      "loss": 2.6992,
      "theoretical_loss": 3.376088536361621,
      "tokens_seen": 2430861312
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.34580102137318e-05,
      "loss": 2.4842,
      "theoretical_loss": 3.3760812353651457,
      "tokens_seen": 2430926848
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.344146018536032e-05,
      "loss": 2.6165,
      "theoretical_loss": 3.3760739346206075,
      "tokens_seen": 2430992384
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.342491015698884e-05,
      "loss": 2.6143,
      "theoretical_loss": 3.376066634127991,
      "tokens_seen": 2431057920
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.340836012861737e-05,
      "loss": 2.6045,
      "theoretical_loss": 3.376059333887281,
      "tokens_seen": 2431123456
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.339181010024589e-05,
      "loss": 2.4922,
      "theoretical_loss": 3.3760520338984623,
      "tokens_seen": 2431188992
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.337526007187442e-05,
      "loss": 2.8797,
      "theoretical_loss": 3.3760447341615185,
      "tokens_seen": 2431254528
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.335871004350294e-05,
      "loss": 2.4426,
      "theoretical_loss": 3.376037434676435,
      "tokens_seen": 2431320064
    },
    {
      "epoch": 0.87,
      "objective/train/avg_token_score": 0.016843337565660477,
      "objective/train/avg_weight": 0.980171263217926,
      "objective/train/docs_used": 1366375,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.9875917434692383,
      "objective/train/std_weight": 0.10577838122844696,
      "objective/train/theoretical_loss": 3.3760301354431963,
      "objective/train/tokens_used": 2451845600,
      "theoretical_loss": 3.3760301354431963,
      "tokens_seen": 2431385600
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.334216001513146e-05,
      "loss": 2.6056,
      "theoretical_loss": 3.3760301354431963,
      "tokens_seen": 2431385600
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.332560998675999e-05,
      "loss": 2.553,
      "theoretical_loss": 3.376022836461786,
      "tokens_seen": 2431451136
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.330905995838851e-05,
      "loss": 2.4877,
      "theoretical_loss": 3.37601553773219,
      "tokens_seen": 2431516672
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.329250993001704e-05,
      "loss": 2.6359,
      "theoretical_loss": 3.3760082392543915,
      "tokens_seen": 2431582208
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.327595990164553e-05,
      "loss": 2.4749,
      "theoretical_loss": 3.3760009410283756,
      "tokens_seen": 2431647744
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.325940987327406e-05,
      "loss": 2.6812,
      "theoretical_loss": 3.3759936430541275,
      "tokens_seen": 2431713280
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.324285984490258e-05,
      "loss": 2.5611,
      "theoretical_loss": 3.3759863453316306,
      "tokens_seen": 2431778816
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.32263098165311e-05,
      "loss": 2.6398,
      "theoretical_loss": 3.37597904786087,
      "tokens_seen": 2431844352
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.320975978815963e-05,
      "loss": 2.4796,
      "theoretical_loss": 3.37597175064183,
      "tokens_seen": 2431909888
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.319320975978815e-05,
      "loss": 2.5958,
      "theoretical_loss": 3.375964453674495,
      "tokens_seen": 2431975424
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.317665973141668e-05,
      "loss": 2.4637,
      "theoretical_loss": 3.3759571569588505,
      "tokens_seen": 2432040960
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.31601097030452e-05,
      "loss": 2.4578,
      "theoretical_loss": 3.37594986049488,
      "tokens_seen": 2432106496
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.314355967467372e-05,
      "loss": 2.0829,
      "theoretical_loss": 3.3759425642825684,
      "tokens_seen": 2432172032
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.312700964630225e-05,
      "loss": 2.5345,
      "theoretical_loss": 3.3759352683219004,
      "tokens_seen": 2432237568
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.311045961793077e-05,
      "loss": 2.537,
      "theoretical_loss": 3.37592797261286,
      "tokens_seen": 2432303104
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.30939095895593e-05,
      "loss": 2.6686,
      "theoretical_loss": 3.3759206771554324,
      "tokens_seen": 2432368640
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.307735956118782e-05,
      "loss": 2.2739,
      "theoretical_loss": 3.3759133819496014,
      "tokens_seen": 2432434176
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.306080953281634e-05,
      "loss": 2.301,
      "theoretical_loss": 3.3759060869953528,
      "tokens_seen": 2432499712
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.304425950444487e-05,
      "loss": 2.4529,
      "theoretical_loss": 3.3758987922926695,
      "tokens_seen": 2432565248
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.302770947607338e-05,
      "loss": 2.7861,
      "theoretical_loss": 3.3758914978415375,
      "tokens_seen": 2432630784
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.30111594477019e-05,
      "loss": 2.7412,
      "theoretical_loss": 3.3758842036419408,
      "tokens_seen": 2432696320
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.299460941933043e-05,
      "loss": 2.5464,
      "theoretical_loss": 3.3758769096938632,
      "tokens_seen": 2432761856
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.297805939095895e-05,
      "loss": 2.7183,
      "theoretical_loss": 3.3758696159972903,
      "tokens_seen": 2432827392
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.296150936258747e-05,
      "loss": 2.1557,
      "theoretical_loss": 3.3758623225522064,
      "tokens_seen": 2432892928
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.2944959334216e-05,
      "loss": 2.4167,
      "theoretical_loss": 3.375855029358596,
      "tokens_seen": 2432958464
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.292840930584452e-05,
      "loss": 1.8297,
      "theoretical_loss": 3.3758477364164436,
      "tokens_seen": 2433024000
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.291185927747305e-05,
      "loss": 2.4584,
      "theoretical_loss": 3.3758404437257337,
      "tokens_seen": 2433089536
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.289530924910157e-05,
      "loss": 2.5192,
      "theoretical_loss": 3.3758331512864506,
      "tokens_seen": 2433155072
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.28787592207301e-05,
      "loss": 2.4723,
      "theoretical_loss": 3.3758258590985797,
      "tokens_seen": 2433220608
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.286220919235862e-05,
      "loss": 2.3961,
      "theoretical_loss": 3.3758185671621046,
      "tokens_seen": 2433286144
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.284565916398714e-05,
      "loss": 2.6116,
      "theoretical_loss": 3.3758112754770107,
      "tokens_seen": 2433351680
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.282910913561567e-05,
      "loss": 2.5383,
      "theoretical_loss": 3.375803984043282,
      "tokens_seen": 2433417216
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.281255910724419e-05,
      "loss": 2.6131,
      "theoretical_loss": 3.375796692860903,
      "tokens_seen": 2433482752
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.279600907887271e-05,
      "loss": 2.5928,
      "theoretical_loss": 3.3757894019298584,
      "tokens_seen": 2433548288
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.277945905050123e-05,
      "loss": 3.0576,
      "theoretical_loss": 3.375782111250133,
      "tokens_seen": 2433613824
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.276290902212975e-05,
      "loss": 2.8674,
      "theoretical_loss": 3.3757748208217118,
      "tokens_seen": 2433679360
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.274635899375827e-05,
      "loss": 2.6175,
      "theoretical_loss": 3.3757675306445782,
      "tokens_seen": 2433744896
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.27298089653868e-05,
      "loss": 2.4033,
      "theoretical_loss": 3.3757602407187175,
      "tokens_seen": 2433810432
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.271325893701532e-05,
      "loss": 2.6084,
      "theoretical_loss": 3.3757529510441144,
      "tokens_seen": 2433875968
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.269670890864385e-05,
      "loss": 2.7562,
      "theoretical_loss": 3.375745661620753,
      "tokens_seen": 2433941504
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.268015888027237e-05,
      "loss": 2.7391,
      "theoretical_loss": 3.375738372448618,
      "tokens_seen": 2434007040
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.266360885190089e-05,
      "loss": 2.3963,
      "theoretical_loss": 3.375731083527694,
      "tokens_seen": 2434072576
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.264705882352942e-05,
      "loss": 2.6954,
      "theoretical_loss": 3.375723794857966,
      "tokens_seen": 2434138112
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.263050879515794e-05,
      "loss": 2.7112,
      "theoretical_loss": 3.375716506439418,
      "tokens_seen": 2434203648
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.261395876678647e-05,
      "loss": 2.4558,
      "theoretical_loss": 3.375709218272035,
      "tokens_seen": 2434269184
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.259740873841499e-05,
      "loss": 1.8686,
      "theoretical_loss": 3.3757019303558016,
      "tokens_seen": 2434334720
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.258085871004351e-05,
      "loss": 2.5791,
      "theoretical_loss": 3.3756946426907017,
      "tokens_seen": 2434400256
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.256430868167204e-05,
      "loss": 2.7164,
      "theoretical_loss": 3.375687355276721,
      "tokens_seen": 2434465792
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.254775865330056e-05,
      "loss": 2.4279,
      "theoretical_loss": 3.375680068113843,
      "tokens_seen": 2434531328
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.253120862492906e-05,
      "loss": 2.6441,
      "theoretical_loss": 3.375672781202053,
      "tokens_seen": 2434596864
    },
    {
      "epoch": 0.87,
      "objective/train/avg_token_score": 0.014580557122826576,
      "objective/train/avg_weight": 0.9828335642814636,
      "objective/train/docs_used": 1367967,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.415879249572754,
      "objective/train/std_weight": 0.09336385875940323,
      "objective/train/theoretical_loss": 3.375665494541335,
      "objective/train/tokens_used": 2455122400,
      "theoretical_loss": 3.375665494541335,
      "tokens_seen": 2434662400
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.251465859655758e-05,
      "loss": 2.4276,
      "theoretical_loss": 3.375665494541335,
      "tokens_seen": 2434662400
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.24981085681861e-05,
      "loss": 2.4439,
      "theoretical_loss": 3.375658208131674,
      "tokens_seen": 2434727936
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.248155853981463e-05,
      "loss": 2.7385,
      "theoretical_loss": 3.3756509219730546,
      "tokens_seen": 2434793472
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.246500851144315e-05,
      "loss": 2.4615,
      "theoretical_loss": 3.3756436360654614,
      "tokens_seen": 2434859008
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.244845848307168e-05,
      "loss": 2.5582,
      "theoretical_loss": 3.375636350408879,
      "tokens_seen": 2434924544
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.24319084547002e-05,
      "loss": 2.5645,
      "theoretical_loss": 3.3756290650032916,
      "tokens_seen": 2434990080
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.241535842632873e-05,
      "loss": 2.2957,
      "theoretical_loss": 3.3756217798486845,
      "tokens_seen": 2435055616
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.239880839795725e-05,
      "loss": 2.4928,
      "theoretical_loss": 3.3756144949450415,
      "tokens_seen": 2435121152
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.238225836958577e-05,
      "loss": 2.3562,
      "theoretical_loss": 3.375607210292348,
      "tokens_seen": 2435186688
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.23657083412143e-05,
      "loss": 2.6364,
      "theoretical_loss": 3.375599925890588,
      "tokens_seen": 2435252224
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.234915831284282e-05,
      "loss": 2.6506,
      "theoretical_loss": 3.3755926417397464,
      "tokens_seen": 2435317760
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.233260828447135e-05,
      "loss": 2.4726,
      "theoretical_loss": 3.3755853578398076,
      "tokens_seen": 2435383296
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.231605825609987e-05,
      "loss": 2.3759,
      "theoretical_loss": 3.375578074190756,
      "tokens_seen": 2435448832
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.22995082277284e-05,
      "loss": 1.9186,
      "theoretical_loss": 3.375570790792577,
      "tokens_seen": 2435514368
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.22829581993569e-05,
      "loss": 2.5864,
      "theoretical_loss": 3.375563507645255,
      "tokens_seen": 2435579904
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.226640817098543e-05,
      "loss": 2.5316,
      "theoretical_loss": 3.375556224748774,
      "tokens_seen": 2435645440
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.224985814261395e-05,
      "loss": 2.6572,
      "theoretical_loss": 3.375548942103119,
      "tokens_seen": 2435710976
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.223330811424248e-05,
      "loss": 2.4885,
      "theoretical_loss": 3.3755416597082744,
      "tokens_seen": 2435776512
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.2216758085871e-05,
      "loss": 2.5278,
      "theoretical_loss": 3.3755343775642253,
      "tokens_seen": 2435842048
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.220020805749952e-05,
      "loss": 2.4031,
      "theoretical_loss": 3.3755270956709555,
      "tokens_seen": 2435907584
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.218365802912805e-05,
      "loss": 2.3684,
      "theoretical_loss": 3.3755198140284506,
      "tokens_seen": 2435973120
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.216710800075657e-05,
      "loss": 2.3918,
      "theoretical_loss": 3.3755125326366944,
      "tokens_seen": 2436038656
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.21505579723851e-05,
      "loss": 2.4812,
      "theoretical_loss": 3.3755052514956723,
      "tokens_seen": 2436104192
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.213400794401362e-05,
      "loss": 2.4896,
      "theoretical_loss": 3.3754979706053683,
      "tokens_seen": 2436169728
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.211745791564214e-05,
      "loss": 2.4008,
      "theoretical_loss": 3.3754906899657673,
      "tokens_seen": 2436235264
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.210090788727067e-05,
      "loss": 2.4575,
      "theoretical_loss": 3.3754834095768533,
      "tokens_seen": 2436300800
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.208435785889919e-05,
      "loss": 1.9228,
      "theoretical_loss": 3.375476129438612,
      "tokens_seen": 2436366336
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.206780783052772e-05,
      "loss": 2.793,
      "theoretical_loss": 3.3754688495510274,
      "tokens_seen": 2436431872
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.205125780215624e-05,
      "loss": 2.6149,
      "theoretical_loss": 3.3754615699140844,
      "tokens_seen": 2436497408
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.203470777378475e-05,
      "loss": 2.4238,
      "theoretical_loss": 3.3754542905277667,
      "tokens_seen": 2436562944
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.201815774541327e-05,
      "loss": 2.223,
      "theoretical_loss": 3.3754470113920605,
      "tokens_seen": 2436628480
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.20016077170418e-05,
      "loss": 2.5563,
      "theoretical_loss": 3.3754397325069494,
      "tokens_seen": 2436694016
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.198505768867032e-05,
      "loss": 2.563,
      "theoretical_loss": 3.375432453872418,
      "tokens_seen": 2436759552
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.196850766029885e-05,
      "loss": 2.5454,
      "theoretical_loss": 3.3754251754884512,
      "tokens_seen": 2436825088
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.195195763192737e-05,
      "loss": 2.6359,
      "theoretical_loss": 3.3754178973550335,
      "tokens_seen": 2436890624
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.19354076035559e-05,
      "loss": 2.4524,
      "theoretical_loss": 3.37541061947215,
      "tokens_seen": 2436956160
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.191885757518442e-05,
      "loss": 2.6346,
      "theoretical_loss": 3.375403341839785,
      "tokens_seen": 2437021696
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.190230754681294e-05,
      "loss": 3.0881,
      "theoretical_loss": 3.375396064457923,
      "tokens_seen": 2437087232
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.188575751844147e-05,
      "loss": 2.6095,
      "theoretical_loss": 3.3753887873265485,
      "tokens_seen": 2437152768
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.186920749006999e-05,
      "loss": 2.4626,
      "theoretical_loss": 3.3753815104456466,
      "tokens_seen": 2437218304
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.185265746169851e-05,
      "loss": 2.393,
      "theoretical_loss": 3.3753742338152017,
      "tokens_seen": 2437283840
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.183610743332704e-05,
      "loss": 2.3775,
      "theoretical_loss": 3.3753669574351983,
      "tokens_seen": 2437349376
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.181955740495556e-05,
      "loss": 2.5889,
      "theoretical_loss": 3.3753596813056213,
      "tokens_seen": 2437414912
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.180300737658409e-05,
      "loss": 2.3075,
      "theoretical_loss": 3.375352405426456,
      "tokens_seen": 2437480448
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.178645734821258e-05,
      "loss": 2.3222,
      "theoretical_loss": 3.3753451297976858,
      "tokens_seen": 2437545984
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.176990731984111e-05,
      "loss": 2.4481,
      "theoretical_loss": 3.375337854419296,
      "tokens_seen": 2437611520
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.175335729146963e-05,
      "loss": 2.6961,
      "theoretical_loss": 3.375330579291271,
      "tokens_seen": 2437677056
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.173680726309816e-05,
      "loss": 2.3447,
      "theoretical_loss": 3.3753233044135955,
      "tokens_seen": 2437742592
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.172025723472668e-05,
      "loss": 2.2507,
      "theoretical_loss": 3.375316029786254,
      "tokens_seen": 2437808128
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.17037072063552e-05,
      "loss": 2.6491,
      "theoretical_loss": 3.375308755409232,
      "tokens_seen": 2437873664
    },
    {
      "epoch": 0.87,
      "objective/train/avg_token_score": 0.023234326392412186,
      "objective/train/avg_weight": 0.9726607799530029,
      "objective/train/docs_used": 1369908,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7476868629455566,
      "objective/train/std_weight": 0.09353917837142944,
      "objective/train/theoretical_loss": 3.3753014812825133,
      "objective/train/tokens_used": 2458399200,
      "theoretical_loss": 3.3753014812825133,
      "tokens_seen": 2437939200
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.168715717798373e-05,
      "loss": 2.3474,
      "theoretical_loss": 3.3753014812825133,
      "tokens_seen": 2437939200
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.167060714961225e-05,
      "loss": 2.4541,
      "theoretical_loss": 3.3752942074060828,
      "tokens_seen": 2438004736
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.165405712124078e-05,
      "loss": 2.2755,
      "theoretical_loss": 3.3752869337799254,
      "tokens_seen": 2438070272
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.16375070928693e-05,
      "loss": 2.439,
      "theoretical_loss": 3.3752796604040256,
      "tokens_seen": 2438135808
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.162095706449782e-05,
      "loss": 2.4586,
      "theoretical_loss": 3.375272387278368,
      "tokens_seen": 2438201344
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.160440703612635e-05,
      "loss": 2.3871,
      "theoretical_loss": 3.3752651144029366,
      "tokens_seen": 2438266880
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.158785700775487e-05,
      "loss": 2.2842,
      "theoretical_loss": 3.3752578417777177,
      "tokens_seen": 2438332416
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.15713069793834e-05,
      "loss": 2.5787,
      "theoretical_loss": 3.375250569402694,
      "tokens_seen": 2438397952
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.15547569510119e-05,
      "loss": 2.3842,
      "theoretical_loss": 3.3752432972778523,
      "tokens_seen": 2438463488
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.153820692264043e-05,
      "loss": 2.4271,
      "theoretical_loss": 3.3752360254031757,
      "tokens_seen": 2438529024
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.152165689426895e-05,
      "loss": 2.3963,
      "theoretical_loss": 3.375228753778649,
      "tokens_seen": 2438594560
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.150510686589748e-05,
      "loss": 2.5126,
      "theoretical_loss": 3.3752214824042572,
      "tokens_seen": 2438660096
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.1488556837526e-05,
      "loss": 2.3737,
      "theoretical_loss": 3.375214211279985,
      "tokens_seen": 2438725632
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.147200680915453e-05,
      "loss": 2.3271,
      "theoretical_loss": 3.3752069404058176,
      "tokens_seen": 2438791168
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.145545678078305e-05,
      "loss": 2.5537,
      "theoretical_loss": 3.375199669781739,
      "tokens_seen": 2438856704
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.143890675241157e-05,
      "loss": 2.242,
      "theoretical_loss": 3.3751923994077337,
      "tokens_seen": 2438922240
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.14223567240401e-05,
      "loss": 2.5473,
      "theoretical_loss": 3.3751851292837864,
      "tokens_seen": 2438987776
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.140580669566862e-05,
      "loss": 2.4789,
      "theoretical_loss": 3.3751778594098822,
      "tokens_seen": 2439053312
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.138925666729715e-05,
      "loss": 2.3567,
      "theoretical_loss": 3.375170589786006,
      "tokens_seen": 2439118848
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.137270663892567e-05,
      "loss": 2.5459,
      "theoretical_loss": 3.375163320412142,
      "tokens_seen": 2439184384
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.13561566105542e-05,
      "loss": 2.3558,
      "theoretical_loss": 3.3751560512882746,
      "tokens_seen": 2439249920
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.133960658218272e-05,
      "loss": 2.4995,
      "theoretical_loss": 3.3751487824143895,
      "tokens_seen": 2439315456
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.132305655381124e-05,
      "loss": 2.392,
      "theoretical_loss": 3.3751415137904703,
      "tokens_seen": 2439380992
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.130650652543975e-05,
      "loss": 2.4145,
      "theoretical_loss": 3.3751342454165023,
      "tokens_seen": 2439446528
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.128995649706828e-05,
      "loss": 2.7548,
      "theoretical_loss": 3.3751269772924704,
      "tokens_seen": 2439512064
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.12734064686968e-05,
      "loss": 2.5809,
      "theoretical_loss": 3.3751197094183585,
      "tokens_seen": 2439577600
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.125685644032532e-05,
      "loss": 2.6945,
      "theoretical_loss": 3.375112441794152,
      "tokens_seen": 2439643136
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.124030641195385e-05,
      "loss": 2.4917,
      "theoretical_loss": 3.3751051744198355,
      "tokens_seen": 2439708672
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.122375638358237e-05,
      "loss": 2.3411,
      "theoretical_loss": 3.375097907295393,
      "tokens_seen": 2439774208
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.12072063552109e-05,
      "loss": 2.702,
      "theoretical_loss": 3.37509064042081,
      "tokens_seen": 2439839744
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.119065632683942e-05,
      "loss": 2.3935,
      "theoretical_loss": 3.375083373796071,
      "tokens_seen": 2439905280
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.117410629846794e-05,
      "loss": 2.5743,
      "theoretical_loss": 3.3750761074211604,
      "tokens_seen": 2439970816
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.115755627009647e-05,
      "loss": 2.4378,
      "theoretical_loss": 3.3750688412960637,
      "tokens_seen": 2440036352
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.114100624172499e-05,
      "loss": 2.3626,
      "theoretical_loss": 3.3750615754207645,
      "tokens_seen": 2440101888
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.112445621335352e-05,
      "loss": 2.6614,
      "theoretical_loss": 3.375054309795248,
      "tokens_seen": 2440167424
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.110790618498204e-05,
      "loss": 2.8148,
      "theoretical_loss": 3.375047044419499,
      "tokens_seen": 2440232960
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.109135615661056e-05,
      "loss": 2.3701,
      "theoretical_loss": 3.3750397792935023,
      "tokens_seen": 2440298496
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.107480612823909e-05,
      "loss": 1.623,
      "theoretical_loss": 3.375032514417242,
      "tokens_seen": 2440364032
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.105825609986758e-05,
      "loss": 2.6156,
      "theoretical_loss": 3.375025249790704,
      "tokens_seen": 2440429568
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.104170607149611e-05,
      "loss": 2.5539,
      "theoretical_loss": 3.375017985413872,
      "tokens_seen": 2440495104
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.102515604312463e-05,
      "loss": 2.2927,
      "theoretical_loss": 3.375010721286731,
      "tokens_seen": 2440560640
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.100860601475316e-05,
      "loss": 2.4163,
      "theoretical_loss": 3.3750034574092656,
      "tokens_seen": 2440626176
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.099205598638168e-05,
      "loss": 2.5073,
      "theoretical_loss": 3.3749961937814605,
      "tokens_seen": 2440691712
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.09755059580102e-05,
      "loss": 2.5051,
      "theoretical_loss": 3.3749889304033007,
      "tokens_seen": 2440757248
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.095895592963873e-05,
      "loss": 2.3192,
      "theoretical_loss": 3.3749816672747706,
      "tokens_seen": 2440822784
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.094240590126725e-05,
      "loss": 2.5458,
      "theoretical_loss": 3.3749744043958554,
      "tokens_seen": 2440888320
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.092585587289578e-05,
      "loss": 2.4347,
      "theoretical_loss": 3.374967141766539,
      "tokens_seen": 2440953856
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.09093058445243e-05,
      "loss": 2.6285,
      "theoretical_loss": 3.374959879386807,
      "tokens_seen": 2441019392
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.089275581615282e-05,
      "loss": 2.7723,
      "theoretical_loss": 3.3749526172566435,
      "tokens_seen": 2441084928
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.087620578778135e-05,
      "loss": 2.679,
      "theoretical_loss": 3.3749453553760334,
      "tokens_seen": 2441150464
    },
    {
      "epoch": 0.87,
      "objective/train/avg_token_score": 0.021364537999033928,
      "objective/train/avg_weight": 0.9748502969741821,
      "objective/train/docs_used": 1371503,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.7499465942382812,
      "objective/train/std_weight": 0.11291147023439407,
      "objective/train/theoretical_loss": 3.374938093744962,
      "objective/train/tokens_used": 2461676000,
      "theoretical_loss": 3.374938093744962,
      "tokens_seen": 2441216000
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.085965575940987e-05,
      "loss": 2.3883,
      "theoretical_loss": 3.374938093744962,
      "tokens_seen": 2441216000
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.08431057310384e-05,
      "loss": 2.6676,
      "theoretical_loss": 3.374930832363413,
      "tokens_seen": 2441281536
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.082655570266692e-05,
      "loss": 2.3973,
      "theoretical_loss": 3.3749235712313714,
      "tokens_seen": 2441347072
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.081000567429543e-05,
      "loss": 2.0214,
      "theoretical_loss": 3.3749163103488224,
      "tokens_seen": 2441412608
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.079345564592396e-05,
      "loss": 2.4176,
      "theoretical_loss": 3.374909049715751,
      "tokens_seen": 2441478144
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.077690561755248e-05,
      "loss": 2.8187,
      "theoretical_loss": 3.3749017893321405,
      "tokens_seen": 2441543680
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.0760355589181e-05,
      "loss": 2.5379,
      "theoretical_loss": 3.3748945291979773,
      "tokens_seen": 2441609216
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.074380556080953e-05,
      "loss": 2.4251,
      "theoretical_loss": 3.374887269313245,
      "tokens_seen": 2441674752
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.072725553243805e-05,
      "loss": 2.5925,
      "theoretical_loss": 3.374880009677929,
      "tokens_seen": 2441740288
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.071070550406658e-05,
      "loss": 2.5846,
      "theoretical_loss": 3.3748727502920133,
      "tokens_seen": 2441805824
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.06941554756951e-05,
      "loss": 2.4917,
      "theoretical_loss": 3.3748654911554836,
      "tokens_seen": 2441871360
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.067760544732362e-05,
      "loss": 2.5259,
      "theoretical_loss": 3.374858232268324,
      "tokens_seen": 2441936896
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.066105541895215e-05,
      "loss": 2.486,
      "theoretical_loss": 3.3748509736305192,
      "tokens_seen": 2442002432
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.064450539058067e-05,
      "loss": 2.4005,
      "theoretical_loss": 3.3748437152420543,
      "tokens_seen": 2442067968
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.06279553622092e-05,
      "loss": 2.4707,
      "theoretical_loss": 3.374836457102914,
      "tokens_seen": 2442133504
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.061140533383772e-05,
      "loss": 2.2038,
      "theoretical_loss": 3.3748291992130826,
      "tokens_seen": 2442199040
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.059485530546624e-05,
      "loss": 2.9274,
      "theoretical_loss": 3.374821941572545,
      "tokens_seen": 2442264576
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.057830527709477e-05,
      "loss": 2.722,
      "theoretical_loss": 3.3748146841812865,
      "tokens_seen": 2442330112
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.056175524872328e-05,
      "loss": 2.4716,
      "theoretical_loss": 3.3748074270392916,
      "tokens_seen": 2442395648
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.05452052203518e-05,
      "loss": 2.85,
      "theoretical_loss": 3.3748001701465444,
      "tokens_seen": 2442461184
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.052865519198033e-05,
      "loss": 2.4985,
      "theoretical_loss": 3.374792913503031,
      "tokens_seen": 2442526720
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.051210516360885e-05,
      "loss": 2.5941,
      "theoretical_loss": 3.3747856571087342,
      "tokens_seen": 2442592256
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.049555513523737e-05,
      "loss": 2.8114,
      "theoretical_loss": 3.3747784009636406,
      "tokens_seen": 2442657792
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.04790051068659e-05,
      "loss": 2.5429,
      "theoretical_loss": 3.3747711450677342,
      "tokens_seen": 2442723328
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.046245507849442e-05,
      "loss": 2.5536,
      "theoretical_loss": 3.3747638894209997,
      "tokens_seen": 2442788864
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.044590505012295e-05,
      "loss": 2.4329,
      "theoretical_loss": 3.374756634023422,
      "tokens_seen": 2442854400
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.042935502175147e-05,
      "loss": 1.559,
      "theoretical_loss": 3.374749378874985,
      "tokens_seen": 2442919936
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.041280499338e-05,
      "loss": 2.4981,
      "theoretical_loss": 3.3747421239756754,
      "tokens_seen": 2442985472
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.039625496500852e-05,
      "loss": 2.3526,
      "theoretical_loss": 3.374734869325476,
      "tokens_seen": 2443051008
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.037970493663704e-05,
      "loss": 2.5675,
      "theoretical_loss": 3.374727614924373,
      "tokens_seen": 2443116544
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.036315490826557e-05,
      "loss": 2.2162,
      "theoretical_loss": 3.3747203607723506,
      "tokens_seen": 2443182080
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.034660487989409e-05,
      "loss": 2.5955,
      "theoretical_loss": 3.374713106869393,
      "tokens_seen": 2443247616
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.033005485152261e-05,
      "loss": 2.498,
      "theoretical_loss": 3.3747058532154854,
      "tokens_seen": 2443313152
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.031350482315111e-05,
      "loss": 2.6634,
      "theoretical_loss": 3.374698599810613,
      "tokens_seen": 2443378688
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.029695479477963e-05,
      "loss": 2.5116,
      "theoretical_loss": 3.3746913466547603,
      "tokens_seen": 2443444224
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.028040476640816e-05,
      "loss": 2.4576,
      "theoretical_loss": 3.374684093747912,
      "tokens_seen": 2443509760
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.026385473803668e-05,
      "loss": 2.4936,
      "theoretical_loss": 3.3746768410900527,
      "tokens_seen": 2443575296
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.02473047096652e-05,
      "loss": 2.589,
      "theoretical_loss": 3.3746695886811677,
      "tokens_seen": 2443640832
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.023075468129373e-05,
      "loss": 2.7219,
      "theoretical_loss": 3.374662336521241,
      "tokens_seen": 2443706368
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.021420465292225e-05,
      "loss": 2.6162,
      "theoretical_loss": 3.374655084610258,
      "tokens_seen": 2443771904
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.019765462455078e-05,
      "loss": 2.5423,
      "theoretical_loss": 3.3746478329482033,
      "tokens_seen": 2443837440
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.01811045961793e-05,
      "loss": 2.6731,
      "theoretical_loss": 3.3746405815350613,
      "tokens_seen": 2443902976
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.016455456780783e-05,
      "loss": 2.4632,
      "theoretical_loss": 3.3746333303708176,
      "tokens_seen": 2443968512
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.014800453943635e-05,
      "loss": 2.3462,
      "theoretical_loss": 3.3746260794554566,
      "tokens_seen": 2444034048
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.013145451106487e-05,
      "loss": 2.5788,
      "theoretical_loss": 3.3746188287889627,
      "tokens_seen": 2444099584
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.011490448269341e-05,
      "loss": 2.4761,
      "theoretical_loss": 3.3746115783713213,
      "tokens_seen": 2444165120
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.009835445432194e-05,
      "loss": 2.2134,
      "theoretical_loss": 3.3746043282025164,
      "tokens_seen": 2444230656
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.008180442595046e-05,
      "loss": 2.5469,
      "theoretical_loss": 3.3745970782825334,
      "tokens_seen": 2444296192
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.006525439757896e-05,
      "loss": 2.4656,
      "theoretical_loss": 3.3745898286113576,
      "tokens_seen": 2444361728
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.004870436920748e-05,
      "loss": 2.2076,
      "theoretical_loss": 3.3745825791889725,
      "tokens_seen": 2444427264
    },
    {
      "epoch": 0.87,
      "objective/train/avg_token_score": 0.020413454622030258,
      "objective/train/avg_weight": 0.9759781956672668,
      "objective/train/docs_used": 1373215,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.439959526062012,
      "objective/train/std_weight": 0.10474928468465805,
      "objective/train/theoretical_loss": 3.3745753300153636,
      "objective/train/tokens_used": 2464952800,
      "theoretical_loss": 3.3745753300153636,
      "tokens_seen": 2444492800
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.0032154340836e-05,
      "loss": 2.4134,
      "theoretical_loss": 3.3745753300153636,
      "tokens_seen": 2444492800
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.001560431246453e-05,
      "loss": 2.5258,
      "theoretical_loss": 3.3745680810905156,
      "tokens_seen": 2444558336
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.999905428409305e-05,
      "loss": 2.6824,
      "theoretical_loss": 3.3745608324144136,
      "tokens_seen": 2444623872
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.998250425572158e-05,
      "loss": 2.4815,
      "theoretical_loss": 3.374553583987042,
      "tokens_seen": 2444689408
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.99659542273501e-05,
      "loss": 2.6605,
      "theoretical_loss": 3.374546335808386,
      "tokens_seen": 2444754944
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.994940419897862e-05,
      "loss": 2.4471,
      "theoretical_loss": 3.3745390878784294,
      "tokens_seen": 2444820480
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.993285417060715e-05,
      "loss": 2.3603,
      "theoretical_loss": 3.3745318401971582,
      "tokens_seen": 2444886016
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.991630414223567e-05,
      "loss": 2.5459,
      "theoretical_loss": 3.3745245927645566,
      "tokens_seen": 2444951552
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.98997541138642e-05,
      "loss": 2.5221,
      "theoretical_loss": 3.37451734558061,
      "tokens_seen": 2445017088
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.988320408549272e-05,
      "loss": 2.5042,
      "theoretical_loss": 3.3745100986453025,
      "tokens_seen": 2445082624
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.986665405712124e-05,
      "loss": 2.634,
      "theoretical_loss": 3.374502851958619,
      "tokens_seen": 2445148160
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.985010402874977e-05,
      "loss": 2.2506,
      "theoretical_loss": 3.3744956055205444,
      "tokens_seen": 2445213696
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.983355400037828e-05,
      "loss": 2.6061,
      "theoretical_loss": 3.3744883593310635,
      "tokens_seen": 2445279232
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.98170039720068e-05,
      "loss": 2.8055,
      "theoretical_loss": 3.3744811133901615,
      "tokens_seen": 2445344768
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.980045394363533e-05,
      "loss": 2.516,
      "theoretical_loss": 3.3744738676978225,
      "tokens_seen": 2445410304
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.978390391526385e-05,
      "loss": 2.4201,
      "theoretical_loss": 3.374466622254032,
      "tokens_seen": 2445475840
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.976735388689237e-05,
      "loss": 2.7098,
      "theoretical_loss": 3.3744593770587743,
      "tokens_seen": 2445541376
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.97508038585209e-05,
      "loss": 2.6941,
      "theoretical_loss": 3.3744521321120344,
      "tokens_seen": 2445606912
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.973425383014942e-05,
      "loss": 2.4126,
      "theoretical_loss": 3.374444887413797,
      "tokens_seen": 2445672448
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.971770380177795e-05,
      "loss": 2.205,
      "theoretical_loss": 3.3744376429640477,
      "tokens_seen": 2445737984
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.970115377340647e-05,
      "loss": 2.312,
      "theoretical_loss": 3.37443039876277,
      "tokens_seen": 2445803520
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.9684603745035e-05,
      "loss": 2.4163,
      "theoretical_loss": 3.3744231548099495,
      "tokens_seen": 2445869056
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.966805371666352e-05,
      "loss": 2.5204,
      "theoretical_loss": 3.3744159111055714,
      "tokens_seen": 2445934592
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.965150368829204e-05,
      "loss": 2.4214,
      "theoretical_loss": 3.3744086676496194,
      "tokens_seen": 2446000128
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.963495365992057e-05,
      "loss": 2.4124,
      "theoretical_loss": 3.3744014244420795,
      "tokens_seen": 2446065664
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.961840363154909e-05,
      "loss": 2.4777,
      "theoretical_loss": 3.3743941814829355,
      "tokens_seen": 2446131200
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.960185360317761e-05,
      "loss": 2.6196,
      "theoretical_loss": 3.374386938772173,
      "tokens_seen": 2446196736
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.958530357480611e-05,
      "loss": 2.8217,
      "theoretical_loss": 3.3743796963097763,
      "tokens_seen": 2446262272
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.956875354643464e-05,
      "loss": 2.4332,
      "theoretical_loss": 3.374372454095731,
      "tokens_seen": 2446327808
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.955220351806316e-05,
      "loss": 2.5259,
      "theoretical_loss": 3.3743652121300207,
      "tokens_seen": 2446393344
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.953565348969168e-05,
      "loss": 2.6839,
      "theoretical_loss": 3.3743579704126314,
      "tokens_seen": 2446458880
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.951910346132021e-05,
      "loss": 2.48,
      "theoretical_loss": 3.374350728943547,
      "tokens_seen": 2446524416
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.950255343294873e-05,
      "loss": 2.5902,
      "theoretical_loss": 3.3743434877227534,
      "tokens_seen": 2446589952
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.948600340457727e-05,
      "loss": 2.5218,
      "theoretical_loss": 3.3743362467502345,
      "tokens_seen": 2446655488
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.946945337620579e-05,
      "loss": 2.6336,
      "theoretical_loss": 3.3743290060259756,
      "tokens_seen": 2446721024
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.945290334783432e-05,
      "loss": 2.556,
      "theoretical_loss": 3.3743217655499613,
      "tokens_seen": 2446786560
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.943635331946284e-05,
      "loss": 2.8402,
      "theoretical_loss": 3.374314525322177,
      "tokens_seen": 2446852096
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.941980329109137e-05,
      "loss": 2.4116,
      "theoretical_loss": 3.3743072853426064,
      "tokens_seen": 2446917632
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.940325326271989e-05,
      "loss": 2.7906,
      "theoretical_loss": 3.3743000456112355,
      "tokens_seen": 2446983168
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.938670323434841e-05,
      "loss": 2.4685,
      "theoretical_loss": 3.3742928061280484,
      "tokens_seen": 2447048704
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.937015320597694e-05,
      "loss": 2.351,
      "theoretical_loss": 3.3742855668930307,
      "tokens_seen": 2447114240
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.935360317760546e-05,
      "loss": 2.5551,
      "theoretical_loss": 3.374278327906166,
      "tokens_seen": 2447179776
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.933705314923396e-05,
      "loss": 2.4064,
      "theoretical_loss": 3.374271089167441,
      "tokens_seen": 2447245312
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.932050312086248e-05,
      "loss": 2.6071,
      "theoretical_loss": 3.3742638506768383,
      "tokens_seen": 2447310848
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.9303953092491e-05,
      "loss": 2.4763,
      "theoretical_loss": 3.3742566124343445,
      "tokens_seen": 2447376384
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.928740306411953e-05,
      "loss": 2.4876,
      "theoretical_loss": 3.3742493744399438,
      "tokens_seen": 2447441920
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.927085303574805e-05,
      "loss": 2.5667,
      "theoretical_loss": 3.3742421366936215,
      "tokens_seen": 2447507456
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.925430300737658e-05,
      "loss": 2.5557,
      "theoretical_loss": 3.3742348991953617,
      "tokens_seen": 2447572992
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.92377529790051e-05,
      "loss": 2.5435,
      "theoretical_loss": 3.3742276619451497,
      "tokens_seen": 2447638528
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.922120295063363e-05,
      "loss": 2.2152,
      "theoretical_loss": 3.3742204249429704,
      "tokens_seen": 2447704064
    },
    {
      "epoch": 0.87,
      "objective/train/avg_token_score": 0.007849251851439476,
      "objective/train/avg_weight": 0.9907592535018921,
      "objective/train/docs_used": 1375012,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4958853721618652,
      "objective/train/std_weight": 0.05522432178258896,
      "objective/train/theoretical_loss": 3.374213188188808,
      "objective/train/tokens_used": 2468229600,
      "theoretical_loss": 3.374213188188808,
      "tokens_seen": 2447769600
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.920465292226215e-05,
      "loss": 2.5342,
      "theoretical_loss": 3.374213188188808,
      "tokens_seen": 2447769600
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.918810289389067e-05,
      "loss": 2.4398,
      "theoretical_loss": 3.374205951682649,
      "tokens_seen": 2447835136
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.91715528655192e-05,
      "loss": 2.0888,
      "theoretical_loss": 3.3741987154244764,
      "tokens_seen": 2447900672
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.915500283714772e-05,
      "loss": 2.5901,
      "theoretical_loss": 3.374191479414276,
      "tokens_seen": 2447966208
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.913845280877625e-05,
      "loss": 2.3915,
      "theoretical_loss": 3.3741842436520324,
      "tokens_seen": 2448031744
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.912190278040477e-05,
      "loss": 2.7931,
      "theoretical_loss": 3.374177008137731,
      "tokens_seen": 2448097280
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.91053527520333e-05,
      "loss": 2.5539,
      "theoretical_loss": 3.3741697728713556,
      "tokens_seen": 2448162816
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.90888027236618e-05,
      "loss": 2.1428,
      "theoretical_loss": 3.374162537852892,
      "tokens_seen": 2448228352
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.907225269529033e-05,
      "loss": 2.4576,
      "theoretical_loss": 3.374155303082325,
      "tokens_seen": 2448293888
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.905570266691885e-05,
      "loss": 2.4783,
      "theoretical_loss": 3.374148068559639,
      "tokens_seen": 2448359424
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.903915263854738e-05,
      "loss": 2.6086,
      "theoretical_loss": 3.374140834284819,
      "tokens_seen": 2448424960
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.90226026101759e-05,
      "loss": 2.4651,
      "theoretical_loss": 3.37413360025785,
      "tokens_seen": 2448490496
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.900605258180442e-05,
      "loss": 2.4117,
      "theoretical_loss": 3.3741263664787176,
      "tokens_seen": 2448556032
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.898950255343295e-05,
      "loss": 2.4129,
      "theoretical_loss": 3.3741191329474054,
      "tokens_seen": 2448621568
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.897295252506147e-05,
      "loss": 2.3052,
      "theoretical_loss": 3.3741118996638986,
      "tokens_seen": 2448687104
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.895640249669e-05,
      "loss": 2.4037,
      "theoretical_loss": 3.3741046666281824,
      "tokens_seen": 2448752640
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.893985246831852e-05,
      "loss": 2.6009,
      "theoretical_loss": 3.374097433840242,
      "tokens_seen": 2448818176
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.892330243994704e-05,
      "loss": 2.4569,
      "theoretical_loss": 3.3740902013000613,
      "tokens_seen": 2448883712
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.890675241157557e-05,
      "loss": 2.4408,
      "theoretical_loss": 3.374082969007626,
      "tokens_seen": 2448949248
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.889020238320409e-05,
      "loss": 2.526,
      "theoretical_loss": 3.374075736962921,
      "tokens_seen": 2449014784
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.887365235483262e-05,
      "loss": 2.2753,
      "theoretical_loss": 3.3740685051659307,
      "tokens_seen": 2449080320
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.885710232646114e-05,
      "loss": 2.7385,
      "theoretical_loss": 3.3740612736166398,
      "tokens_seen": 2449145856
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.884055229808965e-05,
      "loss": 2.6354,
      "theoretical_loss": 3.3740540423150343,
      "tokens_seen": 2449211392
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.882400226971817e-05,
      "loss": 2.2668,
      "theoretical_loss": 3.3740468112610977,
      "tokens_seen": 2449276928
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.88074522413467e-05,
      "loss": 2.6652,
      "theoretical_loss": 3.374039580454816,
      "tokens_seen": 2449342464
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.879090221297522e-05,
      "loss": 2.3854,
      "theoretical_loss": 3.374032349896174,
      "tokens_seen": 2449408000
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.877435218460375e-05,
      "loss": 2.6248,
      "theoretical_loss": 3.3740251195851556,
      "tokens_seen": 2449473536
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.875780215623227e-05,
      "loss": 2.407,
      "theoretical_loss": 3.3740178895217463,
      "tokens_seen": 2449539072
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.87412521278608e-05,
      "loss": 2.3482,
      "theoretical_loss": 3.3740106597059314,
      "tokens_seen": 2449604608
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.872470209948932e-05,
      "loss": 2.5248,
      "theoretical_loss": 3.3740034301376953,
      "tokens_seen": 2449670144
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.870815207111784e-05,
      "loss": 2.6406,
      "theoretical_loss": 3.3739962008170234,
      "tokens_seen": 2449735680
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.869160204274637e-05,
      "loss": 2.5531,
      "theoretical_loss": 3.3739889717439,
      "tokens_seen": 2449801216
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.867505201437489e-05,
      "loss": 2.4488,
      "theoretical_loss": 3.37398174291831,
      "tokens_seen": 2449866752
    },
    {
      "epoch": 0.87,
      "learning_rate": 8.865850198600341e-05,
      "loss": 2.4609,
      "theoretical_loss": 3.3739745143402384,
      "tokens_seen": 2449932288
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.864195195763194e-05,
      "loss": 2.4653,
      "theoretical_loss": 3.373967286009671,
      "tokens_seen": 2449997824
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.862540192926046e-05,
      "loss": 2.5789,
      "theoretical_loss": 3.3739600579265914,
      "tokens_seen": 2450063360
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.860885190088899e-05,
      "loss": 2.7348,
      "theoretical_loss": 3.3739528300909853,
      "tokens_seen": 2450128896
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.859230187251748e-05,
      "loss": 2.6432,
      "theoretical_loss": 3.3739456025028374,
      "tokens_seen": 2450194432
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.857575184414601e-05,
      "loss": 2.4958,
      "theoretical_loss": 3.3739383751621324,
      "tokens_seen": 2450259968
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.855920181577453e-05,
      "loss": 2.7457,
      "theoretical_loss": 3.3739311480688556,
      "tokens_seen": 2450325504
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.854265178740306e-05,
      "loss": 2.4705,
      "theoretical_loss": 3.3739239212229912,
      "tokens_seen": 2450391040
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.852610175903158e-05,
      "loss": 2.4491,
      "theoretical_loss": 3.373916694624525,
      "tokens_seen": 2450456576
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.85095517306601e-05,
      "loss": 2.7347,
      "theoretical_loss": 3.3739094682734416,
      "tokens_seen": 2450522112
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.849300170228863e-05,
      "loss": 2.54,
      "theoretical_loss": 3.373902242169726,
      "tokens_seen": 2450587648
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.847645167391715e-05,
      "loss": 2.4828,
      "theoretical_loss": 3.3738950163133623,
      "tokens_seen": 2450653184
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.845990164554568e-05,
      "loss": 2.7318,
      "theoretical_loss": 3.3738877907043365,
      "tokens_seen": 2450718720
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.84433516171742e-05,
      "loss": 2.7256,
      "theoretical_loss": 3.3738805653426334,
      "tokens_seen": 2450784256
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.842680158880272e-05,
      "loss": 2.5089,
      "theoretical_loss": 3.3738733402282373,
      "tokens_seen": 2450849792
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.841025156043125e-05,
      "loss": 2.0625,
      "theoretical_loss": 3.3738661153611336,
      "tokens_seen": 2450915328
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.839370153205977e-05,
      "loss": 2.4102,
      "theoretical_loss": 3.3738588907413067,
      "tokens_seen": 2450980864
    },
    {
      "epoch": 0.88,
      "objective/train/avg_token_score": 0.017230749130249023,
      "objective/train/avg_weight": 0.9797260761260986,
      "objective/train/docs_used": 1376896,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.810290813446045,
      "objective/train/std_weight": 0.08198007196187973,
      "objective/train/theoretical_loss": 3.373851666368742,
      "objective/train/tokens_used": 2471506400,
      "theoretical_loss": 3.373851666368742,
      "tokens_seen": 2451046400
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.83771515036883e-05,
      "loss": 2.5421,
      "theoretical_loss": 3.373851666368742,
      "tokens_seen": 2451046400
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.836060147531682e-05,
      "loss": 2.5257,
      "theoretical_loss": 3.373844442243425,
      "tokens_seen": 2451111936
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.834405144694533e-05,
      "loss": 2.6722,
      "theoretical_loss": 3.373837218365339,
      "tokens_seen": 2451177472
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.832750141857385e-05,
      "loss": 2.6758,
      "theoretical_loss": 3.3738299947344705,
      "tokens_seen": 2451243008
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.831095139020238e-05,
      "loss": 2.4478,
      "theoretical_loss": 3.3738227713508038,
      "tokens_seen": 2451308544
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.82944013618309e-05,
      "loss": 2.4853,
      "theoretical_loss": 3.3738155482143233,
      "tokens_seen": 2451374080
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.827785133345943e-05,
      "loss": 2.5706,
      "theoretical_loss": 3.373808325325015,
      "tokens_seen": 2451439616
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.826130130508795e-05,
      "loss": 2.7649,
      "theoretical_loss": 3.3738011026828634,
      "tokens_seen": 2451505152
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.824475127671647e-05,
      "loss": 2.6008,
      "theoretical_loss": 3.373793880287853,
      "tokens_seen": 2451570688
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.8228201248345e-05,
      "loss": 2.698,
      "theoretical_loss": 3.373786658139969,
      "tokens_seen": 2451636224
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.821165121997352e-05,
      "loss": 2.7296,
      "theoretical_loss": 3.3737794362391966,
      "tokens_seen": 2451701760
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.819510119160205e-05,
      "loss": 2.4308,
      "theoretical_loss": 3.3737722145855207,
      "tokens_seen": 2451767296
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.817855116323057e-05,
      "loss": 2.4023,
      "theoretical_loss": 3.3737649931789258,
      "tokens_seen": 2451832832
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.81620011348591e-05,
      "loss": 2.5816,
      "theoretical_loss": 3.373757772019397,
      "tokens_seen": 2451898368
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.814545110648762e-05,
      "loss": 2.946,
      "theoretical_loss": 3.37375055110692,
      "tokens_seen": 2451963904
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.812890107811614e-05,
      "loss": 2.6082,
      "theoretical_loss": 3.373743330441479,
      "tokens_seen": 2452029440
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.811235104974465e-05,
      "loss": 2.1795,
      "theoretical_loss": 3.3737361100230587,
      "tokens_seen": 2452094976
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.809580102137318e-05,
      "loss": 2.6425,
      "theoretical_loss": 3.3737288898516447,
      "tokens_seen": 2452160512
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.80792509930017e-05,
      "loss": 2.4101,
      "theoretical_loss": 3.3737216699272214,
      "tokens_seen": 2452226048
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.806270096463022e-05,
      "loss": 2.3898,
      "theoretical_loss": 3.3737144502497745,
      "tokens_seen": 2452291584
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.804615093625875e-05,
      "loss": 2.5471,
      "theoretical_loss": 3.373707230819288,
      "tokens_seen": 2452357120
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.802960090788727e-05,
      "loss": 2.5331,
      "theoretical_loss": 3.3737000116357474,
      "tokens_seen": 2452422656
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.80130508795158e-05,
      "loss": 2.7664,
      "theoretical_loss": 3.3736927926991376,
      "tokens_seen": 2452488192
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.799650085114432e-05,
      "loss": 2.7884,
      "theoretical_loss": 3.373685574009444,
      "tokens_seen": 2452553728
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.797995082277284e-05,
      "loss": 2.6072,
      "theoretical_loss": 3.3736783555666507,
      "tokens_seen": 2452619264
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.796340079440137e-05,
      "loss": 2.5012,
      "theoretical_loss": 3.373671137370743,
      "tokens_seen": 2452684800
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.794685076602989e-05,
      "loss": 2.3495,
      "theoretical_loss": 3.373663919421706,
      "tokens_seen": 2452750336
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.793030073765842e-05,
      "loss": 2.7079,
      "theoretical_loss": 3.3736567017195247,
      "tokens_seen": 2452815872
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.791375070928694e-05,
      "loss": 2.2259,
      "theoretical_loss": 3.3736494842641838,
      "tokens_seen": 2452881408
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.789720068091546e-05,
      "loss": 2.5707,
      "theoretical_loss": 3.3736422670556685,
      "tokens_seen": 2452946944
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.788065065254399e-05,
      "loss": 2.4944,
      "theoretical_loss": 3.373635050093964,
      "tokens_seen": 2453012480
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.786410062417248e-05,
      "loss": 2.2461,
      "theoretical_loss": 3.373627833379054,
      "tokens_seen": 2453078016
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.784755059580101e-05,
      "loss": 2.4446,
      "theoretical_loss": 3.373620616910925,
      "tokens_seen": 2453143552
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.783100056742953e-05,
      "loss": 2.2522,
      "theoretical_loss": 3.3736134006895613,
      "tokens_seen": 2453209088
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.781445053905806e-05,
      "loss": 2.3829,
      "theoretical_loss": 3.373606184714948,
      "tokens_seen": 2453274624
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.779790051068658e-05,
      "loss": 2.3611,
      "theoretical_loss": 3.37359896898707,
      "tokens_seen": 2453340160
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.77813504823151e-05,
      "loss": 2.2967,
      "theoretical_loss": 3.373591753505912,
      "tokens_seen": 2453405696
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.776480045394363e-05,
      "loss": 2.2658,
      "theoretical_loss": 3.3735845382714595,
      "tokens_seen": 2453471232
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.774825042557215e-05,
      "loss": 2.5277,
      "theoretical_loss": 3.373577323283697,
      "tokens_seen": 2453536768
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.773170039720068e-05,
      "loss": 2.3597,
      "theoretical_loss": 3.3735701085426104,
      "tokens_seen": 2453602304
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.77151503688292e-05,
      "loss": 2.5685,
      "theoretical_loss": 3.3735628940481837,
      "tokens_seen": 2453667840
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.769860034045772e-05,
      "loss": 2.6294,
      "theoretical_loss": 3.3735556798004014,
      "tokens_seen": 2453733376
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.768205031208625e-05,
      "loss": 2.4709,
      "theoretical_loss": 3.3735484657992503,
      "tokens_seen": 2453798912
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.766550028371477e-05,
      "loss": 2.4431,
      "theoretical_loss": 3.3735412520447134,
      "tokens_seen": 2453864448
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.76489502553433e-05,
      "loss": 1.7203,
      "theoretical_loss": 3.373534038536777,
      "tokens_seen": 2453929984
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.763240022697182e-05,
      "loss": 2.4851,
      "theoretical_loss": 3.3735268252754262,
      "tokens_seen": 2453995520
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.761585019860033e-05,
      "loss": 2.5535,
      "theoretical_loss": 3.373519612260645,
      "tokens_seen": 2454061056
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.759930017022886e-05,
      "loss": 2.503,
      "theoretical_loss": 3.373512399492419,
      "tokens_seen": 2454126592
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.758275014185738e-05,
      "loss": 2.3455,
      "theoretical_loss": 3.373505186970733,
      "tokens_seen": 2454192128
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.75662001134859e-05,
      "loss": 2.2923,
      "theoretical_loss": 3.3734979746955718,
      "tokens_seen": 2454257664
    },
    {
      "epoch": 0.88,
      "objective/train/avg_token_score": 0.008946981281042099,
      "objective/train/avg_weight": 0.9894744753837585,
      "objective/train/docs_used": 1379050,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7515578269958496,
      "objective/train/std_weight": 0.04179955646395683,
      "objective/train/theoretical_loss": 3.373490762666921,
      "objective/train/tokens_used": 2474783200,
      "theoretical_loss": 3.373490762666921,
      "tokens_seen": 2454323200
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.754965008511443e-05,
      "loss": 2.2791,
      "theoretical_loss": 3.373490762666921,
      "tokens_seen": 2454323200
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.753310005674295e-05,
      "loss": 2.2479,
      "theoretical_loss": 3.373483550884765,
      "tokens_seen": 2454388736
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.751655002837148e-05,
      "loss": 2.4511,
      "theoretical_loss": 3.3734763393490894,
      "tokens_seen": 2454454272
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.75e-05,
      "loss": 2.7304,
      "theoretical_loss": 3.3734691280598783,
      "tokens_seen": 2454519808
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.748344997162852e-05,
      "loss": 2.5549,
      "theoretical_loss": 3.3734619170171176,
      "tokens_seen": 2454585344
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.746689994325705e-05,
      "loss": 2.8631,
      "theoretical_loss": 3.373454706220792,
      "tokens_seen": 2454650880
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.745034991488557e-05,
      "loss": 2.4966,
      "theoretical_loss": 3.3734474956708858,
      "tokens_seen": 2454716416
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.743379988651408e-05,
      "loss": 2.4833,
      "theoretical_loss": 3.373440285367385,
      "tokens_seen": 2454781952
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.74172498581426e-05,
      "loss": 2.5178,
      "theoretical_loss": 3.3734330753102744,
      "tokens_seen": 2454847488
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.740069982977113e-05,
      "loss": 2.5791,
      "theoretical_loss": 3.3734258654995384,
      "tokens_seen": 2454913024
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.738414980139965e-05,
      "loss": 2.6926,
      "theoretical_loss": 3.3734186559351627,
      "tokens_seen": 2454978560
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.736759977302818e-05,
      "loss": 2.4595,
      "theoretical_loss": 3.373411446617132,
      "tokens_seen": 2455044096
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.73510497446567e-05,
      "loss": 2.6113,
      "theoretical_loss": 3.373404237545431,
      "tokens_seen": 2455109632
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.733449971628523e-05,
      "loss": 2.635,
      "theoretical_loss": 3.3733970287200457,
      "tokens_seen": 2455175168
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.731794968791375e-05,
      "loss": 2.6959,
      "theoretical_loss": 3.37338982014096,
      "tokens_seen": 2455240704
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.730139965954227e-05,
      "loss": 2.7655,
      "theoretical_loss": 3.3733826118081596,
      "tokens_seen": 2455306240
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.72848496311708e-05,
      "loss": 2.3521,
      "theoretical_loss": 3.3733754037216293,
      "tokens_seen": 2455371776
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.726829960279932e-05,
      "loss": 2.5222,
      "theoretical_loss": 3.373368195881354,
      "tokens_seen": 2455437312
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.725174957442785e-05,
      "loss": 2.5203,
      "theoretical_loss": 3.3733609882873186,
      "tokens_seen": 2455502848
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.723519954605637e-05,
      "loss": 2.7258,
      "theoretical_loss": 3.3733537809395084,
      "tokens_seen": 2455568384
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.72186495176849e-05,
      "loss": 2.4637,
      "theoretical_loss": 3.3733465738379085,
      "tokens_seen": 2455633920
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.720209948931342e-05,
      "loss": 2.6398,
      "theoretical_loss": 3.3733393669825036,
      "tokens_seen": 2455699456
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.718554946094193e-05,
      "loss": 2.4715,
      "theoretical_loss": 3.3733321603732787,
      "tokens_seen": 2455764992
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.716899943257045e-05,
      "loss": 2.7495,
      "theoretical_loss": 3.3733249540102195,
      "tokens_seen": 2455830528
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.715244940419898e-05,
      "loss": 2.7267,
      "theoretical_loss": 3.37331774789331,
      "tokens_seen": 2455896064
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.71358993758275e-05,
      "loss": 2.6159,
      "theoretical_loss": 3.373310542022536,
      "tokens_seen": 2455961600
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.711934934745602e-05,
      "loss": 2.7701,
      "theoretical_loss": 3.3733033363978824,
      "tokens_seen": 2456027136
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.710279931908455e-05,
      "loss": 2.7321,
      "theoretical_loss": 3.373296131019334,
      "tokens_seen": 2456092672
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.708624929071307e-05,
      "loss": 2.4888,
      "theoretical_loss": 3.3732889258868757,
      "tokens_seen": 2456158208
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.706969926234158e-05,
      "loss": 2.6541,
      "theoretical_loss": 3.373281721000493,
      "tokens_seen": 2456223744
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.70531492339701e-05,
      "loss": 2.6558,
      "theoretical_loss": 3.3732745163601705,
      "tokens_seen": 2456289280
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.703659920559863e-05,
      "loss": 2.3718,
      "theoretical_loss": 3.373267311965894,
      "tokens_seen": 2456354816
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.702004917722715e-05,
      "loss": 2.3345,
      "theoretical_loss": 3.373260107817647,
      "tokens_seen": 2456420352
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.700349914885568e-05,
      "loss": 2.8162,
      "theoretical_loss": 3.373252903915416,
      "tokens_seen": 2456485888
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.69869491204842e-05,
      "loss": 2.3567,
      "theoretical_loss": 3.3732457002591856,
      "tokens_seen": 2456551424
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.697039909211273e-05,
      "loss": 2.4899,
      "theoretical_loss": 3.3732384968489404,
      "tokens_seen": 2456616960
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.695384906374125e-05,
      "loss": 2.6981,
      "theoretical_loss": 3.3732312936846665,
      "tokens_seen": 2456682496
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.693729903536977e-05,
      "loss": 2.647,
      "theoretical_loss": 3.3732240907663473,
      "tokens_seen": 2456748032
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.69207490069983e-05,
      "loss": 2.5145,
      "theoretical_loss": 3.3732168880939692,
      "tokens_seen": 2456813568
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.690419897862682e-05,
      "loss": 2.5091,
      "theoretical_loss": 3.373209685667517,
      "tokens_seen": 2456879104
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.688764895025535e-05,
      "loss": 2.3848,
      "theoretical_loss": 3.3732024834869754,
      "tokens_seen": 2456944640
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.687109892188387e-05,
      "loss": 2.2383,
      "theoretical_loss": 3.37319528155233,
      "tokens_seen": 2457010176
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.68545488935124e-05,
      "loss": 2.6802,
      "theoretical_loss": 3.3731880798635645,
      "tokens_seen": 2457075712
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.683799886514092e-05,
      "loss": 2.6457,
      "theoretical_loss": 3.3731808784206656,
      "tokens_seen": 2457141248
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.682144883676943e-05,
      "loss": 2.7027,
      "theoretical_loss": 3.3731736772236176,
      "tokens_seen": 2457206784
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.680489880839795e-05,
      "loss": 2.6509,
      "theoretical_loss": 3.3731664762724054,
      "tokens_seen": 2457272320
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.678834878002648e-05,
      "loss": 2.3836,
      "theoretical_loss": 3.3731592755670143,
      "tokens_seen": 2457337856
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.6771798751655e-05,
      "loss": 2.472,
      "theoretical_loss": 3.3731520751074293,
      "tokens_seen": 2457403392
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.675524872328352e-05,
      "loss": 2.3867,
      "theoretical_loss": 3.3731448748936357,
      "tokens_seen": 2457468928
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.673869869491205e-05,
      "loss": 2.3229,
      "theoretical_loss": 3.3731376749256183,
      "tokens_seen": 2457534464
    },
    {
      "epoch": 0.88,
      "objective/train/avg_token_score": 0.009017497301101685,
      "objective/train/avg_weight": 0.9893826246261597,
      "objective/train/docs_used": 1380428,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.4602129459381104,
      "objective/train/std_weight": 0.08208950608968735,
      "objective/train/theoretical_loss": 3.3731304752033617,
      "objective/train/tokens_used": 2478060000,
      "theoretical_loss": 3.3731304752033617,
      "tokens_seen": 2457600000
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.672214866654057e-05,
      "loss": 2.4282,
      "theoretical_loss": 3.3731304752033617,
      "tokens_seen": 2457600000
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.67055986381691e-05,
      "loss": 2.4179,
      "theoretical_loss": 3.3731232757268517,
      "tokens_seen": 2457665536
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.668904860979761e-05,
      "loss": 2.1871,
      "theoretical_loss": 3.373116076496073,
      "tokens_seen": 2457731072
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.667249858142613e-05,
      "loss": 2.5427,
      "theoretical_loss": 3.373108877511011,
      "tokens_seen": 2457796608
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.665594855305465e-05,
      "loss": 2.7266,
      "theoretical_loss": 3.3731016787716506,
      "tokens_seen": 2457862144
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.663939852468318e-05,
      "loss": 2.1827,
      "theoretical_loss": 3.3730944802779765,
      "tokens_seen": 2457927680
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.66228484963117e-05,
      "loss": 2.4199,
      "theoretical_loss": 3.373087282029974,
      "tokens_seen": 2457993216
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.660629846794023e-05,
      "loss": 2.8082,
      "theoretical_loss": 3.3730800840276283,
      "tokens_seen": 2458058752
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.658974843956875e-05,
      "loss": 2.2923,
      "theoretical_loss": 3.3730728862709243,
      "tokens_seen": 2458124288
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.657319841119727e-05,
      "loss": 2.5068,
      "theoretical_loss": 3.373065688759847,
      "tokens_seen": 2458189824
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.65566483828258e-05,
      "loss": 2.2942,
      "theoretical_loss": 3.373058491494382,
      "tokens_seen": 2458255360
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.654009835445432e-05,
      "loss": 2.3411,
      "theoretical_loss": 3.3730512944745135,
      "tokens_seen": 2458320896
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.652354832608285e-05,
      "loss": 2.451,
      "theoretical_loss": 3.3730440977002276,
      "tokens_seen": 2458386432
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.650699829771137e-05,
      "loss": 2.5753,
      "theoretical_loss": 3.3730369011715085,
      "tokens_seen": 2458451968
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.64904482693399e-05,
      "loss": 2.5148,
      "theoretical_loss": 3.3730297048883413,
      "tokens_seen": 2458517504
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.647389824096842e-05,
      "loss": 2.4927,
      "theoretical_loss": 3.373022508850712,
      "tokens_seen": 2458583040
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.645734821259694e-05,
      "loss": 2.3652,
      "theoretical_loss": 3.3730153130586045,
      "tokens_seen": 2458648576
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.644079818422545e-05,
      "loss": 2.6932,
      "theoretical_loss": 3.373008117512005,
      "tokens_seen": 2458714112
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.642424815585398e-05,
      "loss": 2.5015,
      "theoretical_loss": 3.373000922210897,
      "tokens_seen": 2458779648
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.64076981274825e-05,
      "loss": 2.5583,
      "theoretical_loss": 3.3729937271552677,
      "tokens_seen": 2458845184
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.639114809911103e-05,
      "loss": 3.6513,
      "theoretical_loss": 3.3729865323451005,
      "tokens_seen": 2458910720
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.637459807073955e-05,
      "loss": 2.5942,
      "theoretical_loss": 3.3729793377803814,
      "tokens_seen": 2458976256
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.635804804236807e-05,
      "loss": 2.7251,
      "theoretical_loss": 3.372972143461095,
      "tokens_seen": 2459041792
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.63414980139966e-05,
      "loss": 2.4908,
      "theoretical_loss": 3.3729649493872262,
      "tokens_seen": 2459107328
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.632494798562511e-05,
      "loss": 2.8408,
      "theoretical_loss": 3.372957755558761,
      "tokens_seen": 2459172864
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.630839795725363e-05,
      "loss": 2.4086,
      "theoretical_loss": 3.3729505619756837,
      "tokens_seen": 2459238400
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.629184792888216e-05,
      "loss": 2.4613,
      "theoretical_loss": 3.372943368637979,
      "tokens_seen": 2459303936
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.627529790051068e-05,
      "loss": 2.6711,
      "theoretical_loss": 3.3729361755456333,
      "tokens_seen": 2459369472
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.62587478721392e-05,
      "loss": 2.7296,
      "theoretical_loss": 3.372928982698631,
      "tokens_seen": 2459435008
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.624219784376773e-05,
      "loss": 2.5633,
      "theoretical_loss": 3.372921790096957,
      "tokens_seen": 2459500544
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.622564781539625e-05,
      "loss": 2.2927,
      "theoretical_loss": 3.3729145977405963,
      "tokens_seen": 2459566080
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.620909778702478e-05,
      "loss": 2.6254,
      "theoretical_loss": 3.372907405629535,
      "tokens_seen": 2459631616
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.61925477586533e-05,
      "loss": 2.544,
      "theoretical_loss": 3.3729002137637565,
      "tokens_seen": 2459697152
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.617599773028182e-05,
      "loss": 2.5045,
      "theoretical_loss": 3.3728930221432476,
      "tokens_seen": 2459762688
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.615944770191035e-05,
      "loss": 2.6365,
      "theoretical_loss": 3.3728858307679923,
      "tokens_seen": 2459828224
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.614289767353887e-05,
      "loss": 2.3841,
      "theoretical_loss": 3.3728786396379764,
      "tokens_seen": 2459893760
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.61263476451674e-05,
      "loss": 2.5146,
      "theoretical_loss": 3.3728714487531843,
      "tokens_seen": 2459959296
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.610979761679592e-05,
      "loss": 2.8448,
      "theoretical_loss": 3.372864258113602,
      "tokens_seen": 2460024832
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.609324758842444e-05,
      "loss": 2.4752,
      "theoretical_loss": 3.3728570677192136,
      "tokens_seen": 2460090368
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.607669756005295e-05,
      "loss": 2.5847,
      "theoretical_loss": 3.3728498775700047,
      "tokens_seen": 2460155904
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.606014753168148e-05,
      "loss": 2.5027,
      "theoretical_loss": 3.3728426876659605,
      "tokens_seen": 2460221440
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.604359750331e-05,
      "loss": 2.7225,
      "theoretical_loss": 3.372835498007066,
      "tokens_seen": 2460286976
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.602704747493853e-05,
      "loss": 2.3793,
      "theoretical_loss": 3.3728283085933066,
      "tokens_seen": 2460352512
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.601049744656705e-05,
      "loss": 2.6307,
      "theoretical_loss": 3.3728211194246667,
      "tokens_seen": 2460418048
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.599394741819557e-05,
      "loss": 2.546,
      "theoretical_loss": 3.3728139305011324,
      "tokens_seen": 2460483584
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.59773973898241e-05,
      "loss": 2.5277,
      "theoretical_loss": 3.3728067418226875,
      "tokens_seen": 2460549120
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.596084736145261e-05,
      "loss": 2.4693,
      "theoretical_loss": 3.372799553389319,
      "tokens_seen": 2460614656
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.594429733308113e-05,
      "loss": 2.5083,
      "theoretical_loss": 3.37279236520101,
      "tokens_seen": 2460680192
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.592774730470966e-05,
      "loss": 2.411,
      "theoretical_loss": 3.3727851772577466,
      "tokens_seen": 2460745728
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.591119727633818e-05,
      "loss": 2.4676,
      "theoretical_loss": 3.372777989559514,
      "tokens_seen": 2460811264
    },
    {
      "epoch": 0.88,
      "objective/train/avg_token_score": 0.014037031680345535,
      "objective/train/avg_weight": 0.9834821224212646,
      "objective/train/docs_used": 1382125,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.952324867248535,
      "objective/train/std_weight": 0.08315713703632355,
      "objective/train/theoretical_loss": 3.372770802106297,
      "objective/train/tokens_used": 2481336800,
      "theoretical_loss": 3.372770802106297,
      "tokens_seen": 2460876800
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.58946472479667e-05,
      "loss": 2.6805,
      "theoretical_loss": 3.372770802106297,
      "tokens_seen": 2460876800
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.587809721959523e-05,
      "loss": 2.2967,
      "theoretical_loss": 3.3727636148980813,
      "tokens_seen": 2460942336
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.586154719122375e-05,
      "loss": 2.0272,
      "theoretical_loss": 3.3727564279348514,
      "tokens_seen": 2461007872
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.584499716285228e-05,
      "loss": 2.9148,
      "theoretical_loss": 3.3727492412165927,
      "tokens_seen": 2461073408
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.58284471344808e-05,
      "loss": 2.5297,
      "theoretical_loss": 3.37274205474329,
      "tokens_seen": 2461138944
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.581189710610932e-05,
      "loss": 2.8646,
      "theoretical_loss": 3.3727348685149288,
      "tokens_seen": 2461204480
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.579534707773785e-05,
      "loss": 2.6549,
      "theoretical_loss": 3.3727276825314942,
      "tokens_seen": 2461270016
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.577879704936637e-05,
      "loss": 2.4684,
      "theoretical_loss": 3.3727204967929714,
      "tokens_seen": 2461335552
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.57622470209949e-05,
      "loss": 2.4852,
      "theoretical_loss": 3.372713311299345,
      "tokens_seen": 2461401088
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.574569699262342e-05,
      "loss": 2.3469,
      "theoretical_loss": 3.3727061260506006,
      "tokens_seen": 2461466624
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.572914696425194e-05,
      "loss": 2.5125,
      "theoretical_loss": 3.3726989410467234,
      "tokens_seen": 2461532160
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.571259693588045e-05,
      "loss": 2.5665,
      "theoretical_loss": 3.372691756287699,
      "tokens_seen": 2461597696
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.569604690750898e-05,
      "loss": 2.6627,
      "theoretical_loss": 3.3726845717735108,
      "tokens_seen": 2461663232
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.56794968791375e-05,
      "loss": 2.4911,
      "theoretical_loss": 3.3726773875041456,
      "tokens_seen": 2461728768
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.566294685076603e-05,
      "loss": 2.7961,
      "theoretical_loss": 3.372670203479588,
      "tokens_seen": 2461794304
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.564639682239455e-05,
      "loss": 2.5341,
      "theoretical_loss": 3.3726630196998233,
      "tokens_seen": 2461859840
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.562984679402307e-05,
      "loss": 2.5772,
      "theoretical_loss": 3.372655836164836,
      "tokens_seen": 2461925376
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.56132967656516e-05,
      "loss": 2.7075,
      "theoretical_loss": 3.372648652874612,
      "tokens_seen": 2461990912
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.559674673728012e-05,
      "loss": 2.4332,
      "theoretical_loss": 3.3726414698291363,
      "tokens_seen": 2462056448
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.558019670890863e-05,
      "loss": 2.443,
      "theoretical_loss": 3.372634287028394,
      "tokens_seen": 2462121984
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.556364668053716e-05,
      "loss": 2.6536,
      "theoretical_loss": 3.37262710447237,
      "tokens_seen": 2462187520
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.554709665216568e-05,
      "loss": 2.573,
      "theoretical_loss": 3.3726199221610496,
      "tokens_seen": 2462253056
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.55305466237942e-05,
      "loss": 2.2015,
      "theoretical_loss": 3.3726127400944184,
      "tokens_seen": 2462318592
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.551399659542273e-05,
      "loss": 2.6355,
      "theoretical_loss": 3.3726055582724603,
      "tokens_seen": 2462384128
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.549744656705127e-05,
      "loss": 2.4809,
      "theoretical_loss": 3.372598376695162,
      "tokens_seen": 2462449664
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.548089653867979e-05,
      "loss": 2.5507,
      "theoretical_loss": 3.3725911953625074,
      "tokens_seen": 2462515200
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.54643465103083e-05,
      "loss": 2.7102,
      "theoretical_loss": 3.3725840142744827,
      "tokens_seen": 2462580736
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.544779648193682e-05,
      "loss": 2.1436,
      "theoretical_loss": 3.372576833431072,
      "tokens_seen": 2462646272
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.543124645356535e-05,
      "loss": 2.6347,
      "theoretical_loss": 3.3725696528322615,
      "tokens_seen": 2462711808
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.541469642519387e-05,
      "loss": 2.3566,
      "theoretical_loss": 3.3725624724780356,
      "tokens_seen": 2462777344
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.53981463968224e-05,
      "loss": 2.4979,
      "theoretical_loss": 3.37255529236838,
      "tokens_seen": 2462842880
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.538159636845092e-05,
      "loss": 2.6301,
      "theoretical_loss": 3.3725481125032792,
      "tokens_seen": 2462908416
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.536504634007944e-05,
      "loss": 2.3837,
      "theoretical_loss": 3.372540932882719,
      "tokens_seen": 2462973952
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.534849631170796e-05,
      "loss": 2.5002,
      "theoretical_loss": 3.3725337535066844,
      "tokens_seen": 2463039488
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.533194628333648e-05,
      "loss": 2.4445,
      "theoretical_loss": 3.37252657437516,
      "tokens_seen": 2463105024
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.5315396254965e-05,
      "loss": 2.468,
      "theoretical_loss": 3.3725193954881316,
      "tokens_seen": 2463170560
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.529884622659353e-05,
      "loss": 2.701,
      "theoretical_loss": 3.3725122168455846,
      "tokens_seen": 2463236096
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.528229619822205e-05,
      "loss": 2.6178,
      "theoretical_loss": 3.3725050384475033,
      "tokens_seen": 2463301632
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.526574616985058e-05,
      "loss": 2.5807,
      "theoretical_loss": 3.372497860293874,
      "tokens_seen": 2463367168
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.52491961414791e-05,
      "loss": 2.1341,
      "theoretical_loss": 3.3724906823846807,
      "tokens_seen": 2463432704
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.523264611310762e-05,
      "loss": 2.5977,
      "theoretical_loss": 3.3724835047199093,
      "tokens_seen": 2463498240
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.521609608473613e-05,
      "loss": 2.7292,
      "theoretical_loss": 3.3724763272995446,
      "tokens_seen": 2463563776
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.519954605636466e-05,
      "loss": 2.5351,
      "theoretical_loss": 3.372469150123572,
      "tokens_seen": 2463629312
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.51829960279932e-05,
      "loss": 2.1969,
      "theoretical_loss": 3.3724619731919767,
      "tokens_seen": 2463694848
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.516644599962172e-05,
      "loss": 2.3836,
      "theoretical_loss": 3.372454796504744,
      "tokens_seen": 2463760384
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.514989597125024e-05,
      "loss": 2.5065,
      "theoretical_loss": 3.3724476200618585,
      "tokens_seen": 2463825920
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.513334594287877e-05,
      "loss": 2.6866,
      "theoretical_loss": 3.372440443863306,
      "tokens_seen": 2463891456
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.511679591450729e-05,
      "loss": 2.4549,
      "theoretical_loss": 3.3724332679090714,
      "tokens_seen": 2463956992
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.51002458861358e-05,
      "loss": 2.4812,
      "theoretical_loss": 3.37242609219914,
      "tokens_seen": 2464022528
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.508369585776433e-05,
      "loss": 2.5197,
      "theoretical_loss": 3.372418916733497,
      "tokens_seen": 2464088064
    },
    {
      "epoch": 0.88,
      "objective/train/avg_token_score": 0.014757266268134117,
      "objective/train/avg_weight": 0.9826393127441406,
      "objective/train/docs_used": 1384145,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8568100929260254,
      "objective/train/std_weight": 0.07782649248838425,
      "objective/train/theoretical_loss": 3.3724117415121273,
      "objective/train/tokens_used": 2484613600,
      "theoretical_loss": 3.3724117415121273,
      "tokens_seen": 2464153600
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.506714582939285e-05,
      "loss": 2.2638,
      "theoretical_loss": 3.3724117415121273,
      "tokens_seen": 2464153600
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.505059580102137e-05,
      "loss": 2.5617,
      "theoretical_loss": 3.372404566535016,
      "tokens_seen": 2464219136
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.50340457726499e-05,
      "loss": 2.7504,
      "theoretical_loss": 3.3723973918021493,
      "tokens_seen": 2464284672
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.501749574427842e-05,
      "loss": 2.491,
      "theoretical_loss": 3.3723902173135114,
      "tokens_seen": 2464350208
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.500094571590695e-05,
      "loss": 2.6284,
      "theoretical_loss": 3.3723830430690875,
      "tokens_seen": 2464415744
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.498439568753547e-05,
      "loss": 2.4455,
      "theoretical_loss": 3.3723758690688634,
      "tokens_seen": 2464481280
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.496784565916398e-05,
      "loss": 2.3014,
      "theoretical_loss": 3.372368695312824,
      "tokens_seen": 2464546816
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.49512956307925e-05,
      "loss": 2.6944,
      "theoretical_loss": 3.372361521800954,
      "tokens_seen": 2464612352
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.493474560242103e-05,
      "loss": 2.4594,
      "theoretical_loss": 3.3723543485332392,
      "tokens_seen": 2464677888
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.491819557404955e-05,
      "loss": 2.5039,
      "theoretical_loss": 3.372347175509665,
      "tokens_seen": 2464743424
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.490164554567808e-05,
      "loss": 2.6262,
      "theoretical_loss": 3.3723400027302164,
      "tokens_seen": 2464808960
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.48850955173066e-05,
      "loss": 2.4771,
      "theoretical_loss": 3.372332830194878,
      "tokens_seen": 2464874496
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.486854548893512e-05,
      "loss": 2.6865,
      "theoretical_loss": 3.3723256579036356,
      "tokens_seen": 2464940032
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.485199546056365e-05,
      "loss": 1.2136,
      "theoretical_loss": 3.372318485856474,
      "tokens_seen": 2465005568
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.483544543219217e-05,
      "loss": 2.4651,
      "theoretical_loss": 3.3723113140533787,
      "tokens_seen": 2465071104
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.48188954038207e-05,
      "loss": 2.5277,
      "theoretical_loss": 3.372304142494335,
      "tokens_seen": 2465136640
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.480234537544922e-05,
      "loss": 2.4087,
      "theoretical_loss": 3.372296971179328,
      "tokens_seen": 2465202176
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.478579534707774e-05,
      "loss": 2.6316,
      "theoretical_loss": 3.372289800108343,
      "tokens_seen": 2465267712
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.476924531870627e-05,
      "loss": 2.2939,
      "theoretical_loss": 3.372282629281365,
      "tokens_seen": 2465333248
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.475269529033479e-05,
      "loss": 2.2644,
      "theoretical_loss": 3.372275458698379,
      "tokens_seen": 2465398784
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.473614526196332e-05,
      "loss": 2.4863,
      "theoretical_loss": 3.372268288359371,
      "tokens_seen": 2465464320
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.471959523359183e-05,
      "loss": 2.5411,
      "theoretical_loss": 3.3722611182643254,
      "tokens_seen": 2465529856
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.470304520522035e-05,
      "loss": 2.209,
      "theoretical_loss": 3.3722539484132277,
      "tokens_seen": 2465595392
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.468649517684887e-05,
      "loss": 2.6252,
      "theoretical_loss": 3.3722467788060633,
      "tokens_seen": 2465660928
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.46699451484774e-05,
      "loss": 2.6512,
      "theoretical_loss": 3.3722396094428175,
      "tokens_seen": 2465726464
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.465339512010592e-05,
      "loss": 2.1864,
      "theoretical_loss": 3.372232440323475,
      "tokens_seen": 2465792000
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.463684509173445e-05,
      "loss": 2.4739,
      "theoretical_loss": 3.372225271448021,
      "tokens_seen": 2465857536
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.462029506336297e-05,
      "loss": 2.3011,
      "theoretical_loss": 3.3722181028164417,
      "tokens_seen": 2465923072
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.460374503499148e-05,
      "loss": 2.5995,
      "theoretical_loss": 3.3722109344287214,
      "tokens_seen": 2465988608
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.458719500662e-05,
      "loss": 2.493,
      "theoretical_loss": 3.3722037662848456,
      "tokens_seen": 2466054144
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.457064497824853e-05,
      "loss": 2.3942,
      "theoretical_loss": 3.3721965983847992,
      "tokens_seen": 2466119680
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.455409494987705e-05,
      "loss": 2.5401,
      "theoretical_loss": 3.372189430728568,
      "tokens_seen": 2466185216
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.453754492150558e-05,
      "loss": 2.4817,
      "theoretical_loss": 3.3721822633161374,
      "tokens_seen": 2466250752
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.45209948931341e-05,
      "loss": 2.4351,
      "theoretical_loss": 3.3721750961474917,
      "tokens_seen": 2466316288
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.450444486476262e-05,
      "loss": 2.786,
      "theoretical_loss": 3.3721679292226163,
      "tokens_seen": 2466381824
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.448789483639115e-05,
      "loss": 2.3426,
      "theoretical_loss": 3.3721607625414975,
      "tokens_seen": 2466447360
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.447134480801967e-05,
      "loss": 2.5016,
      "theoretical_loss": 3.3721535961041194,
      "tokens_seen": 2466512896
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.44547947796482e-05,
      "loss": 2.5782,
      "theoretical_loss": 3.3721464299104675,
      "tokens_seen": 2466578432
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.443824475127672e-05,
      "loss": 2.5986,
      "theoretical_loss": 3.3721392639605274,
      "tokens_seen": 2466643968
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.442169472290524e-05,
      "loss": 2.6826,
      "theoretical_loss": 3.372132098254284,
      "tokens_seen": 2466709504
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.440514469453377e-05,
      "loss": 2.5724,
      "theoretical_loss": 3.3721249327917224,
      "tokens_seen": 2466775040
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.438859466616229e-05,
      "loss": 2.655,
      "theoretical_loss": 3.3721177675728287,
      "tokens_seen": 2466840576
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.437204463779082e-05,
      "loss": 2.2807,
      "theoretical_loss": 3.372110602597587,
      "tokens_seen": 2466906112
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.435549460941933e-05,
      "loss": 2.4418,
      "theoretical_loss": 3.372103437865983,
      "tokens_seen": 2466971648
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.433894458104785e-05,
      "loss": 2.5914,
      "theoretical_loss": 3.372096273378002,
      "tokens_seen": 2467037184
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.432239455267638e-05,
      "loss": 2.5234,
      "theoretical_loss": 3.3720891091336296,
      "tokens_seen": 2467102720
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.43058445243049e-05,
      "loss": 2.5756,
      "theoretical_loss": 3.3720819451328508,
      "tokens_seen": 2467168256
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.428929449593342e-05,
      "loss": 2.5226,
      "theoretical_loss": 3.37207478137565,
      "tokens_seen": 2467233792
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.427274446756195e-05,
      "loss": 0.9565,
      "theoretical_loss": 3.372067617862014,
      "tokens_seen": 2467299328
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.425619443919047e-05,
      "loss": 2.5312,
      "theoretical_loss": 3.3720604545919266,
      "tokens_seen": 2467364864
    },
    {
      "epoch": 0.88,
      "objective/train/avg_token_score": 0.016153153032064438,
      "objective/train/avg_weight": 0.9809982180595398,
      "objective/train/docs_used": 1386568,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2636666297912598,
      "objective/train/std_weight": 0.08673574775457382,
      "objective/train/theoretical_loss": 3.372053291565374,
      "objective/train/tokens_used": 2487890400,
      "theoretical_loss": 3.372053291565374,
      "tokens_seen": 2467430400
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.423964441081898e-05,
      "loss": 2.5717,
      "theoretical_loss": 3.372053291565374,
      "tokens_seen": 2467430400
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.42230943824475e-05,
      "loss": 2.744,
      "theoretical_loss": 3.3720461287823413,
      "tokens_seen": 2467495936
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.420654435407603e-05,
      "loss": 2.2839,
      "theoretical_loss": 3.372038966242813,
      "tokens_seen": 2467561472
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.418999432570455e-05,
      "loss": 2.6776,
      "theoretical_loss": 3.3720318039467756,
      "tokens_seen": 2467627008
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.417344429733308e-05,
      "loss": 2.6499,
      "theoretical_loss": 3.372024641894213,
      "tokens_seen": 2467692544
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.41568942689616e-05,
      "loss": 2.4824,
      "theoretical_loss": 3.3720174800851117,
      "tokens_seen": 2467758080
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.414034424059013e-05,
      "loss": 2.211,
      "theoretical_loss": 3.3720103185194565,
      "tokens_seen": 2467823616
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.412379421221865e-05,
      "loss": 2.9591,
      "theoretical_loss": 3.3720031571972324,
      "tokens_seen": 2467889152
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.410724418384717e-05,
      "loss": 2.8191,
      "theoretical_loss": 3.3719959961184247,
      "tokens_seen": 2467954688
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.40906941554757e-05,
      "loss": 2.5624,
      "theoretical_loss": 3.371988835283019,
      "tokens_seen": 2468020224
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.407414412710422e-05,
      "loss": 2.601,
      "theoretical_loss": 3.3719816746910003,
      "tokens_seen": 2468085760
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.405759409873275e-05,
      "loss": 2.5517,
      "theoretical_loss": 3.371974514342354,
      "tokens_seen": 2468151296
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.404104407036127e-05,
      "loss": 2.5741,
      "theoretical_loss": 3.371967354237065,
      "tokens_seen": 2468216832
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.40244940419898e-05,
      "loss": 2.4186,
      "theoretical_loss": 3.371960194375119,
      "tokens_seen": 2468282368
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.400794401361832e-05,
      "loss": 2.6719,
      "theoretical_loss": 3.371953034756501,
      "tokens_seen": 2468347904
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.399139398524683e-05,
      "loss": 2.3735,
      "theoretical_loss": 3.371945875381197,
      "tokens_seen": 2468413440
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.397484395687535e-05,
      "loss": 2.3647,
      "theoretical_loss": 3.371938716249191,
      "tokens_seen": 2468478976
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.395829392850388e-05,
      "loss": 2.376,
      "theoretical_loss": 3.37193155736047,
      "tokens_seen": 2468544512
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.39417439001324e-05,
      "loss": 2.5706,
      "theoretical_loss": 3.3719243987150174,
      "tokens_seen": 2468610048
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.392519387176092e-05,
      "loss": 3.3951,
      "theoretical_loss": 3.3719172403128193,
      "tokens_seen": 2468675584
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.390864384338945e-05,
      "loss": 2.8286,
      "theoretical_loss": 3.371910082153861,
      "tokens_seen": 2468741120
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.389209381501797e-05,
      "loss": 2.5742,
      "theoretical_loss": 3.3719029242381278,
      "tokens_seen": 2468806656
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.38755437866465e-05,
      "loss": 2.3748,
      "theoretical_loss": 3.3718957665656046,
      "tokens_seen": 2468872192
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.3858993758275e-05,
      "loss": 2.6144,
      "theoretical_loss": 3.3718886091362776,
      "tokens_seen": 2468937728
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.384244372990353e-05,
      "loss": 2.7038,
      "theoretical_loss": 3.3718814519501312,
      "tokens_seen": 2469003264
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.382589370153205e-05,
      "loss": 2.7376,
      "theoretical_loss": 3.371874295007151,
      "tokens_seen": 2469068800
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.380934367316058e-05,
      "loss": 1.2638,
      "theoretical_loss": 3.371867138307322,
      "tokens_seen": 2469134336
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.37927936447891e-05,
      "loss": 2.5163,
      "theoretical_loss": 3.3718599818506303,
      "tokens_seen": 2469199872
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.377624361641763e-05,
      "loss": 2.6245,
      "theoretical_loss": 3.3718528256370606,
      "tokens_seen": 2469265408
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.375969358804615e-05,
      "loss": 2.2818,
      "theoretical_loss": 3.3718456696665977,
      "tokens_seen": 2469330944
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.374314355967467e-05,
      "loss": 2.7357,
      "theoretical_loss": 3.371838513939228,
      "tokens_seen": 2469396480
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.37265935313032e-05,
      "loss": 2.5681,
      "theoretical_loss": 3.3718313584549358,
      "tokens_seen": 2469462016
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.371004350293172e-05,
      "loss": 2.6568,
      "theoretical_loss": 3.371824203213707,
      "tokens_seen": 2469527552
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.369349347456025e-05,
      "loss": 1.9733,
      "theoretical_loss": 3.3718170482155267,
      "tokens_seen": 2469593088
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.367694344618877e-05,
      "loss": 2.5779,
      "theoretical_loss": 3.3718098934603797,
      "tokens_seen": 2469658624
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.36603934178173e-05,
      "loss": 2.5402,
      "theoretical_loss": 3.3718027389482526,
      "tokens_seen": 2469724160
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.364384338944582e-05,
      "loss": 2.5266,
      "theoretical_loss": 3.371795584679129,
      "tokens_seen": 2469789696
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.362729336107433e-05,
      "loss": 2.5009,
      "theoretical_loss": 3.3717884306529955,
      "tokens_seen": 2469855232
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.361074333270285e-05,
      "loss": 2.5142,
      "theoretical_loss": 3.371781276869837,
      "tokens_seen": 2469920768
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.359419330433138e-05,
      "loss": 2.2835,
      "theoretical_loss": 3.371774123329639,
      "tokens_seen": 2469986304
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.35776432759599e-05,
      "loss": 2.3818,
      "theoretical_loss": 3.3717669700323865,
      "tokens_seen": 2470051840
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.356109324758842e-05,
      "loss": 2.6387,
      "theoretical_loss": 3.3717598169780647,
      "tokens_seen": 2470117376
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.354454321921695e-05,
      "loss": 2.6077,
      "theoretical_loss": 3.3717526641666593,
      "tokens_seen": 2470182912
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.352799319084547e-05,
      "loss": 2.4841,
      "theoretical_loss": 3.371745511598155,
      "tokens_seen": 2470248448
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.3511443162474e-05,
      "loss": 2.6188,
      "theoretical_loss": 3.3717383592725376,
      "tokens_seen": 2470313984
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.349489313410251e-05,
      "loss": 2.7124,
      "theoretical_loss": 3.371731207189793,
      "tokens_seen": 2470379520
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.347834310573103e-05,
      "loss": 2.6555,
      "theoretical_loss": 3.371724055349905,
      "tokens_seen": 2470445056
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.346179307735955e-05,
      "loss": 2.7163,
      "theoretical_loss": 3.3717169037528603,
      "tokens_seen": 2470510592
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.344524304898808e-05,
      "loss": 2.6187,
      "theoretical_loss": 3.371709752398643,
      "tokens_seen": 2470576128
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.34286930206166e-05,
      "loss": 2.9887,
      "theoretical_loss": 3.3717026012872395,
      "tokens_seen": 2470641664
    },
    {
      "epoch": 0.88,
      "objective/train/avg_token_score": 0.02585647627711296,
      "objective/train/avg_weight": 0.9695557951927185,
      "objective/train/docs_used": 1387925,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 6.326072692871094,
      "objective/train/std_weight": 0.13805271685123444,
      "objective/train/theoretical_loss": 3.3716954504186347,
      "objective/train/tokens_used": 2491167200,
      "theoretical_loss": 3.3716954504186347,
      "tokens_seen": 2470707200
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.341214299224513e-05,
      "loss": 2.3286,
      "theoretical_loss": 3.3716954504186347,
      "tokens_seen": 2470707200
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.339559296387365e-05,
      "loss": 2.5055,
      "theoretical_loss": 3.3716882997928135,
      "tokens_seen": 2470772736
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.337904293550217e-05,
      "loss": 2.6273,
      "theoretical_loss": 3.371681149409762,
      "tokens_seen": 2470838272
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.33624929071307e-05,
      "loss": 2.5748,
      "theoretical_loss": 3.371673999269465,
      "tokens_seen": 2470903808
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.334594287875922e-05,
      "loss": 2.2876,
      "theoretical_loss": 3.371666849371908,
      "tokens_seen": 2470969344
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.332939285038775e-05,
      "loss": 2.6458,
      "theoretical_loss": 3.3716596997170765,
      "tokens_seen": 2471034880
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.331284282201627e-05,
      "loss": 2.6611,
      "theoretical_loss": 3.371652550304955,
      "tokens_seen": 2471100416
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.32962927936448e-05,
      "loss": 2.7242,
      "theoretical_loss": 3.3716454011355297,
      "tokens_seen": 2471165952
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.327974276527332e-05,
      "loss": 2.3985,
      "theoretical_loss": 3.3716382522087858,
      "tokens_seen": 2471231488
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.326319273690184e-05,
      "loss": 2.291,
      "theoretical_loss": 3.371631103524708,
      "tokens_seen": 2471297024
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.324664270853035e-05,
      "loss": 2.3885,
      "theoretical_loss": 3.3716239550832827,
      "tokens_seen": 2471362560
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.323009268015888e-05,
      "loss": 2.6236,
      "theoretical_loss": 3.3716168068844943,
      "tokens_seen": 2471428096
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.32135426517874e-05,
      "loss": 2.4864,
      "theoretical_loss": 3.3716096589283286,
      "tokens_seen": 2471493632
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.319699262341593e-05,
      "loss": 2.6516,
      "theoretical_loss": 3.3716025112147707,
      "tokens_seen": 2471559168
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.318044259504445e-05,
      "loss": 2.2793,
      "theoretical_loss": 3.371595363743806,
      "tokens_seen": 2471624704
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.316389256667297e-05,
      "loss": 2.3891,
      "theoretical_loss": 3.37158821651542,
      "tokens_seen": 2471690240
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.31473425383015e-05,
      "loss": 2.5907,
      "theoretical_loss": 3.3715810695295976,
      "tokens_seen": 2471755776
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.313079250993001e-05,
      "loss": 2.4799,
      "theoretical_loss": 3.3715739227863244,
      "tokens_seen": 2471821312
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.311424248155853e-05,
      "loss": 2.5685,
      "theoretical_loss": 3.371566776285586,
      "tokens_seen": 2471886848
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.309769245318706e-05,
      "loss": 2.3484,
      "theoretical_loss": 3.3715596300273676,
      "tokens_seen": 2471952384
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.308114242481558e-05,
      "loss": 2.5079,
      "theoretical_loss": 3.371552484011654,
      "tokens_seen": 2472017920
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.30645923964441e-05,
      "loss": 2.7883,
      "theoretical_loss": 3.3715453382384317,
      "tokens_seen": 2472083456
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.304804236807263e-05,
      "loss": 2.7087,
      "theoretical_loss": 3.3715381927076846,
      "tokens_seen": 2472148992
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.303149233970115e-05,
      "loss": 2.3679,
      "theoretical_loss": 3.371531047419399,
      "tokens_seen": 2472214528
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.301494231132968e-05,
      "loss": 2.438,
      "theoretical_loss": 3.3715239023735606,
      "tokens_seen": 2472280064
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.29983922829582e-05,
      "loss": 2.3702,
      "theoretical_loss": 3.3715167575701535,
      "tokens_seen": 2472345600
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.298184225458672e-05,
      "loss": 2.4428,
      "theoretical_loss": 3.371509613009164,
      "tokens_seen": 2472411136
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.296529222621525e-05,
      "loss": 2.5557,
      "theoretical_loss": 3.371502468690577,
      "tokens_seen": 2472476672
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.294874219784377e-05,
      "loss": 2.592,
      "theoretical_loss": 3.371495324614378,
      "tokens_seen": 2472542208
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.29321921694723e-05,
      "loss": 2.5247,
      "theoretical_loss": 3.3714881807805526,
      "tokens_seen": 2472607744
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.291564214110082e-05,
      "loss": 2.6957,
      "theoretical_loss": 3.3714810371890858,
      "tokens_seen": 2472673280
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.289909211272934e-05,
      "loss": 2.2515,
      "theoretical_loss": 3.371473893839963,
      "tokens_seen": 2472738816
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.288254208435785e-05,
      "loss": 2.261,
      "theoretical_loss": 3.3714667507331697,
      "tokens_seen": 2472804352
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.286599205598638e-05,
      "loss": 2.7046,
      "theoretical_loss": 3.371459607868691,
      "tokens_seen": 2472869888
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.28494420276149e-05,
      "loss": 2.3186,
      "theoretical_loss": 3.3714524652465125,
      "tokens_seen": 2472935424
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.283289199924343e-05,
      "loss": 2.5488,
      "theoretical_loss": 3.3714453228666197,
      "tokens_seen": 2473000960
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.281634197087195e-05,
      "loss": 2.2432,
      "theoretical_loss": 3.3714381807289975,
      "tokens_seen": 2473066496
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.279979194250047e-05,
      "loss": 2.4752,
      "theoretical_loss": 3.3714310388336317,
      "tokens_seen": 2473132032
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.2783241914129e-05,
      "loss": 2.2345,
      "theoretical_loss": 3.3714238971805077,
      "tokens_seen": 2473197568
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.276669188575751e-05,
      "loss": 2.702,
      "theoretical_loss": 3.3714167557696104,
      "tokens_seen": 2473263104
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.275014185738603e-05,
      "loss": 2.5954,
      "theoretical_loss": 3.3714096146009256,
      "tokens_seen": 2473328640
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.273359182901456e-05,
      "loss": 2.6611,
      "theoretical_loss": 3.371402473674438,
      "tokens_seen": 2473394176
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.271704180064308e-05,
      "loss": 2.6837,
      "theoretical_loss": 3.3713953329901343,
      "tokens_seen": 2473459712
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.27004917722716e-05,
      "loss": 2.6151,
      "theoretical_loss": 3.3713881925479985,
      "tokens_seen": 2473525248
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.268394174390013e-05,
      "loss": 2.5361,
      "theoretical_loss": 3.3713810523480165,
      "tokens_seen": 2473590784
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.266739171552865e-05,
      "loss": 2.5035,
      "theoretical_loss": 3.371373912390174,
      "tokens_seen": 2473656320
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.265084168715718e-05,
      "loss": 2.6161,
      "theoretical_loss": 3.3713667726744556,
      "tokens_seen": 2473721856
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.26342916587857e-05,
      "loss": 2.0726,
      "theoretical_loss": 3.3713596332008473,
      "tokens_seen": 2473787392
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.261774163041422e-05,
      "loss": 2.5301,
      "theoretical_loss": 3.3713524939693347,
      "tokens_seen": 2473852928
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.260119160204275e-05,
      "loss": 2.4288,
      "theoretical_loss": 3.371345354979902,
      "tokens_seen": 2473918464
    },
    {
      "epoch": 0.88,
      "objective/train/avg_token_score": 0.007353783585131168,
      "objective/train/avg_weight": 0.9913420677185059,
      "objective/train/docs_used": 1390126,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8221664428710938,
      "objective/train/std_weight": 0.05953098461031914,
      "objective/train/theoretical_loss": 3.371338216232536,
      "objective/train/tokens_used": 2494444000,
      "theoretical_loss": 3.371338216232536,
      "tokens_seen": 2473984000
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.258464157367127e-05,
      "loss": 2.5785,
      "theoretical_loss": 3.371338216232536,
      "tokens_seen": 2473984000
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.25680915452998e-05,
      "loss": 2.7824,
      "theoretical_loss": 3.371331077727221,
      "tokens_seen": 2474049536
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.255154151692832e-05,
      "loss": 2.785,
      "theoretical_loss": 3.371323939463943,
      "tokens_seen": 2474115072
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.253499148855684e-05,
      "loss": 2.427,
      "theoretical_loss": 3.371316801442687,
      "tokens_seen": 2474180608
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.251844146018535e-05,
      "loss": 2.3558,
      "theoretical_loss": 3.371309663663439,
      "tokens_seen": 2474246144
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.250189143181388e-05,
      "loss": 2.701,
      "theoretical_loss": 3.371302526126184,
      "tokens_seen": 2474311680
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.24853414034424e-05,
      "loss": 2.5637,
      "theoretical_loss": 3.371295388830907,
      "tokens_seen": 2474377216
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.246879137507093e-05,
      "loss": 2.4394,
      "theoretical_loss": 3.3712882517775937,
      "tokens_seen": 2474442752
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.245224134669945e-05,
      "loss": 2.4826,
      "theoretical_loss": 3.3712811149662296,
      "tokens_seen": 2474508288
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.243569131832797e-05,
      "loss": 2.6892,
      "theoretical_loss": 3.3712739783968004,
      "tokens_seen": 2474573824
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.24191412899565e-05,
      "loss": 2.5911,
      "theoretical_loss": 3.371266842069291,
      "tokens_seen": 2474639360
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.240259126158502e-05,
      "loss": 2.3974,
      "theoretical_loss": 3.3712597059836864,
      "tokens_seen": 2474704896
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.238604123321353e-05,
      "loss": 2.6524,
      "theoretical_loss": 3.371252570139973,
      "tokens_seen": 2474770432
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.236949120484206e-05,
      "loss": 2.2384,
      "theoretical_loss": 3.3712454345381353,
      "tokens_seen": 2474835968
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.235294117647058e-05,
      "loss": 2.6049,
      "theoretical_loss": 3.3712382991781595,
      "tokens_seen": 2474901504
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.23363911480991e-05,
      "loss": 2.6194,
      "theoretical_loss": 3.3712311640600303,
      "tokens_seen": 2474967040
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.231984111972763e-05,
      "loss": 2.6424,
      "theoretical_loss": 3.3712240291837334,
      "tokens_seen": 2475032576
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.230329109135615e-05,
      "loss": 2.4714,
      "theoretical_loss": 3.3712168945492547,
      "tokens_seen": 2475098112
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.228674106298468e-05,
      "loss": 2.4162,
      "theoretical_loss": 3.3712097601565785,
      "tokens_seen": 2475163648
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.22701910346132e-05,
      "loss": 2.4394,
      "theoretical_loss": 3.371202626005691,
      "tokens_seen": 2475229184
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.225364100624172e-05,
      "loss": 2.4527,
      "theoretical_loss": 3.3711954920965774,
      "tokens_seen": 2475294720
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.223709097787025e-05,
      "loss": 2.7728,
      "theoretical_loss": 3.3711883584292233,
      "tokens_seen": 2475360256
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.222054094949877e-05,
      "loss": 2.5255,
      "theoretical_loss": 3.3711812250036135,
      "tokens_seen": 2475425792
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.22039909211273e-05,
      "loss": 2.1598,
      "theoretical_loss": 3.371174091819734,
      "tokens_seen": 2475491328
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.218744089275582e-05,
      "loss": 2.4726,
      "theoretical_loss": 3.37116695887757,
      "tokens_seen": 2475556864
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.217089086438434e-05,
      "loss": 2.7927,
      "theoretical_loss": 3.371159826177107,
      "tokens_seen": 2475622400
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.215434083601287e-05,
      "loss": 2.6963,
      "theoretical_loss": 3.3711526937183303,
      "tokens_seen": 2475687936
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.213779080764138e-05,
      "loss": 2.2841,
      "theoretical_loss": 3.371145561501226,
      "tokens_seen": 2475753472
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.21212407792699e-05,
      "loss": 2.5664,
      "theoretical_loss": 3.371138429525778,
      "tokens_seen": 2475819008
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.210469075089843e-05,
      "loss": 2.321,
      "theoretical_loss": 3.371131297791973,
      "tokens_seen": 2475884544
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.208814072252695e-05,
      "loss": 2.5371,
      "theoretical_loss": 3.3711241662997957,
      "tokens_seen": 2475950080
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.207159069415548e-05,
      "loss": 2.8203,
      "theoretical_loss": 3.371117035049232,
      "tokens_seen": 2476015616
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.2055040665784e-05,
      "loss": 2.5894,
      "theoretical_loss": 3.3711099040402672,
      "tokens_seen": 2476081152
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.203849063741252e-05,
      "loss": 2.7377,
      "theoretical_loss": 3.3711027732728867,
      "tokens_seen": 2476146688
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.202194060904103e-05,
      "loss": 2.4999,
      "theoretical_loss": 3.3710956427470755,
      "tokens_seen": 2476212224
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.200539058066956e-05,
      "loss": 2.748,
      "theoretical_loss": 3.37108851246282,
      "tokens_seen": 2476277760
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.198884055229808e-05,
      "loss": 2.9227,
      "theoretical_loss": 3.3710813824201047,
      "tokens_seen": 2476343296
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.19722905239266e-05,
      "loss": 2.5086,
      "theoretical_loss": 3.3710742526189152,
      "tokens_seen": 2476408832
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.195574049555513e-05,
      "loss": 2.4142,
      "theoretical_loss": 3.3710671230592375,
      "tokens_seen": 2476474368
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.193919046718365e-05,
      "loss": 2.4588,
      "theoretical_loss": 3.3710599937410564,
      "tokens_seen": 2476539904
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.192264043881218e-05,
      "loss": 2.4739,
      "theoretical_loss": 3.3710528646643576,
      "tokens_seen": 2476605440
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.19060904104407e-05,
      "loss": 9.465,
      "theoretical_loss": 3.3710457358291266,
      "tokens_seen": 2476670976
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.188954038206923e-05,
      "loss": 2.0905,
      "theoretical_loss": 3.3710386072353486,
      "tokens_seen": 2476736512
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.187299035369775e-05,
      "loss": 2.3738,
      "theoretical_loss": 3.371031478883009,
      "tokens_seen": 2476802048
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.185644032532627e-05,
      "loss": 2.6996,
      "theoretical_loss": 3.371024350772094,
      "tokens_seen": 2476867584
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.18398902969548e-05,
      "loss": 2.4732,
      "theoretical_loss": 3.3710172229025877,
      "tokens_seen": 2476933120
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.182334026858332e-05,
      "loss": 2.4566,
      "theoretical_loss": 3.3710100952744764,
      "tokens_seen": 2476998656
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.180679024021185e-05,
      "loss": 2.5056,
      "theoretical_loss": 3.371002967887746,
      "tokens_seen": 2477064192
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.179024021184037e-05,
      "loss": 2.4431,
      "theoretical_loss": 3.3709958407423803,
      "tokens_seen": 2477129728
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.177369018346888e-05,
      "loss": 2.5814,
      "theoretical_loss": 3.3709887138383663,
      "tokens_seen": 2477195264
    },
    {
      "epoch": 0.88,
      "objective/train/avg_token_score": 0.013918811455368996,
      "objective/train/avg_weight": 0.9836359024047852,
      "objective/train/docs_used": 1392376,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.676112651824951,
      "objective/train/std_weight": 0.08957678824663162,
      "objective/train/theoretical_loss": 3.370981587175689,
      "objective/train/tokens_used": 2497720800,
      "theoretical_loss": 3.370981587175689,
      "tokens_seen": 2477260800
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.17571401550974e-05,
      "loss": 2.6422,
      "theoretical_loss": 3.370981587175689,
      "tokens_seen": 2477260800
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.174059012672593e-05,
      "loss": 2.2752,
      "theoretical_loss": 3.3709744607543337,
      "tokens_seen": 2477326336
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.172404009835445e-05,
      "loss": 2.5445,
      "theoretical_loss": 3.370967334574286,
      "tokens_seen": 2477391872
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.170749006998298e-05,
      "loss": 2.4358,
      "theoretical_loss": 3.3709602086355313,
      "tokens_seen": 2477457408
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.16909400416115e-05,
      "loss": 2.4916,
      "theoretical_loss": 3.3709530829380543,
      "tokens_seen": 2477522944
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.167439001324002e-05,
      "loss": 2.5351,
      "theoretical_loss": 3.370945957481842,
      "tokens_seen": 2477588480
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.165783998486853e-05,
      "loss": 2.7359,
      "theoretical_loss": 3.3709388322668783,
      "tokens_seen": 2477654016
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.164128995649706e-05,
      "loss": 2.6395,
      "theoretical_loss": 3.37093170729315,
      "tokens_seen": 2477719552
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.162473992812558e-05,
      "loss": 2.4973,
      "theoretical_loss": 3.3709245825606415,
      "tokens_seen": 2477785088
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.16081898997541e-05,
      "loss": 2.844,
      "theoretical_loss": 3.370917458069339,
      "tokens_seen": 2477850624
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.159163987138263e-05,
      "loss": 2.5105,
      "theoretical_loss": 3.370910333819227,
      "tokens_seen": 2477916160
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.157508984301115e-05,
      "loss": 2.6842,
      "theoretical_loss": 3.370903209810292,
      "tokens_seen": 2477981696
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.155853981463968e-05,
      "loss": 2.2264,
      "theoretical_loss": 3.370896086042519,
      "tokens_seen": 2478047232
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.15419897862682e-05,
      "loss": 2.4956,
      "theoretical_loss": 3.3708889625158935,
      "tokens_seen": 2478112768
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.152543975789673e-05,
      "loss": 2.5518,
      "theoretical_loss": 3.3708818392304005,
      "tokens_seen": 2478178304
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.150888972952525e-05,
      "loss": 2.6734,
      "theoretical_loss": 3.3708747161860266,
      "tokens_seen": 2478243840
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.149233970115377e-05,
      "loss": 2.5207,
      "theoretical_loss": 3.370867593382756,
      "tokens_seen": 2478309376
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.14757896727823e-05,
      "loss": 2.4247,
      "theoretical_loss": 3.370860470820575,
      "tokens_seen": 2478374912
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.145923964441082e-05,
      "loss": 2.5776,
      "theoretical_loss": 3.3708533484994687,
      "tokens_seen": 2478440448
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.144268961603935e-05,
      "loss": 2.4694,
      "theoretical_loss": 3.3708462264194226,
      "tokens_seen": 2478505984
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.142613958766787e-05,
      "loss": 2.4865,
      "theoretical_loss": 3.3708391045804222,
      "tokens_seen": 2478571520
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.140958955929638e-05,
      "loss": 2.6004,
      "theoretical_loss": 3.370831982982453,
      "tokens_seen": 2478637056
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.13930395309249e-05,
      "loss": 2.3658,
      "theoretical_loss": 3.3708248616255005,
      "tokens_seen": 2478702592
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.137648950255343e-05,
      "loss": 2.654,
      "theoretical_loss": 3.3708177405095503,
      "tokens_seen": 2478768128
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.135993947418195e-05,
      "loss": 2.5488,
      "theoretical_loss": 3.3708106196345877,
      "tokens_seen": 2478833664
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.134338944581048e-05,
      "loss": 2.4167,
      "theoretical_loss": 3.370803499000598,
      "tokens_seen": 2478899200
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.1326839417439e-05,
      "loss": 2.4419,
      "theoretical_loss": 3.370796378607567,
      "tokens_seen": 2478964736
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.131028938906752e-05,
      "loss": 2.5213,
      "theoretical_loss": 3.37078925845548,
      "tokens_seen": 2479030272
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.129373936069604e-05,
      "loss": 2.5872,
      "theoretical_loss": 3.3707821385443224,
      "tokens_seen": 2479095808
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.127718933232456e-05,
      "loss": 2.3886,
      "theoretical_loss": 3.3707750188740797,
      "tokens_seen": 2479161344
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.126063930395308e-05,
      "loss": 2.4201,
      "theoretical_loss": 3.3707678994447376,
      "tokens_seen": 2479226880
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.124408927558161e-05,
      "loss": 2.6358,
      "theoretical_loss": 3.3707607802562816,
      "tokens_seen": 2479292416
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.122753924721013e-05,
      "loss": 2.6698,
      "theoretical_loss": 3.370753661308697,
      "tokens_seen": 2479357952
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.121098921883866e-05,
      "loss": 2.6401,
      "theoretical_loss": 3.370746542601969,
      "tokens_seen": 2479423488
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.119443919046718e-05,
      "loss": 2.5741,
      "theoretical_loss": 3.3707394241360835,
      "tokens_seen": 2479489024
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.117788916209572e-05,
      "loss": 2.7162,
      "theoretical_loss": 3.370732305911026,
      "tokens_seen": 2479554560
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.116133913372423e-05,
      "loss": 2.2821,
      "theoretical_loss": 3.370725187926782,
      "tokens_seen": 2479620096
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.114478910535275e-05,
      "loss": 2.5766,
      "theoretical_loss": 3.3707180701833366,
      "tokens_seen": 2479685632
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.112823907698128e-05,
      "loss": 2.5472,
      "theoretical_loss": 3.3707109526806756,
      "tokens_seen": 2479751168
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.11116890486098e-05,
      "loss": 2.5956,
      "theoretical_loss": 3.3707038354187846,
      "tokens_seen": 2479816704
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.109513902023832e-05,
      "loss": 0.2515,
      "theoretical_loss": 3.3706967183976486,
      "tokens_seen": 2479882240
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.107858899186685e-05,
      "loss": 2.4833,
      "theoretical_loss": 3.3706896016172543,
      "tokens_seen": 2479947776
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.106203896349537e-05,
      "loss": 2.3683,
      "theoretical_loss": 3.370682485077585,
      "tokens_seen": 2480013312
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.104548893512388e-05,
      "loss": 2.5643,
      "theoretical_loss": 3.3706753687786284,
      "tokens_seen": 2480078848
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.10289389067524e-05,
      "loss": 2.4153,
      "theoretical_loss": 3.3706682527203693,
      "tokens_seen": 2480144384
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.101238887838093e-05,
      "loss": 2.1473,
      "theoretical_loss": 3.3706611369027923,
      "tokens_seen": 2480209920
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.099583885000945e-05,
      "loss": 2.4481,
      "theoretical_loss": 3.370654021325884,
      "tokens_seen": 2480275456
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.097928882163798e-05,
      "loss": 2.037,
      "theoretical_loss": 3.37064690598963,
      "tokens_seen": 2480340992
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.09627387932665e-05,
      "loss": 2.6033,
      "theoretical_loss": 3.3706397908940144,
      "tokens_seen": 2480406528
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.094618876489503e-05,
      "loss": 2.1706,
      "theoretical_loss": 3.3706326760390244,
      "tokens_seen": 2480472064
    },
    {
      "epoch": 0.89,
      "objective/train/avg_token_score": 0.013687331229448318,
      "objective/train/avg_weight": 0.9839262366294861,
      "objective/train/docs_used": 1394461,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.901820182800293,
      "objective/train/std_weight": 0.0735962837934494,
      "objective/train/theoretical_loss": 3.370625561424644,
      "objective/train/tokens_used": 2500997600,
      "theoretical_loss": 3.370625561424644,
      "tokens_seen": 2480537600
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.092963873652355e-05,
      "loss": 2.6699,
      "theoretical_loss": 3.370625561424644,
      "tokens_seen": 2480537600
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.091308870815206e-05,
      "loss": 2.0775,
      "theoretical_loss": 3.37061844705086,
      "tokens_seen": 2480603136
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.089653867978058e-05,
      "loss": 2.5386,
      "theoretical_loss": 3.3706113329176572,
      "tokens_seen": 2480668672
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.087998865140911e-05,
      "loss": 2.3994,
      "theoretical_loss": 3.370604219025021,
      "tokens_seen": 2480734208
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.086343862303765e-05,
      "loss": 2.4139,
      "theoretical_loss": 3.3705971053729376,
      "tokens_seen": 2480799744
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.084688859466617e-05,
      "loss": 2.6641,
      "theoretical_loss": 3.3705899919613915,
      "tokens_seen": 2480865280
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.08303385662947e-05,
      "loss": 2.4512,
      "theoretical_loss": 3.370582878790369,
      "tokens_seen": 2480930816
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.081378853792322e-05,
      "loss": 2.5095,
      "theoretical_loss": 3.370575765859856,
      "tokens_seen": 2480996352
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.079723850955173e-05,
      "loss": 2.4555,
      "theoretical_loss": 3.370568653169837,
      "tokens_seen": 2481061888
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.078068848118025e-05,
      "loss": 2.0186,
      "theoretical_loss": 3.3705615407202973,
      "tokens_seen": 2481127424
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.076413845280878e-05,
      "loss": 2.3031,
      "theoretical_loss": 3.370554428511224,
      "tokens_seen": 2481192960
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.07475884244373e-05,
      "loss": -3.2665,
      "theoretical_loss": 3.370547316542601,
      "tokens_seen": 2481258496
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.073103839606582e-05,
      "loss": 2.459,
      "theoretical_loss": 3.3705402048144153,
      "tokens_seen": 2481324032
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.071448836769435e-05,
      "loss": 2.4,
      "theoretical_loss": 3.370533093326651,
      "tokens_seen": 2481389568
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.069793833932287e-05,
      "loss": 2.5493,
      "theoretical_loss": 3.370525982079294,
      "tokens_seen": 2481455104
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.06813883109514e-05,
      "loss": 2.4268,
      "theoretical_loss": 3.3705188710723304,
      "tokens_seen": 2481520640
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.06648382825799e-05,
      "loss": 2.3719,
      "theoretical_loss": 3.3705117603057455,
      "tokens_seen": 2481586176
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.064828825420843e-05,
      "loss": 2.4835,
      "theoretical_loss": 3.370504649779525,
      "tokens_seen": 2481651712
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.063173822583695e-05,
      "loss": 2.6673,
      "theoretical_loss": 3.3704975394936536,
      "tokens_seen": 2481717248
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.061518819746548e-05,
      "loss": 2.29,
      "theoretical_loss": 3.3704904294481173,
      "tokens_seen": 2481782784
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.0598638169094e-05,
      "loss": 2.3714,
      "theoretical_loss": 3.3704833196429016,
      "tokens_seen": 2481848320
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.058208814072253e-05,
      "loss": 2.5034,
      "theoretical_loss": 3.370476210077993,
      "tokens_seen": 2481913856
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.056553811235105e-05,
      "loss": 2.3877,
      "theoretical_loss": 3.3704691007533754,
      "tokens_seen": 2481979392
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.054898808397957e-05,
      "loss": 2.2884,
      "theoretical_loss": 3.370461991669035,
      "tokens_seen": 2482044928
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.05324380556081e-05,
      "loss": 2.7482,
      "theoretical_loss": 3.370454882824958,
      "tokens_seen": 2482110464
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.051588802723662e-05,
      "loss": 2.7634,
      "theoretical_loss": 3.370447774221129,
      "tokens_seen": 2482176000
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.049933799886515e-05,
      "loss": 2.7167,
      "theoretical_loss": 3.370440665857534,
      "tokens_seen": 2482241536
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.048278797049367e-05,
      "loss": 2.4018,
      "theoretical_loss": 3.3704335577341586,
      "tokens_seen": 2482307072
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.04662379421222e-05,
      "loss": 2.7311,
      "theoretical_loss": 3.3704264498509877,
      "tokens_seen": 2482372608
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.044968791375072e-05,
      "loss": 2.344,
      "theoretical_loss": 3.3704193422080078,
      "tokens_seen": 2482438144
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.043313788537923e-05,
      "loss": 2.5789,
      "theoretical_loss": 3.3704122348052037,
      "tokens_seen": 2482503680
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.041658785700775e-05,
      "loss": 2.3625,
      "theoretical_loss": 3.3704051276425613,
      "tokens_seen": 2482569216
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.040003782863628e-05,
      "loss": 2.2405,
      "theoretical_loss": 3.3703980207200663,
      "tokens_seen": 2482634752
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.03834878002648e-05,
      "loss": 2.6106,
      "theoretical_loss": 3.3703909140377037,
      "tokens_seen": 2482700288
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.036693777189332e-05,
      "loss": 2.4163,
      "theoretical_loss": 3.3703838075954593,
      "tokens_seen": 2482765824
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.035038774352185e-05,
      "loss": 2.3586,
      "theoretical_loss": 3.370376701393319,
      "tokens_seen": 2482831360
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.033383771515037e-05,
      "loss": 2.5291,
      "theoretical_loss": 3.3703695954312676,
      "tokens_seen": 2482896896
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.03172876867789e-05,
      "loss": 2.2425,
      "theoretical_loss": 3.3703624897092914,
      "tokens_seen": 2482962432
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.030073765840741e-05,
      "loss": 2.8838,
      "theoretical_loss": 3.3703553842273757,
      "tokens_seen": 2483027968
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.028418763003593e-05,
      "loss": 2.4718,
      "theoretical_loss": 3.370348278985506,
      "tokens_seen": 2483093504
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.026763760166445e-05,
      "loss": 2.5077,
      "theoretical_loss": 3.3703411739836677,
      "tokens_seen": 2483159040
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.025108757329298e-05,
      "loss": 2.6173,
      "theoretical_loss": 3.3703340692218466,
      "tokens_seen": 2483224576
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.02345375449215e-05,
      "loss": 2.3578,
      "theoretical_loss": 3.3703269647000282,
      "tokens_seen": 2483290112
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.021798751655003e-05,
      "loss": 2.5768,
      "theoretical_loss": 3.3703198604181983,
      "tokens_seen": 2483355648
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.020143748817855e-05,
      "loss": 2.4765,
      "theoretical_loss": 3.3703127563763418,
      "tokens_seen": 2483421184
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.018488745980707e-05,
      "loss": 2.595,
      "theoretical_loss": 3.370305652574445,
      "tokens_seen": 2483486720
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.01683374314356e-05,
      "loss": 2.6122,
      "theoretical_loss": 3.370298549012493,
      "tokens_seen": 2483552256
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.015178740306412e-05,
      "loss": 2.3785,
      "theoretical_loss": 3.370291445690471,
      "tokens_seen": 2483617792
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.013523737469265e-05,
      "loss": 2.67,
      "theoretical_loss": 3.370284342608366,
      "tokens_seen": 2483683328
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.011868734632117e-05,
      "loss": 2.3694,
      "theoretical_loss": 3.370277239766162,
      "tokens_seen": 2483748864
    },
    {
      "epoch": 0.89,
      "objective/train/avg_token_score": 0.022078968584537506,
      "objective/train/avg_weight": 0.9740187525749207,
      "objective/train/docs_used": 1396627,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8319101333618164,
      "objective/train/std_weight": 0.10205477476119995,
      "objective/train/theoretical_loss": 3.3702701371638453,
      "objective/train/tokens_used": 2504274400,
      "theoretical_loss": 3.3702701371638453,
      "tokens_seen": 2483814400
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.01021373179497e-05,
      "loss": 2.4919,
      "theoretical_loss": 3.3702701371638453,
      "tokens_seen": 2483814400
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.008558728957822e-05,
      "loss": 2.7294,
      "theoretical_loss": 3.3702630348014013,
      "tokens_seen": 2483879936
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.006903726120674e-05,
      "loss": 2.2039,
      "theoretical_loss": 3.3702559326788157,
      "tokens_seen": 2483945472
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.005248723283525e-05,
      "loss": 2.5446,
      "theoretical_loss": 3.370248830796074,
      "tokens_seen": 2484011008
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.003593720446378e-05,
      "loss": 2.7532,
      "theoretical_loss": 3.3702417291531614,
      "tokens_seen": 2484076544
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.00193871760923e-05,
      "loss": 2.3532,
      "theoretical_loss": 3.3702346277500643,
      "tokens_seen": 2484142080
    },
    {
      "epoch": 0.89,
      "learning_rate": 8.000283714772083e-05,
      "loss": 2.3187,
      "theoretical_loss": 3.370227526586768,
      "tokens_seen": 2484207616
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.998628711934935e-05,
      "loss": 2.3575,
      "theoretical_loss": 3.3702204256632573,
      "tokens_seen": 2484273152
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.996973709097787e-05,
      "loss": 2.5045,
      "theoretical_loss": 3.3702133249795185,
      "tokens_seen": 2484338688
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.99531870626064e-05,
      "loss": 2.4529,
      "theoretical_loss": 3.3702062245355373,
      "tokens_seen": 2484404224
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.993663703423491e-05,
      "loss": 2.5405,
      "theoretical_loss": 3.370199124331299,
      "tokens_seen": 2484469760
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.992008700586343e-05,
      "loss": 2.4669,
      "theoretical_loss": 3.3701920243667893,
      "tokens_seen": 2484535296
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.990353697749196e-05,
      "loss": 2.3824,
      "theoretical_loss": 3.3701849246419937,
      "tokens_seen": 2484600832
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.988698694912048e-05,
      "loss": 2.564,
      "theoretical_loss": 3.3701778251568975,
      "tokens_seen": 2484666368
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.9870436920749e-05,
      "loss": 2.6786,
      "theoretical_loss": 3.370170725911487,
      "tokens_seen": 2484731904
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.985388689237753e-05,
      "loss": 2.59,
      "theoretical_loss": 3.3701636269057467,
      "tokens_seen": 2484797440
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.983733686400605e-05,
      "loss": 2.5195,
      "theoretical_loss": 3.3701565281396633,
      "tokens_seen": 2484862976
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.982078683563458e-05,
      "loss": 2.5678,
      "theoretical_loss": 3.370149429613222,
      "tokens_seen": 2484928512
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.98042368072631e-05,
      "loss": 0.915,
      "theoretical_loss": 3.370142331326408,
      "tokens_seen": 2484994048
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.978768677889162e-05,
      "loss": 2.4501,
      "theoretical_loss": 3.3701352332792074,
      "tokens_seen": 2485059584
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.977113675052015e-05,
      "loss": 2.1827,
      "theoretical_loss": 3.3701281354716057,
      "tokens_seen": 2485125120
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.975458672214867e-05,
      "loss": 2.6356,
      "theoretical_loss": 3.370121037903588,
      "tokens_seen": 2485190656
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.97380366937772e-05,
      "loss": 2.4312,
      "theoretical_loss": 3.370113940575141,
      "tokens_seen": 2485256192
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.972148666540572e-05,
      "loss": 2.3635,
      "theoretical_loss": 3.3701068434862487,
      "tokens_seen": 2485321728
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.970493663703424e-05,
      "loss": 2.5325,
      "theoretical_loss": 3.3700997466368983,
      "tokens_seen": 2485387264
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.968838660866275e-05,
      "loss": 2.6207,
      "theoretical_loss": 3.3700926500270745,
      "tokens_seen": 2485452800
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.967183658029128e-05,
      "loss": 2.4851,
      "theoretical_loss": 3.3700855536567627,
      "tokens_seen": 2485518336
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.96552865519198e-05,
      "loss": 2.596,
      "theoretical_loss": 3.3700784575259495,
      "tokens_seen": 2485583872
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.963873652354833e-05,
      "loss": 2.4858,
      "theoretical_loss": 3.3700713616346194,
      "tokens_seen": 2485649408
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.962218649517685e-05,
      "loss": 2.3215,
      "theoretical_loss": 3.3700642659827587,
      "tokens_seen": 2485714944
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.960563646680537e-05,
      "loss": 2.73,
      "theoretical_loss": 3.370057170570353,
      "tokens_seen": 2485780480
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.95890864384339e-05,
      "loss": 2.375,
      "theoretical_loss": 3.3700500753973874,
      "tokens_seen": 2485846016
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.957253641006241e-05,
      "loss": 2.242,
      "theoretical_loss": 3.370042980463848,
      "tokens_seen": 2485911552
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.955598638169093e-05,
      "loss": 2.5792,
      "theoretical_loss": 3.3700358857697204,
      "tokens_seen": 2485977088
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.953943635331946e-05,
      "loss": 2.2693,
      "theoretical_loss": 3.3700287913149896,
      "tokens_seen": 2486042624
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.952288632494798e-05,
      "loss": 2.4382,
      "theoretical_loss": 3.3700216970996415,
      "tokens_seen": 2486108160
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.95063362965765e-05,
      "loss": 2.4716,
      "theoretical_loss": 3.3700146031236624,
      "tokens_seen": 2486173696
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.948978626820503e-05,
      "loss": 2.3332,
      "theoretical_loss": 3.370007509387037,
      "tokens_seen": 2486239232
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.947323623983355e-05,
      "loss": 2.469,
      "theoretical_loss": 3.3700004158897516,
      "tokens_seen": 2486304768
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.945668621146208e-05,
      "loss": 2.4543,
      "theoretical_loss": 3.3699933226317915,
      "tokens_seen": 2486370304
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.94401361830906e-05,
      "loss": 2.5318,
      "theoretical_loss": 3.3699862296131418,
      "tokens_seen": 2486435840
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.942358615471912e-05,
      "loss": 2.1025,
      "theoretical_loss": 3.3699791368337895,
      "tokens_seen": 2486501376
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.940703612634765e-05,
      "loss": 2.6133,
      "theoretical_loss": 3.3699720442937187,
      "tokens_seen": 2486566912
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.939048609797617e-05,
      "loss": 2.4728,
      "theoretical_loss": 3.369964951992916,
      "tokens_seen": 2486632448
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.93739360696047e-05,
      "loss": 2.6306,
      "theoretical_loss": 3.3699578599313664,
      "tokens_seen": 2486697984
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.935738604123322e-05,
      "loss": 2.4103,
      "theoretical_loss": 3.3699507681090557,
      "tokens_seen": 2486763520
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.934083601286174e-05,
      "loss": 2.5474,
      "theoretical_loss": 3.36994367652597,
      "tokens_seen": 2486829056
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.932428598449025e-05,
      "loss": 2.5628,
      "theoretical_loss": 3.369936585182095,
      "tokens_seen": 2486894592
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.930773595611878e-05,
      "loss": 2.4397,
      "theoretical_loss": 3.3699294940774154,
      "tokens_seen": 2486960128
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.92911859277473e-05,
      "loss": 2.5641,
      "theoretical_loss": 3.369922403211917,
      "tokens_seen": 2487025664
    },
    {
      "epoch": 0.89,
      "objective/train/avg_token_score": 0.022526346147060394,
      "objective/train/avg_weight": 0.9734890460968018,
      "objective/train/docs_used": 1398483,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.166034698486328,
      "objective/train/std_weight": 0.12133806943893433,
      "objective/train/theoretical_loss": 3.369915312585586,
      "objective/train/tokens_used": 2507551200,
      "theoretical_loss": 3.369915312585586,
      "tokens_seen": 2487091200
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.927463589937583e-05,
      "loss": 2.2985,
      "theoretical_loss": 3.369915312585586,
      "tokens_seen": 2487091200
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.925808587100435e-05,
      "loss": 1.9524,
      "theoretical_loss": 3.3699082221984082,
      "tokens_seen": 2487156736
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.924153584263287e-05,
      "loss": 2.2608,
      "theoretical_loss": 3.3699011320503685,
      "tokens_seen": 2487222272
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.92249858142614e-05,
      "loss": 2.0231,
      "theoretical_loss": 3.369894042141453,
      "tokens_seen": 2487287808
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.920843578588992e-05,
      "loss": 2.2882,
      "theoretical_loss": 3.369886952471647,
      "tokens_seen": 2487353344
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.919188575751843e-05,
      "loss": 2.4752,
      "theoretical_loss": 3.3698798630409366,
      "tokens_seen": 2487418880
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.917533572914696e-05,
      "loss": 2.4896,
      "theoretical_loss": 3.369872773849307,
      "tokens_seen": 2487484416
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.915878570077548e-05,
      "loss": 2.5282,
      "theoretical_loss": 3.3698656848967437,
      "tokens_seen": 2487549952
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.9142235672404e-05,
      "loss": 2.5709,
      "theoretical_loss": 3.3698585961832332,
      "tokens_seen": 2487615488
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.912568564403253e-05,
      "loss": 2.5904,
      "theoretical_loss": 3.36985150770876,
      "tokens_seen": 2487681024
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.910913561566105e-05,
      "loss": 2.5225,
      "theoretical_loss": 3.369844419473311,
      "tokens_seen": 2487746560
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.909258558728958e-05,
      "loss": 2.3642,
      "theoretical_loss": 3.369837331476871,
      "tokens_seen": 2487812096
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.90760355589181e-05,
      "loss": 2.5475,
      "theoretical_loss": 3.3698302437194254,
      "tokens_seen": 2487877632
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.905948553054662e-05,
      "loss": 2.5712,
      "theoretical_loss": 3.369823156200961,
      "tokens_seen": 2487943168
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.904293550217515e-05,
      "loss": 2.6073,
      "theoretical_loss": 3.369816068921462,
      "tokens_seen": 2488008704
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.902638547380367e-05,
      "loss": 2.3577,
      "theoretical_loss": 3.369808981880915,
      "tokens_seen": 2488074240
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.90098354454322e-05,
      "loss": 2.6417,
      "theoretical_loss": 3.3698018950793056,
      "tokens_seen": 2488139776
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.899328541706072e-05,
      "loss": 2.5738,
      "theoretical_loss": 3.369794808516619,
      "tokens_seen": 2488205312
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.897673538868924e-05,
      "loss": 2.5118,
      "theoretical_loss": 3.369787722192841,
      "tokens_seen": 2488270848
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.896018536031777e-05,
      "loss": 2.6841,
      "theoretical_loss": 3.369780636107958,
      "tokens_seen": 2488336384
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.894363533194628e-05,
      "loss": 2.6541,
      "theoretical_loss": 3.3697735502619546,
      "tokens_seen": 2488401920
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.89270853035748e-05,
      "loss": 2.6307,
      "theoretical_loss": 3.369766464654817,
      "tokens_seen": 2488467456
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.891053527520333e-05,
      "loss": 2.4432,
      "theoretical_loss": 3.369759379286531,
      "tokens_seen": 2488532992
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.889398524683185e-05,
      "loss": 2.3603,
      "theoretical_loss": 3.3697522941570814,
      "tokens_seen": 2488598528
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.887743521846038e-05,
      "loss": 2.6424,
      "theoretical_loss": 3.369745209266455,
      "tokens_seen": 2488664064
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.88608851900889e-05,
      "loss": 2.3737,
      "theoretical_loss": 3.3697381246146363,
      "tokens_seen": 2488729600
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.884433516171742e-05,
      "loss": 2.4117,
      "theoretical_loss": 3.369731040201612,
      "tokens_seen": 2488795136
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.882778513334593e-05,
      "loss": 2.236,
      "theoretical_loss": 3.369723956027367,
      "tokens_seen": 2488860672
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.881123510497446e-05,
      "loss": 2.612,
      "theoretical_loss": 3.369716872091888,
      "tokens_seen": 2488926208
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.879468507660298e-05,
      "loss": 2.4309,
      "theoretical_loss": 3.3697097883951597,
      "tokens_seen": 2488991744
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.87781350482315e-05,
      "loss": 2.4514,
      "theoretical_loss": 3.3697027049371675,
      "tokens_seen": 2489057280
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.876158501986003e-05,
      "loss": 2.3281,
      "theoretical_loss": 3.369695621717898,
      "tokens_seen": 2489122816
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.874503499148855e-05,
      "loss": 2.641,
      "theoretical_loss": 3.369688538737337,
      "tokens_seen": 2489188352
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.872848496311708e-05,
      "loss": 2.4673,
      "theoretical_loss": 3.369681455995469,
      "tokens_seen": 2489253888
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.87119349347456e-05,
      "loss": 2.4369,
      "theoretical_loss": 3.3696743734922805,
      "tokens_seen": 2489319424
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.869538490637413e-05,
      "loss": 2.2776,
      "theoretical_loss": 3.369667291227757,
      "tokens_seen": 2489384960
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.867883487800265e-05,
      "loss": 2.8102,
      "theoretical_loss": 3.369660209201884,
      "tokens_seen": 2489450496
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.866228484963117e-05,
      "loss": 2.1974,
      "theoretical_loss": 3.3696531274146477,
      "tokens_seen": 2489516032
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.86457348212597e-05,
      "loss": 2.6347,
      "theoretical_loss": 3.369646045866033,
      "tokens_seen": 2489581568
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.862918479288822e-05,
      "loss": 2.3733,
      "theoretical_loss": 3.3696389645560263,
      "tokens_seen": 2489647104
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.861263476451675e-05,
      "loss": 2.5571,
      "theoretical_loss": 3.369631883484613,
      "tokens_seen": 2489712640
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.859608473614527e-05,
      "loss": 2.4254,
      "theoretical_loss": 3.3696248026517788,
      "tokens_seen": 2489778176
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.857953470777378e-05,
      "loss": 2.3478,
      "theoretical_loss": 3.3696177220575096,
      "tokens_seen": 2489843712
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.85629846794023e-05,
      "loss": 2.3519,
      "theoretical_loss": 3.36961064170179,
      "tokens_seen": 2489909248
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.854643465103083e-05,
      "loss": 2.4656,
      "theoretical_loss": 3.369603561584607,
      "tokens_seen": 2489974784
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.852988462265935e-05,
      "loss": 2.6207,
      "theoretical_loss": 3.369596481705946,
      "tokens_seen": 2490040320
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.851333459428788e-05,
      "loss": 2.4008,
      "theoretical_loss": 3.3695894020657926,
      "tokens_seen": 2490105856
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.84967845659164e-05,
      "loss": 2.218,
      "theoretical_loss": 3.3695823226641317,
      "tokens_seen": 2490171392
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.848023453754492e-05,
      "loss": 2.7889,
      "theoretical_loss": 3.3695752435009503,
      "tokens_seen": 2490236928
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.846368450917343e-05,
      "loss": 2.6735,
      "theoretical_loss": 3.369568164576233,
      "tokens_seen": 2490302464
    },
    {
      "epoch": 0.89,
      "objective/train/avg_token_score": 0.01720762997865677,
      "objective/train/avg_weight": 0.9797623157501221,
      "objective/train/docs_used": 1400481,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.1836583614349365,
      "objective/train/std_weight": 0.0744570791721344,
      "objective/train/theoretical_loss": 3.3695610858899663,
      "objective/train/tokens_used": 2510828000,
      "theoretical_loss": 3.3695610858899663,
      "tokens_seen": 2490368000
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.844713448080196e-05,
      "loss": 2.5546,
      "theoretical_loss": 3.3695610858899663,
      "tokens_seen": 2490368000
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.843058445243048e-05,
      "loss": 2.3251,
      "theoretical_loss": 3.3695540074421353,
      "tokens_seen": 2490433536
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.8414034424059e-05,
      "loss": 2.483,
      "theoretical_loss": 3.369546929232726,
      "tokens_seen": 2490499072
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.839748439568753e-05,
      "loss": 2.4809,
      "theoretical_loss": 3.369539851261724,
      "tokens_seen": 2490564608
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.838093436731605e-05,
      "loss": 2.2858,
      "theoretical_loss": 3.369532773529115,
      "tokens_seen": 2490630144
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.836438433894458e-05,
      "loss": 2.2367,
      "theoretical_loss": 3.369525696034885,
      "tokens_seen": 2490695680
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.83478343105731e-05,
      "loss": 2.1702,
      "theoretical_loss": 3.3695186187790194,
      "tokens_seen": 2490761216
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.833128428220163e-05,
      "loss": 2.2235,
      "theoretical_loss": 3.3695115417615034,
      "tokens_seen": 2490826752
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.831473425383015e-05,
      "loss": 2.3509,
      "theoretical_loss": 3.369504464982324,
      "tokens_seen": 2490892288
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.829818422545867e-05,
      "loss": 2.5147,
      "theoretical_loss": 3.3694973884414656,
      "tokens_seen": 2490957824
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.82816341970872e-05,
      "loss": 2.4403,
      "theoretical_loss": 3.3694903121389146,
      "tokens_seen": 2491023360
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.826508416871572e-05,
      "loss": 2.6796,
      "theoretical_loss": 3.3694832360746565,
      "tokens_seen": 2491088896
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.824853414034425e-05,
      "loss": 2.4483,
      "theoretical_loss": 3.369476160248677,
      "tokens_seen": 2491154432
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.823198411197277e-05,
      "loss": 2.3342,
      "theoretical_loss": 3.369469084660962,
      "tokens_seen": 2491219968
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.821543408360128e-05,
      "loss": 2.3137,
      "theoretical_loss": 3.369462009311497,
      "tokens_seen": 2491285504
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.81988840552298e-05,
      "loss": 2.346,
      "theoretical_loss": 3.3694549342002675,
      "tokens_seen": 2491351040
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.818233402685833e-05,
      "loss": 2.5111,
      "theoretical_loss": 3.36944785932726,
      "tokens_seen": 2491416576
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.816578399848685e-05,
      "loss": 2.65,
      "theoretical_loss": 3.3694407846924594,
      "tokens_seen": 2491482112
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.814923397011538e-05,
      "loss": 2.4675,
      "theoretical_loss": 3.369433710295852,
      "tokens_seen": 2491547648
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.81326839417439e-05,
      "loss": 2.2346,
      "theoretical_loss": 3.3694266361374225,
      "tokens_seen": 2491613184
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.811613391337242e-05,
      "loss": 2.3428,
      "theoretical_loss": 3.3694195622171583,
      "tokens_seen": 2491678720
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.809958388500095e-05,
      "loss": 1.4633,
      "theoretical_loss": 3.3694124885350436,
      "tokens_seen": 2491744256
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.808303385662946e-05,
      "loss": 2.3408,
      "theoretical_loss": 3.369405415091065,
      "tokens_seen": 2491809792
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.806648382825798e-05,
      "loss": 3.0357,
      "theoretical_loss": 3.3693983418852076,
      "tokens_seen": 2491875328
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.804993379988651e-05,
      "loss": 2.5317,
      "theoretical_loss": 3.3693912689174574,
      "tokens_seen": 2491940864
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.803338377151503e-05,
      "loss": 2.3932,
      "theoretical_loss": 3.3693841961878004,
      "tokens_seen": 2492006400
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.801683374314356e-05,
      "loss": 2.2638,
      "theoretical_loss": 3.3693771236962218,
      "tokens_seen": 2492071936
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.800028371477208e-05,
      "loss": 2.6585,
      "theoretical_loss": 3.369370051442708,
      "tokens_seen": 2492137472
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.79837336864006e-05,
      "loss": 2.4854,
      "theoretical_loss": 3.369362979427244,
      "tokens_seen": 2492203008
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.796718365802913e-05,
      "loss": 2.4995,
      "theoretical_loss": 3.369355907649816,
      "tokens_seen": 2492268544
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.795063362965765e-05,
      "loss": 2.5022,
      "theoretical_loss": 3.3693488361104094,
      "tokens_seen": 2492334080
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.793408360128618e-05,
      "loss": 2.3457,
      "theoretical_loss": 3.3693417648090103,
      "tokens_seen": 2492399616
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.79175335729147e-05,
      "loss": 2.6665,
      "theoretical_loss": 3.3693346937456043,
      "tokens_seen": 2492465152
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.790098354454322e-05,
      "loss": 2.8211,
      "theoretical_loss": 3.369327622920177,
      "tokens_seen": 2492530688
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.788443351617175e-05,
      "loss": 2.3895,
      "theoretical_loss": 3.369320552332714,
      "tokens_seen": 2492596224
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.786788348780027e-05,
      "loss": 2.6889,
      "theoretical_loss": 3.3693134819832014,
      "tokens_seen": 2492661760
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.785133345942878e-05,
      "loss": 2.5388,
      "theoretical_loss": 3.3693064118716247,
      "tokens_seen": 2492727296
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.78347834310573e-05,
      "loss": 2.5979,
      "theoretical_loss": 3.36929934199797,
      "tokens_seen": 2492792832
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.781823340268583e-05,
      "loss": 2.6979,
      "theoretical_loss": 3.3692922723622223,
      "tokens_seen": 2492858368
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.780168337431435e-05,
      "loss": 2.1693,
      "theoretical_loss": 3.3692852029643685,
      "tokens_seen": 2492923904
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.778513334594288e-05,
      "loss": 2.6358,
      "theoretical_loss": 3.369278133804393,
      "tokens_seen": 2492989440
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.77685833175714e-05,
      "loss": 2.3861,
      "theoretical_loss": 3.3692710648822826,
      "tokens_seen": 2493054976
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.775203328919993e-05,
      "loss": 2.6077,
      "theoretical_loss": 3.369263996198022,
      "tokens_seen": 2493120512
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.773548326082845e-05,
      "loss": 2.4215,
      "theoretical_loss": 3.3692569277515982,
      "tokens_seen": 2493186048
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.771893323245696e-05,
      "loss": 2.6245,
      "theoretical_loss": 3.3692498595429963,
      "tokens_seen": 2493251584
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.770238320408548e-05,
      "loss": 2.3726,
      "theoretical_loss": 3.369242791572202,
      "tokens_seen": 2493317120
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.768583317571401e-05,
      "loss": 2.3092,
      "theoretical_loss": 3.369235723839201,
      "tokens_seen": 2493382656
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.766928314734253e-05,
      "loss": 2.26,
      "theoretical_loss": 3.369228656343979,
      "tokens_seen": 2493448192
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.765273311897106e-05,
      "loss": 2.3907,
      "theoretical_loss": 3.3692215890865223,
      "tokens_seen": 2493513728
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.763618309059958e-05,
      "loss": 2.485,
      "theoretical_loss": 3.369214522066816,
      "tokens_seen": 2493579264
    },
    {
      "epoch": 0.89,
      "objective/train/avg_token_score": 0.021028241142630577,
      "objective/train/avg_weight": 0.97524094581604,
      "objective/train/docs_used": 1401835,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8834898471832275,
      "objective/train/std_weight": 0.10205395519733429,
      "objective/train/theoretical_loss": 3.3692074552848457,
      "objective/train/tokens_used": 2514104800,
      "theoretical_loss": 3.3692074552848457,
      "tokens_seen": 2493644800
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.76196330622281e-05,
      "loss": 2.4589,
      "theoretical_loss": 3.3692074552848457,
      "tokens_seen": 2493644800
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.760308303385663e-05,
      "loss": 2.2786,
      "theoretical_loss": 3.369200388740598,
      "tokens_seen": 2493710336
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.758653300548515e-05,
      "loss": 2.2088,
      "theoretical_loss": 3.3691933224340587,
      "tokens_seen": 2493775872
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.756998297711368e-05,
      "loss": 2.3321,
      "theoretical_loss": 3.3691862563652126,
      "tokens_seen": 2493841408
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.75534329487422e-05,
      "loss": 2.5804,
      "theoretical_loss": 3.369179190534046,
      "tokens_seen": 2493906944
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.753688292037072e-05,
      "loss": 2.7141,
      "theoretical_loss": 3.3691721249405444,
      "tokens_seen": 2493972480
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.752033289199925e-05,
      "loss": 2.6659,
      "theoretical_loss": 3.3691650595846943,
      "tokens_seen": 2494038016
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.750378286362777e-05,
      "loss": 2.5178,
      "theoretical_loss": 3.3691579944664802,
      "tokens_seen": 2494103552
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.74872328352563e-05,
      "loss": 2.4096,
      "theoretical_loss": 3.3691509295858895,
      "tokens_seen": 2494169088
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.74872328352563e-05,
      "loss": 9.4324,
      "theoretical_loss": 3.3691438649429064,
      "tokens_seen": 2494234624
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.74706828068848e-05,
      "loss": 2.5772,
      "theoretical_loss": 3.3691368005375173,
      "tokens_seen": 2494300160
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.745413277851333e-05,
      "loss": 2.5004,
      "theoretical_loss": 3.3691297363697084,
      "tokens_seen": 2494365696
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.743758275014185e-05,
      "loss": 2.5756,
      "theoretical_loss": 3.369122672439465,
      "tokens_seen": 2494431232
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.742103272177038e-05,
      "loss": 2.7453,
      "theoretical_loss": 3.3691156087467724,
      "tokens_seen": 2494496768
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.74044826933989e-05,
      "loss": 2.1424,
      "theoretical_loss": 3.3691085452916174,
      "tokens_seen": 2494562304
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.738793266502743e-05,
      "loss": 2.273,
      "theoretical_loss": 3.369101482073985,
      "tokens_seen": 2494627840
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.737138263665595e-05,
      "loss": 2.5147,
      "theoretical_loss": 3.3690944190938614,
      "tokens_seen": 2494693376
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.735483260828446e-05,
      "loss": 2.5026,
      "theoretical_loss": 3.3690873563512325,
      "tokens_seen": 2494758912
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.733828257991298e-05,
      "loss": 2.3964,
      "theoretical_loss": 3.3690802938460833,
      "tokens_seen": 2494824448
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.732173255154151e-05,
      "loss": 2.8607,
      "theoretical_loss": 3.3690732315784,
      "tokens_seen": 2494889984
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.730518252317003e-05,
      "loss": 2.472,
      "theoretical_loss": 3.369066169548169,
      "tokens_seen": 2494955520
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.728863249479856e-05,
      "loss": 2.4443,
      "theoretical_loss": 3.3690591077553753,
      "tokens_seen": 2495021056
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.727208246642708e-05,
      "loss": 2.118,
      "theoretical_loss": 3.3690520462000046,
      "tokens_seen": 2495086592
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.72555324380556e-05,
      "loss": 2.3515,
      "theoretical_loss": 3.369044984882043,
      "tokens_seen": 2495152128
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.723898240968413e-05,
      "loss": 2.5325,
      "theoretical_loss": 3.369037923801477,
      "tokens_seen": 2495217664
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.722243238131265e-05,
      "loss": 2.5726,
      "theoretical_loss": 3.369030862958291,
      "tokens_seen": 2495283200
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.720588235294118e-05,
      "loss": 2.2971,
      "theoretical_loss": 3.369023802352472,
      "tokens_seen": 2495348736
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.71893323245697e-05,
      "loss": 2.6365,
      "theoretical_loss": 3.3690167419840047,
      "tokens_seen": 2495414272
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.717278229619822e-05,
      "loss": 2.5955,
      "theoretical_loss": 3.3690096818528756,
      "tokens_seen": 2495479808
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.715623226782675e-05,
      "loss": 2.5883,
      "theoretical_loss": 3.36900262195907,
      "tokens_seen": 2495545344
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.713968223945527e-05,
      "loss": 2.5837,
      "theoretical_loss": 3.3689955623025747,
      "tokens_seen": 2495610880
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.71231322110838e-05,
      "loss": 2.591,
      "theoretical_loss": 3.368988502883374,
      "tokens_seen": 2495676416
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.710658218271231e-05,
      "loss": 2.241,
      "theoretical_loss": 3.368981443701455,
      "tokens_seen": 2495741952
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.709003215434083e-05,
      "loss": 2.5416,
      "theoretical_loss": 3.368974384756803,
      "tokens_seen": 2495807488
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.707348212596935e-05,
      "loss": 2.4873,
      "theoretical_loss": 3.3689673260494035,
      "tokens_seen": 2495873024
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.705693209759788e-05,
      "loss": 2.589,
      "theoretical_loss": 3.3689602675792427,
      "tokens_seen": 2495938560
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.70403820692264e-05,
      "loss": 2.6891,
      "theoretical_loss": 3.368953209346306,
      "tokens_seen": 2496004096
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.702383204085493e-05,
      "loss": 2.583,
      "theoretical_loss": 3.36894615135058,
      "tokens_seen": 2496069632
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.700728201248345e-05,
      "loss": 2.3822,
      "theoretical_loss": 3.36893909359205,
      "tokens_seen": 2496135168
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.699073198411196e-05,
      "loss": 2.4931,
      "theoretical_loss": 3.368932036070701,
      "tokens_seen": 2496200704
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.697418195574049e-05,
      "loss": 2.5295,
      "theoretical_loss": 3.36892497878652,
      "tokens_seen": 2496266240
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.695763192736901e-05,
      "loss": 2.744,
      "theoretical_loss": 3.3689179217394924,
      "tokens_seen": 2496331776
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.694108189899753e-05,
      "loss": 2.426,
      "theoretical_loss": 3.368910864929604,
      "tokens_seen": 2496397312
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.692453187062606e-05,
      "loss": 2.5121,
      "theoretical_loss": 3.3689038083568406,
      "tokens_seen": 2496462848
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.690798184225458e-05,
      "loss": 2.6616,
      "theoretical_loss": 3.368896752021188,
      "tokens_seen": 2496528384
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.68914318138831e-05,
      "loss": 2.5474,
      "theoretical_loss": 3.3688896959226318,
      "tokens_seen": 2496593920
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.687488178551163e-05,
      "loss": 2.5601,
      "theoretical_loss": 3.368882640061158,
      "tokens_seen": 2496659456
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.685833175714015e-05,
      "loss": 2.6543,
      "theoretical_loss": 3.3688755844367524,
      "tokens_seen": 2496724992
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.684178172876868e-05,
      "loss": 2.7369,
      "theoretical_loss": 3.368868529049401,
      "tokens_seen": 2496790528
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.68252317003972e-05,
      "loss": 2.1811,
      "theoretical_loss": 3.3688614738990896,
      "tokens_seen": 2496856064
    },
    {
      "epoch": 0.89,
      "objective/train/avg_token_score": 0.012381414882838726,
      "objective/train/avg_weight": 0.9854251742362976,
      "objective/train/docs_used": 1403825,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6143651008605957,
      "objective/train/std_weight": 0.05373578518629074,
      "objective/train/theoretical_loss": 3.368854418985804,
      "objective/train/tokens_used": 2517381600,
      "theoretical_loss": 3.368854418985804,
      "tokens_seen": 2496921600
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.680868167202573e-05,
      "loss": 2.3199,
      "theoretical_loss": 3.368854418985804,
      "tokens_seen": 2496921600
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.679213164365425e-05,
      "loss": 2.708,
      "theoretical_loss": 3.3688473643095294,
      "tokens_seen": 2496987136
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.677558161528277e-05,
      "loss": 2.508,
      "theoretical_loss": 3.368840309870252,
      "tokens_seen": 2497052672
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.67590315869113e-05,
      "loss": 2.4912,
      "theoretical_loss": 3.3688332556679583,
      "tokens_seen": 2497118208
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.674248155853981e-05,
      "loss": 2.3741,
      "theoretical_loss": 3.368826201702633,
      "tokens_seen": 2497183744
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.672593153016833e-05,
      "loss": 2.4563,
      "theoretical_loss": 3.368819147974263,
      "tokens_seen": 2497249280
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.670938150179686e-05,
      "loss": 2.6041,
      "theoretical_loss": 3.3688120944828333,
      "tokens_seen": 2497314816
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.669283147342538e-05,
      "loss": 2.2049,
      "theoretical_loss": 3.36880504122833,
      "tokens_seen": 2497380352
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.66762814450539e-05,
      "loss": 2.5768,
      "theoretical_loss": 3.368797988210739,
      "tokens_seen": 2497445888
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.665973141668243e-05,
      "loss": 2.4737,
      "theoretical_loss": 3.3687909354300456,
      "tokens_seen": 2497511424
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.664318138831095e-05,
      "loss": 2.3975,
      "theoretical_loss": 3.3687838828862366,
      "tokens_seen": 2497576960
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.662663135993948e-05,
      "loss": 2.4459,
      "theoretical_loss": 3.368776830579297,
      "tokens_seen": 2497642496
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.661008133156799e-05,
      "loss": 2.7071,
      "theoretical_loss": 3.368769778509213,
      "tokens_seen": 2497708032
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.659353130319651e-05,
      "loss": 2.3706,
      "theoretical_loss": 3.3687627266759703,
      "tokens_seen": 2497773568
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.657698127482503e-05,
      "loss": 2.4093,
      "theoretical_loss": 3.368755675079555,
      "tokens_seen": 2497839104
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.656043124645357e-05,
      "loss": 2.7028,
      "theoretical_loss": 3.3687486237199527,
      "tokens_seen": 2497904640
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.65438812180821e-05,
      "loss": 2.4572,
      "theoretical_loss": 3.3687415725971492,
      "tokens_seen": 2497970176
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.652733118971062e-05,
      "loss": 2.4941,
      "theoretical_loss": 3.36873452171113,
      "tokens_seen": 2498035712
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.651078116133914e-05,
      "loss": 2.522,
      "theoretical_loss": 3.368727471061882,
      "tokens_seen": 2498101248
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.649423113296765e-05,
      "loss": 2.2346,
      "theoretical_loss": 3.3687204206493906,
      "tokens_seen": 2498166784
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.647768110459618e-05,
      "loss": 2.5467,
      "theoretical_loss": 3.3687133704736407,
      "tokens_seen": 2498232320
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.64611310762247e-05,
      "loss": 2.0975,
      "theoretical_loss": 3.368706320534619,
      "tokens_seen": 2498297856
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.644458104785323e-05,
      "loss": 2.5973,
      "theoretical_loss": 3.3686992708323116,
      "tokens_seen": 2498363392
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.642803101948175e-05,
      "loss": 2.4679,
      "theoretical_loss": 3.3686922213667034,
      "tokens_seen": 2498428928
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.641148099111027e-05,
      "loss": 2.3918,
      "theoretical_loss": 3.368685172137781,
      "tokens_seen": 2498494464
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.63949309627388e-05,
      "loss": 2.5095,
      "theoretical_loss": 3.3686781231455303,
      "tokens_seen": 2498560000
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.637838093436732e-05,
      "loss": 2.5879,
      "theoretical_loss": 3.3686710743899364,
      "tokens_seen": 2498625536
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.636183090599583e-05,
      "loss": 2.3302,
      "theoretical_loss": 3.368664025870986,
      "tokens_seen": 2498691072
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.634528087762436e-05,
      "loss": 2.6328,
      "theoretical_loss": 3.3686569775886643,
      "tokens_seen": 2498756608
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.632873084925288e-05,
      "loss": 2.3982,
      "theoretical_loss": 3.3686499295429577,
      "tokens_seen": 2498822144
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.63121808208814e-05,
      "loss": 2.6579,
      "theoretical_loss": 3.3686428817338516,
      "tokens_seen": 2498887680
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.629563079250993e-05,
      "loss": 2.5356,
      "theoretical_loss": 3.368635834161332,
      "tokens_seen": 2498953216
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.627908076413845e-05,
      "loss": 1.8677,
      "theoretical_loss": 3.3686287868253846,
      "tokens_seen": 2499018752
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.626253073576698e-05,
      "loss": 2.5392,
      "theoretical_loss": 3.3686217397259957,
      "tokens_seen": 2499084288
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.62459807073955e-05,
      "loss": 2.3395,
      "theoretical_loss": 3.368614692863151,
      "tokens_seen": 2499149824
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.622943067902402e-05,
      "loss": 2.583,
      "theoretical_loss": 3.368607646236836,
      "tokens_seen": 2499215360
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.621288065065255e-05,
      "loss": 2.4809,
      "theoretical_loss": 3.3686005998470367,
      "tokens_seen": 2499280896
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.619633062228107e-05,
      "loss": 2.4671,
      "theoretical_loss": 3.3685935536937395,
      "tokens_seen": 2499346432
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.61797805939096e-05,
      "loss": 2.4036,
      "theoretical_loss": 3.3685865077769295,
      "tokens_seen": 2499411968
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.616323056553812e-05,
      "loss": 2.3902,
      "theoretical_loss": 3.368579462096593,
      "tokens_seen": 2499477504
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.614668053716664e-05,
      "loss": 2.5446,
      "theoretical_loss": 3.368572416652716,
      "tokens_seen": 2499543040
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.613013050879515e-05,
      "loss": 2.6344,
      "theoretical_loss": 3.3685653714452837,
      "tokens_seen": 2499608576
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.611358048042368e-05,
      "loss": 2.3887,
      "theoretical_loss": 3.3685583264742824,
      "tokens_seen": 2499674112
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.60970304520522e-05,
      "loss": 2.4239,
      "theoretical_loss": 3.3685512817396983,
      "tokens_seen": 2499739648
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.608048042368073e-05,
      "loss": 2.3663,
      "theoretical_loss": 3.3685442372415166,
      "tokens_seen": 2499805184
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.606393039530925e-05,
      "loss": 2.5874,
      "theoretical_loss": 3.3685371929797236,
      "tokens_seen": 2499870720
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.604738036693777e-05,
      "loss": 2.3158,
      "theoretical_loss": 3.368530148954305,
      "tokens_seen": 2499936256
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.60308303385663e-05,
      "loss": 2.4112,
      "theoretical_loss": 3.368523105165247,
      "tokens_seen": 2500001792
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.601428031019482e-05,
      "loss": 2.7045,
      "theoretical_loss": 3.3685160616125347,
      "tokens_seen": 2500067328
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.599773028182333e-05,
      "loss": 2.7873,
      "theoretical_loss": 3.368509018296155,
      "tokens_seen": 2500132864
    },
    {
      "epoch": 0.89,
      "objective/train/avg_token_score": 0.022752679884433746,
      "objective/train/avg_weight": 0.9732218384742737,
      "objective/train/docs_used": 1405631,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.360299587249756,
      "objective/train/std_weight": 0.10321398824453354,
      "objective/train/theoretical_loss": 3.368501975216093,
      "objective/train/tokens_used": 2520658400,
      "theoretical_loss": 3.368501975216093,
      "tokens_seen": 2500198400
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.598118025345186e-05,
      "loss": 2.0101,
      "theoretical_loss": 3.368501975216093,
      "tokens_seen": 2500198400
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.596463022508038e-05,
      "loss": 2.563,
      "theoretical_loss": 3.368494932372335,
      "tokens_seen": 2500263936
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.59480801967089e-05,
      "loss": 2.5427,
      "theoretical_loss": 3.3684878897648662,
      "tokens_seen": 2500329472
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.593153016833743e-05,
      "loss": 2.8087,
      "theoretical_loss": 3.3684808473936734,
      "tokens_seen": 2500395008
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.591498013996595e-05,
      "loss": 2.3737,
      "theoretical_loss": 3.368473805258742,
      "tokens_seen": 2500460544
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.589843011159448e-05,
      "loss": 2.2428,
      "theoretical_loss": 3.3684667633600585,
      "tokens_seen": 2500526080
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.5881880083223e-05,
      "loss": 2.4281,
      "theoretical_loss": 3.3684597216976075,
      "tokens_seen": 2500591616
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.586533005485152e-05,
      "loss": 2.5187,
      "theoretical_loss": 3.3684526802713757,
      "tokens_seen": 2500657152
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.584878002648005e-05,
      "loss": 2.1608,
      "theoretical_loss": 3.368445639081349,
      "tokens_seen": 2500722688
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.583222999810857e-05,
      "loss": 2.5,
      "theoretical_loss": 3.3684385981275136,
      "tokens_seen": 2500788224
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.58156799697371e-05,
      "loss": 2.3627,
      "theoretical_loss": 3.3684315574098544,
      "tokens_seen": 2500853760
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.579912994136562e-05,
      "loss": 2.5386,
      "theoretical_loss": 3.368424516928358,
      "tokens_seen": 2500919296
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.578257991299414e-05,
      "loss": 2.5569,
      "theoretical_loss": 3.3684174766830104,
      "tokens_seen": 2500984832
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.576602988462267e-05,
      "loss": 2.4219,
      "theoretical_loss": 3.3684104366737975,
      "tokens_seen": 2501050368
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.574947985625118e-05,
      "loss": 2.6983,
      "theoretical_loss": 3.368403396900704,
      "tokens_seen": 2501115904
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.57329298278797e-05,
      "loss": 2.5375,
      "theoretical_loss": 3.368396357363718,
      "tokens_seen": 2501181440
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.571637979950823e-05,
      "loss": 2.5709,
      "theoretical_loss": 3.368389318062823,
      "tokens_seen": 2501246976
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.569982977113675e-05,
      "loss": 2.2884,
      "theoretical_loss": 3.3683822789980065,
      "tokens_seen": 2501312512
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.568327974276528e-05,
      "loss": 2.656,
      "theoretical_loss": 3.3683752401692537,
      "tokens_seen": 2501378048
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.56667297143938e-05,
      "loss": 2.4897,
      "theoretical_loss": 3.3683682015765513,
      "tokens_seen": 2501443584
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.565017968602232e-05,
      "loss": 2.4065,
      "theoretical_loss": 3.368361163219884,
      "tokens_seen": 2501509120
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.563362965765083e-05,
      "loss": 2.467,
      "theoretical_loss": 3.3683541250992386,
      "tokens_seen": 2501574656
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.561707962927936e-05,
      "loss": 2.5232,
      "theoretical_loss": 3.368347087214601,
      "tokens_seen": 2501640192
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.560052960090788e-05,
      "loss": 2.6201,
      "theoretical_loss": 3.368340049565956,
      "tokens_seen": 2501705728
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.55839795725364e-05,
      "loss": 2.6728,
      "theoretical_loss": 3.368333012153291,
      "tokens_seen": 2501771264
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.556742954416493e-05,
      "loss": 2.5793,
      "theoretical_loss": 3.368325974976591,
      "tokens_seen": 2501836800
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.555087951579345e-05,
      "loss": 2.6183,
      "theoretical_loss": 3.3683189380358423,
      "tokens_seen": 2501902336
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.553432948742198e-05,
      "loss": 2.4689,
      "theoretical_loss": 3.3683119013310305,
      "tokens_seen": 2501967872
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.55177794590505e-05,
      "loss": 2.4534,
      "theoretical_loss": 3.368304864862142,
      "tokens_seen": 2502033408
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.550122943067903e-05,
      "loss": 2.52,
      "theoretical_loss": 3.3682978286291623,
      "tokens_seen": 2502098944
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.548467940230755e-05,
      "loss": 2.5271,
      "theoretical_loss": 3.3682907926320773,
      "tokens_seen": 2502164480
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.546812937393607e-05,
      "loss": 2.4651,
      "theoretical_loss": 3.3682837568708726,
      "tokens_seen": 2502230016
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.54515793455646e-05,
      "loss": 2.3941,
      "theoretical_loss": 3.368276721345535,
      "tokens_seen": 2502295552
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.543502931719312e-05,
      "loss": 1.5163,
      "theoretical_loss": 3.36826968605605,
      "tokens_seen": 2502361088
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.541847928882165e-05,
      "loss": 2.5958,
      "theoretical_loss": 3.368262651002403,
      "tokens_seen": 2502426624
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.540192926045017e-05,
      "loss": 2.1127,
      "theoretical_loss": 3.368255616184581,
      "tokens_seen": 2502492160
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.538537923207868e-05,
      "loss": 2.5689,
      "theoretical_loss": 3.3682485816025687,
      "tokens_seen": 2502557696
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.53688292037072e-05,
      "loss": 2.0528,
      "theoretical_loss": 3.368241547256353,
      "tokens_seen": 2502623232
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.535227917533573e-05,
      "loss": 2.1688,
      "theoretical_loss": 3.3682345131459193,
      "tokens_seen": 2502688768
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.533572914696425e-05,
      "loss": 2.4369,
      "theoretical_loss": 3.3682274792712534,
      "tokens_seen": 2502754304
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.531917911859278e-05,
      "loss": 2.4263,
      "theoretical_loss": 3.3682204456323417,
      "tokens_seen": 2502819840
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.53026290902213e-05,
      "loss": 2.2901,
      "theoretical_loss": 3.36821341222917,
      "tokens_seen": 2502885376
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.528607906184982e-05,
      "loss": 2.6163,
      "theoretical_loss": 3.368206379061724,
      "tokens_seen": 2502950912
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.526952903347833e-05,
      "loss": 2.4369,
      "theoretical_loss": 3.36819934612999,
      "tokens_seen": 2503016448
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.525297900510686e-05,
      "loss": 2.6656,
      "theoretical_loss": 3.3681923134339535,
      "tokens_seen": 2503081984
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.523642897673538e-05,
      "loss": 2.6787,
      "theoretical_loss": 3.3681852809736004,
      "tokens_seen": 2503147520
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.52198789483639e-05,
      "loss": 2.5236,
      "theoretical_loss": 3.368178248748917,
      "tokens_seen": 2503213056
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.520332891999243e-05,
      "loss": 2.5482,
      "theoretical_loss": 3.3681712167598894,
      "tokens_seen": 2503278592
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.518677889162095e-05,
      "loss": 2.7303,
      "theoretical_loss": 3.3681641850065027,
      "tokens_seen": 2503344128
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.517022886324948e-05,
      "loss": 2.4376,
      "theoretical_loss": 3.3681571534887436,
      "tokens_seen": 2503409664
    },
    {
      "epoch": 0.89,
      "objective/train/avg_token_score": 0.02296050451695919,
      "objective/train/avg_weight": 0.9729732275009155,
      "objective/train/docs_used": 1407612,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8035898208618164,
      "objective/train/std_weight": 0.0967830941081047,
      "objective/train/theoretical_loss": 3.3681501222065977,
      "objective/train/tokens_used": 2523935200,
      "theoretical_loss": 3.3681501222065977,
      "tokens_seen": 2503475200
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.5153678834878e-05,
      "loss": 2.6437,
      "theoretical_loss": 3.3681501222065977,
      "tokens_seen": 2503475200
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.513712880650653e-05,
      "loss": 2.511,
      "theoretical_loss": 3.3681430911600514,
      "tokens_seen": 2503540736
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.512057877813505e-05,
      "loss": 2.7724,
      "theoretical_loss": 3.36813606034909,
      "tokens_seen": 2503606272
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.510402874976357e-05,
      "loss": 2.7369,
      "theoretical_loss": 3.3681290297736997,
      "tokens_seen": 2503671808
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.50874787213921e-05,
      "loss": 2.144,
      "theoretical_loss": 3.3681219994338663,
      "tokens_seen": 2503737344
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.507092869302062e-05,
      "loss": 2.5555,
      "theoretical_loss": 3.368114969329576,
      "tokens_seen": 2503802880
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.505437866464915e-05,
      "loss": 2.4985,
      "theoretical_loss": 3.3681079394608147,
      "tokens_seen": 2503868416
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.503782863627767e-05,
      "loss": 2.5452,
      "theoretical_loss": 3.368100909827568,
      "tokens_seen": 2503933952
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.502127860790618e-05,
      "loss": 2.7111,
      "theoretical_loss": 3.3680938804298224,
      "tokens_seen": 2503999488
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.50047285795347e-05,
      "loss": 2.4862,
      "theoretical_loss": 3.3680868512675635,
      "tokens_seen": 2504065024
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.498817855116323e-05,
      "loss": 2.5528,
      "theoretical_loss": 3.3680798223407775,
      "tokens_seen": 2504130560
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.497162852279175e-05,
      "loss": 2.5189,
      "theoretical_loss": 3.36807279364945,
      "tokens_seen": 2504196096
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.495507849442028e-05,
      "loss": 2.4306,
      "theoretical_loss": 3.368065765193567,
      "tokens_seen": 2504261632
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.49385284660488e-05,
      "loss": 2.5977,
      "theoretical_loss": 3.3680587369731145,
      "tokens_seen": 2504327168
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.492197843767732e-05,
      "loss": 2.5584,
      "theoretical_loss": 3.3680517089880784,
      "tokens_seen": 2504392704
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.490542840930585e-05,
      "loss": 2.4116,
      "theoretical_loss": 3.3680446812384455,
      "tokens_seen": 2504458240
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.488887838093436e-05,
      "loss": 2.5222,
      "theoretical_loss": 3.3680376537242003,
      "tokens_seen": 2504523776
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.487232835256288e-05,
      "loss": 2.3274,
      "theoretical_loss": 3.36803062644533,
      "tokens_seen": 2504589312
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.485577832419141e-05,
      "loss": 1.7527,
      "theoretical_loss": 3.3680235994018197,
      "tokens_seen": 2504654848
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.483922829581993e-05,
      "loss": 1.3412,
      "theoretical_loss": 3.368016572593656,
      "tokens_seen": 2504720384
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.482267826744846e-05,
      "loss": 2.5994,
      "theoretical_loss": 3.3680095460208244,
      "tokens_seen": 2504785920
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.480612823907698e-05,
      "loss": 2.3451,
      "theoretical_loss": 3.3680025196833108,
      "tokens_seen": 2504851456
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.47895782107055e-05,
      "loss": 1.9872,
      "theoretical_loss": 3.367995493581102,
      "tokens_seen": 2504916992
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.477302818233403e-05,
      "loss": 2.5653,
      "theoretical_loss": 3.3679884677141825,
      "tokens_seen": 2504982528
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.475647815396255e-05,
      "loss": 2.4513,
      "theoretical_loss": 3.36798144208254,
      "tokens_seen": 2505048064
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.473992812559108e-05,
      "loss": 2.4822,
      "theoretical_loss": 3.3679744166861587,
      "tokens_seen": 2505113600
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.47233780972196e-05,
      "loss": 2.3442,
      "theoretical_loss": 3.367967391525026,
      "tokens_seen": 2505179136
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.470682806884812e-05,
      "loss": 2.5025,
      "theoretical_loss": 3.367960366599127,
      "tokens_seen": 2505244672
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.469027804047665e-05,
      "loss": 2.4617,
      "theoretical_loss": 3.3679533419084486,
      "tokens_seen": 2505310208
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.467372801210517e-05,
      "loss": 2.41,
      "theoretical_loss": 3.367946317452976,
      "tokens_seen": 2505375744
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.465717798373368e-05,
      "loss": 2.4675,
      "theoretical_loss": 3.367939293232695,
      "tokens_seen": 2505441280
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.46406279553622e-05,
      "loss": 2.3745,
      "theoretical_loss": 3.367932269247592,
      "tokens_seen": 2505506816
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.462407792699073e-05,
      "loss": 2.3002,
      "theoretical_loss": 3.367925245497653,
      "tokens_seen": 2505572352
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.460752789861925e-05,
      "loss": 2.4637,
      "theoretical_loss": 3.3679182219828636,
      "tokens_seen": 2505637888
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.459097787024778e-05,
      "loss": 2.4714,
      "theoretical_loss": 3.36791119870321,
      "tokens_seen": 2505703424
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.45744278418763e-05,
      "loss": 2.4029,
      "theoretical_loss": 3.3679041756586785,
      "tokens_seen": 2505768960
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.455787781350483e-05,
      "loss": 2.6714,
      "theoretical_loss": 3.3678971528492547,
      "tokens_seen": 2505834496
    },
    {
      "epoch": 0.89,
      "learning_rate": 7.454132778513335e-05,
      "loss": 2.0099,
      "theoretical_loss": 3.3678901302749247,
      "tokens_seen": 2505900032
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.452477775676186e-05,
      "loss": 2.5237,
      "theoretical_loss": 3.3678831079356746,
      "tokens_seen": 2505965568
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.450822772839038e-05,
      "loss": 2.7315,
      "theoretical_loss": 3.3678760858314902,
      "tokens_seen": 2506031104
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.449167770001891e-05,
      "loss": 2.2499,
      "theoretical_loss": 3.367869063962357,
      "tokens_seen": 2506096640
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.447512767164743e-05,
      "loss": 2.4433,
      "theoretical_loss": 3.3678620423282624,
      "tokens_seen": 2506162176
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.445857764327596e-05,
      "loss": 2.5347,
      "theoretical_loss": 3.367855020929191,
      "tokens_seen": 2506227712
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.444202761490448e-05,
      "loss": 2.7341,
      "theoretical_loss": 3.367847999765129,
      "tokens_seen": 2506293248
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.4425477586533e-05,
      "loss": 2.6447,
      "theoretical_loss": 3.367840978836063,
      "tokens_seen": 2506358784
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.440892755816153e-05,
      "loss": 2.6352,
      "theoretical_loss": 3.3678339581419787,
      "tokens_seen": 2506424320
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.439237752979005e-05,
      "loss": 2.6483,
      "theoretical_loss": 3.367826937682862,
      "tokens_seen": 2506489856
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.437582750141858e-05,
      "loss": 2.49,
      "theoretical_loss": 3.367819917458699,
      "tokens_seen": 2506555392
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.43592774730471e-05,
      "loss": 2.6566,
      "theoretical_loss": 3.3678128974694754,
      "tokens_seen": 2506620928
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.434272744467562e-05,
      "loss": 2.1686,
      "theoretical_loss": 3.367805877715178,
      "tokens_seen": 2506686464
    },
    {
      "epoch": 0.9,
      "objective/train/avg_token_score": 0.01709447056055069,
      "objective/train/avg_weight": 0.9798818230628967,
      "objective/train/docs_used": 1409433,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.7807059288024902,
      "objective/train/std_weight": 0.09102560579776764,
      "objective/train/theoretical_loss": 3.367798858195792,
      "objective/train/tokens_used": 2527212000,
      "theoretical_loss": 3.367798858195792,
      "tokens_seen": 2506752000
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.432617741630415e-05,
      "loss": 1.7438,
      "theoretical_loss": 3.367798858195792,
      "tokens_seen": 2506752000
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.430962738793267e-05,
      "loss": 2.6117,
      "theoretical_loss": 3.3677918389113035,
      "tokens_seen": 2506817536
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.42930773595612e-05,
      "loss": 2.0791,
      "theoretical_loss": 3.3677848198616984,
      "tokens_seen": 2506883072
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.42765273311897e-05,
      "loss": 2.614,
      "theoretical_loss": 3.3677778010469632,
      "tokens_seen": 2506948608
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.425997730281823e-05,
      "loss": 2.3034,
      "theoretical_loss": 3.3677707824670837,
      "tokens_seen": 2507014144
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.424342727444675e-05,
      "loss": 2.6018,
      "theoretical_loss": 3.3677637641220457,
      "tokens_seen": 2507079680
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.422687724607528e-05,
      "loss": 3.3137,
      "theoretical_loss": 3.3677567460118354,
      "tokens_seen": 2507145216
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.42103272177038e-05,
      "loss": 2.6748,
      "theoretical_loss": 3.3677497281364386,
      "tokens_seen": 2507210752
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.419377718933233e-05,
      "loss": 2.5207,
      "theoretical_loss": 3.3677427104958415,
      "tokens_seen": 2507276288
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.417722716096085e-05,
      "loss": 2.3215,
      "theoretical_loss": 3.3677356930900304,
      "tokens_seen": 2507341824
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.416067713258936e-05,
      "loss": 2.4888,
      "theoretical_loss": 3.3677286759189906,
      "tokens_seen": 2507407360
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.414412710421788e-05,
      "loss": 2.5985,
      "theoretical_loss": 3.367721658982709,
      "tokens_seen": 2507472896
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.412757707584641e-05,
      "loss": 3.363,
      "theoretical_loss": 3.3677146422811703,
      "tokens_seen": 2507538432
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.411102704747493e-05,
      "loss": 2.5505,
      "theoretical_loss": 3.3677076258143614,
      "tokens_seen": 2507603968
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.409447701910346e-05,
      "loss": 2.2774,
      "theoretical_loss": 3.367700609582269,
      "tokens_seen": 2507669504
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.407792699073198e-05,
      "loss": 2.4802,
      "theoretical_loss": 3.3676935935848777,
      "tokens_seen": 2507735040
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.40613769623605e-05,
      "loss": 2.6515,
      "theoretical_loss": 3.367686577822174,
      "tokens_seen": 2507800576
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.404482693398903e-05,
      "loss": 2.5714,
      "theoretical_loss": 3.3676795622941444,
      "tokens_seen": 2507866112
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.402827690561755e-05,
      "loss": 2.5324,
      "theoretical_loss": 3.3676725470007747,
      "tokens_seen": 2507931648
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.401172687724608e-05,
      "loss": 2.5075,
      "theoretical_loss": 3.3676655319420505,
      "tokens_seen": 2507997184
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.39951768488746e-05,
      "loss": 2.6768,
      "theoretical_loss": 3.3676585171179583,
      "tokens_seen": 2508062720
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.397862682050312e-05,
      "loss": 2.4129,
      "theoretical_loss": 3.3676515025284837,
      "tokens_seen": 2508128256
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.396207679213165e-05,
      "loss": 2.4665,
      "theoretical_loss": 3.367644488173613,
      "tokens_seen": 2508193792
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.394552676376017e-05,
      "loss": 2.4651,
      "theoretical_loss": 3.3676374740533324,
      "tokens_seen": 2508259328
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.39289767353887e-05,
      "loss": 2.1162,
      "theoretical_loss": 3.3676304601676277,
      "tokens_seen": 2508324864
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.391242670701721e-05,
      "loss": 2.596,
      "theoretical_loss": 3.3676234465164847,
      "tokens_seen": 2508390400
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.389587667864573e-05,
      "loss": 2.2558,
      "theoretical_loss": 3.3676164330998897,
      "tokens_seen": 2508455936
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.387932665027425e-05,
      "loss": 2.6874,
      "theoretical_loss": 3.367609419917829,
      "tokens_seen": 2508521472
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.386277662190278e-05,
      "loss": 2.0554,
      "theoretical_loss": 3.367602406970288,
      "tokens_seen": 2508587008
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.38462265935313e-05,
      "loss": 2.2512,
      "theoretical_loss": 3.367595394257253,
      "tokens_seen": 2508652544
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.382967656515983e-05,
      "loss": 2.5276,
      "theoretical_loss": 3.3675883817787104,
      "tokens_seen": 2508718080
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.381312653678835e-05,
      "loss": 2.5899,
      "theoretical_loss": 3.367581369534646,
      "tokens_seen": 2508783616
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.379657650841686e-05,
      "loss": 2.332,
      "theoretical_loss": 3.3675743575250454,
      "tokens_seen": 2508849152
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.378002648004539e-05,
      "loss": 2.4819,
      "theoretical_loss": 3.367567345749895,
      "tokens_seen": 2508914688
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.376347645167391e-05,
      "loss": 2.6083,
      "theoretical_loss": 3.3675603342091813,
      "tokens_seen": 2508980224
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.374692642330243e-05,
      "loss": 2.4316,
      "theoretical_loss": 3.3675533229028893,
      "tokens_seen": 2509045760
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.373037639493096e-05,
      "loss": 2.4839,
      "theoretical_loss": 3.367546311831006,
      "tokens_seen": 2509111296
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.371382636655948e-05,
      "loss": 2.4052,
      "theoretical_loss": 3.3675393009935166,
      "tokens_seen": 2509176832
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.3697276338188e-05,
      "loss": 2.8695,
      "theoretical_loss": 3.367532290390408,
      "tokens_seen": 2509242368
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.368072630981653e-05,
      "loss": 2.3432,
      "theoretical_loss": 3.3675252800216655,
      "tokens_seen": 2509307904
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.366417628144505e-05,
      "loss": 2.675,
      "theoretical_loss": 3.3675182698872757,
      "tokens_seen": 2509373440
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.364762625307358e-05,
      "loss": 2.6907,
      "theoretical_loss": 3.3675112599872246,
      "tokens_seen": 2509438976
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.36310762247021e-05,
      "loss": 2.5564,
      "theoretical_loss": 3.367504250321498,
      "tokens_seen": 2509504512
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.361452619633063e-05,
      "loss": 2.2457,
      "theoretical_loss": 3.367497240890082,
      "tokens_seen": 2509570048
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.359797616795915e-05,
      "loss": 2.9412,
      "theoretical_loss": 3.3674902316929622,
      "tokens_seen": 2509635584
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.358142613958767e-05,
      "loss": 2.4718,
      "theoretical_loss": 3.3674832227301255,
      "tokens_seen": 2509701120
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.35648761112162e-05,
      "loss": 2.5084,
      "theoretical_loss": 3.3674762140015577,
      "tokens_seen": 2509766656
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.354832608284471e-05,
      "loss": 2.5883,
      "theoretical_loss": 3.3674692055072444,
      "tokens_seen": 2509832192
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.353177605447323e-05,
      "loss": 2.5855,
      "theoretical_loss": 3.367462197247172,
      "tokens_seen": 2509897728
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.351522602610176e-05,
      "loss": 2.5999,
      "theoretical_loss": 3.367455189221327,
      "tokens_seen": 2509963264
    },
    {
      "epoch": 0.9,
      "objective/train/avg_token_score": 0.012440720573067665,
      "objective/train/avg_weight": 0.985359787940979,
      "objective/train/docs_used": 1411323,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.811461925506592,
      "objective/train/std_weight": 0.06410184502601624,
      "objective/train/theoretical_loss": 3.367448181429695,
      "objective/train/tokens_used": 2530488800,
      "theoretical_loss": 3.367448181429695,
      "tokens_seen": 2510028800
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.349867599773028e-05,
      "loss": 2.8135,
      "theoretical_loss": 3.367448181429695,
      "tokens_seen": 2510028800
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.34821259693588e-05,
      "loss": 2.4682,
      "theoretical_loss": 3.3674411738722614,
      "tokens_seen": 2510094336
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.346557594098733e-05,
      "loss": 3.3066,
      "theoretical_loss": 3.367434166549013,
      "tokens_seen": 2510159872
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.344902591261585e-05,
      "loss": 2.6948,
      "theoretical_loss": 3.367427159459936,
      "tokens_seen": 2510225408
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.343247588424438e-05,
      "loss": 2.4408,
      "theoretical_loss": 3.3674201526050163,
      "tokens_seen": 2510290944
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.341592585587289e-05,
      "loss": 2.6824,
      "theoretical_loss": 3.36741314598424,
      "tokens_seen": 2510356480
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.339937582750141e-05,
      "loss": 2.6526,
      "theoretical_loss": 3.3674061395975925,
      "tokens_seen": 2510422016
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.338282579912993e-05,
      "loss": 2.6288,
      "theoretical_loss": 3.3673991334450606,
      "tokens_seen": 2510487552
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.336627577075846e-05,
      "loss": 2.5191,
      "theoretical_loss": 3.3673921275266303,
      "tokens_seen": 2510553088
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.334972574238698e-05,
      "loss": 2.4127,
      "theoretical_loss": 3.367385121842288,
      "tokens_seen": 2510618624
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.33331757140155e-05,
      "loss": 2.0997,
      "theoretical_loss": 3.3673781163920187,
      "tokens_seen": 2510684160
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.331662568564403e-05,
      "loss": 2.7469,
      "theoretical_loss": 3.3673711111758093,
      "tokens_seen": 2510749696
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.330007565727255e-05,
      "loss": 2.6157,
      "theoretical_loss": 3.3673641061936457,
      "tokens_seen": 2510815232
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.328352562890108e-05,
      "loss": 2.4989,
      "theoretical_loss": 3.3673571014455135,
      "tokens_seen": 2510880768
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.32669756005296e-05,
      "loss": 2.4805,
      "theoretical_loss": 3.3673500969314,
      "tokens_seen": 2510946304
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.325042557215813e-05,
      "loss": 2.516,
      "theoretical_loss": 3.3673430926512897,
      "tokens_seen": 2511011840
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.323387554378665e-05,
      "loss": 2.5792,
      "theoretical_loss": 3.3673360886051698,
      "tokens_seen": 2511077376
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.321732551541517e-05,
      "loss": 2.4944,
      "theoretical_loss": 3.367329084793026,
      "tokens_seen": 2511142912
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.32007754870437e-05,
      "loss": 1.4413,
      "theoretical_loss": 3.3673220812148443,
      "tokens_seen": 2511208448
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.318422545867222e-05,
      "loss": 2.4761,
      "theoretical_loss": 3.3673150778706114,
      "tokens_seen": 2511273984
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.316767543030073e-05,
      "loss": 1.9828,
      "theoretical_loss": 3.367308074760312,
      "tokens_seen": 2511339520
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.315112540192926e-05,
      "loss": 2.5395,
      "theoretical_loss": 3.367301071883934,
      "tokens_seen": 2511405056
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.313457537355778e-05,
      "loss": 2.6101,
      "theoretical_loss": 3.3672940692414617,
      "tokens_seen": 2511470592
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.31180253451863e-05,
      "loss": 2.7294,
      "theoretical_loss": 3.3672870668328825,
      "tokens_seen": 2511536128
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.310147531681483e-05,
      "loss": 2.5104,
      "theoretical_loss": 3.367280064658182,
      "tokens_seen": 2511601664
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.308492528844335e-05,
      "loss": 2.6472,
      "theoretical_loss": 3.367273062717346,
      "tokens_seen": 2511667200
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.306837526007188e-05,
      "loss": 2.6874,
      "theoretical_loss": 3.367266061010361,
      "tokens_seen": 2511732736
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.305182523170039e-05,
      "loss": 2.6297,
      "theoretical_loss": 3.367259059537213,
      "tokens_seen": 2511798272
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.303527520332891e-05,
      "loss": 2.5504,
      "theoretical_loss": 3.367252058297888,
      "tokens_seen": 2511863808
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.301872517495743e-05,
      "loss": 2.5265,
      "theoretical_loss": 3.3672450572923722,
      "tokens_seen": 2511929344
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.300217514658596e-05,
      "loss": 2.5206,
      "theoretical_loss": 3.3672380565206517,
      "tokens_seen": 2511994880
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.298562511821448e-05,
      "loss": 2.4082,
      "theoretical_loss": 3.367231055982712,
      "tokens_seen": 2512060416
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.2969075089843e-05,
      "loss": 2.5642,
      "theoretical_loss": 3.36722405567854,
      "tokens_seen": 2512125952
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.295252506147153e-05,
      "loss": 2.5401,
      "theoretical_loss": 3.367217055608122,
      "tokens_seen": 2512191488
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.293597503310005e-05,
      "loss": 1.7952,
      "theoretical_loss": 3.367210055771443,
      "tokens_seen": 2512257024
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.291942500472858e-05,
      "loss": 2.6973,
      "theoretical_loss": 3.3672030561684902,
      "tokens_seen": 2512322560
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.29028749763571e-05,
      "loss": 2.5927,
      "theoretical_loss": 3.3671960567992487,
      "tokens_seen": 2512388096
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.288632494798563e-05,
      "loss": 2.3319,
      "theoretical_loss": 3.3671890576637056,
      "tokens_seen": 2512453632
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.286977491961415e-05,
      "loss": 2.5988,
      "theoretical_loss": 3.367182058761846,
      "tokens_seen": 2512519168
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.285322489124267e-05,
      "loss": 2.5608,
      "theoretical_loss": 3.367175060093657,
      "tokens_seen": 2512584704
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.28366748628712e-05,
      "loss": 2.6262,
      "theoretical_loss": 3.367168061659124,
      "tokens_seen": 2512650240
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.282012483449972e-05,
      "loss": 2.297,
      "theoretical_loss": 3.3671610634582336,
      "tokens_seen": 2512715776
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.280357480612823e-05,
      "loss": 2.7395,
      "theoretical_loss": 3.367154065490971,
      "tokens_seen": 2512781312
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.278702477775676e-05,
      "loss": 2.7821,
      "theoretical_loss": 3.3671470677573234,
      "tokens_seen": 2512846848
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.277047474938528e-05,
      "loss": 2.5672,
      "theoretical_loss": 3.3671400702572765,
      "tokens_seen": 2512912384
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.27539247210138e-05,
      "loss": 2.5956,
      "theoretical_loss": 3.367133072990816,
      "tokens_seen": 2512977920
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.273737469264233e-05,
      "loss": 2.3984,
      "theoretical_loss": 3.3671260759579287,
      "tokens_seen": 2513043456
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.272082466427085e-05,
      "loss": 2.437,
      "theoretical_loss": 3.3671190791586003,
      "tokens_seen": 2513108992
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.270427463589938e-05,
      "loss": 2.5474,
      "theoretical_loss": 3.3671120825928167,
      "tokens_seen": 2513174528
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.268772460752789e-05,
      "loss": 2.6437,
      "theoretical_loss": 3.367105086260565,
      "tokens_seen": 2513240064
    },
    {
      "epoch": 0.9,
      "objective/train/avg_token_score": 0.02077416703104973,
      "objective/train/avg_weight": 0.9755679965019226,
      "objective/train/docs_used": 1413432,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.6178030967712402,
      "objective/train/std_weight": 0.11042772978544235,
      "objective/train/theoretical_loss": 3.36709809016183,
      "objective/train/tokens_used": 2533765600,
      "theoretical_loss": 3.36709809016183,
      "tokens_seen": 2513305600
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.267117457915641e-05,
      "loss": 2.3826,
      "theoretical_loss": 3.36709809016183,
      "tokens_seen": 2513305600
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.265462455078494e-05,
      "loss": 2.5088,
      "theoretical_loss": 3.367091094296599,
      "tokens_seen": 2513371136
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.263807452241346e-05,
      "loss": 2.8004,
      "theoretical_loss": 3.367084098664857,
      "tokens_seen": 2513436672
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.262152449404198e-05,
      "loss": 2.6462,
      "theoretical_loss": 3.367077103266591,
      "tokens_seen": 2513502208
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.260497446567051e-05,
      "loss": 2.7065,
      "theoretical_loss": 3.367070108101786,
      "tokens_seen": 2513567744
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.258842443729903e-05,
      "loss": 2.7457,
      "theoretical_loss": 3.36706311317043,
      "tokens_seen": 2513633280
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.257187440892756e-05,
      "loss": 2.3473,
      "theoretical_loss": 3.3670561184725076,
      "tokens_seen": 2513698816
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.255532438055608e-05,
      "loss": 2.4642,
      "theoretical_loss": 3.3670491240080054,
      "tokens_seen": 2513764352
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.25387743521846e-05,
      "loss": 2.4085,
      "theoretical_loss": 3.3670421297769093,
      "tokens_seen": 2513829888
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.252222432381313e-05,
      "loss": 2.3478,
      "theoretical_loss": 3.367035135779206,
      "tokens_seen": 2513895424
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.250567429544165e-05,
      "loss": 2.6868,
      "theoretical_loss": 3.367028142014881,
      "tokens_seen": 2513960960
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.248912426707018e-05,
      "loss": 2.7095,
      "theoretical_loss": 3.3670211484839205,
      "tokens_seen": 2514026496
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.24725742386987e-05,
      "loss": 2.5911,
      "theoretical_loss": 3.367014155186311,
      "tokens_seen": 2514092032
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.245602421032722e-05,
      "loss": 2.2944,
      "theoretical_loss": 3.3670071621220385,
      "tokens_seen": 2514157568
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.243947418195573e-05,
      "loss": 2.3041,
      "theoretical_loss": 3.367000169291089,
      "tokens_seen": 2514223104
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.242292415358426e-05,
      "loss": 2.4684,
      "theoretical_loss": 3.366993176693449,
      "tokens_seen": 2514288640
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.240637412521278e-05,
      "loss": 2.3902,
      "theoretical_loss": 3.3669861843291042,
      "tokens_seen": 2514354176
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.23898240968413e-05,
      "loss": 2.4673,
      "theoretical_loss": 3.3669791921980408,
      "tokens_seen": 2514419712
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.237327406846983e-05,
      "loss": 2.5985,
      "theoretical_loss": 3.366972200300245,
      "tokens_seen": 2514485248
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.235672404009835e-05,
      "loss": 2.737,
      "theoretical_loss": 3.366965208635703,
      "tokens_seen": 2514550784
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.234017401172688e-05,
      "loss": 2.4042,
      "theoretical_loss": 3.3669582172044006,
      "tokens_seen": 2514616320
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.23236239833554e-05,
      "loss": 2.5802,
      "theoretical_loss": 3.3669512260063246,
      "tokens_seen": 2514681856
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.230707395498391e-05,
      "loss": 2.526,
      "theoretical_loss": 3.3669442350414602,
      "tokens_seen": 2514747392
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.229052392661244e-05,
      "loss": 2.2642,
      "theoretical_loss": 3.3669372443097947,
      "tokens_seen": 2514812928
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.227397389824096e-05,
      "loss": 2.5814,
      "theoretical_loss": 3.3669302538113137,
      "tokens_seen": 2514878464
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.225742386986948e-05,
      "loss": 2.7207,
      "theoretical_loss": 3.366923263546003,
      "tokens_seen": 2514944000
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.224087384149802e-05,
      "loss": 2.7591,
      "theoretical_loss": 3.366916273513849,
      "tokens_seen": 2515009536
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.222432381312655e-05,
      "loss": 2.5621,
      "theoretical_loss": 3.366909283714838,
      "tokens_seen": 2515075072
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.220777378475507e-05,
      "loss": 2.2809,
      "theoretical_loss": 3.3669022941489564,
      "tokens_seen": 2515140608
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.219122375638358e-05,
      "loss": 2.5643,
      "theoretical_loss": 3.3668953048161896,
      "tokens_seen": 2515206144
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.21746737280121e-05,
      "loss": 2.5099,
      "theoretical_loss": 3.3668883157165244,
      "tokens_seen": 2515271680
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.215812369964063e-05,
      "loss": 2.2905,
      "theoretical_loss": 3.3668813268499465,
      "tokens_seen": 2515337216
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.214157367126915e-05,
      "loss": 2.5231,
      "theoretical_loss": 3.366874338216442,
      "tokens_seen": 2515402752
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.212502364289768e-05,
      "loss": 1.9906,
      "theoretical_loss": 3.3668673498159976,
      "tokens_seen": 2515468288
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.21084736145262e-05,
      "loss": 2.5777,
      "theoretical_loss": 3.3668603616485995,
      "tokens_seen": 2515533824
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.209192358615472e-05,
      "loss": 2.5828,
      "theoretical_loss": 3.366853373714233,
      "tokens_seen": 2515599360
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.207537355778323e-05,
      "loss": 2.077,
      "theoretical_loss": 3.366846386012885,
      "tokens_seen": 2515664896
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.205882352941176e-05,
      "loss": 2.6326,
      "theoretical_loss": 3.3668393985445415,
      "tokens_seen": 2515730432
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.204227350104028e-05,
      "loss": 2.0248,
      "theoretical_loss": 3.3668324113091885,
      "tokens_seen": 2515795968
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.20257234726688e-05,
      "loss": 2.7141,
      "theoretical_loss": 3.3668254243068123,
      "tokens_seen": 2515861504
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.200917344429733e-05,
      "loss": 2.5021,
      "theoretical_loss": 3.366818437537399,
      "tokens_seen": 2515927040
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.199262341592585e-05,
      "loss": 1.9408,
      "theoretical_loss": 3.366811451000935,
      "tokens_seen": 2515992576
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.197607338755438e-05,
      "loss": 2.4873,
      "theoretical_loss": 3.366804464697406,
      "tokens_seen": 2516058112
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.19595233591829e-05,
      "loss": 2.5665,
      "theoretical_loss": 3.3667974786267987,
      "tokens_seen": 2516123648
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.194297333081141e-05,
      "loss": 2.6803,
      "theoretical_loss": 3.366790492789099,
      "tokens_seen": 2516189184
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.192642330243995e-05,
      "loss": 2.263,
      "theoretical_loss": 3.3667835071842926,
      "tokens_seen": 2516254720
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.190987327406847e-05,
      "loss": 2.7166,
      "theoretical_loss": 3.3667765218123664,
      "tokens_seen": 2516320256
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.1893323245697e-05,
      "loss": 1.7128,
      "theoretical_loss": 3.3667695366733064,
      "tokens_seen": 2516385792
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.187677321732552e-05,
      "loss": 2.7815,
      "theoretical_loss": 3.3667625517670987,
      "tokens_seen": 2516451328
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.186022318895405e-05,
      "loss": 2.5516,
      "theoretical_loss": 3.3667555670937297,
      "tokens_seen": 2516516864
    },
    {
      "epoch": 0.9,
      "objective/train/avg_token_score": 0.010843781754374504,
      "objective/train/avg_weight": 0.9872377514839172,
      "objective/train/docs_used": 1415580,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6447348594665527,
      "objective/train/std_weight": 0.06308945268392563,
      "objective/train/theoretical_loss": 3.3667485826531847,
      "objective/train/tokens_used": 2537042400,
      "theoretical_loss": 3.3667485826531847,
      "tokens_seen": 2516582400
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.184367316058257e-05,
      "loss": 2.5257,
      "theoretical_loss": 3.3667485826531847,
      "tokens_seen": 2516582400
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.182712313221108e-05,
      "loss": 2.6017,
      "theoretical_loss": 3.3667415984454507,
      "tokens_seen": 2516647936
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.18105731038396e-05,
      "loss": 2.5615,
      "theoretical_loss": 3.3667346144705137,
      "tokens_seen": 2516713472
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.179402307546813e-05,
      "loss": 2.5773,
      "theoretical_loss": 3.3667276307283602,
      "tokens_seen": 2516779008
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.177747304709665e-05,
      "loss": 2.5049,
      "theoretical_loss": 3.3667206472189757,
      "tokens_seen": 2516844544
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.176092301872518e-05,
      "loss": 2.2086,
      "theoretical_loss": 3.366713663942347,
      "tokens_seen": 2516910080
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.17443729903537e-05,
      "loss": 2.7796,
      "theoretical_loss": 3.3667066808984596,
      "tokens_seen": 2516975616
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.172782296198222e-05,
      "loss": 2.4698,
      "theoretical_loss": 3.3666996980873005,
      "tokens_seen": 2517041152
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.171127293361075e-05,
      "loss": 2.5083,
      "theoretical_loss": 3.3666927155088553,
      "tokens_seen": 2517106688
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.169472290523926e-05,
      "loss": 2.4306,
      "theoretical_loss": 3.36668573316311,
      "tokens_seen": 2517172224
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.167817287686778e-05,
      "loss": 2.5726,
      "theoretical_loss": 3.3666787510500518,
      "tokens_seen": 2517237760
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.166162284849631e-05,
      "loss": 1.5534,
      "theoretical_loss": 3.366671769169666,
      "tokens_seen": 2517303296
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.164507282012483e-05,
      "loss": 2.5712,
      "theoretical_loss": 3.366664787521939,
      "tokens_seen": 2517368832
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.162852279175336e-05,
      "loss": 2.5676,
      "theoretical_loss": 3.366657806106857,
      "tokens_seen": 2517434368
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.161197276338188e-05,
      "loss": 2.6144,
      "theoretical_loss": 3.3666508249244056,
      "tokens_seen": 2517499904
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.15954227350104e-05,
      "loss": 2.4934,
      "theoretical_loss": 3.3666438439745723,
      "tokens_seen": 2517565440
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.157887270663893e-05,
      "loss": 2.4978,
      "theoretical_loss": 3.3666368632573427,
      "tokens_seen": 2517630976
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.156232267826745e-05,
      "loss": 2.6096,
      "theoretical_loss": 3.366629882772702,
      "tokens_seen": 2517696512
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.154577264989598e-05,
      "loss": 2.5418,
      "theoretical_loss": 3.3666229025206382,
      "tokens_seen": 2517762048
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.15292226215245e-05,
      "loss": 2.7211,
      "theoretical_loss": 3.3666159225011363,
      "tokens_seen": 2517827584
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.151267259315302e-05,
      "loss": 2.1374,
      "theoretical_loss": 3.3666089427141825,
      "tokens_seen": 2517893120
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.149612256478155e-05,
      "loss": 2.5865,
      "theoretical_loss": 3.3666019631597637,
      "tokens_seen": 2517958656
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.147957253641007e-05,
      "loss": 2.3079,
      "theoretical_loss": 3.3665949838378655,
      "tokens_seen": 2518024192
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.147957253641007e-05,
      "loss": 6.7965,
      "theoretical_loss": 3.3665880047484746,
      "tokens_seen": 2518089728
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.14630225080386e-05,
      "loss": 2.6198,
      "theoretical_loss": 3.3665810258915765,
      "tokens_seen": 2518155264
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.14464724796671e-05,
      "loss": 2.4176,
      "theoretical_loss": 3.3665740472671577,
      "tokens_seen": 2518220800
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.142992245129563e-05,
      "loss": 2.4079,
      "theoretical_loss": 3.366567068875205,
      "tokens_seen": 2518286336
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.141337242292415e-05,
      "loss": 2.3647,
      "theoretical_loss": 3.366560090715703,
      "tokens_seen": 2518351872
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.139682239455268e-05,
      "loss": 2.7985,
      "theoretical_loss": 3.3665531127886403,
      "tokens_seen": 2518417408
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.13802723661812e-05,
      "loss": 2.3101,
      "theoretical_loss": 3.3665461350940014,
      "tokens_seen": 2518482944
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.136372233780973e-05,
      "loss": 2.4664,
      "theoretical_loss": 3.3665391576317725,
      "tokens_seen": 2518548480
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.134717230943825e-05,
      "loss": 2.5319,
      "theoretical_loss": 3.366532180401941,
      "tokens_seen": 2518614016
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.133062228106676e-05,
      "loss": 2.2161,
      "theoretical_loss": 3.3665252034044917,
      "tokens_seen": 2518679552
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.131407225269528e-05,
      "loss": 2.597,
      "theoretical_loss": 3.3665182266394114,
      "tokens_seen": 2518745088
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.129752222432381e-05,
      "loss": 2.1973,
      "theoretical_loss": 3.366511250106687,
      "tokens_seen": 2518810624
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.128097219595233e-05,
      "loss": 2.4354,
      "theoretical_loss": 3.366504273806304,
      "tokens_seen": 2518876160
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.126442216758086e-05,
      "loss": 2.7098,
      "theoretical_loss": 3.366497297738248,
      "tokens_seen": 2518941696
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.124787213920938e-05,
      "loss": 2.7636,
      "theoretical_loss": 3.3664903219025066,
      "tokens_seen": 2519007232
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.12313221108379e-05,
      "loss": 2.4859,
      "theoretical_loss": 3.366483346299065,
      "tokens_seen": 2519072768
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.121477208246643e-05,
      "loss": 2.6883,
      "theoretical_loss": 3.36647637092791,
      "tokens_seen": 2519138304
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.119822205409495e-05,
      "loss": 3.0338,
      "theoretical_loss": 3.3664693957890277,
      "tokens_seen": 2519203840
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.118167202572348e-05,
      "loss": 2.7965,
      "theoretical_loss": 3.366462420882404,
      "tokens_seen": 2519269376
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.1165121997352e-05,
      "loss": 2.3835,
      "theoretical_loss": 3.3664554462080254,
      "tokens_seen": 2519334912
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.114857196898052e-05,
      "loss": 2.416,
      "theoretical_loss": 3.366448471765878,
      "tokens_seen": 2519400448
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.113202194060905e-05,
      "loss": 2.3939,
      "theoretical_loss": 3.3664414975559485,
      "tokens_seen": 2519465984
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.111547191223757e-05,
      "loss": 2.3636,
      "theoretical_loss": 3.366434523578222,
      "tokens_seen": 2519531520
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.10989218838661e-05,
      "loss": 2.5239,
      "theoretical_loss": 3.366427549832686,
      "tokens_seen": 2519597056
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.10823718554946e-05,
      "loss": 1.7169,
      "theoretical_loss": 3.3664205763193262,
      "tokens_seen": 2519662592
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.106582182712313e-05,
      "loss": 2.5969,
      "theoretical_loss": 3.3664136030381284,
      "tokens_seen": 2519728128
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.104927179875165e-05,
      "loss": 2.556,
      "theoretical_loss": 3.3664066299890796,
      "tokens_seen": 2519793664
    },
    {
      "epoch": 0.9,
      "objective/train/avg_token_score": 0.011458439752459526,
      "objective/train/avg_weight": 0.9865118861198425,
      "objective/train/docs_used": 1417591,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.7381911277771,
      "objective/train/std_weight": 0.08198677748441696,
      "objective/train/theoretical_loss": 3.366399657172166,
      "objective/train/tokens_used": 2540319200,
      "theoretical_loss": 3.366399657172166,
      "tokens_seen": 2519859200
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.103272177038018e-05,
      "loss": 2.4841,
      "theoretical_loss": 3.366399657172166,
      "tokens_seen": 2519859200
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.10161717420087e-05,
      "loss": 3.026,
      "theoretical_loss": 3.366392684587373,
      "tokens_seen": 2519924736
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.099962171363723e-05,
      "loss": 2.5463,
      "theoretical_loss": 3.3663857122346874,
      "tokens_seen": 2519990272
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.098307168526575e-05,
      "loss": 3.8688,
      "theoretical_loss": 3.3663787401140954,
      "tokens_seen": 2520055808
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.096652165689426e-05,
      "loss": 2.484,
      "theoretical_loss": 3.3663717682255836,
      "tokens_seen": 2520121344
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.094997162852278e-05,
      "loss": 3.6661,
      "theoretical_loss": 3.3663647965691377,
      "tokens_seen": 2520186880
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.093342160015131e-05,
      "loss": 2.6155,
      "theoretical_loss": 3.3663578251447435,
      "tokens_seen": 2520252416
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.091687157177983e-05,
      "loss": 2.3886,
      "theoretical_loss": 3.3663508539523885,
      "tokens_seen": 2520317952
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.090032154340836e-05,
      "loss": 2.4453,
      "theoretical_loss": 3.3663438829920582,
      "tokens_seen": 2520383488
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.088377151503688e-05,
      "loss": 2.5401,
      "theoretical_loss": 3.3663369122637388,
      "tokens_seen": 2520449024
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.08672214866654e-05,
      "loss": 2.1429,
      "theoretical_loss": 3.366329941767417,
      "tokens_seen": 2520514560
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.085067145829393e-05,
      "loss": 2.5408,
      "theoretical_loss": 3.3663229715030787,
      "tokens_seen": 2520580096
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.083412142992245e-05,
      "loss": 2.4171,
      "theoretical_loss": 3.3663160014707096,
      "tokens_seen": 2520645632
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.081757140155098e-05,
      "loss": 2.477,
      "theoretical_loss": 3.3663090316702973,
      "tokens_seen": 2520711168
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.08010213731795e-05,
      "loss": 2.5391,
      "theoretical_loss": 3.366302062101827,
      "tokens_seen": 2520776704
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.078447134480802e-05,
      "loss": 2.3586,
      "theoretical_loss": 3.366295092765285,
      "tokens_seen": 2520842240
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.076792131643655e-05,
      "loss": 2.4371,
      "theoretical_loss": 3.366288123660658,
      "tokens_seen": 2520907776
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.075137128806507e-05,
      "loss": 2.6521,
      "theoretical_loss": 3.3662811547879317,
      "tokens_seen": 2520973312
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.07348212596936e-05,
      "loss": 2.61,
      "theoretical_loss": 3.366274186147093,
      "tokens_seen": 2521038848
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.071827123132211e-05,
      "loss": 2.4893,
      "theoretical_loss": 3.366267217738128,
      "tokens_seen": 2521104384
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.070172120295063e-05,
      "loss": 2.5336,
      "theoretical_loss": 3.366260249561023,
      "tokens_seen": 2521169920
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.068517117457915e-05,
      "loss": 2.462,
      "theoretical_loss": 3.3662532816157635,
      "tokens_seen": 2521235456
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.066862114620768e-05,
      "loss": 2.4859,
      "theoretical_loss": 3.3662463139023364,
      "tokens_seen": 2521300992
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.06520711178362e-05,
      "loss": 2.1763,
      "theoretical_loss": 3.3662393464207283,
      "tokens_seen": 2521366528
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.063552108946473e-05,
      "loss": 2.3385,
      "theoretical_loss": 3.366232379170925,
      "tokens_seen": 2521432064
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.061897106109325e-05,
      "loss": 2.5448,
      "theoretical_loss": 3.3662254121529123,
      "tokens_seen": 2521497600
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.060242103272177e-05,
      "loss": 2.5165,
      "theoretical_loss": 3.3662184453666777,
      "tokens_seen": 2521563136
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.058587100435029e-05,
      "loss": 2.3594,
      "theoretical_loss": 3.366211478812206,
      "tokens_seen": 2521628672
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.056932097597881e-05,
      "loss": 2.6575,
      "theoretical_loss": 3.3662045124894844,
      "tokens_seen": 2521694208
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.055277094760733e-05,
      "loss": 2.6466,
      "theoretical_loss": 3.3661975463984994,
      "tokens_seen": 2521759744
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.053622091923586e-05,
      "loss": 2.6909,
      "theoretical_loss": 3.3661905805392367,
      "tokens_seen": 2521825280
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.051967089086438e-05,
      "loss": 2.5855,
      "theoretical_loss": 3.3661836149116824,
      "tokens_seen": 2521890816
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.05031208624929e-05,
      "loss": 2.5917,
      "theoretical_loss": 3.3661766495158236,
      "tokens_seen": 2521956352
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.048657083412143e-05,
      "loss": 2.4095,
      "theoretical_loss": 3.3661696843516458,
      "tokens_seen": 2522021888
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.047002080574995e-05,
      "loss": 2.4936,
      "theoretical_loss": 3.3661627194191355,
      "tokens_seen": 2522087424
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.045347077737848e-05,
      "loss": 2.7738,
      "theoretical_loss": 3.366155754718279,
      "tokens_seen": 2522152960
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.0436920749007e-05,
      "loss": 3.1794,
      "theoretical_loss": 3.3661487902490626,
      "tokens_seen": 2522218496
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.042037072063553e-05,
      "loss": 2.5599,
      "theoretical_loss": 3.3661418260114724,
      "tokens_seen": 2522284032
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.040382069226405e-05,
      "loss": 2.6233,
      "theoretical_loss": 3.366134862005495,
      "tokens_seen": 2522349568
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.038727066389257e-05,
      "loss": 2.6169,
      "theoretical_loss": 3.3661278982311167,
      "tokens_seen": 2522415104
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.03707206355211e-05,
      "loss": 2.3367,
      "theoretical_loss": 3.366120934688323,
      "tokens_seen": 2522480640
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.035417060714961e-05,
      "loss": 2.9233,
      "theoretical_loss": 3.3661139713771018,
      "tokens_seen": 2522546176
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.033762057877813e-05,
      "loss": 2.5348,
      "theoretical_loss": 3.366107008297438,
      "tokens_seen": 2522611712
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.032107055040666e-05,
      "loss": 2.4346,
      "theoretical_loss": 3.3661000454493175,
      "tokens_seen": 2522677248
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.030452052203518e-05,
      "loss": 2.5716,
      "theoretical_loss": 3.366093082832728,
      "tokens_seen": 2522742784
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.02879704936637e-05,
      "loss": 2.3248,
      "theoretical_loss": 3.3660861204476555,
      "tokens_seen": 2522808320
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.027142046529223e-05,
      "loss": 2.5962,
      "theoretical_loss": 3.3660791582940854,
      "tokens_seen": 2522873856
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.025487043692075e-05,
      "loss": 2.2682,
      "theoretical_loss": 3.3660721963720044,
      "tokens_seen": 2522939392
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.023832040854928e-05,
      "loss": 2.6832,
      "theoretical_loss": 3.366065234681399,
      "tokens_seen": 2523004928
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.022177038017779e-05,
      "loss": 2.5513,
      "theoretical_loss": 3.3660582732222553,
      "tokens_seen": 2523070464
    },
    {
      "epoch": 0.9,
      "objective/train/avg_token_score": 0.016614004969596863,
      "objective/train/avg_weight": 0.9804458618164062,
      "objective/train/docs_used": 1418957,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7174596786499023,
      "objective/train/std_weight": 0.08568922430276871,
      "objective/train/theoretical_loss": 3.36605131199456,
      "objective/train/tokens_used": 2543596000,
      "theoretical_loss": 3.36605131199456,
      "tokens_seen": 2523136000
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.020522035180631e-05,
      "loss": 2.399,
      "theoretical_loss": 3.36605131199456,
      "tokens_seen": 2523136000
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.018867032343483e-05,
      "loss": 2.4319,
      "theoretical_loss": 3.366044350998299,
      "tokens_seen": 2523201536
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.017212029506336e-05,
      "loss": 2.3956,
      "theoretical_loss": 3.3660373902334584,
      "tokens_seen": 2523267072
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.015557026669188e-05,
      "loss": 2.4963,
      "theoretical_loss": 3.366030429700025,
      "tokens_seen": 2523332608
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.01390202383204e-05,
      "loss": 2.5592,
      "theoretical_loss": 3.3660234693979847,
      "tokens_seen": 2523398144
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.012247020994893e-05,
      "loss": 2.391,
      "theoretical_loss": 3.366016509327324,
      "tokens_seen": 2523463680
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.010592018157745e-05,
      "loss": 2.3024,
      "theoretical_loss": 3.3660095494880293,
      "tokens_seen": 2523529216
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.008937015320598e-05,
      "loss": 2.4754,
      "theoretical_loss": 3.366002589880087,
      "tokens_seen": 2523594752
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.00728201248345e-05,
      "loss": 2.1759,
      "theoretical_loss": 3.3659956305034826,
      "tokens_seen": 2523660288
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.005627009646303e-05,
      "loss": 2.5147,
      "theoretical_loss": 3.365988671358203,
      "tokens_seen": 2523725824
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.003972006809155e-05,
      "loss": 2.5255,
      "theoretical_loss": 3.365981712444235,
      "tokens_seen": 2523791360
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.002317003972007e-05,
      "loss": 2.5912,
      "theoretical_loss": 3.365974753761564,
      "tokens_seen": 2523856896
    },
    {
      "epoch": 0.9,
      "learning_rate": 7.00066200113486e-05,
      "loss": 2.1053,
      "theoretical_loss": 3.3659677953101763,
      "tokens_seen": 2523922432
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.999006998297712e-05,
      "loss": 2.529,
      "theoretical_loss": 3.3659608370900593,
      "tokens_seen": 2523987968
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.997351995460563e-05,
      "loss": 2.4579,
      "theoretical_loss": 3.3659538791011983,
      "tokens_seen": 2524053504
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.995696992623416e-05,
      "loss": 2.5097,
      "theoretical_loss": 3.3659469213435798,
      "tokens_seen": 2524119040
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.994041989786268e-05,
      "loss": 2.6072,
      "theoretical_loss": 3.3659399638171905,
      "tokens_seen": 2524184576
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.99238698694912e-05,
      "loss": 2.4085,
      "theoretical_loss": 3.365933006522016,
      "tokens_seen": 2524250112
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.990731984111973e-05,
      "loss": 2.1133,
      "theoretical_loss": 3.3659260494580434,
      "tokens_seen": 2524315648
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.989076981274825e-05,
      "loss": 2.5663,
      "theoretical_loss": 3.3659190926252585,
      "tokens_seen": 2524381184
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.987421978437678e-05,
      "loss": 2.3625,
      "theoretical_loss": 3.365912136023648,
      "tokens_seen": 2524446720
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.985766975600529e-05,
      "loss": 3.2758,
      "theoretical_loss": 3.365905179653198,
      "tokens_seen": 2524512256
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.984111972763381e-05,
      "loss": 2.5761,
      "theoretical_loss": 3.365898223513894,
      "tokens_seen": 2524577792
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.982456969926233e-05,
      "loss": 2.383,
      "theoretical_loss": 3.3658912676057238,
      "tokens_seen": 2524643328
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.980801967089086e-05,
      "loss": 2.6418,
      "theoretical_loss": 3.3658843119286734,
      "tokens_seen": 2524708864
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.979146964251938e-05,
      "loss": 2.1671,
      "theoretical_loss": 3.365877356482728,
      "tokens_seen": 2524774400
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.97749196141479e-05,
      "loss": 2.3791,
      "theoretical_loss": 3.365870401267875,
      "tokens_seen": 2524839936
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.975836958577643e-05,
      "loss": 2.2215,
      "theoretical_loss": 3.365863446284101,
      "tokens_seen": 2524905472
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.974181955740495e-05,
      "loss": 2.6379,
      "theoretical_loss": 3.365856491531391,
      "tokens_seen": 2524971008
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.972526952903348e-05,
      "loss": 2.5496,
      "theoretical_loss": 3.3658495370097326,
      "tokens_seen": 2525036544
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.9708719500662e-05,
      "loss": 2.5168,
      "theoretical_loss": 3.365842582719111,
      "tokens_seen": 2525102080
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.969216947229053e-05,
      "loss": 2.3409,
      "theoretical_loss": 3.3658356286595135,
      "tokens_seen": 2525167616
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.967561944391905e-05,
      "loss": 2.1859,
      "theoretical_loss": 3.365828674830926,
      "tokens_seen": 2525233152
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.965906941554757e-05,
      "loss": 2.1523,
      "theoretical_loss": 3.365821721233335,
      "tokens_seen": 2525298688
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.96425193871761e-05,
      "loss": 2.3064,
      "theoretical_loss": 3.3658147678667265,
      "tokens_seen": 2525364224
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.962596935880462e-05,
      "loss": 2.203,
      "theoretical_loss": 3.365807814731087,
      "tokens_seen": 2525429760
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.960941933043313e-05,
      "loss": 2.2218,
      "theoretical_loss": 3.3658008618264033,
      "tokens_seen": 2525495296
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.959286930206166e-05,
      "loss": 2.4075,
      "theoretical_loss": 3.3657939091526607,
      "tokens_seen": 2525560832
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.957631927369018e-05,
      "loss": 2.5702,
      "theoretical_loss": 3.3657869567098464,
      "tokens_seen": 2525626368
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.95597692453187e-05,
      "loss": 2.1956,
      "theoretical_loss": 3.365780004497947,
      "tokens_seen": 2525691904
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.954321921694723e-05,
      "loss": 2.2273,
      "theoretical_loss": 3.3657730525169476,
      "tokens_seen": 2525757440
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.952666918857575e-05,
      "loss": 2.5301,
      "theoretical_loss": 3.3657661007668356,
      "tokens_seen": 2525822976
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.951011916020428e-05,
      "loss": 2.4187,
      "theoretical_loss": 3.365759149247597,
      "tokens_seen": 2525888512
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.949356913183279e-05,
      "loss": 2.4288,
      "theoretical_loss": 3.3657521979592184,
      "tokens_seen": 2525954048
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.947701910346131e-05,
      "loss": 2.2611,
      "theoretical_loss": 3.3657452469016853,
      "tokens_seen": 2526019584
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.946046907508984e-05,
      "loss": 2.191,
      "theoretical_loss": 3.3657382960749853,
      "tokens_seen": 2526085120
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.944391904671836e-05,
      "loss": 2.8097,
      "theoretical_loss": 3.365731345479104,
      "tokens_seen": 2526150656
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.942736901834688e-05,
      "loss": 2.5072,
      "theoretical_loss": 3.3657243951140274,
      "tokens_seen": 2526216192
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.941081898997541e-05,
      "loss": 2.6666,
      "theoretical_loss": 3.365717444979743,
      "tokens_seen": 2526281728
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.939426896160393e-05,
      "loss": 2.52,
      "theoretical_loss": 3.3657104950762355,
      "tokens_seen": 2526347264
    },
    {
      "epoch": 0.9,
      "objective/train/avg_token_score": 0.012301560491323471,
      "objective/train/avg_weight": 0.9855243563652039,
      "objective/train/docs_used": 1420931,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.752307891845703,
      "objective/train/std_weight": 0.07627375423908234,
      "objective/train/theoretical_loss": 3.365703545403493,
      "objective/train/tokens_used": 2546872800,
      "theoretical_loss": 3.365703545403493,
      "tokens_seen": 2526412800
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.937771893323246e-05,
      "loss": 2.5587,
      "theoretical_loss": 3.365703545403493,
      "tokens_seen": 2526412800
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.936116890486098e-05,
      "loss": 2.3794,
      "theoretical_loss": 3.3656965959615004,
      "tokens_seen": 2526478336
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.93446188764895e-05,
      "loss": 1.7991,
      "theoretical_loss": 3.365689646750245,
      "tokens_seen": 2526543872
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.932806884811803e-05,
      "loss": 2.4992,
      "theoretical_loss": 3.365682697769713,
      "tokens_seen": 2526609408
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.931151881974655e-05,
      "loss": 2.5537,
      "theoretical_loss": 3.3656757490198905,
      "tokens_seen": 2526674944
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.929496879137508e-05,
      "loss": 2.3631,
      "theoretical_loss": 3.365668800500764,
      "tokens_seen": 2526740480
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.92784187630036e-05,
      "loss": 2.0415,
      "theoretical_loss": 3.3656618522123196,
      "tokens_seen": 2526806016
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.926186873463212e-05,
      "loss": 2.2817,
      "theoretical_loss": 3.3656549041545443,
      "tokens_seen": 2526871552
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.924531870626063e-05,
      "loss": 2.2952,
      "theoretical_loss": 3.365647956327424,
      "tokens_seen": 2526937088
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.922876867788916e-05,
      "loss": 2.5043,
      "theoretical_loss": 3.3656410087309445,
      "tokens_seen": 2527002624
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.921221864951768e-05,
      "loss": 2.08,
      "theoretical_loss": 3.3656340613650935,
      "tokens_seen": 2527068160
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.91956686211462e-05,
      "loss": 2.45,
      "theoretical_loss": 3.365627114229856,
      "tokens_seen": 2527133696
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.917911859277473e-05,
      "loss": 2.5494,
      "theoretical_loss": 3.3656201673252193,
      "tokens_seen": 2527199232
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.916256856440325e-05,
      "loss": 2.4347,
      "theoretical_loss": 3.3656132206511695,
      "tokens_seen": 2527264768
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.914601853603178e-05,
      "loss": 2.74,
      "theoretical_loss": 3.3656062742076926,
      "tokens_seen": 2527330304
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.91294685076603e-05,
      "loss": 2.5054,
      "theoretical_loss": 3.365599327994776,
      "tokens_seen": 2527395840
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.911291847928881e-05,
      "loss": 2.3499,
      "theoretical_loss": 3.3655923820124047,
      "tokens_seen": 2527461376
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.909636845091734e-05,
      "loss": 2.4684,
      "theoretical_loss": 3.3655854362605657,
      "tokens_seen": 2527526912
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.907981842254586e-05,
      "loss": 2.3932,
      "theoretical_loss": 3.365578490739246,
      "tokens_seen": 2527592448
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.906326839417438e-05,
      "loss": 2.5013,
      "theoretical_loss": 3.365571545448431,
      "tokens_seen": 2527657984
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.904671836580291e-05,
      "loss": 2.617,
      "theoretical_loss": 3.3655646003881072,
      "tokens_seen": 2527723520
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.903016833743143e-05,
      "loss": 2.4989,
      "theoretical_loss": 3.365557655558262,
      "tokens_seen": 2527789056
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.901361830905996e-05,
      "loss": 2.19,
      "theoretical_loss": 3.36555071095888,
      "tokens_seen": 2527854592
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.899706828068848e-05,
      "loss": 2.4938,
      "theoretical_loss": 3.3655437665899495,
      "tokens_seen": 2527920128
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.8980518252317e-05,
      "loss": 2.2314,
      "theoretical_loss": 3.365536822451455,
      "tokens_seen": 2527985664
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.896396822394553e-05,
      "loss": 2.4197,
      "theoretical_loss": 3.3655298785433847,
      "tokens_seen": 2528051200
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.894741819557405e-05,
      "loss": 2.3136,
      "theoretical_loss": 3.365522934865724,
      "tokens_seen": 2528116736
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.893086816720258e-05,
      "loss": 2.2984,
      "theoretical_loss": 3.3655159914184587,
      "tokens_seen": 2528182272
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.89143181388311e-05,
      "loss": 2.5314,
      "theoretical_loss": 3.3655090482015764,
      "tokens_seen": 2528247808
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.889776811045962e-05,
      "loss": 2.3603,
      "theoretical_loss": 3.365502105215063,
      "tokens_seen": 2528313344
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.888121808208813e-05,
      "loss": 2.401,
      "theoretical_loss": 3.3654951624589047,
      "tokens_seen": 2528378880
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.886466805371666e-05,
      "loss": 2.3298,
      "theoretical_loss": 3.365488219933088,
      "tokens_seen": 2528444416
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.884811802534518e-05,
      "loss": 2.294,
      "theoretical_loss": 3.3654812776376,
      "tokens_seen": 2528509952
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.88315679969737e-05,
      "loss": 2.368,
      "theoretical_loss": 3.3654743355724257,
      "tokens_seen": 2528575488
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.881501796860223e-05,
      "loss": 2.5814,
      "theoretical_loss": 3.365467393737552,
      "tokens_seen": 2528641024
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.879846794023075e-05,
      "loss": 2.3883,
      "theoretical_loss": 3.3654604521329663,
      "tokens_seen": 2528706560
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.878191791185928e-05,
      "loss": 2.1593,
      "theoretical_loss": 3.3654535107586536,
      "tokens_seen": 2528772096
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.87653678834878e-05,
      "loss": 2.7028,
      "theoretical_loss": 3.365446569614601,
      "tokens_seen": 2528837632
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.874881785511631e-05,
      "loss": 2.5623,
      "theoretical_loss": 3.365439628700795,
      "tokens_seen": 2528903168
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.873226782674484e-05,
      "loss": 2.4141,
      "theoretical_loss": 3.365432688017221,
      "tokens_seen": 2528968704
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.871571779837336e-05,
      "loss": 2.4686,
      "theoretical_loss": 3.3654257475638674,
      "tokens_seen": 2529034240
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.869916777000188e-05,
      "loss": 2.0315,
      "theoretical_loss": 3.3654188073407187,
      "tokens_seen": 2529099776
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.868261774163041e-05,
      "loss": 2.5574,
      "theoretical_loss": 3.365411867347762,
      "tokens_seen": 2529165312
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.866606771325893e-05,
      "loss": 2.3496,
      "theoretical_loss": 3.3654049275849838,
      "tokens_seen": 2529230848
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.864951768488746e-05,
      "loss": 2.4369,
      "theoretical_loss": 3.36539798805237,
      "tokens_seen": 2529296384
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.863296765651598e-05,
      "loss": 2.5757,
      "theoretical_loss": 3.365391048749908,
      "tokens_seen": 2529361920
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.86164176281445e-05,
      "loss": 2.4193,
      "theoretical_loss": 3.365384109677583,
      "tokens_seen": 2529427456
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.859986759977303e-05,
      "loss": 2.3259,
      "theoretical_loss": 3.365377170835382,
      "tokens_seen": 2529492992
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.858331757140155e-05,
      "loss": 2.4128,
      "theoretical_loss": 3.3653702322232917,
      "tokens_seen": 2529558528
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.856676754303008e-05,
      "loss": 2.3981,
      "theoretical_loss": 3.3653632938412983,
      "tokens_seen": 2529624064
    },
    {
      "epoch": 0.9,
      "objective/train/avg_token_score": 0.012975168414413929,
      "objective/train/avg_weight": 0.9847372174263,
      "objective/train/docs_used": 1422953,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.132035255432129,
      "objective/train/std_weight": 0.08461438864469528,
      "objective/train/theoretical_loss": 3.3653563556893875,
      "objective/train/tokens_used": 2550149600,
      "theoretical_loss": 3.3653563556893875,
      "tokens_seen": 2529689600
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.85502175146586e-05,
      "loss": 2.4824,
      "theoretical_loss": 3.3653563556893875,
      "tokens_seen": 2529689600
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.853366748628712e-05,
      "loss": 2.2052,
      "theoretical_loss": 3.365349417767547,
      "tokens_seen": 2529755136
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.851711745791565e-05,
      "loss": 2.2918,
      "theoretical_loss": 3.3653424800757623,
      "tokens_seen": 2529820672
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.850056742954416e-05,
      "loss": 2.4966,
      "theoretical_loss": 3.3653355426140195,
      "tokens_seen": 2529886208
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.848401740117268e-05,
      "loss": 2.2682,
      "theoretical_loss": 3.3653286053823064,
      "tokens_seen": 2529951744
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.846746737280121e-05,
      "loss": 2.3078,
      "theoretical_loss": 3.365321668380608,
      "tokens_seen": 2530017280
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.845091734442973e-05,
      "loss": 2.4855,
      "theoretical_loss": 3.365314731608912,
      "tokens_seen": 2530082816
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.843436731605826e-05,
      "loss": 2.0806,
      "theoretical_loss": 3.3653077950672032,
      "tokens_seen": 2530148352
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.841781728768678e-05,
      "loss": 2.5429,
      "theoretical_loss": 3.3653008587554694,
      "tokens_seen": 2530213888
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.84012672593153e-05,
      "loss": 2.3307,
      "theoretical_loss": 3.3652939226736964,
      "tokens_seen": 2530279424
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.838471723094381e-05,
      "loss": 2.3666,
      "theoretical_loss": 3.3652869868218707,
      "tokens_seen": 2530344960
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.836816720257234e-05,
      "loss": 2.5034,
      "theoretical_loss": 3.3652800511999788,
      "tokens_seen": 2530410496
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.835161717420086e-05,
      "loss": 2.2635,
      "theoretical_loss": 3.3652731158080074,
      "tokens_seen": 2530476032
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.833506714582939e-05,
      "loss": 2.488,
      "theoretical_loss": 3.3652661806459423,
      "tokens_seen": 2530541568
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.831851711745791e-05,
      "loss": 2.4542,
      "theoretical_loss": 3.36525924571377,
      "tokens_seen": 2530607104
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.830196708908643e-05,
      "loss": 2.3024,
      "theoretical_loss": 3.365252311011478,
      "tokens_seen": 2530672640
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.828541706071496e-05,
      "loss": 2.347,
      "theoretical_loss": 3.3652453765390513,
      "tokens_seen": 2530738176
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.826886703234348e-05,
      "loss": 2.6682,
      "theoretical_loss": 3.365238442296477,
      "tokens_seen": 2530803712
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.8252317003972e-05,
      "loss": 2.6899,
      "theoretical_loss": 3.3652315082837414,
      "tokens_seen": 2530869248
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.823576697560053e-05,
      "loss": 2.504,
      "theoretical_loss": 3.3652245745008313,
      "tokens_seen": 2530934784
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.821921694722905e-05,
      "loss": 2.4478,
      "theoretical_loss": 3.3652176409477326,
      "tokens_seen": 2531000320
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.820266691885758e-05,
      "loss": 2.6607,
      "theoretical_loss": 3.3652107076244318,
      "tokens_seen": 2531065856
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.81861168904861e-05,
      "loss": 2.5302,
      "theoretical_loss": 3.3652037745309156,
      "tokens_seen": 2531131392
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.816956686211463e-05,
      "loss": 2.1454,
      "theoretical_loss": 3.3651968416671707,
      "tokens_seen": 2531196928
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.815301683374315e-05,
      "loss": 2.4791,
      "theoretical_loss": 3.3651899090331825,
      "tokens_seen": 2531262464
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.813646680537166e-05,
      "loss": 2.3864,
      "theoretical_loss": 3.3651829766289385,
      "tokens_seen": 2531328000
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.811991677700018e-05,
      "loss": 2.4904,
      "theoretical_loss": 3.365176044454425,
      "tokens_seen": 2531393536
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.810336674862871e-05,
      "loss": 2.3239,
      "theoretical_loss": 3.3651691125096277,
      "tokens_seen": 2531459072
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.808681672025723e-05,
      "loss": 2.5288,
      "theoretical_loss": 3.3651621807945338,
      "tokens_seen": 2531524608
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.807026669188576e-05,
      "loss": 2.5151,
      "theoretical_loss": 3.3651552493091295,
      "tokens_seen": 2531590144
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.805371666351428e-05,
      "loss": 2.5412,
      "theoretical_loss": 3.365148318053401,
      "tokens_seen": 2531655680
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.80371666351428e-05,
      "loss": 2.3861,
      "theoretical_loss": 3.365141387027335,
      "tokens_seen": 2531721216
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.802061660677131e-05,
      "loss": 2.5707,
      "theoretical_loss": 3.365134456230918,
      "tokens_seen": 2531786752
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.800406657839984e-05,
      "loss": 2.4429,
      "theoretical_loss": 3.365127525664136,
      "tokens_seen": 2531852288
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.798751655002836e-05,
      "loss": 2.4623,
      "theoretical_loss": 3.365120595326976,
      "tokens_seen": 2531917824
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.797096652165689e-05,
      "loss": 2.6307,
      "theoretical_loss": 3.365113665219424,
      "tokens_seen": 2531983360
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.795441649328541e-05,
      "loss": 2.3099,
      "theoretical_loss": 3.365106735341467,
      "tokens_seen": 2532048896
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.793786646491393e-05,
      "loss": 2.5491,
      "theoretical_loss": 3.365099805693091,
      "tokens_seen": 2532114432
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.792131643654247e-05,
      "loss": 2.3086,
      "theoretical_loss": 3.365092876274283,
      "tokens_seen": 2532179968
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.7904766408171e-05,
      "loss": 2.0662,
      "theoretical_loss": 3.3650859470850287,
      "tokens_seen": 2532245504
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.78882163797995e-05,
      "loss": 2.5243,
      "theoretical_loss": 3.3650790181253147,
      "tokens_seen": 2532311040
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.787166635142803e-05,
      "loss": 2.5667,
      "theoretical_loss": 3.3650720893951283,
      "tokens_seen": 2532376576
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.785511632305655e-05,
      "loss": 2.3944,
      "theoretical_loss": 3.3650651608944546,
      "tokens_seen": 2532442112
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.783856629468508e-05,
      "loss": 2.3965,
      "theoretical_loss": 3.365058232623281,
      "tokens_seen": 2532507648
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.78220162663136e-05,
      "loss": 2.5573,
      "theoretical_loss": 3.365051304581594,
      "tokens_seen": 2532573184
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.780546623794213e-05,
      "loss": 2.0238,
      "theoretical_loss": 3.3650443767693794,
      "tokens_seen": 2532638720
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.778891620957065e-05,
      "loss": 2.4669,
      "theoretical_loss": 3.365037449186624,
      "tokens_seen": 2532704256
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.777236618119916e-05,
      "loss": 2.3841,
      "theoretical_loss": 3.3650305218333147,
      "tokens_seen": 2532769792
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.775581615282768e-05,
      "loss": 2.4722,
      "theoretical_loss": 3.365023594709437,
      "tokens_seen": 2532835328
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.773926612445621e-05,
      "loss": 2.6349,
      "theoretical_loss": 3.3650166678149787,
      "tokens_seen": 2532900864
    },
    {
      "epoch": 0.9,
      "objective/train/avg_token_score": 0.023488521575927734,
      "objective/train/avg_weight": 0.9724090099334717,
      "objective/train/docs_used": 1425467,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 29.891071319580078,
      "objective/train/std_weight": 0.13901712000370026,
      "objective/train/theoretical_loss": 3.365009741149925,
      "objective/train/tokens_used": 2553426400,
      "theoretical_loss": 3.365009741149925,
      "tokens_seen": 2532966400
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.772271609608473e-05,
      "loss": 3.3653,
      "theoretical_loss": 3.365009741149925,
      "tokens_seen": 2532966400
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.770616606771326e-05,
      "loss": 2.4677,
      "theoretical_loss": 3.365002814714263,
      "tokens_seen": 2533031936
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.768961603934178e-05,
      "loss": 2.5207,
      "theoretical_loss": 3.3649958885079796,
      "tokens_seen": 2533097472
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.76730660109703e-05,
      "loss": 2.3898,
      "theoretical_loss": 3.36498896253106,
      "tokens_seen": 2533163008
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.765651598259883e-05,
      "loss": 2.5106,
      "theoretical_loss": 3.3649820367834917,
      "tokens_seen": 2533228544
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.763996595422734e-05,
      "loss": 2.3823,
      "theoretical_loss": 3.364975111265261,
      "tokens_seen": 2533294080
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.762341592585588e-05,
      "loss": 2.2769,
      "theoretical_loss": 3.3649681859763536,
      "tokens_seen": 2533359616
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.76068658974844e-05,
      "loss": 1.8619,
      "theoretical_loss": 3.3649612609167567,
      "tokens_seen": 2533425152
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.759031586911292e-05,
      "loss": 2.634,
      "theoretical_loss": 3.3649543360864573,
      "tokens_seen": 2533490688
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.757376584074145e-05,
      "loss": 2.5051,
      "theoretical_loss": 3.3649474114854407,
      "tokens_seen": 2533556224
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.755721581236997e-05,
      "loss": 2.5208,
      "theoretical_loss": 3.3649404871136945,
      "tokens_seen": 2533621760
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.75406657839985e-05,
      "loss": 2.5135,
      "theoretical_loss": 3.364933562971204,
      "tokens_seen": 2533687296
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.752411575562701e-05,
      "loss": 2.5122,
      "theoretical_loss": 3.364926639057957,
      "tokens_seen": 2533752832
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.750756572725553e-05,
      "loss": 2.6357,
      "theoretical_loss": 3.3649197153739383,
      "tokens_seen": 2533818368
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.749101569888405e-05,
      "loss": 2.5611,
      "theoretical_loss": 3.364912791919136,
      "tokens_seen": 2533883904
    },
    {
      "epoch": 0.9,
      "learning_rate": 6.747446567051258e-05,
      "loss": 2.726,
      "theoretical_loss": 3.364905868693536,
      "tokens_seen": 2533949440
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.74579156421411e-05,
      "loss": 2.5324,
      "theoretical_loss": 3.3648989456971243,
      "tokens_seen": 2534014976
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.744136561376963e-05,
      "loss": 2.4803,
      "theoretical_loss": 3.3648920229298884,
      "tokens_seen": 2534080512
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.742481558539815e-05,
      "loss": 2.5482,
      "theoretical_loss": 3.364885100391814,
      "tokens_seen": 2534146048
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.740826555702667e-05,
      "loss": 2.3173,
      "theoretical_loss": 3.3648781780828876,
      "tokens_seen": 2534211584
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.739171552865519e-05,
      "loss": 2.5416,
      "theoretical_loss": 3.364871256003096,
      "tokens_seen": 2534277120
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.737516550028371e-05,
      "loss": 2.8271,
      "theoretical_loss": 3.3648643341524256,
      "tokens_seen": 2534342656
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.735861547191223e-05,
      "loss": 2.5672,
      "theoretical_loss": 3.3648574125308626,
      "tokens_seen": 2534408192
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.734206544354076e-05,
      "loss": 2.6168,
      "theoretical_loss": 3.3648504911383936,
      "tokens_seen": 2534473728
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.732551541516928e-05,
      "loss": 2.6578,
      "theoretical_loss": 3.364843569975006,
      "tokens_seen": 2534539264
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.73089653867978e-05,
      "loss": 2.3957,
      "theoretical_loss": 3.3648366490406847,
      "tokens_seen": 2534604800
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.729241535842633e-05,
      "loss": 2.5788,
      "theoretical_loss": 3.3648297283354176,
      "tokens_seen": 2534670336
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.727586533005485e-05,
      "loss": 2.4092,
      "theoretical_loss": 3.3648228078591904,
      "tokens_seen": 2534735872
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.725931530168338e-05,
      "loss": 2.1833,
      "theoretical_loss": 3.3648158876119902,
      "tokens_seen": 2534801408
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.72427652733119e-05,
      "loss": 2.4687,
      "theoretical_loss": 3.3648089675938024,
      "tokens_seen": 2534866944
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.722621524494043e-05,
      "loss": 2.6874,
      "theoretical_loss": 3.364802047804615,
      "tokens_seen": 2534932480
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.720966521656895e-05,
      "loss": 2.2326,
      "theoretical_loss": 3.3647951282444133,
      "tokens_seen": 2534998016
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.719311518819747e-05,
      "loss": 2.6952,
      "theoretical_loss": 3.3647882089131844,
      "tokens_seen": 2535063552
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.7176565159826e-05,
      "loss": 2.5744,
      "theoretical_loss": 3.3647812898109146,
      "tokens_seen": 2535129088
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.716001513145451e-05,
      "loss": 2.0239,
      "theoretical_loss": 3.36477437093759,
      "tokens_seen": 2535194624
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.714346510308303e-05,
      "loss": 2.5532,
      "theoretical_loss": 3.364767452293198,
      "tokens_seen": 2535260160
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.712691507471156e-05,
      "loss": 2.7453,
      "theoretical_loss": 3.3647605338777247,
      "tokens_seen": 2535325696
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.711036504634008e-05,
      "loss": 2.1929,
      "theoretical_loss": 3.3647536156911566,
      "tokens_seen": 2535391232
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.70938150179686e-05,
      "loss": 2.3298,
      "theoretical_loss": 3.36474669773348,
      "tokens_seen": 2535456768
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.707726498959713e-05,
      "loss": 2.3183,
      "theoretical_loss": 3.364739780004682,
      "tokens_seen": 2535522304
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.706071496122565e-05,
      "loss": 2.6227,
      "theoretical_loss": 3.364732862504748,
      "tokens_seen": 2535587840
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.704416493285418e-05,
      "loss": 2.2054,
      "theoretical_loss": 3.3647259452336655,
      "tokens_seen": 2535653376
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.702761490448269e-05,
      "loss": 2.6362,
      "theoretical_loss": 3.3647190281914208,
      "tokens_seen": 2535718912
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.701106487611121e-05,
      "loss": 2.2362,
      "theoretical_loss": 3.3647121113780005,
      "tokens_seen": 2535784448
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.699451484773973e-05,
      "loss": 2.6898,
      "theoretical_loss": 3.364705194793391,
      "tokens_seen": 2535849984
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.697796481936826e-05,
      "loss": 2.6757,
      "theoretical_loss": 3.3646982784375785,
      "tokens_seen": 2535915520
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.696141479099678e-05,
      "loss": 2.5061,
      "theoretical_loss": 3.36469136231055,
      "tokens_seen": 2535981056
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.69448647626253e-05,
      "loss": 2.556,
      "theoretical_loss": 3.3646844464122916,
      "tokens_seen": 2536046592
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.692831473425383e-05,
      "loss": 2.5628,
      "theoretical_loss": 3.36467753074279,
      "tokens_seen": 2536112128
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.691176470588235e-05,
      "loss": 2.6155,
      "theoretical_loss": 3.364670615302032,
      "tokens_seen": 2536177664
    },
    {
      "epoch": 0.91,
      "objective/train/avg_token_score": 0.04254645109176636,
      "objective/train/avg_weight": 0.9499478340148926,
      "objective/train/docs_used": 1426787,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4801108837127686,
      "objective/train/std_weight": 0.14190773665905,
      "objective/train/theoretical_loss": 3.364663700090004,
      "objective/train/tokens_used": 2556703200,
      "theoretical_loss": 3.364663700090004,
      "tokens_seen": 2536243200
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.689521467751088e-05,
      "loss": 2.34,
      "theoretical_loss": 3.364663700090004,
      "tokens_seen": 2536243200
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.68786646491394e-05,
      "loss": 2.2647,
      "theoretical_loss": 3.3646567851066917,
      "tokens_seen": 2536308736
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.686211462076793e-05,
      "loss": 2.3249,
      "theoretical_loss": 3.364649870352083,
      "tokens_seen": 2536374272
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.684556459239645e-05,
      "loss": 2.4787,
      "theoretical_loss": 3.3646429558261635,
      "tokens_seen": 2536439808
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.682901456402497e-05,
      "loss": 2.4785,
      "theoretical_loss": 3.36463604152892,
      "tokens_seen": 2536505344
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.68124645356535e-05,
      "loss": 2.3927,
      "theoretical_loss": 3.364629127460339,
      "tokens_seen": 2536570880
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.679591450728202e-05,
      "loss": 2.4483,
      "theoretical_loss": 3.3646222136204074,
      "tokens_seen": 2536636416
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.677936447891053e-05,
      "loss": 3.0397,
      "theoretical_loss": 3.364615300009111,
      "tokens_seen": 2536701952
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.676281445053906e-05,
      "loss": 2.3249,
      "theoretical_loss": 3.364608386626437,
      "tokens_seen": 2536767488
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.674626442216758e-05,
      "loss": 2.4616,
      "theoretical_loss": 3.3646014734723715,
      "tokens_seen": 2536833024
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.67297143937961e-05,
      "loss": 2.3396,
      "theoretical_loss": 3.3645945605469008,
      "tokens_seen": 2536898560
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.671316436542463e-05,
      "loss": 2.3218,
      "theoretical_loss": 3.3645876478500125,
      "tokens_seen": 2536964096
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.669661433705315e-05,
      "loss": 2.4584,
      "theoretical_loss": 3.364580735381692,
      "tokens_seen": 2537029632
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.668006430868168e-05,
      "loss": 2.4926,
      "theoretical_loss": 3.364573823141926,
      "tokens_seen": 2537095168
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.666351428031019e-05,
      "loss": 2.3555,
      "theoretical_loss": 3.3645669111307024,
      "tokens_seen": 2537160704
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.664696425193871e-05,
      "loss": 2.6152,
      "theoretical_loss": 3.364559999348006,
      "tokens_seen": 2537226240
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.663041422356723e-05,
      "loss": 2.6097,
      "theoretical_loss": 3.364553087793824,
      "tokens_seen": 2537291776
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.661386419519576e-05,
      "loss": 2.6572,
      "theoretical_loss": 3.364546176468143,
      "tokens_seen": 2537357312
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.659731416682428e-05,
      "loss": 2.2885,
      "theoretical_loss": 3.3645392653709494,
      "tokens_seen": 2537422848
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.65807641384528e-05,
      "loss": 2.5925,
      "theoretical_loss": 3.36453235450223,
      "tokens_seen": 2537488384
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.656421411008133e-05,
      "loss": 2.0967,
      "theoretical_loss": 3.364525443861971,
      "tokens_seen": 2537553920
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.654766408170985e-05,
      "loss": 2.112,
      "theoretical_loss": 3.3645185334501595,
      "tokens_seen": 2537619456
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.653111405333838e-05,
      "loss": 2.6,
      "theoretical_loss": 3.3645116232667815,
      "tokens_seen": 2537684992
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.65145640249669e-05,
      "loss": 2.4177,
      "theoretical_loss": 3.364504713311824,
      "tokens_seen": 2537750528
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.649801399659543e-05,
      "loss": 2.4378,
      "theoretical_loss": 3.3644978035852726,
      "tokens_seen": 2537816064
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.648146396822395e-05,
      "loss": 2.7031,
      "theoretical_loss": 3.364490894087115,
      "tokens_seen": 2537881600
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.646491393985247e-05,
      "loss": 2.5136,
      "theoretical_loss": 3.3644839848173373,
      "tokens_seen": 2537947136
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.6448363911481e-05,
      "loss": 2.5307,
      "theoretical_loss": 3.364477075775926,
      "tokens_seen": 2538012672
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.643181388310952e-05,
      "loss": 2.5771,
      "theoretical_loss": 3.364470166962868,
      "tokens_seen": 2538078208
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.641526385473803e-05,
      "loss": 2.6168,
      "theoretical_loss": 3.364463258378149,
      "tokens_seen": 2538143744
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.639871382636656e-05,
      "loss": 2.7251,
      "theoretical_loss": 3.364456350021756,
      "tokens_seen": 2538209280
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.638216379799508e-05,
      "loss": 2.4174,
      "theoretical_loss": 3.3644494418936763,
      "tokens_seen": 2538274816
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.63656137696236e-05,
      "loss": 2.3591,
      "theoretical_loss": 3.3644425339938957,
      "tokens_seen": 2538340352
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.634906374125213e-05,
      "loss": 2.6156,
      "theoretical_loss": 3.364435626322401,
      "tokens_seen": 2538405888
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.633251371288065e-05,
      "loss": 2.5237,
      "theoretical_loss": 3.3644287188791786,
      "tokens_seen": 2538471424
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.631596368450918e-05,
      "loss": 2.3647,
      "theoretical_loss": 3.364421811664215,
      "tokens_seen": 2538536960
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.629941365613769e-05,
      "loss": 2.2155,
      "theoretical_loss": 3.3644149046774965,
      "tokens_seen": 2538602496
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.628286362776621e-05,
      "loss": 2.37,
      "theoretical_loss": 3.3644079979190105,
      "tokens_seen": 2538668032
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.626631359939474e-05,
      "loss": 2.4627,
      "theoretical_loss": 3.364401091388743,
      "tokens_seen": 2538733568
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.624976357102326e-05,
      "loss": 2.4992,
      "theoretical_loss": 3.364394185086681,
      "tokens_seen": 2538799104
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.623321354265178e-05,
      "loss": 2.6833,
      "theoretical_loss": 3.3643872790128104,
      "tokens_seen": 2538864640
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.621666351428031e-05,
      "loss": 2.3927,
      "theoretical_loss": 3.3643803731671182,
      "tokens_seen": 2538930176
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.620011348590883e-05,
      "loss": 2.3223,
      "theoretical_loss": 3.364373467549591,
      "tokens_seen": 2538995712
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.618356345753736e-05,
      "loss": 2.4782,
      "theoretical_loss": 3.3643665621602157,
      "tokens_seen": 2539061248
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.616701342916588e-05,
      "loss": 2.7526,
      "theoretical_loss": 3.3643596569989778,
      "tokens_seen": 2539126784
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.61504634007944e-05,
      "loss": 2.3013,
      "theoretical_loss": 3.3643527520658645,
      "tokens_seen": 2539192320
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.613391337242293e-05,
      "loss": 2.4719,
      "theoretical_loss": 3.3643458473608625,
      "tokens_seen": 2539257856
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.611736334405145e-05,
      "loss": 2.7189,
      "theoretical_loss": 3.3643389428839585,
      "tokens_seen": 2539323392
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.610081331567998e-05,
      "loss": 2.4696,
      "theoretical_loss": 3.3643320386351387,
      "tokens_seen": 2539388928
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.60842632873085e-05,
      "loss": 2.4319,
      "theoretical_loss": 3.36432513461439,
      "tokens_seen": 2539454464
    },
    {
      "epoch": 0.91,
      "objective/train/avg_token_score": 0.012374049052596092,
      "objective/train/avg_weight": 0.9854379296302795,
      "objective/train/docs_used": 1428565,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5544328689575195,
      "objective/train/std_weight": 0.06355191767215729,
      "objective/train/theoretical_loss": 3.3643182308216986,
      "objective/train/tokens_used": 2559980000,
      "theoretical_loss": 3.3643182308216986,
      "tokens_seen": 2539520000
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.606771325893702e-05,
      "loss": 2.4198,
      "theoretical_loss": 3.3643182308216986,
      "tokens_seen": 2539520000
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.605116323056553e-05,
      "loss": 2.5151,
      "theoretical_loss": 3.3643113272570515,
      "tokens_seen": 2539585536
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.603461320219406e-05,
      "loss": 3.5819,
      "theoretical_loss": 3.3643044239204345,
      "tokens_seen": 2539651072
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.601806317382258e-05,
      "loss": 2.6618,
      "theoretical_loss": 3.3642975208118355,
      "tokens_seen": 2539716608
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.60015131454511e-05,
      "loss": 2.4067,
      "theoretical_loss": 3.36429061793124,
      "tokens_seen": 2539782144
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.598496311707963e-05,
      "loss": 2.4808,
      "theoretical_loss": 3.364283715278635,
      "tokens_seen": 2539847680
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.596841308870815e-05,
      "loss": 2.2619,
      "theoretical_loss": 3.364276812854007,
      "tokens_seen": 2539913216
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.595186306033668e-05,
      "loss": 2.2526,
      "theoretical_loss": 3.3642699106573426,
      "tokens_seen": 2539978752
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.59353130319652e-05,
      "loss": 2.5707,
      "theoretical_loss": 3.3642630086886287,
      "tokens_seen": 2540044288
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.591876300359371e-05,
      "loss": 2.6429,
      "theoretical_loss": 3.364256106947851,
      "tokens_seen": 2540109824
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.590221297522224e-05,
      "loss": 2.4808,
      "theoretical_loss": 3.364249205434997,
      "tokens_seen": 2540175360
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.588566294685076e-05,
      "loss": 2.382,
      "theoretical_loss": 3.364242304150053,
      "tokens_seen": 2540240896
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.586911291847928e-05,
      "loss": 2.2719,
      "theoretical_loss": 3.3642354030930055,
      "tokens_seen": 2540306432
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.585256289010781e-05,
      "loss": 2.6931,
      "theoretical_loss": 3.3642285022638414,
      "tokens_seen": 2540371968
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.583601286173633e-05,
      "loss": 2.4229,
      "theoretical_loss": 3.3642216016625466,
      "tokens_seen": 2540437504
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.581946283336486e-05,
      "loss": 2.4466,
      "theoretical_loss": 3.3642147012891086,
      "tokens_seen": 2540503040
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.580291280499338e-05,
      "loss": 2.8649,
      "theoretical_loss": 3.3642078011435133,
      "tokens_seen": 2540568576
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.57863627766219e-05,
      "loss": 1.7719,
      "theoretical_loss": 3.3642009012257477,
      "tokens_seen": 2540634112
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.576981274825043e-05,
      "loss": 2.4315,
      "theoretical_loss": 3.364194001535798,
      "tokens_seen": 2540699648
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.575326271987895e-05,
      "loss": 2.6297,
      "theoretical_loss": 3.3641871020736516,
      "tokens_seen": 2540765184
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.573671269150748e-05,
      "loss": 2.1541,
      "theoretical_loss": 3.364180202839294,
      "tokens_seen": 2540830720
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.5720162663136e-05,
      "loss": 2.3968,
      "theoretical_loss": 3.3641733038327124,
      "tokens_seen": 2540896256
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.570361263476452e-05,
      "loss": 2.5308,
      "theoretical_loss": 3.3641664050538935,
      "tokens_seen": 2540961792
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.568706260639305e-05,
      "loss": 2.6577,
      "theoretical_loss": 3.364159506502824,
      "tokens_seen": 2541027328
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.567051257802156e-05,
      "loss": 2.3926,
      "theoretical_loss": 3.36415260817949,
      "tokens_seen": 2541092864
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.565396254965008e-05,
      "loss": 2.5353,
      "theoretical_loss": 3.3641457100838785,
      "tokens_seen": 2541158400
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.56374125212786e-05,
      "loss": 2.0131,
      "theoretical_loss": 3.3641388122159763,
      "tokens_seen": 2541223936
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.562086249290713e-05,
      "loss": 2.5806,
      "theoretical_loss": 3.3641319145757693,
      "tokens_seen": 2541289472
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.560431246453565e-05,
      "loss": 2.7453,
      "theoretical_loss": 3.3641250171632446,
      "tokens_seen": 2541355008
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.558776243616418e-05,
      "loss": 2.4056,
      "theoretical_loss": 3.364118119978389,
      "tokens_seen": 2541420544
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.55712124077927e-05,
      "loss": 2.3326,
      "theoretical_loss": 3.3641112230211885,
      "tokens_seen": 2541486080
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.555466237942121e-05,
      "loss": 2.1416,
      "theoretical_loss": 3.3641043262916304,
      "tokens_seen": 2541551616
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.553811235104974e-05,
      "loss": 2.1662,
      "theoretical_loss": 3.3640974297897004,
      "tokens_seen": 2541617152
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.552156232267826e-05,
      "loss": 2.5177,
      "theoretical_loss": 3.364090533515386,
      "tokens_seen": 2541682688
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.550501229430678e-05,
      "loss": 2.5645,
      "theoretical_loss": 3.364083637468674,
      "tokens_seen": 2541748224
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.548846226593531e-05,
      "loss": 2.1565,
      "theoretical_loss": 3.36407674164955,
      "tokens_seen": 2541813760
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.547191223756383e-05,
      "loss": 2.2863,
      "theoretical_loss": 3.364069846058001,
      "tokens_seen": 2541879296
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.545536220919236e-05,
      "loss": 2.3503,
      "theoretical_loss": 3.3640629506940147,
      "tokens_seen": 2541944832
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.543881218082088e-05,
      "loss": 2.7756,
      "theoretical_loss": 3.364056055557576,
      "tokens_seen": 2542010368
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.54222621524494e-05,
      "loss": 2.2212,
      "theoretical_loss": 3.3640491606486727,
      "tokens_seen": 2542075904
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.540571212407793e-05,
      "loss": 2.7729,
      "theoretical_loss": 3.364042265967291,
      "tokens_seen": 2542141440
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.538916209570645e-05,
      "loss": 2.3491,
      "theoretical_loss": 3.3640353715134172,
      "tokens_seen": 2542206976
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.537261206733498e-05,
      "loss": 2.5497,
      "theoretical_loss": 3.364028477287039,
      "tokens_seen": 2542272512
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.53560620389635e-05,
      "loss": 2.4431,
      "theoretical_loss": 3.364021583288142,
      "tokens_seen": 2542338048
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.533951201059202e-05,
      "loss": 1.9099,
      "theoretical_loss": 3.364014689516713,
      "tokens_seen": 2542403584
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.532296198222055e-05,
      "loss": 2.2941,
      "theoretical_loss": 3.3640077959727392,
      "tokens_seen": 2542469120
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.530641195384906e-05,
      "loss": 2.5793,
      "theoretical_loss": 3.3640009026562065,
      "tokens_seen": 2542534656
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.528986192547758e-05,
      "loss": 2.4936,
      "theoretical_loss": 3.363994009567102,
      "tokens_seen": 2542600192
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.527331189710611e-05,
      "loss": 2.6462,
      "theoretical_loss": 3.3639871167054123,
      "tokens_seen": 2542665728
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.525676186873463e-05,
      "loss": 2.3192,
      "theoretical_loss": 3.3639802240711236,
      "tokens_seen": 2542731264
    },
    {
      "epoch": 0.91,
      "objective/train/avg_token_score": 0.017111055552959442,
      "objective/train/avg_weight": 0.9798861742019653,
      "objective/train/docs_used": 1430679,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.744759559631348,
      "objective/train/std_weight": 0.08175299316644669,
      "objective/train/theoretical_loss": 3.3639733316642233,
      "objective/train/tokens_used": 2563256800,
      "theoretical_loss": 3.3639733316642233,
      "tokens_seen": 2542796800
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.524021184036316e-05,
      "loss": 2.7001,
      "theoretical_loss": 3.3639733316642233,
      "tokens_seen": 2542796800
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.522366181199168e-05,
      "loss": 2.5243,
      "theoretical_loss": 3.3639664394846975,
      "tokens_seen": 2542862336
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.52071117836202e-05,
      "loss": 4.3465,
      "theoretical_loss": 3.363959547532533,
      "tokens_seen": 2542927872
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.519056175524871e-05,
      "loss": 2.5696,
      "theoretical_loss": 3.363952655807716,
      "tokens_seen": 2542993408
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.517401172687724e-05,
      "loss": 2.5388,
      "theoretical_loss": 3.363945764310234,
      "tokens_seen": 2543058944
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.515746169850576e-05,
      "loss": 2.3816,
      "theoretical_loss": 3.363938873040073,
      "tokens_seen": 2543124480
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.514091167013429e-05,
      "loss": 2.3676,
      "theoretical_loss": 3.36393198199722,
      "tokens_seen": 2543190016
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.512436164176281e-05,
      "loss": 2.4957,
      "theoretical_loss": 3.3639250911816614,
      "tokens_seen": 2543255552
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.510781161339133e-05,
      "loss": 2.5784,
      "theoretical_loss": 3.3639182005933836,
      "tokens_seen": 2543321088
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.509126158501986e-05,
      "loss": 2.3632,
      "theoretical_loss": 3.3639113102323743,
      "tokens_seen": 2543386624
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.507471155664838e-05,
      "loss": 2.5697,
      "theoretical_loss": 3.363904420098619,
      "tokens_seen": 2543452160
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.50581615282769e-05,
      "loss": 2.3471,
      "theoretical_loss": 3.3638975301921046,
      "tokens_seen": 2543517696
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.504161149990543e-05,
      "loss": 2.6474,
      "theoretical_loss": 3.363890640512818,
      "tokens_seen": 2543583232
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.502506147153395e-05,
      "loss": 2.4065,
      "theoretical_loss": 3.3638837510607456,
      "tokens_seen": 2543648768
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.500851144316248e-05,
      "loss": 2.5625,
      "theoretical_loss": 3.3638768618358745,
      "tokens_seen": 2543714304
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.4991961414791e-05,
      "loss": 2.512,
      "theoretical_loss": 3.363869972838191,
      "tokens_seen": 2543779840
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.497541138641953e-05,
      "loss": 2.4497,
      "theoretical_loss": 3.363863084067682,
      "tokens_seen": 2543845376
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.495886135804805e-05,
      "loss": 2.6537,
      "theoretical_loss": 3.363856195524334,
      "tokens_seen": 2543910912
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.494231132967656e-05,
      "loss": 2.4456,
      "theoretical_loss": 3.3638493072081332,
      "tokens_seen": 2543976448
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.492576130130508e-05,
      "loss": 2.4158,
      "theoretical_loss": 3.363842419119067,
      "tokens_seen": 2544041984
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.490921127293361e-05,
      "loss": 2.5322,
      "theoretical_loss": 3.3638355312571218,
      "tokens_seen": 2544107520
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.489266124456213e-05,
      "loss": 1.9289,
      "theoretical_loss": 3.363828643622284,
      "tokens_seen": 2544173056
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.487611121619066e-05,
      "loss": 2.5469,
      "theoretical_loss": 3.363821756214541,
      "tokens_seen": 2544238592
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.485956118781918e-05,
      "loss": 2.4965,
      "theoretical_loss": 3.3638148690338783,
      "tokens_seen": 2544304128
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.48430111594477e-05,
      "loss": 2.4658,
      "theoretical_loss": 3.3638079820802838,
      "tokens_seen": 2544369664
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.482646113107623e-05,
      "loss": 2.4399,
      "theoretical_loss": 3.363801095353743,
      "tokens_seen": 2544435200
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.480991110270474e-05,
      "loss": 2.5358,
      "theoretical_loss": 3.3637942088542436,
      "tokens_seen": 2544500736
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.479336107433326e-05,
      "loss": 2.6308,
      "theoretical_loss": 3.3637873225817714,
      "tokens_seen": 2544566272
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.477681104596179e-05,
      "loss": 2.7171,
      "theoretical_loss": 3.363780436536314,
      "tokens_seen": 2544631808
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.476026101759031e-05,
      "loss": 2.6712,
      "theoretical_loss": 3.363773550717857,
      "tokens_seen": 2544697344
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.474371098921883e-05,
      "loss": 2.2432,
      "theoretical_loss": 3.3637666651263882,
      "tokens_seen": 2544762880
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.472716096084736e-05,
      "loss": 2.6884,
      "theoretical_loss": 3.363759779761893,
      "tokens_seen": 2544828416
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.471061093247588e-05,
      "loss": 2.4312,
      "theoretical_loss": 3.3637528946243593,
      "tokens_seen": 2544893952
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.46940609041044e-05,
      "loss": 2.2995,
      "theoretical_loss": 3.3637460097137732,
      "tokens_seen": 2544959488
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.467751087573293e-05,
      "loss": 2.489,
      "theoretical_loss": 3.363739125030121,
      "tokens_seen": 2545025024
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.466096084736145e-05,
      "loss": 2.6352,
      "theoretical_loss": 3.3637322405733903,
      "tokens_seen": 2545090560
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.464441081898998e-05,
      "loss": 2.4133,
      "theoretical_loss": 3.363725356343567,
      "tokens_seen": 2545156096
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.46278607906185e-05,
      "loss": 2.5239,
      "theoretical_loss": 3.363718472340638,
      "tokens_seen": 2545221632
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.461131076224703e-05,
      "loss": 2.4405,
      "theoretical_loss": 3.3637115885645903,
      "tokens_seen": 2545287168
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.459476073387555e-05,
      "loss": 2.2986,
      "theoretical_loss": 3.3637047050154103,
      "tokens_seen": 2545352704
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.457821070550406e-05,
      "loss": 2.3087,
      "theoretical_loss": 3.363697821693085,
      "tokens_seen": 2545418240
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.456166067713258e-05,
      "loss": 2.4752,
      "theoretical_loss": 3.3636909385976,
      "tokens_seen": 2545483776
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.454511064876111e-05,
      "loss": 2.342,
      "theoretical_loss": 3.3636840557289434,
      "tokens_seen": 2545549312
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.452856062038963e-05,
      "loss": 2.4929,
      "theoretical_loss": 3.3636771730871007,
      "tokens_seen": 2545614848
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.451201059201816e-05,
      "loss": 2.4219,
      "theoretical_loss": 3.3636702906720592,
      "tokens_seen": 2545680384
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.449546056364668e-05,
      "loss": 2.1786,
      "theoretical_loss": 3.3636634084838057,
      "tokens_seen": 2545745920
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.44789105352752e-05,
      "loss": 2.2631,
      "theoretical_loss": 3.363656526522327,
      "tokens_seen": 2545811456
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.446236050690373e-05,
      "loss": 2.4047,
      "theoretical_loss": 3.363649644787609,
      "tokens_seen": 2545876992
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.444581047853224e-05,
      "loss": 1.7975,
      "theoretical_loss": 3.363642763279639,
      "tokens_seen": 2545942528
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.442926045016076e-05,
      "loss": 2.5424,
      "theoretical_loss": 3.3636358819984036,
      "tokens_seen": 2546008064
    },
    {
      "epoch": 0.91,
      "objective/train/avg_token_score": 0.007931648753583431,
      "objective/train/avg_weight": 0.9906690716743469,
      "objective/train/docs_used": 1432418,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6412713527679443,
      "objective/train/std_weight": 0.023402303457260132,
      "objective/train/theoretical_loss": 3.36362900094389,
      "objective/train/tokens_used": 2566533600,
      "theoretical_loss": 3.36362900094389,
      "tokens_seen": 2546073600
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.441271042178929e-05,
      "loss": 2.7024,
      "theoretical_loss": 3.36362900094389,
      "tokens_seen": 2546073600
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.439616039341781e-05,
      "loss": 2.6286,
      "theoretical_loss": 3.3636221201160836,
      "tokens_seen": 2546139136
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.437961036504633e-05,
      "loss": 1.5678,
      "theoretical_loss": 3.3636152395149725,
      "tokens_seen": 2546204672
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.436306033667486e-05,
      "loss": 2.6755,
      "theoretical_loss": 3.3636083591405423,
      "tokens_seen": 2546270208
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.434651030830338e-05,
      "loss": 2.4945,
      "theoretical_loss": 3.36360147899278,
      "tokens_seen": 2546335744
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.432996027993191e-05,
      "loss": 2.8548,
      "theoretical_loss": 3.363594599071673,
      "tokens_seen": 2546401280
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.431341025156043e-05,
      "loss": 2.669,
      "theoretical_loss": 3.3635877193772075,
      "tokens_seen": 2546466816
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.429686022318895e-05,
      "loss": 2.3997,
      "theoretical_loss": 3.36358083990937,
      "tokens_seen": 2546532352
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.428031019481748e-05,
      "loss": 2.4754,
      "theoretical_loss": 3.363573960668147,
      "tokens_seen": 2546597888
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.4263760166446e-05,
      "loss": 2.5572,
      "theoretical_loss": 3.363567081653526,
      "tokens_seen": 2546663424
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.424721013807453e-05,
      "loss": 2.5914,
      "theoretical_loss": 3.363560202865493,
      "tokens_seen": 2546728960
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.423066010970305e-05,
      "loss": 2.5267,
      "theoretical_loss": 3.363553324304035,
      "tokens_seen": 2546794496
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.421411008133157e-05,
      "loss": 2.2729,
      "theoretical_loss": 3.363546445969139,
      "tokens_seen": 2546860032
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.419756005296009e-05,
      "loss": 2.6381,
      "theoretical_loss": 3.363539567860791,
      "tokens_seen": 2546925568
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.418101002458861e-05,
      "loss": 2.4299,
      "theoretical_loss": 3.363532689978978,
      "tokens_seen": 2546991104
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.416445999621713e-05,
      "loss": 2.757,
      "theoretical_loss": 3.3635258123236875,
      "tokens_seen": 2547056640
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.414790996784566e-05,
      "loss": 2.4759,
      "theoretical_loss": 3.363518934894905,
      "tokens_seen": 2547122176
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.413135993947418e-05,
      "loss": 2.809,
      "theoretical_loss": 3.363512057692618,
      "tokens_seen": 2547187712
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.41148099111027e-05,
      "loss": 2.5885,
      "theoretical_loss": 3.3635051807168126,
      "tokens_seen": 2547253248
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.409825988273123e-05,
      "loss": 2.3736,
      "theoretical_loss": 3.363498303967476,
      "tokens_seen": 2547318784
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.408170985435974e-05,
      "loss": 2.4977,
      "theoretical_loss": 3.363491427444595,
      "tokens_seen": 2547384320
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.406515982598826e-05,
      "loss": 2.4465,
      "theoretical_loss": 3.363484551148156,
      "tokens_seen": 2547449856
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.404860979761679e-05,
      "loss": 2.6492,
      "theoretical_loss": 3.363477675078146,
      "tokens_seen": 2547515392
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.403205976924531e-05,
      "loss": 2.3255,
      "theoretical_loss": 3.363470799234551,
      "tokens_seen": 2547580928
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.401550974087384e-05,
      "loss": 2.2369,
      "theoretical_loss": 3.3634639236173585,
      "tokens_seen": 2547646464
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.399895971250236e-05,
      "loss": 2.73,
      "theoretical_loss": 3.363457048226555,
      "tokens_seen": 2547712000
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.398240968413088e-05,
      "loss": 2.2836,
      "theoretical_loss": 3.3634501730621276,
      "tokens_seen": 2547777536
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.396585965575941e-05,
      "loss": 2.272,
      "theoretical_loss": 3.363443298124062,
      "tokens_seen": 2547843072
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.394930962738793e-05,
      "loss": 2.6376,
      "theoretical_loss": 3.3634364234123457,
      "tokens_seen": 2547908608
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.393275959901646e-05,
      "loss": 2.6237,
      "theoretical_loss": 3.363429548926965,
      "tokens_seen": 2547974144
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.391620957064498e-05,
      "loss": -0.4943,
      "theoretical_loss": 3.3634226746679077,
      "tokens_seen": 2548039680
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.38996595422735e-05,
      "loss": 2.4229,
      "theoretical_loss": 3.3634158006351593,
      "tokens_seen": 2548105216
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.388310951390203e-05,
      "loss": 2.5227,
      "theoretical_loss": 3.363408926828707,
      "tokens_seen": 2548170752
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.386655948553055e-05,
      "loss": 3.3157,
      "theoretical_loss": 3.3634020532485374,
      "tokens_seen": 2548236288
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.385000945715908e-05,
      "loss": 2.4769,
      "theoretical_loss": 3.3633951798946375,
      "tokens_seen": 2548301824
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.383345942878759e-05,
      "loss": 2.6902,
      "theoretical_loss": 3.3633883067669936,
      "tokens_seen": 2548367360
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.381690940041611e-05,
      "loss": 2.9742,
      "theoretical_loss": 3.3633814338655927,
      "tokens_seen": 2548432896
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.380035937204463e-05,
      "loss": 2.3052,
      "theoretical_loss": 3.363374561190422,
      "tokens_seen": 2548498432
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.378380934367316e-05,
      "loss": 2.526,
      "theoretical_loss": 3.363367688741467,
      "tokens_seen": 2548563968
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.376725931530168e-05,
      "loss": 2.4178,
      "theoretical_loss": 3.363360816518716,
      "tokens_seen": 2548629504
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.37507092869302e-05,
      "loss": 2.466,
      "theoretical_loss": 3.3633539445221543,
      "tokens_seen": 2548695040
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.373415925855873e-05,
      "loss": 2.5594,
      "theoretical_loss": 3.3633470727517696,
      "tokens_seen": 2548760576
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.371760923018724e-05,
      "loss": 2.4406,
      "theoretical_loss": 3.3633402012075484,
      "tokens_seen": 2548826112
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.370105920181576e-05,
      "loss": 2.5896,
      "theoretical_loss": 3.363333329889477,
      "tokens_seen": 2548891648
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.368450917344429e-05,
      "loss": 2.7019,
      "theoretical_loss": 3.363326458797543,
      "tokens_seen": 2548957184
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.366795914507281e-05,
      "loss": 2.3834,
      "theoretical_loss": 3.363319587931732,
      "tokens_seen": 2549022720
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.365140911670134e-05,
      "loss": 2.4079,
      "theoretical_loss": 3.3633127172920316,
      "tokens_seen": 2549088256
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.363485908832986e-05,
      "loss": 2.4372,
      "theoretical_loss": 3.3633058468784287,
      "tokens_seen": 2549153792
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.36183090599584e-05,
      "loss": 2.3682,
      "theoretical_loss": 3.3632989766909094,
      "tokens_seen": 2549219328
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.360175903158692e-05,
      "loss": 2.4153,
      "theoretical_loss": 3.3632921067294608,
      "tokens_seen": 2549284864
    },
    {
      "epoch": 0.91,
      "objective/train/avg_token_score": 0.046753622591495514,
      "objective/train/avg_weight": 0.9449846148490906,
      "objective/train/docs_used": 1434243,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.012073040008545,
      "objective/train/std_weight": 0.19086547195911407,
      "objective/train/theoretical_loss": 3.3632852369940696,
      "objective/train/tokens_used": 2569810400,
      "theoretical_loss": 3.3632852369940696,
      "tokens_seen": 2549350400
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.358520900321543e-05,
      "loss": 2.8368,
      "theoretical_loss": 3.3632852369940696,
      "tokens_seen": 2549350400
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.356865897484396e-05,
      "loss": 2.7154,
      "theoretical_loss": 3.3632783674847224,
      "tokens_seen": 2549415936
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.355210894647248e-05,
      "loss": 2.0067,
      "theoretical_loss": 3.363271498201406,
      "tokens_seen": 2549481472
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.3535558918101e-05,
      "loss": 2.4771,
      "theoretical_loss": 3.3632646291441075,
      "tokens_seen": 2549547008
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.351900888972953e-05,
      "loss": 2.3607,
      "theoretical_loss": 3.363257760312813,
      "tokens_seen": 2549612544
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.350245886135805e-05,
      "loss": 2.3361,
      "theoretical_loss": 3.36325089170751,
      "tokens_seen": 2549678080
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.348590883298658e-05,
      "loss": 2.2435,
      "theoretical_loss": 3.3632440233281846,
      "tokens_seen": 2549743616
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.346935880461509e-05,
      "loss": 2.7852,
      "theoretical_loss": 3.3632371551748244,
      "tokens_seen": 2549809152
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.345280877624361e-05,
      "loss": 2.6819,
      "theoretical_loss": 3.3632302872474154,
      "tokens_seen": 2549874688
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.343625874787213e-05,
      "loss": 2.069,
      "theoretical_loss": 3.3632234195459443,
      "tokens_seen": 2549940224
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.341970871950066e-05,
      "loss": 2.6244,
      "theoretical_loss": 3.3632165520703983,
      "tokens_seen": 2550005760
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.340315869112918e-05,
      "loss": 2.2372,
      "theoretical_loss": 3.363209684820764,
      "tokens_seen": 2550071296
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.33866086627577e-05,
      "loss": 2.3747,
      "theoretical_loss": 3.363202817797028,
      "tokens_seen": 2550136832
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.337005863438623e-05,
      "loss": 2.5417,
      "theoretical_loss": 3.3631959509991773,
      "tokens_seen": 2550202368
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.335350860601475e-05,
      "loss": 2.2549,
      "theoretical_loss": 3.3631890844271988,
      "tokens_seen": 2550267904
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.333695857764326e-05,
      "loss": 2.7079,
      "theoretical_loss": 3.363182218081079,
      "tokens_seen": 2550333440
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.332040854927179e-05,
      "loss": 8.0772,
      "theoretical_loss": 3.3631753519608045,
      "tokens_seen": 2550398976
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.330385852090033e-05,
      "loss": 2.4651,
      "theoretical_loss": 3.3631684860663627,
      "tokens_seen": 2550464512
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.328730849252885e-05,
      "loss": 2.45,
      "theoretical_loss": 3.3631616203977392,
      "tokens_seen": 2550530048
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.327075846415737e-05,
      "loss": 2.4105,
      "theoretical_loss": 3.3631547549549223,
      "tokens_seen": 2550595584
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.32542084357859e-05,
      "loss": 2.5401,
      "theoretical_loss": 3.3631478897378977,
      "tokens_seen": 2550661120
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.323765840741442e-05,
      "loss": 2.6536,
      "theoretical_loss": 3.3631410247466524,
      "tokens_seen": 2550726656
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.322110837904293e-05,
      "loss": 2.6951,
      "theoretical_loss": 3.3631341599811737,
      "tokens_seen": 2550792192
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.320455835067146e-05,
      "loss": 2.2385,
      "theoretical_loss": 3.3631272954414473,
      "tokens_seen": 2550857728
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.318800832229998e-05,
      "loss": 3.7018,
      "theoretical_loss": 3.3631204311274607,
      "tokens_seen": 2550923264
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.31714582939285e-05,
      "loss": 3.2839,
      "theoretical_loss": 3.3631135670392007,
      "tokens_seen": 2550988800
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.315490826555703e-05,
      "loss": 2.741,
      "theoretical_loss": 3.363106703176654,
      "tokens_seen": 2551054336
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.313835823718555e-05,
      "loss": 2.3522,
      "theoretical_loss": 3.3630998395398075,
      "tokens_seen": 2551119872
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.312180820881408e-05,
      "loss": 2.6566,
      "theoretical_loss": 3.3630929761286477,
      "tokens_seen": 2551185408
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.310525818044259e-05,
      "loss": 2.3666,
      "theoretical_loss": 3.3630861129431615,
      "tokens_seen": 2551250944
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.308870815207111e-05,
      "loss": 2.7053,
      "theoretical_loss": 3.3630792499833357,
      "tokens_seen": 2551316480
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.307215812369964e-05,
      "loss": 2.4724,
      "theoretical_loss": 3.363072387249157,
      "tokens_seen": 2551382016
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.305560809532816e-05,
      "loss": 2.6327,
      "theoretical_loss": 3.363065524740612,
      "tokens_seen": 2551447552
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.303905806695668e-05,
      "loss": 2.6598,
      "theoretical_loss": 3.3630586624576884,
      "tokens_seen": 2551513088
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.302250803858521e-05,
      "loss": 2.8384,
      "theoretical_loss": 3.3630518004003718,
      "tokens_seen": 2551578624
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.300595801021373e-05,
      "loss": 2.7025,
      "theoretical_loss": 3.3630449385686494,
      "tokens_seen": 2551644160
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.298940798184226e-05,
      "loss": 2.578,
      "theoretical_loss": 3.3630380769625083,
      "tokens_seen": 2551709696
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.297285795347078e-05,
      "loss": 1.8542,
      "theoretical_loss": 3.3630312155819353,
      "tokens_seen": 2551775232
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.29563079250993e-05,
      "loss": 2.4606,
      "theoretical_loss": 3.3630243544269165,
      "tokens_seen": 2551840768
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.293975789672783e-05,
      "loss": 2.5537,
      "theoretical_loss": 3.3630174934974395,
      "tokens_seen": 2551906304
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.292320786835635e-05,
      "loss": 2.14,
      "theoretical_loss": 3.363010632793491,
      "tokens_seen": 2551971840
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.290665783998488e-05,
      "loss": 2.3803,
      "theoretical_loss": 3.363003772315057,
      "tokens_seen": 2552037376
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.28901078116134e-05,
      "loss": 2.3817,
      "theoretical_loss": 3.3629969120621257,
      "tokens_seen": 2552102912
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.287355778324192e-05,
      "loss": 2.5748,
      "theoretical_loss": 3.3629900520346823,
      "tokens_seen": 2552168448
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.285700775487043e-05,
      "loss": 2.4594,
      "theoretical_loss": 3.362983192232715,
      "tokens_seen": 2552233984
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.284045772649896e-05,
      "loss": 2.63,
      "theoretical_loss": 3.3629763326562094,
      "tokens_seen": 2552299520
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.282390769812748e-05,
      "loss": 2.4107,
      "theoretical_loss": 3.362969473305153,
      "tokens_seen": 2552365056
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.2807357669756e-05,
      "loss": 2.3546,
      "theoretical_loss": 3.3629626141795326,
      "tokens_seen": 2552430592
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.279080764138453e-05,
      "loss": 2.9081,
      "theoretical_loss": 3.3629557552793345,
      "tokens_seen": 2552496128
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.277425761301305e-05,
      "loss": 2.6715,
      "theoretical_loss": 3.3629488966045464,
      "tokens_seen": 2552561664
    },
    {
      "epoch": 0.91,
      "objective/train/avg_token_score": 0.00827699713408947,
      "objective/train/avg_weight": 0.990268349647522,
      "objective/train/docs_used": 1436345,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5005578994750977,
      "objective/train/std_weight": 0.04153783619403839,
      "objective/train/theoretical_loss": 3.362942038155154,
      "objective/train/tokens_used": 2573087200,
      "theoretical_loss": 3.362942038155154,
      "tokens_seen": 2552627200
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.275770758464158e-05,
      "loss": 2.4331,
      "theoretical_loss": 3.362942038155154,
      "tokens_seen": 2552627200
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.27411575562701e-05,
      "loss": 2.7348,
      "theoretical_loss": 3.362935179931145,
      "tokens_seen": 2552692736
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.272460752789861e-05,
      "loss": 2.293,
      "theoretical_loss": 3.3629283219325057,
      "tokens_seen": 2552758272
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.270805749952714e-05,
      "loss": 1.6562,
      "theoretical_loss": 3.3629214641592236,
      "tokens_seen": 2552823808
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.269150747115566e-05,
      "loss": 2.4087,
      "theoretical_loss": 3.3629146066112847,
      "tokens_seen": 2552889344
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.267495744278418e-05,
      "loss": 2.4469,
      "theoretical_loss": 3.3629077492886763,
      "tokens_seen": 2552954880
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.265840741441271e-05,
      "loss": 2.4199,
      "theoretical_loss": 3.3629008921913845,
      "tokens_seen": 2553020416
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.264185738604123e-05,
      "loss": 2.0519,
      "theoretical_loss": 3.362894035319397,
      "tokens_seen": 2553085952
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.262530735766976e-05,
      "loss": 2.5576,
      "theoretical_loss": 3.3628871786727004,
      "tokens_seen": 2553151488
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.260875732929828e-05,
      "loss": 2.5512,
      "theoretical_loss": 3.362880322251281,
      "tokens_seen": 2553217024
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.25922073009268e-05,
      "loss": 2.3892,
      "theoretical_loss": 3.362873466055126,
      "tokens_seen": 2553282560
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.257565727255533e-05,
      "loss": 2.4713,
      "theoretical_loss": 3.3628666100842226,
      "tokens_seen": 2553348096
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.255910724418385e-05,
      "loss": 2.4851,
      "theoretical_loss": 3.3628597543385568,
      "tokens_seen": 2553413632
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.254255721581238e-05,
      "loss": 2.5169,
      "theoretical_loss": 3.362852898818116,
      "tokens_seen": 2553479168
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.25260071874409e-05,
      "loss": 2.246,
      "theoretical_loss": 3.362846043522887,
      "tokens_seen": 2553544704
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.250945715906942e-05,
      "loss": 2.4809,
      "theoretical_loss": 3.3628391884528566,
      "tokens_seen": 2553610240
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.249290713069795e-05,
      "loss": 2.555,
      "theoretical_loss": 3.362832333608011,
      "tokens_seen": 2553675776
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.247635710232646e-05,
      "loss": 2.2302,
      "theoretical_loss": 3.362825478988338,
      "tokens_seen": 2553741312
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.245980707395498e-05,
      "loss": 2.5605,
      "theoretical_loss": 3.3628186245938236,
      "tokens_seen": 2553806848
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.24432570455835e-05,
      "loss": 2.7309,
      "theoretical_loss": 3.362811770424455,
      "tokens_seen": 2553872384
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.242670701721203e-05,
      "loss": 2.2396,
      "theoretical_loss": 3.3628049164802194,
      "tokens_seen": 2553937920
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.241015698884055e-05,
      "loss": 2.5044,
      "theoretical_loss": 3.362798062761103,
      "tokens_seen": 2554003456
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.239360696046908e-05,
      "loss": 2.3629,
      "theoretical_loss": 3.362791209267093,
      "tokens_seen": 2554068992
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.23770569320976e-05,
      "loss": 2.5878,
      "theoretical_loss": 3.3627843559981763,
      "tokens_seen": 2554134528
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.236050690372611e-05,
      "loss": 2.3873,
      "theoretical_loss": 3.362777502954339,
      "tokens_seen": 2554200064
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.234395687535464e-05,
      "loss": 1.5763,
      "theoretical_loss": 3.3627706501355688,
      "tokens_seen": 2554265600
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.232740684698316e-05,
      "loss": 2.2704,
      "theoretical_loss": 3.362763797541852,
      "tokens_seen": 2554331136
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.231085681861168e-05,
      "loss": 1.1536,
      "theoretical_loss": 3.3627569451731754,
      "tokens_seen": 2554396672
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.229430679024021e-05,
      "loss": 2.5952,
      "theoretical_loss": 3.3627500930295264,
      "tokens_seen": 2554462208
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.227775676186873e-05,
      "loss": 2.4409,
      "theoretical_loss": 3.3627432411108917,
      "tokens_seen": 2554527744
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.226120673349726e-05,
      "loss": 2.4349,
      "theoretical_loss": 3.3627363894172575,
      "tokens_seen": 2554593280
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.224465670512578e-05,
      "loss": 2.3657,
      "theoretical_loss": 3.3627295379486117,
      "tokens_seen": 2554658816
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.22281066767543e-05,
      "loss": 2.5121,
      "theoretical_loss": 3.36272268670494,
      "tokens_seen": 2554724352
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.221155664838283e-05,
      "loss": 2.4521,
      "theoretical_loss": 3.36271583568623,
      "tokens_seen": 2554789888
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.219500662001135e-05,
      "loss": 2.5006,
      "theoretical_loss": 3.362708984892468,
      "tokens_seen": 2554855424
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.217845659163988e-05,
      "loss": 2.3767,
      "theoretical_loss": 3.3627021343236416,
      "tokens_seen": 2554920960
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.21619065632684e-05,
      "loss": 2.4229,
      "theoretical_loss": 3.362695283979737,
      "tokens_seen": 2554986496
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.214535653489692e-05,
      "loss": 2.3592,
      "theoretical_loss": 3.362688433860741,
      "tokens_seen": 2555052032
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.212880650652545e-05,
      "loss": 2.7055,
      "theoretical_loss": 3.362681583966641,
      "tokens_seen": 2555117568
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.211225647815396e-05,
      "loss": 1.924,
      "theoretical_loss": 3.3626747342974235,
      "tokens_seen": 2555183104
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.209570644978248e-05,
      "loss": 2.4455,
      "theoretical_loss": 3.3626678848530753,
      "tokens_seen": 2555248640
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.207915642141101e-05,
      "loss": 2.492,
      "theoretical_loss": 3.3626610356335833,
      "tokens_seen": 2555314176
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.206260639303953e-05,
      "loss": 2.4982,
      "theoretical_loss": 3.3626541866389346,
      "tokens_seen": 2555379712
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.204605636466806e-05,
      "loss": 2.4387,
      "theoretical_loss": 3.3626473378691157,
      "tokens_seen": 2555445248
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.202950633629658e-05,
      "loss": 2.6158,
      "theoretical_loss": 3.3626404893241135,
      "tokens_seen": 2555510784
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.20129563079251e-05,
      "loss": 2.6358,
      "theoretical_loss": 3.362633641003915,
      "tokens_seen": 2555576320
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.199640627955361e-05,
      "loss": 2.4192,
      "theoretical_loss": 3.3626267929085074,
      "tokens_seen": 2555641856
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.197985625118214e-05,
      "loss": 2.4553,
      "theoretical_loss": 3.362619945037877,
      "tokens_seen": 2555707392
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.196330622281066e-05,
      "loss": -0.2575,
      "theoretical_loss": 3.3626130973920105,
      "tokens_seen": 2555772928
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.194675619443919e-05,
      "loss": 2.3916,
      "theoretical_loss": 3.362606249970895,
      "tokens_seen": 2555838464
    },
    {
      "epoch": 0.91,
      "objective/train/avg_token_score": 0.012829352170228958,
      "objective/train/avg_weight": 0.984902024269104,
      "objective/train/docs_used": 1438290,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.4017696380615234,
      "objective/train/std_weight": 0.05978082865476608,
      "objective/train/theoretical_loss": 3.362599402774518,
      "objective/train/tokens_used": 2576364000,
      "theoretical_loss": 3.362599402774518,
      "tokens_seen": 2555904000
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.193020616606771e-05,
      "loss": 2.5084,
      "theoretical_loss": 3.362599402774518,
      "tokens_seen": 2555904000
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.191365613769623e-05,
      "loss": 2.5825,
      "theoretical_loss": 3.3625925558028653,
      "tokens_seen": 2555969536
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.189710610932476e-05,
      "loss": 2.4663,
      "theoretical_loss": 3.362585709055925,
      "tokens_seen": 2556035072
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.188055608095328e-05,
      "loss": 2.5536,
      "theoretical_loss": 3.3625788625336823,
      "tokens_seen": 2556100608
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.18640060525818e-05,
      "loss": 2.457,
      "theoretical_loss": 3.3625720162361254,
      "tokens_seen": 2556166144
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.184745602421033e-05,
      "loss": 2.5506,
      "theoretical_loss": 3.362565170163241,
      "tokens_seen": 2556231680
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.183090599583885e-05,
      "loss": 2.3628,
      "theoretical_loss": 3.3625583243150157,
      "tokens_seen": 2556297216
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.181435596746738e-05,
      "loss": 2.4683,
      "theoretical_loss": 3.362551478691436,
      "tokens_seen": 2556362752
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.17978059390959e-05,
      "loss": 2.4549,
      "theoretical_loss": 3.3625446332924893,
      "tokens_seen": 2556428288
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.178125591072443e-05,
      "loss": 2.5488,
      "theoretical_loss": 3.362537788118163,
      "tokens_seen": 2556493824
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.176470588235295e-05,
      "loss": 1.9282,
      "theoretical_loss": 3.3625309431684425,
      "tokens_seen": 2556559360
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.174815585398146e-05,
      "loss": 2.6137,
      "theoretical_loss": 3.362524098443316,
      "tokens_seen": 2556624896
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.173160582560998e-05,
      "loss": 2.3101,
      "theoretical_loss": 3.3625172539427695,
      "tokens_seen": 2556690432
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.171505579723851e-05,
      "loss": 2.5358,
      "theoretical_loss": 3.36251040966679,
      "tokens_seen": 2556755968
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.169850576886703e-05,
      "loss": 1.9723,
      "theoretical_loss": 3.362503565615365,
      "tokens_seen": 2556821504
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.168195574049556e-05,
      "loss": 2.8655,
      "theoretical_loss": 3.362496721788481,
      "tokens_seen": 2556887040
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.166540571212408e-05,
      "loss": 3.4962,
      "theoretical_loss": 3.3624898781861248,
      "tokens_seen": 2556952576
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.16488556837526e-05,
      "loss": 2.6483,
      "theoretical_loss": 3.3624830348082835,
      "tokens_seen": 2557018112
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.163230565538113e-05,
      "loss": 2.7001,
      "theoretical_loss": 3.3624761916549435,
      "tokens_seen": 2557083648
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.161575562700964e-05,
      "loss": 1.9391,
      "theoretical_loss": 3.3624693487260924,
      "tokens_seen": 2557149184
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.159920559863816e-05,
      "loss": 2.3045,
      "theoretical_loss": 3.3624625060217164,
      "tokens_seen": 2557214720
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.158265557026669e-05,
      "loss": 1.7498,
      "theoretical_loss": 3.3624556635418026,
      "tokens_seen": 2557280256
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.156610554189521e-05,
      "loss": 1.9147,
      "theoretical_loss": 3.362448821286338,
      "tokens_seen": 2557345792
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.154955551352373e-05,
      "loss": 2.2832,
      "theoretical_loss": 3.3624419792553097,
      "tokens_seen": 2557411328
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.153300548515226e-05,
      "loss": 2.5201,
      "theoretical_loss": 3.3624351374487045,
      "tokens_seen": 2557476864
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.151645545678078e-05,
      "loss": 2.4786,
      "theoretical_loss": 3.3624282958665086,
      "tokens_seen": 2557542400
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.14999054284093e-05,
      "loss": 1.938,
      "theoretical_loss": 3.3624214545087097,
      "tokens_seen": 2557607936
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.148335540003783e-05,
      "loss": 2.4846,
      "theoretical_loss": 3.3624146133752943,
      "tokens_seen": 2557673472
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.146680537166635e-05,
      "loss": 2.346,
      "theoretical_loss": 3.3624077724662493,
      "tokens_seen": 2557739008
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.145025534329488e-05,
      "loss": 1.2756,
      "theoretical_loss": 3.362400931781562,
      "tokens_seen": 2557804544
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.14337053149234e-05,
      "loss": 2.5971,
      "theoretical_loss": 3.3623940913212187,
      "tokens_seen": 2557870080
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.141715528655193e-05,
      "loss": 3.5151,
      "theoretical_loss": 3.362387251085207,
      "tokens_seen": 2557935616
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.140060525818045e-05,
      "loss": 2.6055,
      "theoretical_loss": 3.362380411073513,
      "tokens_seen": 2558001152
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.138405522980896e-05,
      "loss": 2.2273,
      "theoretical_loss": 3.3623735712861236,
      "tokens_seen": 2558066688
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.136750520143748e-05,
      "loss": 2.6197,
      "theoretical_loss": 3.362366731723027,
      "tokens_seen": 2558132224
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.135095517306601e-05,
      "loss": 2.5844,
      "theoretical_loss": 3.3623598923842084,
      "tokens_seen": 2558197760
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.133440514469453e-05,
      "loss": 2.612,
      "theoretical_loss": 3.3623530532696555,
      "tokens_seen": 2558263296
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.131785511632306e-05,
      "loss": 2.8492,
      "theoretical_loss": 3.362346214379355,
      "tokens_seen": 2558328832
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.130130508795158e-05,
      "loss": 2.4162,
      "theoretical_loss": 3.3623393757132947,
      "tokens_seen": 2558394368
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.12847550595801e-05,
      "loss": 2.6842,
      "theoretical_loss": 3.3623325372714605,
      "tokens_seen": 2558459904
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.126820503120863e-05,
      "loss": 2.6254,
      "theoretical_loss": 3.362325699053839,
      "tokens_seen": 2558525440
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.125165500283714e-05,
      "loss": 2.8964,
      "theoretical_loss": 3.3623188610604187,
      "tokens_seen": 2558590976
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.123510497446566e-05,
      "loss": 3.0363,
      "theoretical_loss": 3.3623120232911847,
      "tokens_seen": 2558656512
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.121855494609419e-05,
      "loss": 2.5762,
      "theoretical_loss": 3.3623051857461252,
      "tokens_seen": 2558722048
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.120200491772271e-05,
      "loss": 2.4174,
      "theoretical_loss": 3.362298348425226,
      "tokens_seen": 2558787584
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.118545488935123e-05,
      "loss": 2.9458,
      "theoretical_loss": 3.362291511328475,
      "tokens_seen": 2558853120
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.116890486097976e-05,
      "loss": 2.1073,
      "theoretical_loss": 3.3622846744558585,
      "tokens_seen": 2558918656
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.115235483260828e-05,
      "loss": 2.5956,
      "theoretical_loss": 3.362277837807364,
      "tokens_seen": 2558984192
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.113580480423681e-05,
      "loss": 2.5329,
      "theoretical_loss": 3.362271001382978,
      "tokens_seen": 2559049728
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.111925477586533e-05,
      "loss": 3.6327,
      "theoretical_loss": 3.3622641651826872,
      "tokens_seen": 2559115264
    },
    {
      "epoch": 0.91,
      "objective/train/avg_token_score": 0.010222855024039745,
      "objective/train/avg_weight": 0.9879720211029053,
      "objective/train/docs_used": 1439355,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6679317951202393,
      "objective/train/std_weight": 0.06185765564441681,
      "objective/train/theoretical_loss": 3.362257329206479,
      "objective/train/tokens_used": 2579640800,
      "theoretical_loss": 3.362257329206479,
      "tokens_seen": 2559180800
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.110270474749385e-05,
      "loss": 2.5447,
      "theoretical_loss": 3.362257329206479,
      "tokens_seen": 2559180800
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.108615471912238e-05,
      "loss": 2.4407,
      "theoretical_loss": 3.36225049345434,
      "tokens_seen": 2559246336
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.10696046907509e-05,
      "loss": 2.4672,
      "theoretical_loss": 3.362243657926257,
      "tokens_seen": 2559311872
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.105305466237943e-05,
      "loss": 2.5853,
      "theoretical_loss": 3.3622368226222177,
      "tokens_seen": 2559377408
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.103650463400795e-05,
      "loss": 2.6508,
      "theoretical_loss": 3.362229987542208,
      "tokens_seen": 2559442944
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.101995460563647e-05,
      "loss": 2.5733,
      "theoretical_loss": 3.3622231526862154,
      "tokens_seen": 2559508480
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.1003404577264985e-05,
      "loss": 2.201,
      "theoretical_loss": 3.3622163180542266,
      "tokens_seen": 2559574016
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.098685454889351e-05,
      "loss": 2.5202,
      "theoretical_loss": 3.362209483646229,
      "tokens_seen": 2559639552
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.097030452052203e-05,
      "loss": 2.666,
      "theoretical_loss": 3.3622026494622084,
      "tokens_seen": 2559705088
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.095375449215056e-05,
      "loss": 2.2421,
      "theoretical_loss": 3.362195815502153,
      "tokens_seen": 2559770624
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.093720446377908e-05,
      "loss": 2.5861,
      "theoretical_loss": 3.362188981766049,
      "tokens_seen": 2559836160
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.0920654435407605e-05,
      "loss": 2.4643,
      "theoretical_loss": 3.3621821482538836,
      "tokens_seen": 2559901696
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.090410440703613e-05,
      "loss": 2.6751,
      "theoretical_loss": 3.362175314965644,
      "tokens_seen": 2559967232
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.0887554378664646e-05,
      "loss": 2.3859,
      "theoretical_loss": 3.3621684819013167,
      "tokens_seen": 2560032768
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.087100435029317e-05,
      "loss": 2.6428,
      "theoretical_loss": 3.3621616490608885,
      "tokens_seen": 2560098304
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.0854454321921694e-05,
      "loss": 3.3298,
      "theoretical_loss": 3.362154816444346,
      "tokens_seen": 2560163840
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.083790429355022e-05,
      "loss": 2.7172,
      "theoretical_loss": 3.3621479840516777,
      "tokens_seen": 2560229376
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.082135426517874e-05,
      "loss": 2.2304,
      "theoretical_loss": 3.362141151882869,
      "tokens_seen": 2560294912
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.0804804236807266e-05,
      "loss": 2.8491,
      "theoretical_loss": 3.3621343199379075,
      "tokens_seen": 2560360448
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.078825420843579e-05,
      "loss": 1.9109,
      "theoretical_loss": 3.36212748821678,
      "tokens_seen": 2560425984
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.0771704180064314e-05,
      "loss": 2.5355,
      "theoretical_loss": 3.3621206567194735,
      "tokens_seen": 2560491520
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.0755154151692825e-05,
      "loss": 3.628,
      "theoretical_loss": 3.3621138254459746,
      "tokens_seen": 2560557056
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.073860412332135e-05,
      "loss": 2.39,
      "theoretical_loss": 3.3621069943962705,
      "tokens_seen": 2560622592
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.073860412332135e-05,
      "loss": 3.914,
      "theoretical_loss": 3.3621001635703482,
      "tokens_seen": 2560688128
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.072205409494987e-05,
      "loss": 3.5371,
      "theoretical_loss": 3.3620933329681946,
      "tokens_seen": 2560753664
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.07055040665784e-05,
      "loss": 2.6516,
      "theoretical_loss": 3.362086502589797,
      "tokens_seen": 2560819200
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.068895403820692e-05,
      "loss": 2.5113,
      "theoretical_loss": 3.3620796724351414,
      "tokens_seen": 2560884736
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.0672404009835445e-05,
      "loss": 2.612,
      "theoretical_loss": 3.3620728425042157,
      "tokens_seen": 2560950272
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.065585398146397e-05,
      "loss": 1.9178,
      "theoretical_loss": 3.3620660127970066,
      "tokens_seen": 2561015808
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.0639303953092486e-05,
      "loss": 2.5546,
      "theoretical_loss": 3.3620591833135003,
      "tokens_seen": 2561081344
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.062275392472101e-05,
      "loss": 2.8196,
      "theoretical_loss": 3.362052354053685,
      "tokens_seen": 2561146880
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.0606203896349534e-05,
      "loss": 2.6488,
      "theoretical_loss": 3.3620455250175465,
      "tokens_seen": 2561212416
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.058965386797806e-05,
      "loss": 2.5323,
      "theoretical_loss": 3.3620386962050723,
      "tokens_seen": 2561277952
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.057310383960658e-05,
      "loss": 2.5888,
      "theoretical_loss": 3.36203186761625,
      "tokens_seen": 2561343488
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.0556553811235106e-05,
      "loss": 2.3922,
      "theoretical_loss": 3.362025039251065,
      "tokens_seen": 2561409024
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.054000378286363e-05,
      "loss": 1.7683,
      "theoretical_loss": 3.362018211109506,
      "tokens_seen": 2561474560
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.052345375449215e-05,
      "loss": 2.5733,
      "theoretical_loss": 3.3620113831915583,
      "tokens_seen": 2561540096
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.050690372612067e-05,
      "loss": 2.5605,
      "theoretical_loss": 3.36200455549721,
      "tokens_seen": 2561605632
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.0490353697749195e-05,
      "loss": 2.471,
      "theoretical_loss": 3.3619977280264477,
      "tokens_seen": 2561671168
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.047380366937772e-05,
      "loss": 2.2974,
      "theoretical_loss": 3.361990900779258,
      "tokens_seen": 2561736704
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.045725364100624e-05,
      "loss": 2.542,
      "theoretical_loss": 3.361984073755629,
      "tokens_seen": 2561802240
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.044070361263477e-05,
      "loss": 2.4282,
      "theoretical_loss": 3.3619772469555462,
      "tokens_seen": 2561867776
    },
    {
      "epoch": 0.91,
      "learning_rate": 6.042415358426329e-05,
      "loss": 2.2394,
      "theoretical_loss": 3.3619704203789973,
      "tokens_seen": 2561933312
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.0407603555891815e-05,
      "loss": 2.5577,
      "theoretical_loss": 3.361963594025969,
      "tokens_seen": 2561998848
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.0391053527520325e-05,
      "loss": 1.6002,
      "theoretical_loss": 3.361956767896449,
      "tokens_seen": 2562064384
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.037450349914885e-05,
      "loss": 2.3395,
      "theoretical_loss": 3.3619499419904235,
      "tokens_seen": 2562129920
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.0357953470777373e-05,
      "loss": 2.4634,
      "theoretical_loss": 3.3619431163078795,
      "tokens_seen": 2562195456
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.03414034424059e-05,
      "loss": 2.5716,
      "theoretical_loss": 3.3619362908488046,
      "tokens_seen": 2562260992
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.032485341403442e-05,
      "loss": 2.6297,
      "theoretical_loss": 3.361929465613185,
      "tokens_seen": 2562326528
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.030830338566295e-05,
      "loss": 2.6007,
      "theoretical_loss": 3.3619226406010077,
      "tokens_seen": 2562392064
    },
    {
      "epoch": 0.92,
      "objective/train/avg_token_score": 0.010479336604475975,
      "objective/train/avg_weight": 0.9876638054847717,
      "objective/train/docs_used": 1441205,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.739360809326172,
      "objective/train/std_weight": 0.05215159058570862,
      "objective/train/theoretical_loss": 3.36191581581226,
      "objective/train/tokens_used": 2582917600,
      "theoretical_loss": 3.36191581581226,
      "tokens_seen": 2562457600
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.0291753357291476e-05,
      "loss": 2.6387,
      "theoretical_loss": 3.36191581581226,
      "tokens_seen": 2562457600
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.0275203328919987e-05,
      "loss": 2.1886,
      "theoretical_loss": 3.3619089912469295,
      "tokens_seen": 2562523136
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.025865330054851e-05,
      "loss": 2.6017,
      "theoretical_loss": 3.3619021669050024,
      "tokens_seen": 2562588672
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.0242103272177035e-05,
      "loss": 2.3805,
      "theoretical_loss": 3.3618953427864655,
      "tokens_seen": 2562654208
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.022555324380556e-05,
      "loss": 2.6834,
      "theoretical_loss": 3.361888518891306,
      "tokens_seen": 2562719744
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.020900321543408e-05,
      "loss": 2.2614,
      "theoretical_loss": 3.3618816952195107,
      "tokens_seen": 2562785280
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.0192453187062607e-05,
      "loss": 2.418,
      "theoretical_loss": 3.361874871771067,
      "tokens_seen": 2562850816
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.017590315869113e-05,
      "loss": 2.6073,
      "theoretical_loss": 3.361868048545962,
      "tokens_seen": 2562916352
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.0159353130319655e-05,
      "loss": 2.6321,
      "theoretical_loss": 3.3618612255441827,
      "tokens_seen": 2562981888
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.014280310194817e-05,
      "loss": 2.5184,
      "theoretical_loss": 3.361854402765715,
      "tokens_seen": 2563047424
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.0126253073576696e-05,
      "loss": 2.5087,
      "theoretical_loss": 3.3618475802105467,
      "tokens_seen": 2563112960
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.010970304520522e-05,
      "loss": 2.6288,
      "theoretical_loss": 3.361840757878665,
      "tokens_seen": 2563178496
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.0093153016833744e-05,
      "loss": 2.6017,
      "theoretical_loss": 3.361833935770057,
      "tokens_seen": 2563244032
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.007660298846227e-05,
      "loss": 2.6399,
      "theoretical_loss": 3.3618271138847082,
      "tokens_seen": 2563309568
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.006005296009079e-05,
      "loss": 2.7746,
      "theoretical_loss": 3.3618202922226073,
      "tokens_seen": 2563375104
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.0043502931719316e-05,
      "loss": 2.6701,
      "theoretical_loss": 3.3618134707837406,
      "tokens_seen": 2563440640
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.0026952903347826e-05,
      "loss": 2.3516,
      "theoretical_loss": 3.361806649568095,
      "tokens_seen": 2563506176
    },
    {
      "epoch": 0.92,
      "learning_rate": 6.001040287497635e-05,
      "loss": 2.466,
      "theoretical_loss": 3.361799828575658,
      "tokens_seen": 2563571712
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.999385284660488e-05,
      "loss": 2.4736,
      "theoretical_loss": 3.361793007806416,
      "tokens_seen": 2563637248
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9977302818233405e-05,
      "loss": 2.3944,
      "theoretical_loss": 3.361786187260356,
      "tokens_seen": 2563702784
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.996075278986193e-05,
      "loss": 2.3909,
      "theoretical_loss": 3.3617793669374656,
      "tokens_seen": 2563768320
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.994420276149045e-05,
      "loss": 2.3761,
      "theoretical_loss": 3.361772546837731,
      "tokens_seen": 2563833856
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.992765273311898e-05,
      "loss": 2.5221,
      "theoretical_loss": 3.36176572696114,
      "tokens_seen": 2563899392
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.99111027047475e-05,
      "loss": 2.6258,
      "theoretical_loss": 3.361758907307679,
      "tokens_seen": 2563964928
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.989455267637601e-05,
      "loss": 2.6177,
      "theoretical_loss": 3.361752087877335,
      "tokens_seen": 2564030464
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9878002648004535e-05,
      "loss": 2.5844,
      "theoretical_loss": 3.3617452686700955,
      "tokens_seen": 2564096000
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.986145261963306e-05,
      "loss": 2.4774,
      "theoretical_loss": 3.361738449685947,
      "tokens_seen": 2564161536
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.984490259126158e-05,
      "loss": 2.5189,
      "theoretical_loss": 3.3617316309248766,
      "tokens_seen": 2564227072
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.982835256289011e-05,
      "loss": 2.6302,
      "theoretical_loss": 3.3617248123868713,
      "tokens_seen": 2564292608
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.981180253451863e-05,
      "loss": 2.3417,
      "theoretical_loss": 3.3617179940719186,
      "tokens_seen": 2564358144
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9795252506147155e-05,
      "loss": 3.6844,
      "theoretical_loss": 3.361711175980005,
      "tokens_seen": 2564423680
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.977870247777567e-05,
      "loss": 2.6242,
      "theoretical_loss": 3.3617043581111172,
      "tokens_seen": 2564489216
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9762152449404196e-05,
      "loss": 2.5825,
      "theoretical_loss": 3.361697540465243,
      "tokens_seen": 2564554752
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.974560242103272e-05,
      "loss": 2.609,
      "theoretical_loss": 3.361690723042369,
      "tokens_seen": 2564620288
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9729052392661244e-05,
      "loss": 2.7193,
      "theoretical_loss": 3.361683905842482,
      "tokens_seen": 2564685824
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.971250236428977e-05,
      "loss": 2.7781,
      "theoretical_loss": 3.3616770888655694,
      "tokens_seen": 2564751360
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.969595233591829e-05,
      "loss": 2.3402,
      "theoretical_loss": 3.361670272111618,
      "tokens_seen": 2564816896
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9679402307546816e-05,
      "loss": 2.6521,
      "theoretical_loss": 3.3616634555806146,
      "tokens_seen": 2564882432
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9662852279175334e-05,
      "loss": 2.1597,
      "theoretical_loss": 3.3616566392725464,
      "tokens_seen": 2564947968
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.964630225080386e-05,
      "loss": 2.58,
      "theoretical_loss": 3.361649823187401,
      "tokens_seen": 2565013504
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.962975222243238e-05,
      "loss": 2.6297,
      "theoretical_loss": 3.3616430073251644,
      "tokens_seen": 2565079040
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9613202194060906e-05,
      "loss": 2.5607,
      "theoretical_loss": 3.3616361916858244,
      "tokens_seen": 2565144576
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.959665216568943e-05,
      "loss": 2.4185,
      "theoretical_loss": 3.3616293762693674,
      "tokens_seen": 2565210112
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9580102137317954e-05,
      "loss": 2.479,
      "theoretical_loss": 3.361622561075781,
      "tokens_seen": 2565275648
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.956355210894648e-05,
      "loss": 2.5911,
      "theoretical_loss": 3.361615746105052,
      "tokens_seen": 2565341184
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9547002080575e-05,
      "loss": 2.4733,
      "theoretical_loss": 3.361608931357167,
      "tokens_seen": 2565406720
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.953045205220351e-05,
      "loss": 2.667,
      "theoretical_loss": 3.3616021168321137,
      "tokens_seen": 2565472256
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9513902023832036e-05,
      "loss": 2.3891,
      "theoretical_loss": 3.361595302529879,
      "tokens_seen": 2565537792
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.949735199546056e-05,
      "loss": 2.606,
      "theoretical_loss": 3.361588488450449,
      "tokens_seen": 2565603328
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9480801967089084e-05,
      "loss": 2.1677,
      "theoretical_loss": 3.361581674593812,
      "tokens_seen": 2565668864
    },
    {
      "epoch": 0.92,
      "objective/train/avg_token_score": 0.02315736934542656,
      "objective/train/avg_weight": 0.972761869430542,
      "objective/train/docs_used": 1443158,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5192923545837402,
      "objective/train/std_weight": 0.09186911582946777,
      "objective/train/theoretical_loss": 3.3615748609599545,
      "objective/train/tokens_used": 2586194400,
      "theoretical_loss": 3.3615748609599545,
      "tokens_seen": 2565734400
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.946425193871761e-05,
      "loss": 2.5444,
      "theoretical_loss": 3.3615748609599545,
      "tokens_seen": 2565734400
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.944770191034613e-05,
      "loss": 2.5898,
      "theoretical_loss": 3.361568047548863,
      "tokens_seen": 2565799936
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9431151881974656e-05,
      "loss": 2.4899,
      "theoretical_loss": 3.3615612343605257,
      "tokens_seen": 2565865472
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.941460185360317e-05,
      "loss": 2.2043,
      "theoretical_loss": 3.3615544213949287,
      "tokens_seen": 2565931008
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.93980518252317e-05,
      "loss": 2.322,
      "theoretical_loss": 3.3615476086520593,
      "tokens_seen": 2565996544
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.938150179686022e-05,
      "loss": 2.6255,
      "theoretical_loss": 3.3615407961319046,
      "tokens_seen": 2566062080
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9364951768488745e-05,
      "loss": 2.4636,
      "theoretical_loss": 3.3615339838344513,
      "tokens_seen": 2566127616
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.934840174011727e-05,
      "loss": 2.6009,
      "theoretical_loss": 3.361527171759687,
      "tokens_seen": 2566193152
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.933185171174579e-05,
      "loss": 2.4752,
      "theoretical_loss": 3.3615203599075985,
      "tokens_seen": 2566258688
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.931530168337432e-05,
      "loss": 2.2303,
      "theoretical_loss": 3.361513548278172,
      "tokens_seen": 2566324224
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.929875165500284e-05,
      "loss": 2.5369,
      "theoretical_loss": 3.3615067368713962,
      "tokens_seen": 2566389760
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.928220162663136e-05,
      "loss": 2.6895,
      "theoretical_loss": 3.361499925687257,
      "tokens_seen": 2566455296
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.926565159825988e-05,
      "loss": 1.6657,
      "theoretical_loss": 3.361493114725741,
      "tokens_seen": 2566520832
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9249101569888406e-05,
      "loss": 2.6182,
      "theoretical_loss": 3.361486303986837,
      "tokens_seen": 2566586368
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.923255154151693e-05,
      "loss": 2.0651,
      "theoretical_loss": 3.36147949347053,
      "tokens_seen": 2566651904
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9216001513145454e-05,
      "loss": 2.4729,
      "theoretical_loss": 3.3614726831768085,
      "tokens_seen": 2566717440
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.919945148477398e-05,
      "loss": 2.4201,
      "theoretical_loss": 3.361465873105659,
      "tokens_seen": 2566782976
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.91829014564025e-05,
      "loss": 2.5595,
      "theoretical_loss": 3.361459063257068,
      "tokens_seen": 2566848512
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.916635142803101e-05,
      "loss": 1.8436,
      "theoretical_loss": 3.3614522536310236,
      "tokens_seen": 2566914048
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.914980139965954e-05,
      "loss": 2.5186,
      "theoretical_loss": 3.3614454442275123,
      "tokens_seen": 2566979584
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.913325137128806e-05,
      "loss": 2.2815,
      "theoretical_loss": 3.3614386350465213,
      "tokens_seen": 2567045120
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9116701342916585e-05,
      "loss": 2.473,
      "theoretical_loss": 3.3614318260880376,
      "tokens_seen": 2567110656
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.910015131454511e-05,
      "loss": 2.5466,
      "theoretical_loss": 3.361425017352048,
      "tokens_seen": 2567176192
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.908360128617363e-05,
      "loss": 2.5707,
      "theoretical_loss": 3.36141820883854,
      "tokens_seen": 2567241728
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.906705125780216e-05,
      "loss": 2.5341,
      "theoretical_loss": 3.3614114005475004,
      "tokens_seen": 2567307264
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.905050122943068e-05,
      "loss": 2.5459,
      "theoretical_loss": 3.361404592478916,
      "tokens_seen": 2567372800
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.90339512010592e-05,
      "loss": 2.4472,
      "theoretical_loss": 3.361397784632774,
      "tokens_seen": 2567438336
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.901740117268772e-05,
      "loss": 2.3214,
      "theoretical_loss": 3.361390977009062,
      "tokens_seen": 2567503872
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.9000851144316246e-05,
      "loss": 2.7802,
      "theoretical_loss": 3.361384169607766,
      "tokens_seen": 2567569408
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.898430111594477e-05,
      "loss": 2.4625,
      "theoretical_loss": 3.3613773624288745,
      "tokens_seen": 2567634944
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8967751087573294e-05,
      "loss": 2.4976,
      "theoretical_loss": 3.3613705554723734,
      "tokens_seen": 2567700480
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.895120105920182e-05,
      "loss": 2.3914,
      "theoretical_loss": 3.36136374873825,
      "tokens_seen": 2567766016
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.893465103083034e-05,
      "loss": 2.9682,
      "theoretical_loss": 3.3613569422264913,
      "tokens_seen": 2567831552
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.891810100245886e-05,
      "loss": 2.5851,
      "theoretical_loss": 3.361350135937085,
      "tokens_seen": 2567897088
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.890155097408738e-05,
      "loss": 2.2786,
      "theoretical_loss": 3.3613433298700173,
      "tokens_seen": 2567962624
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.888500094571591e-05,
      "loss": 2.5775,
      "theoretical_loss": 3.3613365240252757,
      "tokens_seen": 2568028160
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.886845091734443e-05,
      "loss": 2.8022,
      "theoretical_loss": 3.361329718402847,
      "tokens_seen": 2568093696
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8851900888972955e-05,
      "loss": 2.3665,
      "theoretical_loss": 3.361322913002719,
      "tokens_seen": 2568159232
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.883535086060148e-05,
      "loss": 1.6652,
      "theoretical_loss": 3.361316107824878,
      "tokens_seen": 2568224768
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.881880083223e-05,
      "loss": 2.6147,
      "theoretical_loss": 3.3613093028693113,
      "tokens_seen": 2568290304
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8802250803858513e-05,
      "loss": 2.6932,
      "theoretical_loss": 3.3613024981360056,
      "tokens_seen": 2568355840
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.878570077548704e-05,
      "loss": 2.3672,
      "theoretical_loss": 3.3612956936249487,
      "tokens_seen": 2568421376
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.876915074711556e-05,
      "loss": 2.2472,
      "theoretical_loss": 3.3612888893361275,
      "tokens_seen": 2568486912
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8752600718744085e-05,
      "loss": 2.3829,
      "theoretical_loss": 3.3612820852695284,
      "tokens_seen": 2568552448
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.873605069037261e-05,
      "loss": 2.4432,
      "theoretical_loss": 3.361275281425139,
      "tokens_seen": 2568617984
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.871950066200113e-05,
      "loss": 2.3356,
      "theoretical_loss": 3.361268477802947,
      "tokens_seen": 2568683520
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.870295063362966e-05,
      "loss": 2.5422,
      "theoretical_loss": 3.361261674402938,
      "tokens_seen": 2568749056
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.868640060525818e-05,
      "loss": 2.5553,
      "theoretical_loss": 3.3612548712251002,
      "tokens_seen": 2568814592
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.86698505768867e-05,
      "loss": 2.5867,
      "theoretical_loss": 3.3612480682694206,
      "tokens_seen": 2568880128
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.865330054851522e-05,
      "loss": 2.3193,
      "theoretical_loss": 3.3612412655358854,
      "tokens_seen": 2568945664
    },
    {
      "epoch": 0.92,
      "objective/train/avg_token_score": 0.05397920683026314,
      "objective/train/avg_weight": 0.9364895224571228,
      "objective/train/docs_used": 1445048,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.3496413230895996,
      "objective/train/std_weight": 0.22126464545726776,
      "objective/train/theoretical_loss": 3.361234463024483,
      "objective/train/tokens_used": 2589471200,
      "theoretical_loss": 3.361234463024483,
      "tokens_seen": 2569011200
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8636750520143747e-05,
      "loss": 2.5689,
      "theoretical_loss": 3.361234463024483,
      "tokens_seen": 2569011200
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.862020049177227e-05,
      "loss": 2.4463,
      "theoretical_loss": 3.3612276607351994,
      "tokens_seen": 2569076736
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8603650463400795e-05,
      "loss": 2.4305,
      "theoretical_loss": 3.3612208586680223,
      "tokens_seen": 2569142272
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.858710043502932e-05,
      "loss": 2.5725,
      "theoretical_loss": 3.3612140568229383,
      "tokens_seen": 2569207808
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.857055040665784e-05,
      "loss": 2.3576,
      "theoretical_loss": 3.3612072551999352,
      "tokens_seen": 2569273344
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.855400037828636e-05,
      "loss": 2.7505,
      "theoretical_loss": 3.361200453798999,
      "tokens_seen": 2569338880
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8537450349914884e-05,
      "loss": 2.7021,
      "theoretical_loss": 3.361193652620118,
      "tokens_seen": 2569404416
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.852090032154341e-05,
      "loss": 2.0401,
      "theoretical_loss": 3.361186851663278,
      "tokens_seen": 2569469952
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.850435029317193e-05,
      "loss": 2.9651,
      "theoretical_loss": 3.3611800509284677,
      "tokens_seen": 2569535488
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8487800264800456e-05,
      "loss": 2.7409,
      "theoretical_loss": 3.3611732504156726,
      "tokens_seen": 2569601024
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.847125023642898e-05,
      "loss": 2.3741,
      "theoretical_loss": 3.361166450124881,
      "tokens_seen": 2569666560
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8454700208057504e-05,
      "loss": 1.1495,
      "theoretical_loss": 3.361159650056079,
      "tokens_seen": 2569732096
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.843815017968603e-05,
      "loss": 2.4487,
      "theoretical_loss": 3.361152850209254,
      "tokens_seen": 2569797632
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.842160015131454e-05,
      "loss": 1.4736,
      "theoretical_loss": 3.3611460505843933,
      "tokens_seen": 2569863168
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.840505012294306e-05,
      "loss": 2.2161,
      "theoretical_loss": 3.361139251181484,
      "tokens_seen": 2569928704
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8388500094571586e-05,
      "loss": 2.3448,
      "theoretical_loss": 3.361132452000514,
      "tokens_seen": 2569994240
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.837195006620011e-05,
      "loss": 2.3043,
      "theoretical_loss": 3.3611256530414684,
      "tokens_seen": 2570059776
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8355400037828634e-05,
      "loss": 2.0845,
      "theoretical_loss": 3.361118854304336,
      "tokens_seen": 2570125312
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.833885000945716e-05,
      "loss": 2.5164,
      "theoretical_loss": 3.361112055789103,
      "tokens_seen": 2570190848
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.832229998108568e-05,
      "loss": 2.4743,
      "theoretical_loss": 3.361105257495757,
      "tokens_seen": 2570256384
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.83057499527142e-05,
      "loss": 2.681,
      "theoretical_loss": 3.361098459424285,
      "tokens_seen": 2570321920
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.828919992434272e-05,
      "loss": 2.5299,
      "theoretical_loss": 3.3610916615746738,
      "tokens_seen": 2570387456
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.827264989597125e-05,
      "loss": 2.5416,
      "theoretical_loss": 3.361084863946911,
      "tokens_seen": 2570452992
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.825609986759977e-05,
      "loss": 2.0774,
      "theoretical_loss": 3.3610780665409834,
      "tokens_seen": 2570518528
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8239549839228295e-05,
      "loss": 2.3472,
      "theoretical_loss": 3.361071269356878,
      "tokens_seen": 2570584064
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.822299981085682e-05,
      "loss": 2.1959,
      "theoretical_loss": 3.361064472394582,
      "tokens_seen": 2570649600
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.820644978248534e-05,
      "loss": 2.2217,
      "theoretical_loss": 3.3610576756540826,
      "tokens_seen": 2570715136
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.818989975411386e-05,
      "loss": 2.38,
      "theoretical_loss": 3.361050879135367,
      "tokens_seen": 2570780672
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8173349725742384e-05,
      "loss": 2.6494,
      "theoretical_loss": 3.361044082838422,
      "tokens_seen": 2570846208
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.815679969737091e-05,
      "loss": 2.4993,
      "theoretical_loss": 3.361037286763235,
      "tokens_seen": 2570911744
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.814024966899943e-05,
      "loss": 2.3167,
      "theoretical_loss": 3.361030490909793,
      "tokens_seen": 2570977280
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8123699640627956e-05,
      "loss": 2.6221,
      "theoretical_loss": 3.361023695278083,
      "tokens_seen": 2571042816
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.810714961225648e-05,
      "loss": 2.3891,
      "theoretical_loss": 3.3610168998680923,
      "tokens_seen": 2571108352
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.8090599583885004e-05,
      "loss": 2.9166,
      "theoretical_loss": 3.3610101046798078,
      "tokens_seen": 2571173888
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.807404955551353e-05,
      "loss": 2.3403,
      "theoretical_loss": 3.3610033097132166,
      "tokens_seen": 2571239424
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.805749952714204e-05,
      "loss": 2.4321,
      "theoretical_loss": 3.3609965149683063,
      "tokens_seen": 2571304960
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.804094949877056e-05,
      "loss": 2.4227,
      "theoretical_loss": 3.3609897204450636,
      "tokens_seen": 2571370496
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.802439947039909e-05,
      "loss": 2.4165,
      "theoretical_loss": 3.3609829261434756,
      "tokens_seen": 2571436032
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.800784944202761e-05,
      "loss": 2.1641,
      "theoretical_loss": 3.36097613206353,
      "tokens_seen": 2571501568
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.799129941365614e-05,
      "loss": 2.5011,
      "theoretical_loss": 3.360969338205213,
      "tokens_seen": 2571567104
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7974749385284666e-05,
      "loss": 2.4341,
      "theoretical_loss": 3.3609625445685123,
      "tokens_seen": 2571632640
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.795819935691319e-05,
      "loss": 2.3255,
      "theoretical_loss": 3.3609557511534147,
      "tokens_seen": 2571698176
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.79416493285417e-05,
      "loss": 2.3364,
      "theoretical_loss": 3.3609489579599074,
      "tokens_seen": 2571763712
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7925099300170224e-05,
      "loss": 2.5422,
      "theoretical_loss": 3.360942164987978,
      "tokens_seen": 2571829248
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.790854927179875e-05,
      "loss": 2.9497,
      "theoretical_loss": 3.360935372237613,
      "tokens_seen": 2571894784
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.789199924342727e-05,
      "loss": 2.7307,
      "theoretical_loss": 3.3609285797088004,
      "tokens_seen": 2571960320
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7875449215055796e-05,
      "loss": 3.0676,
      "theoretical_loss": 3.3609217874015265,
      "tokens_seen": 2572025856
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.785889918668432e-05,
      "loss": 2.3955,
      "theoretical_loss": 3.3609149953157784,
      "tokens_seen": 2572091392
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7842349158312844e-05,
      "loss": 2.2815,
      "theoretical_loss": 3.360908203451543,
      "tokens_seen": 2572156928
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.782579912994137e-05,
      "loss": 2.0208,
      "theoretical_loss": 3.3609014118088085,
      "tokens_seen": 2572222464
    },
    {
      "epoch": 0.92,
      "objective/train/avg_token_score": 0.010253781452775002,
      "objective/train/avg_weight": 0.9879311323165894,
      "objective/train/docs_used": 1447122,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.633941650390625,
      "objective/train/std_weight": 0.046553924679756165,
      "objective/train/theoretical_loss": 3.3608946203875614,
      "objective/train/tokens_used": 2592748000,
      "theoretical_loss": 3.3608946203875614,
      "tokens_seen": 2572288000
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7809249101569885e-05,
      "loss": 2.5003,
      "theoretical_loss": 3.3608946203875614,
      "tokens_seen": 2572288000
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.779269907319841e-05,
      "loss": 2.7305,
      "theoretical_loss": 3.360887829187789,
      "tokens_seen": 2572353536
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.777614904482693e-05,
      "loss": 2.5723,
      "theoretical_loss": 3.3608810382094783,
      "tokens_seen": 2572419072
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.775959901645546e-05,
      "loss": 1.9073,
      "theoretical_loss": 3.3608742474526165,
      "tokens_seen": 2572484608
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.774304898808398e-05,
      "loss": 2.5361,
      "theoretical_loss": 3.36086745691719,
      "tokens_seen": 2572550144
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7726498959712505e-05,
      "loss": 2.6242,
      "theoretical_loss": 3.3608606666031875,
      "tokens_seen": 2572615680
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.770994893134103e-05,
      "loss": 2.4941,
      "theoretical_loss": 3.360853876510595,
      "tokens_seen": 2572681216
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.769339890296954e-05,
      "loss": 2.5366,
      "theoretical_loss": 3.3608470866394,
      "tokens_seen": 2572746752
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.767684887459807e-05,
      "loss": 2.3658,
      "theoretical_loss": 3.360840296989589,
      "tokens_seen": 2572812288
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7660298846226594e-05,
      "loss": 2.5167,
      "theoretical_loss": 3.3608335075611504,
      "tokens_seen": 2572877824
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.764374881785512e-05,
      "loss": 2.5495,
      "theoretical_loss": 3.3608267183540703,
      "tokens_seen": 2572943360
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.762719878948364e-05,
      "loss": 2.384,
      "theoretical_loss": 3.3608199293683363,
      "tokens_seen": 2573008896
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7610648761112166e-05,
      "loss": 2.2169,
      "theoretical_loss": 3.360813140603935,
      "tokens_seen": 2573074432
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.759409873274069e-05,
      "loss": 2.4782,
      "theoretical_loss": 3.3608063520608544,
      "tokens_seen": 2573139968
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7577548704369214e-05,
      "loss": 2.2422,
      "theoretical_loss": 3.3607995637390813,
      "tokens_seen": 2573205504
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7560998675997725e-05,
      "loss": 2.3976,
      "theoretical_loss": 3.3607927756386027,
      "tokens_seen": 2573271040
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.754444864762625e-05,
      "loss": 2.5024,
      "theoretical_loss": 3.3607859877594057,
      "tokens_seen": 2573336576
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.752789861925477e-05,
      "loss": 2.5051,
      "theoretical_loss": 3.3607792001014776,
      "tokens_seen": 2573402112
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.75113485908833e-05,
      "loss": 2.2675,
      "theoretical_loss": 3.3607724126648058,
      "tokens_seen": 2573467648
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.749479856251182e-05,
      "loss": 2.7051,
      "theoretical_loss": 3.360765625449377,
      "tokens_seen": 2573533184
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7478248534140345e-05,
      "loss": 2.2443,
      "theoretical_loss": 3.3607588384551788,
      "tokens_seen": 2573598720
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.746169850576887e-05,
      "loss": 2.562,
      "theoretical_loss": 3.3607520516821974,
      "tokens_seen": 2573664256
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7445148477397386e-05,
      "loss": 2.6258,
      "theoretical_loss": 3.3607452651304213,
      "tokens_seen": 2573729792
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.742859844902591e-05,
      "loss": 2.4941,
      "theoretical_loss": 3.3607384787998367,
      "tokens_seen": 2573795328
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7412048420654434e-05,
      "loss": 2.426,
      "theoretical_loss": 3.360731692690431,
      "tokens_seen": 2573860864
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.739549839228296e-05,
      "loss": 1.972,
      "theoretical_loss": 3.360724906802192,
      "tokens_seen": 2573926400
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.737894836391148e-05,
      "loss": 2.7284,
      "theoretical_loss": 3.3607181211351054,
      "tokens_seen": 2573991936
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7362398335540006e-05,
      "loss": 2.5644,
      "theoretical_loss": 3.36071133568916,
      "tokens_seen": 2574057472
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.734584830716853e-05,
      "loss": 1.8477,
      "theoretical_loss": 3.360704550464342,
      "tokens_seen": 2574123008
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.732929827879705e-05,
      "loss": 2.4909,
      "theoretical_loss": 3.3606977654606385,
      "tokens_seen": 2574188544
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.731274825042557e-05,
      "loss": 2.3878,
      "theoretical_loss": 3.360690980678037,
      "tokens_seen": 2574254080
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7296198222054095e-05,
      "loss": 2.5776,
      "theoretical_loss": 3.360684196116525,
      "tokens_seen": 2574319616
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.727964819368262e-05,
      "loss": 2.5536,
      "theoretical_loss": 3.3606774117760887,
      "tokens_seen": 2574385152
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.726309816531114e-05,
      "loss": 2.3494,
      "theoretical_loss": 3.360670627656716,
      "tokens_seen": 2574450688
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.724654813693967e-05,
      "loss": 2.3522,
      "theoretical_loss": 3.3606638437583944,
      "tokens_seen": 2574516224
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.722999810856819e-05,
      "loss": 2.6374,
      "theoretical_loss": 3.36065706008111,
      "tokens_seen": 2574581760
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7213448080196715e-05,
      "loss": 2.3707,
      "theoretical_loss": 3.3606502766248507,
      "tokens_seen": 2574647296
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7196898051825225e-05,
      "loss": 2.5501,
      "theoretical_loss": 3.360643493389604,
      "tokens_seen": 2574712832
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.718034802345375e-05,
      "loss": 2.4213,
      "theoretical_loss": 3.3606367103753563,
      "tokens_seen": 2574778368
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.716379799508227e-05,
      "loss": 2.6353,
      "theoretical_loss": 3.360629927582095,
      "tokens_seen": 2574843904
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.71472479667108e-05,
      "loss": 1.719,
      "theoretical_loss": 3.360623145009807,
      "tokens_seen": 2574909440
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.713069793833932e-05,
      "loss": 2.3577,
      "theoretical_loss": 3.3606163626584804,
      "tokens_seen": 2574974976
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7114147909967845e-05,
      "loss": 2.2124,
      "theoretical_loss": 3.3606095805281018,
      "tokens_seen": 2575040512
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.709759788159637e-05,
      "loss": 2.608,
      "theoretical_loss": 3.3606027986186584,
      "tokens_seen": 2575106048
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7081047853224887e-05,
      "loss": 2.5373,
      "theoretical_loss": 3.360596016930137,
      "tokens_seen": 2575171584
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.706449782485341e-05,
      "loss": 2.372,
      "theoretical_loss": 3.3605892354625255,
      "tokens_seen": 2575237120
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.7047947796481935e-05,
      "loss": 2.4942,
      "theoretical_loss": 3.3605824542158107,
      "tokens_seen": 2575302656
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.703139776811046e-05,
      "loss": 2.2532,
      "theoretical_loss": 3.3605756731899796,
      "tokens_seen": 2575368192
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.701484773973898e-05,
      "loss": 2.4736,
      "theoretical_loss": 3.3605688923850194,
      "tokens_seen": 2575433728
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6998297711367507e-05,
      "loss": 2.8745,
      "theoretical_loss": 3.3605621118009177,
      "tokens_seen": 2575499264
    },
    {
      "epoch": 0.92,
      "objective/train/avg_token_score": 0.009553985670208931,
      "objective/train/avg_weight": 0.9887607097625732,
      "objective/train/docs_used": 1449033,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.488180637359619,
      "objective/train/std_weight": 0.038030873984098434,
      "objective/train/theoretical_loss": 3.360555331437662,
      "objective/train/tokens_used": 2596024800,
      "theoretical_loss": 3.360555331437662,
      "tokens_seen": 2575564800
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.698174768299603e-05,
      "loss": 3.3004,
      "theoretical_loss": 3.360555331437662,
      "tokens_seen": 2575564800
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6965197654624555e-05,
      "loss": 2.4918,
      "theoretical_loss": 3.3605485512952384,
      "tokens_seen": 2575630336
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.694864762625307e-05,
      "loss": 2.3063,
      "theoretical_loss": 3.3605417713736347,
      "tokens_seen": 2575695872
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6932097597881596e-05,
      "loss": 2.552,
      "theoretical_loss": 3.360534991672838,
      "tokens_seen": 2575761408
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.691554756951012e-05,
      "loss": 2.8764,
      "theoretical_loss": 3.3605282121928357,
      "tokens_seen": 2575826944
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6898997541138644e-05,
      "loss": 2.1492,
      "theoretical_loss": 3.3605214329336146,
      "tokens_seen": 2575892480
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.688244751276717e-05,
      "loss": 2.7323,
      "theoretical_loss": 3.3605146538951627,
      "tokens_seen": 2575958016
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.686589748439569e-05,
      "loss": 1.9167,
      "theoretical_loss": 3.360507875077466,
      "tokens_seen": 2576023552
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6849347456024216e-05,
      "loss": 2.4755,
      "theoretical_loss": 3.3605010964805127,
      "tokens_seen": 2576089088
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6832797427652726e-05,
      "loss": 2.5773,
      "theoretical_loss": 3.3604943181042897,
      "tokens_seen": 2576154624
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.681624739928125e-05,
      "loss": 2.4285,
      "theoretical_loss": 3.3604875399487835,
      "tokens_seen": 2576220160
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6799697370909774e-05,
      "loss": 2.0076,
      "theoretical_loss": 3.360480762013982,
      "tokens_seen": 2576285696
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.67831473425383e-05,
      "loss": 2.6431,
      "theoretical_loss": 3.3604739842998725,
      "tokens_seen": 2576351232
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.676659731416682e-05,
      "loss": 2.0755,
      "theoretical_loss": 3.360467206806442,
      "tokens_seen": 2576416768
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6750047285795346e-05,
      "loss": 2.6755,
      "theoretical_loss": 3.3604604295336777,
      "tokens_seen": 2576482304
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.673349725742387e-05,
      "loss": 2.3425,
      "theoretical_loss": 3.360453652481567,
      "tokens_seen": 2576547840
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6716947229052394e-05,
      "loss": 2.2564,
      "theoretical_loss": 3.360446875650097,
      "tokens_seen": 2576613376
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.670039720068091e-05,
      "loss": 2.5191,
      "theoretical_loss": 3.3604400990392547,
      "tokens_seen": 2576678912
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6683847172309435e-05,
      "loss": 2.4452,
      "theoretical_loss": 3.360433322649027,
      "tokens_seen": 2576744448
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.666729714393796e-05,
      "loss": 3.1488,
      "theoretical_loss": 3.3604265464794016,
      "tokens_seen": 2576809984
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.665074711556648e-05,
      "loss": 2.3349,
      "theoretical_loss": 3.360419770530366,
      "tokens_seen": 2576875520
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.663419708719501e-05,
      "loss": 2.8814,
      "theoretical_loss": 3.360412994801907,
      "tokens_seen": 2576941056
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.661764705882353e-05,
      "loss": 2.4328,
      "theoretical_loss": 3.3604062192940116,
      "tokens_seen": 2577006592
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6601097030452055e-05,
      "loss": 2.5928,
      "theoretical_loss": 3.360399444006667,
      "tokens_seen": 2577072128
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.658454700208057e-05,
      "loss": 2.761,
      "theoretical_loss": 3.3603926689398613,
      "tokens_seen": 2577137664
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6567996973709096e-05,
      "loss": 2.5183,
      "theoretical_loss": 3.360385894093581,
      "tokens_seen": 2577203200
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.655144694533762e-05,
      "loss": 2.5079,
      "theoretical_loss": 3.360379119467813,
      "tokens_seen": 2577268736
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6534896916966144e-05,
      "loss": 2.6797,
      "theoretical_loss": 3.3603723450625456,
      "tokens_seen": 2577334272
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.651834688859467e-05,
      "loss": 2.6804,
      "theoretical_loss": 3.3603655708777644,
      "tokens_seen": 2577399808
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.650179686022319e-05,
      "loss": 2.5168,
      "theoretical_loss": 3.3603587969134585,
      "tokens_seen": 2577465344
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6485246831851716e-05,
      "loss": 2.5606,
      "theoretical_loss": 3.3603520231696136,
      "tokens_seen": 2577530880
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.646869680348023e-05,
      "loss": 1.8488,
      "theoretical_loss": 3.3603452496462176,
      "tokens_seen": 2577596416
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.645214677510875e-05,
      "loss": 2.5025,
      "theoretical_loss": 3.3603384763432578,
      "tokens_seen": 2577661952
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6435596746737275e-05,
      "loss": 1.7899,
      "theoretical_loss": 3.360331703260721,
      "tokens_seen": 2577727488
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.64190467183658e-05,
      "loss": 2.5306,
      "theoretical_loss": 3.3603249303985945,
      "tokens_seen": 2577793024
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.640249668999432e-05,
      "loss": 2.7626,
      "theoretical_loss": 3.360318157756866,
      "tokens_seen": 2577858560
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.638594666162285e-05,
      "loss": 2.3092,
      "theoretical_loss": 3.3603113853355224,
      "tokens_seen": 2577924096
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.636939663325137e-05,
      "loss": 2.7385,
      "theoretical_loss": 3.360304613134551,
      "tokens_seen": 2577989632
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6352846604879895e-05,
      "loss": 2.3838,
      "theoretical_loss": 3.3602978411539386,
      "tokens_seen": 2578055168
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.633629657650841e-05,
      "loss": 2.8368,
      "theoretical_loss": 3.360291069393673,
      "tokens_seen": 2578120704
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6319746548136936e-05,
      "loss": 2.4139,
      "theoretical_loss": 3.3602842978537413,
      "tokens_seen": 2578186240
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.630319651976546e-05,
      "loss": 2.2215,
      "theoretical_loss": 3.360277526534131,
      "tokens_seen": 2578251776
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6286646491393984e-05,
      "loss": 2.6995,
      "theoretical_loss": 3.3602707554348283,
      "tokens_seen": 2578317312
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.627009646302251e-05,
      "loss": 2.4137,
      "theoretical_loss": 3.360263984555821,
      "tokens_seen": 2578382848
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.625354643465103e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.360257213897097,
      "tokens_seen": 2578448384
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6236996406279556e-05,
      "loss": 2.2306,
      "theoretical_loss": 3.360250443458643,
      "tokens_seen": 2578513920
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.622044637790807e-05,
      "loss": 2.9494,
      "theoretical_loss": 3.360243673240446,
      "tokens_seen": 2578579456
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.62038963495366e-05,
      "loss": 2.6708,
      "theoretical_loss": 3.3602369032424937,
      "tokens_seen": 2578644992
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.618734632116512e-05,
      "loss": 1.3418,
      "theoretical_loss": 3.3602301334647726,
      "tokens_seen": 2578710528
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6170796292793645e-05,
      "loss": 3.0123,
      "theoretical_loss": 3.360223363907271,
      "tokens_seen": 2578776064
    },
    {
      "epoch": 0.92,
      "objective/train/avg_token_score": 0.02867739088833332,
      "objective/train/avg_weight": 0.9662779569625854,
      "objective/train/docs_used": 1450816,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 12.32082462310791,
      "objective/train/std_weight": 0.13083048164844513,
      "objective/train/theoretical_loss": 3.360216594569975,
      "objective/train/tokens_used": 2599301600,
      "theoretical_loss": 3.360216594569975,
      "tokens_seen": 2578841600
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.615424626442217e-05,
      "loss": 2.1408,
      "theoretical_loss": 3.360216594569975,
      "tokens_seen": 2578841600
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.613769623605069e-05,
      "loss": 2.496,
      "theoretical_loss": 3.3602098254528725,
      "tokens_seen": 2578907136
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.612114620767922e-05,
      "loss": 2.5728,
      "theoretical_loss": 3.3602030565559513,
      "tokens_seen": 2578972672
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.610459617930774e-05,
      "loss": 2.4545,
      "theoretical_loss": 3.360196287879197,
      "tokens_seen": 2579038208
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.608804615093625e-05,
      "loss": 1.8359,
      "theoretical_loss": 3.360189519422599,
      "tokens_seen": 2579103744
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6071496122564776e-05,
      "loss": 2.2087,
      "theoretical_loss": 3.3601827511861426,
      "tokens_seen": 2579169280
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.60549460941933e-05,
      "loss": 2.2942,
      "theoretical_loss": 3.3601759831698157,
      "tokens_seen": 2579234816
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.6038396065821823e-05,
      "loss": 2.3422,
      "theoretical_loss": 3.3601692153736065,
      "tokens_seen": 2579300352
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.602184603745035e-05,
      "loss": 2.1916,
      "theoretical_loss": 3.3601624477975007,
      "tokens_seen": 2579365888
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.600529600907887e-05,
      "loss": 2.4626,
      "theoretical_loss": 3.3601556804414865,
      "tokens_seen": 2579431424
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.59887459807074e-05,
      "loss": 2.2746,
      "theoretical_loss": 3.360148913305551,
      "tokens_seen": 2579496960
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.597219595233591e-05,
      "loss": 2.625,
      "theoretical_loss": 3.3601421463896815,
      "tokens_seen": 2579562496
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.595564592396444e-05,
      "loss": 2.4946,
      "theoretical_loss": 3.3601353796938644,
      "tokens_seen": 2579628032
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.593909589559296e-05,
      "loss": 2.4843,
      "theoretical_loss": 3.3601286132180883,
      "tokens_seen": 2579693568
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5922545867221485e-05,
      "loss": 2.4265,
      "theoretical_loss": 3.36012184696234,
      "tokens_seen": 2579759104
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.590599583885001e-05,
      "loss": 2.0826,
      "theoretical_loss": 3.3601150809266063,
      "tokens_seen": 2579824640
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.588944581047853e-05,
      "loss": 2.8285,
      "theoretical_loss": 3.360108315110875,
      "tokens_seen": 2579890176
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.587289578210706e-05,
      "loss": 4.5872,
      "theoretical_loss": 3.360101549515133,
      "tokens_seen": 2579955712
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.585634575373558e-05,
      "loss": 1.723,
      "theoretical_loss": 3.360094784139368,
      "tokens_seen": 2580021248
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.58397957253641e-05,
      "loss": 2.4047,
      "theoretical_loss": 3.360088018983566,
      "tokens_seen": 2580086784
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.582324569699262e-05,
      "loss": 2.3951,
      "theoretical_loss": 3.3600812540477163,
      "tokens_seen": 2580152320
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5806695668621146e-05,
      "loss": 0.8959,
      "theoretical_loss": 3.3600744893318044,
      "tokens_seen": 2580217856
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.579014564024967e-05,
      "loss": 2.5449,
      "theoretical_loss": 3.3600677248358184,
      "tokens_seen": 2580283392
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5773595611878194e-05,
      "loss": 2.5787,
      "theoretical_loss": 3.3600609605597453,
      "tokens_seen": 2580348928
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.575704558350672e-05,
      "loss": 2.6621,
      "theoretical_loss": 3.3600541965035724,
      "tokens_seen": 2580414464
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.574049555513524e-05,
      "loss": 2.455,
      "theoretical_loss": 3.360047432667287,
      "tokens_seen": 2580480000
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.572394552676375e-05,
      "loss": 3.5599,
      "theoretical_loss": 3.3600406690508766,
      "tokens_seen": 2580545536
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5707395498392276e-05,
      "loss": 2.3294,
      "theoretical_loss": 3.360033905654328,
      "tokens_seen": 2580611072
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.56908454700208e-05,
      "loss": 2.2954,
      "theoretical_loss": 3.360027142477629,
      "tokens_seen": 2580676608
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.567429544164933e-05,
      "loss": 1.2347,
      "theoretical_loss": 3.3600203795207664,
      "tokens_seen": 2580742144
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5657745413277855e-05,
      "loss": 2.4681,
      "theoretical_loss": 3.360013616783728,
      "tokens_seen": 2580807680
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.564119538490638e-05,
      "loss": 2.6113,
      "theoretical_loss": 3.3600068542665005,
      "tokens_seen": 2580873216
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.56246453565349e-05,
      "loss": 0.9209,
      "theoretical_loss": 3.3600000919690713,
      "tokens_seen": 2580938752
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5608095328163413e-05,
      "loss": 2.5982,
      "theoretical_loss": 3.3599933298914277,
      "tokens_seen": 2581004288
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.559154529979194e-05,
      "loss": 1.9489,
      "theoretical_loss": 3.359986568033557,
      "tokens_seen": 2581069824
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.557499527142046e-05,
      "loss": 2.4071,
      "theoretical_loss": 3.3599798063954474,
      "tokens_seen": 2581135360
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5558445243048985e-05,
      "loss": 2.4854,
      "theoretical_loss": 3.3599730449770844,
      "tokens_seen": 2581200896
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.554189521467751e-05,
      "loss": 2.4553,
      "theoretical_loss": 3.3599662837784567,
      "tokens_seen": 2581266432
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.552534518630603e-05,
      "loss": 2.3041,
      "theoretical_loss": 3.359959522799551,
      "tokens_seen": 2581331968
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.550879515793456e-05,
      "loss": 2.2952,
      "theoretical_loss": 3.359952762040354,
      "tokens_seen": 2581397504
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.549224512956308e-05,
      "loss": 2.557,
      "theoretical_loss": 3.359946001500855,
      "tokens_seen": 2581463040
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.54756951011916e-05,
      "loss": 2.921,
      "theoretical_loss": 3.359939241181039,
      "tokens_seen": 2581528576
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.545914507282012e-05,
      "loss": 2.2274,
      "theoretical_loss": 3.3599324810808944,
      "tokens_seen": 2581594112
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5442595044448647e-05,
      "loss": 2.3357,
      "theoretical_loss": 3.3599257212004083,
      "tokens_seen": 2581659648
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.542604501607717e-05,
      "loss": 2.4846,
      "theoretical_loss": 3.3599189615395675,
      "tokens_seen": 2581725184
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5409494987705695e-05,
      "loss": 2.0544,
      "theoretical_loss": 3.3599122020983607,
      "tokens_seen": 2581790720
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.539294495933422e-05,
      "loss": 2.4382,
      "theoretical_loss": 3.359905442876774,
      "tokens_seen": 2581856256
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.537639493096274e-05,
      "loss": 2.6213,
      "theoretical_loss": 3.3598986838747944,
      "tokens_seen": 2581921792
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.535984490259126e-05,
      "loss": 6.0657,
      "theoretical_loss": 3.35989192509241,
      "tokens_seen": 2581987328
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5343294874219784e-05,
      "loss": 2.5014,
      "theoretical_loss": 3.359885166529608,
      "tokens_seen": 2582052864
    },
    {
      "epoch": 0.92,
      "objective/train/avg_token_score": 0.014414283446967602,
      "objective/train/avg_weight": 0.983031153678894,
      "objective/train/docs_used": 1453056,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 9.896429061889648,
      "objective/train/std_weight": 0.08856900781393051,
      "objective/train/theoretical_loss": 3.3598784081863755,
      "objective/train/tokens_used": 2602578400,
      "theoretical_loss": 3.3598784081863755,
      "tokens_seen": 2582118400
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.532674484584831e-05,
      "loss": 4.0912,
      "theoretical_loss": 3.3598784081863755,
      "tokens_seen": 2582118400
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.531019481747683e-05,
      "loss": 5.2834,
      "theoretical_loss": 3.3598716500626997,
      "tokens_seen": 2582183936
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5293644789105356e-05,
      "loss": 2.5208,
      "theoretical_loss": 3.359864892158568,
      "tokens_seen": 2582249472
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.527709476073388e-05,
      "loss": 2.3155,
      "theoretical_loss": 3.359858134473968,
      "tokens_seen": 2582315008
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5260544732362404e-05,
      "loss": 1.7161,
      "theoretical_loss": 3.3598513770088863,
      "tokens_seen": 2582380544
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.524399470399093e-05,
      "loss": 1.6418,
      "theoretical_loss": 3.3598446197633107,
      "tokens_seen": 2582446080
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.522744467561944e-05,
      "loss": 2.5005,
      "theoretical_loss": 3.3598378627372285,
      "tokens_seen": 2582511616
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.521089464724796e-05,
      "loss": 2.7123,
      "theoretical_loss": 3.359831105930627,
      "tokens_seen": 2582577152
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5194344618876486e-05,
      "loss": 2.4693,
      "theoretical_loss": 3.3598243493434934,
      "tokens_seen": 2582642688
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.517779459050501e-05,
      "loss": 2.5657,
      "theoretical_loss": 3.359817592975815,
      "tokens_seen": 2582708224
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5161244562133534e-05,
      "loss": 2.3826,
      "theoretical_loss": 3.359810836827579,
      "tokens_seen": 2582773760
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.514469453376206e-05,
      "loss": 2.4829,
      "theoretical_loss": 3.359804080898773,
      "tokens_seen": 2582839296
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.512814450539058e-05,
      "loss": 2.3288,
      "theoretical_loss": 3.3597973251893842,
      "tokens_seen": 2582904832
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.51115944770191e-05,
      "loss": 2.1066,
      "theoretical_loss": 3.3597905696993995,
      "tokens_seen": 2582970368
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.509504444864762e-05,
      "loss": 2.4469,
      "theoretical_loss": 3.359783814428807,
      "tokens_seen": 2583035904
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.507849442027615e-05,
      "loss": 2.5834,
      "theoretical_loss": 3.359777059377593,
      "tokens_seen": 2583101440
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.506194439190467e-05,
      "loss": 3.2649,
      "theoretical_loss": 3.3597703045457457,
      "tokens_seen": 2583166976
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.5045394363533195e-05,
      "loss": 2.4495,
      "theoretical_loss": 3.359763549933252,
      "tokens_seen": 2583232512
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.502884433516172e-05,
      "loss": 2.6008,
      "theoretical_loss": 3.359756795540099,
      "tokens_seen": 2583298048
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.501229430679024e-05,
      "loss": 2.5987,
      "theoretical_loss": 3.359750041366275,
      "tokens_seen": 2583363584
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.499574427841877e-05,
      "loss": 2.7787,
      "theoretical_loss": 3.359743287411766,
      "tokens_seen": 2583429120
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4979194250047284e-05,
      "loss": 2.286,
      "theoretical_loss": 3.35973653367656,
      "tokens_seen": 2583494656
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.496264422167581e-05,
      "loss": 2.3633,
      "theoretical_loss": 3.3597297801606443,
      "tokens_seen": 2583560192
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.494609419330433e-05,
      "loss": 2.39,
      "theoretical_loss": 3.3597230268640064,
      "tokens_seen": 2583625728
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4929544164932856e-05,
      "loss": 2.5154,
      "theoretical_loss": 3.359716273786633,
      "tokens_seen": 2583691264
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.491299413656138e-05,
      "loss": 2.6483,
      "theoretical_loss": 3.3597095209285124,
      "tokens_seen": 2583756800
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4896444108189904e-05,
      "loss": 2.4871,
      "theoretical_loss": 3.3597027682896305,
      "tokens_seen": 2583822336
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.487989407981843e-05,
      "loss": 2.8047,
      "theoretical_loss": 3.359696015869976,
      "tokens_seen": 2583887872
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.486334405144694e-05,
      "loss": 1.5106,
      "theoretical_loss": 3.3596892636695355,
      "tokens_seen": 2583953408
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.484679402307546e-05,
      "loss": 2.9111,
      "theoretical_loss": 3.3596825116882965,
      "tokens_seen": 2584018944
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.483024399470399e-05,
      "loss": 1.8495,
      "theoretical_loss": 3.3596757599262466,
      "tokens_seen": 2584084480
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.481369396633251e-05,
      "loss": 2.6324,
      "theoretical_loss": 3.3596690083833725,
      "tokens_seen": 2584150016
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4797143937961035e-05,
      "loss": 1.6754,
      "theoretical_loss": 3.3596622570596617,
      "tokens_seen": 2584215552
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.478059390958956e-05,
      "loss": 2.4777,
      "theoretical_loss": 3.3596555059551023,
      "tokens_seen": 2584281088
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.476404388121808e-05,
      "loss": 2.2342,
      "theoretical_loss": 3.3596487550696805,
      "tokens_seen": 2584346624
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.47474938528466e-05,
      "loss": 2.6854,
      "theoretical_loss": 3.3596420044033843,
      "tokens_seen": 2584412160
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4730943824475124e-05,
      "loss": 1.606,
      "theoretical_loss": 3.359635253956201,
      "tokens_seen": 2584477696
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.471439379610365e-05,
      "loss": 2.6936,
      "theoretical_loss": 3.3596285037281177,
      "tokens_seen": 2584543232
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.469784376773217e-05,
      "loss": 2.3778,
      "theoretical_loss": 3.3596217537191215,
      "tokens_seen": 2584608768
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4681293739360696e-05,
      "loss": 2.269,
      "theoretical_loss": 3.3596150039292008,
      "tokens_seen": 2584674304
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.466474371098922e-05,
      "loss": 3.2197,
      "theoretical_loss": 3.359608254358342,
      "tokens_seen": 2584739840
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4648193682617744e-05,
      "loss": 3.8102,
      "theoretical_loss": 3.359601505006532,
      "tokens_seen": 2584805376
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.463164365424627e-05,
      "loss": 2.4465,
      "theoretical_loss": 3.3595947558737596,
      "tokens_seen": 2584870912
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4615093625874785e-05,
      "loss": 5.6952,
      "theoretical_loss": 3.359588006960011,
      "tokens_seen": 2584936448
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.459854359750331e-05,
      "loss": 2.3447,
      "theoretical_loss": 3.359581258265274,
      "tokens_seen": 2585001984
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.458199356913183e-05,
      "loss": 2.2694,
      "theoretical_loss": 3.3595745097895358,
      "tokens_seen": 2585067520
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.456544354076036e-05,
      "loss": 2.3844,
      "theoretical_loss": 3.3595677615327837,
      "tokens_seen": 2585133056
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.454889351238888e-05,
      "loss": 2.7222,
      "theoretical_loss": 3.359561013495005,
      "tokens_seen": 2585198592
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4532343484017405e-05,
      "loss": 2.484,
      "theoretical_loss": 3.359554265676187,
      "tokens_seen": 2585264128
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.451579345564593e-05,
      "loss": 2.6643,
      "theoretical_loss": 3.3595475180763175,
      "tokens_seen": 2585329664
    },
    {
      "epoch": 0.92,
      "objective/train/avg_token_score": 0.023864710703492165,
      "objective/train/avg_weight": 0.9719076156616211,
      "objective/train/docs_used": 1455160,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 31.62553596496582,
      "objective/train/std_weight": 0.13969741761684418,
      "objective/train/theoretical_loss": 3.3595407706953835,
      "objective/train/tokens_used": 2605855200,
      "theoretical_loss": 3.3595407706953835,
      "tokens_seen": 2585395200
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.449924342727444e-05,
      "loss": 0.6818,
      "theoretical_loss": 3.3595407706953835,
      "tokens_seen": 2585395200
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4482693398902964e-05,
      "loss": 2.1649,
      "theoretical_loss": 3.3595340235333726,
      "tokens_seen": 2585460736
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.446614337053149e-05,
      "loss": 2.2255,
      "theoretical_loss": 3.3595272765902715,
      "tokens_seen": 2585526272
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.444959334216001e-05,
      "loss": 2.4947,
      "theoretical_loss": 3.3595205298660678,
      "tokens_seen": 2585591808
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4433043313788535e-05,
      "loss": 2.2293,
      "theoretical_loss": 3.3595137833607494,
      "tokens_seen": 2585657344
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.441649328541706e-05,
      "loss": 2.5493,
      "theoretical_loss": 3.3595070370743034,
      "tokens_seen": 2585722880
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4399943257045583e-05,
      "loss": 2.5535,
      "theoretical_loss": 3.3595002910067167,
      "tokens_seen": 2585788416
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.438339322867411e-05,
      "loss": 2.5004,
      "theoretical_loss": 3.359493545157977,
      "tokens_seen": 2585853952
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4366843200302625e-05,
      "loss": 1.8542,
      "theoretical_loss": 3.3594867995280717,
      "tokens_seen": 2585919488
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.435029317193115e-05,
      "loss": 3.4552,
      "theoretical_loss": 3.3594800541169882,
      "tokens_seen": 2585985024
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.433374314355967e-05,
      "loss": 2.173,
      "theoretical_loss": 3.3594733089247137,
      "tokens_seen": 2586050560
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.43171931151882e-05,
      "loss": 2.5628,
      "theoretical_loss": 3.3594665639512358,
      "tokens_seen": 2586116096
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.430064308681672e-05,
      "loss": 2.7929,
      "theoretical_loss": 3.359459819196541,
      "tokens_seen": 2586181632
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4284093058445245e-05,
      "loss": 2.4817,
      "theoretical_loss": 3.359453074660618,
      "tokens_seen": 2586247168
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.426754303007377e-05,
      "loss": 2.409,
      "theoretical_loss": 3.359446330343453,
      "tokens_seen": 2586312704
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4250993001702286e-05,
      "loss": 2.1646,
      "theoretical_loss": 3.3594395862450344,
      "tokens_seen": 2586378240
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.423444297333081e-05,
      "loss": 2.6125,
      "theoretical_loss": 3.3594328423653486,
      "tokens_seen": 2586443776
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4217892944959334e-05,
      "loss": 2.2614,
      "theoretical_loss": 3.3594260987043834,
      "tokens_seen": 2586509312
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.420134291658786e-05,
      "loss": 1.8916,
      "theoretical_loss": 3.3594193552621263,
      "tokens_seen": 2586574848
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.418479288821638e-05,
      "loss": 2.3934,
      "theoretical_loss": 3.3594126120385646,
      "tokens_seen": 2586640384
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4168242859844906e-05,
      "loss": 2.4757,
      "theoretical_loss": 3.359405869033685,
      "tokens_seen": 2586705920
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.415169283147343e-05,
      "loss": 1.496,
      "theoretical_loss": 3.359399126247476,
      "tokens_seen": 2586771456
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4135142803101954e-05,
      "loss": 2.6453,
      "theoretical_loss": 3.359392383679924,
      "tokens_seen": 2586836992
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4118592774730464e-05,
      "loss": 2.2611,
      "theoretical_loss": 3.3593856413310172,
      "tokens_seen": 2586902528
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.410204274635899e-05,
      "loss": 2.3827,
      "theoretical_loss": 3.3593788992007423,
      "tokens_seen": 2586968064
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.408549271798751e-05,
      "loss": 2.3203,
      "theoretical_loss": 3.359372157289087,
      "tokens_seen": 2587033600
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4068942689616036e-05,
      "loss": 2.8057,
      "theoretical_loss": 3.3593654155960384,
      "tokens_seen": 2587099136
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.405239266124456e-05,
      "loss": 1.5632,
      "theoretical_loss": 3.359358674121584,
      "tokens_seen": 2587164672
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4035842632873084e-05,
      "loss": 2.0289,
      "theoretical_loss": 3.3593519328657115,
      "tokens_seen": 2587230208
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.401929260450161e-05,
      "loss": 1.4673,
      "theoretical_loss": 3.359345191828408,
      "tokens_seen": 2587295744
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.4002742576130125e-05,
      "loss": 2.3811,
      "theoretical_loss": 3.359338451009661,
      "tokens_seen": 2587361280
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.398619254775865e-05,
      "loss": 2.5343,
      "theoretical_loss": 3.3593317104094575,
      "tokens_seen": 2587426816
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.396964251938717e-05,
      "loss": 2.6997,
      "theoretical_loss": 3.359324970027785,
      "tokens_seen": 2587492352
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.39530924910157e-05,
      "loss": 2.5274,
      "theoretical_loss": 3.359318229864632,
      "tokens_seen": 2587557888
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.393654246264422e-05,
      "loss": 2.2832,
      "theoretical_loss": 3.3593114899199836,
      "tokens_seen": 2587623424
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3919992434272745e-05,
      "loss": 3.1642,
      "theoretical_loss": 3.3593047501938296,
      "tokens_seen": 2587688960
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.390344240590127e-05,
      "loss": 2.9952,
      "theoretical_loss": 3.359298010686156,
      "tokens_seen": 2587754496
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3886892377529787e-05,
      "loss": 2.4128,
      "theoretical_loss": 3.35929127139695,
      "tokens_seen": 2587820032
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.387034234915831e-05,
      "loss": 0.8895,
      "theoretical_loss": 3.3592845323261997,
      "tokens_seen": 2587885568
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3853792320786835e-05,
      "loss": 2.1617,
      "theoretical_loss": 3.3592777934738924,
      "tokens_seen": 2587951104
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.383724229241536e-05,
      "loss": 2.2895,
      "theoretical_loss": 3.359271054840015,
      "tokens_seen": 2588016640
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.382069226404388e-05,
      "loss": 1.1457,
      "theoretical_loss": 3.3592643164245555,
      "tokens_seen": 2588082176
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3804142235672407e-05,
      "loss": 2.4057,
      "theoretical_loss": 3.359257578227501,
      "tokens_seen": 2588147712
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.378759220730093e-05,
      "loss": 4.3459,
      "theoretical_loss": 3.3592508402488384,
      "tokens_seen": 2588213248
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3771042178929454e-05,
      "loss": 2.5271,
      "theoretical_loss": 3.3592441024885558,
      "tokens_seen": 2588278784
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3754492150557965e-05,
      "loss": 2.9217,
      "theoretical_loss": 3.3592373649466407,
      "tokens_seen": 2588344320
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.373794212218649e-05,
      "loss": 1.9629,
      "theoretical_loss": 3.35923062762308,
      "tokens_seen": 2588409856
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.372139209381501e-05,
      "loss": 2.4565,
      "theoretical_loss": 3.3592238905178613,
      "tokens_seen": 2588475392
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.370484206544354e-05,
      "loss": 2.3647,
      "theoretical_loss": 3.359217153630972,
      "tokens_seen": 2588540928
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.368829203707206e-05,
      "loss": 2.4952,
      "theoretical_loss": 3.3592104169623997,
      "tokens_seen": 2588606464
    },
    {
      "epoch": 0.92,
      "objective/train/avg_token_score": 0.0073686884716153145,
      "objective/train/avg_weight": 0.9913263320922852,
      "objective/train/docs_used": 1457167,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.54160213470459,
      "objective/train/std_weight": 0.04541981965303421,
      "objective/train/theoretical_loss": 3.359203680512131,
      "objective/train/tokens_used": 2609132000,
      "theoretical_loss": 3.359203680512131,
      "tokens_seen": 2588672000
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.367174200870059e-05,
      "loss": 2.5677,
      "theoretical_loss": 3.359203680512131,
      "tokens_seen": 2588672000
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3655191980329116e-05,
      "loss": 2.3321,
      "theoretical_loss": 3.3591969442801544,
      "tokens_seen": 2588737536
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3638641951957626e-05,
      "loss": 2.1771,
      "theoretical_loss": 3.3591902082664564,
      "tokens_seen": 2588803072
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.362209192358615e-05,
      "loss": 2.3968,
      "theoretical_loss": 3.359183472471025,
      "tokens_seen": 2588868608
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3605541895214674e-05,
      "loss": 2.5003,
      "theoretical_loss": 3.359176736893847,
      "tokens_seen": 2588934144
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.35889918668432e-05,
      "loss": 2.5855,
      "theoretical_loss": 3.3591700015349106,
      "tokens_seen": 2588999680
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.357244183847172e-05,
      "loss": 2.2973,
      "theoretical_loss": 3.3591632663942024,
      "tokens_seen": 2589065216
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3555891810100246e-05,
      "loss": 1.4589,
      "theoretical_loss": 3.3591565314717107,
      "tokens_seen": 2589130752
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.353934178172877e-05,
      "loss": 2.1861,
      "theoretical_loss": 3.3591497967674218,
      "tokens_seen": 2589196288
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3522791753357294e-05,
      "loss": 3.0894,
      "theoretical_loss": 3.3591430622813236,
      "tokens_seen": 2589261824
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.350624172498581e-05,
      "loss": 3.0969,
      "theoretical_loss": 3.3591363280134043,
      "tokens_seen": 2589327360
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3489691696614335e-05,
      "loss": 1.8931,
      "theoretical_loss": 3.35912959396365,
      "tokens_seen": 2589392896
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.347314166824286e-05,
      "loss": 10.4451,
      "theoretical_loss": 3.3591228601320493,
      "tokens_seen": 2589458432
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.345659163987138e-05,
      "loss": 2.2908,
      "theoretical_loss": 3.3591161265185887,
      "tokens_seen": 2589523968
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.344004161149991e-05,
      "loss": 2.7958,
      "theoretical_loss": 3.359109393123256,
      "tokens_seen": 2589589504
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.342349158312843e-05,
      "loss": 2.5238,
      "theoretical_loss": 3.3591026599460383,
      "tokens_seen": 2589655040
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3406941554756955e-05,
      "loss": 1.5144,
      "theoretical_loss": 3.3590959269869236,
      "tokens_seen": 2589720576
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3390391526385466e-05,
      "loss": 2.5462,
      "theoretical_loss": 3.359089194245899,
      "tokens_seen": 2589786112
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.3373841498013996e-05,
      "loss": 1.9601,
      "theoretical_loss": 3.3590824617229513,
      "tokens_seen": 2589851648
    },
    {
      "epoch": 0.92,
      "learning_rate": 5.335729146964252e-05,
      "loss": 2.0484,
      "theoretical_loss": 3.3590757294180693,
      "tokens_seen": 2589917184
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.3340741441271044e-05,
      "loss": 2.2881,
      "theoretical_loss": 3.3590689973312395,
      "tokens_seen": 2589982720
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.332419141289957e-05,
      "loss": 2.3721,
      "theoretical_loss": 3.359062265462449,
      "tokens_seen": 2590048256
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.330764138452809e-05,
      "loss": 2.0074,
      "theoretical_loss": 3.359055533811686,
      "tokens_seen": 2590113792
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.3291091356156616e-05,
      "loss": 2.5187,
      "theoretical_loss": 3.359048802378938,
      "tokens_seen": 2590179328
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.327454132778514e-05,
      "loss": 1.1292,
      "theoretical_loss": 3.3590420711641915,
      "tokens_seen": 2590244864
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.325799129941365e-05,
      "loss": 2.6741,
      "theoretical_loss": 3.3590353401674347,
      "tokens_seen": 2590310400
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.3241441271042175e-05,
      "loss": 2.5767,
      "theoretical_loss": 3.3590286093886546,
      "tokens_seen": 2590375936
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.32248912426707e-05,
      "loss": 2.7667,
      "theoretical_loss": 3.359021878827839,
      "tokens_seen": 2590441472
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.320834121429922e-05,
      "loss": 2.4972,
      "theoretical_loss": 3.359015148484975,
      "tokens_seen": 2590507008
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.319179118592775e-05,
      "loss": 2.5312,
      "theoretical_loss": 3.35900841836005,
      "tokens_seen": 2590572544
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.317524115755627e-05,
      "loss": 2.1036,
      "theoretical_loss": 3.359001688453052,
      "tokens_seen": 2590638080
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.3158691129184795e-05,
      "loss": 2.7151,
      "theoretical_loss": 3.358994958763968,
      "tokens_seen": 2590703616
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.314214110081331e-05,
      "loss": 2.1295,
      "theoretical_loss": 3.3589882292927853,
      "tokens_seen": 2590769152
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.3125591072441836e-05,
      "loss": 1.9066,
      "theoretical_loss": 3.3589815000394916,
      "tokens_seen": 2590834688
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.310904104407036e-05,
      "loss": 1.9681,
      "theoretical_loss": 3.358974771004074,
      "tokens_seen": 2590900224
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.3092491015698884e-05,
      "loss": 2.504,
      "theoretical_loss": 3.35896804218652,
      "tokens_seen": 2590965760
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.307594098732741e-05,
      "loss": 2.0096,
      "theoretical_loss": 3.3589613135868177,
      "tokens_seen": 2591031296
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.305939095895593e-05,
      "loss": 1.7385,
      "theoretical_loss": 3.3589545852049536,
      "tokens_seen": 2591096832
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.3042840930584456e-05,
      "loss": 2.3405,
      "theoretical_loss": 3.358947857040916,
      "tokens_seen": 2591162368
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.302629090221297e-05,
      "loss": 2.8119,
      "theoretical_loss": 3.3589411290946916,
      "tokens_seen": 2591227904
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.30097408738415e-05,
      "loss": 2.4126,
      "theoretical_loss": 3.3589344013662683,
      "tokens_seen": 2591293440
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.299319084547002e-05,
      "loss": 2.6278,
      "theoretical_loss": 3.358927673855633,
      "tokens_seen": 2591358976
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2976640817098545e-05,
      "loss": 2.4582,
      "theoretical_loss": 3.3589209465627743,
      "tokens_seen": 2591424512
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.296009078872707e-05,
      "loss": 1.9816,
      "theoretical_loss": 3.358914219487678,
      "tokens_seen": 2591490048
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.294354076035559e-05,
      "loss": 1.4988,
      "theoretical_loss": 3.3589074926303333,
      "tokens_seen": 2591555584
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.292699073198412e-05,
      "loss": 2.2969,
      "theoretical_loss": 3.358900765990726,
      "tokens_seen": 2591621120
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.291044070361264e-05,
      "loss": 1.2783,
      "theoretical_loss": 3.358894039568845,
      "tokens_seen": 2591686656
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.289389067524115e-05,
      "loss": 2.6596,
      "theoretical_loss": 3.3588873133646766,
      "tokens_seen": 2591752192
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2877340646869675e-05,
      "loss": 2.3628,
      "theoretical_loss": 3.358880587378209,
      "tokens_seen": 2591817728
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.28607906184982e-05,
      "loss": 2.4423,
      "theoretical_loss": 3.358873861609429,
      "tokens_seen": 2591883264
    },
    {
      "epoch": 0.93,
      "objective/train/avg_token_score": 0.011556748300790787,
      "objective/train/avg_weight": 0.9863989949226379,
      "objective/train/docs_used": 1459162,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5854573249816895,
      "objective/train/std_weight": 0.06398551911115646,
      "objective/train/theoretical_loss": 3.3588671360583247,
      "objective/train/tokens_used": 2612408800,
      "theoretical_loss": 3.3588671360583247,
      "tokens_seen": 2591948800
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2844240590126723e-05,
      "loss": 2.4517,
      "theoretical_loss": 3.3588671360583247,
      "tokens_seen": 2591948800
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.282769056175525e-05,
      "loss": 2.6867,
      "theoretical_loss": 3.358860410724883,
      "tokens_seen": 2592014336
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.281114053338377e-05,
      "loss": 2.5292,
      "theoretical_loss": 3.358853685609092,
      "tokens_seen": 2592079872
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2794590505012295e-05,
      "loss": 2.6069,
      "theoretical_loss": 3.358846960710938,
      "tokens_seen": 2592145408
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.277804047664081e-05,
      "loss": 2.5693,
      "theoretical_loss": 3.35884023603041,
      "tokens_seen": 2592210944
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.276149044826934e-05,
      "loss": 2.0643,
      "theoretical_loss": 3.3588335115674943,
      "tokens_seen": 2592276480
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.274494041989786e-05,
      "loss": 2.1644,
      "theoretical_loss": 3.358826787322179,
      "tokens_seen": 2592342016
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2728390391526385e-05,
      "loss": 2.2499,
      "theoretical_loss": 3.358820063294451,
      "tokens_seen": 2592407552
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.271184036315491e-05,
      "loss": 2.3717,
      "theoretical_loss": 3.3588133394842985,
      "tokens_seen": 2592473088
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.269529033478343e-05,
      "loss": 2.5818,
      "theoretical_loss": 3.358806615891708,
      "tokens_seen": 2592538624
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.267874030641196e-05,
      "loss": 3.8755,
      "theoretical_loss": 3.3587998925166676,
      "tokens_seen": 2592604160
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.266219027804048e-05,
      "loss": 2.4412,
      "theoretical_loss": 3.358793169359165,
      "tokens_seen": 2592669696
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2645640249669e-05,
      "loss": 2.4151,
      "theoretical_loss": 3.3587864464191868,
      "tokens_seen": 2592735232
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.262909022129752e-05,
      "loss": 2.3582,
      "theoretical_loss": 3.358779723696721,
      "tokens_seen": 2592800768
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2612540192926046e-05,
      "loss": 2.3638,
      "theoretical_loss": 3.3587730011917554,
      "tokens_seen": 2592866304
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.259599016455457e-05,
      "loss": 2.5998,
      "theoretical_loss": 3.3587662789042767,
      "tokens_seen": 2592931840
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2579440136183094e-05,
      "loss": 2.4228,
      "theoretical_loss": 3.358759556834273,
      "tokens_seen": 2592997376
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.256289010781162e-05,
      "loss": 2.4292,
      "theoretical_loss": 3.3587528349817313,
      "tokens_seen": 2593062912
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.254634007944014e-05,
      "loss": 2.7166,
      "theoretical_loss": 3.3587461133466396,
      "tokens_seen": 2593128448
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.252979005106865e-05,
      "loss": 2.6398,
      "theoretical_loss": 3.358739391928985,
      "tokens_seen": 2593193984
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2513240022697176e-05,
      "loss": 2.2735,
      "theoretical_loss": 3.358732670728755,
      "tokens_seen": 2593259520
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.24966899943257e-05,
      "loss": 2.5311,
      "theoretical_loss": 3.3587259497459367,
      "tokens_seen": 2593325056
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2480139965954224e-05,
      "loss": 2.3308,
      "theoretical_loss": 3.358719228980519,
      "tokens_seen": 2593390592
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.246358993758275e-05,
      "loss": 2.2963,
      "theoretical_loss": 3.3587125084324874,
      "tokens_seen": 2593456128
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.244703990921127e-05,
      "loss": 1.8029,
      "theoretical_loss": 3.3587057881018305,
      "tokens_seen": 2593521664
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2430489880839796e-05,
      "loss": 2.6488,
      "theoretical_loss": 3.358699067988536,
      "tokens_seen": 2593587200
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.241393985246831e-05,
      "loss": 3.0973,
      "theoretical_loss": 3.3586923480925908,
      "tokens_seen": 2593652736
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.239738982409684e-05,
      "loss": 2.5706,
      "theoretical_loss": 3.358685628413982,
      "tokens_seen": 2593718272
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.238083979572536e-05,
      "loss": 2.35,
      "theoretical_loss": 3.3586789089526983,
      "tokens_seen": 2593783808
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2364289767353885e-05,
      "loss": 2.4054,
      "theoretical_loss": 3.3586721897087264,
      "tokens_seen": 2593849344
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.234773973898241e-05,
      "loss": 1.7645,
      "theoretical_loss": 3.358665470682054,
      "tokens_seen": 2593914880
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.233118971061093e-05,
      "loss": 2.0158,
      "theoretical_loss": 3.3586587518726683,
      "tokens_seen": 2593980416
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.231463968223946e-05,
      "loss": 2.0933,
      "theoretical_loss": 3.358652033280557,
      "tokens_seen": 2594045952
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.229808965386798e-05,
      "loss": 2.149,
      "theoretical_loss": 3.3586453149057074,
      "tokens_seen": 2594111488
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.22815396254965e-05,
      "loss": 1.7073,
      "theoretical_loss": 3.3586385967481074,
      "tokens_seen": 2594177024
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.226498959712502e-05,
      "loss": 2.9003,
      "theoretical_loss": 3.3586318788077443,
      "tokens_seen": 2594242560
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2248439568753547e-05,
      "loss": 2.6583,
      "theoretical_loss": 3.358625161084605,
      "tokens_seen": 2594308096
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.223188954038207e-05,
      "loss": 2.5949,
      "theoretical_loss": 3.358618443578678,
      "tokens_seen": 2594373632
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2215339512010595e-05,
      "loss": 2.4578,
      "theoretical_loss": 3.3586117262899506,
      "tokens_seen": 2594439168
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.219878948363912e-05,
      "loss": 2.3507,
      "theoretical_loss": 3.3586050092184094,
      "tokens_seen": 2594504704
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.218223945526764e-05,
      "loss": 1.918,
      "theoretical_loss": 3.358598292364043,
      "tokens_seen": 2594570240
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.216568942689615e-05,
      "loss": 2.7896,
      "theoretical_loss": 3.3585915757268374,
      "tokens_seen": 2594635776
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.214913939852468e-05,
      "loss": 3.4981,
      "theoretical_loss": 3.3585848593067817,
      "tokens_seen": 2594701312
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.21325893701532e-05,
      "loss": 2.6046,
      "theoretical_loss": 3.358578143103863,
      "tokens_seen": 2594766848
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2116039341781725e-05,
      "loss": 2.5232,
      "theoretical_loss": 3.3585714271180684,
      "tokens_seen": 2594832384
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.209948931341025e-05,
      "loss": 2.5813,
      "theoretical_loss": 3.3585647113493855,
      "tokens_seen": 2594897920
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.208293928503877e-05,
      "loss": 2.4546,
      "theoretical_loss": 3.358557995797802,
      "tokens_seen": 2594963456
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.20663892566673e-05,
      "loss": 2.3313,
      "theoretical_loss": 3.358551280463305,
      "tokens_seen": 2595028992
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.204983922829582e-05,
      "loss": 2.4488,
      "theoretical_loss": 3.3585445653458823,
      "tokens_seen": 2595094528
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.203328919992434e-05,
      "loss": 2.5333,
      "theoretical_loss": 3.3585378504455212,
      "tokens_seen": 2595160064
    },
    {
      "epoch": 0.93,
      "objective/train/avg_token_score": 0.012360312044620514,
      "objective/train/avg_weight": 0.9854807257652283,
      "objective/train/docs_used": 1460817,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.9248056411743164,
      "objective/train/std_weight": 0.08471377193927765,
      "objective/train/theoretical_loss": 3.3585311357622096,
      "objective/train/tokens_used": 2615685600,
      "theoretical_loss": 3.3585311357622096,
      "tokens_seen": 2595225600
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.201673917155286e-05,
      "loss": 2.4247,
      "theoretical_loss": 3.3585311357622096,
      "tokens_seen": 2595225600
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.2000189143181386e-05,
      "loss": 2.7371,
      "theoretical_loss": 3.3585244212959346,
      "tokens_seen": 2595291136
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.198363911480991e-05,
      "loss": 2.7543,
      "theoretical_loss": 3.3585177070466843,
      "tokens_seen": 2595356672
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1967089086438434e-05,
      "loss": 2.2439,
      "theoretical_loss": 3.3585109930144452,
      "tokens_seen": 2595422208
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.195053905806696e-05,
      "loss": 2.4624,
      "theoretical_loss": 3.358504279199206,
      "tokens_seen": 2595487744
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.193398902969548e-05,
      "loss": 3.7875,
      "theoretical_loss": 3.358497565600953,
      "tokens_seen": 2595553280
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1917439001324e-05,
      "loss": 2.2026,
      "theoretical_loss": 3.358490852219674,
      "tokens_seen": 2595618816
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.190088897295252e-05,
      "loss": 2.4096,
      "theoretical_loss": 3.3584841390553577,
      "tokens_seen": 2595684352
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.188433894458105e-05,
      "loss": 2.6892,
      "theoretical_loss": 3.3584774261079904,
      "tokens_seen": 2595749888
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.186778891620957e-05,
      "loss": 2.4737,
      "theoretical_loss": 3.3584707133775598,
      "tokens_seen": 2595815424
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1851238887838095e-05,
      "loss": 1.6289,
      "theoretical_loss": 3.3584640008640534,
      "tokens_seen": 2595880960
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.183468885946662e-05,
      "loss": 2.461,
      "theoretical_loss": 3.3584572885674593,
      "tokens_seen": 2595946496
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.181813883109514e-05,
      "loss": 2.5179,
      "theoretical_loss": 3.358450576487764,
      "tokens_seen": 2596012032
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.180158880272367e-05,
      "loss": 1.8565,
      "theoretical_loss": 3.358443864624956,
      "tokens_seen": 2596077568
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.178503877435218e-05,
      "loss": 2.4646,
      "theoretical_loss": 3.3584371529790222,
      "tokens_seen": 2596143104
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.17684887459807e-05,
      "loss": 1.5829,
      "theoretical_loss": 3.3584304415499506,
      "tokens_seen": 2596208640
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1751938717609226e-05,
      "loss": 2.3309,
      "theoretical_loss": 3.3584237303377282,
      "tokens_seen": 2596274176
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.173538868923775e-05,
      "loss": 0.2328,
      "theoretical_loss": 3.3584170193423426,
      "tokens_seen": 2596339712
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1718838660866274e-05,
      "loss": 2.5561,
      "theoretical_loss": 3.3584103085637818,
      "tokens_seen": 2596405248
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.17022886324948e-05,
      "loss": 1.5394,
      "theoretical_loss": 3.358403598002033,
      "tokens_seen": 2596470784
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.168573860412333e-05,
      "loss": 2.1905,
      "theoretical_loss": 3.3583968876570833,
      "tokens_seen": 2596536320
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.166918857575184e-05,
      "loss": 2.5967,
      "theoretical_loss": 3.3583901775289213,
      "tokens_seen": 2596601856
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.165263854738036e-05,
      "loss": 2.6171,
      "theoretical_loss": 3.3583834676175335,
      "tokens_seen": 2596667392
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.163608851900889e-05,
      "loss": 2.5411,
      "theoretical_loss": 3.358376757922908,
      "tokens_seen": 2596732928
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.161953849063741e-05,
      "loss": 2.5729,
      "theoretical_loss": 3.358370048445032,
      "tokens_seen": 2596798464
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1602988462265935e-05,
      "loss": 2.3799,
      "theoretical_loss": 3.3583633391838927,
      "tokens_seen": 2596864000
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.158643843389446e-05,
      "loss": 1.9958,
      "theoretical_loss": 3.3583566301394785,
      "tokens_seen": 2596929536
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.156988840552298e-05,
      "loss": 2.7558,
      "theoretical_loss": 3.3583499213117767,
      "tokens_seen": 2596995072
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.15533383771515e-05,
      "loss": 2.365,
      "theoretical_loss": 3.358343212700774,
      "tokens_seen": 2597060608
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1536788348780024e-05,
      "loss": 2.174,
      "theoretical_loss": 3.3583365043064592,
      "tokens_seen": 2597126144
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.152023832040855e-05,
      "loss": 2.0871,
      "theoretical_loss": 3.358329796128819,
      "tokens_seen": 2597191680
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.150368829203707e-05,
      "loss": 2.5204,
      "theoretical_loss": 3.3583230881678414,
      "tokens_seen": 2597257216
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1487138263665596e-05,
      "loss": -0.5981,
      "theoretical_loss": 3.358316380423513,
      "tokens_seen": 2597322752
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.147058823529412e-05,
      "loss": 2.281,
      "theoretical_loss": 3.358309672895823,
      "tokens_seen": 2597388288
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1454038206922644e-05,
      "loss": 2.729,
      "theoretical_loss": 3.358302965584757,
      "tokens_seen": 2597453824
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.143748817855117e-05,
      "loss": 2.0028,
      "theoretical_loss": 3.358296258490304,
      "tokens_seen": 2597519360
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.142093815017968e-05,
      "loss": 2.6689,
      "theoretical_loss": 3.358289551612451,
      "tokens_seen": 2597584896
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.14043881218082e-05,
      "loss": 2.0903,
      "theoretical_loss": 3.3582828449511855,
      "tokens_seen": 2597650432
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1387838093436726e-05,
      "loss": 2.5162,
      "theoretical_loss": 3.358276138506495,
      "tokens_seen": 2597715968
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.137128806506526e-05,
      "loss": 2.383,
      "theoretical_loss": 3.3582694322783673,
      "tokens_seen": 2597781504
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.135473803669378e-05,
      "loss": 2.7848,
      "theoretical_loss": 3.3582627262667897,
      "tokens_seen": 2597847040
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1338188008322305e-05,
      "loss": 2.5586,
      "theoretical_loss": 3.35825602047175,
      "tokens_seen": 2597912576
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.132163797995083e-05,
      "loss": 2.2947,
      "theoretical_loss": 3.3582493148932357,
      "tokens_seen": 2597978112
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.130508795157934e-05,
      "loss": 2.2988,
      "theoretical_loss": 3.358242609531234,
      "tokens_seen": 2598043648
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1288537923207863e-05,
      "loss": 2.6127,
      "theoretical_loss": 3.3582359043857326,
      "tokens_seen": 2598109184
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.127198789483639e-05,
      "loss": 2.3373,
      "theoretical_loss": 3.3582291994567193,
      "tokens_seen": 2598174720
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.125543786646491e-05,
      "loss": 2.1624,
      "theoretical_loss": 3.3582224947441817,
      "tokens_seen": 2598240256
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1238887838093435e-05,
      "loss": 2.7188,
      "theoretical_loss": 3.3582157902481065,
      "tokens_seen": 2598305792
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.122233780972196e-05,
      "loss": 2.7009,
      "theoretical_loss": 3.3582090859684826,
      "tokens_seen": 2598371328
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1205787781350483e-05,
      "loss": 2.667,
      "theoretical_loss": 3.358202381905296,
      "tokens_seen": 2598436864
    },
    {
      "epoch": 0.93,
      "objective/train/avg_token_score": 0.0131613127887249,
      "objective/train/avg_weight": 0.9845088720321655,
      "objective/train/docs_used": 1462730,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.090369939804077,
      "objective/train/std_weight": 0.071063332259655,
      "objective/train/theoretical_loss": 3.358195678058536,
      "objective/train/tokens_used": 2618962400,
      "theoretical_loss": 3.358195678058536,
      "tokens_seen": 2598502400
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.118923775297901e-05,
      "loss": 2.6059,
      "theoretical_loss": 3.358195678058536,
      "tokens_seen": 2598502400
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1172687724607525e-05,
      "loss": 1.358,
      "theoretical_loss": 3.358188974428189,
      "tokens_seen": 2598567936
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.115613769623605e-05,
      "loss": 4.5894,
      "theoretical_loss": 3.3581822710142424,
      "tokens_seen": 2598633472
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.113958766786457e-05,
      "loss": 1.9922,
      "theoretical_loss": 3.3581755678166845,
      "tokens_seen": 2598699008
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.11230376394931e-05,
      "loss": 2.0032,
      "theoretical_loss": 3.358168864835503,
      "tokens_seen": 2598764544
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.110648761112162e-05,
      "loss": 2.6047,
      "theoretical_loss": 3.358162162070684,
      "tokens_seen": 2598830080
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1089937582750145e-05,
      "loss": 2.3479,
      "theoretical_loss": 3.3581554595222167,
      "tokens_seen": 2598895616
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.107338755437867e-05,
      "loss": 2.5933,
      "theoretical_loss": 3.3581487571900874,
      "tokens_seen": 2598961152
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1056837526007186e-05,
      "loss": 2.8127,
      "theoretical_loss": 3.3581420550742846,
      "tokens_seen": 2599026688
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.104028749763571e-05,
      "loss": 2.564,
      "theoretical_loss": 3.358135353174796,
      "tokens_seen": 2599092224
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.1023737469264234e-05,
      "loss": 2.6101,
      "theoretical_loss": 3.358128651491608,
      "tokens_seen": 2599157760
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.100718744089276e-05,
      "loss": 2.5253,
      "theoretical_loss": 3.3581219500247093,
      "tokens_seen": 2599223296
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.099063741252128e-05,
      "loss": 2.541,
      "theoretical_loss": 3.358115248774087,
      "tokens_seen": 2599288832
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0974087384149806e-05,
      "loss": 2.71,
      "theoretical_loss": 3.3581085477397283,
      "tokens_seen": 2599354368
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.095753735577833e-05,
      "loss": 2.2247,
      "theoretical_loss": 3.3581018469216213,
      "tokens_seen": 2599419904
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0940987327406854e-05,
      "loss": 2.6108,
      "theoretical_loss": 3.3580951463197537,
      "tokens_seen": 2599485440
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0924437299035364e-05,
      "loss": 2.567,
      "theoretical_loss": 3.3580884459341123,
      "tokens_seen": 2599550976
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.090788727066389e-05,
      "loss": 2.4047,
      "theoretical_loss": 3.358081745764686,
      "tokens_seen": 2599616512
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.089133724229241e-05,
      "loss": 2.3171,
      "theoretical_loss": 3.3580750458114608,
      "tokens_seen": 2599682048
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0874787213920936e-05,
      "loss": 1.9927,
      "theoretical_loss": 3.3580683460744254,
      "tokens_seen": 2599747584
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.085823718554946e-05,
      "loss": 0.772,
      "theoretical_loss": 3.3580616465535664,
      "tokens_seen": 2599813120
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0841687157177984e-05,
      "loss": 2.8208,
      "theoretical_loss": 3.3580549472488728,
      "tokens_seen": 2599878656
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.082513712880651e-05,
      "loss": 2.1727,
      "theoretical_loss": 3.3580482481603307,
      "tokens_seen": 2599944192
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0808587100435025e-05,
      "loss": 2.8931,
      "theoretical_loss": 3.358041549287929,
      "tokens_seen": 2600009728
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.079203707206355e-05,
      "loss": 2.6801,
      "theoretical_loss": 3.358034850631654,
      "tokens_seen": 2600075264
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.077548704369207e-05,
      "loss": 2.6734,
      "theoretical_loss": 3.358028152191494,
      "tokens_seen": 2600140800
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.07589370153206e-05,
      "loss": 1.9333,
      "theoretical_loss": 3.3580214539674365,
      "tokens_seen": 2600206336
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.074238698694912e-05,
      "loss": 2.3906,
      "theoretical_loss": 3.358014755959469,
      "tokens_seen": 2600271872
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0725836958577645e-05,
      "loss": 2.5103,
      "theoretical_loss": 3.358008058167579,
      "tokens_seen": 2600337408
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.070928693020617e-05,
      "loss": 2.1507,
      "theoretical_loss": 3.3580013605917545,
      "tokens_seen": 2600402944
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0692736901834687e-05,
      "loss": 2.5607,
      "theoretical_loss": 3.357994663231983,
      "tokens_seen": 2600468480
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.067618687346321e-05,
      "loss": 2.3171,
      "theoretical_loss": 3.3579879660882512,
      "tokens_seen": 2600534016
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0659636845091735e-05,
      "loss": 1.901,
      "theoretical_loss": 3.3579812691605477,
      "tokens_seen": 2600599552
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.064308681672026e-05,
      "loss": 2.5265,
      "theoretical_loss": 3.35797457244886,
      "tokens_seen": 2600665088
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.062653678834878e-05,
      "loss": 2.0949,
      "theoretical_loss": 3.3579678759531753,
      "tokens_seen": 2600730624
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0609986759977306e-05,
      "loss": 2.8143,
      "theoretical_loss": 3.357961179673481,
      "tokens_seen": 2600796160
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.059343673160583e-05,
      "loss": 2.4831,
      "theoretical_loss": 3.357954483609765,
      "tokens_seen": 2600861696
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0576886703234354e-05,
      "loss": 2.2896,
      "theoretical_loss": 3.3579477877620154,
      "tokens_seen": 2600927232
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0560336674862865e-05,
      "loss": 2.5693,
      "theoretical_loss": 3.3579410921302193,
      "tokens_seen": 2600992768
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.054378664649139e-05,
      "loss": 2.6355,
      "theoretical_loss": 3.3579343967143642,
      "tokens_seen": 2601058304
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.052723661811991e-05,
      "loss": 1.9772,
      "theoretical_loss": 3.3579277015144378,
      "tokens_seen": 2601123840
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.051068658974844e-05,
      "loss": 2.2307,
      "theoretical_loss": 3.3579210065304275,
      "tokens_seen": 2601189376
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.049413656137696e-05,
      "loss": 2.0513,
      "theoretical_loss": 3.357914311762321,
      "tokens_seen": 2601254912
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0477586533005485e-05,
      "loss": 2.1331,
      "theoretical_loss": 3.357907617210106,
      "tokens_seen": 2601320448
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.046103650463401e-05,
      "loss": 2.0308,
      "theoretical_loss": 3.3579009228737706,
      "tokens_seen": 2601385984
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0444486476262526e-05,
      "loss": 2.5037,
      "theoretical_loss": 3.3578942287533016,
      "tokens_seen": 2601451520
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.042793644789105e-05,
      "loss": 1.9953,
      "theoretical_loss": 3.3578875348486865,
      "tokens_seen": 2601517056
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0411386419519574e-05,
      "loss": 2.2325,
      "theoretical_loss": 3.357880841159914,
      "tokens_seen": 2601582592
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.03948363911481e-05,
      "loss": 1.6046,
      "theoretical_loss": 3.3578741476869705,
      "tokens_seen": 2601648128
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.037828636277662e-05,
      "loss": 2.2719,
      "theoretical_loss": 3.357867454429844,
      "tokens_seen": 2601713664
    },
    {
      "epoch": 0.93,
      "objective/train/avg_token_score": 0.010727731510996819,
      "objective/train/avg_weight": 0.987374484539032,
      "objective/train/docs_used": 1464443,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.737276554107666,
      "objective/train/std_weight": 0.040072161704301834,
      "objective/train/theoretical_loss": 3.3578607613885225,
      "objective/train/tokens_used": 2622239200,
      "theoretical_loss": 3.3578607613885225,
      "tokens_seen": 2601779200
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0361736334405146e-05,
      "loss": 2.6245,
      "theoretical_loss": 3.3578607613885225,
      "tokens_seen": 2601779200
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.034518630603367e-05,
      "loss": 1.9631,
      "theoretical_loss": 3.357854068562993,
      "tokens_seen": 2601844736
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0328636277662194e-05,
      "loss": 2.5356,
      "theoretical_loss": 3.357847375953244,
      "tokens_seen": 2601910272
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.031208624929071e-05,
      "loss": 2.6036,
      "theoretical_loss": 3.357840683559262,
      "tokens_seen": 2601975808
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0295536220919235e-05,
      "loss": 2.5347,
      "theoretical_loss": 3.3578339913810353,
      "tokens_seen": 2602041344
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.027898619254776e-05,
      "loss": 2.5204,
      "theoretical_loss": 3.3578272994185516,
      "tokens_seen": 2602106880
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.026243616417628e-05,
      "loss": 2.3022,
      "theoretical_loss": 3.357820607671798,
      "tokens_seen": 2602172416
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.024588613580481e-05,
      "loss": 1.6522,
      "theoretical_loss": 3.3578139161407625,
      "tokens_seen": 2602237952
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.022933610743333e-05,
      "loss": 3.0327,
      "theoretical_loss": 3.3578072248254327,
      "tokens_seen": 2602303488
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0212786079061855e-05,
      "loss": 2.5125,
      "theoretical_loss": 3.3578005337257957,
      "tokens_seen": 2602369024
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0196236050690366e-05,
      "loss": 1.7497,
      "theoretical_loss": 3.35779384284184,
      "tokens_seen": 2602434560
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.017968602231889e-05,
      "loss": 2.4516,
      "theoretical_loss": 3.357787152173552,
      "tokens_seen": 2602500096
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0163135993947414e-05,
      "loss": 0.8844,
      "theoretical_loss": 3.3577804617209206,
      "tokens_seen": 2602565632
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.014658596557594e-05,
      "loss": 2.3526,
      "theoretical_loss": 3.3577737714839326,
      "tokens_seen": 2602631168
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.013003593720446e-05,
      "loss": 2.3124,
      "theoretical_loss": 3.357767081462576,
      "tokens_seen": 2602696704
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0113485908832986e-05,
      "loss": 3.4693,
      "theoretical_loss": 3.357760391656839,
      "tokens_seen": 2602762240
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.009693588046151e-05,
      "loss": 2.6847,
      "theoretical_loss": 3.3577537020667076,
      "tokens_seen": 2602827776
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0080385852090034e-05,
      "loss": 2.6431,
      "theoretical_loss": 3.3577470126921707,
      "tokens_seen": 2602893312
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.006383582371855e-05,
      "loss": 2.2711,
      "theoretical_loss": 3.3577403235332155,
      "tokens_seen": 2602958848
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.0047285795347075e-05,
      "loss": 3.4578,
      "theoretical_loss": 3.3577336345898297,
      "tokens_seen": 2603024384
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.00307357669756e-05,
      "loss": 1.3794,
      "theoretical_loss": 3.3577269458620007,
      "tokens_seen": 2603089920
    },
    {
      "epoch": 0.93,
      "learning_rate": 5.001418573860412e-05,
      "loss": 2.4427,
      "theoretical_loss": 3.357720257349717,
      "tokens_seen": 2603155456
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.999763571023265e-05,
      "loss": 2.73,
      "theoretical_loss": 3.357713569052965,
      "tokens_seen": 2603220992
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.998108568186117e-05,
      "loss": 1.5037,
      "theoretical_loss": 3.357706880971733,
      "tokens_seen": 2603286528
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9964535653489695e-05,
      "loss": 2.6711,
      "theoretical_loss": 3.357700193106009,
      "tokens_seen": 2603352064
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.994798562511821e-05,
      "loss": 2.6242,
      "theoretical_loss": 3.3576935054557797,
      "tokens_seen": 2603417600
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9931435596746736e-05,
      "loss": 2.476,
      "theoretical_loss": 3.3576868180210333,
      "tokens_seen": 2603483136
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.991488556837526e-05,
      "loss": 2.7331,
      "theoretical_loss": 3.357680130801757,
      "tokens_seen": 2603548672
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9898335540003784e-05,
      "loss": 2.1171,
      "theoretical_loss": 3.3576734437979394,
      "tokens_seen": 2603614208
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.988178551163231e-05,
      "loss": 2.5435,
      "theoretical_loss": 3.357666757009567,
      "tokens_seen": 2603679744
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.986523548326083e-05,
      "loss": 2.4358,
      "theoretical_loss": 3.357660070436628,
      "tokens_seen": 2603745280
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9848685454889356e-05,
      "loss": 2.3145,
      "theoretical_loss": 3.3576533840791107,
      "tokens_seen": 2603810816
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9832135426517866e-05,
      "loss": 2.9349,
      "theoretical_loss": 3.3576466979370014,
      "tokens_seen": 2603876352
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.981558539814639e-05,
      "loss": 2.4791,
      "theoretical_loss": 3.3576400120102883,
      "tokens_seen": 2603941888
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9799035369774914e-05,
      "loss": 2.521,
      "theoretical_loss": 3.3576333262989593,
      "tokens_seen": 2604007424
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.978248534140344e-05,
      "loss": 1.4733,
      "theoretical_loss": 3.3576266408030015,
      "tokens_seen": 2604072960
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.976593531303196e-05,
      "loss": 2.5446,
      "theoretical_loss": 3.3576199555224036,
      "tokens_seen": 2604138496
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9749385284660486e-05,
      "loss": 2.4577,
      "theoretical_loss": 3.357613270457152,
      "tokens_seen": 2604204032
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.973283525628901e-05,
      "loss": 2.7245,
      "theoretical_loss": 3.357606585607235,
      "tokens_seen": 2604269568
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9716285227917534e-05,
      "loss": 2.4847,
      "theoretical_loss": 3.35759990097264,
      "tokens_seen": 2604335104
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.969973519954605e-05,
      "loss": 2.2962,
      "theoretical_loss": 3.3575932165533544,
      "tokens_seen": 2604400640
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9683185171174575e-05,
      "loss": 2.5575,
      "theoretical_loss": 3.3575865323493668,
      "tokens_seen": 2604466176
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.96666351428031e-05,
      "loss": 2.4698,
      "theoretical_loss": 3.357579848360664,
      "tokens_seen": 2604531712
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9650085114431623e-05,
      "loss": 2.6022,
      "theoretical_loss": 3.357573164587234,
      "tokens_seen": 2604597248
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.963353508606015e-05,
      "loss": 2.5361,
      "theoretical_loss": 3.3575664810290644,
      "tokens_seen": 2604662784
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.961698505768867e-05,
      "loss": 1.8814,
      "theoretical_loss": 3.3575597976861427,
      "tokens_seen": 2604728320
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9600435029317195e-05,
      "loss": 2.4954,
      "theoretical_loss": 3.3575531145584567,
      "tokens_seen": 2604793856
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.958388500094571e-05,
      "loss": 2.1362,
      "theoretical_loss": 3.3575464316459938,
      "tokens_seen": 2604859392
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.956733497257424e-05,
      "loss": 3.1097,
      "theoretical_loss": 3.357539748948742,
      "tokens_seen": 2604924928
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.955078494420276e-05,
      "loss": 2.1438,
      "theoretical_loss": 3.357533066466689,
      "tokens_seen": 2604990464
    },
    {
      "epoch": 0.93,
      "objective/train/avg_token_score": 0.005751020275056362,
      "objective/train/avg_weight": 0.9932690858840942,
      "objective/train/docs_used": 1466332,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6687798500061035,
      "objective/train/std_weight": 0.04212731122970581,
      "objective/train/theoretical_loss": 3.357526384199822,
      "objective/train/tokens_used": 2625516000,
      "theoretical_loss": 3.357526384199822,
      "tokens_seen": 2605056000
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9534234915831285e-05,
      "loss": 1.502,
      "theoretical_loss": 3.357526384199822,
      "tokens_seen": 2605056000
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.951768488745981e-05,
      "loss": 2.5444,
      "theoretical_loss": 3.3575197021481293,
      "tokens_seen": 2605121536
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.950113485908833e-05,
      "loss": 2.3245,
      "theoretical_loss": 3.357513020311598,
      "tokens_seen": 2605187072
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9484584830716857e-05,
      "loss": 2.4683,
      "theoretical_loss": 3.357506338690216,
      "tokens_seen": 2605252608
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.946803480234538e-05,
      "loss": 2.5374,
      "theoretical_loss": 3.3574996572839706,
      "tokens_seen": 2605318144
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.945148477397389e-05,
      "loss": 3.0667,
      "theoretical_loss": 3.3574929760928502,
      "tokens_seen": 2605383680
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9434934745602415e-05,
      "loss": 2.6821,
      "theoretical_loss": 3.357486295116842,
      "tokens_seen": 2605449216
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.941838471723094e-05,
      "loss": 2.5493,
      "theoretical_loss": 3.3574796143559333,
      "tokens_seen": 2605514752
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.940183468885946e-05,
      "loss": 1.7716,
      "theoretical_loss": 3.3574729338101124,
      "tokens_seen": 2605580288
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.938528466048799e-05,
      "loss": 2.1097,
      "theoretical_loss": 3.357466253479367,
      "tokens_seen": 2605645824
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.936873463211652e-05,
      "loss": 2.5543,
      "theoretical_loss": 3.357459573363684,
      "tokens_seen": 2605711360
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.935218460374504e-05,
      "loss": 2.2931,
      "theoretical_loss": 3.357452893463052,
      "tokens_seen": 2605776896
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.933563457537355e-05,
      "loss": 2.0404,
      "theoretical_loss": 3.357446213777458,
      "tokens_seen": 2605842432
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9319084547002076e-05,
      "loss": 1.5419,
      "theoretical_loss": 3.3574395343068897,
      "tokens_seen": 2605907968
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.93025345186306e-05,
      "loss": 2.4883,
      "theoretical_loss": 3.3574328550513353,
      "tokens_seen": 2605973504
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9285984490259124e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.357426176010782,
      "tokens_seen": 2606039040
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.926943446188765e-05,
      "loss": 2.7245,
      "theoretical_loss": 3.3574194971852176,
      "tokens_seen": 2606104576
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.925288443351617e-05,
      "loss": 2.4463,
      "theoretical_loss": 3.35741281857463,
      "tokens_seen": 2606170112
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9236334405144696e-05,
      "loss": 2.4421,
      "theoretical_loss": 3.3574061401790067,
      "tokens_seen": 2606235648
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.921978437677322e-05,
      "loss": 3.3334,
      "theoretical_loss": 3.357399461998335,
      "tokens_seen": 2606301184
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.920323434840174e-05,
      "loss": 1.6253,
      "theoretical_loss": 3.357392784032603,
      "tokens_seen": 2606366720
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.918668432003026e-05,
      "loss": 2.5263,
      "theoretical_loss": 3.3573861062817985,
      "tokens_seen": 2606432256
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9170134291658785e-05,
      "loss": 2.3885,
      "theoretical_loss": 3.357379428745909,
      "tokens_seen": 2606497792
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.915358426328731e-05,
      "loss": 2.1134,
      "theoretical_loss": 3.357372751424922,
      "tokens_seen": 2606563328
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.913703423491583e-05,
      "loss": 2.6361,
      "theoretical_loss": 3.357366074318825,
      "tokens_seen": 2606628864
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.912048420654436e-05,
      "loss": 2.2022,
      "theoretical_loss": 3.3573593974276066,
      "tokens_seen": 2606694400
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.910393417817288e-05,
      "loss": 1.9858,
      "theoretical_loss": 3.357352720751254,
      "tokens_seen": 2606759936
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.908738414980139e-05,
      "loss": 2.1419,
      "theoretical_loss": 3.3573460442897542,
      "tokens_seen": 2606825472
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9070834121429916e-05,
      "loss": 2.3565,
      "theoretical_loss": 3.3573393680430956,
      "tokens_seen": 2606891008
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9054284093058447e-05,
      "loss": 2.5891,
      "theoretical_loss": 3.357332692011266,
      "tokens_seen": 2606956544
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.903773406468697e-05,
      "loss": 2.4703,
      "theoretical_loss": 3.3573260161942526,
      "tokens_seen": 2607022080
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.9021184036315494e-05,
      "loss": 2.1523,
      "theoretical_loss": 3.357319340592044,
      "tokens_seen": 2607087616
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.900463400794402e-05,
      "loss": 2.2806,
      "theoretical_loss": 3.3573126652046263,
      "tokens_seen": 2607153152
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.898808397957254e-05,
      "loss": 2.3355,
      "theoretical_loss": 3.3573059900319886,
      "tokens_seen": 2607218688
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.897153395120105e-05,
      "loss": 3.8354,
      "theoretical_loss": 3.357299315074118,
      "tokens_seen": 2607284224
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.895498392282958e-05,
      "loss": 2.7392,
      "theoretical_loss": 3.357292640331002,
      "tokens_seen": 2607349760
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.89384338944581e-05,
      "loss": 2.1054,
      "theoretical_loss": 3.357285965802629,
      "tokens_seen": 2607415296
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8921883866086625e-05,
      "loss": 2.3825,
      "theoretical_loss": 3.357279291488986,
      "tokens_seen": 2607480832
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.890533383771515e-05,
      "loss": 2.6579,
      "theoretical_loss": 3.3572726173900613,
      "tokens_seen": 2607546368
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.888878380934367e-05,
      "loss": 3.1343,
      "theoretical_loss": 3.3572659435058423,
      "tokens_seen": 2607611904
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.88722337809722e-05,
      "loss": 2.6843,
      "theoretical_loss": 3.3572592698363164,
      "tokens_seen": 2607677440
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.885568375260072e-05,
      "loss": 2.4349,
      "theoretical_loss": 3.3572525963814717,
      "tokens_seen": 2607742976
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.883913372422924e-05,
      "loss": 2.6101,
      "theoretical_loss": 3.3572459231412957,
      "tokens_seen": 2607808512
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.882258369585776e-05,
      "loss": 0.8573,
      "theoretical_loss": 3.357239250115776,
      "tokens_seen": 2607874048
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8806033667486286e-05,
      "loss": 2.3844,
      "theoretical_loss": 3.3572325773049005,
      "tokens_seen": 2607939584
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.878948363911481e-05,
      "loss": 3.1872,
      "theoretical_loss": 3.357225904708657,
      "tokens_seen": 2608005120
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8772933610743334e-05,
      "loss": 2.1091,
      "theoretical_loss": 3.3572192323270333,
      "tokens_seen": 2608070656
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.875638358237186e-05,
      "loss": 2.3512,
      "theoretical_loss": 3.3572125601600167,
      "tokens_seen": 2608136192
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.873983355400038e-05,
      "loss": 0.7715,
      "theoretical_loss": 3.357205888207595,
      "tokens_seen": 2608201728
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.87232835256289e-05,
      "loss": 2.1434,
      "theoretical_loss": 3.357199216469756,
      "tokens_seen": 2608267264
    },
    {
      "epoch": 0.93,
      "objective/train/avg_token_score": 0.011956076137721539,
      "objective/train/avg_weight": 0.9859285950660706,
      "objective/train/docs_used": 1468043,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 11.311209678649902,
      "objective/train/std_weight": 0.07475373148918152,
      "objective/train/theoretical_loss": 3.3571925449464874,
      "objective/train/tokens_used": 2628792800,
      "theoretical_loss": 3.3571925449464874,
      "tokens_seen": 2608332800
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.870673349725742e-05,
      "loss": 2.0187,
      "theoretical_loss": 3.3571925449464874,
      "tokens_seen": 2608332800
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.869018346888595e-05,
      "loss": 2.4396,
      "theoretical_loss": 3.3571858736377767,
      "tokens_seen": 2608398336
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.867363344051447e-05,
      "loss": 2.6938,
      "theoretical_loss": 3.357179202543612,
      "tokens_seen": 2608463872
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8657083412142995e-05,
      "loss": 2.8878,
      "theoretical_loss": 3.3571725316639807,
      "tokens_seen": 2608529408
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.864053338377152e-05,
      "loss": 2.5428,
      "theoretical_loss": 3.357165860998871,
      "tokens_seen": 2608594944
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.862398335540004e-05,
      "loss": 2.7183,
      "theoretical_loss": 3.35715919054827,
      "tokens_seen": 2608660480
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.860743332702857e-05,
      "loss": 2.6771,
      "theoretical_loss": 3.3571525203121655,
      "tokens_seen": 2608726016
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.859088329865708e-05,
      "loss": 1.0273,
      "theoretical_loss": 3.3571458502905456,
      "tokens_seen": 2608791552
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.85743332702856e-05,
      "loss": 0.6133,
      "theoretical_loss": 3.3571391804833977,
      "tokens_seen": 2608857088
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8557783241914126e-05,
      "loss": 1.3543,
      "theoretical_loss": 3.3571325108907097,
      "tokens_seen": 2608922624
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.854123321354265e-05,
      "loss": 2.3019,
      "theoretical_loss": 3.357125841512469,
      "tokens_seen": 2608988160
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8524683185171174e-05,
      "loss": 2.6606,
      "theoretical_loss": 3.3571191723486637,
      "tokens_seen": 2609053696
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.85081331567997e-05,
      "loss": 2.7046,
      "theoretical_loss": 3.3571125033992812,
      "tokens_seen": 2609119232
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.849158312842822e-05,
      "loss": 2.4436,
      "theoretical_loss": 3.35710583466431,
      "tokens_seen": 2609184768
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.847503310005674e-05,
      "loss": 2.0651,
      "theoretical_loss": 3.357099166143737,
      "tokens_seen": 2609250304
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.845848307168526e-05,
      "loss": 1.8297,
      "theoretical_loss": 3.3570924978375496,
      "tokens_seen": 2609315840
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.844193304331379e-05,
      "loss": 1.9843,
      "theoretical_loss": 3.357085829745736,
      "tokens_seen": 2609381376
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.842538301494231e-05,
      "loss": 3.3054,
      "theoretical_loss": 3.3570791618682843,
      "tokens_seen": 2609446912
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8408832986570835e-05,
      "loss": 2.3882,
      "theoretical_loss": 3.357072494205182,
      "tokens_seen": 2609512448
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.839228295819936e-05,
      "loss": 2.335,
      "theoretical_loss": 3.3570658267564166,
      "tokens_seen": 2609577984
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.837573292982788e-05,
      "loss": 1.5482,
      "theoretical_loss": 3.357059159521976,
      "tokens_seen": 2609643520
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.835918290145641e-05,
      "loss": 1.5942,
      "theoretical_loss": 3.3570524925018477,
      "tokens_seen": 2609709056
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8342632873084924e-05,
      "loss": 2.023,
      "theoretical_loss": 3.35704582569602,
      "tokens_seen": 2609774592
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.832608284471345e-05,
      "loss": 2.7003,
      "theoretical_loss": 3.3570391591044797,
      "tokens_seen": 2609840128
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.830953281634197e-05,
      "loss": 2.3134,
      "theoretical_loss": 3.3570324927272157,
      "tokens_seen": 2609905664
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8292982787970496e-05,
      "loss": 2.62,
      "theoretical_loss": 3.3570258265642146,
      "tokens_seen": 2609971200
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.827643275959902e-05,
      "loss": 2.7438,
      "theoretical_loss": 3.3570191606154647,
      "tokens_seen": 2610036736
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8259882731227544e-05,
      "loss": 1.9355,
      "theoretical_loss": 3.3570124948809537,
      "tokens_seen": 2610102272
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.824333270285607e-05,
      "loss": 2.5244,
      "theoretical_loss": 3.3570058293606695,
      "tokens_seen": 2610167808
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.822678267448458e-05,
      "loss": 2.5457,
      "theoretical_loss": 3.3569991640545993,
      "tokens_seen": 2610233344
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.82102326461131e-05,
      "loss": 2.8014,
      "theoretical_loss": 3.3569924989627316,
      "tokens_seen": 2610298880
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8193682617741626e-05,
      "loss": 2.0495,
      "theoretical_loss": 3.3569858340850534,
      "tokens_seen": 2610364416
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.817713258937015e-05,
      "loss": 2.4077,
      "theoretical_loss": 3.3569791694215527,
      "tokens_seen": 2610429952
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8160582560998674e-05,
      "loss": 2.7045,
      "theoretical_loss": 3.3569725049722177,
      "tokens_seen": 2610495488
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.81440325326272e-05,
      "loss": 2.4654,
      "theoretical_loss": 3.3569658407370353,
      "tokens_seen": 2610561024
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.812748250425572e-05,
      "loss": 2.0712,
      "theoretical_loss": 3.356959176715994,
      "tokens_seen": 2610626560
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.811093247588424e-05,
      "loss": 2.2142,
      "theoretical_loss": 3.356952512909081,
      "tokens_seen": 2610692096
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8094382447512763e-05,
      "loss": 2.637,
      "theoretical_loss": 3.3569458493162845,
      "tokens_seen": 2610757632
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.807783241914129e-05,
      "loss": 2.5713,
      "theoretical_loss": 3.356939185937592,
      "tokens_seen": 2610823168
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.806128239076981e-05,
      "loss": 1.883,
      "theoretical_loss": 3.356932522772991,
      "tokens_seen": 2610888704
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8044732362398335e-05,
      "loss": 0.7467,
      "theoretical_loss": 3.3569258598224696,
      "tokens_seen": 2610954240
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.802818233402686e-05,
      "loss": 2.5269,
      "theoretical_loss": 3.3569191970860155,
      "tokens_seen": 2611019776
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.8011632305655383e-05,
      "loss": 2.4657,
      "theoretical_loss": 3.356912534563617,
      "tokens_seen": 2611085312
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.799508227728391e-05,
      "loss": 2.5426,
      "theoretical_loss": 3.3569058722552603,
      "tokens_seen": 2611150848
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7978532248912425e-05,
      "loss": 1.7688,
      "theoretical_loss": 3.3568992101609347,
      "tokens_seen": 2611216384
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.796198222054095e-05,
      "loss": 2.6524,
      "theoretical_loss": 3.3568925482806273,
      "tokens_seen": 2611281920
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.794543219216947e-05,
      "loss": 2.7528,
      "theoretical_loss": 3.3568858866143256,
      "tokens_seen": 2611347456
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7928882163798e-05,
      "loss": 2.9541,
      "theoretical_loss": 3.356879225162018,
      "tokens_seen": 2611412992
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.791233213542652e-05,
      "loss": 2.0215,
      "theoretical_loss": 3.356872563923692,
      "tokens_seen": 2611478528
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7895782107055045e-05,
      "loss": 2.064,
      "theoretical_loss": 3.356865902899335,
      "tokens_seen": 2611544064
    },
    {
      "epoch": 0.93,
      "objective/train/avg_token_score": 0.006494482979178429,
      "objective/train/avg_weight": 0.9923556447029114,
      "objective/train/docs_used": 1469063,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.291926622390747,
      "objective/train/std_weight": 0.050927869975566864,
      "objective/train/theoretical_loss": 3.356859242088935,
      "objective/train/tokens_used": 2632069600,
      "theoretical_loss": 3.356859242088935,
      "tokens_seen": 2611609600
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.787923207868357e-05,
      "loss": 2.2731,
      "theoretical_loss": 3.356859242088935,
      "tokens_seen": 2611609600
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.786268205031208e-05,
      "loss": 2.5804,
      "theoretical_loss": 3.35685258149248,
      "tokens_seen": 2611675136
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.78461320219406e-05,
      "loss": 5.7139,
      "theoretical_loss": 3.3568459211099575,
      "tokens_seen": 2611740672
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.782958199356913e-05,
      "loss": 1.987,
      "theoretical_loss": 3.356839260941355,
      "tokens_seen": 2611806208
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.781303196519765e-05,
      "loss": 2.5688,
      "theoretical_loss": 3.356832600986661,
      "tokens_seen": 2611871744
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7796481936826175e-05,
      "loss": 3.3827,
      "theoretical_loss": 3.356825941245863,
      "tokens_seen": 2611937280
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.77799319084547e-05,
      "loss": 2.6212,
      "theoretical_loss": 3.3568192817189484,
      "tokens_seen": 2612002816
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.776338188008322e-05,
      "loss": 2.2913,
      "theoretical_loss": 3.356812622405905,
      "tokens_seen": 2612068352
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.774683185171175e-05,
      "loss": 3.905,
      "theoretical_loss": 3.356805963306721,
      "tokens_seen": 2612133888
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7730281823340264e-05,
      "loss": 2.6969,
      "theoretical_loss": 3.3567993044213837,
      "tokens_seen": 2612199424
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.771373179496879e-05,
      "loss": 2.2757,
      "theoretical_loss": 3.3567926457498816,
      "tokens_seen": 2612264960
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.769718176659731e-05,
      "loss": 2.5825,
      "theoretical_loss": 3.3567859872922012,
      "tokens_seen": 2612330496
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7680631738225836e-05,
      "loss": 2.8469,
      "theoretical_loss": 3.356779329048331,
      "tokens_seen": 2612396032
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.766408170985436e-05,
      "loss": 1.8858,
      "theoretical_loss": 3.3567726710182595,
      "tokens_seen": 2612461568
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7647531681482884e-05,
      "loss": 2.4326,
      "theoretical_loss": 3.356766013201973,
      "tokens_seen": 2612527104
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.763098165311141e-05,
      "loss": 2.6205,
      "theoretical_loss": 3.356759355599461,
      "tokens_seen": 2612592640
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7614431624739925e-05,
      "loss": 6.2176,
      "theoretical_loss": 3.35675269821071,
      "tokens_seen": 2612658176
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.759788159636845e-05,
      "loss": 2.6853,
      "theoretical_loss": 3.3567460410357075,
      "tokens_seen": 2612723712
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.758133156799697e-05,
      "loss": 2.9033,
      "theoretical_loss": 3.356739384074442,
      "tokens_seen": 2612789248
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.75647815396255e-05,
      "loss": 2.3598,
      "theoretical_loss": 3.3567327273269014,
      "tokens_seen": 2612854784
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.754823151125402e-05,
      "loss": 1.7529,
      "theoretical_loss": 3.356726070793073,
      "tokens_seen": 2612920320
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7531681482882545e-05,
      "loss": 2.4297,
      "theoretical_loss": 3.356719414472945,
      "tokens_seen": 2612985856
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.751513145451107e-05,
      "loss": 1.5112,
      "theoretical_loss": 3.3567127583665046,
      "tokens_seen": 2613051392
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.749858142613959e-05,
      "loss": 2.4947,
      "theoretical_loss": 3.3567061024737406,
      "tokens_seen": 2613116928
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7482031397768104e-05,
      "loss": 2.0585,
      "theoretical_loss": 3.3566994467946394,
      "tokens_seen": 2613182464
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.746548136939663e-05,
      "loss": 2.6737,
      "theoretical_loss": 3.35669279132919,
      "tokens_seen": 2613248000
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.744893134102515e-05,
      "loss": 1.6775,
      "theoretical_loss": 3.3566861360773794,
      "tokens_seen": 2613313536
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7432381312653676e-05,
      "loss": 2.4966,
      "theoretical_loss": 3.356679481039196,
      "tokens_seen": 2613379072
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.74158312842822e-05,
      "loss": 2.2284,
      "theoretical_loss": 3.356672826214627,
      "tokens_seen": 2613444608
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7399281255910724e-05,
      "loss": 2.8309,
      "theoretical_loss": 3.3566661716036608,
      "tokens_seen": 2613510144
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.738273122753925e-05,
      "loss": 2.1369,
      "theoretical_loss": 3.3566595172062845,
      "tokens_seen": 2613575680
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7366181199167765e-05,
      "loss": 2.1469,
      "theoretical_loss": 3.3566528630224863,
      "tokens_seen": 2613641216
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.734963117079629e-05,
      "loss": 2.5213,
      "theoretical_loss": 3.356646209052254,
      "tokens_seen": 2613706752
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.733308114242481e-05,
      "loss": 2.2054,
      "theoretical_loss": 3.3566395552955752,
      "tokens_seen": 2613772288
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.731653111405334e-05,
      "loss": 1.8628,
      "theoretical_loss": 3.356632901752438,
      "tokens_seen": 2613837824
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.729998108568186e-05,
      "loss": 2.3753,
      "theoretical_loss": 3.3566262484228298,
      "tokens_seen": 2613903360
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7283431057310385e-05,
      "loss": 2.4611,
      "theoretical_loss": 3.356619595306739,
      "tokens_seen": 2613968896
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.726688102893891e-05,
      "loss": 1.9605,
      "theoretical_loss": 3.356612942404152,
      "tokens_seen": 2614034432
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7250331000567426e-05,
      "loss": 2.5483,
      "theoretical_loss": 3.3566062897150584,
      "tokens_seen": 2614099968
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.723378097219595e-05,
      "loss": 2.3331,
      "theoretical_loss": 3.3565996372394453,
      "tokens_seen": 2614165504
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7217230943824474e-05,
      "loss": 2.1795,
      "theoretical_loss": 3.3565929849772997,
      "tokens_seen": 2614231040
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7200680915453e-05,
      "loss": 4.2673,
      "theoretical_loss": 3.356586332928611,
      "tokens_seen": 2614296576
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.718413088708152e-05,
      "loss": 1.4122,
      "theoretical_loss": 3.3565796810933652,
      "tokens_seen": 2614362112
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7167580858710046e-05,
      "loss": 2.7687,
      "theoretical_loss": 3.3565730294715515,
      "tokens_seen": 2614427648
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.715103083033857e-05,
      "loss": 1.7156,
      "theoretical_loss": 3.356566378063157,
      "tokens_seen": 2614493184
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7134480801967094e-05,
      "loss": 2.0224,
      "theoretical_loss": 3.35655972686817,
      "tokens_seen": 2614558720
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7117930773595604e-05,
      "loss": 1.9105,
      "theoretical_loss": 3.3565530758865774,
      "tokens_seen": 2614624256
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.710138074522413e-05,
      "loss": 2.2523,
      "theoretical_loss": 3.356546425118368,
      "tokens_seen": 2614689792
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.708483071685265e-05,
      "loss": 2.4562,
      "theoretical_loss": 3.356539774563529,
      "tokens_seen": 2614755328
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7068280688481176e-05,
      "loss": 2.6039,
      "theoretical_loss": 3.3565331242220484,
      "tokens_seen": 2614820864
    },
    {
      "epoch": 0.93,
      "objective/train/avg_token_score": 0.008781389333307743,
      "objective/train/avg_weight": 0.9896641373634338,
      "objective/train/docs_used": 1470949,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.3495616912841797,
      "objective/train/std_weight": 0.048572007566690445,
      "objective/train/theoretical_loss": 3.356526474093914,
      "objective/train/tokens_used": 2635346400,
      "theoretical_loss": 3.356526474093914,
      "tokens_seen": 2614886400
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.705173066010971e-05,
      "loss": 2.5291,
      "theoretical_loss": 3.356526474093914,
      "tokens_seen": 2614886400
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.703518063173823e-05,
      "loss": 4.3744,
      "theoretical_loss": 3.356519824179114,
      "tokens_seen": 2614951936
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7018630603366755e-05,
      "loss": 2.5258,
      "theoretical_loss": 3.3565131744776355,
      "tokens_seen": 2615017472
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.7002080574995266e-05,
      "loss": 2.1658,
      "theoretical_loss": 3.3565065249894666,
      "tokens_seen": 2615083008
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.698553054662379e-05,
      "loss": 2.4993,
      "theoretical_loss": 3.3564998757145954,
      "tokens_seen": 2615148544
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6968980518252314e-05,
      "loss": 2.5467,
      "theoretical_loss": 3.356493226653009,
      "tokens_seen": 2615214080
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.695243048988084e-05,
      "loss": 1.6414,
      "theoretical_loss": 3.356486577804696,
      "tokens_seen": 2615279616
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.693588046150936e-05,
      "loss": 2.4612,
      "theoretical_loss": 3.356479929169644,
      "tokens_seen": 2615345152
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6919330433137886e-05,
      "loss": 1.9228,
      "theoretical_loss": 3.356473280747841,
      "tokens_seen": 2615410688
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.690278040476641e-05,
      "loss": 2.548,
      "theoretical_loss": 3.3564666325392736,
      "tokens_seen": 2615476224
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6886230376394934e-05,
      "loss": 2.382,
      "theoretical_loss": 3.3564599845439314,
      "tokens_seen": 2615541760
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.686968034802345e-05,
      "loss": 2.2627,
      "theoretical_loss": 3.356453336761801,
      "tokens_seen": 2615607296
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6853130319651975e-05,
      "loss": 2.4773,
      "theoretical_loss": 3.356446689192871,
      "tokens_seen": 2615672832
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.68365802912805e-05,
      "loss": 2.3442,
      "theoretical_loss": 3.356440041837128,
      "tokens_seen": 2615738368
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.682003026290902e-05,
      "loss": 2.0334,
      "theoretical_loss": 3.356433394694561,
      "tokens_seen": 2615803904
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.680348023453755e-05,
      "loss": 2.1517,
      "theoretical_loss": 3.3564267477651577,
      "tokens_seen": 2615869440
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.678693020616607e-05,
      "loss": 2.0011,
      "theoretical_loss": 3.356420101048905,
      "tokens_seen": 2615934976
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6770380177794595e-05,
      "loss": 2.6478,
      "theoretical_loss": 3.3564134545457924,
      "tokens_seen": 2616000512
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6753830149423105e-05,
      "loss": 2.7273,
      "theoretical_loss": 3.356406808255806,
      "tokens_seen": 2616066048
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6737280121051636e-05,
      "loss": 2.3186,
      "theoretical_loss": 3.356400162178935,
      "tokens_seen": 2616131584
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.672073009268016e-05,
      "loss": 2.2565,
      "theoretical_loss": 3.356393516315166,
      "tokens_seen": 2616197120
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6704180064308684e-05,
      "loss": 2.3328,
      "theoretical_loss": 3.3563868706644877,
      "tokens_seen": 2616262656
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.668763003593721e-05,
      "loss": 2.3938,
      "theoretical_loss": 3.3563802252268875,
      "tokens_seen": 2616328192
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.667108000756573e-05,
      "loss": 2.5732,
      "theoretical_loss": 3.3563735800023533,
      "tokens_seen": 2616393728
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6654529979194256e-05,
      "loss": 2.86,
      "theoretical_loss": 3.356366934990873,
      "tokens_seen": 2616459264
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6637979950822766e-05,
      "loss": 1.9304,
      "theoretical_loss": 3.3563602901924345,
      "tokens_seen": 2616524800
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.662142992245129e-05,
      "loss": 2.4776,
      "theoretical_loss": 3.3563536456070255,
      "tokens_seen": 2616590336
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6604879894079814e-05,
      "loss": 2.2953,
      "theoretical_loss": 3.356347001234634,
      "tokens_seen": 2616655872
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.658832986570834e-05,
      "loss": 2.8149,
      "theoretical_loss": 3.356340357075248,
      "tokens_seen": 2616721408
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.657177983733686e-05,
      "loss": 2.2042,
      "theoretical_loss": 3.356333713128855,
      "tokens_seen": 2616786944
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6555229808965386e-05,
      "loss": 2.4427,
      "theoretical_loss": 3.3563270693954426,
      "tokens_seen": 2616852480
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.653867978059391e-05,
      "loss": 3.5759,
      "theoretical_loss": 3.356320425874999,
      "tokens_seen": 2616918016
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6522129752222434e-05,
      "loss": 2.1468,
      "theoretical_loss": 3.3563137825675122,
      "tokens_seen": 2616983552
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.650557972385095e-05,
      "loss": 2.6907,
      "theoretical_loss": 3.3563071394729698,
      "tokens_seen": 2617049088
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6489029695479475e-05,
      "loss": 2.3376,
      "theoretical_loss": 3.3563004965913597,
      "tokens_seen": 2617114624
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6472479667108e-05,
      "loss": 1.9416,
      "theoretical_loss": 3.3562938539226694,
      "tokens_seen": 2617180160
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6455929638736523e-05,
      "loss": 2.5799,
      "theoretical_loss": 3.3562872114668876,
      "tokens_seen": 2617245696
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.643937961036505e-05,
      "loss": 2.5255,
      "theoretical_loss": 3.356280569224001,
      "tokens_seen": 2617311232
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.642282958199357e-05,
      "loss": 2.8987,
      "theoretical_loss": 3.3562739271939988,
      "tokens_seen": 2617376768
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6406279553622095e-05,
      "loss": 1.7687,
      "theoretical_loss": 3.356267285376868,
      "tokens_seen": 2617442304
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.638972952525061e-05,
      "loss": 2.5493,
      "theoretical_loss": 3.356260643772596,
      "tokens_seen": 2617507840
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.637317949687914e-05,
      "loss": 2.5381,
      "theoretical_loss": 3.3562540023811716,
      "tokens_seen": 2617573376
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.635662946850766e-05,
      "loss": 2.2632,
      "theoretical_loss": 3.356247361202582,
      "tokens_seen": 2617638912
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6340079440136185e-05,
      "loss": 2.6291,
      "theoretical_loss": 3.356240720236815,
      "tokens_seen": 2617704448
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.632352941176471e-05,
      "loss": 2.2581,
      "theoretical_loss": 3.356234079483859,
      "tokens_seen": 2617769984
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.630697938339323e-05,
      "loss": 2.7253,
      "theoretical_loss": 3.3562274389437023,
      "tokens_seen": 2617835520
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.6290429355021757e-05,
      "loss": 2.3943,
      "theoretical_loss": 3.3562207986163313,
      "tokens_seen": 2617901056
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.627387932665028e-05,
      "loss": 2.4577,
      "theoretical_loss": 3.356214158501735,
      "tokens_seen": 2617966592
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.625732929827879e-05,
      "loss": 2.7047,
      "theoretical_loss": 3.3562075185999007,
      "tokens_seen": 2618032128
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.6240779269907315e-05,
      "loss": 2.62,
      "theoretical_loss": 3.3562008789108164,
      "tokens_seen": 2618097664
    },
    {
      "epoch": 0.94,
      "objective/train/avg_token_score": 0.01849929429590702,
      "objective/train/avg_weight": 0.9782521724700928,
      "objective/train/docs_used": 1472903,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 6.832622528076172,
      "objective/train/std_weight": 0.0914958044886589,
      "objective/train/theoretical_loss": 3.35619423943447,
      "objective/train/tokens_used": 2638623200,
      "theoretical_loss": 3.35619423943447,
      "tokens_seen": 2618163200
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.622422924153584e-05,
      "loss": 2.152,
      "theoretical_loss": 3.35619423943447,
      "tokens_seen": 2618163200
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.620767921316436e-05,
      "loss": 1.6246,
      "theoretical_loss": 3.3561876001708493,
      "tokens_seen": 2618228736
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.619112918479289e-05,
      "loss": 2.2837,
      "theoretical_loss": 3.3561809611199425,
      "tokens_seen": 2618294272
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.617457915642141e-05,
      "loss": 2.3727,
      "theoretical_loss": 3.356174322281737,
      "tokens_seen": 2618359808
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.6158029128049935e-05,
      "loss": 2.5439,
      "theoretical_loss": 3.3561676836562206,
      "tokens_seen": 2618425344
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.614147909967845e-05,
      "loss": 1.9543,
      "theoretical_loss": 3.3561610452433817,
      "tokens_seen": 2618490880
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.6124929071306976e-05,
      "loss": 2.5709,
      "theoretical_loss": 3.3561544070432077,
      "tokens_seen": 2618556416
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.61083790429355e-05,
      "loss": 2.1384,
      "theoretical_loss": 3.3561477690556867,
      "tokens_seen": 2618621952
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.6091829014564024e-05,
      "loss": 2.4108,
      "theoretical_loss": 3.3561411312808067,
      "tokens_seen": 2618687488
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.607527898619255e-05,
      "loss": 1.1755,
      "theoretical_loss": 3.356134493718555,
      "tokens_seen": 2618753024
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.605872895782107e-05,
      "loss": 1.8318,
      "theoretical_loss": 3.3561278563689196,
      "tokens_seen": 2618818560
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.6042178929449596e-05,
      "loss": 2.6896,
      "theoretical_loss": 3.356121219231889,
      "tokens_seen": 2618884096
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.602562890107812e-05,
      "loss": 2.4335,
      "theoretical_loss": 3.3561145823074505,
      "tokens_seen": 2618949632
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.600907887270664e-05,
      "loss": 2.696,
      "theoretical_loss": 3.3561079455955922,
      "tokens_seen": 2619015168
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.599252884433516e-05,
      "loss": 2.7695,
      "theoretical_loss": 3.356101309096302,
      "tokens_seen": 2619080704
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5975978815963685e-05,
      "loss": 2.834,
      "theoretical_loss": 3.3560946728095677,
      "tokens_seen": 2619146240
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.595942878759221e-05,
      "loss": 2.1285,
      "theoretical_loss": 3.356088036735377,
      "tokens_seen": 2619211776
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.594287875922073e-05,
      "loss": 2.4071,
      "theoretical_loss": 3.356081400873718,
      "tokens_seen": 2619277312
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.592632873084926e-05,
      "loss": 2.3635,
      "theoretical_loss": 3.3560747652245784,
      "tokens_seen": 2619342848
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.590977870247778e-05,
      "loss": 2.2966,
      "theoretical_loss": 3.356068129787946,
      "tokens_seen": 2619408384
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.589322867410629e-05,
      "loss": 2.8009,
      "theoretical_loss": 3.3560614945638094,
      "tokens_seen": 2619473920
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5876678645734816e-05,
      "loss": 2.7981,
      "theoretical_loss": 3.3560548595521555,
      "tokens_seen": 2619539456
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.586012861736334e-05,
      "loss": 2.2474,
      "theoretical_loss": 3.356048224752973,
      "tokens_seen": 2619604992
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5843578588991864e-05,
      "loss": 2.4655,
      "theoretical_loss": 3.356041590166249,
      "tokens_seen": 2619670528
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.582702856062039e-05,
      "loss": 2.8043,
      "theoretical_loss": 3.356034955791972,
      "tokens_seen": 2619736064
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.581047853224891e-05,
      "loss": 3.0383,
      "theoretical_loss": 3.3560283216301294,
      "tokens_seen": 2619801600
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5793928503877436e-05,
      "loss": 2.462,
      "theoretical_loss": 3.35602168768071,
      "tokens_seen": 2619867136
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.577737847550595e-05,
      "loss": 2.737,
      "theoretical_loss": 3.3560150539437004,
      "tokens_seen": 2619932672
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.576082844713448e-05,
      "loss": 3.2211,
      "theoretical_loss": 3.356008420419089,
      "tokens_seen": 2619998208
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5744278418763e-05,
      "loss": 1.6324,
      "theoretical_loss": 3.356001787106864,
      "tokens_seen": 2620063744
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5727728390391525e-05,
      "loss": 2.7735,
      "theoretical_loss": 3.3559951540070134,
      "tokens_seen": 2620129280
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.571117836202005e-05,
      "loss": 2.7353,
      "theoretical_loss": 3.3559885211195244,
      "tokens_seen": 2620194816
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.569462833364857e-05,
      "loss": 4.5849,
      "theoretical_loss": 3.3559818884443855,
      "tokens_seen": 2620260352
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.56780783052771e-05,
      "loss": 0.3913,
      "theoretical_loss": 3.3559752559815843,
      "tokens_seen": 2620325888
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.566152827690562e-05,
      "loss": 2.4068,
      "theoretical_loss": 3.355968623731109,
      "tokens_seen": 2620391424
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.564497824853414e-05,
      "loss": -0.6856,
      "theoretical_loss": 3.3559619916929466,
      "tokens_seen": 2620456960
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.562842822016266e-05,
      "loss": -0.0423,
      "theoretical_loss": 3.355955359867086,
      "tokens_seen": 2620522496
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5611878191791186e-05,
      "loss": 1.1534,
      "theoretical_loss": 3.355948728253515,
      "tokens_seen": 2620588032
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.559532816341971e-05,
      "loss": 2.2719,
      "theoretical_loss": 3.355942096852221,
      "tokens_seen": 2620653568
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5578778135048234e-05,
      "loss": 2.541,
      "theoretical_loss": 3.355935465663192,
      "tokens_seen": 2620719104
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.556222810667676e-05,
      "loss": 1.4417,
      "theoretical_loss": 3.355928834686416,
      "tokens_seen": 2620784640
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.554567807830528e-05,
      "loss": -0.3547,
      "theoretical_loss": 3.3559222039218812,
      "tokens_seen": 2620850176
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.552912804993379e-05,
      "loss": 2.7892,
      "theoretical_loss": 3.355915573369575,
      "tokens_seen": 2620915712
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5512578021562316e-05,
      "loss": 2.6078,
      "theoretical_loss": 3.3559089430294855,
      "tokens_seen": 2620981248
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.549602799319084e-05,
      "loss": 1.2804,
      "theoretical_loss": 3.3559023129016006,
      "tokens_seen": 2621046784
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5479477964819364e-05,
      "loss": 2.0012,
      "theoretical_loss": 3.355895682985908,
      "tokens_seen": 2621112320
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.546292793644789e-05,
      "loss": 0.9057,
      "theoretical_loss": 3.3558890532823966,
      "tokens_seen": 2621177856
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.544637790807641e-05,
      "loss": 3.0426,
      "theoretical_loss": 3.355882423791053,
      "tokens_seen": 2621243392
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5429827879704936e-05,
      "loss": 2.5547,
      "theoretical_loss": 3.355875794511866,
      "tokens_seen": 2621308928
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.541327785133346e-05,
      "loss": 2.9662,
      "theoretical_loss": 3.3558691654448225,
      "tokens_seen": 2621374464
    },
    {
      "debugging/Self-BLEU-5": 0.449855913696806,
      "debugging/distinct-1-grams": 0.769778305351557,
      "debugging/distinct-2-grams": 0.9467712904639874,
      "debugging/entropy-1-grams": 5.998633443012235,
      "debugging/entropy-2-grams": 6.951908950814323,
      "debugging/length": 469.11764705882354,
      "debugging/num_segments": 17,
      "epoch": 0.94,
      "objective/train/avg_token_score": 0.039095163345336914,
      "objective/train/avg_weight": 0.9540200233459473,
      "objective/train/docs_used": 1474904,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 29.192790985107422,
      "objective/train/std_weight": 0.13041827082633972,
      "objective/train/theoretical_loss": 3.3558625365899117,
      "objective/train/tokens_used": 2641900000,
      "theoretical_loss": 3.3558625365899117,
      "tokens_seen": 2621440000
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.539672782296198e-05,
      "loss": 4.1209,
      "theoretical_loss": 3.3558625365899117,
      "tokens_seen": 2621440000
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.53801777945905e-05,
      "loss": 2.7267,
      "theoretical_loss": 3.3558559079471206,
      "tokens_seen": 2621505536
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5363627766219026e-05,
      "loss": 1.4517,
      "theoretical_loss": 3.355849279516437,
      "tokens_seen": 2621571072
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.534707773784755e-05,
      "loss": -3.4492,
      "theoretical_loss": 3.35584265129785,
      "tokens_seen": 2621636608
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5330527709476074e-05,
      "loss": 2.1261,
      "theoretical_loss": 3.355836023291346,
      "tokens_seen": 2621702144
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.53139776811046e-05,
      "loss": 2.5807,
      "theoretical_loss": 3.355829395496914,
      "tokens_seen": 2621767680
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.529742765273312e-05,
      "loss": -1.1383,
      "theoretical_loss": 3.355822767914541,
      "tokens_seen": 2621833216
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.528087762436164e-05,
      "loss": -0.0264,
      "theoretical_loss": 3.355816140544216,
      "tokens_seen": 2621898752
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.526432759599016e-05,
      "loss": 2.9852,
      "theoretical_loss": 3.3558095133859265,
      "tokens_seen": 2621964288
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.524777756761869e-05,
      "loss": -2.6141,
      "theoretical_loss": 3.3558028864396596,
      "tokens_seen": 2622029824
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.523122753924721e-05,
      "loss": 1.5247,
      "theoretical_loss": 3.3557962597054045,
      "tokens_seen": 2622095360
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5214677510875735e-05,
      "loss": 2.5364,
      "theoretical_loss": 3.3557896331831483,
      "tokens_seen": 2622160896
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.519812748250426e-05,
      "loss": 2.9055,
      "theoretical_loss": 3.355783006872879,
      "tokens_seen": 2622226432
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.518157745413278e-05,
      "loss": 2.4731,
      "theoretical_loss": 3.3557763807745844,
      "tokens_seen": 2622291968
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.516502742576131e-05,
      "loss": 2.7083,
      "theoretical_loss": 3.355769754888253,
      "tokens_seen": 2622357504
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.514847739738982e-05,
      "loss": 1.7918,
      "theoretical_loss": 3.3557631292138725,
      "tokens_seen": 2622423040
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.513192736901834e-05,
      "loss": 2.7146,
      "theoretical_loss": 3.3557565037514303,
      "tokens_seen": 2622488576
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5115377340646865e-05,
      "loss": 2.7948,
      "theoretical_loss": 3.355749878500915,
      "tokens_seen": 2622554112
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.509882731227539e-05,
      "loss": 1.9579,
      "theoretical_loss": 3.3557432534623146,
      "tokens_seen": 2622619648
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.508227728390391e-05,
      "loss": 2.4316,
      "theoretical_loss": 3.3557366286356163,
      "tokens_seen": 2622685184
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.506572725553244e-05,
      "loss": 2.1386,
      "theoretical_loss": 3.355730004020808,
      "tokens_seen": 2622750720
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.504917722716097e-05,
      "loss": 0.8654,
      "theoretical_loss": 3.3557233796178787,
      "tokens_seen": 2622816256
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.503262719878948e-05,
      "loss": 2.4489,
      "theoretical_loss": 3.3557167554268155,
      "tokens_seen": 2622881792
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.5016077170418e-05,
      "loss": 1.9785,
      "theoretical_loss": 3.3557101314476068,
      "tokens_seen": 2622947328
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4999527142046526e-05,
      "loss": 3.296,
      "theoretical_loss": 3.3557035076802397,
      "tokens_seen": 2623012864
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.498297711367505e-05,
      "loss": 1.8365,
      "theoretical_loss": 3.355696884124703,
      "tokens_seen": 2623078400
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4966427085303574e-05,
      "loss": 1.8205,
      "theoretical_loss": 3.355690260780984,
      "tokens_seen": 2623143936
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.49498770569321e-05,
      "loss": 2.6482,
      "theoretical_loss": 3.3556836376490713,
      "tokens_seen": 2623209472
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.493332702856062e-05,
      "loss": 1.5464,
      "theoretical_loss": 3.3556770147289523,
      "tokens_seen": 2623275008
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.491677700018914e-05,
      "loss": 2.3962,
      "theoretical_loss": 3.355670392020615,
      "tokens_seen": 2623340544
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4900226971817663e-05,
      "loss": 1.5357,
      "theoretical_loss": 3.3556637695240474,
      "tokens_seen": 2623406080
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.488367694344619e-05,
      "loss": 1.518,
      "theoretical_loss": 3.3556571472392376,
      "tokens_seen": 2623471616
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.486712691507471e-05,
      "loss": 2.2905,
      "theoretical_loss": 3.3556505251661735,
      "tokens_seen": 2623537152
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4850576886703235e-05,
      "loss": 1.9419,
      "theoretical_loss": 3.355643903304843,
      "tokens_seen": 2623602688
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.483402685833176e-05,
      "loss": 2.482,
      "theoretical_loss": 3.355637281655234,
      "tokens_seen": 2623668224
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4817476829960283e-05,
      "loss": 2.0695,
      "theoretical_loss": 3.3556306602173342,
      "tokens_seen": 2623733760
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.480092680158881e-05,
      "loss": 2.05,
      "theoretical_loss": 3.3556240389911323,
      "tokens_seen": 2623799296
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.478437677321732e-05,
      "loss": 2.544,
      "theoretical_loss": 3.355617417976615,
      "tokens_seen": 2623864832
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.476782674484584e-05,
      "loss": 3.0438,
      "theoretical_loss": 3.355610797173772,
      "tokens_seen": 2623930368
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4751276716474366e-05,
      "loss": 3.1843,
      "theoretical_loss": 3.355604176582589,
      "tokens_seen": 2623995904
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4734726688102897e-05,
      "loss": -1.7938,
      "theoretical_loss": 3.355597556203056,
      "tokens_seen": 2624061440
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.471817665973142e-05,
      "loss": 1.4029,
      "theoretical_loss": 3.35559093603516,
      "tokens_seen": 2624126976
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4701626631359945e-05,
      "loss": 2.6313,
      "theoretical_loss": 3.3555843160788887,
      "tokens_seen": 2624192512
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.468507660298847e-05,
      "loss": 2.5196,
      "theoretical_loss": 3.355577696334231,
      "tokens_seen": 2624258048
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.466852657461698e-05,
      "loss": 2.1057,
      "theoretical_loss": 3.3555710768011737,
      "tokens_seen": 2624323584
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.46519765462455e-05,
      "loss": -0.4216,
      "theoretical_loss": 3.355564457479706,
      "tokens_seen": 2624389120
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.463542651787403e-05,
      "loss": 2.3161,
      "theoretical_loss": 3.3555578383698146,
      "tokens_seen": 2624454656
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.461887648950255e-05,
      "loss": 2.7561,
      "theoretical_loss": 3.3555512194714883,
      "tokens_seen": 2624520192
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4602326461131075e-05,
      "loss": 4.3749,
      "theoretical_loss": 3.3555446007847145,
      "tokens_seen": 2624585728
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.45857764327596e-05,
      "loss": 2.6599,
      "theoretical_loss": 3.3555379823094817,
      "tokens_seen": 2624651264
    },
    {
      "epoch": 0.94,
      "objective/train/avg_token_score": 0.012923024594783783,
      "objective/train/avg_weight": 0.9847995042800903,
      "objective/train/docs_used": 1477123,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7103452682495117,
      "objective/train/std_weight": 0.05227181315422058,
      "objective/train/theoretical_loss": 3.3555313640457776,
      "objective/train/tokens_used": 2645176800,
      "theoretical_loss": 3.3555313640457776,
      "tokens_seen": 2624716800
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.456922640438812e-05,
      "loss": 2.3858,
      "theoretical_loss": 3.3555313640457776,
      "tokens_seen": 2624716800
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.455267637601665e-05,
      "loss": 1.0755,
      "theoretical_loss": 3.35552474599359,
      "tokens_seen": 2624782336
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4536126347645164e-05,
      "loss": 1.9138,
      "theoretical_loss": 3.355518128152907,
      "tokens_seen": 2624847872
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.451957631927369e-05,
      "loss": 4.3049,
      "theoretical_loss": 3.3555115105237165,
      "tokens_seen": 2624913408
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.450302629090221e-05,
      "loss": 4.1277,
      "theoretical_loss": 3.3555048931060067,
      "tokens_seen": 2624978944
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4486476262530736e-05,
      "loss": 0.9852,
      "theoretical_loss": 3.3554982758997656,
      "tokens_seen": 2625044480
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.446992623415926e-05,
      "loss": 1.6214,
      "theoretical_loss": 3.355491658904981,
      "tokens_seen": 2625110016
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4453376205787784e-05,
      "loss": 2.485,
      "theoretical_loss": 3.3554850421216402,
      "tokens_seen": 2625175552
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.443682617741631e-05,
      "loss": 2.4063,
      "theoretical_loss": 3.3554784255497325,
      "tokens_seen": 2625241088
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4420276149044825e-05,
      "loss": 2.12,
      "theoretical_loss": 3.355471809189245,
      "tokens_seen": 2625306624
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.440372612067335e-05,
      "loss": 1.8041,
      "theoretical_loss": 3.3554651930401658,
      "tokens_seen": 2625372160
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.438717609230187e-05,
      "loss": 0.5415,
      "theoretical_loss": 3.3554585771024827,
      "tokens_seen": 2625437696
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.43706260639304e-05,
      "loss": 2.2598,
      "theoretical_loss": 3.3554519613761835,
      "tokens_seen": 2625503232
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.435407603555892e-05,
      "loss": 2.0049,
      "theoretical_loss": 3.3554453458612574,
      "tokens_seen": 2625568768
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4337526007187445e-05,
      "loss": 2.4391,
      "theoretical_loss": 3.355438730557691,
      "tokens_seen": 2625634304
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.432097597881597e-05,
      "loss": 0.7455,
      "theoretical_loss": 3.355432115465473,
      "tokens_seen": 2625699840
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.430442595044449e-05,
      "loss": 2.095,
      "theoretical_loss": 3.355425500584591,
      "tokens_seen": 2625765376
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4287875922073004e-05,
      "loss": 2.5837,
      "theoretical_loss": 3.355418885915033,
      "tokens_seen": 2625830912
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.427132589370153e-05,
      "loss": 2.3424,
      "theoretical_loss": 3.355412271456787,
      "tokens_seen": 2625896448
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.425477586533005e-05,
      "loss": 0.4427,
      "theoretical_loss": 3.3554056572098414,
      "tokens_seen": 2625961984
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4238225836958576e-05,
      "loss": 2.3545,
      "theoretical_loss": 3.355399043174184,
      "tokens_seen": 2626027520
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.42216758085871e-05,
      "loss": 2.6239,
      "theoretical_loss": 3.355392429349802,
      "tokens_seen": 2626093056
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4205125780215624e-05,
      "loss": 2.6354,
      "theoretical_loss": 3.3553858157366845,
      "tokens_seen": 2626158592
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.418857575184415e-05,
      "loss": 2.4399,
      "theoretical_loss": 3.3553792023348192,
      "tokens_seen": 2626224128
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4172025723472665e-05,
      "loss": 2.7864,
      "theoretical_loss": 3.3553725891441935,
      "tokens_seen": 2626289664
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.415547569510119e-05,
      "loss": 2.8465,
      "theoretical_loss": 3.3553659761647956,
      "tokens_seen": 2626355200
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.413892566672971e-05,
      "loss": 1.0722,
      "theoretical_loss": 3.355359363396614,
      "tokens_seen": 2626420736
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.412237563835824e-05,
      "loss": 2.3016,
      "theoretical_loss": 3.355352750839636,
      "tokens_seen": 2626486272
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.410582560998676e-05,
      "loss": 1.9654,
      "theoretical_loss": 3.35534613849385,
      "tokens_seen": 2626551808
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4089275581615285e-05,
      "loss": 2.0589,
      "theoretical_loss": 3.355339526359244,
      "tokens_seen": 2626617344
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.407272555324381e-05,
      "loss": 2.3982,
      "theoretical_loss": 3.355332914435806,
      "tokens_seen": 2626682880
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4056175524872326e-05,
      "loss": 2.2553,
      "theoretical_loss": 3.3553263027235234,
      "tokens_seen": 2626748416
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.403962549650085e-05,
      "loss": 2.3369,
      "theoretical_loss": 3.355319691222385,
      "tokens_seen": 2626813952
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.4023075468129374e-05,
      "loss": 1.0906,
      "theoretical_loss": 3.3553130799323783,
      "tokens_seen": 2626879488
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.40065254397579e-05,
      "loss": 2.7793,
      "theoretical_loss": 3.3553064688534913,
      "tokens_seen": 2626945024
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.398997541138642e-05,
      "loss": 2.7167,
      "theoretical_loss": 3.3552998579857123,
      "tokens_seen": 2627010560
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3973425383014946e-05,
      "loss": 2.5025,
      "theoretical_loss": 3.3552932473290293,
      "tokens_seen": 2627076096
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.395687535464347e-05,
      "loss": 2.3809,
      "theoretical_loss": 3.35528663688343,
      "tokens_seen": 2627141632
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3940325326271994e-05,
      "loss": 2.6061,
      "theoretical_loss": 3.3552800266489022,
      "tokens_seen": 2627207168
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3923775297900504e-05,
      "loss": 2.5432,
      "theoretical_loss": 3.3552734166254345,
      "tokens_seen": 2627272704
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.390722526952903e-05,
      "loss": 1.1391,
      "theoretical_loss": 3.3552668068130145,
      "tokens_seen": 2627338240
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.389067524115755e-05,
      "loss": 4.4718,
      "theoretical_loss": 3.35526019721163,
      "tokens_seen": 2627403776
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3874125212786076e-05,
      "loss": 2.6975,
      "theoretical_loss": 3.3552535878212697,
      "tokens_seen": 2627469312
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.38575751844146e-05,
      "loss": 2.5596,
      "theoretical_loss": 3.355246978641921,
      "tokens_seen": 2627534848
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3841025156043124e-05,
      "loss": 1.5421,
      "theoretical_loss": 3.3552403696735724,
      "tokens_seen": 2627600384
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.382447512767165e-05,
      "loss": 1.288,
      "theoretical_loss": 3.355233760916211,
      "tokens_seen": 2627665920
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3807925099300166e-05,
      "loss": 2.2606,
      "theoretical_loss": 3.355227152369826,
      "tokens_seen": 2627731456
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.379137507092869e-05,
      "loss": 3.1648,
      "theoretical_loss": 3.3552205440344043,
      "tokens_seen": 2627796992
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3774825042557214e-05,
      "loss": 2.5437,
      "theoretical_loss": 3.355213935909935,
      "tokens_seen": 2627862528
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.375827501418574e-05,
      "loss": 2.5767,
      "theoretical_loss": 3.355207327996405,
      "tokens_seen": 2627928064
    },
    {
      "epoch": 0.94,
      "objective/train/avg_token_score": 0.014310333877801895,
      "objective/train/avg_weight": 0.9831557273864746,
      "objective/train/docs_used": 1478969,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.523054361343384,
      "objective/train/std_weight": 0.07213561236858368,
      "objective/train/theoretical_loss": 3.355200720293803,
      "objective/train/tokens_used": 2648453600,
      "theoretical_loss": 3.355200720293803,
      "tokens_seen": 2627993600
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.374172498581426e-05,
      "loss": 2.471,
      "theoretical_loss": 3.355200720293803,
      "tokens_seen": 2627993600
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3725174957442786e-05,
      "loss": 2.6254,
      "theoretical_loss": 3.3551941128021165,
      "tokens_seen": 2628059136
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.37086249290713e-05,
      "loss": 2.1214,
      "theoretical_loss": 3.355187505521334,
      "tokens_seen": 2628124672
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.369207490069983e-05,
      "loss": 2.6409,
      "theoretical_loss": 3.3551808984514437,
      "tokens_seen": 2628190208
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.367552487232835e-05,
      "loss": 1.8628,
      "theoretical_loss": 3.3551742915924327,
      "tokens_seen": 2628255744
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3658974843956875e-05,
      "loss": 2.5419,
      "theoretical_loss": 3.35516768494429,
      "tokens_seen": 2628321280
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.36424248155854e-05,
      "loss": 2.5307,
      "theoretical_loss": 3.3551610785070034,
      "tokens_seen": 2628386816
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.362587478721392e-05,
      "loss": 2.4414,
      "theoretical_loss": 3.35515447228056,
      "tokens_seen": 2628452352
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.360932475884245e-05,
      "loss": 2.129,
      "theoretical_loss": 3.3551478662649488,
      "tokens_seen": 2628517888
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3592774730470964e-05,
      "loss": 2.6753,
      "theoretical_loss": 3.3551412604601576,
      "tokens_seen": 2628583424
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.357622470209949e-05,
      "loss": 2.243,
      "theoretical_loss": 3.3551346548661742,
      "tokens_seen": 2628648960
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.355967467372801e-05,
      "loss": 1.3249,
      "theoretical_loss": 3.3551280494829867,
      "tokens_seen": 2628714496
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3543124645356536e-05,
      "loss": 2.517,
      "theoretical_loss": 3.3551214443105835,
      "tokens_seen": 2628780032
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.352657461698505e-05,
      "loss": 2.7921,
      "theoretical_loss": 3.3551148393489516,
      "tokens_seen": 2628845568
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.351002458861358e-05,
      "loss": 2.4141,
      "theoretical_loss": 3.35510823459808,
      "tokens_seen": 2628911104
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.34934745602421e-05,
      "loss": 2.4095,
      "theoretical_loss": 3.355101630057957,
      "tokens_seen": 2628976640
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3476924531870625e-05,
      "loss": 3.3854,
      "theoretical_loss": 3.3550950257285694,
      "tokens_seen": 2629042176
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.346037450349915e-05,
      "loss": 2.7303,
      "theoretical_loss": 3.355088421609906,
      "tokens_seen": 2629107712
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.344382447512767e-05,
      "loss": 2.1698,
      "theoretical_loss": 3.355081817701955,
      "tokens_seen": 2629173248
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.34272744467562e-05,
      "loss": 2.4327,
      "theoretical_loss": 3.355075214004704,
      "tokens_seen": 2629238784
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3410724418384714e-05,
      "loss": 2.522,
      "theoretical_loss": 3.3550686105181406,
      "tokens_seen": 2629304320
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.339417439001324e-05,
      "loss": 1.9,
      "theoretical_loss": 3.355062007242254,
      "tokens_seen": 2629369856
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.337762436164176e-05,
      "loss": 2.5193,
      "theoretical_loss": 3.3550554041770315,
      "tokens_seen": 2629435392
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3361074333270286e-05,
      "loss": 2.122,
      "theoretical_loss": 3.355048801322461,
      "tokens_seen": 2629500928
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3344524304898803e-05,
      "loss": 2.4782,
      "theoretical_loss": 3.3550421986785306,
      "tokens_seen": 2629566464
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.332797427652733e-05,
      "loss": 0.4863,
      "theoretical_loss": 3.355035596245229,
      "tokens_seen": 2629632000
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.331142424815585e-05,
      "loss": 2.5593,
      "theoretical_loss": 3.3550289940225433,
      "tokens_seen": 2629697536
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3294874219784375e-05,
      "loss": 1.9249,
      "theoretical_loss": 3.3550223920104623,
      "tokens_seen": 2629763072
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.32783241914129e-05,
      "loss": 2.2625,
      "theoretical_loss": 3.3550157902089737,
      "tokens_seen": 2629828608
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3261774163041423e-05,
      "loss": 1.5235,
      "theoretical_loss": 3.355009188618065,
      "tokens_seen": 2629894144
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.324522413466995e-05,
      "loss": 0.7383,
      "theoretical_loss": 3.3550025872377254,
      "tokens_seen": 2629959680
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.322867410629847e-05,
      "loss": 2.5444,
      "theoretical_loss": 3.3549959860679417,
      "tokens_seen": 2630025216
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.321212407792699e-05,
      "loss": 2.5922,
      "theoretical_loss": 3.354989385108703,
      "tokens_seen": 2630090752
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.319557404955551e-05,
      "loss": 1.3436,
      "theoretical_loss": 3.3549827843599966,
      "tokens_seen": 2630156288
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3179024021184037e-05,
      "loss": 2.3335,
      "theoretical_loss": 3.354976183821811,
      "tokens_seen": 2630221824
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3162473992812554e-05,
      "loss": 1.0828,
      "theoretical_loss": 3.354969583494134,
      "tokens_seen": 2630287360
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.314592396444108e-05,
      "loss": 2.8014,
      "theoretical_loss": 3.3549629833769536,
      "tokens_seen": 2630352896
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.31293739360696e-05,
      "loss": 2.466,
      "theoretical_loss": 3.354956383470258,
      "tokens_seen": 2630418432
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3112823907698126e-05,
      "loss": 2.4023,
      "theoretical_loss": 3.354949783774035,
      "tokens_seen": 2630483968
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.309627387932665e-05,
      "loss": 2.2571,
      "theoretical_loss": 3.354943184288273,
      "tokens_seen": 2630549504
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.3079723850955174e-05,
      "loss": 2.1311,
      "theoretical_loss": 3.3549365850129598,
      "tokens_seen": 2630615040
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.30631738225837e-05,
      "loss": 3.1273,
      "theoretical_loss": 3.3549299859480834,
      "tokens_seen": 2630680576
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.304662379421222e-05,
      "loss": 1.7153,
      "theoretical_loss": 3.354923387093632,
      "tokens_seen": 2630746112
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.303007376584074e-05,
      "loss": 2.2672,
      "theoretical_loss": 3.354916788449594,
      "tokens_seen": 2630811648
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.301352373746926e-05,
      "loss": 2.3361,
      "theoretical_loss": 3.354910190015957,
      "tokens_seen": 2630877184
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.299697370909779e-05,
      "loss": 1.4373,
      "theoretical_loss": 3.354903591792709,
      "tokens_seen": 2630942720
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2980423680726304e-05,
      "loss": 1.9223,
      "theoretical_loss": 3.3548969937798376,
      "tokens_seen": 2631008256
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.296387365235483e-05,
      "loss": 3.0818,
      "theoretical_loss": 3.354890395977332,
      "tokens_seen": 2631073792
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.294732362398335e-05,
      "loss": 1.1675,
      "theoretical_loss": 3.3548837983851794,
      "tokens_seen": 2631139328
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2930773595611876e-05,
      "loss": 2.0264,
      "theoretical_loss": 3.3548772010033687,
      "tokens_seen": 2631204864
    },
    {
      "epoch": 0.94,
      "objective/train/avg_token_score": 0.01801973395049572,
      "objective/train/avg_weight": 0.9788313508033752,
      "objective/train/docs_used": 1481036,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 8.575088500976562,
      "objective/train/std_weight": 0.10016300529241562,
      "objective/train/theoretical_loss": 3.3548706038318867,
      "objective/train/tokens_used": 2651730400,
      "theoretical_loss": 3.3548706038318867,
      "tokens_seen": 2631270400
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.29142235672404e-05,
      "loss": 2.7461,
      "theoretical_loss": 3.3548706038318867,
      "tokens_seen": 2631270400
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2897673538868924e-05,
      "loss": 1.2821,
      "theoretical_loss": 3.3548640068707223,
      "tokens_seen": 2631335936
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.288112351049745e-05,
      "loss": 1.9492,
      "theoretical_loss": 3.3548574101198634,
      "tokens_seen": 2631401472
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.286457348212597e-05,
      "loss": 1.6676,
      "theoretical_loss": 3.354850813579298,
      "tokens_seen": 2631467008
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.284802345375449e-05,
      "loss": 2.004,
      "theoretical_loss": 3.3548442172490147,
      "tokens_seen": 2631532544
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.283147342538301e-05,
      "loss": 2.5759,
      "theoretical_loss": 3.3548376211290005,
      "tokens_seen": 2631598080
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.281492339701154e-05,
      "loss": 1.1505,
      "theoretical_loss": 3.3548310252192444,
      "tokens_seen": 2631663616
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.279837336864006e-05,
      "loss": 2.5906,
      "theoretical_loss": 3.354824429519734,
      "tokens_seen": 2631729152
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.278182334026858e-05,
      "loss": -0.0991,
      "theoretical_loss": 3.354817834030457,
      "tokens_seen": 2631794688
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.27652733118971e-05,
      "loss": 2.2001,
      "theoretical_loss": 3.3548112387514024,
      "tokens_seen": 2631860224
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.274872328352563e-05,
      "loss": 1.0789,
      "theoretical_loss": 3.3548046436825576,
      "tokens_seen": 2631925760
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.273217325515415e-05,
      "loss": 2.5638,
      "theoretical_loss": 3.354798048823911,
      "tokens_seen": 2631991296
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2715623226782674e-05,
      "loss": 0.2377,
      "theoretical_loss": 3.35479145417545,
      "tokens_seen": 2632056832
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.26990731984112e-05,
      "loss": 1.9952,
      "theoretical_loss": 3.3547848597371637,
      "tokens_seen": 2632122368
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.268252317003972e-05,
      "loss": 1.5426,
      "theoretical_loss": 3.3547782655090397,
      "tokens_seen": 2632187904
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.266597314166824e-05,
      "loss": 2.304,
      "theoretical_loss": 3.354771671491066,
      "tokens_seen": 2632253440
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2649423113296764e-05,
      "loss": 0.5746,
      "theoretical_loss": 3.3547650776832305,
      "tokens_seen": 2632318976
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.263287308492529e-05,
      "loss": 1.3511,
      "theoretical_loss": 3.3547584840855214,
      "tokens_seen": 2632384512
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.261632305655381e-05,
      "loss": 2.5534,
      "theoretical_loss": 3.3547518906979272,
      "tokens_seen": 2632450048
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.259977302818233e-05,
      "loss": 2.7379,
      "theoretical_loss": 3.3547452975204353,
      "tokens_seen": 2632515584
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.258322299981086e-05,
      "loss": 1.9115,
      "theoretical_loss": 3.354738704553034,
      "tokens_seen": 2632581120
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2566672971439384e-05,
      "loss": 2.4927,
      "theoretical_loss": 3.3547321117957116,
      "tokens_seen": 2632646656
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.25501229430679e-05,
      "loss": 2.4431,
      "theoretical_loss": 3.354725519248456,
      "tokens_seen": 2632712192
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2533572914696425e-05,
      "loss": 2.1369,
      "theoretical_loss": 3.3547189269112554,
      "tokens_seen": 2632777728
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.251702288632495e-05,
      "loss": 2.3337,
      "theoretical_loss": 3.354712334784098,
      "tokens_seen": 2632843264
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.250047285795347e-05,
      "loss": 2.5779,
      "theoretical_loss": 3.354705742866971,
      "tokens_seen": 2632908800
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.248392282958199e-05,
      "loss": 2.7619,
      "theoretical_loss": 3.354699151159864,
      "tokens_seen": 2632974336
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2467372801210514e-05,
      "loss": 2.0527,
      "theoretical_loss": 3.3546925596627637,
      "tokens_seen": 2633039872
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.245082277283904e-05,
      "loss": 3.4063,
      "theoretical_loss": 3.3546859683756587,
      "tokens_seen": 2633105408
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.243427274446756e-05,
      "loss": 1.1189,
      "theoretical_loss": 3.3546793772985373,
      "tokens_seen": 2633170944
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2417722716096086e-05,
      "loss": 2.0098,
      "theoretical_loss": 3.3546727864313874,
      "tokens_seen": 2633236480
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.240117268772461e-05,
      "loss": 2.3284,
      "theoretical_loss": 3.354666195774197,
      "tokens_seen": 2633302016
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2384622659353134e-05,
      "loss": 3.0815,
      "theoretical_loss": 3.354659605326954,
      "tokens_seen": 2633367552
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.236807263098166e-05,
      "loss": 1.3603,
      "theoretical_loss": 3.354653015089647,
      "tokens_seen": 2633433088
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2351522602610175e-05,
      "loss": 2.3506,
      "theoretical_loss": 3.354646425062264,
      "tokens_seen": 2633498624
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.23349725742387e-05,
      "loss": 2.6888,
      "theoretical_loss": 3.3546398352447926,
      "tokens_seen": 2633564160
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.231842254586722e-05,
      "loss": 0.5924,
      "theoretical_loss": 3.3546332456372214,
      "tokens_seen": 2633629696
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.230187251749574e-05,
      "loss": 1.2714,
      "theoretical_loss": 3.3546266562395384,
      "tokens_seen": 2633695232
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2285322489124264e-05,
      "loss": 2.0131,
      "theoretical_loss": 3.3546200670517314,
      "tokens_seen": 2633760768
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.226877246075279e-05,
      "loss": 2.901,
      "theoretical_loss": 3.354613478073789,
      "tokens_seen": 2633826304
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.225222243238131e-05,
      "loss": 2.3333,
      "theoretical_loss": 3.354606889305699,
      "tokens_seen": 2633891840
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2235672404009836e-05,
      "loss": 1.4236,
      "theoretical_loss": 3.3546003007474487,
      "tokens_seen": 2633957376
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.221912237563836e-05,
      "loss": 2.3068,
      "theoretical_loss": 3.3545937123990273,
      "tokens_seen": 2634022912
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2202572347266884e-05,
      "loss": 1.3369,
      "theoretical_loss": 3.354587124260423,
      "tokens_seen": 2634088448
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.218602231889541e-05,
      "loss": 2.5525,
      "theoretical_loss": 3.354580536331623,
      "tokens_seen": 2634153984
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2169472290523926e-05,
      "loss": 2.5917,
      "theoretical_loss": 3.3545739486126163,
      "tokens_seen": 2634219520
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.215292226215245e-05,
      "loss": 2.3818,
      "theoretical_loss": 3.3545673611033906,
      "tokens_seen": 2634285056
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2136372233780974e-05,
      "loss": 2.193,
      "theoretical_loss": 3.3545607738039336,
      "tokens_seen": 2634350592
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.211982220540949e-05,
      "loss": 0.8564,
      "theoretical_loss": 3.3545541867142337,
      "tokens_seen": 2634416128
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2103272177038015e-05,
      "loss": 2.5655,
      "theoretical_loss": 3.3545475998342793,
      "tokens_seen": 2634481664
    },
    {
      "epoch": 0.94,
      "objective/train/avg_token_score": 0.017899416387081146,
      "objective/train/avg_weight": 0.9789294600486755,
      "objective/train/docs_used": 1482195,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 6.39625883102417,
      "objective/train/std_weight": 0.08549806475639343,
      "objective/train/theoretical_loss": 3.3545410131640585,
      "objective/train/tokens_used": 2655007200,
      "theoretical_loss": 3.3545410131640585,
      "tokens_seen": 2634547200
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.208672214866654e-05,
      "loss": 2.4696,
      "theoretical_loss": 3.3545410131640585,
      "tokens_seen": 2634547200
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.207017212029506e-05,
      "loss": 2.2285,
      "theoretical_loss": 3.354534426703559,
      "tokens_seen": 2634612736
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.205362209192359e-05,
      "loss": 1.6625,
      "theoretical_loss": 3.3545278404527687,
      "tokens_seen": 2634678272
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.203707206355211e-05,
      "loss": 2.7347,
      "theoretical_loss": 3.3545212544116767,
      "tokens_seen": 2634743808
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.2020522035180635e-05,
      "loss": 2.6335,
      "theoretical_loss": 3.3545146685802703,
      "tokens_seen": 2634809344
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.200397200680916e-05,
      "loss": 1.9176,
      "theoretical_loss": 3.354508082958538,
      "tokens_seen": 2634874880
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1987421978437676e-05,
      "loss": 1.9529,
      "theoretical_loss": 3.354501497546467,
      "tokens_seen": 2634940416
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.19708719500662e-05,
      "loss": 3.9322,
      "theoretical_loss": 3.354494912344047,
      "tokens_seen": 2635005952
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1954321921694724e-05,
      "loss": 2.0622,
      "theoretical_loss": 3.3544883273512647,
      "tokens_seen": 2635071488
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.193777189332325e-05,
      "loss": 1.725,
      "theoretical_loss": 3.354481742568109,
      "tokens_seen": 2635137024
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1921221864951765e-05,
      "loss": 2.5508,
      "theoretical_loss": 3.3544751579945675,
      "tokens_seen": 2635202560
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.190467183658029e-05,
      "loss": 1.3748,
      "theoretical_loss": 3.354468573630629,
      "tokens_seen": 2635268096
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.188812180820881e-05,
      "loss": 2.2389,
      "theoretical_loss": 3.354461989476281,
      "tokens_seen": 2635333632
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.187157177983734e-05,
      "loss": 2.4801,
      "theoretical_loss": 3.3544554055315117,
      "tokens_seen": 2635399168
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.185502175146586e-05,
      "loss": 2.5943,
      "theoretical_loss": 3.3544488217963093,
      "tokens_seen": 2635464704
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1838471723094385e-05,
      "loss": 2.336,
      "theoretical_loss": 3.3544422382706625,
      "tokens_seen": 2635530240
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.182192169472291e-05,
      "loss": 2.623,
      "theoretical_loss": 3.354435654954558,
      "tokens_seen": 2635595776
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1805371666351426e-05,
      "loss": 2.493,
      "theoretical_loss": 3.3544290718479854,
      "tokens_seen": 2635661312
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.178882163797995e-05,
      "loss": -1.1901,
      "theoretical_loss": 3.354422488950932,
      "tokens_seen": 2635726848
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1772271609608474e-05,
      "loss": 1.8389,
      "theoretical_loss": 3.3544159062633865,
      "tokens_seen": 2635792384
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1755721581237e-05,
      "loss": 1.5799,
      "theoretical_loss": 3.3544093237853363,
      "tokens_seen": 2635857920
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1739171552865515e-05,
      "loss": 0.352,
      "theoretical_loss": 3.35440274151677,
      "tokens_seen": 2635923456
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.172262152449404e-05,
      "loss": -1.1486,
      "theoretical_loss": 3.3543961594576754,
      "tokens_seen": 2635988992
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1706071496122563e-05,
      "loss": 1.2966,
      "theoretical_loss": 3.3543895776080412,
      "tokens_seen": 2636054528
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.168952146775109e-05,
      "loss": 1.7904,
      "theoretical_loss": 3.354382995967855,
      "tokens_seen": 2636120064
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.167297143937961e-05,
      "loss": 2.487,
      "theoretical_loss": 3.354376414537105,
      "tokens_seen": 2636185600
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1656421411008135e-05,
      "loss": 2.3119,
      "theoretical_loss": 3.354369833315779,
      "tokens_seen": 2636251136
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.163987138263666e-05,
      "loss": 1.138,
      "theoretical_loss": 3.3543632523038664,
      "tokens_seen": 2636316672
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.162332135426518e-05,
      "loss": 2.5674,
      "theoretical_loss": 3.354356671501354,
      "tokens_seen": 2636382208
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.16067713258937e-05,
      "loss": 2.5954,
      "theoretical_loss": 3.3543500909082304,
      "tokens_seen": 2636447744
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1590221297522225e-05,
      "loss": 2.51,
      "theoretical_loss": 3.354343510524484,
      "tokens_seen": 2636513280
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.157367126915075e-05,
      "loss": 2.4946,
      "theoretical_loss": 3.3543369303501027,
      "tokens_seen": 2636578816
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1557121240779266e-05,
      "loss": 1.5733,
      "theoretical_loss": 3.3543303503850748,
      "tokens_seen": 2636644352
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.154057121240779e-05,
      "loss": 2.3687,
      "theoretical_loss": 3.354323770629388,
      "tokens_seen": 2636709888
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1524021184036314e-05,
      "loss": 0.0334,
      "theoretical_loss": 3.3543171910830303,
      "tokens_seen": 2636775424
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.150747115566484e-05,
      "loss": 2.647,
      "theoretical_loss": 3.3543106117459907,
      "tokens_seen": 2636840960
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.149092112729336e-05,
      "loss": 2.4599,
      "theoretical_loss": 3.3543040326182565,
      "tokens_seen": 2636906496
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1474371098921886e-05,
      "loss": 2.36,
      "theoretical_loss": 3.3542974536998162,
      "tokens_seen": 2636972032
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.145782107055041e-05,
      "loss": -1.235,
      "theoretical_loss": 3.3542908749906584,
      "tokens_seen": 2637037568
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.144127104217893e-05,
      "loss": 2.5403,
      "theoretical_loss": 3.3542842964907704,
      "tokens_seen": 2637103104
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.142472101380745e-05,
      "loss": 2.7431,
      "theoretical_loss": 3.354277718200141,
      "tokens_seen": 2637168640
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1408170985435975e-05,
      "loss": 2.6368,
      "theoretical_loss": 3.354271140118758,
      "tokens_seen": 2637234176
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.13916209570645e-05,
      "loss": 2.4386,
      "theoretical_loss": 3.3542645622466094,
      "tokens_seen": 2637299712
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1375070928693016e-05,
      "loss": 2.6689,
      "theoretical_loss": 3.354257984583684,
      "tokens_seen": 2637365248
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.135852090032154e-05,
      "loss": 2.7721,
      "theoretical_loss": 3.354251407129969,
      "tokens_seen": 2637430784
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1341970871950064e-05,
      "loss": 2.2172,
      "theoretical_loss": 3.3542448298854532,
      "tokens_seen": 2637496320
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.132542084357859e-05,
      "loss": 2.638,
      "theoretical_loss": 3.354238252850125,
      "tokens_seen": 2637561856
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.130887081520711e-05,
      "loss": 1.1578,
      "theoretical_loss": 3.3542316760239714,
      "tokens_seen": 2637627392
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1292320786835636e-05,
      "loss": -0.3309,
      "theoretical_loss": 3.3542250994069818,
      "tokens_seen": 2637692928
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.127577075846416e-05,
      "loss": 2.5501,
      "theoretical_loss": 3.354218522999144,
      "tokens_seen": 2637758464
    },
    {
      "epoch": 0.94,
      "objective/train/avg_token_score": 0.015882346779108047,
      "objective/train/avg_weight": 0.9813157320022583,
      "objective/train/docs_used": 1484156,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6117963790893555,
      "objective/train/std_weight": 0.08538983762264252,
      "objective/train/theoretical_loss": 3.354211946800446,
      "objective/train/tokens_used": 2658284000,
      "theoretical_loss": 3.354211946800446,
      "tokens_seen": 2637824000
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.125922073009268e-05,
      "loss": 2.5896,
      "theoretical_loss": 3.354211946800446,
      "tokens_seen": 2637824000
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.12426707017212e-05,
      "loss": 2.4134,
      "theoretical_loss": 3.3542053708108757,
      "tokens_seen": 2637889536
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1226120673349725e-05,
      "loss": 1.4617,
      "theoretical_loss": 3.3541987950304213,
      "tokens_seen": 2637955072
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.120957064497825e-05,
      "loss": 1.4813,
      "theoretical_loss": 3.3541922194590716,
      "tokens_seen": 2638020608
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1193020616606767e-05,
      "loss": 1.4109,
      "theoretical_loss": 3.3541856440968143,
      "tokens_seen": 2638086144
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.117647058823529e-05,
      "loss": 2.2008,
      "theoretical_loss": 3.3541790689436377,
      "tokens_seen": 2638151680
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1159920559863815e-05,
      "loss": -1.7777,
      "theoretical_loss": 3.3541724939995294,
      "tokens_seen": 2638217216
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.114337053149234e-05,
      "loss": 1.5316,
      "theoretical_loss": 3.3541659192644784,
      "tokens_seen": 2638282752
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.112682050312086e-05,
      "loss": 3.1358,
      "theoretical_loss": 3.354159344738472,
      "tokens_seen": 2638348288
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1110270474749386e-05,
      "loss": 1.8652,
      "theoretical_loss": 3.354152770421499,
      "tokens_seen": 2638413824
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.109372044637791e-05,
      "loss": 2.2982,
      "theoretical_loss": 3.3541461963135477,
      "tokens_seen": 2638479360
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1077170418006434e-05,
      "loss": 0.1308,
      "theoretical_loss": 3.354139622414606,
      "tokens_seen": 2638544896
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.106062038963495e-05,
      "loss": 2.4753,
      "theoretical_loss": 3.3541330487246617,
      "tokens_seen": 2638610432
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1044070361263476e-05,
      "loss": 2.5948,
      "theoretical_loss": 3.354126475243703,
      "tokens_seen": 2638675968
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1027520332892e-05,
      "loss": 2.4466,
      "theoretical_loss": 3.354119901971719,
      "tokens_seen": 2638741504
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.101097030452052e-05,
      "loss": 1.3687,
      "theoretical_loss": 3.3541133289086966,
      "tokens_seen": 2638807040
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.099442027614904e-05,
      "loss": 1.9754,
      "theoretical_loss": 3.354106756054625,
      "tokens_seen": 2638872576
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0977870247777565e-05,
      "loss": 2.4807,
      "theoretical_loss": 3.354100183409492,
      "tokens_seen": 2638938112
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.096132021940609e-05,
      "loss": 1.6969,
      "theoretical_loss": 3.3540936109732855,
      "tokens_seen": 2639003648
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.094477019103461e-05,
      "loss": 3.0095,
      "theoretical_loss": 3.354087038745994,
      "tokens_seen": 2639069184
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.092822016266314e-05,
      "loss": 1.8938,
      "theoretical_loss": 3.3540804667276056,
      "tokens_seen": 2639134720
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.091167013429166e-05,
      "loss": 1.5377,
      "theoretical_loss": 3.354073894918108,
      "tokens_seen": 2639200256
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0895120105920185e-05,
      "loss": 2.5578,
      "theoretical_loss": 3.3540673233174902,
      "tokens_seen": 2639265792
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.08785700775487e-05,
      "loss": 2.4392,
      "theoretical_loss": 3.35406075192574,
      "tokens_seen": 2639331328
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0862020049177226e-05,
      "loss": 1.209,
      "theoretical_loss": 3.3540541807428452,
      "tokens_seen": 2639396864
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.084547002080575e-05,
      "loss": 1.9662,
      "theoretical_loss": 3.354047609768795,
      "tokens_seen": 2639462400
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.082891999243427e-05,
      "loss": 2.585,
      "theoretical_loss": 3.3540410390035764,
      "tokens_seen": 2639527936
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.081236996406279e-05,
      "loss": 2.3474,
      "theoretical_loss": 3.354034468447178,
      "tokens_seen": 2639593472
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0795819935691315e-05,
      "loss": 1.9083,
      "theoretical_loss": 3.354027898099588,
      "tokens_seen": 2639659008
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.077926990731984e-05,
      "loss": 2.6975,
      "theoretical_loss": 3.354021327960795,
      "tokens_seen": 2639724544
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.076271987894836e-05,
      "loss": 2.5879,
      "theoretical_loss": 3.354014758030787,
      "tokens_seen": 2639790080
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.074616985057689e-05,
      "loss": 2.4704,
      "theoretical_loss": 3.3540081883095514,
      "tokens_seen": 2639855616
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.072961982220541e-05,
      "loss": 2.1297,
      "theoretical_loss": 3.3540016187970774,
      "tokens_seen": 2639921152
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0713069793833935e-05,
      "loss": 2.7442,
      "theoretical_loss": 3.353995049493353,
      "tokens_seen": 2639986688
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.069651976546245e-05,
      "loss": 2.2651,
      "theoretical_loss": 3.3539884803983657,
      "tokens_seen": 2640052224
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0679969737090976e-05,
      "loss": 3.6359,
      "theoretical_loss": 3.353981911512104,
      "tokens_seen": 2640117760
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.06634197087195e-05,
      "loss": 2.0724,
      "theoretical_loss": 3.353975342834557,
      "tokens_seen": 2640183296
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.064686968034802e-05,
      "loss": 2.5718,
      "theoretical_loss": 3.353968774365711,
      "tokens_seen": 2640248832
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.063031965197654e-05,
      "loss": 2.5155,
      "theoretical_loss": 3.3539622061055563,
      "tokens_seen": 2640314368
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0613769623605066e-05,
      "loss": 2.519,
      "theoretical_loss": 3.35395563805408,
      "tokens_seen": 2640379904
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.059721959523359e-05,
      "loss": 1.603,
      "theoretical_loss": 3.3539490702112698,
      "tokens_seen": 2640445440
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0580669566862114e-05,
      "loss": 2.8479,
      "theoretical_loss": 3.353942502577115,
      "tokens_seen": 2640510976
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.056411953849064e-05,
      "loss": 2.4094,
      "theoretical_loss": 3.353935935151603,
      "tokens_seen": 2640576512
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.054756951011916e-05,
      "loss": 2.6974,
      "theoretical_loss": 3.3539293679347226,
      "tokens_seen": 2640642048
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0531019481747686e-05,
      "loss": -0.2345,
      "theoretical_loss": 3.353922800926461,
      "tokens_seen": 2640707584
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.05144694533762e-05,
      "loss": 2.6081,
      "theoretical_loss": 3.3539162341268076,
      "tokens_seen": 2640773120
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.049791942500473e-05,
      "loss": 2.5815,
      "theoretical_loss": 3.35390966753575,
      "tokens_seen": 2640838656
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.048136939663325e-05,
      "loss": 2.6808,
      "theoretical_loss": 3.3539031011532767,
      "tokens_seen": 2640904192
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0464819368261775e-05,
      "loss": 2.4278,
      "theoretical_loss": 3.3538965349793752,
      "tokens_seen": 2640969728
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.044826933989029e-05,
      "loss": 2.4464,
      "theoretical_loss": 3.3538899690140345,
      "tokens_seen": 2641035264
    },
    {
      "epoch": 0.94,
      "objective/train/avg_token_score": 0.011314010247588158,
      "objective/train/avg_weight": 0.986688494682312,
      "objective/train/docs_used": 1485777,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.53775691986084,
      "objective/train/std_weight": 0.0565415620803833,
      "objective/train/theoretical_loss": 3.3538834032572424,
      "objective/train/tokens_used": 2661560800,
      "theoretical_loss": 3.3538834032572424,
      "tokens_seen": 2641100800
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.043171931151882e-05,
      "loss": 2.5147,
      "theoretical_loss": 3.3538834032572424,
      "tokens_seen": 2641100800
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.041516928314735e-05,
      "loss": 2.0708,
      "theoretical_loss": 3.3538768377089867,
      "tokens_seen": 2641166336
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0398619254775864e-05,
      "loss": 2.4017,
      "theoretical_loss": 3.3538702723692566,
      "tokens_seen": 2641231872
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.038206922640439e-05,
      "loss": 2.396,
      "theoretical_loss": 3.3538637072380393,
      "tokens_seen": 2641297408
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.036551919803291e-05,
      "loss": 3.724,
      "theoretical_loss": 3.3538571423153236,
      "tokens_seen": 2641362944
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0348969169661436e-05,
      "loss": 2.6,
      "theoretical_loss": 3.353850577601098,
      "tokens_seen": 2641428480
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.033241914128995e-05,
      "loss": 3.1104,
      "theoretical_loss": 3.3538440130953497,
      "tokens_seen": 2641494016
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.031586911291848e-05,
      "loss": 2.6186,
      "theoretical_loss": 3.353837448798068,
      "tokens_seen": 2641559552
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0299319084547e-05,
      "loss": 2.4786,
      "theoretical_loss": 3.3538308847092404,
      "tokens_seen": 2641625088
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0282769056175525e-05,
      "loss": 3.0271,
      "theoretical_loss": 3.353824320828855,
      "tokens_seen": 2641690624
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.026621902780405e-05,
      "loss": 1.9634,
      "theoretical_loss": 3.3538177571569006,
      "tokens_seen": 2641756160
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.024966899943257e-05,
      "loss": 2.0633,
      "theoretical_loss": 3.3538111936933652,
      "tokens_seen": 2641821696
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.02331189710611e-05,
      "loss": 2.4829,
      "theoretical_loss": 3.3538046304382365,
      "tokens_seen": 2641887232
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0216568942689614e-05,
      "loss": 2.6936,
      "theoretical_loss": 3.3537980673915033,
      "tokens_seen": 2641952768
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.020001891431814e-05,
      "loss": 2.3506,
      "theoretical_loss": 3.353791504553154,
      "tokens_seen": 2642018304
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.018346888594666e-05,
      "loss": 0.6227,
      "theoretical_loss": 3.3537849419231764,
      "tokens_seen": 2642083840
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0166918857575186e-05,
      "loss": 2.5398,
      "theoretical_loss": 3.3537783795015588,
      "tokens_seen": 2642149376
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0150368829203703e-05,
      "loss": 1.3968,
      "theoretical_loss": 3.353771817288289,
      "tokens_seen": 2642214912
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.013381880083223e-05,
      "loss": 2.9119,
      "theoretical_loss": 3.3537652552833563,
      "tokens_seen": 2642280448
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.011726877246075e-05,
      "loss": 2.237,
      "theoretical_loss": 3.3537586934867476,
      "tokens_seen": 2642345984
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0100718744089275e-05,
      "loss": 2.419,
      "theoretical_loss": 3.3537521318984522,
      "tokens_seen": 2642411520
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.00841687157178e-05,
      "loss": 2.2586,
      "theoretical_loss": 3.3537455705184573,
      "tokens_seen": 2642477056
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.0067618687346323e-05,
      "loss": 2.5888,
      "theoretical_loss": 3.3537390093467523,
      "tokens_seen": 2642542592
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.005106865897485e-05,
      "loss": 2.0536,
      "theoretical_loss": 3.3537324483833246,
      "tokens_seen": 2642608128
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.003451863060337e-05,
      "loss": 4.3823,
      "theoretical_loss": 3.353725887628163,
      "tokens_seen": 2642673664
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.001796860223189e-05,
      "loss": 2.7997,
      "theoretical_loss": 3.3537193270812544,
      "tokens_seen": 2642739200
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.000141857386041e-05,
      "loss": 2.1417,
      "theoretical_loss": 3.3537127667425892,
      "tokens_seen": 2642804736
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9984868545488937e-05,
      "loss": 2.7759,
      "theoretical_loss": 3.353706206612154,
      "tokens_seen": 2642870272
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9968318517117454e-05,
      "loss": 2.538,
      "theoretical_loss": 3.353699646689937,
      "tokens_seen": 2642935808
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.995176848874598e-05,
      "loss": 2.3005,
      "theoretical_loss": 3.353693086975927,
      "tokens_seen": 2643001344
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.99352184603745e-05,
      "loss": 1.6338,
      "theoretical_loss": 3.3536865274701126,
      "tokens_seen": 2643066880
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9918668432003026e-05,
      "loss": 2.5376,
      "theoretical_loss": 3.3536799681724814,
      "tokens_seen": 2643132416
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.990211840363155e-05,
      "loss": 2.498,
      "theoretical_loss": 3.3536734090830214,
      "tokens_seen": 2643197952
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9885568375260074e-05,
      "loss": 3.1332,
      "theoretical_loss": 3.3536668502017215,
      "tokens_seen": 2643263488
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.98690183468886e-05,
      "loss": -0.958,
      "theoretical_loss": 3.35366029152857,
      "tokens_seen": 2643329024
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.985246831851712e-05,
      "loss": 2.3415,
      "theoretical_loss": 3.353653733063554,
      "tokens_seen": 2643394560
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.983591829014564e-05,
      "loss": 0.7147,
      "theoretical_loss": 3.353647174806663,
      "tokens_seen": 2643460096
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.981936826177416e-05,
      "loss": 2.6434,
      "theoretical_loss": 3.3536406167578843,
      "tokens_seen": 2643525632
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.980281823340269e-05,
      "loss": 2.7434,
      "theoretical_loss": 3.353634058917207,
      "tokens_seen": 2643591168
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9786268205031204e-05,
      "loss": 2.7643,
      "theoretical_loss": 3.353627501284619,
      "tokens_seen": 2643656704
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.976971817665973e-05,
      "loss": 2.6293,
      "theoretical_loss": 3.353620943860108,
      "tokens_seen": 2643722240
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.975316814828825e-05,
      "loss": 1.7158,
      "theoretical_loss": 3.353614386643663,
      "tokens_seen": 2643787776
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9736618119916776e-05,
      "loss": 2.6056,
      "theoretical_loss": 3.353607829635272,
      "tokens_seen": 2643853312
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.97200680915453e-05,
      "loss": 1.784,
      "theoretical_loss": 3.3536012728349234,
      "tokens_seen": 2643918848
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9703518063173824e-05,
      "loss": 2.0525,
      "theoretical_loss": 3.3535947162426045,
      "tokens_seen": 2643984384
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.968696803480235e-05,
      "loss": 2.8283,
      "theoretical_loss": 3.353588159858305,
      "tokens_seen": 2644049920
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.967041800643087e-05,
      "loss": 3.1234,
      "theoretical_loss": 3.353581603682012,
      "tokens_seen": 2644115456
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.965386797805939e-05,
      "loss": 0.7752,
      "theoretical_loss": 3.353575047713714,
      "tokens_seen": 2644180992
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.963731794968791e-05,
      "loss": -0.5898,
      "theoretical_loss": 3.3535684919534,
      "tokens_seen": 2644246528
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.962076792131644e-05,
      "loss": 2.6455,
      "theoretical_loss": 3.353561936401057,
      "tokens_seen": 2644312064
    },
    {
      "epoch": 0.94,
      "objective/train/avg_token_score": 0.018379341810941696,
      "objective/train/avg_weight": 0.9783840179443359,
      "objective/train/docs_used": 1488108,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 8.107200622558594,
      "objective/train/std_weight": 0.09923982620239258,
      "objective/train/theoretical_loss": 3.3535553810566743,
      "objective/train/tokens_used": 2664837600,
      "theoretical_loss": 3.3535553810566743,
      "tokens_seen": 2644377600
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.960421789294496e-05,
      "loss": 1.9713,
      "theoretical_loss": 3.3535553810566743,
      "tokens_seen": 2644377600
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.958766786457348e-05,
      "loss": 2.5496,
      "theoretical_loss": 3.3535488259202397,
      "tokens_seen": 2644443136
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9571117836202e-05,
      "loss": 2.4471,
      "theoretical_loss": 3.3535422709917415,
      "tokens_seen": 2644508672
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9554567807830526e-05,
      "loss": 2.6397,
      "theoretical_loss": 3.3535357162711676,
      "tokens_seen": 2644574208
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.953801777945905e-05,
      "loss": 0.2208,
      "theoretical_loss": 3.353529161758507,
      "tokens_seen": 2644639744
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9521467751087574e-05,
      "loss": 1.8587,
      "theoretical_loss": 3.3535226074537476,
      "tokens_seen": 2644705280
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.95049177227161e-05,
      "loss": 2.4575,
      "theoretical_loss": 3.353516053356877,
      "tokens_seen": 2644770816
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.948836769434462e-05,
      "loss": 2.6507,
      "theoretical_loss": 3.3535094994678847,
      "tokens_seen": 2644836352
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.947181766597314e-05,
      "loss": 2.4494,
      "theoretical_loss": 3.353502945786758,
      "tokens_seen": 2644901888
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9455267637601664e-05,
      "loss": 3.1409,
      "theoretical_loss": 3.353496392313486,
      "tokens_seen": 2644967424
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.943871760923019e-05,
      "loss": 2.3512,
      "theoretical_loss": 3.3534898390480556,
      "tokens_seen": 2645032960
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.942216758085871e-05,
      "loss": 2.586,
      "theoretical_loss": 3.353483285990457,
      "tokens_seen": 2645098496
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.940561755248723e-05,
      "loss": 2.7619,
      "theoretical_loss": 3.3534767331406763,
      "tokens_seen": 2645164032
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.938906752411575e-05,
      "loss": 2.3541,
      "theoretical_loss": 3.3534701804987033,
      "tokens_seen": 2645229568
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.937251749574428e-05,
      "loss": 2.1155,
      "theoretical_loss": 3.3534636280645254,
      "tokens_seen": 2645295104
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.93559674673728e-05,
      "loss": 2.5816,
      "theoretical_loss": 3.353457075838132,
      "tokens_seen": 2645360640
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9339417439001325e-05,
      "loss": -0.1587,
      "theoretical_loss": 3.35345052381951,
      "tokens_seen": 2645426176
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.932286741062985e-05,
      "loss": 2.4515,
      "theoretical_loss": 3.353443972008648,
      "tokens_seen": 2645491712
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.930631738225837e-05,
      "loss": -1.6573,
      "theoretical_loss": 3.353437420405535,
      "tokens_seen": 2645557248
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.928976735388689e-05,
      "loss": 2.8237,
      "theoretical_loss": 3.3534308690101584,
      "tokens_seen": 2645622784
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.9273217325515414e-05,
      "loss": 1.786,
      "theoretical_loss": 3.3534243178225074,
      "tokens_seen": 2645688320
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.925666729714394e-05,
      "loss": 2.4159,
      "theoretical_loss": 3.3534177668425698,
      "tokens_seen": 2645753856
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.924011726877246e-05,
      "loss": 1.9723,
      "theoretical_loss": 3.353411216070333,
      "tokens_seen": 2645819392
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.922356724040098e-05,
      "loss": 2.4226,
      "theoretical_loss": 3.3534046655057868,
      "tokens_seen": 2645884928
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.92070172120295e-05,
      "loss": 2.8052,
      "theoretical_loss": 3.3533981151489187,
      "tokens_seen": 2645950464
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.919046718365803e-05,
      "loss": 2.4217,
      "theoretical_loss": 3.3533915649997166,
      "tokens_seen": 2646016000
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.917391715528655e-05,
      "loss": 1.2925,
      "theoretical_loss": 3.3533850150581697,
      "tokens_seen": 2646081536
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.9157367126915075e-05,
      "loss": 2.6279,
      "theoretical_loss": 3.3533784653242655,
      "tokens_seen": 2646147072
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.91408170985436e-05,
      "loss": 1.8557,
      "theoretical_loss": 3.353371915797992,
      "tokens_seen": 2646212608
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.912426707017212e-05,
      "loss": 2.4246,
      "theoretical_loss": 3.3533653664793386,
      "tokens_seen": 2646278144
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.910771704180064e-05,
      "loss": 2.5417,
      "theoretical_loss": 3.353358817368293,
      "tokens_seen": 2646343680
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.9091167013429164e-05,
      "loss": 2.7049,
      "theoretical_loss": 3.3533522684648434,
      "tokens_seen": 2646409216
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.907461698505769e-05,
      "loss": 1.8723,
      "theoretical_loss": 3.353345719768978,
      "tokens_seen": 2646474752
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.905806695668621e-05,
      "loss": 1.1573,
      "theoretical_loss": 3.3533391712806857,
      "tokens_seen": 2646540288
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.904151692831473e-05,
      "loss": 2.3421,
      "theoretical_loss": 3.353332622999954,
      "tokens_seen": 2646605824
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.9024966899943254e-05,
      "loss": 2.5117,
      "theoretical_loss": 3.3533260749267715,
      "tokens_seen": 2646671360
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.900841687157178e-05,
      "loss": -2.3067,
      "theoretical_loss": 3.3533195270611262,
      "tokens_seen": 2646736896
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.89918668432003e-05,
      "loss": 2.4273,
      "theoretical_loss": 3.353312979403007,
      "tokens_seen": 2646802432
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8975316814828826e-05,
      "loss": 1.1737,
      "theoretical_loss": 3.3533064319524017,
      "tokens_seen": 2646867968
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.895876678645735e-05,
      "loss": 2.3927,
      "theoretical_loss": 3.353299884709299,
      "tokens_seen": 2646933504
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8942216758085874e-05,
      "loss": 2.5282,
      "theoretical_loss": 3.3532933376736866,
      "tokens_seen": 2646999040
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.892566672971439e-05,
      "loss": 2.0371,
      "theoretical_loss": 3.3532867908455533,
      "tokens_seen": 2647064576
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8909116701342915e-05,
      "loss": 3.3534,
      "theoretical_loss": 3.353280244224887,
      "tokens_seen": 2647130112
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.889256667297144e-05,
      "loss": 2.6814,
      "theoretical_loss": 3.353273697811676,
      "tokens_seen": 2647195648
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.887601664459996e-05,
      "loss": 2.317,
      "theoretical_loss": 3.353267151605909,
      "tokens_seen": 2647261184
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.885946661622848e-05,
      "loss": 2.7092,
      "theoretical_loss": 3.3532606056075744,
      "tokens_seen": 2647326720
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8842916587857004e-05,
      "loss": 1.2077,
      "theoretical_loss": 3.35325405981666,
      "tokens_seen": 2647392256
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.882636655948553e-05,
      "loss": 1.1592,
      "theoretical_loss": 3.3532475142331535,
      "tokens_seen": 2647457792
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.880981653111405e-05,
      "loss": 2.1189,
      "theoretical_loss": 3.353240968857045,
      "tokens_seen": 2647523328
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8793266502742576e-05,
      "loss": 2.2705,
      "theoretical_loss": 3.353234423688321,
      "tokens_seen": 2647588864
    },
    {
      "epoch": 0.95,
      "objective/train/avg_token_score": 0.011643584817647934,
      "objective/train/avg_weight": 0.9862926602363586,
      "objective/train/docs_used": 1490143,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 14.147200584411621,
      "objective/train/std_weight": 0.08436877280473709,
      "objective/train/theoretical_loss": 3.3532278787269707,
      "objective/train/tokens_used": 2668114400,
      "theoretical_loss": 3.3532278787269707,
      "tokens_seen": 2647654400
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.87767164743711e-05,
      "loss": 1.7374,
      "theoretical_loss": 3.3532278787269707,
      "tokens_seen": 2647654400
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8760166445999624e-05,
      "loss": 1.8174,
      "theoretical_loss": 3.353221333972982,
      "tokens_seen": 2647719936
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.874361641762815e-05,
      "loss": 3.2166,
      "theoretical_loss": 3.353214789426344,
      "tokens_seen": 2647785472
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8727066389256665e-05,
      "loss": 2.3238,
      "theoretical_loss": 3.3532082450870444,
      "tokens_seen": 2647851008
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.871051636088519e-05,
      "loss": 2.4401,
      "theoretical_loss": 3.353201700955071,
      "tokens_seen": 2647916544
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.869396633251371e-05,
      "loss": 2.5713,
      "theoretical_loss": 3.353195157030413,
      "tokens_seen": 2647982080
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.867741630414223e-05,
      "loss": 2.7161,
      "theoretical_loss": 3.3531886133130584,
      "tokens_seen": 2648047616
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8660866275770754e-05,
      "loss": 2.5421,
      "theoretical_loss": 3.3531820698029953,
      "tokens_seen": 2648113152
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.864431624739928e-05,
      "loss": 2.2503,
      "theoretical_loss": 3.353175526500212,
      "tokens_seen": 2648178688
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.86277662190278e-05,
      "loss": 2.275,
      "theoretical_loss": 3.353168983404697,
      "tokens_seen": 2648244224
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8611216190656326e-05,
      "loss": 1.5884,
      "theoretical_loss": 3.3531624405164386,
      "tokens_seen": 2648309760
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.859466616228485e-05,
      "loss": 0.6561,
      "theoretical_loss": 3.353155897835425,
      "tokens_seen": 2648375296
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8578116133913374e-05,
      "loss": 2.4984,
      "theoretical_loss": 3.3531493553616447,
      "tokens_seen": 2648440832
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.85615661055419e-05,
      "loss": 2.9668,
      "theoretical_loss": 3.3531428130950856,
      "tokens_seen": 2648506368
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8545016077170415e-05,
      "loss": 0.1857,
      "theoretical_loss": 3.353136271035736,
      "tokens_seen": 2648571904
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.852846604879894e-05,
      "loss": 2.5339,
      "theoretical_loss": 3.353129729183585,
      "tokens_seen": 2648637440
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8511916020427463e-05,
      "loss": 2.4024,
      "theoretical_loss": 3.3531231875386207,
      "tokens_seen": 2648702976
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.849536599205598e-05,
      "loss": 1.5239,
      "theoretical_loss": 3.3531166461008306,
      "tokens_seen": 2648768512
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8478815963684505e-05,
      "loss": 2.8239,
      "theoretical_loss": 3.3531101048702037,
      "tokens_seen": 2648834048
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.846226593531303e-05,
      "loss": 2.5103,
      "theoretical_loss": 3.353103563846728,
      "tokens_seen": 2648899584
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.844571590694155e-05,
      "loss": 2.2561,
      "theoretical_loss": 3.353097023030392,
      "tokens_seen": 2648965120
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8429165878570077e-05,
      "loss": 1.8302,
      "theoretical_loss": 3.353090482421184,
      "tokens_seen": 2649030656
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.84126158501986e-05,
      "loss": 3.1871,
      "theoretical_loss": 3.3530839420190923,
      "tokens_seen": 2649096192
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8396065821827125e-05,
      "loss": 3.2681,
      "theoretical_loss": 3.353077401824105,
      "tokens_seen": 2649161728
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.837951579345565e-05,
      "loss": 2.7371,
      "theoretical_loss": 3.353070861836211,
      "tokens_seen": 2649227264
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8362965765084166e-05,
      "loss": 2.4634,
      "theoretical_loss": 3.3530643220553977,
      "tokens_seen": 2649292800
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.834641573671269e-05,
      "loss": 2.2754,
      "theoretical_loss": 3.3530577824816543,
      "tokens_seen": 2649358336
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8329865708341214e-05,
      "loss": 1.4485,
      "theoretical_loss": 3.353051243114969,
      "tokens_seen": 2649423872
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.831331567996974e-05,
      "loss": 2.4249,
      "theoretical_loss": 3.3530447039553293,
      "tokens_seen": 2649489408
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8296765651598255e-05,
      "loss": 3.3692,
      "theoretical_loss": 3.3530381650027246,
      "tokens_seen": 2649554944
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8280215623226786e-05,
      "loss": 1.3993,
      "theoretical_loss": 3.3530316262571427,
      "tokens_seen": 2649620480
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.826366559485531e-05,
      "loss": 1.1997,
      "theoretical_loss": 3.3530250877185717,
      "tokens_seen": 2649686016
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.824711556648383e-05,
      "loss": 2.2206,
      "theoretical_loss": 3.3530185493870004,
      "tokens_seen": 2649751552
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.823056553811235e-05,
      "loss": 2.6364,
      "theoretical_loss": 3.3530120112624173,
      "tokens_seen": 2649817088
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8214015509740875e-05,
      "loss": 2.0943,
      "theoretical_loss": 3.3530054733448096,
      "tokens_seen": 2649882624
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.81974654813694e-05,
      "loss": 2.2842,
      "theoretical_loss": 3.352998935634167,
      "tokens_seen": 2649948160
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8180915452997916e-05,
      "loss": 2.659,
      "theoretical_loss": 3.352992398130477,
      "tokens_seen": 2650013696
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.816436542462644e-05,
      "loss": 2.5836,
      "theoretical_loss": 3.352985860833728,
      "tokens_seen": 2650079232
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8147815396254964e-05,
      "loss": 2.3368,
      "theoretical_loss": 3.3529793237439085,
      "tokens_seen": 2650144768
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.813126536788349e-05,
      "loss": 2.0267,
      "theoretical_loss": 3.352972786861007,
      "tokens_seen": 2650210304
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.811471533951201e-05,
      "loss": 2.4671,
      "theoretical_loss": 3.3529662501850117,
      "tokens_seen": 2650275840
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8098165311140536e-05,
      "loss": 1.6051,
      "theoretical_loss": 3.3529597137159106,
      "tokens_seen": 2650341376
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.808161528276906e-05,
      "loss": 1.0493,
      "theoretical_loss": 3.352953177453693,
      "tokens_seen": 2650406912
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.806506525439758e-05,
      "loss": 2.1538,
      "theoretical_loss": 3.3529466413983458,
      "tokens_seen": 2650472448
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.80485152260261e-05,
      "loss": 1.6049,
      "theoretical_loss": 3.352940105549858,
      "tokens_seen": 2650537984
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.8031965197654625e-05,
      "loss": 1.4128,
      "theoretical_loss": 3.3529335699082186,
      "tokens_seen": 2650603520
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.801541516928315e-05,
      "loss": 2.9708,
      "theoretical_loss": 3.3529270344734154,
      "tokens_seen": 2650669056
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7998865140911667e-05,
      "loss": 1.8069,
      "theoretical_loss": 3.3529204992454362,
      "tokens_seen": 2650734592
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.798231511254019e-05,
      "loss": 1.1744,
      "theoretical_loss": 3.3529139642242702,
      "tokens_seen": 2650800128
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7965765084168714e-05,
      "loss": 2.659,
      "theoretical_loss": 3.352907429409905,
      "tokens_seen": 2650865664
    },
    {
      "epoch": 0.95,
      "objective/train/avg_token_score": 0.022484224289655685,
      "objective/train/avg_weight": 0.9735287427902222,
      "objective/train/docs_used": 1492100,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 40.74029541015625,
      "objective/train/std_weight": 0.14419615268707275,
      "objective/train/theoretical_loss": 3.35290089480233,
      "objective/train/tokens_used": 2671391200,
      "theoretical_loss": 3.35290089480233,
      "tokens_seen": 2650931200
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.794921505579724e-05,
      "loss": -0.1972,
      "theoretical_loss": 3.35290089480233,
      "tokens_seen": 2650931200
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.793266502742576e-05,
      "loss": 2.6085,
      "theoretical_loss": 3.352894360401532,
      "tokens_seen": 2650996736
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7916114999054286e-05,
      "loss": 4.3944,
      "theoretical_loss": 3.352887826207501,
      "tokens_seen": 2651062272
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.789956497068281e-05,
      "loss": 2.3718,
      "theoretical_loss": 3.3528812922202245,
      "tokens_seen": 2651127808
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7883014942311334e-05,
      "loss": 2.4474,
      "theoretical_loss": 3.3528747584396905,
      "tokens_seen": 2651193344
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.786646491393985e-05,
      "loss": 2.4941,
      "theoretical_loss": 3.352868224865888,
      "tokens_seen": 2651258880
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7849914885568376e-05,
      "loss": 2.4352,
      "theoretical_loss": 3.352861691498805,
      "tokens_seen": 2651324416
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.78333648571969e-05,
      "loss": 2.6584,
      "theoretical_loss": 3.3528551583384303,
      "tokens_seen": 2651389952
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.781681482882542e-05,
      "loss": -3.0932,
      "theoretical_loss": 3.352848625384752,
      "tokens_seen": 2651455488
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.780026480045394e-05,
      "loss": 2.5941,
      "theoretical_loss": 3.352842092637758,
      "tokens_seen": 2651521024
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7783714772082465e-05,
      "loss": 1.4908,
      "theoretical_loss": 3.352835560097437,
      "tokens_seen": 2651586560
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.776716474371099e-05,
      "loss": 2.5094,
      "theoretical_loss": 3.352829027763778,
      "tokens_seen": 2651652096
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.775061471533951e-05,
      "loss": 2.5636,
      "theoretical_loss": 3.352822495636768,
      "tokens_seen": 2651717632
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.773406468696804e-05,
      "loss": 3.3167,
      "theoretical_loss": 3.352815963716396,
      "tokens_seen": 2651783168
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.771751465859656e-05,
      "loss": 1.6719,
      "theoretical_loss": 3.3528094320026507,
      "tokens_seen": 2651848704
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7700964630225085e-05,
      "loss": 2.378,
      "theoretical_loss": 3.3528029004955204,
      "tokens_seen": 2651914240
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.76844146018536e-05,
      "loss": 2.5373,
      "theoretical_loss": 3.352796369194993,
      "tokens_seen": 2651979776
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7667864573482126e-05,
      "loss": 2.4339,
      "theoretical_loss": 3.3527898381010575,
      "tokens_seen": 2652045312
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.765131454511065e-05,
      "loss": 2.4985,
      "theoretical_loss": 3.352783307213701,
      "tokens_seen": 2652110848
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.763476451673917e-05,
      "loss": 2.7232,
      "theoretical_loss": 3.352776776532914,
      "tokens_seen": 2652176384
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.761821448836769e-05,
      "loss": 3.0793,
      "theoretical_loss": 3.352770246058683,
      "tokens_seen": 2652241920
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7601664459996215e-05,
      "loss": 2.6326,
      "theoretical_loss": 3.3527637157909966,
      "tokens_seen": 2652307456
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.758511443162474e-05,
      "loss": 2.4896,
      "theoretical_loss": 3.352757185729844,
      "tokens_seen": 2652372992
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.756856440325326e-05,
      "loss": 2.8433,
      "theoretical_loss": 3.3527506558752127,
      "tokens_seen": 2652438528
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.755201437488179e-05,
      "loss": 2.5285,
      "theoretical_loss": 3.352744126227092,
      "tokens_seen": 2652504064
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.753546434651031e-05,
      "loss": 2.3759,
      "theoretical_loss": 3.3527375967854693,
      "tokens_seen": 2652569600
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7518914318138835e-05,
      "loss": 2.6043,
      "theoretical_loss": 3.3527310675503337,
      "tokens_seen": 2652635136
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.750236428976735e-05,
      "loss": 3.0857,
      "theoretical_loss": 3.352724538521673,
      "tokens_seen": 2652700672
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7485814261395876e-05,
      "loss": 2.6984,
      "theoretical_loss": 3.352718009699476,
      "tokens_seen": 2652766208
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.74692642330244e-05,
      "loss": 2.1128,
      "theoretical_loss": 3.352711481083731,
      "tokens_seen": 2652831744
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7452714204652924e-05,
      "loss": 2.7174,
      "theoretical_loss": 3.3527049526744257,
      "tokens_seen": 2652897280
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.743616417628144e-05,
      "loss": 2.6161,
      "theoretical_loss": 3.3526984244715496,
      "tokens_seen": 2652962816
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7419614147909966e-05,
      "loss": 2.5738,
      "theoretical_loss": 3.3526918964750907,
      "tokens_seen": 2653028352
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.740306411953849e-05,
      "loss": 1.5103,
      "theoretical_loss": 3.352685368685037,
      "tokens_seen": 2653093888
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7386514091167014e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.3526788411013766,
      "tokens_seen": 2653159424
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.736996406279554e-05,
      "loss": 2.215,
      "theoretical_loss": 3.352672313724099,
      "tokens_seen": 2653224960
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.735341403442406e-05,
      "loss": 2.2049,
      "theoretical_loss": 3.3526657865531915,
      "tokens_seen": 2653290496
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7336864006052586e-05,
      "loss": 2.3402,
      "theoretical_loss": 3.352659259588643,
      "tokens_seen": 2653356032
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.73203139776811e-05,
      "loss": 2.343,
      "theoretical_loss": 3.352652732830442,
      "tokens_seen": 2653421568
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.730376394930963e-05,
      "loss": 2.3174,
      "theoretical_loss": 3.3526462062785765,
      "tokens_seen": 2653487104
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.728721392093815e-05,
      "loss": 2.4184,
      "theoretical_loss": 3.352639679933035,
      "tokens_seen": 2653552640
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7270663892566675e-05,
      "loss": 2.3946,
      "theoretical_loss": 3.352633153793806,
      "tokens_seen": 2653618176
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.725411386419519e-05,
      "loss": 2.5516,
      "theoretical_loss": 3.352626627860878,
      "tokens_seen": 2653683712
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7237563835823716e-05,
      "loss": 2.4341,
      "theoretical_loss": 3.3526201021342388,
      "tokens_seen": 2653749248
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.722101380745224e-05,
      "loss": 2.1803,
      "theoretical_loss": 3.3526135766138774,
      "tokens_seen": 2653814784
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7204463779080764e-05,
      "loss": 2.4802,
      "theoretical_loss": 3.352607051299782,
      "tokens_seen": 2653880320
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.718791375070929e-05,
      "loss": 1.6156,
      "theoretical_loss": 3.352600526191941,
      "tokens_seen": 2653945856
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.717136372233781e-05,
      "loss": 2.4237,
      "theoretical_loss": 3.352594001290343,
      "tokens_seen": 2654011392
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7154813693966336e-05,
      "loss": 1.9634,
      "theoretical_loss": 3.3525874765949752,
      "tokens_seen": 2654076928
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.713826366559485e-05,
      "loss": 2.5143,
      "theoretical_loss": 3.3525809521058276,
      "tokens_seen": 2654142464
    },
    {
      "epoch": 0.95,
      "objective/train/avg_token_score": 0.0101164560765028,
      "objective/train/avg_weight": 0.9881132245063782,
      "objective/train/docs_used": 1494002,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.89601993560791,
      "objective/train/std_weight": 0.05147651955485344,
      "objective/train/theoretical_loss": 3.352574427822888,
      "objective/train/tokens_used": 2674668000,
      "theoretical_loss": 3.352574427822888,
      "tokens_seen": 2654208000
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.712171363722338e-05,
      "loss": 2.6456,
      "theoretical_loss": 3.352574427822888,
      "tokens_seen": 2654208000
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.71051636088519e-05,
      "loss": 2.5367,
      "theoretical_loss": 3.352567903746144,
      "tokens_seen": 2654273536
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7088613580480425e-05,
      "loss": 2.4269,
      "theoretical_loss": 3.352561379875586,
      "tokens_seen": 2654339072
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.707206355210894e-05,
      "loss": 2.655,
      "theoretical_loss": 3.3525548562112,
      "tokens_seen": 2654404608
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7055513523737466e-05,
      "loss": 2.448,
      "theoretical_loss": 3.3525483327529755,
      "tokens_seen": 2654470144
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.703896349536599e-05,
      "loss": 2.0984,
      "theoretical_loss": 3.3525418095009014,
      "tokens_seen": 2654535680
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.7022413466994514e-05,
      "loss": 2.4541,
      "theoretical_loss": 3.352535286454965,
      "tokens_seen": 2654601216
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.700586343862304e-05,
      "loss": 2.2836,
      "theoretical_loss": 3.352528763615156,
      "tokens_seen": 2654666752
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.698931341025156e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.3525222409814615,
      "tokens_seen": 2654732288
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6972763381880086e-05,
      "loss": 2.7284,
      "theoretical_loss": 3.3525157185538705,
      "tokens_seen": 2654797824
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6956213353508603e-05,
      "loss": 2.4133,
      "theoretical_loss": 3.3525091963323717,
      "tokens_seen": 2654863360
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.693966332513713e-05,
      "loss": 2.4979,
      "theoretical_loss": 3.352502674316953,
      "tokens_seen": 2654928896
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.692311329676565e-05,
      "loss": 2.1472,
      "theoretical_loss": 3.352496152507603,
      "tokens_seen": 2654994432
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6906563268394175e-05,
      "loss": 2.4936,
      "theoretical_loss": 3.35248963090431,
      "tokens_seen": 2655059968
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.689001324002269e-05,
      "loss": 2.6494,
      "theoretical_loss": 3.3524831095070624,
      "tokens_seen": 2655125504
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.687346321165122e-05,
      "loss": 2.7513,
      "theoretical_loss": 3.3524765883158487,
      "tokens_seen": 2655191040
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.685691318327974e-05,
      "loss": 2.5512,
      "theoretical_loss": 3.3524700673306573,
      "tokens_seen": 2655256576
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6840363154908265e-05,
      "loss": 2.6042,
      "theoretical_loss": 3.3524635465514767,
      "tokens_seen": 2655322112
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.682381312653679e-05,
      "loss": 1.1202,
      "theoretical_loss": 3.352457025978295,
      "tokens_seen": 2655387648
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.680726309816531e-05,
      "loss": 2.4466,
      "theoretical_loss": 3.3524505056111007,
      "tokens_seen": 2655453184
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6790713069793837e-05,
      "loss": 1.4023,
      "theoretical_loss": 3.3524439854498826,
      "tokens_seen": 2655518720
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6774163041422354e-05,
      "loss": 1.4501,
      "theoretical_loss": 3.352437465494629,
      "tokens_seen": 2655584256
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.675761301305088e-05,
      "loss": 2.6815,
      "theoretical_loss": 3.352430945745328,
      "tokens_seen": 2655649792
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.67410629846794e-05,
      "loss": 2.3993,
      "theoretical_loss": 3.352424426201968,
      "tokens_seen": 2655715328
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6724512956307926e-05,
      "loss": 2.5783,
      "theoretical_loss": 3.3524179068645377,
      "tokens_seen": 2655780864
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.670796292793644e-05,
      "loss": 2.4902,
      "theoretical_loss": 3.3524113877330253,
      "tokens_seen": 2655846400
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.669141289956497e-05,
      "loss": 0.7146,
      "theoretical_loss": 3.3524048688074193,
      "tokens_seen": 2655911936
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.667486287119349e-05,
      "loss": 1.1505,
      "theoretical_loss": 3.352398350087708,
      "tokens_seen": 2655977472
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6658312842822015e-05,
      "loss": 2.4119,
      "theoretical_loss": 3.35239183157388,
      "tokens_seen": 2656043008
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.664176281445054e-05,
      "loss": 2.6493,
      "theoretical_loss": 3.352385313265924,
      "tokens_seen": 2656108544
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.662521278607906e-05,
      "loss": 2.3983,
      "theoretical_loss": 3.3523787951638275,
      "tokens_seen": 2656174080
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.660866275770759e-05,
      "loss": 2.7986,
      "theoretical_loss": 3.35237227726758,
      "tokens_seen": 2656239616
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.659211272933611e-05,
      "loss": 2.1585,
      "theoretical_loss": 3.352365759577169,
      "tokens_seen": 2656305152
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.657556270096463e-05,
      "loss": 2.4999,
      "theoretical_loss": 3.352359242092583,
      "tokens_seen": 2656370688
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.655901267259315e-05,
      "loss": 2.4048,
      "theoretical_loss": 3.3523527248138114,
      "tokens_seen": 2656436224
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6542462644221676e-05,
      "loss": 2.5247,
      "theoretical_loss": 3.3523462077408417,
      "tokens_seen": 2656501760
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.652591261585019e-05,
      "loss": 2.5109,
      "theoretical_loss": 3.352339690873663,
      "tokens_seen": 2656567296
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.650936258747872e-05,
      "loss": 2.2724,
      "theoretical_loss": 3.352333174212263,
      "tokens_seen": 2656632832
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.649281255910724e-05,
      "loss": 2.4047,
      "theoretical_loss": 3.35232665775663,
      "tokens_seen": 2656698368
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6476262530735765e-05,
      "loss": 1.2747,
      "theoretical_loss": 3.3523201415067536,
      "tokens_seen": 2656763904
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.645971250236429e-05,
      "loss": 2.4615,
      "theoretical_loss": 3.352313625462621,
      "tokens_seen": 2656829440
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.644316247399281e-05,
      "loss": 2.3143,
      "theoretical_loss": 3.3523071096242214,
      "tokens_seen": 2656894976
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.642661244562134e-05,
      "loss": 2.5246,
      "theoretical_loss": 3.352300593991543,
      "tokens_seen": 2656960512
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.641006241724986e-05,
      "loss": 2.2174,
      "theoretical_loss": 3.352294078564574,
      "tokens_seen": 2657026048
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.639351238887838e-05,
      "loss": 2.0642,
      "theoretical_loss": 3.352287563343303,
      "tokens_seen": 2657091584
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.63769623605069e-05,
      "loss": 3.0217,
      "theoretical_loss": 3.352281048327719,
      "tokens_seen": 2657157120
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6360412332135426e-05,
      "loss": 2.7419,
      "theoretical_loss": 3.352274533517809,
      "tokens_seen": 2657222656
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6343862303763944e-05,
      "loss": 2.5518,
      "theoretical_loss": 3.3522680189135627,
      "tokens_seen": 2657288192
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.632731227539247e-05,
      "loss": 2.2401,
      "theoretical_loss": 3.3522615045149684,
      "tokens_seen": 2657353728
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.631076224702099e-05,
      "loss": 4.9273,
      "theoretical_loss": 3.352254990322014,
      "tokens_seen": 2657419264
    },
    {
      "epoch": 0.95,
      "objective/train/avg_token_score": 0.011713355779647827,
      "objective/train/avg_weight": 0.9862116575241089,
      "objective/train/docs_used": 1495727,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.093133926391602,
      "objective/train/std_weight": 0.07616729289293289,
      "objective/train/theoretical_loss": 3.3522484763346885,
      "objective/train/tokens_used": 2677944800,
      "theoretical_loss": 3.3522484763346885,
      "tokens_seen": 2657484800
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6294212218649516e-05,
      "loss": 2.0458,
      "theoretical_loss": 3.3522484763346885,
      "tokens_seen": 2657484800
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.627766219027804e-05,
      "loss": 1.8959,
      "theoretical_loss": 3.35224196255298,
      "tokens_seen": 2657550336
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6261112161906564e-05,
      "loss": 2.8119,
      "theoretical_loss": 3.352235448976877,
      "tokens_seen": 2657615872
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.624456213353509e-05,
      "loss": 2.5034,
      "theoretical_loss": 3.3522289356063677,
      "tokens_seen": 2657681408
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.622801210516361e-05,
      "loss": 2.5978,
      "theoretical_loss": 3.3522224224414408,
      "tokens_seen": 2657746944
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.621146207679213e-05,
      "loss": 2.4783,
      "theoretical_loss": 3.352215909482085,
      "tokens_seen": 2657812480
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.619491204842065e-05,
      "loss": 2.0233,
      "theoretical_loss": 3.352209396728288,
      "tokens_seen": 2657878016
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.617836202004918e-05,
      "loss": 2.2928,
      "theoretical_loss": 3.3522028841800395,
      "tokens_seen": 2657943552
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.61618119916777e-05,
      "loss": 2.0001,
      "theoretical_loss": 3.3521963718373264,
      "tokens_seen": 2658009088
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.614526196330622e-05,
      "loss": 1.1186,
      "theoretical_loss": 3.3521898597001387,
      "tokens_seen": 2658074624
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.612871193493474e-05,
      "loss": 2.7166,
      "theoretical_loss": 3.3521833477684635,
      "tokens_seen": 2658140160
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.611216190656327e-05,
      "loss": 4.137,
      "theoretical_loss": 3.3521768360422897,
      "tokens_seen": 2658205696
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.609561187819179e-05,
      "loss": 2.5473,
      "theoretical_loss": 3.352170324521606,
      "tokens_seen": 2658271232
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.6079061849820314e-05,
      "loss": 2.2863,
      "theoretical_loss": 3.352163813206401,
      "tokens_seen": 2658336768
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.606251182144884e-05,
      "loss": 1.0519,
      "theoretical_loss": 3.3521573020966624,
      "tokens_seen": 2658402304
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.604596179307736e-05,
      "loss": 2.5783,
      "theoretical_loss": 3.3521507911923796,
      "tokens_seen": 2658467840
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.602941176470588e-05,
      "loss": 2.1688,
      "theoretical_loss": 3.3521442804935404,
      "tokens_seen": 2658533376
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.60128617363344e-05,
      "loss": 2.4526,
      "theoretical_loss": 3.352137770000133,
      "tokens_seen": 2658598912
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.599631170796293e-05,
      "loss": 2.6568,
      "theoretical_loss": 3.352131259712147,
      "tokens_seen": 2658664448
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.597976167959145e-05,
      "loss": 1.487,
      "theoretical_loss": 3.3521247496295694,
      "tokens_seen": 2658729984
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5963211651219975e-05,
      "loss": 1.7986,
      "theoretical_loss": 3.3521182397523894,
      "tokens_seen": 2658795520
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.59466616228485e-05,
      "loss": 1.9441,
      "theoretical_loss": 3.352111730080596,
      "tokens_seen": 2658861056
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.593011159447702e-05,
      "loss": 2.3481,
      "theoretical_loss": 3.3521052206141766,
      "tokens_seen": 2658926592
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.591356156610554e-05,
      "loss": 2.065,
      "theoretical_loss": 3.35209871135312,
      "tokens_seen": 2658992128
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5897011537734064e-05,
      "loss": 1.6127,
      "theoretical_loss": 3.3520922022974156,
      "tokens_seen": 2659057664
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.588046150936259e-05,
      "loss": 1.9403,
      "theoretical_loss": 3.3520856934470507,
      "tokens_seen": 2659123200
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.586391148099111e-05,
      "loss": 1.4564,
      "theoretical_loss": 3.352079184802014,
      "tokens_seen": 2659188736
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.584736145261963e-05,
      "loss": 3.1429,
      "theoretical_loss": 3.352072676362294,
      "tokens_seen": 2659254272
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5830811424248154e-05,
      "loss": 2.5163,
      "theoretical_loss": 3.3520661681278794,
      "tokens_seen": 2659319808
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.581426139587668e-05,
      "loss": 2.309,
      "theoretical_loss": 3.3520596600987584,
      "tokens_seen": 2659385344
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.57977113675052e-05,
      "loss": 2.5525,
      "theoretical_loss": 3.35205315227492,
      "tokens_seen": 2659450880
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5781161339133726e-05,
      "loss": 2.5766,
      "theoretical_loss": 3.352046644656352,
      "tokens_seen": 2659516416
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.576461131076225e-05,
      "loss": 2.7416,
      "theoretical_loss": 3.352040137243043,
      "tokens_seen": 2659581952
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5748061282390774e-05,
      "loss": 2.0931,
      "theoretical_loss": 3.3520336300349816,
      "tokens_seen": 2659647488
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.57315112540193e-05,
      "loss": 2.5068,
      "theoretical_loss": 3.3520271230321566,
      "tokens_seen": 2659713024
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5714961225647815e-05,
      "loss": 2.4164,
      "theoretical_loss": 3.3520206162345554,
      "tokens_seen": 2659778560
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.569841119727634e-05,
      "loss": 2.6902,
      "theoretical_loss": 3.352014109642168,
      "tokens_seen": 2659844096
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.568186116890486e-05,
      "loss": 2.347,
      "theoretical_loss": 3.3520076032549815,
      "tokens_seen": 2659909632
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.566531114053338e-05,
      "loss": 2.4673,
      "theoretical_loss": 3.352001097072985,
      "tokens_seen": 2659975168
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5648761112161904e-05,
      "loss": 1.9,
      "theoretical_loss": 3.3519945910961675,
      "tokens_seen": 2660040704
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.563221108379043e-05,
      "loss": 1.3476,
      "theoretical_loss": 3.3519880853245163,
      "tokens_seen": 2660106240
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.561566105541895e-05,
      "loss": 2.3856,
      "theoretical_loss": 3.3519815797580206,
      "tokens_seen": 2660171776
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5599111027047476e-05,
      "loss": 2.2523,
      "theoretical_loss": 3.3519750743966688,
      "tokens_seen": 2660237312
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5582560998676e-05,
      "loss": 2.7985,
      "theoretical_loss": 3.3519685692404493,
      "tokens_seen": 2660302848
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5566010970304524e-05,
      "loss": 2.3598,
      "theoretical_loss": 3.3519620642893506,
      "tokens_seen": 2660368384
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.554946094193305e-05,
      "loss": 2.7192,
      "theoretical_loss": 3.351955559543361,
      "tokens_seen": 2660433920
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5532910913561565e-05,
      "loss": 2.3626,
      "theoretical_loss": 3.3519490550024695,
      "tokens_seen": 2660499456
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.551636088519009e-05,
      "loss": 2.3462,
      "theoretical_loss": 3.3519425506666645,
      "tokens_seen": 2660564992
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.549981085681861e-05,
      "loss": 2.3603,
      "theoretical_loss": 3.3519360465359336,
      "tokens_seen": 2660630528
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.548326082844713e-05,
      "loss": 6.346,
      "theoretical_loss": 3.351929542610266,
      "tokens_seen": 2660696064
    },
    {
      "epoch": 0.95,
      "objective/train/avg_token_score": 0.017239362001419067,
      "objective/train/avg_weight": 0.979710578918457,
      "objective/train/docs_used": 1497686,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.3483963012695312,
      "objective/train/std_weight": 0.06445354968309402,
      "objective/train/theoretical_loss": 3.3519230388896504,
      "objective/train/tokens_used": 2681221600,
      "theoretical_loss": 3.3519230388896504,
      "tokens_seen": 2660761600
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5466710800075654e-05,
      "loss": 2.7056,
      "theoretical_loss": 3.3519230388896504,
      "tokens_seen": 2660761600
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.545016077170418e-05,
      "loss": 3.1363,
      "theoretical_loss": 3.351916535374075,
      "tokens_seen": 2660827136
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.54336107433327e-05,
      "loss": 2.0924,
      "theoretical_loss": 3.351910032063528,
      "tokens_seen": 2660892672
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5417060714961226e-05,
      "loss": 2.5063,
      "theoretical_loss": 3.351903528957998,
      "tokens_seen": 2660958208
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.540051068658975e-05,
      "loss": 1.6426,
      "theoretical_loss": 3.3518970260574736,
      "tokens_seen": 2661023744
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5383960658218274e-05,
      "loss": 1.5233,
      "theoretical_loss": 3.351890523361944,
      "tokens_seen": 2661089280
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.53674106298468e-05,
      "loss": 2.5087,
      "theoretical_loss": 3.3518840208713967,
      "tokens_seen": 2661154816
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5350860601475315e-05,
      "loss": 2.2612,
      "theoretical_loss": 3.35187751858582,
      "tokens_seen": 2661220352
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.533431057310384e-05,
      "loss": 2.4829,
      "theoretical_loss": 3.3518710165052035,
      "tokens_seen": 2661285888
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5317760544732363e-05,
      "loss": 3.7528,
      "theoretical_loss": 3.351864514629535,
      "tokens_seen": 2661351424
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.530121051636089e-05,
      "loss": 5.7648,
      "theoretical_loss": 3.351858012958803,
      "tokens_seen": 2661416960
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5284660487989405e-05,
      "loss": 4.0415,
      "theoretical_loss": 3.351851511492996,
      "tokens_seen": 2661482496
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.526811045961793e-05,
      "loss": 2.4016,
      "theoretical_loss": 3.3518450102321027,
      "tokens_seen": 2661548032
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.525156043124645e-05,
      "loss": 2.5066,
      "theoretical_loss": 3.3518385091761114,
      "tokens_seen": 2661613568
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5235010402874977e-05,
      "loss": 3.2962,
      "theoretical_loss": 3.3518320083250104,
      "tokens_seen": 2661679104
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.52184603745035e-05,
      "loss": 2.5341,
      "theoretical_loss": 3.3518255076787886,
      "tokens_seen": 2661744640
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5201910346132025e-05,
      "loss": 2.314,
      "theoretical_loss": 3.3518190072374345,
      "tokens_seen": 2661810176
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.518536031776055e-05,
      "loss": 2.2219,
      "theoretical_loss": 3.3518125070009366,
      "tokens_seen": 2661875712
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5168810289389066e-05,
      "loss": 2.4735,
      "theoretical_loss": 3.351806006969283,
      "tokens_seen": 2661941248
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.515226026101759e-05,
      "loss": 2.419,
      "theoretical_loss": 3.351799507142463,
      "tokens_seen": 2662006784
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5135710232646114e-05,
      "loss": 2.3708,
      "theoretical_loss": 3.351793007520464,
      "tokens_seen": 2662072320
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.511916020427464e-05,
      "loss": -3.447,
      "theoretical_loss": 3.3517865081032747,
      "tokens_seen": 2662137856
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5102610175903155e-05,
      "loss": 2.4176,
      "theoretical_loss": 3.351780008890885,
      "tokens_seen": 2662203392
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.508606014753168e-05,
      "loss": 2.5046,
      "theoretical_loss": 3.351773509883282,
      "tokens_seen": 2662268928
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.50695101191602e-05,
      "loss": 2.6397,
      "theoretical_loss": 3.351767011080454,
      "tokens_seen": 2662334464
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.505296009078873e-05,
      "loss": 2.1507,
      "theoretical_loss": 3.3517605124823904,
      "tokens_seen": 2662400000
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.503641006241725e-05,
      "loss": 2.5341,
      "theoretical_loss": 3.35175401408908,
      "tokens_seen": 2662465536
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.5019860034045775e-05,
      "loss": 2.5433,
      "theoretical_loss": 3.3517475159005103,
      "tokens_seen": 2662531072
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.50033100056743e-05,
      "loss": 0.6746,
      "theoretical_loss": 3.35174101791667,
      "tokens_seen": 2662596608
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4986759977302816e-05,
      "loss": 2.0913,
      "theoretical_loss": 3.351734520137548,
      "tokens_seen": 2662662144
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.497020994893134e-05,
      "loss": 2.3502,
      "theoretical_loss": 3.3517280225631327,
      "tokens_seen": 2662727680
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4953659920559864e-05,
      "loss": 0.9734,
      "theoretical_loss": 3.3517215251934123,
      "tokens_seen": 2662793216
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.493710989218839e-05,
      "loss": 2.525,
      "theoretical_loss": 3.351715028028376,
      "tokens_seen": 2662858752
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4920559863816905e-05,
      "loss": 0.9788,
      "theoretical_loss": 3.3517085310680117,
      "tokens_seen": 2662924288
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.490400983544543e-05,
      "loss": 1.9371,
      "theoretical_loss": 3.351702034312308,
      "tokens_seen": 2662989824
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.488745980707395e-05,
      "loss": 2.6223,
      "theoretical_loss": 3.3516955377612536,
      "tokens_seen": 2663055360
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.487090977870248e-05,
      "loss": 2.4764,
      "theoretical_loss": 3.351689041414837,
      "tokens_seen": 2663120896
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4854359750331e-05,
      "loss": 2.7008,
      "theoretical_loss": 3.3516825452730465,
      "tokens_seen": 2663186432
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4837809721959525e-05,
      "loss": 2.5313,
      "theoretical_loss": 3.351676049335871,
      "tokens_seen": 2663251968
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.482125969358805e-05,
      "loss": 2.4223,
      "theoretical_loss": 3.3516695536032985,
      "tokens_seen": 2663317504
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4804709665216566e-05,
      "loss": 2.4745,
      "theoretical_loss": 3.351663058075318,
      "tokens_seen": 2663383040
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.478815963684509e-05,
      "loss": 2.4262,
      "theoretical_loss": 3.3516565627519177,
      "tokens_seen": 2663448576
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4771609608473614e-05,
      "loss": 2.653,
      "theoretical_loss": 3.3516500676330865,
      "tokens_seen": 2663514112
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.475505958010214e-05,
      "loss": 1.9212,
      "theoretical_loss": 3.3516435727188125,
      "tokens_seen": 2663579648
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4738509551730656e-05,
      "loss": 2.2846,
      "theoretical_loss": 3.3516370780090847,
      "tokens_seen": 2663645184
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.472195952335918e-05,
      "loss": 2.0448,
      "theoretical_loss": 3.3516305835038906,
      "tokens_seen": 2663710720
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4705409494987704e-05,
      "loss": 3.3373,
      "theoretical_loss": 3.35162408920322,
      "tokens_seen": 2663776256
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.468885946661623e-05,
      "loss": 2.2432,
      "theoretical_loss": 3.351617595107061,
      "tokens_seen": 2663841792
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.467230943824475e-05,
      "loss": 2.5577,
      "theoretical_loss": 3.351611101215402,
      "tokens_seen": 2663907328
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4655759409873276e-05,
      "loss": 2.8327,
      "theoretical_loss": 3.3516046075282313,
      "tokens_seen": 2663972864
    },
    {
      "epoch": 0.95,
      "objective/train/avg_token_score": 0.006268182769417763,
      "objective/train/avg_weight": 0.9926273822784424,
      "objective/train/docs_used": 1499388,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.693689346313477,
      "objective/train/std_weight": 0.04854466766119003,
      "objective/train/theoretical_loss": 3.3515981140455375,
      "objective/train/tokens_used": 2684498400,
      "theoretical_loss": 3.3515981140455375,
      "tokens_seen": 2664038400
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.46392093815018e-05,
      "loss": 2.4408,
      "theoretical_loss": 3.3515981140455375,
      "tokens_seen": 2664038400
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.462265935313032e-05,
      "loss": 2.3073,
      "theoretical_loss": 3.35159162076731,
      "tokens_seen": 2664103936
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.460610932475884e-05,
      "loss": 2.791,
      "theoretical_loss": 3.3515851276935362,
      "tokens_seen": 2664169472
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4589559296387365e-05,
      "loss": 1.8601,
      "theoretical_loss": 3.351578634824205,
      "tokens_seen": 2664235008
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.457300926801589e-05,
      "loss": 2.6016,
      "theoretical_loss": 3.351572142159305,
      "tokens_seen": 2664300544
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4556459239644406e-05,
      "loss": 2.2524,
      "theoretical_loss": 3.351565649698825,
      "tokens_seen": 2664366080
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.453990921127293e-05,
      "loss": 2.4713,
      "theoretical_loss": 3.351559157442753,
      "tokens_seen": 2664431616
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4523359182901454e-05,
      "loss": 2.6861,
      "theoretical_loss": 3.351552665391078,
      "tokens_seen": 2664497152
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.450680915452998e-05,
      "loss": 2.3153,
      "theoretical_loss": 3.351546173543788,
      "tokens_seen": 2664562688
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.44902591261585e-05,
      "loss": 2.3127,
      "theoretical_loss": 3.3515396819008725,
      "tokens_seen": 2664628224
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4473709097787026e-05,
      "loss": 2.6031,
      "theoretical_loss": 3.351533190462319,
      "tokens_seen": 2664693760
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.445715906941555e-05,
      "loss": 2.5316,
      "theoretical_loss": 3.351526699228117,
      "tokens_seen": 2664759296
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.444060904104407e-05,
      "loss": 2.3691,
      "theoretical_loss": 3.351520208198254,
      "tokens_seen": 2664824832
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.442405901267259e-05,
      "loss": 2.6754,
      "theoretical_loss": 3.351513717372719,
      "tokens_seen": 2664890368
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4407508984301115e-05,
      "loss": -0.733,
      "theoretical_loss": 3.351507226751501,
      "tokens_seen": 2664955904
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.439095895592964e-05,
      "loss": 2.6045,
      "theoretical_loss": 3.3515007363345877,
      "tokens_seen": 2665021440
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4374408927558156e-05,
      "loss": 2.6506,
      "theoretical_loss": 3.3514942461219683,
      "tokens_seen": 2665086976
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.435785889918668e-05,
      "loss": 1.6948,
      "theoretical_loss": 3.351487756113631,
      "tokens_seen": 2665152512
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4341308870815204e-05,
      "loss": 2.2408,
      "theoretical_loss": 3.3514812663095648,
      "tokens_seen": 2665218048
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.432475884244373e-05,
      "loss": 2.4003,
      "theoretical_loss": 3.3514747767097575,
      "tokens_seen": 2665283584
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.430820881407225e-05,
      "loss": 2.8772,
      "theoretical_loss": 3.3514682873141988,
      "tokens_seen": 2665349120
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4291658785700776e-05,
      "loss": 2.7893,
      "theoretical_loss": 3.351461798122876,
      "tokens_seen": 2665414656
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.42751087573293e-05,
      "loss": 3.932,
      "theoretical_loss": 3.351455309135778,
      "tokens_seen": 2665480192
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4258558728957824e-05,
      "loss": 2.4886,
      "theoretical_loss": 3.351448820352894,
      "tokens_seen": 2665545728
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.424200870058634e-05,
      "loss": 2.6166,
      "theoretical_loss": 3.3514423317742117,
      "tokens_seen": 2665611264
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4225458672214866e-05,
      "loss": 0.9893,
      "theoretical_loss": 3.3514358433997202,
      "tokens_seen": 2665676800
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.420890864384339e-05,
      "loss": 2.3544,
      "theoretical_loss": 3.351429355229408,
      "tokens_seen": 2665742336
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.419235861547191e-05,
      "loss": 1.8319,
      "theoretical_loss": 3.3514228672632633,
      "tokens_seen": 2665807872
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.417580858710043e-05,
      "loss": 2.3546,
      "theoretical_loss": 3.3514163795012752,
      "tokens_seen": 2665873408
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4159258558728955e-05,
      "loss": 2.0393,
      "theoretical_loss": 3.351409891943432,
      "tokens_seen": 2665938944
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.414270853035748e-05,
      "loss": 2.6559,
      "theoretical_loss": 3.3514034045897216,
      "tokens_seen": 2666004480
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4126158501986e-05,
      "loss": 2.4552,
      "theoretical_loss": 3.351396917440134,
      "tokens_seen": 2666070016
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.410960847361453e-05,
      "loss": 2.6398,
      "theoretical_loss": 3.3513904304946562,
      "tokens_seen": 2666135552
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.409305844524305e-05,
      "loss": 2.5643,
      "theoretical_loss": 3.351383943753278,
      "tokens_seen": 2666201088
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4076508416871575e-05,
      "loss": 2.1084,
      "theoretical_loss": 3.351377457215987,
      "tokens_seen": 2666266624
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.405995838850009e-05,
      "loss": 1.5992,
      "theoretical_loss": 3.351370970882773,
      "tokens_seen": 2666332160
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.4043408360128616e-05,
      "loss": 2.3411,
      "theoretical_loss": 3.3513644847536233,
      "tokens_seen": 2666397696
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.402685833175714e-05,
      "loss": -0.4189,
      "theoretical_loss": 3.3513579988285267,
      "tokens_seen": 2666463232
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.401030830338566e-05,
      "loss": 2.6517,
      "theoretical_loss": 3.3513515131074723,
      "tokens_seen": 2666528768
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.399375827501418e-05,
      "loss": 1.5136,
      "theoretical_loss": 3.3513450275904484,
      "tokens_seen": 2666594304
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3977208246642705e-05,
      "loss": 2.143,
      "theoretical_loss": 3.3513385422774435,
      "tokens_seen": 2666659840
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3960658218271236e-05,
      "loss": 2.5921,
      "theoretical_loss": 3.3513320571684466,
      "tokens_seen": 2666725376
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.394410818989975e-05,
      "loss": 2.4799,
      "theoretical_loss": 3.3513255722634456,
      "tokens_seen": 2666790912
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.392755816152828e-05,
      "loss": 2.3728,
      "theoretical_loss": 3.351319087562429,
      "tokens_seen": 2666856448
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.39110081331568e-05,
      "loss": 2.636,
      "theoretical_loss": 3.351312603065386,
      "tokens_seen": 2666921984
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3894458104785325e-05,
      "loss": 1.9896,
      "theoretical_loss": 3.351306118772305,
      "tokens_seen": 2666987520
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.387790807641384e-05,
      "loss": 2.3274,
      "theoretical_loss": 3.3512996346831745,
      "tokens_seen": 2667053056
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3861358048042366e-05,
      "loss": 2.2556,
      "theoretical_loss": 3.351293150797983,
      "tokens_seen": 2667118592
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.384480801967089e-05,
      "loss": 0.3932,
      "theoretical_loss": 3.351286667116719,
      "tokens_seen": 2667184128
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3828257991299414e-05,
      "loss": 2.556,
      "theoretical_loss": 3.3512801836393713,
      "tokens_seen": 2667249664
    },
    {
      "epoch": 0.95,
      "objective/train/avg_token_score": 0.021785810589790344,
      "objective/train/avg_weight": 0.9743589162826538,
      "objective/train/docs_used": 1500922,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.8756556510925293,
      "objective/train/std_weight": 0.09404383599758148,
      "objective/train/theoretical_loss": 3.3512737003659288,
      "objective/train/tokens_used": 2687775200,
      "theoretical_loss": 3.3512737003659288,
      "tokens_seen": 2667315200
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.381170796292794e-05,
      "loss": 2.2332,
      "theoretical_loss": 3.3512737003659288,
      "tokens_seen": 2667315200
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.379515793455646e-05,
      "loss": 2.4191,
      "theoretical_loss": 3.351267217296379,
      "tokens_seen": 2667380736
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3778607906184986e-05,
      "loss": 1.7033,
      "theoretical_loss": 3.3512607344307113,
      "tokens_seen": 2667446272
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3762057877813503e-05,
      "loss": 1.9957,
      "theoretical_loss": 3.351254251768914,
      "tokens_seen": 2667511808
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.374550784944203e-05,
      "loss": 2.0983,
      "theoretical_loss": 3.351247769310976,
      "tokens_seen": 2667577344
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.372895782107055e-05,
      "loss": 2.4101,
      "theoretical_loss": 3.3512412870568857,
      "tokens_seen": 2667642880
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3712407792699075e-05,
      "loss": 1.7204,
      "theoretical_loss": 3.351234805006632,
      "tokens_seen": 2667708416
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.369585776432759e-05,
      "loss": 2.4973,
      "theoretical_loss": 3.3512283231602025,
      "tokens_seen": 2667773952
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3679307735956117e-05,
      "loss": 1.9439,
      "theoretical_loss": 3.3512218415175865,
      "tokens_seen": 2667839488
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.366275770758464e-05,
      "loss": 1.9853,
      "theoretical_loss": 3.351215360078773,
      "tokens_seen": 2667905024
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3646207679213165e-05,
      "loss": 2.3158,
      "theoretical_loss": 3.3512088788437495,
      "tokens_seen": 2667970560
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.362965765084169e-05,
      "loss": 2.2793,
      "theoretical_loss": 3.3512023978125054,
      "tokens_seen": 2668036096
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.361310762247021e-05,
      "loss": -0.176,
      "theoretical_loss": 3.351195916985029,
      "tokens_seen": 2668101632
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3596557594098737e-05,
      "loss": 2.0261,
      "theoretical_loss": 3.351189436361309,
      "tokens_seen": 2668167168
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3580007565727254e-05,
      "loss": 1.6368,
      "theoretical_loss": 3.351182955941334,
      "tokens_seen": 2668232704
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.356345753735578e-05,
      "loss": 2.5887,
      "theoretical_loss": 3.351176475725093,
      "tokens_seen": 2668298240
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.35469075089843e-05,
      "loss": 1.8764,
      "theoretical_loss": 3.3511699957125733,
      "tokens_seen": 2668363776
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3530357480612826e-05,
      "loss": 2.2612,
      "theoretical_loss": 3.3511635159037647,
      "tokens_seen": 2668429312
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.351380745224134e-05,
      "loss": 2.3799,
      "theoretical_loss": 3.3511570362986554,
      "tokens_seen": 2668494848
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.349725742386987e-05,
      "loss": 1.9549,
      "theoretical_loss": 3.3511505568972337,
      "tokens_seen": 2668560384
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.348070739549839e-05,
      "loss": 1.2054,
      "theoretical_loss": 3.351144077699489,
      "tokens_seen": 2668625920
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3464157367126915e-05,
      "loss": 2.5162,
      "theoretical_loss": 3.3511375987054093,
      "tokens_seen": 2668691456
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.344760733875544e-05,
      "loss": 1.8452,
      "theoretical_loss": 3.3511311199149834,
      "tokens_seen": 2668756992
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.343105731038396e-05,
      "loss": 2.8365,
      "theoretical_loss": 3.3511246413281994,
      "tokens_seen": 2668822528
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.341450728201249e-05,
      "loss": 1.8442,
      "theoretical_loss": 3.3511181629450464,
      "tokens_seen": 2668888064
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.339795725364101e-05,
      "loss": 2.4599,
      "theoretical_loss": 3.351111684765513,
      "tokens_seen": 2668953600
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.338140722526953e-05,
      "loss": 2.3477,
      "theoretical_loss": 3.3511052067895877,
      "tokens_seen": 2669019136
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.336485719689805e-05,
      "loss": 2.3583,
      "theoretical_loss": 3.351098729017259,
      "tokens_seen": 2669084672
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3348307168526576e-05,
      "loss": 2.5271,
      "theoretical_loss": 3.3510922514485157,
      "tokens_seen": 2669150208
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.333175714015509e-05,
      "loss": 2.6539,
      "theoretical_loss": 3.351085774083346,
      "tokens_seen": 2669215744
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.331520711178362e-05,
      "loss": 2.7751,
      "theoretical_loss": 3.351079296921739,
      "tokens_seen": 2669281280
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.329865708341214e-05,
      "loss": 2.6119,
      "theoretical_loss": 3.3510728199636834,
      "tokens_seen": 2669346816
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3282107055040665e-05,
      "loss": 2.7025,
      "theoretical_loss": 3.351066343209167,
      "tokens_seen": 2669412352
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.326555702666919e-05,
      "loss": 1.5671,
      "theoretical_loss": 3.3510598666581792,
      "tokens_seen": 2669477888
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.324900699829771e-05,
      "loss": 2.5025,
      "theoretical_loss": 3.3510533903107085,
      "tokens_seen": 2669543424
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.323245696992624e-05,
      "loss": 2.5825,
      "theoretical_loss": 3.3510469141667434,
      "tokens_seen": 2669608960
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.321590694155476e-05,
      "loss": 1.3432,
      "theoretical_loss": 3.351040438226272,
      "tokens_seen": 2669674496
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.319935691318328e-05,
      "loss": 0.5529,
      "theoretical_loss": 3.3510339624892835,
      "tokens_seen": 2669740032
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.31828068848118e-05,
      "loss": 2.4792,
      "theoretical_loss": 3.3510274869557666,
      "tokens_seen": 2669805568
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3166256856440326e-05,
      "loss": 0.5609,
      "theoretical_loss": 3.3510210116257095,
      "tokens_seen": 2669871104
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3149706828068844e-05,
      "loss": 3.5031,
      "theoretical_loss": 3.351014536499101,
      "tokens_seen": 2669936640
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.313315679969737e-05,
      "loss": 2.4963,
      "theoretical_loss": 3.3510080615759295,
      "tokens_seen": 2670002176
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.311660677132589e-05,
      "loss": 3.1337,
      "theoretical_loss": 3.351001586856184,
      "tokens_seen": 2670067712
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3100056742954416e-05,
      "loss": 2.3755,
      "theoretical_loss": 3.350995112339853,
      "tokens_seen": 2670133248
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.308350671458294e-05,
      "loss": 2.6187,
      "theoretical_loss": 3.3509886380269247,
      "tokens_seen": 2670198784
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.3066956686211464e-05,
      "loss": 2.3101,
      "theoretical_loss": 3.3509821639173887,
      "tokens_seen": 2670264320
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.305040665783999e-05,
      "loss": 2.4492,
      "theoretical_loss": 3.3509756900112326,
      "tokens_seen": 2670329856
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.303385662946851e-05,
      "loss": 2.0745,
      "theoretical_loss": 3.3509692163084455,
      "tokens_seen": 2670395392
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.301730660109703e-05,
      "loss": -0.0478,
      "theoretical_loss": 3.350962742809016,
      "tokens_seen": 2670460928
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.300075657272555e-05,
      "loss": 2.4218,
      "theoretical_loss": 3.3509562695129325,
      "tokens_seen": 2670526464
    },
    {
      "epoch": 0.95,
      "objective/train/avg_token_score": 0.02041342854499817,
      "objective/train/avg_weight": 0.9759702086448669,
      "objective/train/docs_used": 1502841,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5080318450927734,
      "objective/train/std_weight": 0.09266188740730286,
      "objective/train/theoretical_loss": 3.350949796420184,
      "objective/train/tokens_used": 2691052000,
      "theoretical_loss": 3.350949796420184,
      "tokens_seen": 2670592000
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.298420654435408e-05,
      "loss": 2.4259,
      "theoretical_loss": 3.350949796420184,
      "tokens_seen": 2670592000
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.29676565159826e-05,
      "loss": 1.6741,
      "theoretical_loss": 3.3509433235307586,
      "tokens_seen": 2670657536
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.295110648761112e-05,
      "loss": 2.3034,
      "theoretical_loss": 3.3509368508446453,
      "tokens_seen": 2670723072
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.293455645923964e-05,
      "loss": 2.7144,
      "theoretical_loss": 3.350930378361833,
      "tokens_seen": 2670788608
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.2918006430868166e-05,
      "loss": 2.3996,
      "theoretical_loss": 3.3509239060823095,
      "tokens_seen": 2670854144
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.290145640249669e-05,
      "loss": 2.4417,
      "theoretical_loss": 3.3509174340060643,
      "tokens_seen": 2670919680
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.2884906374125214e-05,
      "loss": 2.4414,
      "theoretical_loss": 3.3509109621330855,
      "tokens_seen": 2670985216
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.286835634575374e-05,
      "loss": 2.5731,
      "theoretical_loss": 3.350904490463362,
      "tokens_seen": 2671050752
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.285180631738226e-05,
      "loss": 1.5187,
      "theoretical_loss": 3.3508980189968818,
      "tokens_seen": 2671116288
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.283525628901078e-05,
      "loss": 3.6812,
      "theoretical_loss": 3.3508915477336343,
      "tokens_seen": 2671181824
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.28187062606393e-05,
      "loss": 1.7005,
      "theoretical_loss": 3.350885076673608,
      "tokens_seen": 2671247360
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.280215623226783e-05,
      "loss": 2.5249,
      "theoretical_loss": 3.350878605816791,
      "tokens_seen": 2671312896
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.278560620389635e-05,
      "loss": 2.4894,
      "theoretical_loss": 3.350872135163173,
      "tokens_seen": 2671378432
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.276905617552487e-05,
      "loss": 2.4562,
      "theoretical_loss": 3.3508656647127415,
      "tokens_seen": 2671443968
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.275250614715339e-05,
      "loss": 2.0376,
      "theoretical_loss": 3.3508591944654853,
      "tokens_seen": 2671509504
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.2735956118781916e-05,
      "loss": 2.652,
      "theoretical_loss": 3.3508527244213937,
      "tokens_seen": 2671575040
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.271940609041044e-05,
      "loss": 0.9419,
      "theoretical_loss": 3.350846254580455,
      "tokens_seen": 2671640576
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.2702856062038964e-05,
      "loss": 2.5206,
      "theoretical_loss": 3.3508397849426577,
      "tokens_seen": 2671706112
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.268630603366749e-05,
      "loss": 2.1818,
      "theoretical_loss": 3.3508333155079906,
      "tokens_seen": 2671771648
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.266975600529601e-05,
      "loss": 2.6704,
      "theoretical_loss": 3.3508268462764423,
      "tokens_seen": 2671837184
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.265320597692453e-05,
      "loss": 2.4529,
      "theoretical_loss": 3.3508203772480014,
      "tokens_seen": 2671902720
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.2636655948553054e-05,
      "loss": 1.7881,
      "theoretical_loss": 3.3508139084226563,
      "tokens_seen": 2671968256
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.262010592018158e-05,
      "loss": 2.4546,
      "theoretical_loss": 3.350807439800396,
      "tokens_seen": 2672033792
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.26035558918101e-05,
      "loss": 2.533,
      "theoretical_loss": 3.3508009713812092,
      "tokens_seen": 2672099328
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.258700586343862e-05,
      "loss": 2.2254,
      "theoretical_loss": 3.3507945031650843,
      "tokens_seen": 2672164864
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.257045583506714e-05,
      "loss": 1.9338,
      "theoretical_loss": 3.3507880351520103,
      "tokens_seen": 2672230400
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.255390580669567e-05,
      "loss": 2.3599,
      "theoretical_loss": 3.3507815673419756,
      "tokens_seen": 2672295936
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.253735577832419e-05,
      "loss": 2.4054,
      "theoretical_loss": 3.3507750997349683,
      "tokens_seen": 2672361472
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.2520805749952715e-05,
      "loss": 2.4283,
      "theoretical_loss": 3.350768632330978,
      "tokens_seen": 2672427008
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.250425572158124e-05,
      "loss": 2.1503,
      "theoretical_loss": 3.350762165129993,
      "tokens_seen": 2672492544
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.248770569320976e-05,
      "loss": 2.3502,
      "theoretical_loss": 3.3507556981320015,
      "tokens_seen": 2672558080
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.247115566483828e-05,
      "loss": 2.3074,
      "theoretical_loss": 3.3507492313369927,
      "tokens_seen": 2672623616
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.2454605636466804e-05,
      "loss": 2.5198,
      "theoretical_loss": 3.3507427647449552,
      "tokens_seen": 2672689152
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.243805560809533e-05,
      "loss": 2.6143,
      "theoretical_loss": 3.3507362983558773,
      "tokens_seen": 2672754688
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.242150557972385e-05,
      "loss": 1.4031,
      "theoretical_loss": 3.350729832169748,
      "tokens_seen": 2672820224
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.240495555135237e-05,
      "loss": 2.619,
      "theoretical_loss": 3.350723366186556,
      "tokens_seen": 2672885760
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.238840552298089e-05,
      "loss": -0.7099,
      "theoretical_loss": 3.3507169004062893,
      "tokens_seen": 2672951296
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.237185549460942e-05,
      "loss": 2.3741,
      "theoretical_loss": 3.3507104348289376,
      "tokens_seen": 2673016832
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.235530546623794e-05,
      "loss": 2.2771,
      "theoretical_loss": 3.350703969454489,
      "tokens_seen": 2673082368
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.2338755437866465e-05,
      "loss": 2.4828,
      "theoretical_loss": 3.350697504282932,
      "tokens_seen": 2673147904
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.232220540949499e-05,
      "loss": 2.2573,
      "theoretical_loss": 3.3506910393142553,
      "tokens_seen": 2673213440
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.230565538112351e-05,
      "loss": 2.3318,
      "theoretical_loss": 3.3506845745484477,
      "tokens_seen": 2673278976
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.228910535275203e-05,
      "loss": 2.4509,
      "theoretical_loss": 3.3506781099854983,
      "tokens_seen": 2673344512
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.2272555324380554e-05,
      "loss": 1.0251,
      "theoretical_loss": 3.3506716456253947,
      "tokens_seen": 2673410048
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.225600529600908e-05,
      "loss": 2.615,
      "theoretical_loss": 3.3506651814681265,
      "tokens_seen": 2673475584
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.22394552676376e-05,
      "loss": 2.7214,
      "theoretical_loss": 3.350658717513682,
      "tokens_seen": 2673541120
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.222290523926612e-05,
      "loss": 2.2822,
      "theoretical_loss": 3.35065225376205,
      "tokens_seen": 2673606656
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.2206355210894643e-05,
      "loss": 2.4859,
      "theoretical_loss": 3.3506457902132185,
      "tokens_seen": 2673672192
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.218980518252317e-05,
      "loss": 2.566,
      "theoretical_loss": 3.3506393268671775,
      "tokens_seen": 2673737728
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.217325515415169e-05,
      "loss": 1.5202,
      "theoretical_loss": 3.3506328637239147,
      "tokens_seen": 2673803264
    },
    {
      "epoch": 0.95,
      "objective/train/avg_token_score": 0.012939120642840862,
      "objective/train/avg_weight": 0.9847655296325684,
      "objective/train/docs_used": 1504654,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.5472798347473145,
      "objective/train/std_weight": 0.08949597924947739,
      "objective/train/theoretical_loss": 3.3506264007834186,
      "objective/train/tokens_used": 2694328800,
      "theoretical_loss": 3.3506264007834186,
      "tokens_seen": 2673868800
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.2156705125780215e-05,
      "loss": 2.3093,
      "theoretical_loss": 3.3506264007834186,
      "tokens_seen": 2673868800
    },
    {
      "epoch": 0.95,
      "learning_rate": 3.214015509740874e-05,
      "loss": 2.628,
      "theoretical_loss": 3.3506199380456785,
      "tokens_seen": 2673934336
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.212360506903726e-05,
      "loss": 2.7449,
      "theoretical_loss": 3.350613475510683,
      "tokens_seen": 2673999872
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.210705504066579e-05,
      "loss": 2.4169,
      "theoretical_loss": 3.3506070131784202,
      "tokens_seen": 2674065408
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.2090505012294305e-05,
      "loss": 1.7288,
      "theoretical_loss": 3.3506005510488794,
      "tokens_seen": 2674130944
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.207395498392283e-05,
      "loss": 0.8344,
      "theoretical_loss": 3.350594089122049,
      "tokens_seen": 2674196480
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.205740495555135e-05,
      "loss": 2.2241,
      "theoretical_loss": 3.3505876273979176,
      "tokens_seen": 2674262016
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.204085492717987e-05,
      "loss": 2.5292,
      "theoretical_loss": 3.350581165876474,
      "tokens_seen": 2674327552
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.2024304898808394e-05,
      "loss": 2.4657,
      "theoretical_loss": 3.350574704557707,
      "tokens_seen": 2674393088
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.200775487043692e-05,
      "loss": 2.6279,
      "theoretical_loss": 3.3505682434416046,
      "tokens_seen": 2674458624
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.199120484206544e-05,
      "loss": 2.3612,
      "theoretical_loss": 3.3505617825281564,
      "tokens_seen": 2674524160
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1974654813693966e-05,
      "loss": 2.6862,
      "theoretical_loss": 3.3505553218173505,
      "tokens_seen": 2674589696
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.195810478532249e-05,
      "loss": 2.3536,
      "theoretical_loss": 3.350548861309176,
      "tokens_seen": 2674655232
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1941554756951014e-05,
      "loss": 1.1147,
      "theoretical_loss": 3.3505424010036213,
      "tokens_seen": 2674720768
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.192500472857954e-05,
      "loss": 1.4375,
      "theoretical_loss": 3.350535940900675,
      "tokens_seen": 2674786304
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1908454700208055e-05,
      "loss": 1.5267,
      "theoretical_loss": 3.350529481000326,
      "tokens_seen": 2674851840
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.189190467183658e-05,
      "loss": 2.2188,
      "theoretical_loss": 3.3505230213025627,
      "tokens_seen": 2674917376
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.18753546434651e-05,
      "loss": 1.6871,
      "theoretical_loss": 3.3505165618073742,
      "tokens_seen": 2674982912
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.185880461509362e-05,
      "loss": 2.6498,
      "theoretical_loss": 3.350510102514749,
      "tokens_seen": 2675048448
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1842254586722144e-05,
      "loss": 3.1148,
      "theoretical_loss": 3.3505036434246755,
      "tokens_seen": 2675113984
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.182570455835067e-05,
      "loss": 2.1061,
      "theoretical_loss": 3.3504971845371427,
      "tokens_seen": 2675179520
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.18091545299792e-05,
      "loss": 1.6303,
      "theoretical_loss": 3.350490725852139,
      "tokens_seen": 2675245056
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1792604501607716e-05,
      "loss": 2.2178,
      "theoretical_loss": 3.3504842673696538,
      "tokens_seen": 2675310592
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.177605447323624e-05,
      "loss": 3.2255,
      "theoretical_loss": 3.350477809089675,
      "tokens_seen": 2675376128
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1759504444864764e-05,
      "loss": 2.7495,
      "theoretical_loss": 3.3504713510121915,
      "tokens_seen": 2675441664
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.174295441649329e-05,
      "loss": 2.0731,
      "theoretical_loss": 3.3504648931371923,
      "tokens_seen": 2675507200
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1726404388121805e-05,
      "loss": 1.7762,
      "theoretical_loss": 3.350458435464666,
      "tokens_seen": 2675572736
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.170985435975033e-05,
      "loss": 1.7352,
      "theoretical_loss": 3.3504519779946005,
      "tokens_seen": 2675638272
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.169330433137885e-05,
      "loss": 2.1335,
      "theoretical_loss": 3.3504455207269856,
      "tokens_seen": 2675703808
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.167675430300738e-05,
      "loss": 2.6578,
      "theoretical_loss": 3.3504390636618093,
      "tokens_seen": 2675769344
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1660204274635894e-05,
      "loss": 2.6893,
      "theoretical_loss": 3.350432606799061,
      "tokens_seen": 2675834880
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1643654246264425e-05,
      "loss": 2.6303,
      "theoretical_loss": 3.3504261501387282,
      "tokens_seen": 2675900416
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.162710421789295e-05,
      "loss": 2.5486,
      "theoretical_loss": 3.350419693680801,
      "tokens_seen": 2675965952
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1610554189521466e-05,
      "loss": 2.1331,
      "theoretical_loss": 3.350413237425267,
      "tokens_seen": 2676031488
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.159400416114999e-05,
      "loss": 3.2796,
      "theoretical_loss": 3.350406781372116,
      "tokens_seen": 2676097024
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1577454132778514e-05,
      "loss": 2.4208,
      "theoretical_loss": 3.3504003255213353,
      "tokens_seen": 2676162560
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.156090410440704e-05,
      "loss": 2.6508,
      "theoretical_loss": 3.3503938698729145,
      "tokens_seen": 2676228096
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1544354076035556e-05,
      "loss": 2.1256,
      "theoretical_loss": 3.350387414426842,
      "tokens_seen": 2676293632
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.152780404766408e-05,
      "loss": 2.6858,
      "theoretical_loss": 3.350380959183107,
      "tokens_seen": 2676359168
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1511254019292604e-05,
      "loss": 2.5987,
      "theoretical_loss": 3.3503745041416977,
      "tokens_seen": 2676424704
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.149470399092113e-05,
      "loss": 2.958,
      "theoretical_loss": 3.350368049302603,
      "tokens_seen": 2676490240
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.147815396254965e-05,
      "loss": 1.314,
      "theoretical_loss": 3.350361594665811,
      "tokens_seen": 2676555776
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1461603934178176e-05,
      "loss": 2.7227,
      "theoretical_loss": 3.350355140231312,
      "tokens_seen": 2676621312
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.14450539058067e-05,
      "loss": 2.3807,
      "theoretical_loss": 3.3503486859990925,
      "tokens_seen": 2676686848
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.142850387743522e-05,
      "loss": 3.0347,
      "theoretical_loss": 3.350342231969143,
      "tokens_seen": 2676752384
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.141195384906374e-05,
      "loss": 2.498,
      "theoretical_loss": 3.3503357781414516,
      "tokens_seen": 2676817920
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1395403820692265e-05,
      "loss": 1.781,
      "theoretical_loss": 3.350329324516007,
      "tokens_seen": 2676883456
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.137885379232079e-05,
      "loss": 1.5854,
      "theoretical_loss": 3.3503228710927977,
      "tokens_seen": 2676948992
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1362303763949306e-05,
      "loss": 2.1955,
      "theoretical_loss": 3.350316417871813,
      "tokens_seen": 2677014528
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.134575373557783e-05,
      "loss": 2.106,
      "theoretical_loss": 3.350309964853041,
      "tokens_seen": 2677080064
    },
    {
      "epoch": 0.96,
      "objective/train/avg_token_score": 0.015003170818090439,
      "objective/train/avg_weight": 0.9823430776596069,
      "objective/train/docs_used": 1506560,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.458938121795654,
      "objective/train/std_weight": 0.09028258174657822,
      "objective/train/theoretical_loss": 3.3503035120364704,
      "objective/train/tokens_used": 2697605600,
      "theoretical_loss": 3.3503035120364704,
      "tokens_seen": 2677145600
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1329203707206354e-05,
      "loss": 2.4388,
      "theoretical_loss": 3.3503035120364704,
      "tokens_seen": 2677145600
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.131265367883488e-05,
      "loss": 2.3046,
      "theoretical_loss": 3.35029705942209,
      "tokens_seen": 2677211136
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.12961036504634e-05,
      "loss": 2.3203,
      "theoretical_loss": 3.350290607009889,
      "tokens_seen": 2677276672
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1279553622091926e-05,
      "loss": 3.9621,
      "theoretical_loss": 3.350284154799856,
      "tokens_seen": 2677342208
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.126300359372045e-05,
      "loss": 2.1687,
      "theoretical_loss": 3.350277702791979,
      "tokens_seen": 2677407744
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1246453565348974e-05,
      "loss": 2.586,
      "theoretical_loss": 3.3502712509862476,
      "tokens_seen": 2677473280
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.122990353697749e-05,
      "loss": -0.2969,
      "theoretical_loss": 3.35026479938265,
      "tokens_seen": 2677538816
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1213353508606015e-05,
      "loss": 2.1026,
      "theoretical_loss": 3.3502583479811747,
      "tokens_seen": 2677604352
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.119680348023454e-05,
      "loss": 2.1425,
      "theoretical_loss": 3.3502518967818116,
      "tokens_seen": 2677669888
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1180253451863056e-05,
      "loss": 2.4337,
      "theoretical_loss": 3.350245445784548,
      "tokens_seen": 2677735424
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.116370342349158e-05,
      "loss": 2.3219,
      "theoretical_loss": 3.3502389949893736,
      "tokens_seen": 2677800960
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1147153395120104e-05,
      "loss": 2.7466,
      "theoretical_loss": 3.3502325443962766,
      "tokens_seen": 2677866496
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.113060336674863e-05,
      "loss": 2.5133,
      "theoretical_loss": 3.3502260940052455,
      "tokens_seen": 2677932032
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.111405333837715e-05,
      "loss": 2.5154,
      "theoretical_loss": 3.3502196438162697,
      "tokens_seen": 2677997568
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1097503310005676e-05,
      "loss": 2.4154,
      "theoretical_loss": 3.3502131938293376,
      "tokens_seen": 2678063104
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.10809532816342e-05,
      "loss": 2.4674,
      "theoretical_loss": 3.3502067440444376,
      "tokens_seen": 2678128640
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1064403253262724e-05,
      "loss": 2.564,
      "theoretical_loss": 3.350200294461559,
      "tokens_seen": 2678194176
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.104785322489124e-05,
      "loss": 1.3708,
      "theoretical_loss": 3.3501938450806907,
      "tokens_seen": 2678259712
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.1031303196519766e-05,
      "loss": 2.0696,
      "theoretical_loss": 3.3501873959018207,
      "tokens_seen": 2678325248
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.101475316814829e-05,
      "loss": 2.0707,
      "theoretical_loss": 3.3501809469249384,
      "tokens_seen": 2678390784
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.099820313977681e-05,
      "loss": 1.8713,
      "theoretical_loss": 3.350174498150032,
      "tokens_seen": 2678456320
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.098165311140533e-05,
      "loss": 2.2264,
      "theoretical_loss": 3.35016804957709,
      "tokens_seen": 2678521856
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0965103083033855e-05,
      "loss": 2.8435,
      "theoretical_loss": 3.350161601206102,
      "tokens_seen": 2678587392
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.094855305466238e-05,
      "loss": 1.887,
      "theoretical_loss": 3.3501551530370564,
      "tokens_seen": 2678652928
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.09320030262909e-05,
      "loss": 2.7171,
      "theoretical_loss": 3.3501487050699414,
      "tokens_seen": 2678718464
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.091545299791943e-05,
      "loss": 2.7488,
      "theoretical_loss": 3.3501422573047464,
      "tokens_seen": 2678784000
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.089890296954795e-05,
      "loss": 2.1802,
      "theoretical_loss": 3.35013580974146,
      "tokens_seen": 2678849536
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0882352941176475e-05,
      "loss": 1.7715,
      "theoretical_loss": 3.3501293623800708,
      "tokens_seen": 2678915072
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.086580291280499e-05,
      "loss": 1.5902,
      "theoretical_loss": 3.3501229152205676,
      "tokens_seen": 2678980608
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0849252884433516e-05,
      "loss": 2.7304,
      "theoretical_loss": 3.350116468262939,
      "tokens_seen": 2679046144
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.083270285606204e-05,
      "loss": 2.4157,
      "theoretical_loss": 3.350110021507174,
      "tokens_seen": 2679111680
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0816152827690564e-05,
      "loss": 2.7061,
      "theoretical_loss": 3.3501035749532613,
      "tokens_seen": 2679177216
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.079960279931908e-05,
      "loss": 1.7979,
      "theoretical_loss": 3.350097128601189,
      "tokens_seen": 2679242752
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0783052770947605e-05,
      "loss": 2.1543,
      "theoretical_loss": 3.350090682450947,
      "tokens_seen": 2679308288
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.076650274257613e-05,
      "loss": 2.7741,
      "theoretical_loss": 3.350084236502523,
      "tokens_seen": 2679373824
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.074995271420465e-05,
      "loss": 1.9934,
      "theoretical_loss": 3.3500777907559067,
      "tokens_seen": 2679439360
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.073340268583318e-05,
      "loss": 2.3227,
      "theoretical_loss": 3.350071345211086,
      "tokens_seen": 2679504896
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.07168526574617e-05,
      "loss": 2.0655,
      "theoretical_loss": 3.35006489986805,
      "tokens_seen": 2679570432
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0700302629090225e-05,
      "loss": 2.364,
      "theoretical_loss": 3.3500584547267875,
      "tokens_seen": 2679635968
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.068375260071874e-05,
      "loss": 2.6599,
      "theoretical_loss": 3.3500520097872872,
      "tokens_seen": 2679701504
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0667202572347266e-05,
      "loss": 1.8308,
      "theoretical_loss": 3.3500455650495375,
      "tokens_seen": 2679767040
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.065065254397579e-05,
      "loss": 4.1538,
      "theoretical_loss": 3.350039120513528,
      "tokens_seen": 2679832576
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0634102515604314e-05,
      "loss": 2.4629,
      "theoretical_loss": 3.3500326761792465,
      "tokens_seen": 2679898112
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.061755248723283e-05,
      "loss": 2.5958,
      "theoretical_loss": 3.350026232046682,
      "tokens_seen": 2679963648
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0601002458861355e-05,
      "loss": 1.8462,
      "theoretical_loss": 3.3500197881158242,
      "tokens_seen": 2680029184
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.058445243048988e-05,
      "loss": 3.2804,
      "theoretical_loss": 3.35001334438666,
      "tokens_seen": 2680094720
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0567902402118403e-05,
      "loss": 2.188,
      "theoretical_loss": 3.35000690085918,
      "tokens_seen": 2680160256
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.055135237374693e-05,
      "loss": 1.4268,
      "theoretical_loss": 3.3500004575333726,
      "tokens_seen": 2680225792
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.053480234537545e-05,
      "loss": 3.9632,
      "theoretical_loss": 3.3499940144092255,
      "tokens_seen": 2680291328
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0518252317003975e-05,
      "loss": 2.5885,
      "theoretical_loss": 3.349987571486728,
      "tokens_seen": 2680356864
    },
    {
      "epoch": 0.96,
      "objective/train/avg_token_score": 0.041676655411720276,
      "objective/train/avg_weight": 0.9509739875793457,
      "objective/train/docs_used": 1507767,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6341519355773926,
      "objective/train/std_weight": 0.1312674880027771,
      "objective/train/theoretical_loss": 3.349981128765869,
      "objective/train/tokens_used": 2700882400,
      "theoretical_loss": 3.349981128765869,
      "tokens_seen": 2680422400
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0501702288632493e-05,
      "loss": 2.5314,
      "theoretical_loss": 3.349981128765869,
      "tokens_seen": 2680422400
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0485152260261017e-05,
      "loss": 3.4169,
      "theoretical_loss": 3.3499746862466377,
      "tokens_seen": 2680487936
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.046860223188954e-05,
      "loss": 2.0845,
      "theoretical_loss": 3.3499682439290224,
      "tokens_seen": 2680553472
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0452052203518065e-05,
      "loss": 2.5657,
      "theoretical_loss": 3.349961801813011,
      "tokens_seen": 2680619008
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0435502175146585e-05,
      "loss": 2.8541,
      "theoretical_loss": 3.349955359898594,
      "tokens_seen": 2680684544
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.041895214677511e-05,
      "loss": 2.0994,
      "theoretical_loss": 3.349948918185759,
      "tokens_seen": 2680750080
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0402402118403633e-05,
      "loss": 2.5762,
      "theoretical_loss": 3.349942476674495,
      "tokens_seen": 2680815616
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0385852090032157e-05,
      "loss": 2.2054,
      "theoretical_loss": 3.3499360353647907,
      "tokens_seen": 2680881152
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0369302061660674e-05,
      "loss": 2.7324,
      "theoretical_loss": 3.349929594256635,
      "tokens_seen": 2680946688
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.03527520332892e-05,
      "loss": 2.517,
      "theoretical_loss": 3.3499231533500167,
      "tokens_seen": 2681012224
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0336202004917722e-05,
      "loss": 3.0525,
      "theoretical_loss": 3.3499167126449243,
      "tokens_seen": 2681077760
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0319651976546243e-05,
      "loss": 1.6787,
      "theoretical_loss": 3.349910272141347,
      "tokens_seen": 2681143296
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0303101948174767e-05,
      "loss": 0.6252,
      "theoretical_loss": 3.3499038318392733,
      "tokens_seen": 2681208832
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.028655191980329e-05,
      "loss": 2.5677,
      "theoretical_loss": 3.3498973917386925,
      "tokens_seen": 2681274368
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0270001891431815e-05,
      "loss": 2.29,
      "theoretical_loss": 3.3498909518395923,
      "tokens_seen": 2681339904
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0253451863060336e-05,
      "loss": 2.2251,
      "theoretical_loss": 3.349884512141962,
      "tokens_seen": 2681405440
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.023690183468886e-05,
      "loss": 2.5534,
      "theoretical_loss": 3.3498780726457906,
      "tokens_seen": 2681470976
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0220351806317384e-05,
      "loss": 3.2341,
      "theoretical_loss": 3.349871633351067,
      "tokens_seen": 2681536512
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0203801777945907e-05,
      "loss": 2.7747,
      "theoretical_loss": 3.349865194257779,
      "tokens_seen": 2681602048
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0187251749574425e-05,
      "loss": 2.5827,
      "theoretical_loss": 3.3498587553659167,
      "tokens_seen": 2681667584
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.017070172120295e-05,
      "loss": 2.2901,
      "theoretical_loss": 3.349852316675468,
      "tokens_seen": 2681733120
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0154151692831476e-05,
      "loss": 2.137,
      "theoretical_loss": 3.349845878186422,
      "tokens_seen": 2681798656
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0137601664459993e-05,
      "loss": 2.114,
      "theoretical_loss": 3.3498394398987674,
      "tokens_seen": 2681864192
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0121051636088517e-05,
      "loss": 2.246,
      "theoretical_loss": 3.3498330018124927,
      "tokens_seen": 2681929728
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.010450160771704e-05,
      "loss": 2.4755,
      "theoretical_loss": 3.349826563927587,
      "tokens_seen": 2681995264
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0087951579345565e-05,
      "loss": 2.9623,
      "theoretical_loss": 3.3498201262440395,
      "tokens_seen": 2682060800
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0071401550974086e-05,
      "loss": 2.063,
      "theoretical_loss": 3.349813688761838,
      "tokens_seen": 2682126336
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.005485152260261e-05,
      "loss": 2.6209,
      "theoretical_loss": 3.349807251480972,
      "tokens_seen": 2682191872
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0038301494231134e-05,
      "loss": 2.5485,
      "theoretical_loss": 3.34980081440143,
      "tokens_seen": 2682257408
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0021751465859658e-05,
      "loss": 3.0478,
      "theoretical_loss": 3.349794377523201,
      "tokens_seen": 2682322944
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.0005201437488175e-05,
      "loss": 1.9477,
      "theoretical_loss": 3.3497879408462734,
      "tokens_seen": 2682388480
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9988651409116702e-05,
      "loss": 2.0197,
      "theoretical_loss": 3.3497815043706365,
      "tokens_seen": 2682454016
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9972101380745226e-05,
      "loss": 0.8994,
      "theoretical_loss": 3.3497750680962786,
      "tokens_seen": 2682519552
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.995555135237375e-05,
      "loss": 2.746,
      "theoretical_loss": 3.349768632023189,
      "tokens_seen": 2682585088
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9939001324002268e-05,
      "loss": 1.2358,
      "theoretical_loss": 3.3497621961513557,
      "tokens_seen": 2682650624
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.992245129563079e-05,
      "loss": 2.4153,
      "theoretical_loss": 3.3497557604807686,
      "tokens_seen": 2682716160
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9905901267259316e-05,
      "loss": 2.1108,
      "theoretical_loss": 3.3497493250114156,
      "tokens_seen": 2682781696
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9889351238887836e-05,
      "loss": 2.7297,
      "theoretical_loss": 3.3497428897432857,
      "tokens_seen": 2682847232
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.987280121051636e-05,
      "loss": 2.1788,
      "theoretical_loss": 3.3497364546763677,
      "tokens_seen": 2682912768
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9856251182144884e-05,
      "loss": 2.3783,
      "theoretical_loss": 3.349730019810651,
      "tokens_seen": 2682978304
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9839701153773408e-05,
      "loss": 2.3187,
      "theoretical_loss": 3.3497235851461236,
      "tokens_seen": 2683043840
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.982315112540193e-05,
      "loss": 2.3004,
      "theoretical_loss": 3.3497171506827743,
      "tokens_seen": 2683109376
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9806601097030453e-05,
      "loss": 2.6703,
      "theoretical_loss": 3.349710716420592,
      "tokens_seen": 2683174912
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9790051068658977e-05,
      "loss": 2.1496,
      "theoretical_loss": 3.349704282359566,
      "tokens_seen": 2683240448
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.97735010402875e-05,
      "loss": 2.6766,
      "theoretical_loss": 3.3496978484996847,
      "tokens_seen": 2683305984
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9756951011916018e-05,
      "loss": 1.4303,
      "theoretical_loss": 3.349691414840937,
      "tokens_seen": 2683371520
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9740400983544542e-05,
      "loss": 2.4909,
      "theoretical_loss": 3.349684981383312,
      "tokens_seen": 2683437056
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9723850955173066e-05,
      "loss": 2.5744,
      "theoretical_loss": 3.349678548126797,
      "tokens_seen": 2683502592
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9707300926801587e-05,
      "loss": 2.2814,
      "theoretical_loss": 3.349672115071383,
      "tokens_seen": 2683568128
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.969075089843011e-05,
      "loss": 2.4983,
      "theoretical_loss": 3.349665682217058,
      "tokens_seen": 2683633664
    },
    {
      "epoch": 0.96,
      "objective/train/avg_token_score": 0.014631098136305809,
      "objective/train/avg_weight": 0.9828047752380371,
      "objective/train/docs_used": 1509839,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6503665447235107,
      "objective/train/std_weight": 0.0726546123623848,
      "objective/train/theoretical_loss": 3.3496592495638096,
      "objective/train/tokens_used": 2704159200,
      "theoretical_loss": 3.3496592495638096,
      "tokens_seen": 2683699200
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9674200870058635e-05,
      "loss": 2.6189,
      "theoretical_loss": 3.3496592495638096,
      "tokens_seen": 2683699200
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.965765084168716e-05,
      "loss": 2.6677,
      "theoretical_loss": 3.349652817111628,
      "tokens_seen": 2683764736
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.964110081331568e-05,
      "loss": 2.4854,
      "theoretical_loss": 3.3496463848605016,
      "tokens_seen": 2683830272
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9624550784944203e-05,
      "loss": 2.6829,
      "theoretical_loss": 3.349639952810419,
      "tokens_seen": 2683895808
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9608000756572727e-05,
      "loss": 2.6766,
      "theoretical_loss": 3.3496335209613695,
      "tokens_seen": 2683961344
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.959145072820125e-05,
      "loss": 1.5877,
      "theoretical_loss": 3.349627089313341,
      "tokens_seen": 2684026880
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.957490069982977e-05,
      "loss": 2.5773,
      "theoretical_loss": 3.3496206578663235,
      "tokens_seen": 2684092416
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9558350671458292e-05,
      "loss": 2.304,
      "theoretical_loss": 3.349614226620305,
      "tokens_seen": 2684157952
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9541800643086816e-05,
      "loss": 2.3048,
      "theoretical_loss": 3.3496077955752743,
      "tokens_seen": 2684223488
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.952525061471534e-05,
      "loss": 2.3309,
      "theoretical_loss": 3.3496013647312206,
      "tokens_seen": 2684289024
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.950870058634386e-05,
      "loss": 2.3852,
      "theoretical_loss": 3.349594934088133,
      "tokens_seen": 2684354560
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9492150557972385e-05,
      "loss": 2.6552,
      "theoretical_loss": 3.349588503645999,
      "tokens_seen": 2684420096
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.947560052960091e-05,
      "loss": 2.096,
      "theoretical_loss": 3.3495820734048087,
      "tokens_seen": 2684485632
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.945905050122943e-05,
      "loss": 2.1713,
      "theoretical_loss": 3.3495756433645507,
      "tokens_seen": 2684551168
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9442500472857954e-05,
      "loss": 2.2762,
      "theoretical_loss": 3.3495692135252133,
      "tokens_seen": 2684616704
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9425950444486477e-05,
      "loss": 2.3945,
      "theoretical_loss": 3.3495627838867854,
      "tokens_seen": 2684682240
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9409400416115e-05,
      "loss": 2.8391,
      "theoretical_loss": 3.3495563544492564,
      "tokens_seen": 2684747776
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.939285038774352e-05,
      "loss": 1.6932,
      "theoretical_loss": 3.3495499252126146,
      "tokens_seen": 2684813312
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9376300359372043e-05,
      "loss": 1.0768,
      "theoretical_loss": 3.3495434961768495,
      "tokens_seen": 2684878848
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9359750331000567e-05,
      "loss": 2.3849,
      "theoretical_loss": 3.3495370673419487,
      "tokens_seen": 2684944384
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.934320030262909e-05,
      "loss": 2.5724,
      "theoretical_loss": 3.3495306387079022,
      "tokens_seen": 2685009920
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.932665027425761e-05,
      "loss": 3.1218,
      "theoretical_loss": 3.349524210274698,
      "tokens_seen": 2685075456
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9310100245886135e-05,
      "loss": 2.1424,
      "theoretical_loss": 3.3495177820423256,
      "tokens_seen": 2685140992
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.929355021751466e-05,
      "loss": 1.6024,
      "theoretical_loss": 3.3495113540107733,
      "tokens_seen": 2685206528
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.927700018914318e-05,
      "loss": 2.3175,
      "theoretical_loss": 3.34950492618003,
      "tokens_seen": 2685272064
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9260450160771704e-05,
      "loss": 2.4148,
      "theoretical_loss": 3.3494984985500844,
      "tokens_seen": 2685337600
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9243900132400228e-05,
      "loss": 2.1921,
      "theoretical_loss": 3.349492071120926,
      "tokens_seen": 2685403136
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9227350104028752e-05,
      "loss": 1.6627,
      "theoretical_loss": 3.3494856438925433,
      "tokens_seen": 2685468672
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.921080007565727e-05,
      "loss": 2.6167,
      "theoretical_loss": 3.349479216864925,
      "tokens_seen": 2685534208
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9194250047285793e-05,
      "loss": 2.531,
      "theoretical_loss": 3.3494727900380594,
      "tokens_seen": 2685599744
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9177700018914317e-05,
      "loss": 2.5203,
      "theoretical_loss": 3.3494663634119366,
      "tokens_seen": 2685665280
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.916114999054284e-05,
      "loss": 1.716,
      "theoretical_loss": 3.3494599369865443,
      "tokens_seen": 2685730816
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.914459996217136e-05,
      "loss": 2.4141,
      "theoretical_loss": 3.349453510761872,
      "tokens_seen": 2685796352
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9128049933799886e-05,
      "loss": 2.2522,
      "theoretical_loss": 3.349447084737908,
      "tokens_seen": 2685861888
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.911149990542841e-05,
      "loss": 2.5194,
      "theoretical_loss": 3.3494406589146415,
      "tokens_seen": 2685927424
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.909494987705693e-05,
      "loss": 2.76,
      "theoretical_loss": 3.3494342332920612,
      "tokens_seen": 2685992960
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9078399848685454e-05,
      "loss": 2.3504,
      "theoretical_loss": 3.349427807870156,
      "tokens_seen": 2686058496
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9061849820313978e-05,
      "loss": 2.4897,
      "theoretical_loss": 3.3494213826489148,
      "tokens_seen": 2686124032
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9045299791942502e-05,
      "loss": 2.6797,
      "theoretical_loss": 3.3494149576283263,
      "tokens_seen": 2686189568
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.902874976357102e-05,
      "loss": 2.4754,
      "theoretical_loss": 3.3494085328083796,
      "tokens_seen": 2686255104
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.9012199735199543e-05,
      "loss": 2.3701,
      "theoretical_loss": 3.3494021081890635,
      "tokens_seen": 2686320640
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.899564970682807e-05,
      "loss": 0.3626,
      "theoretical_loss": 3.349395683770366,
      "tokens_seen": 2686386176
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8979099678456595e-05,
      "loss": 2.1799,
      "theoretical_loss": 3.349389259552277,
      "tokens_seen": 2686451712
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8962549650085112e-05,
      "loss": 1.8919,
      "theoretical_loss": 3.349382835534785,
      "tokens_seen": 2686517248
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8945999621713636e-05,
      "loss": 0.1769,
      "theoretical_loss": 3.349376411717879,
      "tokens_seen": 2686582784
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.892944959334216e-05,
      "loss": 2.2948,
      "theoretical_loss": 3.3493699881015475,
      "tokens_seen": 2686648320
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8912899564970684e-05,
      "loss": 1.7486,
      "theoretical_loss": 3.3493635646857793,
      "tokens_seen": 2686713856
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8896349536599205e-05,
      "loss": 2.4686,
      "theoretical_loss": 3.3493571414705636,
      "tokens_seen": 2686779392
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.887979950822773e-05,
      "loss": -0.8767,
      "theoretical_loss": 3.349350718455889,
      "tokens_seen": 2686844928
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8863249479856253e-05,
      "loss": 2.1106,
      "theoretical_loss": 3.3493442956417443,
      "tokens_seen": 2686910464
    },
    {
      "epoch": 0.96,
      "objective/train/avg_token_score": 0.008735906332731247,
      "objective/train/avg_weight": 0.9897167086601257,
      "objective/train/docs_used": 1511903,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.847769260406494,
      "objective/train/std_weight": 0.06517183780670166,
      "objective/train/theoretical_loss": 3.349337873028119,
      "objective/train/tokens_used": 2707436000,
      "theoretical_loss": 3.349337873028119,
      "tokens_seen": 2686976000
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.884669945148477e-05,
      "loss": 2.4874,
      "theoretical_loss": 3.349337873028119,
      "tokens_seen": 2686976000
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8830149423113297e-05,
      "loss": 1.6812,
      "theoretical_loss": 3.349331450615001,
      "tokens_seen": 2687041536
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.881359939474182e-05,
      "loss": 2.4901,
      "theoretical_loss": 3.3493250284023794,
      "tokens_seen": 2687107072
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8797049366370345e-05,
      "loss": 2.3196,
      "theoretical_loss": 3.349318606390244,
      "tokens_seen": 2687172608
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8780499337998862e-05,
      "loss": 2.5743,
      "theoretical_loss": 3.349312184578582,
      "tokens_seen": 2687238144
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8763949309627386e-05,
      "loss": 2.3811,
      "theoretical_loss": 3.3493057629673837,
      "tokens_seen": 2687303680
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.874739928125591e-05,
      "loss": 2.4371,
      "theoretical_loss": 3.3492993415566374,
      "tokens_seen": 2687369216
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8730849252884434e-05,
      "loss": 2.3881,
      "theoretical_loss": 3.3492929203463317,
      "tokens_seen": 2687434752
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8714299224512955e-05,
      "loss": 1.6678,
      "theoretical_loss": 3.349286499336456,
      "tokens_seen": 2687500288
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.869774919614148e-05,
      "loss": 2.4202,
      "theoretical_loss": 3.3492800785269985,
      "tokens_seen": 2687565824
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8681199167770003e-05,
      "loss": 2.7459,
      "theoretical_loss": 3.3492736579179487,
      "tokens_seen": 2687631360
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8664649139398524e-05,
      "loss": 2.3869,
      "theoretical_loss": 3.349267237509295,
      "tokens_seen": 2687696896
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8648099111027048e-05,
      "loss": 2.6208,
      "theoretical_loss": 3.3492608173010265,
      "tokens_seen": 2687762432
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.863154908265557e-05,
      "loss": 2.0142,
      "theoretical_loss": 3.349254397293132,
      "tokens_seen": 2687827968
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8614999054284095e-05,
      "loss": 2.7522,
      "theoretical_loss": 3.3492479774856005,
      "tokens_seen": 2687893504
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8598449025912613e-05,
      "loss": 1.5279,
      "theoretical_loss": 3.3492415578784205,
      "tokens_seen": 2687959040
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8581898997541137e-05,
      "loss": 2.233,
      "theoretical_loss": 3.349235138471581,
      "tokens_seen": 2688024576
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.856534896916966e-05,
      "loss": 3.3219,
      "theoretical_loss": 3.3492287192650716,
      "tokens_seen": 2688090112
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8548798940798185e-05,
      "loss": 2.8497,
      "theoretical_loss": 3.3492223002588797,
      "tokens_seen": 2688155648
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8532248912426705e-05,
      "loss": 2.4771,
      "theoretical_loss": 3.349215881452995,
      "tokens_seen": 2688221184
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.851569888405523e-05,
      "loss": 1.4825,
      "theoretical_loss": 3.3492094628474067,
      "tokens_seen": 2688286720
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8499148855683753e-05,
      "loss": 1.7491,
      "theoretical_loss": 3.3492030444421035,
      "tokens_seen": 2688352256
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8482598827312277e-05,
      "loss": 2.4726,
      "theoretical_loss": 3.349196626237074,
      "tokens_seen": 2688417792
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8466048798940798e-05,
      "loss": 2.4869,
      "theoretical_loss": 3.349190208232307,
      "tokens_seen": 2688483328
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8449498770569322e-05,
      "loss": 2.5367,
      "theoretical_loss": 3.3491837904277912,
      "tokens_seen": 2688548864
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8432948742197846e-05,
      "loss": 1.0371,
      "theoretical_loss": 3.349177372823516,
      "tokens_seen": 2688614400
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8416398713826363e-05,
      "loss": 2.5626,
      "theoretical_loss": 3.34917095541947,
      "tokens_seen": 2688679936
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8399848685454887e-05,
      "loss": -0.8105,
      "theoretical_loss": 3.3491645382156428,
      "tokens_seen": 2688745472
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.838329865708341e-05,
      "loss": 2.4609,
      "theoretical_loss": 3.349158121212022,
      "tokens_seen": 2688811008
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8366748628711935e-05,
      "loss": 2.3587,
      "theoretical_loss": 3.349151704408597,
      "tokens_seen": 2688876544
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8350198600340456e-05,
      "loss": 1.9528,
      "theoretical_loss": 3.349145287805357,
      "tokens_seen": 2688942080
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.833364857196898e-05,
      "loss": 2.5644,
      "theoretical_loss": 3.3491388714022907,
      "tokens_seen": 2689007616
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8317098543597504e-05,
      "loss": 1.6233,
      "theoretical_loss": 3.3491324551993866,
      "tokens_seen": 2689073152
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8300548515226028e-05,
      "loss": 2.3581,
      "theoretical_loss": 3.349126039196634,
      "tokens_seen": 2689138688
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8283998486854548e-05,
      "loss": 2.6199,
      "theoretical_loss": 3.3491196233940217,
      "tokens_seen": 2689204224
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8267448458483072e-05,
      "loss": 2.4654,
      "theoretical_loss": 3.349113207791538,
      "tokens_seen": 2689269760
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8250898430111596e-05,
      "loss": 3.0348,
      "theoretical_loss": 3.3491067923891733,
      "tokens_seen": 2689335296
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8234348401740113e-05,
      "loss": 2.5394,
      "theoretical_loss": 3.349100377186915,
      "tokens_seen": 2689400832
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8217798373368637e-05,
      "loss": 2.5721,
      "theoretical_loss": 3.3490939621847526,
      "tokens_seen": 2689466368
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.820124834499716e-05,
      "loss": 2.5426,
      "theoretical_loss": 3.3490875473826747,
      "tokens_seen": 2689531904
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8184698316625685e-05,
      "loss": 2.4249,
      "theoretical_loss": 3.3490811327806704,
      "tokens_seen": 2689597440
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8168148288254206e-05,
      "loss": 2.49,
      "theoretical_loss": 3.349074718378729,
      "tokens_seen": 2689662976
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.815159825988273e-05,
      "loss": 1.9751,
      "theoretical_loss": 3.3490683041768383,
      "tokens_seen": 2689728512
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8135048231511254e-05,
      "loss": 2.6397,
      "theoretical_loss": 3.349061890174988,
      "tokens_seen": 2689794048
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8118498203139778e-05,
      "loss": 2.4894,
      "theoretical_loss": 3.3490554763731666,
      "tokens_seen": 2689859584
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.81019481747683e-05,
      "loss": 2.7073,
      "theoretical_loss": 3.3490490627713636,
      "tokens_seen": 2689925120
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8085398146396823e-05,
      "loss": 2.0642,
      "theoretical_loss": 3.3490426493695673,
      "tokens_seen": 2689990656
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8068848118025347e-05,
      "loss": 2.5467,
      "theoretical_loss": 3.349036236167767,
      "tokens_seen": 2690056192
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.805229808965387e-05,
      "loss": 3.0052,
      "theoretical_loss": 3.349029823165951,
      "tokens_seen": 2690121728
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8035748061282388e-05,
      "loss": 2.1401,
      "theoretical_loss": 3.3490234103641088,
      "tokens_seen": 2690187264
    },
    {
      "epoch": 0.96,
      "objective/train/avg_token_score": 0.00910988450050354,
      "objective/train/avg_weight": 0.9892780780792236,
      "objective/train/docs_used": 1514263,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.3688480854034424,
      "objective/train/std_weight": 0.03889713063836098,
      "objective/train/theoretical_loss": 3.349016997762229,
      "objective/train/tokens_used": 2710712800,
      "theoretical_loss": 3.349016997762229,
      "tokens_seen": 2690252800
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8019198032910912e-05,
      "loss": 2.2886,
      "theoretical_loss": 3.349016997762229,
      "tokens_seen": 2690252800
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.8002648004539436e-05,
      "loss": 2.9377,
      "theoretical_loss": 3.3490105853603005,
      "tokens_seen": 2690318336
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7986097976167956e-05,
      "loss": 2.409,
      "theoretical_loss": 3.349004173158312,
      "tokens_seen": 2690383872
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.796954794779648e-05,
      "loss": 1.9998,
      "theoretical_loss": 3.348997761156253,
      "tokens_seen": 2690449408
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7952997919425004e-05,
      "loss": 2.5934,
      "theoretical_loss": 3.3489913493541117,
      "tokens_seen": 2690514944
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.793644789105353e-05,
      "loss": 1.7068,
      "theoretical_loss": 3.3489849377518777,
      "tokens_seen": 2690580480
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.791989786268205e-05,
      "loss": 2.4564,
      "theoretical_loss": 3.3489785263495393,
      "tokens_seen": 2690646016
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7903347834310573e-05,
      "loss": 1.5833,
      "theoretical_loss": 3.3489721151470855,
      "tokens_seen": 2690711552
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7886797805939097e-05,
      "loss": 3.061,
      "theoretical_loss": 3.3489657041445056,
      "tokens_seen": 2690777088
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.787024777756762e-05,
      "loss": 2.9474,
      "theoretical_loss": 3.348959293341788,
      "tokens_seen": 2690842624
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7853697749196138e-05,
      "loss": 2.1727,
      "theoretical_loss": 3.3489528827389217,
      "tokens_seen": 2690908160
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7837147720824665e-05,
      "loss": 2.4683,
      "theoretical_loss": 3.348946472335896,
      "tokens_seen": 2690973696
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.782059769245319e-05,
      "loss": 2.5283,
      "theoretical_loss": 3.3489400621326997,
      "tokens_seen": 2691039232
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7804047664081707e-05,
      "loss": 2.5131,
      "theoretical_loss": 3.348933652129321,
      "tokens_seen": 2691104768
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.778749763571023e-05,
      "loss": 1.8635,
      "theoretical_loss": 3.3489272423257495,
      "tokens_seen": 2691170304
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7770947607338755e-05,
      "loss": 2.2353,
      "theoretical_loss": 3.348920832721974,
      "tokens_seen": 2691235840
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.775439757896728e-05,
      "loss": 2.5538,
      "theoretical_loss": 3.3489144233179835,
      "tokens_seen": 2691301376
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.77378475505958e-05,
      "loss": 0.3136,
      "theoretical_loss": 3.348908014113767,
      "tokens_seen": 2691366912
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7721297522224323e-05,
      "loss": 2.1021,
      "theoretical_loss": 3.3489016051093126,
      "tokens_seen": 2691432448
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7704747493852847e-05,
      "loss": 2.1797,
      "theoretical_loss": 3.34889519630461,
      "tokens_seen": 2691497984
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.768819746548137e-05,
      "loss": 2.4339,
      "theoretical_loss": 3.3488887876996483,
      "tokens_seen": 2691563520
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7671647437109892e-05,
      "loss": 2.4187,
      "theoretical_loss": 3.348882379294415,
      "tokens_seen": 2691629056
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7655097408738416e-05,
      "loss": 2.2357,
      "theoretical_loss": 3.348875971088901,
      "tokens_seen": 2691694592
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.763854738036694e-05,
      "loss": 2.5537,
      "theoretical_loss": 3.3488695630830936,
      "tokens_seen": 2691760128
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7621997351995464e-05,
      "loss": 2.2494,
      "theoretical_loss": 3.348863155276983,
      "tokens_seen": 2691825664
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.760544732362398e-05,
      "loss": 2.0525,
      "theoretical_loss": 3.348856747670557,
      "tokens_seen": 2691891200
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7588897295252505e-05,
      "loss": 2.5791,
      "theoretical_loss": 3.3488503402638052,
      "tokens_seen": 2691956736
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.757234726688103e-05,
      "loss": 2.4384,
      "theoretical_loss": 3.348843933056716,
      "tokens_seen": 2692022272
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.755579723850955e-05,
      "loss": 2.5608,
      "theoretical_loss": 3.3488375260492784,
      "tokens_seen": 2692087808
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7539247210138074e-05,
      "loss": 2.5916,
      "theoretical_loss": 3.348831119241482,
      "tokens_seen": 2692153344
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7522697181766598e-05,
      "loss": 2.355,
      "theoretical_loss": 3.348824712633315,
      "tokens_seen": 2692218880
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.750614715339512e-05,
      "loss": 2.4717,
      "theoretical_loss": 3.348818306224767,
      "tokens_seen": 2692284416
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7489597125023642e-05,
      "loss": 2.6966,
      "theoretical_loss": 3.348811900015826,
      "tokens_seen": 2692349952
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7473047096652166e-05,
      "loss": 2.662,
      "theoretical_loss": 3.3488054940064815,
      "tokens_seen": 2692415488
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.745649706828069e-05,
      "loss": 2.2769,
      "theoretical_loss": 3.3487990881967225,
      "tokens_seen": 2692481024
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7439947039909214e-05,
      "loss": 2.3959,
      "theoretical_loss": 3.3487926825865375,
      "tokens_seen": 2692546560
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.742339701153773e-05,
      "loss": 1.9209,
      "theoretical_loss": 3.348786277175916,
      "tokens_seen": 2692612096
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7406846983166255e-05,
      "loss": 0.56,
      "theoretical_loss": 3.3487798719648465,
      "tokens_seen": 2692677632
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.739029695479478e-05,
      "loss": 3.4672,
      "theoretical_loss": 3.348773466953318,
      "tokens_seen": 2692743168
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.73737469264233e-05,
      "loss": 3.176,
      "theoretical_loss": 3.3487670621413193,
      "tokens_seen": 2692808704
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7357196898051824e-05,
      "loss": 2.3383,
      "theoretical_loss": 3.34876065752884,
      "tokens_seen": 2692874240
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7340646869680348e-05,
      "loss": 4.451,
      "theoretical_loss": 3.348754253115868,
      "tokens_seen": 2692939776
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7324096841308872e-05,
      "loss": 2.6779,
      "theoretical_loss": 3.3487478489023927,
      "tokens_seen": 2693005312
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7307546812937393e-05,
      "loss": 2.2446,
      "theoretical_loss": 3.348741444888403,
      "tokens_seen": 2693070848
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7290996784565917e-05,
      "loss": 2.7027,
      "theoretical_loss": 3.3487350410738883,
      "tokens_seen": 2693136384
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.727444675619444e-05,
      "loss": 2.2527,
      "theoretical_loss": 3.348728637458837,
      "tokens_seen": 2693201920
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7257896727822965e-05,
      "loss": 2.5321,
      "theoretical_loss": 3.3487222340432385,
      "tokens_seen": 2693267456
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7257896727822965e-05,
      "loss": 2.0465,
      "theoretical_loss": 3.3487158308270812,
      "tokens_seen": 2693332992
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7241346699451482e-05,
      "loss": 1.5747,
      "theoretical_loss": 3.3487094278103537,
      "tokens_seen": 2693398528
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7224796671080006e-05,
      "loss": 2.0578,
      "theoretical_loss": 3.3487030249930463,
      "tokens_seen": 2693464064
    },
    {
      "epoch": 0.96,
      "objective/train/avg_token_score": 0.012582475319504738,
      "objective/train/avg_weight": 0.9851948022842407,
      "objective/train/docs_used": 1516259,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.217837810516357,
      "objective/train/std_weight": 0.09302075952291489,
      "objective/train/theoretical_loss": 3.348696622375147,
      "objective/train/tokens_used": 2713989600,
      "theoretical_loss": 3.348696622375147,
      "tokens_seen": 2693529600
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.720824664270853e-05,
      "loss": 2.2636,
      "theoretical_loss": 3.348696622375147,
      "tokens_seen": 2693529600
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7191696614337054e-05,
      "loss": 2.4523,
      "theoretical_loss": 3.3486902199566444,
      "tokens_seen": 2693595136
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7175146585965574e-05,
      "loss": 1.8314,
      "theoretical_loss": 3.348683817737528,
      "tokens_seen": 2693660672
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.71585965575941e-05,
      "loss": 2.1499,
      "theoretical_loss": 3.3486774157177868,
      "tokens_seen": 2693726208
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7142046529222622e-05,
      "loss": 2.4944,
      "theoretical_loss": 3.34867101389741,
      "tokens_seen": 2693791744
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7125496500851143e-05,
      "loss": 4.2949,
      "theoretical_loss": 3.3486646122763855,
      "tokens_seen": 2693857280
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7108946472479667e-05,
      "loss": 2.4002,
      "theoretical_loss": 3.348658210854703,
      "tokens_seen": 2693922816
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.709239644410819e-05,
      "loss": 1.0305,
      "theoretical_loss": 3.348651809632351,
      "tokens_seen": 2693988352
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7075846415736715e-05,
      "loss": 2.4973,
      "theoretical_loss": 3.348645408609319,
      "tokens_seen": 2694053888
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7059296387365232e-05,
      "loss": 2.5723,
      "theoretical_loss": 3.348639007785596,
      "tokens_seen": 2694119424
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7042746358993756e-05,
      "loss": 2.7249,
      "theoretical_loss": 3.3486326071611705,
      "tokens_seen": 2694184960
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.702619633062228e-05,
      "loss": 2.4658,
      "theoretical_loss": 3.3486262067360317,
      "tokens_seen": 2694250496
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.7009646302250804e-05,
      "loss": 2.9617,
      "theoretical_loss": 3.348619806510168,
      "tokens_seen": 2694316032
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6993096273879325e-05,
      "loss": 0.3201,
      "theoretical_loss": 3.3486134064835693,
      "tokens_seen": 2694381568
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.697654624550785e-05,
      "loss": 2.4859,
      "theoretical_loss": 3.348607006656224,
      "tokens_seen": 2694447104
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6959996217136373e-05,
      "loss": 2.4139,
      "theoretical_loss": 3.3486006070281205,
      "tokens_seen": 2694512640
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6943446188764893e-05,
      "loss": 2.3127,
      "theoretical_loss": 3.3485942075992488,
      "tokens_seen": 2694578176
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6926896160393417e-05,
      "loss": 2.2904,
      "theoretical_loss": 3.3485878083695972,
      "tokens_seen": 2694643712
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.691034613202194e-05,
      "loss": 2.744,
      "theoretical_loss": 3.3485814093391553,
      "tokens_seen": 2694709248
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6893796103650465e-05,
      "loss": 1.5419,
      "theoretical_loss": 3.3485750105079113,
      "tokens_seen": 2694774784
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6877246075278982e-05,
      "loss": 2.3106,
      "theoretical_loss": 3.348568611875854,
      "tokens_seen": 2694840320
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6860696046907506e-05,
      "loss": 2.3653,
      "theoretical_loss": 3.3485622134429733,
      "tokens_seen": 2694905856
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.684414601853603e-05,
      "loss": 0.3151,
      "theoretical_loss": 3.3485558152092576,
      "tokens_seen": 2694971392
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6827595990164558e-05,
      "loss": 2.4383,
      "theoretical_loss": 3.348549417174696,
      "tokens_seen": 2695036928
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6811045961793075e-05,
      "loss": 1.5554,
      "theoretical_loss": 3.348543019339277,
      "tokens_seen": 2695102464
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.67944959334216e-05,
      "loss": 2.7593,
      "theoretical_loss": 3.3485366217029906,
      "tokens_seen": 2695168000
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6777945905050123e-05,
      "loss": 2.573,
      "theoretical_loss": 3.348530224265825,
      "tokens_seen": 2695233536
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6761395876678647e-05,
      "loss": 2.5626,
      "theoretical_loss": 3.3485238270277686,
      "tokens_seen": 2695299072
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6744845848307168e-05,
      "loss": 2.016,
      "theoretical_loss": 3.3485174299888114,
      "tokens_seen": 2695364608
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.672829581993569e-05,
      "loss": 1.8849,
      "theoretical_loss": 3.348511033148942,
      "tokens_seen": 2695430144
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6711745791564216e-05,
      "loss": 2.366,
      "theoretical_loss": 3.3485046365081494,
      "tokens_seen": 2695495680
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6695195763192733e-05,
      "loss": 2.5584,
      "theoretical_loss": 3.3484982400664225,
      "tokens_seen": 2695561216
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.667864573482126e-05,
      "loss": 2.3696,
      "theoretical_loss": 3.34849184382375,
      "tokens_seen": 2695626752
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6662095706449784e-05,
      "loss": 1.7166,
      "theoretical_loss": 3.3484854477801216,
      "tokens_seen": 2695692288
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6645545678078308e-05,
      "loss": 2.5152,
      "theoretical_loss": 3.3484790519355254,
      "tokens_seen": 2695757824
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6628995649706825e-05,
      "loss": 2.8527,
      "theoretical_loss": 3.3484726562899514,
      "tokens_seen": 2695823360
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.661244562133535e-05,
      "loss": 1.9119,
      "theoretical_loss": 3.348466260843387,
      "tokens_seen": 2695888896
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6595895592963873e-05,
      "loss": 2.6168,
      "theoretical_loss": 3.3484598655958226,
      "tokens_seen": 2695954432
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6579345564592397e-05,
      "loss": 2.1615,
      "theoretical_loss": 3.348453470547247,
      "tokens_seen": 2696019968
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6562795536220918e-05,
      "loss": 1.6748,
      "theoretical_loss": 3.3484470756976488,
      "tokens_seen": 2696085504
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6546245507849442e-05,
      "loss": 2.2719,
      "theoretical_loss": 3.3484406810470166,
      "tokens_seen": 2696151040
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6529695479477966e-05,
      "loss": 2.3312,
      "theoretical_loss": 3.34843428659534,
      "tokens_seen": 2696216576
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6513145451106487e-05,
      "loss": 2.3212,
      "theoretical_loss": 3.348427892342608,
      "tokens_seen": 2696282112
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.649659542273501e-05,
      "loss": 2.3744,
      "theoretical_loss": 3.348421498288809,
      "tokens_seen": 2696347648
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6480045394363535e-05,
      "loss": 2.2712,
      "theoretical_loss": 3.3484151044339328,
      "tokens_seen": 2696413184
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.646349536599206e-05,
      "loss": 2.7709,
      "theoretical_loss": 3.348408710777967,
      "tokens_seen": 2696478720
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6446945337620576e-05,
      "loss": 2.05,
      "theoretical_loss": 3.3484023173209025,
      "tokens_seen": 2696544256
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.64303953092491e-05,
      "loss": 0.3161,
      "theoretical_loss": 3.348395924062727,
      "tokens_seen": 2696609792
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6413845280877624e-05,
      "loss": 2.2869,
      "theoretical_loss": 3.3483895310034297,
      "tokens_seen": 2696675328
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6397295252506148e-05,
      "loss": 4.5276,
      "theoretical_loss": 3.3483831381429994,
      "tokens_seen": 2696740864
    },
    {
      "epoch": 0.96,
      "objective/train/avg_token_score": 0.015651635825634003,
      "objective/train/avg_weight": 0.9815731644630432,
      "objective/train/docs_used": 1518290,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 9.534809112548828,
      "objective/train/std_weight": 0.09632422029972076,
      "objective/train/theoretical_loss": 3.3483767454814255,
      "objective/train/tokens_used": 2717266400,
      "theoretical_loss": 3.3483767454814255,
      "tokens_seen": 2696806400
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.638074522413467e-05,
      "loss": 1.9616,
      "theoretical_loss": 3.3483767454814255,
      "tokens_seen": 2696806400
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6364195195763192e-05,
      "loss": 1.2076,
      "theoretical_loss": 3.3483703530186966,
      "tokens_seen": 2696871936
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6347645167391716e-05,
      "loss": 2.5272,
      "theoretical_loss": 3.348363960754802,
      "tokens_seen": 2696937472
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.633109513902024e-05,
      "loss": 2.9434,
      "theoretical_loss": 3.348357568689731,
      "tokens_seen": 2697003008
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.631454511064876e-05,
      "loss": 2.4449,
      "theoretical_loss": 3.3483511768234715,
      "tokens_seen": 2697068544
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6297995082277285e-05,
      "loss": 2.3669,
      "theoretical_loss": 3.348344785156013,
      "tokens_seen": 2697134080
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.628144505390581e-05,
      "loss": 2.5985,
      "theoretical_loss": 3.348338393687345,
      "tokens_seen": 2697199616
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6264895025534326e-05,
      "loss": 2.7875,
      "theoretical_loss": 3.3483320024174565,
      "tokens_seen": 2697265152
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.624834499716285e-05,
      "loss": 1.9049,
      "theoretical_loss": 3.3483256113463353,
      "tokens_seen": 2697330688
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6231794968791374e-05,
      "loss": 0.3975,
      "theoretical_loss": 3.3483192204739716,
      "tokens_seen": 2697396224
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6215244940419898e-05,
      "loss": 1.6047,
      "theoretical_loss": 3.3483128298003537,
      "tokens_seen": 2697461760
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.619869491204842e-05,
      "loss": 1.4772,
      "theoretical_loss": 3.348306439325471,
      "tokens_seen": 2697527296
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6182144883676943e-05,
      "loss": 2.3737,
      "theoretical_loss": 3.3483000490493127,
      "tokens_seen": 2697592832
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6165594855305467e-05,
      "loss": 2.3425,
      "theoretical_loss": 3.3482936589718673,
      "tokens_seen": 2697658368
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.614904482693399e-05,
      "loss": 1.5696,
      "theoretical_loss": 3.348287269093124,
      "tokens_seen": 2697723904
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.613249479856251e-05,
      "loss": 2.1874,
      "theoretical_loss": 3.348280879413071,
      "tokens_seen": 2697789440
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6115944770191035e-05,
      "loss": 2.5786,
      "theoretical_loss": 3.3482744899316987,
      "tokens_seen": 2697854976
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.609939474181956e-05,
      "loss": 1.8852,
      "theoretical_loss": 3.3482681006489954,
      "tokens_seen": 2697920512
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6082844713448076e-05,
      "loss": 2.5442,
      "theoretical_loss": 3.34826171156495,
      "tokens_seen": 2697986048
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.60662946850766e-05,
      "loss": 1.6645,
      "theoretical_loss": 3.3482553226795515,
      "tokens_seen": 2698051584
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6049744656705124e-05,
      "loss": 2.5819,
      "theoretical_loss": 3.348248933992789,
      "tokens_seen": 2698117120
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.603319462833365e-05,
      "loss": 2.2498,
      "theoretical_loss": 3.3482425455046516,
      "tokens_seen": 2698182656
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.601664459996217e-05,
      "loss": 2.5514,
      "theoretical_loss": 3.3482361572151285,
      "tokens_seen": 2698248192
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.6000094571590693e-05,
      "loss": 2.3116,
      "theoretical_loss": 3.3482297691242078,
      "tokens_seen": 2698313728
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5983544543219217e-05,
      "loss": 2.6387,
      "theoretical_loss": 3.34822338123188,
      "tokens_seen": 2698379264
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.596699451484774e-05,
      "loss": 2.1753,
      "theoretical_loss": 3.3482169935381325,
      "tokens_seen": 2698444800
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.595044448647626e-05,
      "loss": 2.3768,
      "theoretical_loss": 3.3482106060429553,
      "tokens_seen": 2698510336
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5933894458104786e-05,
      "loss": 1.6345,
      "theoretical_loss": 3.348204218746337,
      "tokens_seen": 2698575872
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.591734442973331e-05,
      "loss": 2.7642,
      "theoretical_loss": 3.348197831648267,
      "tokens_seen": 2698641408
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5900794401361834e-05,
      "loss": 2.5776,
      "theoretical_loss": 3.3481914447487338,
      "tokens_seen": 2698706944
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.588424437299035e-05,
      "loss": 2.5646,
      "theoretical_loss": 3.3481850580477266,
      "tokens_seen": 2698772480
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5867694344618875e-05,
      "loss": 2.3316,
      "theoretical_loss": 3.3481786715452344,
      "tokens_seen": 2698838016
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.58511443162474e-05,
      "loss": 2.3701,
      "theoretical_loss": 3.3481722852412465,
      "tokens_seen": 2698903552
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.583459428787592e-05,
      "loss": -0.5594,
      "theoretical_loss": 3.348165899135752,
      "tokens_seen": 2698969088
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5818044259504443e-05,
      "loss": 2.2493,
      "theoretical_loss": 3.348159513228739,
      "tokens_seen": 2699034624
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5801494231132967e-05,
      "loss": 2.3419,
      "theoretical_loss": 3.348153127520197,
      "tokens_seen": 2699100160
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.578494420276149e-05,
      "loss": 2.4009,
      "theoretical_loss": 3.348146742010116,
      "tokens_seen": 2699165696
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5768394174390012e-05,
      "loss": 1.7761,
      "theoretical_loss": 3.348140356698483,
      "tokens_seen": 2699231232
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5751844146018536e-05,
      "loss": 2.4229,
      "theoretical_loss": 3.348133971585289,
      "tokens_seen": 2699296768
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.573529411764706e-05,
      "loss": 2.3131,
      "theoretical_loss": 3.348127586670522,
      "tokens_seen": 2699362304
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5718744089275584e-05,
      "loss": 2.5779,
      "theoretical_loss": 3.348121201954171,
      "tokens_seen": 2699427840
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.57021940609041e-05,
      "loss": 2.319,
      "theoretical_loss": 3.348114817436225,
      "tokens_seen": 2699493376
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.568564403253263e-05,
      "loss": 2.1722,
      "theoretical_loss": 3.3481084331166735,
      "tokens_seen": 2699558912
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5669094004161153e-05,
      "loss": 2.4324,
      "theoretical_loss": 3.3481020489955053,
      "tokens_seen": 2699624448
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.565254397578967e-05,
      "loss": 2.5919,
      "theoretical_loss": 3.3480956650727096,
      "tokens_seen": 2699689984
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5635993947418194e-05,
      "loss": 2.0587,
      "theoretical_loss": 3.3480892813482743,
      "tokens_seen": 2699755520
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5619443919046718e-05,
      "loss": 2.9063,
      "theoretical_loss": 3.34808289782219,
      "tokens_seen": 2699821056
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5602893890675242e-05,
      "loss": 2.3738,
      "theoretical_loss": 3.348076514494445,
      "tokens_seen": 2699886592
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5586343862303762e-05,
      "loss": 6.1621,
      "theoretical_loss": 3.348070131365028,
      "tokens_seen": 2699952128
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5569793833932286e-05,
      "loss": 2.8146,
      "theoretical_loss": 3.348063748433929,
      "tokens_seen": 2700017664
    },
    {
      "epoch": 0.96,
      "objective/train/avg_token_score": 0.008578348904848099,
      "objective/train/avg_weight": 0.9899012446403503,
      "objective/train/docs_used": 1520209,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.485742568969727,
      "objective/train/std_weight": 0.0711994543671608,
      "objective/train/theoretical_loss": 3.3480573657011354,
      "objective/train/tokens_used": 2720543200,
      "theoretical_loss": 3.3480573657011354,
      "tokens_seen": 2700083200
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.555324380556081e-05,
      "loss": 2.7495,
      "theoretical_loss": 3.3480573657011354,
      "tokens_seen": 2700083200
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5536693777189334e-05,
      "loss": 3.0443,
      "theoretical_loss": 3.348050983166638,
      "tokens_seen": 2700148736
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5520143748817855e-05,
      "loss": 2.5059,
      "theoretical_loss": 3.3480446008304248,
      "tokens_seen": 2700214272
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.550359372044638e-05,
      "loss": 2.4578,
      "theoretical_loss": 3.348038218692485,
      "tokens_seen": 2700279808
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5487043692074903e-05,
      "loss": 2.6215,
      "theoretical_loss": 3.3480318367528077,
      "tokens_seen": 2700345344
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5470493663703427e-05,
      "loss": 1.7373,
      "theoretical_loss": 3.3480254550113817,
      "tokens_seen": 2700410880
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5453943635331944e-05,
      "loss": 0.7546,
      "theoretical_loss": 3.3480190734681967,
      "tokens_seen": 2700476416
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5437393606960468e-05,
      "loss": -1.0053,
      "theoretical_loss": 3.348012692123241,
      "tokens_seen": 2700541952
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5420843578588992e-05,
      "loss": 3.2707,
      "theoretical_loss": 3.348006310976504,
      "tokens_seen": 2700607488
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5404293550217513e-05,
      "loss": 2.3841,
      "theoretical_loss": 3.347999930027975,
      "tokens_seen": 2700673024
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5387743521846037e-05,
      "loss": 2.1827,
      "theoretical_loss": 3.3479935492776423,
      "tokens_seen": 2700738560
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.537119349347456e-05,
      "loss": 2.4749,
      "theoretical_loss": 3.347987168725495,
      "tokens_seen": 2700804096
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5354643465103085e-05,
      "loss": 1.8872,
      "theoretical_loss": 3.3479807883715234,
      "tokens_seen": 2700869632
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5338093436731605e-05,
      "loss": 2.2456,
      "theoretical_loss": 3.347974408215715,
      "tokens_seen": 2700935168
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.532154340836013e-05,
      "loss": 2.3932,
      "theoretical_loss": 3.3479680282580593,
      "tokens_seen": 2701000704
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5304993379988653e-05,
      "loss": 2.346,
      "theoretical_loss": 3.3479616484985457,
      "tokens_seen": 2701066240
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5288443351617177e-05,
      "loss": 2.629,
      "theoretical_loss": 3.347955268937163,
      "tokens_seen": 2701131776
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5271893323245694e-05,
      "loss": -3.6444,
      "theoretical_loss": 3.3479488895739,
      "tokens_seen": 2701197312
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.525534329487422e-05,
      "loss": -1.5367,
      "theoretical_loss": 3.347942510408746,
      "tokens_seen": 2701262848
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5238793266502742e-05,
      "loss": 2.2278,
      "theoretical_loss": 3.3479361314416907,
      "tokens_seen": 2701328384
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5222243238131263e-05,
      "loss": 2.496,
      "theoretical_loss": 3.347929752672722,
      "tokens_seen": 2701393920
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5205693209759787e-05,
      "loss": 2.2815,
      "theoretical_loss": 3.3479233741018293,
      "tokens_seen": 2701459456
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.518914318138831e-05,
      "loss": 2.5526,
      "theoretical_loss": 3.347916995729002,
      "tokens_seen": 2701524992
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5172593153016835e-05,
      "loss": 2.4125,
      "theoretical_loss": 3.347910617554229,
      "tokens_seen": 2701590528
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5156043124645356e-05,
      "loss": 2.5096,
      "theoretical_loss": 3.3479042395774985,
      "tokens_seen": 2701656064
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.513949309627388e-05,
      "loss": 2.23,
      "theoretical_loss": 3.347897861798801,
      "tokens_seen": 2701721600
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5122943067902404e-05,
      "loss": -0.3901,
      "theoretical_loss": 3.347891484218125,
      "tokens_seen": 2701787136
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5106393039530928e-05,
      "loss": 2.4662,
      "theoretical_loss": 3.3478851068354585,
      "tokens_seen": 2701852672
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.5089843011159445e-05,
      "loss": 1.911,
      "theoretical_loss": 3.3478787296507924,
      "tokens_seen": 2701918208
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.507329298278797e-05,
      "loss": 2.4962,
      "theoretical_loss": 3.347872352664114,
      "tokens_seen": 2701983744
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.5056742954416493e-05,
      "loss": 2.5909,
      "theoretical_loss": 3.347865975875414,
      "tokens_seen": 2702049280
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.5040192926045017e-05,
      "loss": 2.2189,
      "theoretical_loss": 3.34785959928468,
      "tokens_seen": 2702114816
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.5023642897673537e-05,
      "loss": 2.664,
      "theoretical_loss": 3.3478532228919016,
      "tokens_seen": 2702180352
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.500709286930206e-05,
      "loss": 2.562,
      "theoretical_loss": 3.347846846697068,
      "tokens_seen": 2702245888
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4990542840930585e-05,
      "loss": 0.2769,
      "theoretical_loss": 3.3478404707001683,
      "tokens_seen": 2702311424
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4973992812559106e-05,
      "loss": 2.6396,
      "theoretical_loss": 3.3478340949011915,
      "tokens_seen": 2702376960
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.495744278418763e-05,
      "loss": 2.5604,
      "theoretical_loss": 3.3478277193001267,
      "tokens_seen": 2702442496
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4940892755816154e-05,
      "loss": 2.3126,
      "theoretical_loss": 3.347821343896962,
      "tokens_seen": 2702508032
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4924342727444678e-05,
      "loss": 2.4917,
      "theoretical_loss": 3.3478149686916883,
      "tokens_seen": 2702573568
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4907792699073195e-05,
      "loss": 2.1406,
      "theoretical_loss": 3.3478085936842934,
      "tokens_seen": 2702639104
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.489124267070172e-05,
      "loss": 1.8353,
      "theoretical_loss": 3.347802218874766,
      "tokens_seen": 2702704640
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4874692642330243e-05,
      "loss": 2.2354,
      "theoretical_loss": 3.3477958442630964,
      "tokens_seen": 2702770176
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4858142613958767e-05,
      "loss": 2.255,
      "theoretical_loss": 3.347789469849273,
      "tokens_seen": 2702835712
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4841592585587288e-05,
      "loss": 2.4328,
      "theoretical_loss": 3.3477830956332846,
      "tokens_seen": 2702901248
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4825042557215812e-05,
      "loss": 2.624,
      "theoretical_loss": 3.3477767216151206,
      "tokens_seen": 2702966784
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4808492528844336e-05,
      "loss": 2.293,
      "theoretical_loss": 3.34777034779477,
      "tokens_seen": 2703032320
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4791942500472856e-05,
      "loss": 2.1,
      "theoretical_loss": 3.347763974172222,
      "tokens_seen": 2703097856
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.477539247210138e-05,
      "loss": 1.4304,
      "theoretical_loss": 3.3477576007474656,
      "tokens_seen": 2703163392
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4758842443729904e-05,
      "loss": 2.6535,
      "theoretical_loss": 3.3477512275204897,
      "tokens_seen": 2703228928
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4742292415358428e-05,
      "loss": 2.1813,
      "theoretical_loss": 3.3477448544912836,
      "tokens_seen": 2703294464
    },
    {
      "epoch": 0.97,
      "objective/train/avg_token_score": 0.022442694753408432,
      "objective/train/avg_weight": 0.9736106991767883,
      "objective/train/docs_used": 1521508,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.038576126098633,
      "objective/train/std_weight": 0.11040513217449188,
      "objective/train/theoretical_loss": 3.347738481659836,
      "objective/train/tokens_used": 2723820000,
      "theoretical_loss": 3.347738481659836,
      "tokens_seen": 2703360000
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4725742386986946e-05,
      "loss": 2.6975,
      "theoretical_loss": 3.347738481659836,
      "tokens_seen": 2703360000
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.470919235861547e-05,
      "loss": 1.7284,
      "theoretical_loss": 3.3477321090261363,
      "tokens_seen": 2703425536
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4692642330243993e-05,
      "loss": 2.5966,
      "theoretical_loss": 3.347725736590174,
      "tokens_seen": 2703491072
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.467609230187252e-05,
      "loss": 2.3794,
      "theoretical_loss": 3.347719364351937,
      "tokens_seen": 2703556608
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4659542273501038e-05,
      "loss": 1.2121,
      "theoretical_loss": 3.3477129923114153,
      "tokens_seen": 2703622144
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4642992245129562e-05,
      "loss": 2.2893,
      "theoretical_loss": 3.347706620468598,
      "tokens_seen": 2703687680
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4626442216758086e-05,
      "loss": 2.3765,
      "theoretical_loss": 3.3477002488234735,
      "tokens_seen": 2703753216
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.460989218838661e-05,
      "loss": 2.0853,
      "theoretical_loss": 3.347693877376031,
      "tokens_seen": 2703818752
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.459334216001513e-05,
      "loss": 2.057,
      "theoretical_loss": 3.3476875061262605,
      "tokens_seen": 2703884288
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4576792131643655e-05,
      "loss": 2.3555,
      "theoretical_loss": 3.3476811350741498,
      "tokens_seen": 2703949824
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.456024210327218e-05,
      "loss": 1.9387,
      "theoretical_loss": 3.347674764219689,
      "tokens_seen": 2704015360
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4543692074900696e-05,
      "loss": 2.5459,
      "theoretical_loss": 3.3476683935628664,
      "tokens_seen": 2704080896
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4527142046529223e-05,
      "loss": 2.2383,
      "theoretical_loss": 3.3476620231036716,
      "tokens_seen": 2704146432
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4510592018157747e-05,
      "loss": 2.8749,
      "theoretical_loss": 3.3476556528420938,
      "tokens_seen": 2704211968
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.449404198978627e-05,
      "loss": 2.1425,
      "theoretical_loss": 3.3476492827781215,
      "tokens_seen": 2704277504
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.447749196141479e-05,
      "loss": 2.3953,
      "theoretical_loss": 3.3476429129117444,
      "tokens_seen": 2704343040
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4460941933043312e-05,
      "loss": 2.8117,
      "theoretical_loss": 3.347636543242951,
      "tokens_seen": 2704408576
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4444391904671836e-05,
      "loss": 2.5665,
      "theoretical_loss": 3.347630173771731,
      "tokens_seen": 2704474112
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.442784187630036e-05,
      "loss": 2.3086,
      "theoretical_loss": 3.3476238044980726,
      "tokens_seen": 2704539648
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.441129184792888e-05,
      "loss": 2.5378,
      "theoretical_loss": 3.3476174354219657,
      "tokens_seen": 2704605184
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4394741819557405e-05,
      "loss": 2.4432,
      "theoretical_loss": 3.347611066543399,
      "tokens_seen": 2704670720
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.437819179118593e-05,
      "loss": 2.6048,
      "theoretical_loss": 3.347604697862362,
      "tokens_seen": 2704736256
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.436164176281445e-05,
      "loss": 2.3618,
      "theoretical_loss": 3.347598329378843,
      "tokens_seen": 2704801792
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4345091734442974e-05,
      "loss": 5.0908,
      "theoretical_loss": 3.347591961092832,
      "tokens_seen": 2704867328
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4328541706071498e-05,
      "loss": 1.0612,
      "theoretical_loss": 3.3475855930043172,
      "tokens_seen": 2704932864
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.431199167770002e-05,
      "loss": 2.0065,
      "theoretical_loss": 3.3475792251132885,
      "tokens_seen": 2704998400
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.429544164932854e-05,
      "loss": 2.4182,
      "theoretical_loss": 3.3475728574197348,
      "tokens_seen": 2705063936
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4278891620957063e-05,
      "loss": 2.4482,
      "theoretical_loss": 3.347566489923645,
      "tokens_seen": 2705129472
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4262341592585587e-05,
      "loss": 2.2193,
      "theoretical_loss": 3.347560122625008,
      "tokens_seen": 2705195008
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.424579156421411e-05,
      "loss": 1.9971,
      "theoretical_loss": 3.3475537555238133,
      "tokens_seen": 2705260544
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.422924153584263e-05,
      "loss": 2.0906,
      "theoretical_loss": 3.3475473886200495,
      "tokens_seen": 2705326080
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4212691507471155e-05,
      "loss": 3.2944,
      "theoretical_loss": 3.347541021913706,
      "tokens_seen": 2705391616
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.419614147909968e-05,
      "loss": 2.9577,
      "theoretical_loss": 3.3475346554047722,
      "tokens_seen": 2705457152
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4179591450728203e-05,
      "loss": 2.6739,
      "theoretical_loss": 3.347528289093237,
      "tokens_seen": 2705522688
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4163041422356724e-05,
      "loss": 2.5553,
      "theoretical_loss": 3.347521922979089,
      "tokens_seen": 2705588224
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4146491393985248e-05,
      "loss": 2.4215,
      "theoretical_loss": 3.347515557062318,
      "tokens_seen": 2705653760
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4129941365613772e-05,
      "loss": 2.4638,
      "theoretical_loss": 3.3475091913429127,
      "tokens_seen": 2705719296
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.411339133724229e-05,
      "loss": 2.8174,
      "theoretical_loss": 3.347502825820863,
      "tokens_seen": 2705784832
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4096841308870813e-05,
      "loss": 2.5956,
      "theoretical_loss": 3.3474964604961563,
      "tokens_seen": 2705850368
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4080291280499337e-05,
      "loss": 1.8425,
      "theoretical_loss": 3.3474900953687827,
      "tokens_seen": 2705915904
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.406374125212786e-05,
      "loss": 2.3467,
      "theoretical_loss": 3.3474837304387317,
      "tokens_seen": 2705981440
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4047191223756382e-05,
      "loss": 2.3529,
      "theoretical_loss": 3.347477365705992,
      "tokens_seen": 2706046976
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.4030641195384906e-05,
      "loss": 2.2603,
      "theoretical_loss": 3.3474710011705526,
      "tokens_seen": 2706112512
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.401409116701343e-05,
      "loss": 3.3768,
      "theoretical_loss": 3.3474646368324033,
      "tokens_seen": 2706178048
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3997541138641954e-05,
      "loss": 2.9332,
      "theoretical_loss": 3.347458272691532,
      "tokens_seen": 2706243584
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3980991110270474e-05,
      "loss": 2.6299,
      "theoretical_loss": 3.3474519087479284,
      "tokens_seen": 2706309120
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3964441081899e-05,
      "loss": 2.5412,
      "theoretical_loss": 3.347445545001582,
      "tokens_seen": 2706374656
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3947891053527522e-05,
      "loss": 2.0597,
      "theoretical_loss": 3.347439181452481,
      "tokens_seen": 2706440192
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.393134102515604e-05,
      "loss": 2.4724,
      "theoretical_loss": 3.3474328181006157,
      "tokens_seen": 2706505728
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3914790996784564e-05,
      "loss": 2.0789,
      "theoretical_loss": 3.3474264549459742,
      "tokens_seen": 2706571264
    },
    {
      "epoch": 0.97,
      "objective/train/avg_token_score": 0.017278797924518585,
      "objective/train/avg_weight": 0.9796614050865173,
      "objective/train/docs_used": 1523597,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.061285972595215,
      "objective/train/std_weight": 0.08276185393333435,
      "objective/train/theoretical_loss": 3.347420091988546,
      "objective/train/tokens_used": 2727096800,
      "theoretical_loss": 3.347420091988546,
      "tokens_seen": 2706636800
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3898240968413087e-05,
      "loss": 2.289,
      "theoretical_loss": 3.347420091988546,
      "tokens_seen": 2706636800
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.388169094004161e-05,
      "loss": 2.747,
      "theoretical_loss": 3.347413729228321,
      "tokens_seen": 2706702336
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3865140911670132e-05,
      "loss": 2.5358,
      "theoretical_loss": 3.3474073666652866,
      "tokens_seen": 2706767872
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3848590883298656e-05,
      "loss": 2.626,
      "theoretical_loss": 3.3474010042994333,
      "tokens_seen": 2706833408
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.383204085492718e-05,
      "loss": 1.74,
      "theoretical_loss": 3.347394642130749,
      "tokens_seen": 2706898944
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3815490826555704e-05,
      "loss": 0.3928,
      "theoretical_loss": 3.3473882801592243,
      "tokens_seen": 2706964480
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3798940798184225e-05,
      "loss": 2.3437,
      "theoretical_loss": 3.3473819183848477,
      "tokens_seen": 2707030016
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.378239076981275e-05,
      "loss": 2.5569,
      "theoretical_loss": 3.3473755568076076,
      "tokens_seen": 2707095552
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3765840741441273e-05,
      "loss": 2.1257,
      "theoretical_loss": 3.3473691954274942,
      "tokens_seen": 2707161088
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3749290713069797e-05,
      "loss": 2.6247,
      "theoretical_loss": 3.3473628342444957,
      "tokens_seen": 2707226624
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3732740684698314e-05,
      "loss": 2.1033,
      "theoretical_loss": 3.347356473258602,
      "tokens_seen": 2707292160
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3716190656326838e-05,
      "loss": 2.4555,
      "theoretical_loss": 3.347350112469802,
      "tokens_seen": 2707357696
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3699640627955362e-05,
      "loss": 4.1514,
      "theoretical_loss": 3.347343751878084,
      "tokens_seen": 2707423232
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3683090599583882e-05,
      "loss": -2.8161,
      "theoretical_loss": 3.3473373914834386,
      "tokens_seen": 2707488768
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3666540571212406e-05,
      "loss": 2.2486,
      "theoretical_loss": 3.3473310312858535,
      "tokens_seen": 2707554304
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.364999054284093e-05,
      "loss": -0.6687,
      "theoretical_loss": 3.3473246712853193,
      "tokens_seen": 2707619840
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3633440514469454e-05,
      "loss": 2.5554,
      "theoretical_loss": 3.347318311481824,
      "tokens_seen": 2707685376
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3616890486097975e-05,
      "loss": 1.774,
      "theoretical_loss": 3.3473119518753567,
      "tokens_seen": 2707750912
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.36003404577265e-05,
      "loss": 2.5117,
      "theoretical_loss": 3.347305592465907,
      "tokens_seen": 2707816448
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3583790429355023e-05,
      "loss": 2.4999,
      "theoretical_loss": 3.3472992332534637,
      "tokens_seen": 2707881984
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3567240400983547e-05,
      "loss": 2.6321,
      "theoretical_loss": 3.347292874238016,
      "tokens_seen": 2707947520
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3550690372612064e-05,
      "loss": 0.9779,
      "theoretical_loss": 3.3472865154195537,
      "tokens_seen": 2708013056
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3534140344240588e-05,
      "loss": 2.5423,
      "theoretical_loss": 3.3472801567980652,
      "tokens_seen": 2708078592
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3517590315869116e-05,
      "loss": 2.1017,
      "theoretical_loss": 3.34727379837354,
      "tokens_seen": 2708144128
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3501040287497633e-05,
      "loss": 2.0234,
      "theoretical_loss": 3.3472674401459663,
      "tokens_seen": 2708209664
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3484490259126157e-05,
      "loss": 2.3442,
      "theoretical_loss": 3.3472610821153346,
      "tokens_seen": 2708275200
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.346794023075468e-05,
      "loss": 1.6686,
      "theoretical_loss": 3.3472547242816333,
      "tokens_seen": 2708340736
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3451390202383205e-05,
      "loss": 2.6125,
      "theoretical_loss": 3.347248366644852,
      "tokens_seen": 2708406272
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3434840174011725e-05,
      "loss": 2.2992,
      "theoretical_loss": 3.3472420092049786,
      "tokens_seen": 2708471808
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.341829014564025e-05,
      "loss": 0.6817,
      "theoretical_loss": 3.347235651962004,
      "tokens_seen": 2708537344
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3401740117268773e-05,
      "loss": 3.0621,
      "theoretical_loss": 3.3472292949159157,
      "tokens_seen": 2708602880
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3385190088897297e-05,
      "loss": 2.4453,
      "theoretical_loss": 3.347222938066704,
      "tokens_seen": 2708668416
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3368640060525818e-05,
      "loss": 2.5264,
      "theoretical_loss": 3.3472165814143575,
      "tokens_seen": 2708733952
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3352090032154342e-05,
      "loss": 2.5389,
      "theoretical_loss": 3.3472102249588653,
      "tokens_seen": 2708799488
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3335540003782866e-05,
      "loss": 2.7663,
      "theoretical_loss": 3.347203868700217,
      "tokens_seen": 2708865024
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3318989975411383e-05,
      "loss": 2.9199,
      "theoretical_loss": 3.3471975126384015,
      "tokens_seen": 2708930560
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3302439947039907e-05,
      "loss": 3.5285,
      "theoretical_loss": 3.347191156773408,
      "tokens_seen": 2708996096
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.328588991866843e-05,
      "loss": 2.2098,
      "theoretical_loss": 3.347184801105225,
      "tokens_seen": 2709061632
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3269339890296955e-05,
      "loss": 2.2942,
      "theoretical_loss": 3.3471784456338427,
      "tokens_seen": 2709127168
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3252789861925476e-05,
      "loss": 2.1254,
      "theoretical_loss": 3.3471720903592495,
      "tokens_seen": 2709192704
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3236239833554e-05,
      "loss": 1.8582,
      "theoretical_loss": 3.3471657352814352,
      "tokens_seen": 2709258240
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3219689805182524e-05,
      "loss": 2.3251,
      "theoretical_loss": 3.347159380400388,
      "tokens_seen": 2709323776
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3203139776811048e-05,
      "loss": 1.9006,
      "theoretical_loss": 3.3471530257160977,
      "tokens_seen": 2709389312
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.318658974843957e-05,
      "loss": 1.5359,
      "theoretical_loss": 3.347146671228553,
      "tokens_seen": 2709454848
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3170039720068092e-05,
      "loss": 3.0328,
      "theoretical_loss": 3.347140316937744,
      "tokens_seen": 2709520384
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3153489691696616e-05,
      "loss": 1.9858,
      "theoretical_loss": 3.347133962843659,
      "tokens_seen": 2709585920
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.313693966332514e-05,
      "loss": 2.4273,
      "theoretical_loss": 3.3471276089462876,
      "tokens_seen": 2709651456
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3120389634953658e-05,
      "loss": 1.5516,
      "theoretical_loss": 3.3471212552456184,
      "tokens_seen": 2709716992
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.310383960658218e-05,
      "loss": 2.782,
      "theoretical_loss": 3.347114901741641,
      "tokens_seen": 2709782528
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3087289578210705e-05,
      "loss": 2.3745,
      "theoretical_loss": 3.347108548434344,
      "tokens_seen": 2709848064
    },
    {
      "epoch": 0.97,
      "objective/train/avg_token_score": 0.010975640267133713,
      "objective/train/avg_weight": 0.9870799779891968,
      "objective/train/docs_used": 1525609,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.741077899932861,
      "objective/train/std_weight": 0.0604972243309021,
      "objective/train/theoretical_loss": 3.3471021953237177,
      "objective/train/tokens_used": 2730373600,
      "theoretical_loss": 3.3471021953237177,
      "tokens_seen": 2709913600
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3070739549839226e-05,
      "loss": 3.3843,
      "theoretical_loss": 3.3471021953237177,
      "tokens_seen": 2709913600
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.305418952146775e-05,
      "loss": 2.4996,
      "theoretical_loss": 3.34709584240975,
      "tokens_seen": 2709979136
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3037639493096274e-05,
      "loss": 1.0985,
      "theoretical_loss": 3.347089489692431,
      "tokens_seen": 2710044672
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.3021089464724798e-05,
      "loss": 2.5884,
      "theoretical_loss": 3.347083137171749,
      "tokens_seen": 2710110208
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.300453943635332e-05,
      "loss": 2.2164,
      "theoretical_loss": 3.347076784847694,
      "tokens_seen": 2710175744
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2987989407981843e-05,
      "loss": 3.5088,
      "theoretical_loss": 3.3470704327202547,
      "tokens_seen": 2710241280
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2971439379610367e-05,
      "loss": 2.2473,
      "theoretical_loss": 3.3470640807894205,
      "tokens_seen": 2710306816
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.295488935123889e-05,
      "loss": 1.1807,
      "theoretical_loss": 3.34705772905518,
      "tokens_seen": 2710372352
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2938339322867408e-05,
      "loss": 2.4401,
      "theoretical_loss": 3.347051377517523,
      "tokens_seen": 2710437888
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2921789294495932e-05,
      "loss": 2.6653,
      "theoretical_loss": 3.3470450261764384,
      "tokens_seen": 2710503424
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2905239266124456e-05,
      "loss": 2.0758,
      "theoretical_loss": 3.3470386750319157,
      "tokens_seen": 2710568960
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2888689237752976e-05,
      "loss": 0.6794,
      "theoretical_loss": 3.3470323240839432,
      "tokens_seen": 2710634496
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.28721392093815e-05,
      "loss": 2.5207,
      "theoretical_loss": 3.347025973332511,
      "tokens_seen": 2710700032
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2855589181010024e-05,
      "loss": 0.4934,
      "theoretical_loss": 3.3470196227776077,
      "tokens_seen": 2710765568
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.283903915263855e-05,
      "loss": 2.2934,
      "theoretical_loss": 3.3470132724192228,
      "tokens_seen": 2710831104
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.282248912426707e-05,
      "loss": 2.363,
      "theoretical_loss": 3.3470069222573455,
      "tokens_seen": 2710896640
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2805939095895593e-05,
      "loss": 2.6212,
      "theoretical_loss": 3.3470005722919645,
      "tokens_seen": 2710962176
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2789389067524117e-05,
      "loss": 2.4395,
      "theoretical_loss": 3.346994222523069,
      "tokens_seen": 2711027712
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.277283903915264e-05,
      "loss": 2.0527,
      "theoretical_loss": 3.346987872950649,
      "tokens_seen": 2711093248
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2756289010781158e-05,
      "loss": 2.8349,
      "theoretical_loss": 3.3469815235746925,
      "tokens_seen": 2711158784
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2739738982409682e-05,
      "loss": 1.3816,
      "theoretical_loss": 3.34697517439519,
      "tokens_seen": 2711224320
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2723188954038206e-05,
      "loss": 2.5334,
      "theoretical_loss": 3.3469688254121293,
      "tokens_seen": 2711289856
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.270663892566673e-05,
      "loss": 3.1212,
      "theoretical_loss": 3.3469624766255004,
      "tokens_seen": 2711355392
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.269008889729525e-05,
      "loss": 2.4306,
      "theoretical_loss": 3.346956128035292,
      "tokens_seen": 2711420928
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2673538868923775e-05,
      "loss": 2.5062,
      "theoretical_loss": 3.3469497796414944,
      "tokens_seen": 2711486464
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.26569888405523e-05,
      "loss": 3.0835,
      "theoretical_loss": 3.3469434314440956,
      "tokens_seen": 2711552000
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.264043881218082e-05,
      "loss": 2.6537,
      "theoretical_loss": 3.346937083443085,
      "tokens_seen": 2711617536
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2623888783809343e-05,
      "loss": 1.9313,
      "theoretical_loss": 3.346930735638452,
      "tokens_seen": 2711683072
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2607338755437867e-05,
      "loss": 2.4934,
      "theoretical_loss": 3.3469243880301853,
      "tokens_seen": 2711748608
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.259078872706639e-05,
      "loss": 2.4244,
      "theoretical_loss": 3.3469180406182746,
      "tokens_seen": 2711814144
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.257423869869491e-05,
      "loss": 2.7249,
      "theoretical_loss": 3.3469116934027094,
      "tokens_seen": 2711879680
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2557688670323433e-05,
      "loss": 2.5453,
      "theoretical_loss": 3.346905346383478,
      "tokens_seen": 2711945216
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2541138641951957e-05,
      "loss": 2.5533,
      "theoretical_loss": 3.3468989995605702,
      "tokens_seen": 2712010752
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2524588613580484e-05,
      "loss": 0.8944,
      "theoretical_loss": 3.346892652933975,
      "tokens_seen": 2712076288
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2508038585209e-05,
      "loss": 2.5937,
      "theoretical_loss": 3.3468863065036816,
      "tokens_seen": 2712141824
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2491488556837525e-05,
      "loss": 2.4389,
      "theoretical_loss": 3.346879960269679,
      "tokens_seen": 2712207360
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.247493852846605e-05,
      "loss": -1.6245,
      "theoretical_loss": 3.3468736142319564,
      "tokens_seen": 2712272896
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.245838850009457e-05,
      "loss": 2.3647,
      "theoretical_loss": 3.346867268390503,
      "tokens_seen": 2712338432
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2441838471723094e-05,
      "loss": 2.5753,
      "theoretical_loss": 3.346860922745309,
      "tokens_seen": 2712403968
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2425288443351618e-05,
      "loss": 2.3033,
      "theoretical_loss": 3.3468545772963623,
      "tokens_seen": 2712469504
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2408738414980142e-05,
      "loss": 2.3528,
      "theoretical_loss": 3.3468482320436523,
      "tokens_seen": 2712535040
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.239218838660866e-05,
      "loss": 2.8671,
      "theoretical_loss": 3.3468418869871686,
      "tokens_seen": 2712600576
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2375638358237183e-05,
      "loss": 1.6474,
      "theoretical_loss": 3.3468355421269,
      "tokens_seen": 2712666112
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.235908832986571e-05,
      "loss": 2.1166,
      "theoretical_loss": 3.346829197462836,
      "tokens_seen": 2712731648
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2342538301494234e-05,
      "loss": 2.6372,
      "theoretical_loss": 3.3468228529949653,
      "tokens_seen": 2712797184
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.232598827312275e-05,
      "loss": 2.5526,
      "theoretical_loss": 3.3468165087232777,
      "tokens_seen": 2712862720
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2309438244751275e-05,
      "loss": 2.3701,
      "theoretical_loss": 3.3468101646477626,
      "tokens_seen": 2712928256
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.22928882163798e-05,
      "loss": 2.2506,
      "theoretical_loss": 3.3468038207684083,
      "tokens_seen": 2712993792
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2276338188008323e-05,
      "loss": 2.3039,
      "theoretical_loss": 3.3467974770852043,
      "tokens_seen": 2713059328
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2259788159636844e-05,
      "loss": 1.0724,
      "theoretical_loss": 3.34679113359814,
      "tokens_seen": 2713124864
    },
    {
      "epoch": 0.97,
      "objective/train/avg_token_score": 0.03382943198084831,
      "objective/train/avg_weight": 0.9602087140083313,
      "objective/train/docs_used": 1527506,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.290359020233154,
      "objective/train/std_weight": 0.12285007536411285,
      "objective/train/theoretical_loss": 3.346784790307205,
      "objective/train/tokens_used": 2733650400,
      "theoretical_loss": 3.346784790307205,
      "tokens_seen": 2713190400
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2243238131265368e-05,
      "loss": 2.3791,
      "theoretical_loss": 3.346784790307205,
      "tokens_seen": 2713190400
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2226688102893892e-05,
      "loss": 0.3244,
      "theoretical_loss": 3.3467784472123876,
      "tokens_seen": 2713255936
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2210138074522413e-05,
      "loss": 1.8424,
      "theoretical_loss": 3.3467721043136773,
      "tokens_seen": 2713321472
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2193588046150937e-05,
      "loss": 2.4463,
      "theoretical_loss": 3.346765761611064,
      "tokens_seen": 2713387008
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.217703801777946e-05,
      "loss": 1.7056,
      "theoretical_loss": 3.346759419104536,
      "tokens_seen": 2713452544
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2160487989407985e-05,
      "loss": 2.9002,
      "theoretical_loss": 3.3467530767940827,
      "tokens_seen": 2713518080
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2143937961036502e-05,
      "loss": 0.2858,
      "theoretical_loss": 3.346746734679694,
      "tokens_seen": 2713583616
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2127387932665026e-05,
      "loss": 1.9666,
      "theoretical_loss": 3.346740392761358,
      "tokens_seen": 2713649152
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.211083790429355e-05,
      "loss": 1.2955,
      "theoretical_loss": 3.3467340510390646,
      "tokens_seen": 2713714688
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2094287875922074e-05,
      "loss": 2.4313,
      "theoretical_loss": 3.346727709512803,
      "tokens_seen": 2713780224
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2077737847550594e-05,
      "loss": 2.5464,
      "theoretical_loss": 3.3467213681825623,
      "tokens_seen": 2713845760
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.206118781917912e-05,
      "loss": 2.7381,
      "theoretical_loss": 3.3467150270483312,
      "tokens_seen": 2713911296
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2044637790807642e-05,
      "loss": 2.4488,
      "theoretical_loss": 3.3467086861101,
      "tokens_seen": 2713976832
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2028087762436163e-05,
      "loss": 2.497,
      "theoretical_loss": 3.346702345367857,
      "tokens_seen": 2714042368
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.2011537734064687e-05,
      "loss": 1.4177,
      "theoretical_loss": 3.3466960048215917,
      "tokens_seen": 2714107904
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.199498770569321e-05,
      "loss": 2.717,
      "theoretical_loss": 3.3466896644712936,
      "tokens_seen": 2714173440
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1978437677321735e-05,
      "loss": 2.7707,
      "theoretical_loss": 3.3466833243169516,
      "tokens_seen": 2714238976
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1961887648950252e-05,
      "loss": 2.388,
      "theoretical_loss": 3.3466769843585547,
      "tokens_seen": 2714304512
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1945337620578776e-05,
      "loss": 0.673,
      "theoretical_loss": 3.346670644596092,
      "tokens_seen": 2714370048
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.19287875922073e-05,
      "loss": 1.8357,
      "theoretical_loss": 3.3466643050295537,
      "tokens_seen": 2714435584
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1912237563835824e-05,
      "loss": 2.3887,
      "theoretical_loss": 3.346657965658928,
      "tokens_seen": 2714501120
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1895687535464345e-05,
      "loss": 2.3712,
      "theoretical_loss": 3.3466516264842046,
      "tokens_seen": 2714566656
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.187913750709287e-05,
      "loss": 2.3012,
      "theoretical_loss": 3.346645287505373,
      "tokens_seen": 2714632192
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1862587478721393e-05,
      "loss": 2.413,
      "theoretical_loss": 3.3466389487224215,
      "tokens_seen": 2714697728
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1846037450349913e-05,
      "loss": 2.5406,
      "theoretical_loss": 3.34663261013534,
      "tokens_seen": 2714763264
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1829487421978437e-05,
      "loss": 1.1145,
      "theoretical_loss": 3.3466262717441175,
      "tokens_seen": 2714828800
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.181293739360696e-05,
      "loss": 1.8418,
      "theoretical_loss": 3.346619933548743,
      "tokens_seen": 2714894336
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1796387365235482e-05,
      "loss": 0.8248,
      "theoretical_loss": 3.346613595549207,
      "tokens_seen": 2714959872
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1779837336864006e-05,
      "loss": 2.4551,
      "theoretical_loss": 3.3466072577454966,
      "tokens_seen": 2715025408
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1763287308492527e-05,
      "loss": 1.0955,
      "theoretical_loss": 3.346600920137603,
      "tokens_seen": 2715090944
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.174673728012105e-05,
      "loss": 2.5722,
      "theoretical_loss": 3.346594582725514,
      "tokens_seen": 2715156480
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1730187251749575e-05,
      "loss": 2.1482,
      "theoretical_loss": 3.3465882455092197,
      "tokens_seen": 2715222016
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.17136372233781e-05,
      "loss": 2.3979,
      "theoretical_loss": 3.346581908488709,
      "tokens_seen": 2715287552
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.169708719500662e-05,
      "loss": 1.8286,
      "theoretical_loss": 3.3465755716639713,
      "tokens_seen": 2715353088
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1680537166635143e-05,
      "loss": 2.8981,
      "theoretical_loss": 3.3465692350349956,
      "tokens_seen": 2715418624
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1663987138263664e-05,
      "loss": 2.5562,
      "theoretical_loss": 3.3465628986017713,
      "tokens_seen": 2715484160
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1647437109892188e-05,
      "loss": 2.161,
      "theoretical_loss": 3.3465565623642872,
      "tokens_seen": 2715549696
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1630887081520712e-05,
      "loss": 2.5039,
      "theoretical_loss": 3.3465502263225333,
      "tokens_seen": 2715615232
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1614337053149236e-05,
      "loss": 1.6126,
      "theoretical_loss": 3.346543890476498,
      "tokens_seen": 2715680768
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1597787024777756e-05,
      "loss": 2.3482,
      "theoretical_loss": 3.346537554826171,
      "tokens_seen": 2715746304
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1581236996406277e-05,
      "loss": 2.4202,
      "theoretical_loss": 3.3465312193715415,
      "tokens_seen": 2715811840
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.15646869680348e-05,
      "loss": 1.6656,
      "theoretical_loss": 3.3465248841125987,
      "tokens_seen": 2715877376
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1548136939663325e-05,
      "loss": 2.5267,
      "theoretical_loss": 3.346518549049332,
      "tokens_seen": 2715942912
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.153158691129185e-05,
      "loss": 2.212,
      "theoretical_loss": 3.3465122141817303,
      "tokens_seen": 2716008448
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.151503688292037e-05,
      "loss": 2.3842,
      "theoretical_loss": 3.3465058795097833,
      "tokens_seen": 2716073984
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1498486854548893e-05,
      "loss": 2.3003,
      "theoretical_loss": 3.3464995450334794,
      "tokens_seen": 2716139520
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1481936826177414e-05,
      "loss": 2.558,
      "theoretical_loss": 3.3464932107528087,
      "tokens_seen": 2716205056
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1465386797805938e-05,
      "loss": 2.3529,
      "theoretical_loss": 3.3464868766677602,
      "tokens_seen": 2716270592
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1448836769434462e-05,
      "loss": -0.3898,
      "theoretical_loss": 3.346480542778323,
      "tokens_seen": 2716336128
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1432286741062986e-05,
      "loss": 2.3554,
      "theoretical_loss": 3.3464742090844863,
      "tokens_seen": 2716401664
    },
    {
      "epoch": 0.97,
      "objective/train/avg_token_score": 0.021883871406316757,
      "objective/train/avg_weight": 0.9742408990859985,
      "objective/train/docs_used": 1529013,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.770952224731445,
      "objective/train/std_weight": 0.09317489713430405,
      "objective/train/theoretical_loss": 3.346467875586239,
      "objective/train/tokens_used": 2736927200,
      "theoretical_loss": 3.346467875586239,
      "tokens_seen": 2716467200
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1415736712691507e-05,
      "loss": 3.3048,
      "theoretical_loss": 3.346467875586239,
      "tokens_seen": 2716467200
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.139918668432003e-05,
      "loss": 2.4541,
      "theoretical_loss": 3.3464615422835715,
      "tokens_seen": 2716532736
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.138263665594855e-05,
      "loss": 2.1875,
      "theoretical_loss": 3.346455209176472,
      "tokens_seen": 2716598272
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1366086627577075e-05,
      "loss": 2.4516,
      "theoretical_loss": 3.34644887626493,
      "tokens_seen": 2716663808
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.13495365992056e-05,
      "loss": 2.5336,
      "theoretical_loss": 3.346442543548935,
      "tokens_seen": 2716729344
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.133298657083412e-05,
      "loss": 2.4313,
      "theoretical_loss": 3.346436211028476,
      "tokens_seen": 2716794880
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1316436542462644e-05,
      "loss": 2.4506,
      "theoretical_loss": 3.346429878703542,
      "tokens_seen": 2716860416
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1299886514091164e-05,
      "loss": 2.5773,
      "theoretical_loss": 3.346423546574123,
      "tokens_seen": 2716925952
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1283336485719692e-05,
      "loss": 2.707,
      "theoretical_loss": 3.3464172146402076,
      "tokens_seen": 2716991488
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1266786457348212e-05,
      "loss": 1.2266,
      "theoretical_loss": 3.346410882901785,
      "tokens_seen": 2717057024
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1250236428976736e-05,
      "loss": 0.2772,
      "theoretical_loss": 3.346404551358845,
      "tokens_seen": 2717122560
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1233686400605257e-05,
      "loss": 2.0127,
      "theoretical_loss": 3.3463982200113764,
      "tokens_seen": 2717188096
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.121713637223378e-05,
      "loss": 3.9578,
      "theoretical_loss": 3.3463918888593684,
      "tokens_seen": 2717253632
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1200586343862305e-05,
      "loss": 3.2703,
      "theoretical_loss": 3.346385557902811,
      "tokens_seen": 2717319168
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.118403631549083e-05,
      "loss": 2.2458,
      "theoretical_loss": 3.3463792271416923,
      "tokens_seen": 2717384704
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.116748628711935e-05,
      "loss": -3.07,
      "theoretical_loss": 3.3463728965760025,
      "tokens_seen": 2717450240
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.115093625874787e-05,
      "loss": 2.7743,
      "theoretical_loss": 3.34636656620573,
      "tokens_seen": 2717515776
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1134386230376394e-05,
      "loss": 1.0003,
      "theoretical_loss": 3.3463602360308653,
      "tokens_seen": 2717581312
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1117836202004918e-05,
      "loss": 1.7366,
      "theoretical_loss": 3.3463539060513963,
      "tokens_seen": 2717646848
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1101286173633442e-05,
      "loss": 2.4966,
      "theoretical_loss": 3.346347576267313,
      "tokens_seen": 2717712384
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1084736145261963e-05,
      "loss": 2.7517,
      "theoretical_loss": 3.346341246678605,
      "tokens_seen": 2717777920
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1068186116890487e-05,
      "loss": 2.4597,
      "theoretical_loss": 3.346334917285261,
      "tokens_seen": 2717843456
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1051636088519007e-05,
      "loss": 2.3482,
      "theoretical_loss": 3.3463285880872697,
      "tokens_seen": 2717908992
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.103508606014753e-05,
      "loss": 2.5594,
      "theoretical_loss": 3.346322259084621,
      "tokens_seen": 2717974528
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.1018536031776055e-05,
      "loss": 2.5402,
      "theoretical_loss": 3.3463159302773047,
      "tokens_seen": 2718040064
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.100198600340458e-05,
      "loss": -1.0033,
      "theoretical_loss": 3.3463096016653093,
      "tokens_seen": 2718105600
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.09854359750331e-05,
      "loss": 1.4887,
      "theoretical_loss": 3.3463032732486244,
      "tokens_seen": 2718171136
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0968885946661624e-05,
      "loss": 0.9264,
      "theoretical_loss": 3.3462969450272393,
      "tokens_seen": 2718236672
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0952335918290145e-05,
      "loss": 1.6449,
      "theoretical_loss": 3.346290617001143,
      "tokens_seen": 2718302208
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.093578588991867e-05,
      "loss": 0.9209,
      "theoretical_loss": 3.3462842891703244,
      "tokens_seen": 2718367744
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0919235861547193e-05,
      "loss": 3.2226,
      "theoretical_loss": 3.3462779615347737,
      "tokens_seen": 2718433280
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0902685833175713e-05,
      "loss": 2.7709,
      "theoretical_loss": 3.34627163409448,
      "tokens_seen": 2718498816
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0886135804804237e-05,
      "loss": 1.0864,
      "theoretical_loss": 3.3462653068494315,
      "tokens_seen": 2718564352
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0869585776432758e-05,
      "loss": 2.488,
      "theoretical_loss": 3.346258979799619,
      "tokens_seen": 2718629888
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0853035748061282e-05,
      "loss": 2.8031,
      "theoretical_loss": 3.3462526529450303,
      "tokens_seen": 2718695424
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0836485719689806e-05,
      "loss": 2.5399,
      "theoretical_loss": 3.346246326285656,
      "tokens_seen": 2718760960
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.081993569131833e-05,
      "loss": 2.0912,
      "theoretical_loss": 3.3462399998214845,
      "tokens_seen": 2718826496
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.080338566294685e-05,
      "loss": 2.668,
      "theoretical_loss": 3.3462336735525056,
      "tokens_seen": 2718892032
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0786835634575374e-05,
      "loss": 2.7104,
      "theoretical_loss": 3.346227347478708,
      "tokens_seen": 2718957568
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0770285606203895e-05,
      "loss": 2.8367,
      "theoretical_loss": 3.3462210216000816,
      "tokens_seen": 2719023104
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.075373557783242e-05,
      "loss": 2.5093,
      "theoretical_loss": 3.346214695916615,
      "tokens_seen": 2719088640
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0737185549460943e-05,
      "loss": 2.7848,
      "theoretical_loss": 3.3462083704282977,
      "tokens_seen": 2719154176
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0720635521089463e-05,
      "loss": 2.7953,
      "theoretical_loss": 3.3462020451351195,
      "tokens_seen": 2719219712
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0704085492717987e-05,
      "loss": 1.8785,
      "theoretical_loss": 3.3461957200370693,
      "tokens_seen": 2719285248
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0687535464346508e-05,
      "loss": 2.5182,
      "theoretical_loss": 3.346189395134136,
      "tokens_seen": 2719350784
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0670985435975032e-05,
      "loss": 2.4559,
      "theoretical_loss": 3.3461830704263096,
      "tokens_seen": 2719416320
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0654435407603556e-05,
      "loss": 2.7089,
      "theoretical_loss": 3.3461767459135787,
      "tokens_seen": 2719481856
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.063788537923208e-05,
      "loss": 2.7836,
      "theoretical_loss": 3.346170421595933,
      "tokens_seen": 2719547392
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.06213353508606e-05,
      "loss": 2.4485,
      "theoretical_loss": 3.346164097473362,
      "tokens_seen": 2719612928
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0604785322489125e-05,
      "loss": 2.408,
      "theoretical_loss": 3.3461577735458548,
      "tokens_seen": 2719678464
    },
    {
      "epoch": 0.97,
      "objective/train/avg_token_score": 0.0413028746843338,
      "objective/train/avg_weight": 0.9513960480690002,
      "objective/train/docs_used": 1529588,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7023887634277344,
      "objective/train/std_weight": 0.1566411554813385,
      "objective/train/theoretical_loss": 3.3461514498133997,
      "objective/train/tokens_used": 2740204000,
      "theoretical_loss": 3.3461514498133997,
      "tokens_seen": 2719744000
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0588235294117645e-05,
      "loss": 2.4135,
      "theoretical_loss": 3.3461514498133997,
      "tokens_seen": 2719744000
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.057168526574617e-05,
      "loss": 1.3268,
      "theoretical_loss": 3.346145126275988,
      "tokens_seen": 2719809536
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0555135237374693e-05,
      "loss": 2.7561,
      "theoretical_loss": 3.3461388029336065,
      "tokens_seen": 2719875072
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0538585209003217e-05,
      "loss": 2.6194,
      "theoretical_loss": 3.3461324797862466,
      "tokens_seen": 2719940608
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0522035180631738e-05,
      "loss": 2.7163,
      "theoretical_loss": 3.346126156833897,
      "tokens_seen": 2720006144
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.050548515226026e-05,
      "loss": 1.4381,
      "theoretical_loss": 3.3461198340765463,
      "tokens_seen": 2720071680
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0488935123888782e-05,
      "loss": 2.4391,
      "theoretical_loss": 3.3461135115141842,
      "tokens_seen": 2720137216
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0472385095517306e-05,
      "loss": 1.4254,
      "theoretical_loss": 3.3461071891468004,
      "tokens_seen": 2720202752
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.045583506714583e-05,
      "loss": 2.8299,
      "theoretical_loss": 3.3461008669743837,
      "tokens_seen": 2720268288
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.043928503877435e-05,
      "loss": 1.8772,
      "theoretical_loss": 3.3460945449969235,
      "tokens_seen": 2720333824
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0422735010402875e-05,
      "loss": 2.6405,
      "theoretical_loss": 3.346088223214409,
      "tokens_seen": 2720399360
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0406184982031396e-05,
      "loss": 2.7922,
      "theoretical_loss": 3.34608190162683,
      "tokens_seen": 2720464896
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.038963495365992e-05,
      "loss": 2.4536,
      "theoretical_loss": 3.3460755802341753,
      "tokens_seen": 2720530432
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0373084925288444e-05,
      "loss": 2.5932,
      "theoretical_loss": 3.346069259036434,
      "tokens_seen": 2720595968
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0356534896916968e-05,
      "loss": 1.1818,
      "theoretical_loss": 3.346062938033596,
      "tokens_seen": 2720661504
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0339984868545488e-05,
      "loss": 1.8338,
      "theoretical_loss": 3.34605661722565,
      "tokens_seen": 2720727040
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.032343484017401e-05,
      "loss": 2.1742,
      "theoretical_loss": 3.346050296612586,
      "tokens_seen": 2720792576
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0306884811802533e-05,
      "loss": 2.7283,
      "theoretical_loss": 3.3460439761943928,
      "tokens_seen": 2720858112
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0290334783431057e-05,
      "loss": 2.8373,
      "theoretical_loss": 3.3460376559710596,
      "tokens_seen": 2720923648
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.027378475505958e-05,
      "loss": 2.3807,
      "theoretical_loss": 3.346031335942576,
      "tokens_seen": 2720989184
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.02572347266881e-05,
      "loss": 1.4251,
      "theoretical_loss": 3.346025016108931,
      "tokens_seen": 2721054720
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0240684698316625e-05,
      "loss": 2.2946,
      "theoretical_loss": 3.3460186964701144,
      "tokens_seen": 2721120256
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0224134669945146e-05,
      "loss": 2.4791,
      "theoretical_loss": 3.3460123770261148,
      "tokens_seen": 2721185792
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0207584641573673e-05,
      "loss": 1.9376,
      "theoretical_loss": 3.3460060577769224,
      "tokens_seen": 2721251328
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0191034613202194e-05,
      "loss": 2.4274,
      "theoretical_loss": 3.3459997387225258,
      "tokens_seen": 2721316864
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0174484584830718e-05,
      "loss": 7.4436,
      "theoretical_loss": 3.345993419862914,
      "tokens_seen": 2721382400
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.015793455645924e-05,
      "loss": 2.6641,
      "theoretical_loss": 3.3459871011980775,
      "tokens_seen": 2721447936
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0141384528087763e-05,
      "loss": 2.6979,
      "theoretical_loss": 3.3459807827280046,
      "tokens_seen": 2721513472
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0124834499716287e-05,
      "loss": 2.8537,
      "theoretical_loss": 3.345974464452685,
      "tokens_seen": 2721579008
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0108284471344807e-05,
      "loss": 1.9368,
      "theoretical_loss": 3.3459681463721083,
      "tokens_seen": 2721644544
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.009173444297333e-05,
      "loss": 1.5491,
      "theoretical_loss": 3.3459618284862627,
      "tokens_seen": 2721710080
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0075184414601852e-05,
      "loss": 1.7553,
      "theoretical_loss": 3.3459555107951386,
      "tokens_seen": 2721775616
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0058634386230376e-05,
      "loss": 2.4649,
      "theoretical_loss": 3.345949193298725,
      "tokens_seen": 2721841152
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.00420843578589e-05,
      "loss": 1.4035,
      "theoretical_loss": 3.3459428759970113,
      "tokens_seen": 2721906688
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0025534329487424e-05,
      "loss": 2.3853,
      "theoretical_loss": 3.3459365588899863,
      "tokens_seen": 2721972224
    },
    {
      "epoch": 0.97,
      "learning_rate": 2.0008984301115944e-05,
      "loss": 2.8738,
      "theoretical_loss": 3.34593024197764,
      "tokens_seen": 2722037760
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9992434272744468e-05,
      "loss": 1.3431,
      "theoretical_loss": 3.3459239252599615,
      "tokens_seen": 2722103296
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.997588424437299e-05,
      "loss": 2.6559,
      "theoretical_loss": 3.34591760873694,
      "tokens_seen": 2722168832
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9959334216001513e-05,
      "loss": 2.7636,
      "theoretical_loss": 3.3459112924085646,
      "tokens_seen": 2722234368
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9942784187630037e-05,
      "loss": 2.2308,
      "theoretical_loss": 3.345904976274825,
      "tokens_seen": 2722299904
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.992623415925856e-05,
      "loss": 1.8442,
      "theoretical_loss": 3.34589866033571,
      "tokens_seen": 2722365440
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.990968413088708e-05,
      "loss": 2.1538,
      "theoretical_loss": 3.34589234459121,
      "tokens_seen": 2722430976
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9893134102515602e-05,
      "loss": 4.6377,
      "theoretical_loss": 3.345886029041313,
      "tokens_seen": 2722496512
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9876584074144126e-05,
      "loss": 2.3252,
      "theoretical_loss": 3.345879713686009,
      "tokens_seen": 2722562048
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.986003404577265e-05,
      "loss": 2.6695,
      "theoretical_loss": 3.3458733985252875,
      "tokens_seen": 2722627584
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9843484017401174e-05,
      "loss": 2.4178,
      "theoretical_loss": 3.3458670835591375,
      "tokens_seen": 2722693120
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9826933989029695e-05,
      "loss": 1.8578,
      "theoretical_loss": 3.3458607687875483,
      "tokens_seen": 2722758656
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.981038396065822e-05,
      "loss": 2.1623,
      "theoretical_loss": 3.3458544542105093,
      "tokens_seen": 2722824192
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.979383393228674e-05,
      "loss": 3.729,
      "theoretical_loss": 3.3458481398280098,
      "tokens_seen": 2722889728
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9777283903915263e-05,
      "loss": 2.1765,
      "theoretical_loss": 3.345841825640039,
      "tokens_seen": 2722955264
    },
    {
      "epoch": 0.97,
      "objective/train/avg_token_score": 0.01714947447180748,
      "objective/train/avg_weight": 0.9798213243484497,
      "objective/train/docs_used": 1531474,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9805803298950195,
      "objective/train/std_weight": 0.0731111392378807,
      "objective/train/theoretical_loss": 3.345835511646587,
      "objective/train/tokens_used": 2743480800,
      "theoretical_loss": 3.345835511646587,
      "tokens_seen": 2723020800
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9760733875543787e-05,
      "loss": 2.5644,
      "theoretical_loss": 3.345835511646587,
      "tokens_seen": 2723020800
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.974418384717231e-05,
      "loss": 2.6096,
      "theoretical_loss": 3.345829197847642,
      "tokens_seen": 2723086336
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9727633818800832e-05,
      "loss": 2.4503,
      "theoretical_loss": 3.345822884243194,
      "tokens_seen": 2723151872
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9711083790429356e-05,
      "loss": 2.1077,
      "theoretical_loss": 3.3458165708332324,
      "tokens_seen": 2723217408
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9694533762057876e-05,
      "loss": 2.5655,
      "theoretical_loss": 3.345810257617746,
      "tokens_seen": 2723282944
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.96779837336864e-05,
      "loss": 2.5581,
      "theoretical_loss": 3.3458039445967245,
      "tokens_seen": 2723348480
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9661433705314924e-05,
      "loss": 2.5669,
      "theoretical_loss": 3.3457976317701568,
      "tokens_seen": 2723414016
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9644883676943445e-05,
      "loss": 2.3722,
      "theoretical_loss": 3.3457913191380326,
      "tokens_seen": 2723479552
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.962833364857197e-05,
      "loss": 2.1812,
      "theoretical_loss": 3.345785006700342,
      "tokens_seen": 2723545088
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.961178362020049e-05,
      "loss": 2.4039,
      "theoretical_loss": 3.345778694457073,
      "tokens_seen": 2723610624
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9595233591829014e-05,
      "loss": 2.6159,
      "theoretical_loss": 3.3457723824082155,
      "tokens_seen": 2723676160
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9578683563457538e-05,
      "loss": 2.3598,
      "theoretical_loss": 3.3457660705537586,
      "tokens_seen": 2723741696
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.956213353508606e-05,
      "loss": 2.3368,
      "theoretical_loss": 3.3457597588936925,
      "tokens_seen": 2723807232
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9545583506714582e-05,
      "loss": 4.1532,
      "theoretical_loss": 3.3457534474280055,
      "tokens_seen": 2723872768
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9529033478343106e-05,
      "loss": 2.5672,
      "theoretical_loss": 3.3457471361566875,
      "tokens_seen": 2723938304
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9512483449971627e-05,
      "loss": 2.469,
      "theoretical_loss": 3.3457408250797274,
      "tokens_seen": 2724003840
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.949593342160015e-05,
      "loss": 1.9031,
      "theoretical_loss": 3.345734514197115,
      "tokens_seen": 2724069376
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9479383393228675e-05,
      "loss": 1.9755,
      "theoretical_loss": 3.3457282035088394,
      "tokens_seen": 2724134912
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9462833364857195e-05,
      "loss": 2.5529,
      "theoretical_loss": 3.3457218930148898,
      "tokens_seen": 2724200448
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.944628333648572e-05,
      "loss": 2.5702,
      "theoretical_loss": 3.345715582715256,
      "tokens_seen": 2724265984
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.942973330811424e-05,
      "loss": 2.1649,
      "theoretical_loss": 3.345709272609927,
      "tokens_seen": 2724331520
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9413183279742764e-05,
      "loss": 2.1889,
      "theoretical_loss": 3.3457029626988923,
      "tokens_seen": 2724397056
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9396633251371288e-05,
      "loss": 2.376,
      "theoretical_loss": 3.345696652982141,
      "tokens_seen": 2724462592
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9380083222999812e-05,
      "loss": 2.2625,
      "theoretical_loss": 3.3456903434596628,
      "tokens_seen": 2724528128
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9363533194628333e-05,
      "loss": 2.5127,
      "theoretical_loss": 3.3456840341314464,
      "tokens_seen": 2724593664
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9346983166256857e-05,
      "loss": 2.6162,
      "theoretical_loss": 3.345677724997482,
      "tokens_seen": 2724659200
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9330433137885377e-05,
      "loss": 2.6432,
      "theoretical_loss": 3.3456714160577583,
      "tokens_seen": 2724724736
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.93138831095139e-05,
      "loss": 2.2448,
      "theoretical_loss": 3.345665107312265,
      "tokens_seen": 2724790272
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9297333081142425e-05,
      "loss": 2.3137,
      "theoretical_loss": 3.3456587987609914,
      "tokens_seen": 2724855808
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.928078305277095e-05,
      "loss": 2.2959,
      "theoretical_loss": 3.3456524904039266,
      "tokens_seen": 2724921344
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.926423302439947e-05,
      "loss": 2.4365,
      "theoretical_loss": 3.3456461822410604,
      "tokens_seen": 2724986880
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.924768299602799e-05,
      "loss": 2.264,
      "theoretical_loss": 3.345639874272382,
      "tokens_seen": 2725052416
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9231132967656514e-05,
      "loss": 2.35,
      "theoretical_loss": 3.3456335664978805,
      "tokens_seen": 2725117952
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9214582939285038e-05,
      "loss": 2.1112,
      "theoretical_loss": 3.345627258917545,
      "tokens_seen": 2725183488
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9198032910913562e-05,
      "loss": 2.0517,
      "theoretical_loss": 3.345620951531366,
      "tokens_seen": 2725249024
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9181482882542083e-05,
      "loss": 2.4944,
      "theoretical_loss": 3.3456146443393315,
      "tokens_seen": 2725314560
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9164932854170607e-05,
      "loss": 2.7402,
      "theoretical_loss": 3.3456083373414316,
      "tokens_seen": 2725380096
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9148382825799127e-05,
      "loss": 2.7449,
      "theoretical_loss": 3.3456020305376555,
      "tokens_seen": 2725445632
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9131832797427655e-05,
      "loss": 2.184,
      "theoretical_loss": 3.3455957239279925,
      "tokens_seen": 2725511168
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9115282769056175e-05,
      "loss": 2.2597,
      "theoretical_loss": 3.345589417512432,
      "tokens_seen": 2725576704
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.90987327406847e-05,
      "loss": 2.6101,
      "theoretical_loss": 3.345583111290964,
      "tokens_seen": 2725642240
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.908218271231322e-05,
      "loss": 2.4754,
      "theoretical_loss": 3.3455768052635766,
      "tokens_seen": 2725707776
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9065632683941744e-05,
      "loss": 2.9504,
      "theoretical_loss": 3.34557049943026,
      "tokens_seen": 2725773312
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9049082655570268e-05,
      "loss": 2.3799,
      "theoretical_loss": 3.3455641937910037,
      "tokens_seen": 2725838848
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.903253262719879e-05,
      "loss": 2.9372,
      "theoretical_loss": 3.3455578883457964,
      "tokens_seen": 2725904384
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.9015982598827313e-05,
      "loss": 2.5206,
      "theoretical_loss": 3.3455515830946276,
      "tokens_seen": 2725969920
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8999432570455833e-05,
      "loss": 2.5847,
      "theoretical_loss": 3.3455452780374872,
      "tokens_seen": 2726035456
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8982882542084357e-05,
      "loss": -0.624,
      "theoretical_loss": 3.345538973174364,
      "tokens_seen": 2726100992
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.896633251371288e-05,
      "loss": 2.2867,
      "theoretical_loss": 3.345532668505248,
      "tokens_seen": 2726166528
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8949782485341405e-05,
      "loss": 2.6868,
      "theoretical_loss": 3.345526364030128,
      "tokens_seen": 2726232064
    },
    {
      "epoch": 0.97,
      "objective/train/avg_token_score": 0.018288597464561462,
      "objective/train/avg_weight": 0.978481113910675,
      "objective/train/docs_used": 1533341,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.6067776679992676,
      "objective/train/std_weight": 0.07417289167642593,
      "objective/train/theoretical_loss": 3.345520059748993,
      "objective/train/tokens_used": 2746757600,
      "theoretical_loss": 3.345520059748993,
      "tokens_seen": 2726297600
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8933232456969926e-05,
      "loss": 3.0526,
      "theoretical_loss": 3.345520059748993,
      "tokens_seen": 2726297600
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.891668242859845e-05,
      "loss": 2.0069,
      "theoretical_loss": 3.3455137556618335,
      "tokens_seen": 2726363136
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.890013240022697e-05,
      "loss": 1.6017,
      "theoretical_loss": 3.345507451768638,
      "tokens_seen": 2726428672
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8883582371855494e-05,
      "loss": 2.6092,
      "theoretical_loss": 3.3455011480693964,
      "tokens_seen": 2726494208
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.886703234348402e-05,
      "loss": 2.1593,
      "theoretical_loss": 3.3454948445640973,
      "tokens_seen": 2726559744
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8850482315112542e-05,
      "loss": 2.6544,
      "theoretical_loss": 3.345488541252731,
      "tokens_seen": 2726625280
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8833932286741063e-05,
      "loss": 2.6074,
      "theoretical_loss": 3.3454822381352862,
      "tokens_seen": 2726690816
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8817382258369584e-05,
      "loss": 2.6576,
      "theoretical_loss": 3.345475935211753,
      "tokens_seen": 2726756352
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8800832229998108e-05,
      "loss": 1.3012,
      "theoretical_loss": 3.3454696324821196,
      "tokens_seen": 2726821888
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.878428220162663e-05,
      "loss": 2.7274,
      "theoretical_loss": 3.3454633299463765,
      "tokens_seen": 2726887424
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8767732173255156e-05,
      "loss": 2.3589,
      "theoretical_loss": 3.345457027604513,
      "tokens_seen": 2726952960
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8751182144883676e-05,
      "loss": 3.3003,
      "theoretical_loss": 3.3454507254565176,
      "tokens_seen": 2727018496
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.87346321165122e-05,
      "loss": 2.4198,
      "theoretical_loss": 3.3454444235023804,
      "tokens_seen": 2727084032
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.871808208814072e-05,
      "loss": 1.5507,
      "theoretical_loss": 3.3454381217420903,
      "tokens_seen": 2727149568
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8701532059769245e-05,
      "loss": 2.7835,
      "theoretical_loss": 3.3454318201756372,
      "tokens_seen": 2727215104
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.868498203139777e-05,
      "loss": 1.6796,
      "theoretical_loss": 3.34542551880301,
      "tokens_seen": 2727280640
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8668432003026293e-05,
      "loss": 0.9109,
      "theoretical_loss": 3.345419217624199,
      "tokens_seen": 2727346176
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8651881974654813e-05,
      "loss": 2.519,
      "theoretical_loss": 3.345412916639192,
      "tokens_seen": 2727411712
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8635331946283337e-05,
      "loss": 2.5439,
      "theoretical_loss": 3.34540661584798,
      "tokens_seen": 2727477248
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8618781917911858e-05,
      "loss": 2.4587,
      "theoretical_loss": 3.345400315250551,
      "tokens_seen": 2727542784
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8602231889540382e-05,
      "loss": 2.5587,
      "theoretical_loss": 3.3453940148468955,
      "tokens_seen": 2727608320
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8585681861168906e-05,
      "loss": 1.9415,
      "theoretical_loss": 3.3453877146370026,
      "tokens_seen": 2727673856
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8569131832797427e-05,
      "loss": 2.5865,
      "theoretical_loss": 3.345381414620861,
      "tokens_seen": 2727739392
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.855258180442595e-05,
      "loss": 2.3949,
      "theoretical_loss": 3.345375114798461,
      "tokens_seen": 2727804928
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.853603177605447e-05,
      "loss": 2.4145,
      "theoretical_loss": 3.345368815169792,
      "tokens_seen": 2727870464
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8519481747682995e-05,
      "loss": 2.5124,
      "theoretical_loss": 3.345362515734842,
      "tokens_seen": 2727936000
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.850293171931152e-05,
      "loss": 2.4405,
      "theoretical_loss": 3.345356216493602,
      "tokens_seen": 2728001536
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8486381690940043e-05,
      "loss": 2.1391,
      "theoretical_loss": 3.3453499174460606,
      "tokens_seen": 2728067072
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8469831662568564e-05,
      "loss": 1.3935,
      "theoretical_loss": 3.3453436185922074,
      "tokens_seen": 2728132608
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8453281634197088e-05,
      "loss": 2.5058,
      "theoretical_loss": 3.3453373199320318,
      "tokens_seen": 2728198144
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.843673160582561e-05,
      "loss": -0.0315,
      "theoretical_loss": 3.345331021465523,
      "tokens_seen": 2728263680
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8420181577454132e-05,
      "loss": 2.3952,
      "theoretical_loss": 3.3453247231926704,
      "tokens_seen": 2728329216
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8403631549082656e-05,
      "loss": 2.6119,
      "theoretical_loss": 3.345318425113464,
      "tokens_seen": 2728394752
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8387081520711177e-05,
      "loss": 2.5797,
      "theoretical_loss": 3.345312127227892,
      "tokens_seen": 2728460288
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.83705314923397e-05,
      "loss": 2.4239,
      "theoretical_loss": 3.3453058295359455,
      "tokens_seen": 2728525824
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.835398146396822e-05,
      "loss": 1.954,
      "theoretical_loss": 3.345299532037612,
      "tokens_seen": 2728591360
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8337431435596745e-05,
      "loss": 2.536,
      "theoretical_loss": 3.3452932347328823,
      "tokens_seen": 2728656896
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.832088140722527e-05,
      "loss": 2.5666,
      "theoretical_loss": 3.3452869376217453,
      "tokens_seen": 2728722432
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8304331378853793e-05,
      "loss": 1.9454,
      "theoretical_loss": 3.3452806407041904,
      "tokens_seen": 2728787968
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8287781350482314e-05,
      "loss": 2.5001,
      "theoretical_loss": 3.345274343980207,
      "tokens_seen": 2728853504
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8271231322110838e-05,
      "loss": 1.9734,
      "theoretical_loss": 3.3452680474497845,
      "tokens_seen": 2728919040
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.825468129373936e-05,
      "loss": 3.01,
      "theoretical_loss": 3.345261751112912,
      "tokens_seen": 2728984576
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8238131265367883e-05,
      "loss": 2.2142,
      "theoretical_loss": 3.3452554549695797,
      "tokens_seen": 2729050112
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8221581236996407e-05,
      "loss": 2.6325,
      "theoretical_loss": 3.345249159019776,
      "tokens_seen": 2729115648
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.820503120862493e-05,
      "loss": 2.3039,
      "theoretical_loss": 3.3452428632634916,
      "tokens_seen": 2729181184
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.818848118025345e-05,
      "loss": 2.9709,
      "theoretical_loss": 3.3452365677007143,
      "tokens_seen": 2729246720
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8171931151881972e-05,
      "loss": 2.5813,
      "theoretical_loss": 3.3452302723314347,
      "tokens_seen": 2729312256
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8155381123510496e-05,
      "loss": 2.4638,
      "theoretical_loss": 3.345223977155642,
      "tokens_seen": 2729377792
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.813883109513902e-05,
      "loss": 1.8435,
      "theoretical_loss": 3.3452176821733253,
      "tokens_seen": 2729443328
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8122281066767544e-05,
      "loss": 2.1975,
      "theoretical_loss": 3.345211387384474,
      "tokens_seen": 2729508864
    },
    {
      "epoch": 0.97,
      "objective/train/avg_token_score": 0.015583922155201435,
      "objective/train/avg_weight": 0.9816559553146362,
      "objective/train/docs_used": 1535248,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.857468843460083,
      "objective/train/std_weight": 0.0839780941605568,
      "objective/train/theoretical_loss": 3.3452050927890777,
      "objective/train/tokens_used": 2750034400,
      "theoretical_loss": 3.3452050927890777,
      "tokens_seen": 2729574400
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8105731038396064e-05,
      "loss": 2.5043,
      "theoretical_loss": 3.3452050927890777,
      "tokens_seen": 2729574400
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.808918101002459e-05,
      "loss": 2.7682,
      "theoretical_loss": 3.345198798387126,
      "tokens_seen": 2729639936
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.807263098165311e-05,
      "loss": 2.2278,
      "theoretical_loss": 3.345192504178608,
      "tokens_seen": 2729705472
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8056080953281636e-05,
      "loss": 2.6814,
      "theoretical_loss": 3.3451862101635133,
      "tokens_seen": 2729771008
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.8039530924910157e-05,
      "loss": 2.429,
      "theoretical_loss": 3.345179916341831,
      "tokens_seen": 2729836544
    },
    {
      "epoch": 0.97,
      "learning_rate": 1.802298089653868e-05,
      "loss": 2.5646,
      "theoretical_loss": 3.345173622713551,
      "tokens_seen": 2729902080
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.80064308681672e-05,
      "loss": 2.4783,
      "theoretical_loss": 3.345167329278662,
      "tokens_seen": 2729967616
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7989880839795726e-05,
      "loss": 2.6779,
      "theoretical_loss": 3.345161036037154,
      "tokens_seen": 2730033152
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.797333081142425e-05,
      "loss": 2.1905,
      "theoretical_loss": 3.3451547429890165,
      "tokens_seen": 2730098688
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.795678078305277e-05,
      "loss": 2.2151,
      "theoretical_loss": 3.3451484501342383,
      "tokens_seen": 2730164224
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7940230754681294e-05,
      "loss": 2.5779,
      "theoretical_loss": 3.34514215747281,
      "tokens_seen": 2730229760
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7923680726309815e-05,
      "loss": 2.3361,
      "theoretical_loss": 3.3451358650047194,
      "tokens_seen": 2730295296
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.790713069793834e-05,
      "loss": 2.6395,
      "theoretical_loss": 3.3451295727299573,
      "tokens_seen": 2730360832
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7890580669566863e-05,
      "loss": 2.1108,
      "theoretical_loss": 3.345123280648512,
      "tokens_seen": 2730426368
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7874030641195387e-05,
      "loss": 2.7152,
      "theoretical_loss": 3.345116988760374,
      "tokens_seen": 2730491904
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7857480612823907e-05,
      "loss": 2.2314,
      "theoretical_loss": 3.345110697065532,
      "tokens_seen": 2730557440
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.784093058445243e-05,
      "loss": 2.2916,
      "theoretical_loss": 3.3451044055639754,
      "tokens_seen": 2730622976
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7824380556080952e-05,
      "loss": 2.3518,
      "theoretical_loss": 3.345098114255694,
      "tokens_seen": 2730688512
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7807830527709476e-05,
      "loss": 0.6722,
      "theoretical_loss": 3.3450918231406774,
      "tokens_seen": 2730754048
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7791280499338e-05,
      "loss": 2.1075,
      "theoretical_loss": 3.3450855322189144,
      "tokens_seen": 2730819584
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7774730470966524e-05,
      "loss": 2.7362,
      "theoretical_loss": 3.345079241490395,
      "tokens_seen": 2730885120
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7758180442595045e-05,
      "loss": 0.111,
      "theoretical_loss": 3.345072950955108,
      "tokens_seen": 2730950656
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7741630414223565e-05,
      "loss": 2.5365,
      "theoretical_loss": 3.345066660613043,
      "tokens_seen": 2731016192
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.772508038585209e-05,
      "loss": 3.352,
      "theoretical_loss": 3.34506037046419,
      "tokens_seen": 2731081728
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7708530357480613e-05,
      "loss": 2.2056,
      "theoretical_loss": 3.345054080508538,
      "tokens_seen": 2731147264
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7691980329109137e-05,
      "loss": 1.9327,
      "theoretical_loss": 3.3450477907460767,
      "tokens_seen": 2731212800
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7675430300737658e-05,
      "loss": 1.9755,
      "theoretical_loss": 3.345041501176795,
      "tokens_seen": 2731278336
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7658880272366182e-05,
      "loss": 2.7429,
      "theoretical_loss": 3.3450352118006825,
      "tokens_seen": 2731343872
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7642330243994702e-05,
      "loss": 2.6761,
      "theoretical_loss": 3.345028922617729,
      "tokens_seen": 2731409408
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7625780215623226e-05,
      "loss": 2.7711,
      "theoretical_loss": 3.3450226336279236,
      "tokens_seen": 2731474944
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.760923018725175e-05,
      "loss": 0.965,
      "theoretical_loss": 3.3450163448312558,
      "tokens_seen": 2731540480
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7592680158880274e-05,
      "loss": 2.3973,
      "theoretical_loss": 3.345010056227715,
      "tokens_seen": 2731606016
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7576130130508795e-05,
      "loss": 2.4801,
      "theoretical_loss": 3.345003767817291,
      "tokens_seen": 2731671552
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.755958010213732e-05,
      "loss": 2.6449,
      "theoretical_loss": 3.3449974795999724,
      "tokens_seen": 2731737088
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.754303007376584e-05,
      "loss": 2.1426,
      "theoretical_loss": 3.3449911915757498,
      "tokens_seen": 2731802624
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7526480045394363e-05,
      "loss": 2.3648,
      "theoretical_loss": 3.3449849037446118,
      "tokens_seen": 2731868160
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7509930017022887e-05,
      "loss": 1.631,
      "theoretical_loss": 3.344978616106548,
      "tokens_seen": 2731933696
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7493379988651408e-05,
      "loss": 2.5716,
      "theoretical_loss": 3.344972328661548,
      "tokens_seen": 2731999232
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7476829960279932e-05,
      "loss": 0.2372,
      "theoretical_loss": 3.344966041409601,
      "tokens_seen": 2732064768
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7460279931908453e-05,
      "loss": 2.6214,
      "theoretical_loss": 3.3449597543506964,
      "tokens_seen": 2732130304
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7443729903536977e-05,
      "loss": 2.4468,
      "theoretical_loss": 3.344953467484824,
      "tokens_seen": 2732195840
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.74271798751655e-05,
      "loss": 1.8536,
      "theoretical_loss": 3.3449471808119733,
      "tokens_seen": 2732261376
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7410629846794025e-05,
      "loss": 2.5698,
      "theoretical_loss": 3.344940894332133,
      "tokens_seen": 2732326912
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7394079818422545e-05,
      "loss": 2.8075,
      "theoretical_loss": 3.3449346080452935,
      "tokens_seen": 2732392448
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.737752979005107e-05,
      "loss": 2.5048,
      "theoretical_loss": 3.3449283219514436,
      "tokens_seen": 2732457984
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.736097976167959e-05,
      "loss": 2.0595,
      "theoretical_loss": 3.344922036050573,
      "tokens_seen": 2732523520
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7344429733308114e-05,
      "loss": 2.856,
      "theoretical_loss": 3.344915750342671,
      "tokens_seen": 2732589056
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7327879704936638e-05,
      "loss": 3.5187,
      "theoretical_loss": 3.3449094648277273,
      "tokens_seen": 2732654592
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.731132967656516e-05,
      "loss": 1.4103,
      "theoretical_loss": 3.344903179505731,
      "tokens_seen": 2732720128
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7294779648193682e-05,
      "loss": 2.6161,
      "theoretical_loss": 3.344896894376672,
      "tokens_seen": 2732785664
    },
    {
      "epoch": 0.98,
      "objective/train/avg_token_score": 0.04129261150956154,
      "objective/train/avg_weight": 0.9513999223709106,
      "objective/train/docs_used": 1536975,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 8.93538761138916,
      "objective/train/std_weight": 0.15044286847114563,
      "objective/train/theoretical_loss": 3.3448906094405393,
      "objective/train/tokens_used": 2753311200,
      "theoretical_loss": 3.3448906094405393,
      "tokens_seen": 2732851200
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7278229619822203e-05,
      "loss": 2.9664,
      "theoretical_loss": 3.3448906094405393,
      "tokens_seen": 2732851200
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7261679591450727e-05,
      "loss": 2.4559,
      "theoretical_loss": 3.3448843246973228,
      "tokens_seen": 2732916736
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.724512956307925e-05,
      "loss": 1.545,
      "theoretical_loss": 3.3448780401470115,
      "tokens_seen": 2732982272
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7228579534707775e-05,
      "loss": 1.1656,
      "theoretical_loss": 3.344871755789595,
      "tokens_seen": 2733047808
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7212029506336296e-05,
      "loss": 2.5732,
      "theoretical_loss": 3.344865471625063,
      "tokens_seen": 2733113344
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.719547947796482e-05,
      "loss": 2.53,
      "theoretical_loss": 3.3448591876534044,
      "tokens_seen": 2733178880
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.717892944959334e-05,
      "loss": 1.0938,
      "theoretical_loss": 3.3448529038746093,
      "tokens_seen": 2733244416
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7162379421221864e-05,
      "loss": 1.7089,
      "theoretical_loss": 3.3448466202886666,
      "tokens_seen": 2733309952
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7145829392850388e-05,
      "loss": 0.8749,
      "theoretical_loss": 3.344840336895566,
      "tokens_seen": 2733375488
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7129279364478912e-05,
      "loss": 2.5288,
      "theoretical_loss": 3.3448340536952976,
      "tokens_seen": 2733441024
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7112729336107433e-05,
      "loss": 1.4651,
      "theoretical_loss": 3.3448277706878495,
      "tokens_seen": 2733506560
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7096179307735953e-05,
      "loss": 2.2059,
      "theoretical_loss": 3.3448214878732125,
      "tokens_seen": 2733572096
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7079629279364477e-05,
      "loss": 2.4924,
      "theoretical_loss": 3.344815205251375,
      "tokens_seen": 2733637632
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7063079250993e-05,
      "loss": 2.0149,
      "theoretical_loss": 3.3448089228223274,
      "tokens_seen": 2733703168
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7046529222621525e-05,
      "loss": 2.3227,
      "theoretical_loss": 3.344802640586058,
      "tokens_seen": 2733768704
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.7029979194250046e-05,
      "loss": 0.4409,
      "theoretical_loss": 3.3447963585425575,
      "tokens_seen": 2733834240
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.701342916587857e-05,
      "loss": 2.1679,
      "theoretical_loss": 3.344790076691815,
      "tokens_seen": 2733899776
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.699687913750709e-05,
      "loss": 2.3222,
      "theoretical_loss": 3.3447837950338193,
      "tokens_seen": 2733965312
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6980329109135618e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.3447775135685607,
      "tokens_seen": 2734030848
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.696377908076414e-05,
      "loss": 1.7008,
      "theoretical_loss": 3.344771232296028,
      "tokens_seen": 2734096384
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6947229052392663e-05,
      "loss": 2.2333,
      "theoretical_loss": 3.3447649512162108,
      "tokens_seen": 2734161920
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6930679024021183e-05,
      "loss": 2.536,
      "theoretical_loss": 3.3447586703290995,
      "tokens_seen": 2734227456
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6914128995649707e-05,
      "loss": 2.5059,
      "theoretical_loss": 3.344752389634682,
      "tokens_seen": 2734292992
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.689757896727823e-05,
      "loss": 1.6256,
      "theoretical_loss": 3.344746109132949,
      "tokens_seen": 2734358528
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6881028938906752e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.3447398288238896,
      "tokens_seen": 2734424064
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6864478910535276e-05,
      "loss": 2.2821,
      "theoretical_loss": 3.344733548707493,
      "tokens_seen": 2734489600
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6847928882163796e-05,
      "loss": 2.2553,
      "theoretical_loss": 3.3447272687837493,
      "tokens_seen": 2734555136
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.683137885379232e-05,
      "loss": 3.1646,
      "theoretical_loss": 3.344720989052647,
      "tokens_seen": 2734620672
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6814828825420844e-05,
      "loss": 2.3117,
      "theoretical_loss": 3.344714709514177,
      "tokens_seen": 2734686208
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6798278797049368e-05,
      "loss": 2.1844,
      "theoretical_loss": 3.344708430168327,
      "tokens_seen": 2734751744
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.678172876867789e-05,
      "loss": 2.2419,
      "theoretical_loss": 3.344702151015088,
      "tokens_seen": 2734817280
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6765178740306413e-05,
      "loss": 0.5649,
      "theoretical_loss": 3.3446958720544484,
      "tokens_seen": 2734882816
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6748628711934933e-05,
      "loss": 1.5269,
      "theoretical_loss": 3.3446895932863985,
      "tokens_seen": 2734948352
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6732078683563457e-05,
      "loss": 1.6985,
      "theoretical_loss": 3.3446833147109274,
      "tokens_seen": 2735013888
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.671552865519198e-05,
      "loss": 2.5012,
      "theoretical_loss": 3.3446770363280245,
      "tokens_seen": 2735079424
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6698978626820505e-05,
      "loss": 3.3367,
      "theoretical_loss": 3.3446707581376796,
      "tokens_seen": 2735144960
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6682428598449026e-05,
      "loss": 2.2098,
      "theoretical_loss": 3.344664480139882,
      "tokens_seen": 2735210496
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6665878570077547e-05,
      "loss": 2.2429,
      "theoretical_loss": 3.344658202334621,
      "tokens_seen": 2735276032
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.664932854170607e-05,
      "loss": 3.1108,
      "theoretical_loss": 3.344651924721886,
      "tokens_seen": 2735341568
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6632778513334595e-05,
      "loss": 2.7237,
      "theoretical_loss": 3.3446456473016672,
      "tokens_seen": 2735407104
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.661622848496312e-05,
      "loss": 2.7902,
      "theoretical_loss": 3.344639370073953,
      "tokens_seen": 2735472640
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.659967845659164e-05,
      "loss": 0.3796,
      "theoretical_loss": 3.344633093038734,
      "tokens_seen": 2735538176
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6583128428220163e-05,
      "loss": 2.2727,
      "theoretical_loss": 3.344626816195999,
      "tokens_seen": 2735603712
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6566578399848684e-05,
      "loss": 2.5297,
      "theoretical_loss": 3.344620539545738,
      "tokens_seen": 2735669248
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6550028371477208e-05,
      "loss": 1.9146,
      "theoretical_loss": 3.3446142630879394,
      "tokens_seen": 2735734784
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6533478343105732e-05,
      "loss": 3.0202,
      "theoretical_loss": 3.344607986822594,
      "tokens_seen": 2735800320
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6516928314734256e-05,
      "loss": 1.6616,
      "theoretical_loss": 3.3446017107496906,
      "tokens_seen": 2735865856
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6500378286362776e-05,
      "loss": 2.4135,
      "theoretical_loss": 3.344595434869219,
      "tokens_seen": 2735931392
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.64838282579913e-05,
      "loss": 2.4489,
      "theoretical_loss": 3.3445891591811683,
      "tokens_seen": 2735996928
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.646727822961982e-05,
      "loss": 0.6705,
      "theoretical_loss": 3.3445828836855283,
      "tokens_seen": 2736062464
    },
    {
      "epoch": 0.98,
      "objective/train/avg_token_score": 0.008229759521782398,
      "objective/train/avg_weight": 0.9903144240379333,
      "objective/train/docs_used": 1538829,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.9060709476470947,
      "objective/train/std_weight": 0.052204377949237823,
      "objective/train/theoretical_loss": 3.344576608382288,
      "objective/train/tokens_used": 2756588000,
      "theoretical_loss": 3.344576608382288,
      "tokens_seen": 2736128000
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6450728201248345e-05,
      "loss": 2.1057,
      "theoretical_loss": 3.344576608382288,
      "tokens_seen": 2736128000
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.643417817287687e-05,
      "loss": 2.7584,
      "theoretical_loss": 3.3445703332714376,
      "tokens_seen": 2736193536
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.641762814450539e-05,
      "loss": 2.5333,
      "theoretical_loss": 3.344564058352966,
      "tokens_seen": 2736259072
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6401078116133914e-05,
      "loss": 2.6926,
      "theoretical_loss": 3.3445577836268634,
      "tokens_seen": 2736324608
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6384528087762434e-05,
      "loss": 2.1421,
      "theoretical_loss": 3.344551509093119,
      "tokens_seen": 2736390144
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6367978059390958e-05,
      "loss": 3.0063,
      "theoretical_loss": 3.3445452347517217,
      "tokens_seen": 2736455680
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6351428031019482e-05,
      "loss": 2.1708,
      "theoretical_loss": 3.3445389606026614,
      "tokens_seen": 2736521216
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6334878002648006e-05,
      "loss": 2.4053,
      "theoretical_loss": 3.3445326866459277,
      "tokens_seen": 2736586752
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6318327974276527e-05,
      "loss": 2.5855,
      "theoretical_loss": 3.3445264128815104,
      "tokens_seen": 2736652288
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.630177794590505e-05,
      "loss": 2.2771,
      "theoretical_loss": 3.3445201393093984,
      "tokens_seen": 2736717824
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.628522791753357e-05,
      "loss": 2.665,
      "theoretical_loss": 3.344513865929582,
      "tokens_seen": 2736783360
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6268677889162095e-05,
      "loss": 2.5411,
      "theoretical_loss": 3.3445075927420493,
      "tokens_seen": 2736848896
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.625212786079062e-05,
      "loss": 4.1324,
      "theoretical_loss": 3.344501319746791,
      "tokens_seen": 2736914432
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.623557783241914e-05,
      "loss": -0.1821,
      "theoretical_loss": 3.3444950469437966,
      "tokens_seen": 2736979968
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6219027804047664e-05,
      "loss": 0.6961,
      "theoretical_loss": 3.3444887743330547,
      "tokens_seen": 2737045504
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6202477775676185e-05,
      "loss": 2.1158,
      "theoretical_loss": 3.344482501914556,
      "tokens_seen": 2737111040
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.618592774730471e-05,
      "loss": 2.0219,
      "theoretical_loss": 3.344476229688289,
      "tokens_seen": 2737176576
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6169377718933233e-05,
      "loss": 2.2947,
      "theoretical_loss": 3.3444699576542436,
      "tokens_seen": 2737242112
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6152827690561757e-05,
      "loss": 2.2856,
      "theoretical_loss": 3.3444636858124093,
      "tokens_seen": 2737307648
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6136277662190277e-05,
      "loss": 2.7304,
      "theoretical_loss": 3.3444574141627754,
      "tokens_seen": 2737373184
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.61197276338188e-05,
      "loss": 3.398,
      "theoretical_loss": 3.344451142705332,
      "tokens_seen": 2737438720
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6103177605447322e-05,
      "loss": 1.9818,
      "theoretical_loss": 3.344444871440068,
      "tokens_seen": 2737504256
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6086627577075846e-05,
      "loss": 2.5095,
      "theoretical_loss": 3.344438600366973,
      "tokens_seen": 2737569792
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.607007754870437e-05,
      "loss": 2.5511,
      "theoretical_loss": 3.3444323294860365,
      "tokens_seen": 2737635328
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6053527520332894e-05,
      "loss": 2.5994,
      "theoretical_loss": 3.3444260587972487,
      "tokens_seen": 2737700864
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6036977491961414e-05,
      "loss": 3.9062,
      "theoretical_loss": 3.3444197883005984,
      "tokens_seen": 2737766400
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.6020427463589935e-05,
      "loss": 2.6552,
      "theoretical_loss": 3.344413517996075,
      "tokens_seen": 2737831936
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.600387743521846e-05,
      "loss": 2.1265,
      "theoretical_loss": 3.3444072478836686,
      "tokens_seen": 2737897472
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5987327406846983e-05,
      "loss": 2.1396,
      "theoretical_loss": 3.3444009779633683,
      "tokens_seen": 2737963008
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5970777378475507e-05,
      "loss": 2.3947,
      "theoretical_loss": 3.3443947082351633,
      "tokens_seen": 2738028544
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5954227350104027e-05,
      "loss": 2.7247,
      "theoretical_loss": 3.344388438699044,
      "tokens_seen": 2738094080
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.593767732173255e-05,
      "loss": 0.6903,
      "theoretical_loss": 3.3443821693549993,
      "tokens_seen": 2738159616
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5921127293361072e-05,
      "loss": 0.6311,
      "theoretical_loss": 3.344375900203019,
      "tokens_seen": 2738225152
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.59045772649896e-05,
      "loss": 2.7425,
      "theoretical_loss": 3.3443696312430924,
      "tokens_seen": 2738290688
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.588802723661812e-05,
      "loss": 2.4766,
      "theoretical_loss": 3.3443633624752094,
      "tokens_seen": 2738356224
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5871477208246644e-05,
      "loss": 1.6367,
      "theoretical_loss": 3.344357093899359,
      "tokens_seen": 2738421760
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5854927179875165e-05,
      "loss": 3.0166,
      "theoretical_loss": 3.344350825515531,
      "tokens_seen": 2738487296
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.583837715150369e-05,
      "loss": 2.5107,
      "theoretical_loss": 3.3443445573237147,
      "tokens_seen": 2738552832
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5821827123132213e-05,
      "loss": 2.1893,
      "theoretical_loss": 3.3443382893238995,
      "tokens_seen": 2738618368
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5805277094760733e-05,
      "loss": 2.6605,
      "theoretical_loss": 3.3443320215160757,
      "tokens_seen": 2738683904
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5788727066389257e-05,
      "loss": 2.5514,
      "theoretical_loss": 3.3443257539002325,
      "tokens_seen": 2738749440
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5772177038017778e-05,
      "loss": 2.2488,
      "theoretical_loss": 3.3443194864763592,
      "tokens_seen": 2738814976
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5755627009646302e-05,
      "loss": 0.6349,
      "theoretical_loss": 3.3443132192444454,
      "tokens_seen": 2738880512
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5739076981274826e-05,
      "loss": 2.3845,
      "theoretical_loss": 3.34430695220448,
      "tokens_seen": 2738946048
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.572252695290335e-05,
      "loss": 2.7143,
      "theoretical_loss": 3.344300685356454,
      "tokens_seen": 2739011584
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.570597692453187e-05,
      "loss": 2.2355,
      "theoretical_loss": 3.344294418700356,
      "tokens_seen": 2739077120
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5689426896160394e-05,
      "loss": 3.6484,
      "theoretical_loss": 3.3442881522361754,
      "tokens_seen": 2739142656
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5672876867788915e-05,
      "loss": 0.5002,
      "theoretical_loss": 3.344281885963902,
      "tokens_seen": 2739208192
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.565632683941744e-05,
      "loss": 5.1403,
      "theoretical_loss": 3.344275619883525,
      "tokens_seen": 2739273728
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5639776811045963e-05,
      "loss": 2.4733,
      "theoretical_loss": 3.3442693539950343,
      "tokens_seen": 2739339264
    },
    {
      "epoch": 0.98,
      "objective/train/avg_token_score": 0.009753314778208733,
      "objective/train/avg_weight": 0.9885247945785522,
      "objective/train/docs_used": 1540914,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.608168601989746,
      "objective/train/std_weight": 0.04081965237855911,
      "objective/train/theoretical_loss": 3.34426308829842,
      "objective/train/tokens_used": 2759864800,
      "theoretical_loss": 3.34426308829842,
      "tokens_seen": 2739404800
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5623226782674487e-05,
      "loss": 2.8514,
      "theoretical_loss": 3.34426308829842,
      "tokens_seen": 2739404800
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5606676754303008e-05,
      "loss": 2.1628,
      "theoretical_loss": 3.34425682279367,
      "tokens_seen": 2739470336
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5590126725931528e-05,
      "loss": 2.7976,
      "theoretical_loss": 3.3442505574807755,
      "tokens_seen": 2739535872
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5573576697560052e-05,
      "loss": 2.4057,
      "theoretical_loss": 3.344244292359725,
      "tokens_seen": 2739601408
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5557026669188576e-05,
      "loss": 2.4141,
      "theoretical_loss": 3.3442380274305084,
      "tokens_seen": 2739666944
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.55404766408171e-05,
      "loss": -1.4748,
      "theoretical_loss": 3.3442317626931155,
      "tokens_seen": 2739732480
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.552392661244562e-05,
      "loss": 2.4649,
      "theoretical_loss": 3.3442254981475354,
      "tokens_seen": 2739798016
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5507376584074145e-05,
      "loss": 2.1848,
      "theoretical_loss": 3.344219233793758,
      "tokens_seen": 2739863552
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5490826555702665e-05,
      "loss": 2.4208,
      "theoretical_loss": 3.344212969631773,
      "tokens_seen": 2739929088
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.547427652733119e-05,
      "loss": 2.3569,
      "theoretical_loss": 3.3442067056615685,
      "tokens_seen": 2739994624
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5457726498959713e-05,
      "loss": 2.6699,
      "theoretical_loss": 3.3442004418831357,
      "tokens_seen": 2740060160
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5441176470588237e-05,
      "loss": 0.5544,
      "theoretical_loss": 3.3441941782964637,
      "tokens_seen": 2740125696
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5424626442216758e-05,
      "loss": 2.2292,
      "theoretical_loss": 3.344187914901542,
      "tokens_seen": 2740191232
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5408076413845282e-05,
      "loss": 0.691,
      "theoretical_loss": 3.3441816516983596,
      "tokens_seen": 2740256768
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5391526385473803e-05,
      "loss": 2.2827,
      "theoretical_loss": 3.3441753886869066,
      "tokens_seen": 2740322304
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5374976357102327e-05,
      "loss": 2.0296,
      "theoretical_loss": 3.3441691258671726,
      "tokens_seen": 2740387840
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.535842632873085e-05,
      "loss": 2.6666,
      "theoretical_loss": 3.3441628632391476,
      "tokens_seen": 2740453376
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.534187630035937e-05,
      "loss": 2.6821,
      "theoretical_loss": 3.34415660080282,
      "tokens_seen": 2740518912
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5325326271987895e-05,
      "loss": 2.6233,
      "theoretical_loss": 3.3441503385581797,
      "tokens_seen": 2740584448
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5308776243616416e-05,
      "loss": 2.6587,
      "theoretical_loss": 3.3441440765052164,
      "tokens_seen": 2740649984
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.529222621524494e-05,
      "loss": 2.125,
      "theoretical_loss": 3.34413781464392,
      "tokens_seen": 2740715520
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5275676186873464e-05,
      "loss": 2.6676,
      "theoretical_loss": 3.3441315529742797,
      "tokens_seen": 2740781056
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5259126158501988e-05,
      "loss": 0.787,
      "theoretical_loss": 3.344125291496285,
      "tokens_seen": 2740846592
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5242576130130508e-05,
      "loss": 2.2995,
      "theoretical_loss": 3.3441190302099253,
      "tokens_seen": 2740912128
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5226026101759032e-05,
      "loss": 2.6701,
      "theoretical_loss": 3.3441127691151906,
      "tokens_seen": 2740977664
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5209476073387555e-05,
      "loss": 3.5622,
      "theoretical_loss": 3.3441065082120707,
      "tokens_seen": 2741043200
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5192926045016079e-05,
      "loss": 2.0029,
      "theoretical_loss": 3.344100247500554,
      "tokens_seen": 2741108736
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.51763760166446e-05,
      "loss": 2.4925,
      "theoretical_loss": 3.344093986980631,
      "tokens_seen": 2741174272
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5159825988273121e-05,
      "loss": 2.3233,
      "theoretical_loss": 3.3440877266522913,
      "tokens_seen": 2741239808
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5143275959901645e-05,
      "loss": 2.1563,
      "theoretical_loss": 3.344081466515524,
      "tokens_seen": 2741305344
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5126725931530168e-05,
      "loss": 2.5273,
      "theoretical_loss": 3.3440752065703183,
      "tokens_seen": 2741370880
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5110175903158692e-05,
      "loss": 2.3713,
      "theoretical_loss": 3.344068946816665,
      "tokens_seen": 2741436416
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5093625874787212e-05,
      "loss": -1.094,
      "theoretical_loss": 3.344062687254553,
      "tokens_seen": 2741501952
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5077075846415738e-05,
      "loss": 0.2301,
      "theoretical_loss": 3.344056427883971,
      "tokens_seen": 2741567488
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5060525818044259e-05,
      "loss": 2.8549,
      "theoretical_loss": 3.3440501687049102,
      "tokens_seen": 2741633024
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5043975789672783e-05,
      "loss": 1.149,
      "theoretical_loss": 3.3440439097173593,
      "tokens_seen": 2741698560
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5027425761301305e-05,
      "loss": 1.8885,
      "theoretical_loss": 3.3440376509213072,
      "tokens_seen": 2741764096
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.5010875732929829e-05,
      "loss": 2.3268,
      "theoretical_loss": 3.344031392316745,
      "tokens_seen": 2741829632
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4994325704558351e-05,
      "loss": 2.6664,
      "theoretical_loss": 3.344025133903661,
      "tokens_seen": 2741895168
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4977775676186875e-05,
      "loss": 2.6704,
      "theoretical_loss": 3.3440188756820453,
      "tokens_seen": 2741960704
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4961225647815396e-05,
      "loss": 2.533,
      "theoretical_loss": 3.344012617651887,
      "tokens_seen": 2742026240
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4944675619443918e-05,
      "loss": 2.3804,
      "theoretical_loss": 3.3440063598131764,
      "tokens_seen": 2742091776
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4928125591072442e-05,
      "loss": 2.6951,
      "theoretical_loss": 3.3440001021659027,
      "tokens_seen": 2742157312
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4911575562700964e-05,
      "loss": 2.4012,
      "theoretical_loss": 3.3439938447100555,
      "tokens_seen": 2742222848
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4895025534329488e-05,
      "loss": 2.6418,
      "theoretical_loss": 3.343987587445624,
      "tokens_seen": 2742288384
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4878475505958009e-05,
      "loss": 5.2119,
      "theoretical_loss": 3.3439813303725985,
      "tokens_seen": 2742353920
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4861925477586533e-05,
      "loss": 2.6016,
      "theoretical_loss": 3.343975073490968,
      "tokens_seen": 2742419456
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4845375449215055e-05,
      "loss": 2.6133,
      "theoretical_loss": 3.343968816800722,
      "tokens_seen": 2742484992
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.482882542084358e-05,
      "loss": 2.5393,
      "theoretical_loss": 3.3439625603018506,
      "tokens_seen": 2742550528
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4812275392472102e-05,
      "loss": 2.5374,
      "theoretical_loss": 3.343956303994343,
      "tokens_seen": 2742616064
    },
    {
      "epoch": 0.98,
      "objective/train/avg_token_score": 0.023074422031641006,
      "objective/train/avg_weight": 0.9728428721427917,
      "objective/train/docs_used": 1542635,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 24.927230834960938,
      "objective/train/std_weight": 0.1285254955291748,
      "objective/train/theoretical_loss": 3.343950047878189,
      "objective/train/tokens_used": 2763141600,
      "theoretical_loss": 3.343950047878189,
      "tokens_seen": 2742681600
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4795725364100626e-05,
      "loss": 0.4989,
      "theoretical_loss": 3.343950047878189,
      "tokens_seen": 2742681600
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4779175335729146e-05,
      "loss": 2.3246,
      "theoretical_loss": 3.343943791953378,
      "tokens_seen": 2742747136
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.476262530735767e-05,
      "loss": 2.5962,
      "theoretical_loss": 3.3439375362198995,
      "tokens_seen": 2742812672
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4746075278986192e-05,
      "loss": 2.284,
      "theoretical_loss": 3.3439312806777433,
      "tokens_seen": 2742878208
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4729525250614715e-05,
      "loss": 2.5731,
      "theoretical_loss": 3.343925025326899,
      "tokens_seen": 2742943744
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4712975222243239e-05,
      "loss": 2.7097,
      "theoretical_loss": 3.3439187701673556,
      "tokens_seen": 2743009280
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.469642519387176e-05,
      "loss": 2.4264,
      "theoretical_loss": 3.3439125151991034,
      "tokens_seen": 2743074816
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4679875165500283e-05,
      "loss": 2.5113,
      "theoretical_loss": 3.3439062604221315,
      "tokens_seen": 2743140352
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4663325137128806e-05,
      "loss": 1.5161,
      "theoretical_loss": 3.34390000583643,
      "tokens_seen": 2743205888
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.464677510875733e-05,
      "loss": 0.6087,
      "theoretical_loss": 3.343893751441988,
      "tokens_seen": 2743271424
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4630225080385852e-05,
      "loss": 2.2634,
      "theoretical_loss": 3.3438874972387955,
      "tokens_seen": 2743336960
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4613675052014376e-05,
      "loss": 2.6239,
      "theoretical_loss": 3.3438812432268414,
      "tokens_seen": 2743402496
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4597125023642897e-05,
      "loss": 0.7755,
      "theoretical_loss": 3.343874989406116,
      "tokens_seen": 2743468032
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.458057499527142e-05,
      "loss": 2.7084,
      "theoretical_loss": 3.3438687357766086,
      "tokens_seen": 2743533568
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4564024966899943e-05,
      "loss": 2.5457,
      "theoretical_loss": 3.343862482338309,
      "tokens_seen": 2743599104
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4547474938528465e-05,
      "loss": 2.7329,
      "theoretical_loss": 3.343856229091206,
      "tokens_seen": 2743664640
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4530924910156989e-05,
      "loss": 2.9907,
      "theoretical_loss": 3.3438499760352904,
      "tokens_seen": 2743730176
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.451437488178551e-05,
      "loss": 2.3913,
      "theoretical_loss": 3.3438437231705507,
      "tokens_seen": 2743795712
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4497824853414035e-05,
      "loss": 4.3486,
      "theoretical_loss": 3.343837470496977,
      "tokens_seen": 2743861248
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4481274825042556e-05,
      "loss": 2.2342,
      "theoretical_loss": 3.343831218014559,
      "tokens_seen": 2743926784
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4481274825042556e-05,
      "loss": -0.1536,
      "theoretical_loss": 3.343824965723286,
      "tokens_seen": 2743992320
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.446472479667108e-05,
      "loss": 0.8306,
      "theoretical_loss": 3.343818713623148,
      "tokens_seen": 2744057856
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4448174768299602e-05,
      "loss": 2.739,
      "theoretical_loss": 3.343812461714134,
      "tokens_seen": 2744123392
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4431624739928126e-05,
      "loss": 1.7401,
      "theoretical_loss": 3.343806209996234,
      "tokens_seen": 2744188928
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4415074711556649e-05,
      "loss": 2.8877,
      "theoretical_loss": 3.3437999584694373,
      "tokens_seen": 2744254464
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4398524683185173e-05,
      "loss": 0.4119,
      "theoretical_loss": 3.343793707133734,
      "tokens_seen": 2744320000
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4381974654813693e-05,
      "loss": 2.7166,
      "theoretical_loss": 3.343787455989113,
      "tokens_seen": 2744385536
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4365424626442217e-05,
      "loss": 2.5369,
      "theoretical_loss": 3.343781205035565,
      "tokens_seen": 2744451072
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.434887459807074e-05,
      "loss": -0.4658,
      "theoretical_loss": 3.343774954273078,
      "tokens_seen": 2744516608
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4332324569699262e-05,
      "loss": 2.5928,
      "theoretical_loss": 3.343768703701643,
      "tokens_seen": 2744582144
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4315774541327786e-05,
      "loss": 2.787,
      "theoretical_loss": 3.343762453321249,
      "tokens_seen": 2744647680
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4299224512956306e-05,
      "loss": 2.5008,
      "theoretical_loss": 3.3437562031318855,
      "tokens_seen": 2744713216
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.428267448458483e-05,
      "loss": 2.62,
      "theoretical_loss": 3.3437499531335426,
      "tokens_seen": 2744778752
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4266124456213353e-05,
      "loss": 2.4609,
      "theoretical_loss": 3.3437437033262096,
      "tokens_seen": 2744844288
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4249574427841877e-05,
      "loss": -0.6696,
      "theoretical_loss": 3.343737453709876,
      "tokens_seen": 2744909824
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4233024399470399e-05,
      "loss": 1.2318,
      "theoretical_loss": 3.3437312042845306,
      "tokens_seen": 2744975360
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4216474371098923e-05,
      "loss": 2.6289,
      "theoretical_loss": 3.343724955050165,
      "tokens_seen": 2745040896
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4199924342727444e-05,
      "loss": 2.8368,
      "theoretical_loss": 3.3437187060067672,
      "tokens_seen": 2745106432
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4183374314355968e-05,
      "loss": 2.2395,
      "theoretical_loss": 3.3437124571543273,
      "tokens_seen": 2745171968
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.416682428598449e-05,
      "loss": 2.7036,
      "theoretical_loss": 3.343706208492835,
      "tokens_seen": 2745237504
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4150274257613014e-05,
      "loss": 2.4324,
      "theoretical_loss": 3.34369996002228,
      "tokens_seen": 2745303040
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4133724229241536e-05,
      "loss": 3.2025,
      "theoretical_loss": 3.3436937117426515,
      "tokens_seen": 2745368576
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4117174200870057e-05,
      "loss": 1.8236,
      "theoretical_loss": 3.3436874636539393,
      "tokens_seen": 2745434112
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.410062417249858e-05,
      "loss": 2.4605,
      "theoretical_loss": 3.343681215756133,
      "tokens_seen": 2745499648
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4084074144127103e-05,
      "loss": 2.558,
      "theoretical_loss": 3.3436749680492226,
      "tokens_seen": 2745565184
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4067524115755627e-05,
      "loss": 0.9991,
      "theoretical_loss": 3.3436687205331967,
      "tokens_seen": 2745630720
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.405097408738415e-05,
      "loss": 2.5411,
      "theoretical_loss": 3.343662473208046,
      "tokens_seen": 2745696256
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4034424059012673e-05,
      "loss": 2.4184,
      "theoretical_loss": 3.34365622607376,
      "tokens_seen": 2745761792
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4017874030641194e-05,
      "loss": 1.2656,
      "theoretical_loss": 3.3436499791303276,
      "tokens_seen": 2745827328
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.4001324002269718e-05,
      "loss": 2.7296,
      "theoretical_loss": 3.343643732377739,
      "tokens_seen": 2745892864
    },
    {
      "epoch": 0.98,
      "objective/train/avg_token_score": 0.009051799774169922,
      "objective/train/avg_weight": 0.989345133304596,
      "objective/train/docs_used": 1544447,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.776102066040039,
      "objective/train/std_weight": 0.06221960484981537,
      "objective/train/theoretical_loss": 3.343637485815983,
      "objective/train/tokens_used": 2766418400,
      "theoretical_loss": 3.343637485815983,
      "tokens_seen": 2745958400
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.398477397389824e-05,
      "loss": 2.6928,
      "theoretical_loss": 3.343637485815983,
      "tokens_seen": 2745958400
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3968223945526764e-05,
      "loss": 1.5058,
      "theoretical_loss": 3.3436312394450507,
      "tokens_seen": 2746023936
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3951673917155286e-05,
      "loss": 1.967,
      "theoretical_loss": 3.34362499326493,
      "tokens_seen": 2746089472
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.393512388878381e-05,
      "loss": 0.262,
      "theoretical_loss": 3.3436187472756123,
      "tokens_seen": 2746155008
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3918573860412333e-05,
      "loss": 2.0201,
      "theoretical_loss": 3.343612501477086,
      "tokens_seen": 2746220544
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3902023832040853e-05,
      "loss": 0.3329,
      "theoretical_loss": 3.3436062558693402,
      "tokens_seen": 2746286080
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3885473803669377e-05,
      "loss": 1.9952,
      "theoretical_loss": 3.3436000104523664,
      "tokens_seen": 2746351616
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.38689237752979e-05,
      "loss": 2.5189,
      "theoretical_loss": 3.3435937652261525,
      "tokens_seen": 2746417152
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3852373746926424e-05,
      "loss": 3.7157,
      "theoretical_loss": 3.343587520190689,
      "tokens_seen": 2746482688
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3835823718554946e-05,
      "loss": 2.5322,
      "theoretical_loss": 3.343581275345965,
      "tokens_seen": 2746548224
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.381927369018347e-05,
      "loss": 2.412,
      "theoretical_loss": 3.343575030691971,
      "tokens_seen": 2746613760
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.380272366181199e-05,
      "loss": 2.6666,
      "theoretical_loss": 3.343568786228696,
      "tokens_seen": 2746679296
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3786173633440515e-05,
      "loss": 2.7383,
      "theoretical_loss": 3.3435625419561292,
      "tokens_seen": 2746744832
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3769623605069037e-05,
      "loss": 2.7045,
      "theoretical_loss": 3.343556297874261,
      "tokens_seen": 2746810368
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.375307357669756e-05,
      "loss": 2.7825,
      "theoretical_loss": 3.3435500539830807,
      "tokens_seen": 2746875904
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3736523548326083e-05,
      "loss": 1.7926,
      "theoretical_loss": 3.3435438102825774,
      "tokens_seen": 2746941440
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3719973519954607e-05,
      "loss": 2.6341,
      "theoretical_loss": 3.343537566772742,
      "tokens_seen": 2747006976
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3703423491583128e-05,
      "loss": 2.6557,
      "theoretical_loss": 3.343531323453563,
      "tokens_seen": 2747072512
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.368687346321165e-05,
      "loss": 1.3977,
      "theoretical_loss": 3.3435250803250307,
      "tokens_seen": 2747138048
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3670323434840174e-05,
      "loss": 2.2136,
      "theoretical_loss": 3.3435188373871343,
      "tokens_seen": 2747203584
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3653773406468696e-05,
      "loss": 2.3178,
      "theoretical_loss": 3.3435125946398636,
      "tokens_seen": 2747269120
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.363722337809722e-05,
      "loss": 2.7773,
      "theoretical_loss": 3.3435063520832085,
      "tokens_seen": 2747334656
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3620673349725741e-05,
      "loss": 2.1598,
      "theoretical_loss": 3.3435001097171577,
      "tokens_seen": 2747400192
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3604123321354265e-05,
      "loss": 0.693,
      "theoretical_loss": 3.343493867541702,
      "tokens_seen": 2747465728
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3587573292982787e-05,
      "loss": 4.4684,
      "theoretical_loss": 3.3434876255568304,
      "tokens_seen": 2747531264
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3571023264611311e-05,
      "loss": 2.684,
      "theoretical_loss": 3.3434813837625326,
      "tokens_seen": 2747596800
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3554473236239833e-05,
      "loss": 2.6946,
      "theoretical_loss": 3.3434751421587987,
      "tokens_seen": 2747662336
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3537923207868357e-05,
      "loss": 2.2856,
      "theoretical_loss": 3.3434689007456173,
      "tokens_seen": 2747727872
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3521373179496878e-05,
      "loss": 2.587,
      "theoretical_loss": 3.343462659522979,
      "tokens_seen": 2747793408
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3504823151125402e-05,
      "loss": 2.4047,
      "theoretical_loss": 3.3434564184908733,
      "tokens_seen": 2747858944
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3488273122753924e-05,
      "loss": 2.1733,
      "theoretical_loss": 3.3434501776492898,
      "tokens_seen": 2747924480
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3471723094382447e-05,
      "loss": 2.6149,
      "theoretical_loss": 3.3434439369982174,
      "tokens_seen": 2747990016
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.345517306601097e-05,
      "loss": 2.0985,
      "theoretical_loss": 3.343437696537647,
      "tokens_seen": 2748055552
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3438623037639491e-05,
      "loss": 1.0813,
      "theoretical_loss": 3.3434314562675667,
      "tokens_seen": 2748121088
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3422073009268015e-05,
      "loss": 2.5066,
      "theoretical_loss": 3.343425216187968,
      "tokens_seen": 2748186624
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3405522980896538e-05,
      "loss": 2.3669,
      "theoretical_loss": 3.343418976298839,
      "tokens_seen": 2748252160
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3388972952525062e-05,
      "loss": 2.959,
      "theoretical_loss": 3.3434127366001696,
      "tokens_seen": 2748317696
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3372422924153584e-05,
      "loss": 0.7014,
      "theoretical_loss": 3.34340649709195,
      "tokens_seen": 2748383232
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3355872895782108e-05,
      "loss": 2.4072,
      "theoretical_loss": 3.34340025777417,
      "tokens_seen": 2748448768
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.333932286741063e-05,
      "loss": 2.6794,
      "theoretical_loss": 3.3433940186468187,
      "tokens_seen": 2748514304
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3322772839039154e-05,
      "loss": 2.0521,
      "theoretical_loss": 3.343387779709886,
      "tokens_seen": 2748579840
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3306222810667675e-05,
      "loss": 2.7246,
      "theoretical_loss": 3.3433815409633607,
      "tokens_seen": 2748645376
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3289672782296199e-05,
      "loss": 2.5283,
      "theoretical_loss": 3.3433753024072335,
      "tokens_seen": 2748710912
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3273122753924721e-05,
      "loss": 2.6584,
      "theoretical_loss": 3.3433690640414944,
      "tokens_seen": 2748776448
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3256572725553243e-05,
      "loss": 2.8638,
      "theoretical_loss": 3.343362825866132,
      "tokens_seen": 2748841984
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3240022697181767e-05,
      "loss": 2.7758,
      "theoretical_loss": 3.3433565878811358,
      "tokens_seen": 2748907520
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3223472668810288e-05,
      "loss": 0.9522,
      "theoretical_loss": 3.3433503500864967,
      "tokens_seen": 2748973056
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3206922640438812e-05,
      "loss": 1.3745,
      "theoretical_loss": 3.343344112482203,
      "tokens_seen": 2749038592
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3190372612067334e-05,
      "loss": 2.1998,
      "theoretical_loss": 3.343337875068246,
      "tokens_seen": 2749104128
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3173822583695858e-05,
      "loss": 1.6919,
      "theoretical_loss": 3.3433316378446136,
      "tokens_seen": 2749169664
    },
    {
      "epoch": 0.98,
      "objective/train/avg_token_score": 0.020770728588104248,
      "objective/train/avg_weight": 0.975556492805481,
      "objective/train/docs_used": 1546277,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7739100456237793,
      "objective/train/std_weight": 0.07999619841575623,
      "objective/train/theoretical_loss": 3.3433254008112963,
      "objective/train/tokens_used": 2769695200,
      "theoretical_loss": 3.3433254008112963,
      "tokens_seen": 2749235200
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.315727255532438e-05,
      "loss": 2.7446,
      "theoretical_loss": 3.3433254008112963,
      "tokens_seen": 2749235200
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3140722526952904e-05,
      "loss": 2.4783,
      "theoretical_loss": 3.3433191639682835,
      "tokens_seen": 2749300736
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3124172498581425e-05,
      "loss": 2.3479,
      "theoretical_loss": 3.3433129273155657,
      "tokens_seen": 2749366272
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3107622470209949e-05,
      "loss": 2.1644,
      "theoretical_loss": 3.3433066908531313,
      "tokens_seen": 2749431808
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3091072441838471e-05,
      "loss": 2.4951,
      "theoretical_loss": 3.343300454580971,
      "tokens_seen": 2749497344
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3074522413466995e-05,
      "loss": 2.6372,
      "theoretical_loss": 3.3432942184990733,
      "tokens_seen": 2749562880
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3057972385095518e-05,
      "loss": 1.4045,
      "theoretical_loss": 3.343287982607429,
      "tokens_seen": 2749628416
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3041422356724038e-05,
      "loss": 2.6917,
      "theoretical_loss": 3.3432817469060274,
      "tokens_seen": 2749693952
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3024872328352562e-05,
      "loss": 3.0841,
      "theoretical_loss": 3.343275511394858,
      "tokens_seen": 2749759488
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.3008322299981085e-05,
      "loss": 2.1853,
      "theoretical_loss": 3.3432692760739102,
      "tokens_seen": 2749825024
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2991772271609609e-05,
      "loss": 2.5189,
      "theoretical_loss": 3.3432630409431745,
      "tokens_seen": 2749890560
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.297522224323813e-05,
      "loss": 2.2159,
      "theoretical_loss": 3.34325680600264,
      "tokens_seen": 2749956096
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2958672214866655e-05,
      "loss": 1.6384,
      "theoretical_loss": 3.3432505712522964,
      "tokens_seen": 2750021632
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2942122186495175e-05,
      "loss": 2.4218,
      "theoretical_loss": 3.343244336692133,
      "tokens_seen": 2750087168
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.29255721581237e-05,
      "loss": 2.6741,
      "theoretical_loss": 3.3432381023221405,
      "tokens_seen": 2750152704
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2909022129752222e-05,
      "loss": 2.4088,
      "theoretical_loss": 3.3432318681423077,
      "tokens_seen": 2750218240
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2892472101380746e-05,
      "loss": -0.7536,
      "theoretical_loss": 3.3432256341526245,
      "tokens_seen": 2750283776
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2875922073009268e-05,
      "loss": 2.489,
      "theoretical_loss": 3.343219400353081,
      "tokens_seen": 2750349312
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2859372044637792e-05,
      "loss": 2.3593,
      "theoretical_loss": 3.3432131667436655,
      "tokens_seen": 2750414848
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2842822016266314e-05,
      "loss": 1.9352,
      "theoretical_loss": 3.3432069333243692,
      "tokens_seen": 2750480384
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2826271987894835e-05,
      "loss": 2.7309,
      "theoretical_loss": 3.3432007000951813,
      "tokens_seen": 2750545920
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2809721959523359e-05,
      "loss": 3.3337,
      "theoretical_loss": 3.343194467056091,
      "tokens_seen": 2750611456
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2793171931151881e-05,
      "loss": 2.8829,
      "theoretical_loss": 3.343188234207089,
      "tokens_seen": 2750676992
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2776621902780405e-05,
      "loss": 2.4303,
      "theoretical_loss": 3.3431820015481635,
      "tokens_seen": 2750742528
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2760071874408927e-05,
      "loss": 2.7779,
      "theoretical_loss": 3.3431757690793056,
      "tokens_seen": 2750808064
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2743521846037451e-05,
      "loss": 1.099,
      "theoretical_loss": 3.343169536800504,
      "tokens_seen": 2750873600
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2726971817665972e-05,
      "loss": 2.5945,
      "theoretical_loss": 3.343163304711749,
      "tokens_seen": 2750939136
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2710421789294496e-05,
      "loss": 2.7287,
      "theoretical_loss": 3.3431570728130295,
      "tokens_seen": 2751004672
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2693871760923018e-05,
      "loss": -0.9239,
      "theoretical_loss": 3.3431508411043365,
      "tokens_seen": 2751070208
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2677321732551542e-05,
      "loss": 2.4769,
      "theoretical_loss": 3.3431446095856585,
      "tokens_seen": 2751135744
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2660771704180065e-05,
      "loss": 2.5268,
      "theoretical_loss": 3.3431383782569855,
      "tokens_seen": 2751201280
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2644221675808589e-05,
      "loss": 2.6664,
      "theoretical_loss": 3.343132147118307,
      "tokens_seen": 2751266816
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.262767164743711e-05,
      "loss": 2.613,
      "theoretical_loss": 3.343125916169613,
      "tokens_seen": 2751332352
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2611121619065632e-05,
      "loss": 2.1044,
      "theoretical_loss": 3.3431196854108935,
      "tokens_seen": 2751397888
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2594571590694156e-05,
      "loss": 2.9129,
      "theoretical_loss": 3.3431134548421375,
      "tokens_seen": 2751463424
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2578021562322678e-05,
      "loss": 1.7208,
      "theoretical_loss": 3.3431072244633353,
      "tokens_seen": 2751528960
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2561471533951202e-05,
      "loss": 2.4514,
      "theoretical_loss": 3.343100994274476,
      "tokens_seen": 2751594496
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2544921505579722e-05,
      "loss": 2.0169,
      "theoretical_loss": 3.3430947642755493,
      "tokens_seen": 2751660032
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2528371477208246e-05,
      "loss": 2.512,
      "theoretical_loss": 3.3430885344665455,
      "tokens_seen": 2751725568
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2511821448836769e-05,
      "loss": 2.7543,
      "theoretical_loss": 3.343082304847454,
      "tokens_seen": 2751791104
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2495271420465293e-05,
      "loss": 2.6273,
      "theoretical_loss": 3.343076075418264,
      "tokens_seen": 2751856640
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2478721392093815e-05,
      "loss": 0.378,
      "theoretical_loss": 3.343069846178966,
      "tokens_seen": 2751922176
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2462171363722339e-05,
      "loss": 2.6812,
      "theoretical_loss": 3.3430636171295487,
      "tokens_seen": 2751987712
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.244562133535086e-05,
      "loss": 2.3556,
      "theoretical_loss": 3.3430573882700028,
      "tokens_seen": 2752053248
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2429071306979384e-05,
      "loss": 2.4299,
      "theoretical_loss": 3.3430511596003174,
      "tokens_seen": 2752118784
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2412521278607906e-05,
      "loss": 0.8845,
      "theoretical_loss": 3.3430449311204824,
      "tokens_seen": 2752184320
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2395971250236428e-05,
      "loss": 2.5959,
      "theoretical_loss": 3.3430387028304875,
      "tokens_seen": 2752249856
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2379421221864952e-05,
      "loss": 2.0973,
      "theoretical_loss": 3.343032474730322,
      "tokens_seen": 2752315392
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2362871193493473e-05,
      "loss": 2.1791,
      "theoretical_loss": 3.3430262468199765,
      "tokens_seen": 2752380928
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2346321165121997e-05,
      "loss": 1.5633,
      "theoretical_loss": 3.34302001909944,
      "tokens_seen": 2752446464
    },
    {
      "epoch": 0.98,
      "objective/train/avg_token_score": 0.026518136262893677,
      "objective/train/avg_weight": 0.9688000082969666,
      "objective/train/docs_used": 1547942,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.914694309234619,
      "objective/train/std_weight": 0.10628005117177963,
      "objective/train/theoretical_loss": 3.343013791568702,
      "objective/train/tokens_used": 2772972000,
      "theoretical_loss": 3.343013791568702,
      "tokens_seen": 2752512000
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2329771136750519e-05,
      "loss": 3.9489,
      "theoretical_loss": 3.343013791568702,
      "tokens_seen": 2752512000
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2313221108379043e-05,
      "loss": -0.8338,
      "theoretical_loss": 3.3430075642277526,
      "tokens_seen": 2752577536
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2296671080007565e-05,
      "loss": 2.406,
      "theoretical_loss": 3.343001337076582,
      "tokens_seen": 2752643072
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.228012105163609e-05,
      "loss": 2.3579,
      "theoretical_loss": 3.3429951101151785,
      "tokens_seen": 2752708608
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2263571023264612e-05,
      "loss": 2.8821,
      "theoretical_loss": 3.342988883343533,
      "tokens_seen": 2752774144
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2247020994893136e-05,
      "loss": 3.0061,
      "theoretical_loss": 3.342982656761635,
      "tokens_seen": 2752839680
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2230470966521656e-05,
      "loss": 2.3721,
      "theoretical_loss": 3.3429764303694736,
      "tokens_seen": 2752905216
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.221392093815018e-05,
      "loss": 2.3687,
      "theoretical_loss": 3.342970204167039,
      "tokens_seen": 2752970752
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2197370909778703e-05,
      "loss": -1.6611,
      "theoretical_loss": 3.3429639781543212,
      "tokens_seen": 2753036288
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2180820881407225e-05,
      "loss": 2.1034,
      "theoretical_loss": 3.342957752331309,
      "tokens_seen": 2753101824
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2164270853035749e-05,
      "loss": 2.4954,
      "theoretical_loss": 3.342951526697993,
      "tokens_seen": 2753167360
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.214772082466427e-05,
      "loss": 2.1185,
      "theoretical_loss": 3.342945301254362,
      "tokens_seen": 2753232896
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2131170796292793e-05,
      "loss": 2.3275,
      "theoretical_loss": 3.342939076000407,
      "tokens_seen": 2753298432
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2114620767921316e-05,
      "loss": 2.4224,
      "theoretical_loss": 3.3429328509361165,
      "tokens_seen": 2753363968
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.209807073954984e-05,
      "loss": 2.4402,
      "theoretical_loss": 3.342926626061481,
      "tokens_seen": 2753429504
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2081520711178362e-05,
      "loss": 2.6125,
      "theoretical_loss": 3.3429204013764897,
      "tokens_seen": 2753495040
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2064970682806886e-05,
      "loss": 2.8557,
      "theoretical_loss": 3.3429141768811323,
      "tokens_seen": 2753560576
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2048420654435407e-05,
      "loss": 2.3279,
      "theoretical_loss": 3.342907952575399,
      "tokens_seen": 2753626112
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.203187062606393e-05,
      "loss": 2.4682,
      "theoretical_loss": 3.3429017284592786,
      "tokens_seen": 2753691648
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.2015320597692453e-05,
      "loss": 2.5081,
      "theoretical_loss": 3.3428955045327617,
      "tokens_seen": 2753757184
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1998770569320977e-05,
      "loss": 1.3248,
      "theoretical_loss": 3.3428892807958377,
      "tokens_seen": 2753822720
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.19822205409495e-05,
      "loss": 1.0681,
      "theoretical_loss": 3.3428830572484967,
      "tokens_seen": 2753888256
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.196567051257802e-05,
      "loss": 1.6203,
      "theoretical_loss": 3.3428768338907275,
      "tokens_seen": 2753953792
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1949120484206544e-05,
      "loss": 2.7047,
      "theoretical_loss": 3.342870610722521,
      "tokens_seen": 2754019328
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1932570455835066e-05,
      "loss": 2.6415,
      "theoretical_loss": 3.3428643877438655,
      "tokens_seen": 2754084864
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.191602042746359e-05,
      "loss": 2.5182,
      "theoretical_loss": 3.3428581649547517,
      "tokens_seen": 2754150400
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1899470399092112e-05,
      "loss": 2.0248,
      "theoretical_loss": 3.3428519423551695,
      "tokens_seen": 2754215936
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1882920370720636e-05,
      "loss": 2.3971,
      "theoretical_loss": 3.342845719945108,
      "tokens_seen": 2754281472
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1866370342349157e-05,
      "loss": 2.5576,
      "theoretical_loss": 3.3428394977245572,
      "tokens_seen": 2754347008
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1849820313977681e-05,
      "loss": 2.3721,
      "theoretical_loss": 3.3428332756935064,
      "tokens_seen": 2754412544
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1833270285606203e-05,
      "loss": 2.4898,
      "theoretical_loss": 3.342827053851946,
      "tokens_seen": 2754478080
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1816720257234727e-05,
      "loss": 2.418,
      "theoretical_loss": 3.3428208321998656,
      "tokens_seen": 2754543616
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.180017022886325e-05,
      "loss": 2.719,
      "theoretical_loss": 3.3428146107372543,
      "tokens_seen": 2754609152
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1783620200491773e-05,
      "loss": 2.5694,
      "theoretical_loss": 3.3428083894641025,
      "tokens_seen": 2754674688
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1767070172120294e-05,
      "loss": 2.2572,
      "theoretical_loss": 3.3428021683803992,
      "tokens_seen": 2754740224
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1750520143748816e-05,
      "loss": 2.4025,
      "theoretical_loss": 3.342795947486135,
      "tokens_seen": 2754805760
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.173397011537734e-05,
      "loss": 2.5947,
      "theoretical_loss": 3.3427897267812994,
      "tokens_seen": 2754871296
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1717420087005863e-05,
      "loss": 1.8365,
      "theoretical_loss": 3.3427835062658815,
      "tokens_seen": 2754936832
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1700870058634387e-05,
      "loss": 1.6073,
      "theoretical_loss": 3.3427772859398717,
      "tokens_seen": 2755002368
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1684320030262909e-05,
      "loss": 3.1684,
      "theoretical_loss": 3.3427710658032597,
      "tokens_seen": 2755067904
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1667770001891433e-05,
      "loss": 2.4337,
      "theoretical_loss": 3.342764845856035,
      "tokens_seen": 2755133440
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1651219973519954e-05,
      "loss": 2.1824,
      "theoretical_loss": 3.342758626098187,
      "tokens_seen": 2755198976
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1634669945148478e-05,
      "loss": 2.9475,
      "theoretical_loss": 3.342752406529706,
      "tokens_seen": 2755264512
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1618119916777e-05,
      "loss": 2.7252,
      "theoretical_loss": 3.3427461871505812,
      "tokens_seen": 2755330048
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1601569888405524e-05,
      "loss": 2.6865,
      "theoretical_loss": 3.342739967960803,
      "tokens_seen": 2755395584
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1585019860034046e-05,
      "loss": 1.9167,
      "theoretical_loss": 3.3427337489603604,
      "tokens_seen": 2755461120
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.156846983166257e-05,
      "loss": 2.5521,
      "theoretical_loss": 3.3427275301492436,
      "tokens_seen": 2755526656
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.155191980329109e-05,
      "loss": 2.5788,
      "theoretical_loss": 3.342721311527443,
      "tokens_seen": 2755592192
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1535369774919613e-05,
      "loss": 2.359,
      "theoretical_loss": 3.342715093094947,
      "tokens_seen": 2755657728
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1518819746548137e-05,
      "loss": 2.4966,
      "theoretical_loss": 3.3427088748517457,
      "tokens_seen": 2755723264
    },
    {
      "epoch": 0.98,
      "objective/train/avg_token_score": 0.030750572681427002,
      "objective/train/avg_weight": 0.963820219039917,
      "objective/train/docs_used": 1550559,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.7077412605285645,
      "objective/train/std_weight": 0.12629997730255127,
      "objective/train/theoretical_loss": 3.342702656797829,
      "objective/train/tokens_used": 2776248800,
      "theoretical_loss": 3.342702656797829,
      "tokens_seen": 2755788800
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.150226971817666e-05,
      "loss": 2.5742,
      "theoretical_loss": 3.342702656797829,
      "tokens_seen": 2755788800
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1485719689805183e-05,
      "loss": 2.3524,
      "theoretical_loss": 3.3426964389331872,
      "tokens_seen": 2755854336
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1469169661433704e-05,
      "loss": 0.148,
      "theoretical_loss": 3.342690221257809,
      "tokens_seen": 2755919872
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1452619633062228e-05,
      "loss": 5.5293,
      "theoretical_loss": 3.342684003771685,
      "tokens_seen": 2755985408
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.143606960469075e-05,
      "loss": 2.6905,
      "theoretical_loss": 3.3426777864748045,
      "tokens_seen": 2756050944
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1419519576319274e-05,
      "loss": 2.3245,
      "theoretical_loss": 3.3426715693671576,
      "tokens_seen": 2756116480
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1402969547947797e-05,
      "loss": 2.3163,
      "theoretical_loss": 3.3426653524487335,
      "tokens_seen": 2756182016
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.138641951957632e-05,
      "loss": 2.2832,
      "theoretical_loss": 3.342659135719522,
      "tokens_seen": 2756247552
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1369869491204841e-05,
      "loss": 1.1558,
      "theoretical_loss": 3.342652919179513,
      "tokens_seen": 2756313088
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1353319462833365e-05,
      "loss": 2.8769,
      "theoretical_loss": 3.342646702828697,
      "tokens_seen": 2756378624
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1336769434461887e-05,
      "loss": 0.5573,
      "theoretical_loss": 3.342640486667063,
      "tokens_seen": 2756444160
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.132021940609041e-05,
      "loss": 2.3405,
      "theoretical_loss": 3.3426342706946004,
      "tokens_seen": 2756509696
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1303669377718934e-05,
      "loss": 2.7068,
      "theoretical_loss": 3.342628054911299,
      "tokens_seen": 2756575232
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1287119349347454e-05,
      "loss": 2.5395,
      "theoretical_loss": 3.3426218393171494,
      "tokens_seen": 2756640768
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1270569320975978e-05,
      "loss": 1.8114,
      "theoretical_loss": 3.342615623912141,
      "tokens_seen": 2756706304
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.12540192926045e-05,
      "loss": 2.458,
      "theoretical_loss": 3.342609408696263,
      "tokens_seen": 2756771840
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1237469264233025e-05,
      "loss": 2.605,
      "theoretical_loss": 3.3426031936695058,
      "tokens_seen": 2756837376
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1220919235861547e-05,
      "loss": 0.9936,
      "theoretical_loss": 3.3425969788318586,
      "tokens_seen": 2756902912
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1204369207490071e-05,
      "loss": 2.5214,
      "theoretical_loss": 3.3425907641833117,
      "tokens_seen": 2756968448
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1187819179118591e-05,
      "loss": 2.0037,
      "theoretical_loss": 3.3425845497238544,
      "tokens_seen": 2757033984
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1171269150747117e-05,
      "loss": 2.8915,
      "theoretical_loss": 3.342578335453477,
      "tokens_seen": 2757099520
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1154719122375638e-05,
      "loss": 1.1596,
      "theoretical_loss": 3.3425721213721684,
      "tokens_seen": 2757165056
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1138169094004162e-05,
      "loss": 2.9083,
      "theoretical_loss": 3.342565907479919,
      "tokens_seen": 2757230592
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1121619065632684e-05,
      "loss": 2.545,
      "theoretical_loss": 3.3425596937767184,
      "tokens_seen": 2757296128
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1105069037261206e-05,
      "loss": 2.3329,
      "theoretical_loss": 3.3425534802625565,
      "tokens_seen": 2757361664
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.108851900888973e-05,
      "loss": 1.1853,
      "theoretical_loss": 3.342547266937423,
      "tokens_seen": 2757427200
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1071968980518251e-05,
      "loss": 2.3754,
      "theoretical_loss": 3.342541053801307,
      "tokens_seen": 2757492736
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1055418952146775e-05,
      "loss": 2.5209,
      "theoretical_loss": 3.3425348408541993,
      "tokens_seen": 2757558272
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1038868923775297e-05,
      "loss": 2.7055,
      "theoretical_loss": 3.3425286280960895,
      "tokens_seen": 2757623808
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1022318895403821e-05,
      "loss": 2.414,
      "theoretical_loss": 3.3425224155269664,
      "tokens_seen": 2757689344
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.1005768867032344e-05,
      "loss": 2.1052,
      "theoretical_loss": 3.3425162031468205,
      "tokens_seen": 2757754880
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0989218838660867e-05,
      "loss": 2.6872,
      "theoretical_loss": 3.3425099909556417,
      "tokens_seen": 2757820416
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0972668810289388e-05,
      "loss": 2.3581,
      "theoretical_loss": 3.3425037789534198,
      "tokens_seen": 2757885952
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.0956118781917912e-05,
      "loss": 2.5952,
      "theoretical_loss": 3.3424975671401436,
      "tokens_seen": 2757951488
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0939568753546434e-05,
      "loss": 1.9359,
      "theoretical_loss": 3.342491355515804,
      "tokens_seen": 2758017024
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0923018725174957e-05,
      "loss": 2.365,
      "theoretical_loss": 3.34248514408039,
      "tokens_seen": 2758082560
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.090646869680348e-05,
      "loss": -2.4155,
      "theoretical_loss": 3.3424789328338917,
      "tokens_seen": 2758148096
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0889918668432003e-05,
      "loss": 2.3747,
      "theoretical_loss": 3.342472721776299,
      "tokens_seen": 2758213632
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0873368640060525e-05,
      "loss": 2.6423,
      "theoretical_loss": 3.342466510907602,
      "tokens_seen": 2758279168
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.085681861168905e-05,
      "loss": 2.5014,
      "theoretical_loss": 3.3424603002277893,
      "tokens_seen": 2758344704
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0840268583317572e-05,
      "loss": 2.2368,
      "theoretical_loss": 3.3424540897368518,
      "tokens_seen": 2758410240
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0823718554946094e-05,
      "loss": 2.6528,
      "theoretical_loss": 3.3424478794347783,
      "tokens_seen": 2758475776
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0807168526574618e-05,
      "loss": 1.8992,
      "theoretical_loss": 3.3424416693215595,
      "tokens_seen": 2758541312
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0790618498203138e-05,
      "loss": 2.6957,
      "theoretical_loss": 3.3424354593971843,
      "tokens_seen": 2758606848
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0774068469831662e-05,
      "loss": 2.3874,
      "theoretical_loss": 3.3424292496616435,
      "tokens_seen": 2758672384
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0757518441460185e-05,
      "loss": 2.4113,
      "theoretical_loss": 3.342423040114926,
      "tokens_seen": 2758737920
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0740968413088707e-05,
      "loss": 2.5261,
      "theoretical_loss": 3.342416830757022,
      "tokens_seen": 2758803456
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0724418384717231e-05,
      "loss": 2.4814,
      "theoretical_loss": 3.342410621587921,
      "tokens_seen": 2758868992
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0707868356345753e-05,
      "loss": 2.0661,
      "theoretical_loss": 3.342404412607613,
      "tokens_seen": 2758934528
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0691318327974276e-05,
      "loss": 1.9824,
      "theoretical_loss": 3.3423982038160878,
      "tokens_seen": 2759000064
    },
    {
      "epoch": 0.99,
      "objective/train/avg_token_score": 0.013318187557160854,
      "objective/train/avg_weight": 0.9843593835830688,
      "objective/train/docs_used": 1552546,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 12.667391777038574,
      "objective/train/std_weight": 0.1004774272441864,
      "objective/train/theoretical_loss": 3.342391995213335,
      "objective/train/tokens_used": 2779525600,
      "theoretical_loss": 3.342391995213335,
      "tokens_seen": 2759065600
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.06747682996028e-05,
      "loss": 2.1077,
      "theoretical_loss": 3.342391995213335,
      "tokens_seen": 2759065600
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0658218271231322e-05,
      "loss": 2.1099,
      "theoretical_loss": 3.342385786799344,
      "tokens_seen": 2759131136
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0641668242859846e-05,
      "loss": 2.3868,
      "theoretical_loss": 3.3423795785741057,
      "tokens_seen": 2759196672
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0625118214488368e-05,
      "loss": 2.6614,
      "theoretical_loss": 3.342373370537609,
      "tokens_seen": 2759262208
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.060856818611689e-05,
      "loss": 2.5713,
      "theoretical_loss": 3.3423671626898437,
      "tokens_seen": 2759327744
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0592018157745414e-05,
      "loss": 1.7952,
      "theoretical_loss": 3.3423609550308,
      "tokens_seen": 2759393280
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0575468129373935e-05,
      "loss": 3.0998,
      "theoretical_loss": 3.342354747560467,
      "tokens_seen": 2759458816
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0558918101002459e-05,
      "loss": 2.5479,
      "theoretical_loss": 3.3423485402788353,
      "tokens_seen": 2759524352
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0542368072630981e-05,
      "loss": 1.8733,
      "theoretical_loss": 3.3423423331858944,
      "tokens_seen": 2759589888
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0525818044259504e-05,
      "loss": 2.5478,
      "theoretical_loss": 3.342336126281634,
      "tokens_seen": 2759655424
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0509268015888028e-05,
      "loss": 2.048,
      "theoretical_loss": 3.342329919566044,
      "tokens_seen": 2759720960
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.049271798751655e-05,
      "loss": 2.4492,
      "theoretical_loss": 3.3423237130391135,
      "tokens_seen": 2759786496
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0476167959145072e-05,
      "loss": 1.9091,
      "theoretical_loss": 3.3423175067008333,
      "tokens_seen": 2759852032
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0459617930773596e-05,
      "loss": 2.2804,
      "theoretical_loss": 3.3423113005511924,
      "tokens_seen": 2759917568
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0443067902402119e-05,
      "loss": 2.5908,
      "theoretical_loss": 3.3423050945901815,
      "tokens_seen": 2759983104
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0426517874030641e-05,
      "loss": 2.4368,
      "theoretical_loss": 3.342298888817789,
      "tokens_seen": 2760048640
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0409967845659165e-05,
      "loss": 1.5803,
      "theoretical_loss": 3.3422926832340063,
      "tokens_seen": 2760114176
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0393417817287687e-05,
      "loss": 2.3209,
      "theoretical_loss": 3.342286477838822,
      "tokens_seen": 2760179712
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.037686778891621e-05,
      "loss": -0.7996,
      "theoretical_loss": 3.3422802726322263,
      "tokens_seen": 2760245248
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0360317760544732e-05,
      "loss": 2.5467,
      "theoretical_loss": 3.3422740676142086,
      "tokens_seen": 2760310784
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0343767732173254e-05,
      "loss": 2.4457,
      "theoretical_loss": 3.3422678627847597,
      "tokens_seen": 2760376320
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0327217703801778e-05,
      "loss": 2.7698,
      "theoretical_loss": 3.3422616581438684,
      "tokens_seen": 2760441856
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.03106676754303e-05,
      "loss": 2.3377,
      "theoretical_loss": 3.342255453691525,
      "tokens_seen": 2760507392
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0294117647058823e-05,
      "loss": 2.86,
      "theoretical_loss": 3.342249249427719,
      "tokens_seen": 2760572928
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0277567618687347e-05,
      "loss": 2.5541,
      "theoretical_loss": 3.3422430453524403,
      "tokens_seen": 2760638464
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0261017590315869e-05,
      "loss": 4.365,
      "theoretical_loss": 3.342236841465679,
      "tokens_seen": 2760704000
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0244467561944391e-05,
      "loss": 1.0657,
      "theoretical_loss": 3.3422306377674245,
      "tokens_seen": 2760769536
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0227917533572915e-05,
      "loss": -0.6391,
      "theoretical_loss": 3.3422244342576666,
      "tokens_seen": 2760835072
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0211367505201438e-05,
      "loss": 3.7487,
      "theoretical_loss": 3.3422182309363953,
      "tokens_seen": 2760900608
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.019481747682996e-05,
      "loss": -0.122,
      "theoretical_loss": 3.3422120278036003,
      "tokens_seen": 2760966144
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0178267448458484e-05,
      "loss": 2.7131,
      "theoretical_loss": 3.3422058248592714,
      "tokens_seen": 2761031680
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0161717420087004e-05,
      "loss": 1.3906,
      "theoretical_loss": 3.3421996221033985,
      "tokens_seen": 2761097216
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0145167391715528e-05,
      "loss": 2.2581,
      "theoretical_loss": 3.3421934195359713,
      "tokens_seen": 2761162752
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.012861736334405e-05,
      "loss": 1.1354,
      "theoretical_loss": 3.3421872171569795,
      "tokens_seen": 2761228288
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0112067334972573e-05,
      "loss": 1.6083,
      "theoretical_loss": 3.3421810149664135,
      "tokens_seen": 2761293824
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0095517306601097e-05,
      "loss": 2.7935,
      "theoretical_loss": 3.342174812964262,
      "tokens_seen": 2761359360
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.007896727822962e-05,
      "loss": 2.1369,
      "theoretical_loss": 3.342168611150516,
      "tokens_seen": 2761424896
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0062417249858143e-05,
      "loss": 2.94,
      "theoretical_loss": 3.3421624095251645,
      "tokens_seen": 2761490432
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0045867221486666e-05,
      "loss": 2.46,
      "theoretical_loss": 3.3421562080881975,
      "tokens_seen": 2761555968
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0029317193115188e-05,
      "loss": 2.5804,
      "theoretical_loss": 3.3421500068396046,
      "tokens_seen": 2761621504
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.0012767164743712e-05,
      "loss": 2.6309,
      "theoretical_loss": 3.3421438057793766,
      "tokens_seen": 2761687040
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.996217136372234e-06,
      "loss": 1.9004,
      "theoretical_loss": 3.342137604907502,
      "tokens_seen": 2761752576
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.979667108000756e-06,
      "loss": 2.5428,
      "theoretical_loss": 3.3421314042239714,
      "tokens_seen": 2761818112
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.96311707962928e-06,
      "loss": 2.6524,
      "theoretical_loss": 3.3421252037287745,
      "tokens_seen": 2761883648
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.946567051257801e-06,
      "loss": 1.8826,
      "theoretical_loss": 3.3421190034219004,
      "tokens_seen": 2761949184
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.930017022886325e-06,
      "loss": 2.6103,
      "theoretical_loss": 3.34211280330334,
      "tokens_seen": 2762014720
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.913466994514847e-06,
      "loss": 0.3165,
      "theoretical_loss": 3.342106603373083,
      "tokens_seen": 2762080256
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.89691696614337e-06,
      "loss": 2.3245,
      "theoretical_loss": 3.342100403631118,
      "tokens_seen": 2762145792
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.880366937771894e-06,
      "loss": 3.1062,
      "theoretical_loss": 3.342094204077436,
      "tokens_seen": 2762211328
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.863816909400416e-06,
      "loss": 0.9928,
      "theoretical_loss": 3.3420880047120267,
      "tokens_seen": 2762276864
    },
    {
      "epoch": 0.99,
      "objective/train/avg_token_score": 0.013883264735341072,
      "objective/train/avg_weight": 0.9836724996566772,
      "objective/train/docs_used": 1554602,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.86199951171875,
      "objective/train/std_weight": 0.07113757729530334,
      "objective/train/theoretical_loss": 3.34208180553488,
      "objective/train/tokens_used": 2782802400,
      "theoretical_loss": 3.34208180553488,
      "tokens_seen": 2762342400
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.847266881028938e-06,
      "loss": 2.2695,
      "theoretical_loss": 3.34208180553488,
      "tokens_seen": 2762342400
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.830716852657462e-06,
      "loss": 2.1305,
      "theoretical_loss": 3.3420756065459845,
      "tokens_seen": 2762407936
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.814166824285985e-06,
      "loss": 1.6076,
      "theoretical_loss": 3.3420694077453317,
      "tokens_seen": 2762473472
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.797616795914507e-06,
      "loss": 2.5382,
      "theoretical_loss": 3.34206320913291,
      "tokens_seen": 2762539008
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.78106676754303e-06,
      "loss": 2.7085,
      "theoretical_loss": 3.3420570107087104,
      "tokens_seen": 2762604544
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.764516739171553e-06,
      "loss": 1.7251,
      "theoretical_loss": 3.342050812472722,
      "tokens_seen": 2762670080
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.747966710800075e-06,
      "loss": 2.3826,
      "theoretical_loss": 3.3420446144249345,
      "tokens_seen": 2762735616
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.731416682428598e-06,
      "loss": 1.9317,
      "theoretical_loss": 3.3420384165653383,
      "tokens_seen": 2762801152
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.71486665405712e-06,
      "loss": 2.6593,
      "theoretical_loss": 3.342032218893923,
      "tokens_seen": 2762866688
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.698316625685644e-06,
      "loss": 2.3514,
      "theoretical_loss": 3.3420260214106783,
      "tokens_seen": 2762932224
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.681766597314166e-06,
      "loss": 3.2305,
      "theoretical_loss": 3.342019824115594,
      "tokens_seen": 2762997760
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.665216568942689e-06,
      "loss": 2.7778,
      "theoretical_loss": 3.3420136270086602,
      "tokens_seen": 2763063296
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.648666540571213e-06,
      "loss": 2.1529,
      "theoretical_loss": 3.3420074300898666,
      "tokens_seen": 2763128832
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.632116512199735e-06,
      "loss": 2.1774,
      "theoretical_loss": 3.342001233359203,
      "tokens_seen": 2763194368
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.615566483828257e-06,
      "loss": 2.3567,
      "theoretical_loss": 3.341995036816659,
      "tokens_seen": 2763259904
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.599016455456781e-06,
      "loss": 2.3254,
      "theoretical_loss": 3.3419888404622244,
      "tokens_seen": 2763325440
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.582466427085303e-06,
      "loss": 2.7987,
      "theoretical_loss": 3.3419826442958893,
      "tokens_seen": 2763390976
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.565916398713827e-06,
      "loss": 2.2835,
      "theoretical_loss": 3.3419764483176437,
      "tokens_seen": 2763456512
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.54936637034235e-06,
      "loss": 2.434,
      "theoretical_loss": 3.3419702525274775,
      "tokens_seen": 2763522048
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.532816341970872e-06,
      "loss": 2.0594,
      "theoretical_loss": 3.34196405692538,
      "tokens_seen": 2763587584
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.516266313599394e-06,
      "loss": 2.7302,
      "theoretical_loss": 3.341957861511341,
      "tokens_seen": 2763653120
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.499716285227917e-06,
      "loss": 2.0364,
      "theoretical_loss": 3.341951666285351,
      "tokens_seen": 2763718656
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.48316625685644e-06,
      "loss": 4.2386,
      "theoretical_loss": 3.341945471247399,
      "tokens_seen": 2763784192
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.466616228484963e-06,
      "loss": 2.6,
      "theoretical_loss": 3.3419392763974756,
      "tokens_seen": 2763849728
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.450066200113485e-06,
      "loss": 1.9855,
      "theoretical_loss": 3.34193308173557,
      "tokens_seen": 2763915264
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.43351617174201e-06,
      "loss": 3.7688,
      "theoretical_loss": 3.3419268872616725,
      "tokens_seen": 2763980800
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.416966143370532e-06,
      "loss": 1.7111,
      "theoretical_loss": 3.3419206929757728,
      "tokens_seen": 2764046336
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.400416114999054e-06,
      "loss": 2.4329,
      "theoretical_loss": 3.341914498877861,
      "tokens_seen": 2764111872
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.383866086627578e-06,
      "loss": 2.0061,
      "theoretical_loss": 3.341908304967926,
      "tokens_seen": 2764177408
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.3673160582561e-06,
      "loss": 2.2226,
      "theoretical_loss": 3.341902111245959,
      "tokens_seen": 2764242944
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.350766029884622e-06,
      "loss": 2.9879,
      "theoretical_loss": 3.3418959177119487,
      "tokens_seen": 2764308480
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.334216001513146e-06,
      "loss": 2.465,
      "theoretical_loss": 3.3418897243658856,
      "tokens_seen": 2764374016
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.317665973141669e-06,
      "loss": 2.289,
      "theoretical_loss": 3.341883531207759,
      "tokens_seen": 2764439552
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.301115944770191e-06,
      "loss": 2.2954,
      "theoretical_loss": 3.341877338237559,
      "tokens_seen": 2764505088
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.284565916398713e-06,
      "loss": -0.2385,
      "theoretical_loss": 3.341871145455276,
      "tokens_seen": 2764570624
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.268015888027236e-06,
      "loss": 2.9084,
      "theoretical_loss": 3.341864952860899,
      "tokens_seen": 2764636160
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.25146585965576e-06,
      "loss": 2.4527,
      "theoretical_loss": 3.3418587604544183,
      "tokens_seen": 2764701696
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.234915831284282e-06,
      "loss": 2.1434,
      "theoretical_loss": 3.3418525682358236,
      "tokens_seen": 2764767232
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.218365802912804e-06,
      "loss": 2.5584,
      "theoretical_loss": 3.3418463762051047,
      "tokens_seen": 2764832768
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.201815774541328e-06,
      "loss": 2.6957,
      "theoretical_loss": 3.3418401843622516,
      "tokens_seen": 2764898304
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.18526574616985e-06,
      "loss": 2.0192,
      "theoretical_loss": 3.3418339927072545,
      "tokens_seen": 2764963840
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.168715717798373e-06,
      "loss": 2.4089,
      "theoretical_loss": 3.341827801240102,
      "tokens_seen": 2765029376
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.152165689426897e-06,
      "loss": 2.102,
      "theoretical_loss": 3.3418216099607854,
      "tokens_seen": 2765094912
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.135615661055419e-06,
      "loss": 2.5668,
      "theoretical_loss": 3.3418154188692935,
      "tokens_seen": 2765160448
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.119065632683941e-06,
      "loss": 1.6089,
      "theoretical_loss": 3.3418092279656166,
      "tokens_seen": 2765225984
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.102515604312465e-06,
      "loss": 2.7014,
      "theoretical_loss": 3.3418030372497447,
      "tokens_seen": 2765291520
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.085965575940986e-06,
      "loss": 2.4664,
      "theoretical_loss": 3.3417968467216674,
      "tokens_seen": 2765357056
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.06941554756951e-06,
      "loss": 2.6751,
      "theoretical_loss": 3.3417906563813746,
      "tokens_seen": 2765422592
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.052865519198032e-06,
      "loss": 2.7206,
      "theoretical_loss": 3.341784466228856,
      "tokens_seen": 2765488128
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.036315490826555e-06,
      "loss": 1.3445,
      "theoretical_loss": 3.341778276264102,
      "tokens_seen": 2765553664
    },
    {
      "epoch": 0.99,
      "objective/train/avg_token_score": 0.019885048270225525,
      "objective/train/avg_weight": 0.9765907526016235,
      "objective/train/docs_used": 1555428,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 8.301036834716797,
      "objective/train/std_weight": 0.0837535634636879,
      "objective/train/theoretical_loss": 3.341772086487102,
      "objective/train/tokens_used": 2786079200,
      "theoretical_loss": 3.341772086487102,
      "tokens_seen": 2765619200
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.019765462455078e-06,
      "loss": 2.302,
      "theoretical_loss": 3.341772086487102,
      "tokens_seen": 2765619200
    },
    {
      "epoch": 0.99,
      "learning_rate": 9.0032154340836e-06,
      "loss": 2.5781,
      "theoretical_loss": 3.341765896897846,
      "tokens_seen": 2765684736
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.986665405712125e-06,
      "loss": 1.5001,
      "theoretical_loss": 3.3417597074963235,
      "tokens_seen": 2765750272
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.970115377340647e-06,
      "loss": 2.8363,
      "theoretical_loss": 3.341753518282525,
      "tokens_seen": 2765815808
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.95356534896917e-06,
      "loss": 2.3607,
      "theoretical_loss": 3.3417473292564397,
      "tokens_seen": 2765881344
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.937015320597693e-06,
      "loss": 2.1525,
      "theoretical_loss": 3.3417411404180584,
      "tokens_seen": 2765946880
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.920465292226216e-06,
      "loss": 2.6689,
      "theoretical_loss": 3.3417349517673696,
      "tokens_seen": 2766012416
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.903915263854738e-06,
      "loss": 1.7999,
      "theoretical_loss": 3.3417287633043644,
      "tokens_seen": 2766077952
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.887365235483262e-06,
      "loss": 2.2536,
      "theoretical_loss": 3.341722575029032,
      "tokens_seen": 2766143488
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.870815207111783e-06,
      "loss": 2.5836,
      "theoretical_loss": 3.341716386941362,
      "tokens_seen": 2766209024
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.854265178740307e-06,
      "loss": 2.7858,
      "theoretical_loss": 3.341710199041345,
      "tokens_seen": 2766274560
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.837715150368829e-06,
      "loss": 1.6282,
      "theoretical_loss": 3.3417040113289707,
      "tokens_seen": 2766340096
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.821165121997351e-06,
      "loss": 2.3311,
      "theoretical_loss": 3.3416978238042283,
      "tokens_seen": 2766405632
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.804615093625875e-06,
      "loss": 2.7326,
      "theoretical_loss": 3.3416916364671088,
      "tokens_seen": 2766471168
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.788065065254397e-06,
      "loss": 2.7177,
      "theoretical_loss": 3.341685449317601,
      "tokens_seen": 2766536704
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.77151503688292e-06,
      "loss": 2.3997,
      "theoretical_loss": 3.3416792623556955,
      "tokens_seen": 2766602240
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.754965008511444e-06,
      "loss": 2.1455,
      "theoretical_loss": 3.341673075581382,
      "tokens_seen": 2766667776
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.738414980139966e-06,
      "loss": 1.3453,
      "theoretical_loss": 3.34166688899465,
      "tokens_seen": 2766733312
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.721864951768488e-06,
      "loss": 2.7775,
      "theoretical_loss": 3.3416607025954894,
      "tokens_seen": 2766798848
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.705314923397012e-06,
      "loss": 3.7606,
      "theoretical_loss": 3.341654516383891,
      "tokens_seen": 2766864384
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.688764895025535e-06,
      "loss": 2.5532,
      "theoretical_loss": 3.341648330359843,
      "tokens_seen": 2766929920
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.672214866654057e-06,
      "loss": 1.9335,
      "theoretical_loss": 3.341642144523337,
      "tokens_seen": 2766995456
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.65566483828258e-06,
      "loss": 2.8139,
      "theoretical_loss": 3.341635958874362,
      "tokens_seen": 2767060992
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.639114809911102e-06,
      "loss": 2.9099,
      "theoretical_loss": 3.3416297734129077,
      "tokens_seen": 2767126528
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.622564781539625e-06,
      "loss": 2.4487,
      "theoretical_loss": 3.3416235881389644,
      "tokens_seen": 2767192064
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.606014753168148e-06,
      "loss": 2.5285,
      "theoretical_loss": 3.341617403052522,
      "tokens_seen": 2767257600
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.58946472479667e-06,
      "loss": 2.6551,
      "theoretical_loss": 3.3416112181535698,
      "tokens_seen": 2767323136
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.572914696425194e-06,
      "loss": 2.475,
      "theoretical_loss": 3.3416050334420984,
      "tokens_seen": 2767388672
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.556364668053716e-06,
      "loss": 2.1449,
      "theoretical_loss": 3.341598848918097,
      "tokens_seen": 2767454208
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.539814639682239e-06,
      "loss": 1.8598,
      "theoretical_loss": 3.3415926645815563,
      "tokens_seen": 2767519744
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.523264611310763e-06,
      "loss": 2.5813,
      "theoretical_loss": 3.3415864804324653,
      "tokens_seen": 2767585280
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.506714582939285e-06,
      "loss": 2.3564,
      "theoretical_loss": 3.3415802964708146,
      "tokens_seen": 2767650816
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.490164554567809e-06,
      "loss": 2.438,
      "theoretical_loss": 3.3415741126965934,
      "tokens_seen": 2767716352
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.473614526196331e-06,
      "loss": 2.3507,
      "theoretical_loss": 3.341567929109792,
      "tokens_seen": 2767781888
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.457064497824854e-06,
      "loss": -2.5719,
      "theoretical_loss": 3.3415617457104005,
      "tokens_seen": 2767847424
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.440514469453376e-06,
      "loss": 1.9809,
      "theoretical_loss": 3.341555562498409,
      "tokens_seen": 2767912960
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.423964441081898e-06,
      "loss": 2.3476,
      "theoretical_loss": 3.341549379473806,
      "tokens_seen": 2767978496
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.407414412710422e-06,
      "loss": 2.8058,
      "theoretical_loss": 3.3415431966365827,
      "tokens_seen": 2768044032
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.390864384338944e-06,
      "loss": 2.9367,
      "theoretical_loss": 3.3415370139867284,
      "tokens_seen": 2768109568
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.374314355967467e-06,
      "loss": 2.7068,
      "theoretical_loss": 3.341530831524233,
      "tokens_seen": 2768175104
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.35776432759599e-06,
      "loss": 1.5934,
      "theoretical_loss": 3.341524649249087,
      "tokens_seen": 2768240640
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.341214299224513e-06,
      "loss": 1.7021,
      "theoretical_loss": 3.34151846716128,
      "tokens_seen": 2768306176
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.324664270853035e-06,
      "loss": 2.6696,
      "theoretical_loss": 3.3415122852608015,
      "tokens_seen": 2768371712
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.30811424248156e-06,
      "loss": 2.6784,
      "theoretical_loss": 3.3415061035476414,
      "tokens_seen": 2768437248
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.291564214110082e-06,
      "loss": 2.5442,
      "theoretical_loss": 3.34149992202179,
      "tokens_seen": 2768502784
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.275014185738604e-06,
      "loss": 2.5421,
      "theoretical_loss": 3.3414937406832363,
      "tokens_seen": 2768568320
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.258464157367128e-06,
      "loss": 2.7135,
      "theoretical_loss": 3.3414875595319717,
      "tokens_seen": 2768633856
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.24191412899565e-06,
      "loss": 2.0782,
      "theoretical_loss": 3.341481378567985,
      "tokens_seen": 2768699392
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.225364100624172e-06,
      "loss": 3.3063,
      "theoretical_loss": 3.3414751977912664,
      "tokens_seen": 2768764928
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.208814072252695e-06,
      "loss": 2.0383,
      "theoretical_loss": 3.341469017201806,
      "tokens_seen": 2768830464
    },
    {
      "epoch": 0.99,
      "objective/train/avg_token_score": 0.012067302130162716,
      "objective/train/avg_weight": 0.9858004450798035,
      "objective/train/docs_used": 1557592,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.304986953735352,
      "objective/train/std_weight": 0.056058939546346664,
      "objective/train/theoretical_loss": 3.341462836799593,
      "objective/train/tokens_used": 2789356000,
      "theoretical_loss": 3.341462836799593,
      "tokens_seen": 2768896000
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.192264043881217e-06,
      "loss": 2.3868,
      "theoretical_loss": 3.341462836799593,
      "tokens_seen": 2768896000
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.175714015509741e-06,
      "loss": 2.2609,
      "theoretical_loss": 3.341456656584618,
      "tokens_seen": 2768961536
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.159163987138263e-06,
      "loss": 2.5265,
      "theoretical_loss": 3.3414504765568704,
      "tokens_seen": 2769027072
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.142613958766786e-06,
      "loss": 2.6148,
      "theoretical_loss": 3.341444296716341,
      "tokens_seen": 2769092608
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.12606393039531e-06,
      "loss": 2.7175,
      "theoretical_loss": 3.3414381170630185,
      "tokens_seen": 2769158144
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.109513902023832e-06,
      "loss": 2.6431,
      "theoretical_loss": 3.3414319375968935,
      "tokens_seen": 2769223680
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.092963873652354e-06,
      "loss": 1.8787,
      "theoretical_loss": 3.341425758317955,
      "tokens_seen": 2769289216
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.076413845280878e-06,
      "loss": 5.5129,
      "theoretical_loss": 3.3414195792261947,
      "tokens_seen": 2769354752
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.0598638169094e-06,
      "loss": 2.6887,
      "theoretical_loss": 3.341413400321601,
      "tokens_seen": 2769420288
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.043313788537923e-06,
      "loss": 2.5342,
      "theoretical_loss": 3.3414072216041646,
      "tokens_seen": 2769485824
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.026763760166447e-06,
      "loss": 1.9498,
      "theoretical_loss": 3.3414010430738745,
      "tokens_seen": 2769551360
    },
    {
      "epoch": 0.99,
      "learning_rate": 8.010213731794967e-06,
      "loss": 2.5506,
      "theoretical_loss": 3.3413948647307214,
      "tokens_seen": 2769616896
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.993663703423491e-06,
      "loss": 2.5546,
      "theoretical_loss": 3.341388686574695,
      "tokens_seen": 2769682432
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.977113675052014e-06,
      "loss": 2.614,
      "theoretical_loss": 3.3413825086057853,
      "tokens_seen": 2769747968
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.960563646680536e-06,
      "loss": 0.5955,
      "theoretical_loss": 3.341376330823982,
      "tokens_seen": 2769813504
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.94401361830906e-06,
      "loss": 2.6882,
      "theoretical_loss": 3.341370153229275,
      "tokens_seen": 2769879040
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.927463589937582e-06,
      "loss": 1.7684,
      "theoretical_loss": 3.341363975821654,
      "tokens_seen": 2769944576
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.910913561566106e-06,
      "loss": 1.7889,
      "theoretical_loss": 3.341357798601109,
      "tokens_seen": 2770010112
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.894363533194629e-06,
      "loss": 2.3483,
      "theoretical_loss": 3.3413516215676307,
      "tokens_seen": 2770075648
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.877813504823151e-06,
      "loss": 2.9978,
      "theoretical_loss": 3.3413454447212083,
      "tokens_seen": 2770141184
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.861263476451675e-06,
      "loss": 2.6981,
      "theoretical_loss": 3.341339268061832,
      "tokens_seen": 2770206720
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.844713448080197e-06,
      "loss": 2.6448,
      "theoretical_loss": 3.341333091589491,
      "tokens_seen": 2770272256
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.82816341970872e-06,
      "loss": 2.6755,
      "theoretical_loss": 3.341326915304176,
      "tokens_seen": 2770337792
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.811613391337243e-06,
      "loss": 2.437,
      "theoretical_loss": 3.3413207392058766,
      "tokens_seen": 2770403328
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.795063362965764e-06,
      "loss": 2.6881,
      "theoretical_loss": 3.3413145632945827,
      "tokens_seen": 2770468864
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.778513334594288e-06,
      "loss": 2.3044,
      "theoretical_loss": 3.3413083875702845,
      "tokens_seen": 2770534400
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.76196330622281e-06,
      "loss": 2.3683,
      "theoretical_loss": 3.3413022120329714,
      "tokens_seen": 2770599936
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.745413277851333e-06,
      "loss": 2.6046,
      "theoretical_loss": 3.3412960366826336,
      "tokens_seen": 2770665472
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.728863249479857e-06,
      "loss": 2.5686,
      "theoretical_loss": 3.3412898615192614,
      "tokens_seen": 2770731008
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.712313221108379e-06,
      "loss": 2.2852,
      "theoretical_loss": 3.341283686542844,
      "tokens_seen": 2770796544
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.695763192736901e-06,
      "loss": 1.8181,
      "theoretical_loss": 3.341277511753372,
      "tokens_seen": 2770862080
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.679213164365425e-06,
      "loss": 2.4093,
      "theoretical_loss": 3.3412713371508347,
      "tokens_seen": 2770927616
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.662663135993948e-06,
      "loss": 2.6177,
      "theoretical_loss": 3.3412651627352226,
      "tokens_seen": 2770993152
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.64611310762247e-06,
      "loss": 2.556,
      "theoretical_loss": 3.341258988506525,
      "tokens_seen": 2771058688
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.629563079250994e-06,
      "loss": 2.6066,
      "theoretical_loss": 3.341252814464732,
      "tokens_seen": 2771124224
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.613013050879516e-06,
      "loss": 3.0846,
      "theoretical_loss": 3.341246640609834,
      "tokens_seen": 2771189760
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.596463022508039e-06,
      "loss": 1.7028,
      "theoretical_loss": 3.3412404669418203,
      "tokens_seen": 2771255296
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.579912994136561e-06,
      "loss": 2.7525,
      "theoretical_loss": 3.3412342934606816,
      "tokens_seen": 2771320832
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.563362965765084e-06,
      "loss": 2.7834,
      "theoretical_loss": 3.3412281201664067,
      "tokens_seen": 2771386368
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.546812937393606e-06,
      "loss": 1.8754,
      "theoretical_loss": 3.341221947058987,
      "tokens_seen": 2771451904
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.530262909022129e-06,
      "loss": 2.5742,
      "theoretical_loss": 3.3412157741384108,
      "tokens_seen": 2771517440
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.5137128806506525e-06,
      "loss": 2.5974,
      "theoretical_loss": 3.341209601404669,
      "tokens_seen": 2771582976
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.497162852279176e-06,
      "loss": 2.1639,
      "theoretical_loss": 3.3412034288577512,
      "tokens_seen": 2771648512
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.480612823907698e-06,
      "loss": 2.2111,
      "theoretical_loss": 3.3411972564976478,
      "tokens_seen": 2771714048
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.464062795536221e-06,
      "loss": 2.5566,
      "theoretical_loss": 3.341191084324348,
      "tokens_seen": 2771779584
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.447512767164744e-06,
      "loss": 3.4093,
      "theoretical_loss": 3.3411849123378428,
      "tokens_seen": 2771845120
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.4309627387932665e-06,
      "loss": 2.643,
      "theoretical_loss": 3.341178740538121,
      "tokens_seen": 2771910656
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.41441271042179e-06,
      "loss": 2.9449,
      "theoretical_loss": 3.341172568925173,
      "tokens_seen": 2771976192
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.397862682050313e-06,
      "loss": 2.1983,
      "theoretical_loss": 3.3411663974989887,
      "tokens_seen": 2772041728
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.381312653678835e-06,
      "loss": 2.5094,
      "theoretical_loss": 3.341160226259558,
      "tokens_seen": 2772107264
    },
    {
      "epoch": 0.99,
      "objective/train/avg_token_score": 0.00911212619394064,
      "objective/train/avg_weight": 0.9892752170562744,
      "objective/train/docs_used": 1559642,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.720419406890869,
      "objective/train/std_weight": 0.03886338695883751,
      "objective/train/theoretical_loss": 3.341154055206871,
      "objective/train/tokens_used": 2792632800,
      "theoretical_loss": 3.341154055206871,
      "tokens_seen": 2772172800
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.364762625307357e-06,
      "loss": 2.5244,
      "theoretical_loss": 3.341154055206871,
      "tokens_seen": 2772172800
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.34821259693588e-06,
      "loss": 2.473,
      "theoretical_loss": 3.3411478843409177,
      "tokens_seen": 2772238336
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.331662568564403e-06,
      "loss": 2.4152,
      "theoretical_loss": 3.3411417136616874,
      "tokens_seen": 2772303872
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.315112540192926e-06,
      "loss": 3.4944,
      "theoretical_loss": 3.341135543169171,
      "tokens_seen": 2772369408
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.298562511821448e-06,
      "loss": 2.6503,
      "theoretical_loss": 3.3411293728633575,
      "tokens_seen": 2772434944
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.282012483449971e-06,
      "loss": 2.7402,
      "theoretical_loss": 3.3411232027442375,
      "tokens_seen": 2772500480
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.2654624550784946e-06,
      "loss": -0.0784,
      "theoretical_loss": 3.341117032811801,
      "tokens_seen": 2772566016
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.248912426707018e-06,
      "loss": 2.5221,
      "theoretical_loss": 3.341110863066037,
      "tokens_seen": 2772631552
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.23236239833554e-06,
      "loss": 3.0652,
      "theoretical_loss": 3.3411046935069364,
      "tokens_seen": 2772697088
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.215812369964063e-06,
      "loss": 2.8808,
      "theoretical_loss": 3.341098524134489,
      "tokens_seen": 2772762624
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.199262341592586e-06,
      "loss": 2.5815,
      "theoretical_loss": 3.3410923549486844,
      "tokens_seen": 2772828160
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.182712313221109e-06,
      "loss": 2.627,
      "theoretical_loss": 3.3410861859495125,
      "tokens_seen": 2772893696
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.166162284849631e-06,
      "loss": 1.923,
      "theoretical_loss": 3.3410800171369637,
      "tokens_seen": 2772959232
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.149612256478153e-06,
      "loss": 2.6191,
      "theoretical_loss": 3.3410738485110274,
      "tokens_seen": 2773024768
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.133062228106676e-06,
      "loss": 2.2112,
      "theoretical_loss": 3.3410676800716943,
      "tokens_seen": 2773090304
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.1165121997351995e-06,
      "loss": 2.6804,
      "theoretical_loss": 3.3410615118189537,
      "tokens_seen": 2773155840
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.099962171363722e-06,
      "loss": 2.6228,
      "theoretical_loss": 3.341055343752796,
      "tokens_seen": 2773221376
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.083412142992245e-06,
      "loss": 2.6075,
      "theoretical_loss": 3.34104917587321,
      "tokens_seen": 2773286912
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.066862114620768e-06,
      "loss": 2.0573,
      "theoretical_loss": 3.341043008180187,
      "tokens_seen": 2773352448
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.05031208624929e-06,
      "loss": 2.0175,
      "theoretical_loss": 3.3410368406737168,
      "tokens_seen": 2773417984
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.0337620578778135e-06,
      "loss": 2.5801,
      "theoretical_loss": 3.3410306733537887,
      "tokens_seen": 2773483520
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.017212029506337e-06,
      "loss": 2.6608,
      "theoretical_loss": 3.3410245062203936,
      "tokens_seen": 2773549056
    },
    {
      "epoch": 0.99,
      "learning_rate": 7.000662001134859e-06,
      "loss": 2.4861,
      "theoretical_loss": 3.34101833927352,
      "tokens_seen": 2773614592
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.984111972763382e-06,
      "loss": 2.8429,
      "theoretical_loss": 3.341012172513159,
      "tokens_seen": 2773680128
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.967561944391905e-06,
      "loss": 2.6156,
      "theoretical_loss": 3.3410060059393003,
      "tokens_seen": 2773745664
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.951011916020427e-06,
      "loss": 2.5289,
      "theoretical_loss": 3.340999839551934,
      "tokens_seen": 2773811200
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.93446188764895e-06,
      "loss": 2.5447,
      "theoretical_loss": 3.3409936733510492,
      "tokens_seen": 2773876736
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.917911859277473e-06,
      "loss": 1.8024,
      "theoretical_loss": 3.340987507336637,
      "tokens_seen": 2773942272
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.901361830905995e-06,
      "loss": 2.1114,
      "theoretical_loss": 3.3409813415086864,
      "tokens_seen": 2774007808
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.884811802534518e-06,
      "loss": 2.7953,
      "theoretical_loss": 3.340975175867188,
      "tokens_seen": 2774073344
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.8682617741630416e-06,
      "loss": 2.8015,
      "theoretical_loss": 3.340969010412132,
      "tokens_seen": 2774138880
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.851711745791564e-06,
      "loss": 2.0088,
      "theoretical_loss": 3.3409628451435074,
      "tokens_seen": 2774204416
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.835161717420087e-06,
      "loss": 2.4632,
      "theoretical_loss": 3.340956680061305,
      "tokens_seen": 2774269952
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.81861168904861e-06,
      "loss": 2.4489,
      "theoretical_loss": 3.340950515165514,
      "tokens_seen": 2774335488
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.8020616606771324e-06,
      "loss": 2.7064,
      "theoretical_loss": 3.340944350456125,
      "tokens_seen": 2774401024
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.785511632305656e-06,
      "loss": 2.7215,
      "theoretical_loss": 3.340938185933128,
      "tokens_seen": 2774466560
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.768961603934179e-06,
      "loss": 2.7719,
      "theoretical_loss": 3.3409320215965126,
      "tokens_seen": 2774532096
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.752411575562701e-06,
      "loss": 2.4407,
      "theoretical_loss": 3.340925857446269,
      "tokens_seen": 2774597632
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.735861547191223e-06,
      "loss": 2.6178,
      "theoretical_loss": 3.3409196934823866,
      "tokens_seen": 2774663168
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.719311518819746e-06,
      "loss": 2.3011,
      "theoretical_loss": 3.3409135297048564,
      "tokens_seen": 2774728704
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.702761490448269e-06,
      "loss": 2.0649,
      "theoretical_loss": 3.340907366113667,
      "tokens_seen": 2774794240
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.686211462076792e-06,
      "loss": 3.2094,
      "theoretical_loss": 3.34090120270881,
      "tokens_seen": 2774859776
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.669661433705315e-06,
      "loss": 2.5946,
      "theoretical_loss": 3.340895039490274,
      "tokens_seen": 2774925312
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.653111405333837e-06,
      "loss": 2.3743,
      "theoretical_loss": 3.3408888764580498,
      "tokens_seen": 2774990848
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.6365613769623605e-06,
      "loss": 2.4946,
      "theoretical_loss": 3.3408827136121264,
      "tokens_seen": 2775056384
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.620011348590884e-06,
      "loss": 0.1162,
      "theoretical_loss": 3.340876550952495,
      "tokens_seen": 2775121920
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.603461320219406e-06,
      "loss": 2.9005,
      "theoretical_loss": 3.340870388479145,
      "tokens_seen": 2775187456
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.586911291847929e-06,
      "loss": 2.8915,
      "theoretical_loss": 3.3408642261920662,
      "tokens_seen": 2775252992
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.570361263476452e-06,
      "loss": 2.7292,
      "theoretical_loss": 3.3408580640912486,
      "tokens_seen": 2775318528
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.5538112351049745e-06,
      "loss": 2.8254,
      "theoretical_loss": 3.3408519021766825,
      "tokens_seen": 2775384064
    },
    {
      "epoch": 0.99,
      "objective/train/avg_token_score": 0.010162509977817535,
      "objective/train/avg_weight": 0.9880372285842896,
      "objective/train/docs_used": 1561371,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 11.961865425109863,
      "objective/train/std_weight": 0.07398331165313721,
      "objective/train/theoretical_loss": 3.3408457404483576,
      "objective/train/tokens_used": 2795909600,
      "theoretical_loss": 3.3408457404483576,
      "tokens_seen": 2775449600
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.537261206733498e-06,
      "loss": 1.2157,
      "theoretical_loss": 3.3408457404483576,
      "tokens_seen": 2775449600
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.520711178362019e-06,
      "loss": 2.7915,
      "theoretical_loss": 3.3408395789062637,
      "tokens_seen": 2775515136
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.504161149990542e-06,
      "loss": 2.6911,
      "theoretical_loss": 3.3408334175503915,
      "tokens_seen": 2775580672
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.487611121619065e-06,
      "loss": 1.5503,
      "theoretical_loss": 3.34082725638073,
      "tokens_seen": 2775646208
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.471061093247588e-06,
      "loss": 2.5535,
      "theoretical_loss": 3.3408210953972697,
      "tokens_seen": 2775711744
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.454511064876111e-06,
      "loss": 2.472,
      "theoretical_loss": 3.340814934600001,
      "tokens_seen": 2775777280
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.437961036504634e-06,
      "loss": 2.5622,
      "theoretical_loss": 3.340808773988913,
      "tokens_seen": 2775842816
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.421411008133157e-06,
      "loss": 2.2147,
      "theoretical_loss": 3.3408026135639965,
      "tokens_seen": 2775908352
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.4048609797616794e-06,
      "loss": 2.3584,
      "theoretical_loss": 3.3407964533252406,
      "tokens_seen": 2775973888
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.388310951390203e-06,
      "loss": 2.5862,
      "theoretical_loss": 3.340790293272636,
      "tokens_seen": 2776039424
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.371760923018726e-06,
      "loss": 2.4538,
      "theoretical_loss": 3.3407841334061725,
      "tokens_seen": 2776104960
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.355210894647248e-06,
      "loss": 1.3664,
      "theoretical_loss": 3.34077797372584,
      "tokens_seen": 2776170496
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.338660866275771e-06,
      "loss": 2.3358,
      "theoretical_loss": 3.3407718142316285,
      "tokens_seen": 2776236032
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.322110837904294e-06,
      "loss": 1.8418,
      "theoretical_loss": 3.340765654923528,
      "tokens_seen": 2776301568
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.305560809532816e-06,
      "loss": 2.2369,
      "theoretical_loss": 3.340759495801528,
      "tokens_seen": 2776367104
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.289010781161339e-06,
      "loss": 4.2473,
      "theoretical_loss": 3.3407533368656197,
      "tokens_seen": 2776432640
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.272460752789861e-06,
      "loss": 2.6109,
      "theoretical_loss": 3.340747178115792,
      "tokens_seen": 2776498176
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.255910724418384e-06,
      "loss": 2.7366,
      "theoretical_loss": 3.340741019552035,
      "tokens_seen": 2776563712
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.2393606960469075e-06,
      "loss": 2.4511,
      "theoretical_loss": 3.3407348611743393,
      "tokens_seen": 2776629248
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.22281066767543e-06,
      "loss": 2.5772,
      "theoretical_loss": 3.340728702982694,
      "tokens_seen": 2776694784
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.206260639303953e-06,
      "loss": 3.0951,
      "theoretical_loss": 3.34072254497709,
      "tokens_seen": 2776760320
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.189710610932476e-06,
      "loss": 1.0789,
      "theoretical_loss": 3.340716387157517,
      "tokens_seen": 2776825856
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.173160582560998e-06,
      "loss": 2.3072,
      "theoretical_loss": 3.3407102295239643,
      "tokens_seen": 2776891392
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.1566105541895215e-06,
      "loss": 1.7463,
      "theoretical_loss": 3.340704072076423,
      "tokens_seen": 2776956928
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.140060525818045e-06,
      "loss": 2.5885,
      "theoretical_loss": 3.340697914814882,
      "tokens_seen": 2777022464
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.123510497446568e-06,
      "loss": 2.8096,
      "theoretical_loss": 3.3406917577393322,
      "tokens_seen": 2777088000
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.10696046907509e-06,
      "loss": 2.5793,
      "theoretical_loss": 3.3406856008497634,
      "tokens_seen": 2777153536
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.090410440703612e-06,
      "loss": 2.68,
      "theoretical_loss": 3.3406794441461654,
      "tokens_seen": 2777219072
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.073860412332135e-06,
      "loss": 1.0235,
      "theoretical_loss": 3.340673287628528,
      "tokens_seen": 2777284608
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.057310383960658e-06,
      "loss": 2.0872,
      "theoretical_loss": 3.3406671312968417,
      "tokens_seen": 2777350144
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.040760355589181e-06,
      "loss": 2.7045,
      "theoretical_loss": 3.3406609751510956,
      "tokens_seen": 2777415680
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.024210327217703e-06,
      "loss": 2.1656,
      "theoretical_loss": 3.340654819191281,
      "tokens_seen": 2777481216
    },
    {
      "epoch": 0.99,
      "learning_rate": 6.0076602988462264e-06,
      "loss": 2.4635,
      "theoretical_loss": 3.3406486634173866,
      "tokens_seen": 2777546752
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.99111027047475e-06,
      "loss": 2.5164,
      "theoretical_loss": 3.3406425078294033,
      "tokens_seen": 2777612288
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.974560242103272e-06,
      "loss": 2.5611,
      "theoretical_loss": 3.3406363524273206,
      "tokens_seen": 2777677824
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.958010213731795e-06,
      "loss": 2.4684,
      "theoretical_loss": 3.340630197211129,
      "tokens_seen": 2777743360
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.941460185360318e-06,
      "loss": 2.5958,
      "theoretical_loss": 3.340624042180818,
      "tokens_seen": 2777808896
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.9249101569888405e-06,
      "loss": 2.5953,
      "theoretical_loss": 3.340617887336378,
      "tokens_seen": 2777874432
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.908360128617364e-06,
      "loss": 2.5212,
      "theoretical_loss": 3.3406117326777984,
      "tokens_seen": 2777939968
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.891810100245887e-06,
      "loss": 2.6038,
      "theoretical_loss": 3.34060557820507,
      "tokens_seen": 2778005504
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.875260071874408e-06,
      "loss": 1.8683,
      "theoretical_loss": 3.3405994239181824,
      "tokens_seen": 2778071040
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.858710043502931e-06,
      "loss": 2.3756,
      "theoretical_loss": 3.3405932698171257,
      "tokens_seen": 2778136576
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.8421600151314545e-06,
      "loss": 2.4545,
      "theoretical_loss": 3.3405871159018896,
      "tokens_seen": 2778202112
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.825609986759977e-06,
      "loss": 2.7031,
      "theoretical_loss": 3.3405809621724645,
      "tokens_seen": 2778267648
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.8090599583885e-06,
      "loss": 2.4234,
      "theoretical_loss": 3.3405748086288396,
      "tokens_seen": 2778333184
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.792509930017023e-06,
      "loss": 2.4834,
      "theoretical_loss": 3.340568655271006,
      "tokens_seen": 2778398720
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.775959901645545e-06,
      "loss": 2.7303,
      "theoretical_loss": 3.340562502098954,
      "tokens_seen": 2778464256
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.7594098732740685e-06,
      "loss": 2.5445,
      "theoretical_loss": 3.3405563491126715,
      "tokens_seen": 2778529792
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.742859844902592e-06,
      "loss": 2.6281,
      "theoretical_loss": 3.340550196312151,
      "tokens_seen": 2778595328
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.726309816531114e-06,
      "loss": 2.6175,
      "theoretical_loss": 3.340544043697381,
      "tokens_seen": 2778660864
    },
    {
      "epoch": 0.99,
      "objective/train/avg_token_score": 0.016881663352251053,
      "objective/train/avg_weight": 0.9801218509674072,
      "objective/train/docs_used": 1563208,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 5.208395004272461,
      "objective/train/std_weight": 0.10705780982971191,
      "objective/train/theoretical_loss": 3.3405378912683514,
      "objective/train/tokens_used": 2799186400,
      "theoretical_loss": 3.3405378912683514,
      "tokens_seen": 2778726400
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.709759788159637e-06,
      "loss": 0.5403,
      "theoretical_loss": 3.3405378912683514,
      "tokens_seen": 2778726400
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.69320975978816e-06,
      "loss": 2.6091,
      "theoretical_loss": 3.340531739025053,
      "tokens_seen": 2778791936
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.6766597314166825e-06,
      "loss": 2.6219,
      "theoretical_loss": 3.340525586967476,
      "tokens_seen": 2778857472
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.660109703045205e-06,
      "loss": 3.4666,
      "theoretical_loss": 3.3405194350956093,
      "tokens_seen": 2778923008
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.643559674673727e-06,
      "loss": 1.076,
      "theoretical_loss": 3.340513283409444,
      "tokens_seen": 2778988544
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.62700964630225e-06,
      "loss": 2.1425,
      "theoretical_loss": 3.340507131908969,
      "tokens_seen": 2779054080
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.6104596179307734e-06,
      "loss": 2.2585,
      "theoretical_loss": 3.3405009805941757,
      "tokens_seen": 2779119616
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.593909589559296e-06,
      "loss": 2.1398,
      "theoretical_loss": 3.340494829465053,
      "tokens_seen": 2779185152
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.577359561187819e-06,
      "loss": 2.0448,
      "theoretical_loss": 3.3404886785215915,
      "tokens_seen": 2779250688
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.560809532816342e-06,
      "loss": 2.32,
      "theoretical_loss": 3.340482527763781,
      "tokens_seen": 2779316224
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.544259504444865e-06,
      "loss": 2.6973,
      "theoretical_loss": 3.3404763771916115,
      "tokens_seen": 2779381760
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.5277094760733875e-06,
      "loss": 2.6313,
      "theoretical_loss": 3.340470226805073,
      "tokens_seen": 2779447296
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.511159447701911e-06,
      "loss": 2.6805,
      "theoretical_loss": 3.340464076604156,
      "tokens_seen": 2779512832
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.494609419330434e-06,
      "loss": 2.5291,
      "theoretical_loss": 3.3404579265888494,
      "tokens_seen": 2779578368
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.478059390958956e-06,
      "loss": 2.5051,
      "theoretical_loss": 3.3404517767591444,
      "tokens_seen": 2779643904
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.461509362587478e-06,
      "loss": 2.8469,
      "theoretical_loss": 3.3404456271150305,
      "tokens_seen": 2779709440
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.4449593342160015e-06,
      "loss": 2.1791,
      "theoretical_loss": 3.3404394776564974,
      "tokens_seen": 2779774976
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.428409305844525e-06,
      "loss": 2.3165,
      "theoretical_loss": 3.340433328383536,
      "tokens_seen": 2779840512
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.411859277473047e-06,
      "loss": 2.5001,
      "theoretical_loss": 3.3404271792961358,
      "tokens_seen": 2779906048
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.395309249101569e-06,
      "loss": 1.96,
      "theoretical_loss": 3.3404210303942863,
      "tokens_seen": 2779971584
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.378759220730092e-06,
      "loss": 2.7306,
      "theoretical_loss": 3.3404148816779786,
      "tokens_seen": 2780037120
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.3622091923586155e-06,
      "loss": 2.3392,
      "theoretical_loss": 3.3404087331472025,
      "tokens_seen": 2780102656
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.345659163987138e-06,
      "loss": 2.4623,
      "theoretical_loss": 3.340402584801947,
      "tokens_seen": 2780168192
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.329109135615661e-06,
      "loss": 2.1609,
      "theoretical_loss": 3.3403964366422034,
      "tokens_seen": 2780233728
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.312559107244184e-06,
      "loss": 1.4383,
      "theoretical_loss": 3.340390288667961,
      "tokens_seen": 2780299264
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.296009078872707e-06,
      "loss": 1.9504,
      "theoretical_loss": 3.34038414087921,
      "tokens_seen": 2780364800
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.2794590505012295e-06,
      "loss": 2.4592,
      "theoretical_loss": 3.3403779932759403,
      "tokens_seen": 2780430336
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.262909022129752e-06,
      "loss": 1.6092,
      "theoretical_loss": 3.3403718458581424,
      "tokens_seen": 2780495872
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.246358993758275e-06,
      "loss": 2.6433,
      "theoretical_loss": 3.340365698625806,
      "tokens_seen": 2780561408
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.229808965386798e-06,
      "loss": 2.6279,
      "theoretical_loss": 3.340359551578921,
      "tokens_seen": 2780626944
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.2132589370153204e-06,
      "loss": 1.267,
      "theoretical_loss": 3.3403534047174777,
      "tokens_seen": 2780692480
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.1967089086438436e-06,
      "loss": 2.3218,
      "theoretical_loss": 3.340347258041466,
      "tokens_seen": 2780758016
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.180158880272366e-06,
      "loss": 2.6423,
      "theoretical_loss": 3.3403411115508757,
      "tokens_seen": 2780823552
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.163608851900889e-06,
      "loss": 2.8662,
      "theoretical_loss": 3.3403349652456975,
      "tokens_seen": 2780889088
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.147058823529411e-06,
      "loss": 2.6422,
      "theoretical_loss": 3.340328819125921,
      "tokens_seen": 2780954624
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.1305087951579345e-06,
      "loss": 2.6708,
      "theoretical_loss": 3.340322673191536,
      "tokens_seen": 2781020160
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.113958766786458e-06,
      "loss": 2.185,
      "theoretical_loss": 3.340316527442533,
      "tokens_seen": 2781085696
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.09740873841498e-06,
      "loss": 2.3635,
      "theoretical_loss": 3.340310381878902,
      "tokens_seen": 2781151232
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.080858710043502e-06,
      "loss": 2.722,
      "theoretical_loss": 3.3403042365006326,
      "tokens_seen": 2781216768
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.064308681672025e-06,
      "loss": 2.6962,
      "theoretical_loss": 3.3402980913077154,
      "tokens_seen": 2781282304
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.0477586533005485e-06,
      "loss": 1.7628,
      "theoretical_loss": 3.34029194630014,
      "tokens_seen": 2781347840
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.031208624929072e-06,
      "loss": 2.8472,
      "theoretical_loss": 3.3402858014778967,
      "tokens_seen": 2781413376
    },
    {
      "epoch": 0.99,
      "learning_rate": 5.014658596557594e-06,
      "loss": 1.9679,
      "theoretical_loss": 3.3402796568409756,
      "tokens_seen": 2781478912
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.998108568186117e-06,
      "loss": 1.7974,
      "theoretical_loss": 3.3402735123893663,
      "tokens_seen": 2781544448
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.98155853981464e-06,
      "loss": 2.6161,
      "theoretical_loss": 3.3402673681230595,
      "tokens_seen": 2781609984
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.9650085114431625e-06,
      "loss": 2.6096,
      "theoretical_loss": 3.3402612240420444,
      "tokens_seen": 2781675520
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.948458483071685e-06,
      "loss": 2.6994,
      "theoretical_loss": 3.340255080146312,
      "tokens_seen": 2781741056
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.931908454700208e-06,
      "loss": 2.5854,
      "theoretical_loss": 3.340248936435852,
      "tokens_seen": 2781806592
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.915358426328731e-06,
      "loss": 2.2661,
      "theoretical_loss": 3.340242792910654,
      "tokens_seen": 2781872128
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.898808397957253e-06,
      "loss": 1.9964,
      "theoretical_loss": 3.3402366495707083,
      "tokens_seen": 2781937664
    },
    {
      "epoch": 0.99,
      "objective/train/avg_token_score": 0.010796599090099335,
      "objective/train/avg_weight": 0.9872885942459106,
      "objective/train/docs_used": 1564970,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.532985210418701,
      "objective/train/std_weight": 0.07559353858232498,
      "objective/train/theoretical_loss": 3.340230506416005,
      "objective/train/tokens_used": 2802463200,
      "theoretical_loss": 3.340230506416005,
      "tokens_seen": 2782003200
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.8822583695857765e-06,
      "loss": 2.6873,
      "theoretical_loss": 3.340230506416005,
      "tokens_seen": 2782003200
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.865708341214299e-06,
      "loss": 2.3714,
      "theoretical_loss": 3.3402243634465347,
      "tokens_seen": 2782068736
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.849158312842822e-06,
      "loss": 2.396,
      "theoretical_loss": 3.340218220662287,
      "tokens_seen": 2782134272
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.832608284471344e-06,
      "loss": 2.6454,
      "theoretical_loss": 3.340212078063251,
      "tokens_seen": 2782199808
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.8160582560998674e-06,
      "loss": 2.3394,
      "theoretical_loss": 3.3402059356494185,
      "tokens_seen": 2782265344
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.7995082277283906e-06,
      "loss": 1.075,
      "theoretical_loss": 3.3401997934207786,
      "tokens_seen": 2782330880
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.782958199356914e-06,
      "loss": 1.1484,
      "theoretical_loss": 3.340193651377321,
      "tokens_seen": 2782396416
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.766408170985436e-06,
      "loss": 2.6075,
      "theoretical_loss": 3.340187509519037,
      "tokens_seen": 2782461952
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.749858142613958e-06,
      "loss": 2.5811,
      "theoretical_loss": 3.3401813678459154,
      "tokens_seen": 2782527488
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.7333081142424815e-06,
      "loss": 2.7884,
      "theoretical_loss": 3.3401752263579465,
      "tokens_seen": 2782593024
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.716758085871005e-06,
      "loss": 2.5397,
      "theoretical_loss": 3.3401690850551207,
      "tokens_seen": 2782658560
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.700208057499527e-06,
      "loss": 2.6494,
      "theoretical_loss": 3.340162943937428,
      "tokens_seen": 2782724096
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.68365802912805e-06,
      "loss": 3.4057,
      "theoretical_loss": 3.3401568030048585,
      "tokens_seen": 2782789632
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.667108000756573e-06,
      "loss": 2.3987,
      "theoretical_loss": 3.340150662257402,
      "tokens_seen": 2782855168
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.6505579723850955e-06,
      "loss": 2.6442,
      "theoretical_loss": 3.3401445216950494,
      "tokens_seen": 2782920704
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.634007944013618e-06,
      "loss": 2.584,
      "theoretical_loss": 3.3401383813177894,
      "tokens_seen": 2782986240
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.617457915642141e-06,
      "loss": 1.3658,
      "theoretical_loss": 3.340132241125613,
      "tokens_seen": 2783051776
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.600907887270664e-06,
      "loss": 2.542,
      "theoretical_loss": 3.3401261011185097,
      "tokens_seen": 2783117312
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.584357858899186e-06,
      "loss": 2.3644,
      "theoretical_loss": 3.3401199612964705,
      "tokens_seen": 2783182848
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.5678078305277095e-06,
      "loss": 2.5652,
      "theoretical_loss": 3.340113821659484,
      "tokens_seen": 2783248384
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.551257802156233e-06,
      "loss": 1.9824,
      "theoretical_loss": 3.3401076822075417,
      "tokens_seen": 2783313920
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.534707773784755e-06,
      "loss": 2.6896,
      "theoretical_loss": 3.3401015429406327,
      "tokens_seen": 2783379456
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.518157745413277e-06,
      "loss": 2.6834,
      "theoretical_loss": 3.340095403858748,
      "tokens_seen": 2783444992
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.5016077170418e-06,
      "loss": 1.8722,
      "theoretical_loss": 3.3400892649618767,
      "tokens_seen": 2783510528
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.4850576886703235e-06,
      "loss": 2.5335,
      "theoretical_loss": 3.3400831262500095,
      "tokens_seen": 2783576064
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.468507660298847e-06,
      "loss": 1.9555,
      "theoretical_loss": 3.340076987723136,
      "tokens_seen": 2783641600
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.451957631927369e-06,
      "loss": 2.4816,
      "theoretical_loss": 3.3400708493812465,
      "tokens_seen": 2783707136
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.435407603555891e-06,
      "loss": 3.5722,
      "theoretical_loss": 3.3400647112243314,
      "tokens_seen": 2783772672
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.4188575751844144e-06,
      "loss": 2.3677,
      "theoretical_loss": 3.3400585732523798,
      "tokens_seen": 2783838208
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.4023075468129376e-06,
      "loss": 2.0714,
      "theoretical_loss": 3.340052435465383,
      "tokens_seen": 2783903744
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.38575751844146e-06,
      "loss": 2.6091,
      "theoretical_loss": 3.3400462978633305,
      "tokens_seen": 2783969280
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.369207490069983e-06,
      "loss": 2.807,
      "theoretical_loss": 3.340040160446212,
      "tokens_seen": 2784034816
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.352657461698506e-06,
      "loss": 2.0782,
      "theoretical_loss": 3.3400340232140184,
      "tokens_seen": 2784100352
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.3361074333270285e-06,
      "loss": 2.5291,
      "theoretical_loss": 3.340027886166739,
      "tokens_seen": 2784165888
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.319557404955551e-06,
      "loss": 2.5501,
      "theoretical_loss": 3.3400217493043645,
      "tokens_seen": 2784231424
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.303007376584074e-06,
      "loss": 2.4521,
      "theoretical_loss": 3.3400156126268845,
      "tokens_seen": 2784296960
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.286457348212597e-06,
      "loss": 2.443,
      "theoretical_loss": 3.340009476134289,
      "tokens_seen": 2784362496
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.269907319841119e-06,
      "loss": 1.685,
      "theoretical_loss": 3.340003339826569,
      "tokens_seen": 2784428032
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.2533572914696425e-06,
      "loss": 2.0233,
      "theoretical_loss": 3.3399972037037133,
      "tokens_seen": 2784493568
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.236807263098166e-06,
      "loss": 2.4266,
      "theoretical_loss": 3.3399910677657125,
      "tokens_seen": 2784559104
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.220257234726688e-06,
      "loss": 2.5737,
      "theoretical_loss": 3.339984932012557,
      "tokens_seen": 2784624640
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.203707206355211e-06,
      "loss": 2.9532,
      "theoretical_loss": 3.339978796444237,
      "tokens_seen": 2784690176
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.187157177983733e-06,
      "loss": 2.7243,
      "theoretical_loss": 3.339972661060742,
      "tokens_seen": 2784755712
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.1706071496122565e-06,
      "loss": 2.4285,
      "theoretical_loss": 3.3399665258620623,
      "tokens_seen": 2784821248
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.15405712124078e-06,
      "loss": 1.9188,
      "theoretical_loss": 3.339960390848188,
      "tokens_seen": 2784886784
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.137507092869302e-06,
      "loss": 2.44,
      "theoretical_loss": 3.339954256019109,
      "tokens_seen": 2784952320
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.120957064497825e-06,
      "loss": 1.8795,
      "theoretical_loss": 3.3399481213748157,
      "tokens_seen": 2785017856
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.104407036126347e-06,
      "loss": 2.7488,
      "theoretical_loss": 3.3399419869152984,
      "tokens_seen": 2785083392
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.0878570077548705e-06,
      "loss": 2.8426,
      "theoretical_loss": 3.3399358526405463,
      "tokens_seen": 2785148928
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.071306979383393e-06,
      "loss": 2.2785,
      "theoretical_loss": 3.3399297185505503,
      "tokens_seen": 2785214464
    },
    {
      "epoch": 0.99,
      "objective/train/avg_token_score": 0.025402138009667397,
      "objective/train/avg_weight": 0.970130443572998,
      "objective/train/docs_used": 1566749,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 4.634153366088867,
      "objective/train/std_weight": 0.13682498037815094,
      "objective/train/theoretical_loss": 3.3399235846453004,
      "objective/train/tokens_used": 2805740000,
      "theoretical_loss": 3.3399235846453004,
      "tokens_seen": 2785280000
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.054756951011916e-06,
      "loss": 1.6316,
      "theoretical_loss": 3.3399235846453004,
      "tokens_seen": 2785280000
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.038206922640439e-06,
      "loss": 2.3038,
      "theoretical_loss": 3.339917450924786,
      "tokens_seen": 2785345536
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.0216568942689614e-06,
      "loss": 3.1006,
      "theoretical_loss": 3.3399113173889985,
      "tokens_seen": 2785411072
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.005106865897484e-06,
      "loss": 2.6613,
      "theoretical_loss": 3.3399051840379266,
      "tokens_seen": 2785476608
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.988556837526007e-06,
      "loss": 2.5361,
      "theoretical_loss": 3.339899050871561,
      "tokens_seen": 2785542144
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.97200680915453e-06,
      "loss": 2.7686,
      "theoretical_loss": 3.339892917889892,
      "tokens_seen": 2785607680
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.955456780783053e-06,
      "loss": 2.6089,
      "theoretical_loss": 3.3398867850929093,
      "tokens_seen": 2785673216
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.9389067524115755e-06,
      "loss": 0.9942,
      "theoretical_loss": 3.339880652480603,
      "tokens_seen": 2785738752
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.922356724040099e-06,
      "loss": 2.6929,
      "theoretical_loss": 3.339874520052964,
      "tokens_seen": 2785804288
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.905806695668622e-06,
      "loss": 2.8398,
      "theoretical_loss": 3.339868387809981,
      "tokens_seen": 2785869824
    },
    {
      "epoch": 0.99,
      "learning_rate": 3.889256667297144e-06,
      "loss": 2.4422,
      "theoretical_loss": 3.3398622557516453,
      "tokens_seen": 2785935360
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.872706638925666e-06,
      "loss": 2.5012,
      "theoretical_loss": 3.3398561238779463,
      "tokens_seen": 2786000896
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.8561566105541895e-06,
      "loss": 2.5337,
      "theoretical_loss": 3.3398499921888742,
      "tokens_seen": 2786066432
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.839606582182713e-06,
      "loss": 2.4859,
      "theoretical_loss": 3.33984386068442,
      "tokens_seen": 2786131968
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.823056553811235e-06,
      "loss": 3.0853,
      "theoretical_loss": 3.339837729364572,
      "tokens_seen": 2786197504
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.806506525439758e-06,
      "loss": 0.5227,
      "theoretical_loss": 3.3398315982293223,
      "tokens_seen": 2786263040
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.7899564970682804e-06,
      "loss": 2.5286,
      "theoretical_loss": 3.3398254672786596,
      "tokens_seen": 2786328576
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.773406468696803e-06,
      "loss": 2.3832,
      "theoretical_loss": 3.339819336512574,
      "tokens_seen": 2786394112
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.7568564403253262e-06,
      "loss": 2.5255,
      "theoretical_loss": 3.3398132059310566,
      "tokens_seen": 2786459648
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.740306411953849e-06,
      "loss": 2.8828,
      "theoretical_loss": 3.3398070755340967,
      "tokens_seen": 2786525184
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.723756383582372e-06,
      "loss": 2.5122,
      "theoretical_loss": 3.3398009453216853,
      "tokens_seen": 2786590720
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.707206355210895e-06,
      "loss": 2.4648,
      "theoretical_loss": 3.339794815293811,
      "tokens_seen": 2786656256
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.6906563268394175e-06,
      "loss": 2.4381,
      "theoretical_loss": 3.3397886854504653,
      "tokens_seen": 2786721792
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.67410629846794e-06,
      "loss": 2.3624,
      "theoretical_loss": 3.339782555791637,
      "tokens_seen": 2786787328
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.657556270096463e-06,
      "loss": 2.4327,
      "theoretical_loss": 3.339776426317318,
      "tokens_seen": 2786852864
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.6410062417249857e-06,
      "loss": 2.3216,
      "theoretical_loss": 3.339770297027497,
      "tokens_seen": 2786918400
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.624456213353509e-06,
      "loss": 2.7223,
      "theoretical_loss": 3.3397641679221644,
      "tokens_seen": 2786983936
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.6079061849820316e-06,
      "loss": 2.8176,
      "theoretical_loss": 3.3397580390013104,
      "tokens_seen": 2787049472
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.5913561566105543e-06,
      "loss": -0.554,
      "theoretical_loss": 3.339751910264925,
      "tokens_seen": 2787115008
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.5748061282390766e-06,
      "loss": 2.267,
      "theoretical_loss": 3.3397457817129985,
      "tokens_seen": 2787180544
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.5582560998675997e-06,
      "loss": 2.1762,
      "theoretical_loss": 3.339739653345521,
      "tokens_seen": 2787246080
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.5417060714961225e-06,
      "loss": 2.7787,
      "theoretical_loss": 3.3397335251624822,
      "tokens_seen": 2787311616
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.525156043124645e-06,
      "loss": 2.5223,
      "theoretical_loss": 3.3397273971638732,
      "tokens_seen": 2787377152
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.5086060147531683e-06,
      "loss": 2.5898,
      "theoretical_loss": 3.339721269349683,
      "tokens_seen": 2787442688
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.492055986381691e-06,
      "loss": 2.3371,
      "theoretical_loss": 3.339715141719902,
      "tokens_seen": 2787508224
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.4755059580102133e-06,
      "loss": 2.624,
      "theoretical_loss": 3.339709014274521,
      "tokens_seen": 2787573760
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.4589559296387365e-06,
      "loss": 2.9167,
      "theoretical_loss": 3.3397028870135292,
      "tokens_seen": 2787639296
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.442405901267259e-06,
      "loss": 1.7322,
      "theoretical_loss": 3.3396967599369174,
      "tokens_seen": 2787704832
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.425855872895782e-06,
      "loss": 2.9505,
      "theoretical_loss": 3.3396906330446754,
      "tokens_seen": 2787770368
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.409305844524305e-06,
      "loss": 2.5284,
      "theoretical_loss": 3.339684506336793,
      "tokens_seen": 2787835904
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.392755816152828e-06,
      "loss": 0.5088,
      "theoretical_loss": 3.339678379813261,
      "tokens_seen": 2787901440
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.3762057877813505e-06,
      "loss": 2.6578,
      "theoretical_loss": 3.3396722534740695,
      "tokens_seen": 2787966976
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.359655759409873e-06,
      "loss": 2.0748,
      "theoretical_loss": 3.339666127319208,
      "tokens_seen": 2788032512
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.343105731038396e-06,
      "loss": 2.946,
      "theoretical_loss": 3.339660001348667,
      "tokens_seen": 2788098048
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.3265557026669187e-06,
      "loss": 2.5355,
      "theoretical_loss": 3.3396538755624365,
      "tokens_seen": 2788163584
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.310005674295442e-06,
      "loss": 1.2743,
      "theoretical_loss": 3.339647749960507,
      "tokens_seen": 2788229120
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.2934556459239645e-06,
      "loss": 2.4519,
      "theoretical_loss": 3.3396416245428675,
      "tokens_seen": 2788294656
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.2769056175524873e-06,
      "loss": 2.5062,
      "theoretical_loss": 3.3396354993095096,
      "tokens_seen": 2788360192
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.2603555891810096e-06,
      "loss": 2.2074,
      "theoretical_loss": 3.3396293742604226,
      "tokens_seen": 2788425728
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.2438055608095327e-06,
      "loss": 2.4412,
      "theoretical_loss": 3.339623249395597,
      "tokens_seen": 2788491264
    },
    {
      "epoch": 1.0,
      "objective/train/avg_token_score": 0.01672363467514515,
      "objective/train/avg_weight": 0.9803543090820312,
      "objective/train/docs_used": 1568853,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.733901023864746,
      "objective/train/std_weight": 0.09641505032777786,
      "objective/train/theoretical_loss": 3.3396171247150224,
      "objective/train/tokens_used": 2809016800,
      "theoretical_loss": 3.3396171247150224,
      "tokens_seen": 2788556800
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.2272555324380554e-06,
      "loss": 2.8336,
      "theoretical_loss": 3.3396171247150224,
      "tokens_seen": 2788556800
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.2107055040665786e-06,
      "loss": 2.3305,
      "theoretical_loss": 3.339611000218689,
      "tokens_seen": 2788622336
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.1941554756951013e-06,
      "loss": 0.8462,
      "theoretical_loss": 3.3396048759065877,
      "tokens_seen": 2788687872
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.177605447323624e-06,
      "loss": 2.6561,
      "theoretical_loss": 3.3395987517787082,
      "tokens_seen": 2788753408
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.161055418952147e-06,
      "loss": 2.6401,
      "theoretical_loss": 3.33959262783504,
      "tokens_seen": 2788818944
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.1445053905806695e-06,
      "loss": 2.5768,
      "theoretical_loss": 3.339586504075574,
      "tokens_seen": 2788884480
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.127955362209192e-06,
      "loss": 2.5681,
      "theoretical_loss": 3.3395803805003004,
      "tokens_seen": 2788950016
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.111405333837715e-06,
      "loss": 1.6737,
      "theoretical_loss": 3.3395742571092084,
      "tokens_seen": 2789015552
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.094855305466238e-06,
      "loss": 2.5369,
      "theoretical_loss": 3.339568133902289,
      "tokens_seen": 2789081088
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.0783052770947608e-06,
      "loss": 2.4978,
      "theoretical_loss": 3.3395620108795323,
      "tokens_seen": 2789146624
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.061755248723284e-06,
      "loss": 2.29,
      "theoretical_loss": 3.339555888040928,
      "tokens_seen": 2789212160
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.045205220351806e-06,
      "loss": 2.2556,
      "theoretical_loss": 3.3395497653864665,
      "tokens_seen": 2789277696
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.028655191980329e-06,
      "loss": 2.6272,
      "theoretical_loss": 3.339543642916138,
      "tokens_seen": 2789343232
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.0121051636088516e-06,
      "loss": 2.1037,
      "theoretical_loss": 3.339537520629932,
      "tokens_seen": 2789408768
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.995555135237375e-06,
      "loss": 2.8203,
      "theoretical_loss": 3.33953139852784,
      "tokens_seen": 2789474304
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.9790051068658975e-06,
      "loss": 2.5084,
      "theoretical_loss": 3.3395252766098507,
      "tokens_seen": 2789539840
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.9624550784944202e-06,
      "loss": 2.4428,
      "theoretical_loss": 3.339519154875955,
      "tokens_seen": 2789605376
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.9459050501229434e-06,
      "loss": 2.6627,
      "theoretical_loss": 3.339513033326143,
      "tokens_seen": 2789670912
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.9293550217514657e-06,
      "loss": 2.4175,
      "theoretical_loss": 3.3395069119604046,
      "tokens_seen": 2789736448
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.9128049933799884e-06,
      "loss": -1.2309,
      "theoretical_loss": 3.3395007907787297,
      "tokens_seen": 2789801984
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.8962549650085115e-06,
      "loss": 2.3667,
      "theoretical_loss": 3.3394946697811094,
      "tokens_seen": 2789867520
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.8797049366370343e-06,
      "loss": 0.2873,
      "theoretical_loss": 3.339488548967533,
      "tokens_seen": 2789933056
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.863154908265557e-06,
      "loss": 2.5136,
      "theoretical_loss": 3.339482428337991,
      "tokens_seen": 2789998592
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.84660487989408e-06,
      "loss": 2.1502,
      "theoretical_loss": 3.3394763078924736,
      "tokens_seen": 2790064128
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.8300548515226024e-06,
      "loss": 2.3738,
      "theoretical_loss": 3.3394701876309703,
      "tokens_seen": 2790129664
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.813504823151125e-06,
      "loss": 2.523,
      "theoretical_loss": 3.3394640675534717,
      "tokens_seen": 2790195200
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.796954794779648e-06,
      "loss": 3.8726,
      "theoretical_loss": 3.3394579476599686,
      "tokens_seen": 2790260736
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.780404766408171e-06,
      "loss": 2.571,
      "theoretical_loss": 3.33945182795045,
      "tokens_seen": 2790326272
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.7638547380366937e-06,
      "loss": 2.2607,
      "theoretical_loss": 3.3394457084249067,
      "tokens_seen": 2790391808
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.747304709665217e-06,
      "loss": 2.3582,
      "theoretical_loss": 3.3394395890833284,
      "tokens_seen": 2790457344
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.730754681293739e-06,
      "loss": 1.5899,
      "theoretical_loss": 3.339433469925706,
      "tokens_seen": 2790522880
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.7142046529222623e-06,
      "loss": 1.8935,
      "theoretical_loss": 3.339427350952029,
      "tokens_seen": 2790588416
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.6976546245507846e-06,
      "loss": 2.7819,
      "theoretical_loss": 3.339421232162288,
      "tokens_seen": 2790653952
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.6811045961793078e-06,
      "loss": 2.1644,
      "theoretical_loss": 3.3394151135564725,
      "tokens_seen": 2790719488
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.6645545678078305e-06,
      "loss": 1.9403,
      "theoretical_loss": 3.3394089951345736,
      "tokens_seen": 2790785024
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.6480045394363536e-06,
      "loss": 2.3876,
      "theoretical_loss": 3.3394028768965804,
      "tokens_seen": 2790850560
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.631454511064876e-06,
      "loss": 2.3729,
      "theoretical_loss": 3.3393967588424838,
      "tokens_seen": 2790916096
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.614904482693399e-06,
      "loss": 2.6548,
      "theoretical_loss": 3.3393906409722733,
      "tokens_seen": 2790981632
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.5983544543219218e-06,
      "loss": 2.6677,
      "theoretical_loss": 3.3393845232859403,
      "tokens_seen": 2791047168
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.5818044259504445e-06,
      "loss": 2.4269,
      "theoretical_loss": 3.3393784057834734,
      "tokens_seen": 2791112704
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.5652543975789672e-06,
      "loss": 2.5319,
      "theoretical_loss": 3.339372288464864,
      "tokens_seen": 2791178240
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.54870436920749e-06,
      "loss": 2.4453,
      "theoretical_loss": 3.3393661713301013,
      "tokens_seen": 2791243776
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.5321543408360127e-06,
      "loss": 2.4779,
      "theoretical_loss": 3.3393600543791764,
      "tokens_seen": 2791309312
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.515604312464536e-06,
      "loss": 2.5862,
      "theoretical_loss": 3.3393539376120787,
      "tokens_seen": 2791374848
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.4990542840930585e-06,
      "loss": 2.4537,
      "theoretical_loss": 3.339347821028799,
      "tokens_seen": 2791440384
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.4825042557215813e-06,
      "loss": 0.964,
      "theoretical_loss": 3.3393417046293266,
      "tokens_seen": 2791505920
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.465954227350104e-06,
      "loss": 2.4866,
      "theoretical_loss": 3.339335588413652,
      "tokens_seen": 2791571456
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.4494041989786267e-06,
      "loss": 2.4937,
      "theoretical_loss": 3.339329472381766,
      "tokens_seen": 2791636992
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.4328541706071494e-06,
      "loss": 2.4994,
      "theoretical_loss": 3.339323356533658,
      "tokens_seen": 2791702528
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.416304142235672e-06,
      "loss": 2.629,
      "theoretical_loss": 3.3393172408693186,
      "tokens_seen": 2791768064
    },
    {
      "epoch": 1.0,
      "objective/train/avg_token_score": 0.015544552356004715,
      "objective/train/avg_weight": 0.9817254543304443,
      "objective/train/docs_used": 1571029,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.7813515663146973,
      "objective/train/std_weight": 0.08305560052394867,
      "objective/train/theoretical_loss": 3.3393111253887375,
      "objective/train/tokens_used": 2812293600,
      "theoretical_loss": 3.3393111253887375,
      "tokens_seen": 2791833600
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.3997541138641953e-06,
      "loss": 2.8002,
      "theoretical_loss": 3.3393111253887375,
      "tokens_seen": 2791833600
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.383204085492718e-06,
      "loss": 1.4599,
      "theoretical_loss": 3.3393050100919055,
      "tokens_seen": 2791899136
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.3666540571212407e-06,
      "loss": 2.4183,
      "theoretical_loss": 3.3392988949788123,
      "tokens_seen": 2791964672
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.3501040287497634e-06,
      "loss": 2.3815,
      "theoretical_loss": 3.3392927800494485,
      "tokens_seen": 2792030208
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.3335540003782866e-06,
      "loss": 0.2252,
      "theoretical_loss": 3.3392866653038036,
      "tokens_seen": 2792095744
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.317003972006809e-06,
      "loss": 2.5032,
      "theoretical_loss": 3.339280550741868,
      "tokens_seen": 2792161280
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.300453943635332e-06,
      "loss": 3.1875,
      "theoretical_loss": 3.3392744363636324,
      "tokens_seen": 2792226816
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.2839039152638548e-06,
      "loss": 2.3858,
      "theoretical_loss": 3.3392683221690866,
      "tokens_seen": 2792292352
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.2673538868923775e-06,
      "loss": 2.6608,
      "theoretical_loss": 3.33926220815822,
      "tokens_seen": 2792357888
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.2508038585209e-06,
      "loss": 2.5628,
      "theoretical_loss": 3.3392560943310245,
      "tokens_seen": 2792423424
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.2342538301494233e-06,
      "loss": 2.5151,
      "theoretical_loss": 3.339249980687489,
      "tokens_seen": 2792488960
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.2177038017779456e-06,
      "loss": 2.4222,
      "theoretical_loss": 3.3392438672276037,
      "tokens_seen": 2792554496
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.2011537734064688e-06,
      "loss": 2.8552,
      "theoretical_loss": 3.339237753951359,
      "tokens_seen": 2792620032
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.1846037450349915e-06,
      "loss": 2.415,
      "theoretical_loss": 3.3392316408587455,
      "tokens_seen": 2792685568
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.1680537166635142e-06,
      "loss": 3.0265,
      "theoretical_loss": 3.3392255279497527,
      "tokens_seen": 2792751104
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.151503688292037e-06,
      "loss": 0.9319,
      "theoretical_loss": 3.3392194152243713,
      "tokens_seen": 2792816640
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.1349536599205597e-06,
      "loss": 2.687,
      "theoretical_loss": 3.339213302682591,
      "tokens_seen": 2792882176
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.118403631549083e-06,
      "loss": 2.1592,
      "theoretical_loss": 3.339207190324402,
      "tokens_seen": 2792947712
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.1018536031776055e-06,
      "loss": 2.3568,
      "theoretical_loss": 3.339201078149795,
      "tokens_seen": 2793013248
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.0853035748061283e-06,
      "loss": 1.8671,
      "theoretical_loss": 3.33919496615876,
      "tokens_seen": 2793078784
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.068753546434651e-06,
      "loss": 1.3255,
      "theoretical_loss": 3.3391888543512867,
      "tokens_seen": 2793144320
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.0522035180631737e-06,
      "loss": 2.1603,
      "theoretical_loss": 3.339182742727366,
      "tokens_seen": 2793209856
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.0356534896916964e-06,
      "loss": 2.413,
      "theoretical_loss": 3.3391766312869873,
      "tokens_seen": 2793275392
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.0191034613202196e-06,
      "loss": 2.3366,
      "theoretical_loss": 3.339170520030142,
      "tokens_seen": 2793340928
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.002553432948742e-06,
      "loss": 2.4746,
      "theoretical_loss": 3.339164408956819,
      "tokens_seen": 2793406464
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.986003404577265e-06,
      "loss": 2.5416,
      "theoretical_loss": 3.3391582980670087,
      "tokens_seen": 2793472000
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.9694533762057877e-06,
      "loss": 1.7487,
      "theoretical_loss": 3.3391521873607015,
      "tokens_seen": 2793537536
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.952903347834311e-06,
      "loss": 2.6778,
      "theoretical_loss": 3.3391460768378876,
      "tokens_seen": 2793603072
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.936353319462833e-06,
      "loss": 2.2017,
      "theoretical_loss": 3.3391399664985575,
      "tokens_seen": 2793668608
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.9198032910913563e-06,
      "loss": 0.3848,
      "theoretical_loss": 3.339133856342701,
      "tokens_seen": 2793734144
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.903253262719879e-06,
      "loss": 3.0009,
      "theoretical_loss": 3.3391277463703086,
      "tokens_seen": 2793799680
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.8867032343484015e-06,
      "loss": 2.5696,
      "theoretical_loss": 3.3391216365813703,
      "tokens_seen": 2793865216
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.8701532059769245e-06,
      "loss": 1.8777,
      "theoretical_loss": 3.339115526975876,
      "tokens_seen": 2793930752
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.8536031776054474e-06,
      "loss": 1.447,
      "theoretical_loss": 3.339109417553816,
      "tokens_seen": 2793996288
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.83705314923397e-06,
      "loss": 2.6507,
      "theoretical_loss": 3.339103308315181,
      "tokens_seen": 2794061824
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.8205031208624929e-06,
      "loss": 2.6846,
      "theoretical_loss": 3.3390971992599607,
      "tokens_seen": 2794127360
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.8039530924910158e-06,
      "loss": 0.6843,
      "theoretical_loss": 3.3390910903881457,
      "tokens_seen": 2794192896
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.7874030641195383e-06,
      "loss": 2.8165,
      "theoretical_loss": 3.3390849816997257,
      "tokens_seen": 2794258432
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.7708530357480612e-06,
      "loss": 2.8822,
      "theoretical_loss": 3.339078873194691,
      "tokens_seen": 2794323968
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.7543030073765842e-06,
      "loss": 2.0655,
      "theoretical_loss": 3.339072764873032,
      "tokens_seen": 2794389504
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.7377529790051067e-06,
      "loss": 2.5111,
      "theoretical_loss": 3.3390666567347385,
      "tokens_seen": 2794455040
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.7212029506336296e-06,
      "loss": 2.6082,
      "theoretical_loss": 3.3390605487798015,
      "tokens_seen": 2794520576
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.7046529222621525e-06,
      "loss": 2.5652,
      "theoretical_loss": 3.3390544410082104,
      "tokens_seen": 2794586112
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.6881028938906753e-06,
      "loss": 2.7349,
      "theoretical_loss": 3.3390483334199557,
      "tokens_seen": 2794651648
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.671552865519198e-06,
      "loss": 2.1497,
      "theoretical_loss": 3.3390422260150276,
      "tokens_seen": 2794717184
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.655002837147721e-06,
      "loss": 2.4901,
      "theoretical_loss": 3.3390361187934166,
      "tokens_seen": 2794782720
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.6384528087762436e-06,
      "loss": 1.8033,
      "theoretical_loss": 3.339030011755112,
      "tokens_seen": 2794848256
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.6219027804047664e-06,
      "loss": 2.1186,
      "theoretical_loss": 3.339023904900105,
      "tokens_seen": 2794913792
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.6053527520332893e-06,
      "loss": 2.6714,
      "theoretical_loss": 3.3390177982283853,
      "tokens_seen": 2794979328
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.588802723661812e-06,
      "loss": 2.5352,
      "theoretical_loss": 3.339011691739943,
      "tokens_seen": 2795044864
    },
    {
      "epoch": 1.0,
      "objective/train/avg_token_score": 0.017288485541939735,
      "objective/train/avg_weight": 0.9797190427780151,
      "objective/train/docs_used": 1572728,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 3.2091853618621826,
      "objective/train/std_weight": 0.09506263583898544,
      "objective/train/theoretical_loss": 3.339005585434769,
      "objective/train/tokens_used": 2815570400,
      "theoretical_loss": 3.339005585434769,
      "tokens_seen": 2795110400
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.5722526952903347e-06,
      "loss": 2.1881,
      "theoretical_loss": 3.339005585434769,
      "tokens_seen": 2795110400
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.5557026669188574e-06,
      "loss": 2.4197,
      "theoretical_loss": 3.3389994793128523,
      "tokens_seen": 2795175936
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.5391526385473804e-06,
      "loss": 1.7935,
      "theoretical_loss": 3.3389933733741843,
      "tokens_seen": 2795241472
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.522602610175903e-06,
      "loss": 2.5046,
      "theoretical_loss": 3.3389872676187546,
      "tokens_seen": 2795307008
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.5060525818044258e-06,
      "loss": 2.4008,
      "theoretical_loss": 3.3389811620465535,
      "tokens_seen": 2795372544
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.4895025534329488e-06,
      "loss": 2.7848,
      "theoretical_loss": 3.338975056657571,
      "tokens_seen": 2795438080
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.4729525250614717e-06,
      "loss": 2.6792,
      "theoretical_loss": 3.3389689514517977,
      "tokens_seen": 2795503616
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.4564024966899942e-06,
      "loss": 1.8556,
      "theoretical_loss": 3.3389628464292236,
      "tokens_seen": 2795569152
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.4398524683185171e-06,
      "loss": 0.5813,
      "theoretical_loss": 3.338956741589839,
      "tokens_seen": 2795634688
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.42330243994704e-06,
      "loss": 1.5245,
      "theoretical_loss": 3.338950636933634,
      "tokens_seen": 2795700224
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.4067524115755626e-06,
      "loss": 2.2869,
      "theoretical_loss": 3.338944532460599,
      "tokens_seen": 2795765760
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.3902023832040855e-06,
      "loss": 1.596,
      "theoretical_loss": 3.3389384281707235,
      "tokens_seen": 2795831296
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.3736523548326084e-06,
      "loss": 2.4606,
      "theoretical_loss": 3.338932324063999,
      "tokens_seen": 2795896832
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.3571023264611312e-06,
      "loss": 2.0129,
      "theoretical_loss": 3.338926220140414,
      "tokens_seen": 2795962368
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.3405522980896539e-06,
      "loss": 1.2284,
      "theoretical_loss": 3.3389201163999607,
      "tokens_seen": 2796027904
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.3240022697181768e-06,
      "loss": 2.6307,
      "theoretical_loss": 3.3389140128426282,
      "tokens_seen": 2796093440
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.3074522413466995e-06,
      "loss": 2.4568,
      "theoretical_loss": 3.338907909468406,
      "tokens_seen": 2796158976
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.2909022129752223e-06,
      "loss": 1.704,
      "theoretical_loss": 3.338901806277286,
      "tokens_seen": 2796224512
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.274352184603745e-06,
      "loss": 2.4631,
      "theoretical_loss": 3.3388957032692574,
      "tokens_seen": 2796290048
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.257802156232268e-06,
      "loss": 2.8175,
      "theoretical_loss": 3.3388896004443107,
      "tokens_seen": 2796355584
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.2412521278607906e-06,
      "loss": 1.6675,
      "theoretical_loss": 3.3388834978024358,
      "tokens_seen": 2796421120
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.2247020994893133e-06,
      "loss": 2.1189,
      "theoretical_loss": 3.338877395343623,
      "tokens_seen": 2796486656
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.208152071117836e-06,
      "loss": 2.3404,
      "theoretical_loss": 3.3388712930678626,
      "tokens_seen": 2796552192
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.191602042746359e-06,
      "loss": 2.483,
      "theoretical_loss": 3.338865190975145,
      "tokens_seen": 2796617728
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.1750520143748817e-06,
      "loss": 2.6331,
      "theoretical_loss": 3.33885908906546,
      "tokens_seen": 2796683264
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.1585019860034044e-06,
      "loss": 2.3372,
      "theoretical_loss": 3.3388529873387984,
      "tokens_seen": 2796748800
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.1419519576319274e-06,
      "loss": 2.3062,
      "theoretical_loss": 3.33884688579515,
      "tokens_seen": 2796814336
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.12540192926045e-06,
      "loss": 2.1422,
      "theoretical_loss": 3.338840784434505,
      "tokens_seen": 2796879872
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.1088519008889728e-06,
      "loss": 2.2693,
      "theoretical_loss": 3.338834683256854,
      "tokens_seen": 2796945408
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.0923018725174958e-06,
      "loss": 0.2125,
      "theoretical_loss": 3.338828582262187,
      "tokens_seen": 2797010944
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.0757518441460185e-06,
      "loss": 2.1086,
      "theoretical_loss": 3.338822481450494,
      "tokens_seen": 2797076480
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.0592018157745414e-06,
      "loss": 2.7088,
      "theoretical_loss": 3.3388163808217657,
      "tokens_seen": 2797142016
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.0426517874030641e-06,
      "loss": 2.4809,
      "theoretical_loss": 3.3388102803759914,
      "tokens_seen": 2797207552
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.0261017590315868e-06,
      "loss": 2.5976,
      "theoretical_loss": 3.3388041801131627,
      "tokens_seen": 2797273088
    },
    {
      "epoch": 1.0,
      "learning_rate": 1.0095517306601098e-06,
      "loss": 1.3558,
      "theoretical_loss": 3.3387980800332686,
      "tokens_seen": 2797338624
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.930017022886325e-07,
      "loss": 2.7252,
      "theoretical_loss": 3.3387919801363,
      "tokens_seen": 2797404160
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.764516739171554e-07,
      "loss": 2.7252,
      "theoretical_loss": 3.338785880422247,
      "tokens_seen": 2797469696
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.599016455456782e-07,
      "loss": 2.6096,
      "theoretical_loss": 3.3387797808911,
      "tokens_seen": 2797535232
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.433516171742008e-07,
      "loss": 2.2802,
      "theoretical_loss": 3.338773681542849,
      "tokens_seen": 2797600768
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.268015888027237e-07,
      "loss": 2.4027,
      "theoretical_loss": 3.3387675823774843,
      "tokens_seen": 2797666304
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.102515604312464e-07,
      "loss": 2.7425,
      "theoretical_loss": 3.338761483394996,
      "tokens_seen": 2797731840
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.937015320597691e-07,
      "loss": 1.876,
      "theoretical_loss": 3.338755384595374,
      "tokens_seen": 2797797376
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.771515036882921e-07,
      "loss": 3.183,
      "theoretical_loss": 3.3387492859786096,
      "tokens_seen": 2797862912
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.606014753168148e-07,
      "loss": 1.9668,
      "theoretical_loss": 3.338743187544692,
      "tokens_seen": 2797928448
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.440514469453376e-07,
      "loss": 2.6136,
      "theoretical_loss": 3.338737089293612,
      "tokens_seen": 2797993984
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.275014185738605e-07,
      "loss": 2.5782,
      "theoretical_loss": 3.3387309912253595,
      "tokens_seen": 2798059520
    },
    {
      "epoch": 1.0,
      "learning_rate": 8.109513902023832e-07,
      "loss": 2.7897,
      "theoretical_loss": 3.338724893339925,
      "tokens_seen": 2798125056
    },
    {
      "epoch": 1.0,
      "learning_rate": 7.94401361830906e-07,
      "loss": 2.5053,
      "theoretical_loss": 3.338718795637299,
      "tokens_seen": 2798190592
    },
    {
      "epoch": 1.0,
      "learning_rate": 7.778513334594287e-07,
      "loss": 2.4287,
      "theoretical_loss": 3.3387126981174706,
      "tokens_seen": 2798256128
    },
    {
      "epoch": 1.0,
      "learning_rate": 7.613013050879516e-07,
      "loss": 2.5573,
      "theoretical_loss": 3.338706600780432,
      "tokens_seen": 2798321664
    },
    {
      "epoch": 1.0,
      "objective/train/avg_token_score": 0.016539547592401505,
      "objective/train/avg_weight": 0.9805448651313782,
      "objective/train/docs_used": 1574561,
      "objective/train/instantaneous_batch_size": 32,
      "objective/train/instantaneous_microbatch_size": 32768,
      "objective/train/original_loss": 2.2078425884246826,
      "objective/train/std_weight": 0.07985614240169525,
      "objective/train/theoretical_loss": 3.338700503626171,
      "objective/train/tokens_used": 2818847200,
      "theoretical_loss": 3.338700503626171,
      "tokens_seen": 2798387200
    },
    {
      "epoch": 1.0,
      "learning_rate": 7.447512767164744e-07,
      "loss": 2.3821,
      "theoretical_loss": 3.338700503626171,
      "tokens_seen": 2798387200
    },
    {
      "epoch": 1.0,
      "learning_rate": 7.282012483449971e-07,
      "loss": 1.3599,
      "theoretical_loss": 3.3386944066546795,
      "tokens_seen": 2798452736
    },
    {
      "epoch": 1.0,
      "learning_rate": 7.1165121997352e-07,
      "loss": 1.4062,
      "theoretical_loss": 3.3386883098659474,
      "tokens_seen": 2798518272
    },
    {
      "epoch": 1.0,
      "learning_rate": 6.951011916020428e-07,
      "loss": 2.2436,
      "theoretical_loss": 3.338682213259965,
      "tokens_seen": 2798583808
    },
    {
      "epoch": 1.0,
      "learning_rate": 6.785511632305656e-07,
      "loss": 2.2385,
      "theoretical_loss": 3.3386761168367225,
      "tokens_seen": 2798649344
    },
    {
      "epoch": 1.0,
      "learning_rate": 6.620011348590884e-07,
      "loss": 2.1161,
      "theoretical_loss": 3.33867002059621,
      "tokens_seen": 2798714880
    },
    {
      "epoch": 1.0,
      "learning_rate": 6.454511064876111e-07,
      "loss": 3.5789,
      "theoretical_loss": 3.3386639245384178,
      "tokens_seen": 2798780416
    },
    {
      "epoch": 1.0,
      "learning_rate": 6.28901078116134e-07,
      "loss": 2.5555,
      "theoretical_loss": 3.3386578286633357,
      "tokens_seen": 2798845952
    },
    {
      "epoch": 1.0,
      "learning_rate": 6.123510497446567e-07,
      "loss": 2.2576,
      "theoretical_loss": 3.338651732970955,
      "tokens_seen": 2798911488
    },
    {
      "epoch": 1.0,
      "learning_rate": 5.958010213731795e-07,
      "loss": 3.0204,
      "theoretical_loss": 3.338645637461265,
      "tokens_seen": 2798977024
    },
    {
      "epoch": 1.0,
      "learning_rate": 5.792509930017022e-07,
      "loss": 1.4826,
      "theoretical_loss": 3.3386395421342567,
      "tokens_seen": 2799042560
    },
    {
      "epoch": 1.0,
      "learning_rate": 5.62700964630225e-07,
      "loss": 2.5572,
      "theoretical_loss": 3.3386334469899195,
      "tokens_seen": 2799108096
    },
    {
      "epoch": 1.0,
      "learning_rate": 5.461509362587479e-07,
      "loss": 2.5402,
      "theoretical_loss": 3.3386273520282446,
      "tokens_seen": 2799173632
    },
    {
      "epoch": 1.0,
      "learning_rate": 5.296009078872707e-07,
      "loss": 2.6131,
      "theoretical_loss": 3.3386212572492213,
      "tokens_seen": 2799239168
    },
    {
      "epoch": 1.0,
      "learning_rate": 5.130508795157934e-07,
      "loss": 2.47,
      "theoretical_loss": 3.3386151626528404,
      "tokens_seen": 2799304704
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.965008511443163e-07,
      "loss": 3.0364,
      "theoretical_loss": 3.338609068239092,
      "tokens_seen": 2799370240
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.799508227728391e-07,
      "loss": 2.4016,
      "theoretical_loss": 3.3386029740079666,
      "tokens_seen": 2799435776
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.6340079440136185e-07,
      "loss": 1.9436,
      "theoretical_loss": 3.338596879959454,
      "tokens_seen": 2799501312
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.4685076602988457e-07,
      "loss": 3.5308,
      "theoretical_loss": 3.338590786093545,
      "tokens_seen": 2799566848
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.303007376584074e-07,
      "loss": 2.1503,
      "theoretical_loss": 3.338584692410229,
      "tokens_seen": 2799632384
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.1375070928693023e-07,
      "loss": 2.9983,
      "theoretical_loss": 3.338578598909497,
      "tokens_seen": 2799697920
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.97200680915453e-07,
      "loss": 2.3686,
      "theoretical_loss": 3.3385725055913396,
      "tokens_seen": 2799763456
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.806506525439758e-07,
      "loss": 2.0552,
      "theoretical_loss": 3.338566412455746,
      "tokens_seen": 2799828992
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.6410062417249855e-07,
      "loss": -1.3429,
      "theoretical_loss": 3.338560319502707,
      "tokens_seen": 2799894528
    }
  ],
  "max_steps": 42724,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 1.428886973841408e+18,
  "trial_name": null,
  "trial_params": null
}